Pull request #4648 updated
Connecting to https://api.github.com using dgl-bot-personal-access-key220906
Connecting to https://api.github.com to check permissions of obtain list of yaox12 for dmlc/dgl
Obtained Jenkinsfile from 37ed78035f406940ed88a4b60cadd2b7cfde5fea+d78a3a4baf611b90871a849f58647160c7cd9ab4 (2a6b7c3d5b495238d0d9cc8640b20ee0facfdf2f)
[Pipeline] Start of Pipeline
[Pipeline] node
Running on dgl-manual-cpu-worker in /home/ubuntu/jenkins_new/workspace/dgl_PR-4648
[Pipeline] {
[Pipeline] stage
[Pipeline] { (Declarative: Checkout SCM)
[Pipeline] checkout
The recommended git tool is: git
using credential 150de63f-189c-4717-bcaf-010460d2f51a
Fetching changes from the remote Git repository
Cleaning workspace
Fetching without tags
Merging remotes/origin/master commit d78a3a4baf611b90871a849f58647160c7cd9ab4 into PR head commit 37ed78035f406940ed88a4b60cadd2b7cfde5fea
Merge succeeded, producing 37ed78035f406940ed88a4b60cadd2b7cfde5fea
Checking out Revision 37ed78035f406940ed88a4b60cadd2b7cfde5fea (PR-4648)
Commit message: "fix for pytorch < 1.12"
 > git rev-parse --resolve-git-dir /home/ubuntu/jenkins_new/workspace/dgl_PR-4648/.git # timeout=10
 > git config remote.origin.url https://github.com/dmlc/dgl.git # timeout=10
 > git rev-parse --verify HEAD # timeout=10
Resetting working tree
 > git reset --hard # timeout=10
 > git clean -fdx # timeout=10
Fetching upstream changes from https://github.com/dmlc/dgl.git
 > git --version # timeout=10
 > git --version # 'git version 2.17.1'
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git fetch --no-tags --progress -- https://github.com/dmlc/dgl.git +refs/pull/4648/head:refs/remotes/origin/PR-4648 +refs/heads/master:refs/remotes/origin/master # timeout=10
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git remote # timeout=10
 > git config --get remote.origin.url # timeout=10
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git merge d78a3a4baf611b90871a849f58647160c7cd9ab4 # timeout=10
 > git rev-parse HEAD^{commit} # timeout=10
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git branch -a -v --no-abbrev # timeout=10
 > git branch -D PR-4648 # timeout=10
 > git checkout -b PR-4648 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git rev-list --no-walk 213b27ce5e5aec17b70af3c5ce56e3d7abfbda0f # timeout=10
Cleaning workspace
 > git rev-parse --verify HEAD # timeout=10
Resetting working tree
 > git reset --hard # timeout=10
 > git clean -fdx # timeout=10
[Pipeline] }
[Pipeline] // stage
[Pipeline] withEnv
[Pipeline] {
[Pipeline] stage
[Pipeline] { (Authentication)
[Pipeline] node
Running on dgci-worker-permanent in /home/ubuntu/jenkins_benchmark/workspace/dgl_PR-4648
[Pipeline] {
[Pipeline] checkout
The recommended git tool is: git
using credential 150de63f-189c-4717-bcaf-010460d2f51a
Fetching changes from the remote Git repository
Cleaning workspace
Fetching without tags
 > git rev-parse --resolve-git-dir /home/ubuntu/jenkins_benchmark/workspace/dgl_PR-4648/.git # timeout=10
 > git config remote.origin.url https://github.com/dmlc/dgl.git # timeout=10
 > git rev-parse --verify HEAD # timeout=10
Resetting working tree
 > git reset --hard # timeout=10
 > git clean -fdx # timeout=10
Fetching upstream changes from https://github.com/dmlc/dgl.git
 > git --version # timeout=10
 > git --version # 'git version 2.25.1'
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git fetch --no-tags --force --progress -- https://github.com/dmlc/dgl.git +refs/pull/4648/head:refs/remotes/origin/PR-4648 +refs/heads/master:refs/remotes/origin/master # timeout=10
Merging remotes/origin/master commit d78a3a4baf611b90871a849f58647160c7cd9ab4 into PR head commit 37ed78035f406940ed88a4b60cadd2b7cfde5fea
Merge succeeded, producing 37ed78035f406940ed88a4b60cadd2b7cfde5fea
Checking out Revision 37ed78035f406940ed88a4b60cadd2b7cfde5fea (PR-4648)
Commit message: "fix for pytorch < 1.12"
Cleaning workspace
[Pipeline] withEnv
[Pipeline] {
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git remote # timeout=10
 > git config --get remote.origin.url # timeout=10
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git merge d78a3a4baf611b90871a849f58647160c7cd9ab4 # timeout=10
 > git rev-parse HEAD^{commit} # timeout=10
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git branch -a -v --no-abbrev # timeout=10
 > git branch -D PR-4648 # timeout=10
 > git checkout -b PR-4648 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git rev-parse --verify HEAD # timeout=10
Resetting working tree
 > git reset --hard # timeout=10
 > git clean -fdx # timeout=10
[Pipeline] isUnix
[Pipeline] withEnv
[Pipeline] {
[Pipeline] sh
+ docker pull dgllib/dgl-ci-lint
Using default tag: latest
latest: Pulling from dgllib/dgl-ci-lint
Digest: sha256:11b87b4425630143a2560fe6ad584f245ee0a86bc58b40cda3c2be93656b0d5e
Status: Image is up to date for dgllib/dgl-ci-lint:latest
docker.io/dgllib/dgl-ci-lint:latest
[Pipeline] }
[Pipeline] // withEnv
[Pipeline] isUnix
[Pipeline] withEnv
[Pipeline] {
[Pipeline] sh
+ docker inspect -f . dgllib/dgl-ci-lint
.
[Pipeline] }
[Pipeline] // withEnv
[Pipeline] withDockerContainer
dgci-worker-permanent does not seem to be running inside a container
$ docker run -t -d -u 0:0 -w /home/ubuntu/jenkins_benchmark/workspace/dgl_PR-4648 -v /home/ubuntu/jenkins_benchmark/workspace/dgl_PR-4648:/home/ubuntu/jenkins_benchmark/workspace/dgl_PR-4648:rw,z -v /home/ubuntu/jenkins_benchmark/workspace/dgl_PR-4648@tmp:/home/ubuntu/jenkins_benchmark/workspace/dgl_PR-4648@tmp:rw,z -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** dgllib/dgl-ci-lint cat
$ docker top 415d53530fec5aed6c0040e81b3796a94de3e7f8cb1dcc4d15e5f1518126325d -eo pid,comm
[Pipeline] {
[Pipeline] script
[Pipeline] {
[Pipeline] }
[Pipeline] // script
[Pipeline] }
$ docker stop --time=1 415d53530fec5aed6c0040e81b3796a94de3e7f8cb1dcc4d15e5f1518126325d
$ docker rm -f 415d53530fec5aed6c0040e81b3796a94de3e7f8cb1dcc4d15e5f1518126325d
[Pipeline] // withDockerContainer
[Pipeline] }
[Pipeline] // withEnv
[Pipeline] }
[Pipeline] // node
[Pipeline] }
[Pipeline] // stage
[Pipeline] stage
[Pipeline] { (AuthenticationComment)
[Pipeline] node
Running on dgci-worker-permanent in /home/ubuntu/jenkins_benchmark/workspace/dgl_PR-4648
[Pipeline] {
[Pipeline] checkout
The recommended git tool is: git
using credential 150de63f-189c-4717-bcaf-010460d2f51a
Fetching changes from the remote Git repository
Cleaning workspace
Fetching without tags
Merging remotes/origin/master commit d78a3a4baf611b90871a849f58647160c7cd9ab4 into PR head commit 37ed78035f406940ed88a4b60cadd2b7cfde5fea
Merge succeeded, producing 37ed78035f406940ed88a4b60cadd2b7cfde5fea
Checking out Revision 37ed78035f406940ed88a4b60cadd2b7cfde5fea (PR-4648)
Commit message: "fix for pytorch < 1.12"
Cleaning workspace
[Pipeline] withEnv
[Pipeline] {
[Pipeline] isUnix
[Pipeline] withEnv
[Pipeline] {
[Pipeline] sh
 > git rev-parse --resolve-git-dir /home/ubuntu/jenkins_benchmark/workspace/dgl_PR-4648/.git # timeout=10
 > git config remote.origin.url https://github.com/dmlc/dgl.git # timeout=10
 > git rev-parse --verify HEAD # timeout=10
Resetting working tree
 > git reset --hard # timeout=10
 > git clean -fdx # timeout=10
Fetching upstream changes from https://github.com/dmlc/dgl.git
 > git --version # timeout=10
 > git --version # 'git version 2.25.1'
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git fetch --no-tags --force --progress -- https://github.com/dmlc/dgl.git +refs/pull/4648/head:refs/remotes/origin/PR-4648 +refs/heads/master:refs/remotes/origin/master # timeout=10
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git remote # timeout=10
 > git config --get remote.origin.url # timeout=10
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git merge d78a3a4baf611b90871a849f58647160c7cd9ab4 # timeout=10
 > git rev-parse HEAD^{commit} # timeout=10
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git branch -a -v --no-abbrev # timeout=10
 > git branch -D PR-4648 # timeout=10
 > git checkout -b PR-4648 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git rev-parse --verify HEAD # timeout=10
Resetting working tree
 > git reset --hard # timeout=10
 > git clean -fdx # timeout=10
+ docker pull dgllib/dgl-ci-lint
Using default tag: latest
latest: Pulling from dgllib/dgl-ci-lint
Digest: sha256:11b87b4425630143a2560fe6ad584f245ee0a86bc58b40cda3c2be93656b0d5e
Status: Image is up to date for dgllib/dgl-ci-lint:latest
docker.io/dgllib/dgl-ci-lint:latest
[Pipeline] }
[Pipeline] // withEnv
[Pipeline] isUnix
[Pipeline] withEnv
[Pipeline] {
[Pipeline] sh
+ docker inspect -f . dgllib/dgl-ci-lint
.
[Pipeline] }
[Pipeline] // withEnv
[Pipeline] withDockerContainer
dgci-worker-permanent does not seem to be running inside a container
$ docker run -t -d -u 0:0 -w /home/ubuntu/jenkins_benchmark/workspace/dgl_PR-4648 -v /home/ubuntu/jenkins_benchmark/workspace/dgl_PR-4648:/home/ubuntu/jenkins_benchmark/workspace/dgl_PR-4648:rw,z -v /home/ubuntu/jenkins_benchmark/workspace/dgl_PR-4648@tmp:/home/ubuntu/jenkins_benchmark/workspace/dgl_PR-4648@tmp:rw,z -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** dgllib/dgl-ci-lint cat
$ docker top b4f3029846f5165bf12c85fd7be505bbf78db783074ae1a07b14b774d1eb3855 -eo pid,comm
[Pipeline] {
[Pipeline] }
$ docker stop --time=1 b4f3029846f5165bf12c85fd7be505bbf78db783074ae1a07b14b774d1eb3855
$ docker rm -f b4f3029846f5165bf12c85fd7be505bbf78db783074ae1a07b14b774d1eb3855
[Pipeline] // withDockerContainer
[Pipeline] }
[Pipeline] // withEnv
[Pipeline] }
[Pipeline] // node
Stage "AuthenticationComment" skipped due to when conditional
[Pipeline] }
[Pipeline] // stage
[Pipeline] stage
[Pipeline] { (Regression Test)
[Pipeline] node
Running on dgci-worker-permanent in /home/ubuntu/jenkins_benchmark/workspace/dgl_PR-4648
[Pipeline] {
[Pipeline] checkout
The recommended git tool is: git
using credential 150de63f-189c-4717-bcaf-010460d2f51a
Fetching changes from the remote Git repository
Cleaning workspace
Fetching without tags
Merging remotes/origin/master commit d78a3a4baf611b90871a849f58647160c7cd9ab4 into PR head commit 37ed78035f406940ed88a4b60cadd2b7cfde5fea
Merge succeeded, producing 37ed78035f406940ed88a4b60cadd2b7cfde5fea
Checking out Revision 37ed78035f406940ed88a4b60cadd2b7cfde5fea (PR-4648)
Commit message: "fix for pytorch < 1.12"
Cleaning workspace
 > git rev-parse --resolve-git-dir /home/ubuntu/jenkins_benchmark/workspace/dgl_PR-4648/.git # timeout=10
 > git config remote.origin.url https://github.com/dmlc/dgl.git # timeout=10
 > git rev-parse --verify HEAD # timeout=10
Resetting working tree
 > git reset --hard # timeout=10
 > git clean -fdx # timeout=10
Fetching upstream changes from https://github.com/dmlc/dgl.git
 > git --version # timeout=10
 > git --version # 'git version 2.25.1'
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git fetch --no-tags --force --progress -- https://github.com/dmlc/dgl.git +refs/pull/4648/head:refs/remotes/origin/PR-4648 +refs/heads/master:refs/remotes/origin/master # timeout=10
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git remote # timeout=10
 > git config --get remote.origin.url # timeout=10
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git merge d78a3a4baf611b90871a849f58647160c7cd9ab4 # timeout=10
 > git rev-parse HEAD^{commit} # timeout=10
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git branch -a -v --no-abbrev # timeout=10
 > git branch -D PR-4648 # timeout=10
 > git checkout -b PR-4648 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git rev-parse --verify HEAD # timeout=10
Resetting working tree
 > git reset --hard # timeout=10
 > git clean -fdx # timeout=10
[Pipeline] withEnv
[Pipeline] {
[Pipeline] isUnix
[Pipeline] withEnv
[Pipeline] {
[Pipeline] sh
+ docker pull dgllib/dgl-ci-lint
Using default tag: latest
latest: Pulling from dgllib/dgl-ci-lint
Digest: sha256:11b87b4425630143a2560fe6ad584f245ee0a86bc58b40cda3c2be93656b0d5e
Status: Image is up to date for dgllib/dgl-ci-lint:latest
docker.io/dgllib/dgl-ci-lint:latest
[Pipeline] }
[Pipeline] // withEnv
[Pipeline] isUnix
[Pipeline] withEnv
[Pipeline] {
[Pipeline] sh
+ docker inspect -f . dgllib/dgl-ci-lint
.
[Pipeline] }
[Pipeline] // withEnv
[Pipeline] withDockerContainer
dgci-worker-permanent does not seem to be running inside a container
$ docker run -t -d -u 0:0 -w /home/ubuntu/jenkins_benchmark/workspace/dgl_PR-4648 -v /home/ubuntu/jenkins_benchmark/workspace/dgl_PR-4648:/home/ubuntu/jenkins_benchmark/workspace/dgl_PR-4648:rw,z -v /home/ubuntu/jenkins_benchmark/workspace/dgl_PR-4648@tmp:/home/ubuntu/jenkins_benchmark/workspace/dgl_PR-4648@tmp:rw,z -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** dgllib/dgl-ci-lint cat
$ docker top 59fbbb555874ff8ca9167ae061ff4e42e3c3eeed473ccd147a34d4e5fe5fb8fe -eo pid,comm
[Pipeline] {
[Pipeline] }
$ docker stop --time=1 59fbbb555874ff8ca9167ae061ff4e42e3c3eeed473ccd147a34d4e5fe5fb8fe
$ docker rm -f 59fbbb555874ff8ca9167ae061ff4e42e3c3eeed473ccd147a34d4e5fe5fb8fe
[Pipeline] // withDockerContainer
[Pipeline] }
[Pipeline] // withEnv
[Pipeline] }
[Pipeline] // node
Stage "Regression Test" skipped due to when conditional
[Pipeline] }
[Pipeline] // stage
[Pipeline] stage
[Pipeline] { (CI)
[Pipeline] stage
[Pipeline] { (Lint Check)
[Pipeline] node
Running on dgl-manual-large-cpu in /root/jenkins/workspace/dgl_PR-4648
[Pipeline] {
[Pipeline] checkout
The recommended git tool is: git
using credential 150de63f-189c-4717-bcaf-010460d2f51a
Cloning the remote Git repository
Cloning with configured refspecs honoured and without tags
Cloning repository https://github.com/dmlc/dgl.git
 > git init /root/jenkins/workspace/dgl_PR-4648 # timeout=10
Fetching upstream changes from https://github.com/dmlc/dgl.git
 > git --version # timeout=10
 > git --version # 'git version 2.17.1'
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git fetch --no-tags --progress -- https://github.com/dmlc/dgl.git +refs/pull/4648/head:refs/remotes/origin/PR-4648 +refs/heads/master:refs/remotes/origin/master # timeout=10
 > git config remote.origin.url https://github.com/dmlc/dgl.git # timeout=10
 > git config --add remote.origin.fetch +refs/pull/4648/head:refs/remotes/origin/PR-4648 # timeout=10
 > git config --add remote.origin.fetch +refs/heads/master:refs/remotes/origin/master # timeout=10
Cleaning workspace
Fetching without tags
Merging remotes/origin/master commit d78a3a4baf611b90871a849f58647160c7cd9ab4 into PR head commit 37ed78035f406940ed88a4b60cadd2b7cfde5fea
Merge succeeded, producing 37ed78035f406940ed88a4b60cadd2b7cfde5fea
Checking out Revision 37ed78035f406940ed88a4b60cadd2b7cfde5fea (PR-4648)
Commit message: "fix for pytorch < 1.12"
Cleaning workspace
 > git config remote.origin.url https://github.com/dmlc/dgl.git # timeout=10
 > git rev-parse --verify HEAD # timeout=10
No valid HEAD. Skipping the resetting
 > git clean -fdx # timeout=10
Fetching upstream changes from https://github.com/dmlc/dgl.git
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git fetch --no-tags --progress -- https://github.com/dmlc/dgl.git +refs/pull/4648/head:refs/remotes/origin/PR-4648 +refs/heads/master:refs/remotes/origin/master # timeout=10
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git remote # timeout=10
 > git config --get remote.origin.url # timeout=10
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git merge d78a3a4baf611b90871a849f58647160c7cd9ab4 # timeout=10
 > git rev-parse HEAD^{commit} # timeout=10
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git branch -a -v --no-abbrev # timeout=10
 > git checkout -b PR-4648 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git rev-parse --verify HEAD # timeout=10
Resetting working tree
 > git reset --hard # timeout=10
 > git clean -fdx # timeout=10
[Pipeline] withEnv
[Pipeline] {
[Pipeline] isUnix
[Pipeline] withEnv
[Pipeline] {
[Pipeline] sh
+ docker pull dgllib/dgl-ci-lint
Using default tag: latest
latest: Pulling from dgllib/dgl-ci-lint
Digest: sha256:11b87b4425630143a2560fe6ad584f245ee0a86bc58b40cda3c2be93656b0d5e
Status: Image is up to date for dgllib/dgl-ci-lint:latest
docker.io/dgllib/dgl-ci-lint:latest
[Pipeline] }
[Pipeline] // withEnv
[Pipeline] isUnix
[Pipeline] withEnv
[Pipeline] {
[Pipeline] sh
+ docker inspect -f . dgllib/dgl-ci-lint
.
[Pipeline] }
[Pipeline] // withEnv
[Pipeline] withDockerContainer
dgl-manual-large-cpu does not seem to be running inside a container
$ docker run -t -d -u 0:0 -w /root/jenkins/workspace/dgl_PR-4648 -v /root/jenkins/workspace/dgl_PR-4648:/root/jenkins/workspace/dgl_PR-4648:rw,z -v /root/jenkins/workspace/dgl_PR-4648@tmp:/root/jenkins/workspace/dgl_PR-4648@tmp:rw,z -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** dgllib/dgl-ci-lint cat
$ docker top bb5e029d386aa285458fa71b74bfdb1eafdfce2f99026c9cffde6d1346714b02 -eo pid,comm
[Pipeline] {
[Pipeline] sh
+ rm -rf CMakeLists.txt CONTRIBUTORS.md Jenkinsfile LICENSE NEWS.md README.md apps benchmarks cmake conda dglgo docker docs examples featgraph include pyproject.toml python readthedocs.yml src tensoradapter tests third_party tools tutorials
[Pipeline] checkout
The recommended git tool is: git
using credential 150de63f-189c-4717-bcaf-010460d2f51a
Warning: JENKINS-30600: special launcher org.jenkinsci.plugins.docker.workflow.WithContainerStep$Decorator$1@3d5a0fa4; decorates RemoteLauncher[hudson.remoting.Channel@1d9b9638:dgl-manual-large-cpu] will be ignored (a typical symptom is the Git executable not being run inside a designated container)
Fetching changes from the remote Git repository
Cleaning workspace
Fetching without tags
Merging remotes/origin/master commit d78a3a4baf611b90871a849f58647160c7cd9ab4 into PR head commit 37ed78035f406940ed88a4b60cadd2b7cfde5fea
Merge succeeded, producing 37ed78035f406940ed88a4b60cadd2b7cfde5fea
Checking out Revision 37ed78035f406940ed88a4b60cadd2b7cfde5fea (PR-4648)
Commit message: "fix for pytorch < 1.12"
Cleaning workspace
 > git rev-parse --resolve-git-dir /root/jenkins/workspace/dgl_PR-4648/.git # timeout=10
 > git config remote.origin.url https://github.com/dmlc/dgl.git # timeout=10
 > git rev-parse --verify HEAD # timeout=10
Resetting working tree
 > git reset --hard # timeout=10
 > git clean -fdx # timeout=10
Fetching upstream changes from https://github.com/dmlc/dgl.git
 > git --version # timeout=10
 > git --version # 'git version 2.17.1'
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git fetch --no-tags --progress -- https://github.com/dmlc/dgl.git +refs/pull/4648/head:refs/remotes/origin/PR-4648 +refs/heads/master:refs/remotes/origin/master # timeout=10
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git remote # timeout=10
 > git config --get remote.origin.url # timeout=10
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git merge d78a3a4baf611b90871a849f58647160c7cd9ab4 # timeout=10
 > git rev-parse HEAD^{commit} # timeout=10
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git branch -a -v --no-abbrev # timeout=10
 > git branch -D PR-4648 # timeout=10
 > git checkout -b PR-4648 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git rev-parse --verify HEAD # timeout=10
Resetting working tree
 > git reset --hard # timeout=10
 > git clean -fdx # timeout=10
[Pipeline] sh
+ git submodule update --recursive --init
Submodule 'third_party/METIS' (https://github.com/KarypisLab/METIS.git) registered for path 'third_party/METIS'
Submodule 'third_party/dlpack' (https://github.com/dmlc/dlpack.git) registered for path 'third_party/dlpack'
Submodule 'third_party/dmlc-core' (https://github.com/dmlc/dmlc-core.git) registered for path 'third_party/dmlc-core'
Submodule 'third_party/googletest' (https://github.com/google/googletest.git) registered for path 'third_party/googletest'
Submodule 'third_party/libxsmm' (https://github.com/hfp/libxsmm.git) registered for path 'third_party/libxsmm'
Submodule 'third_party/nanoflann' (https://github.com/jlblancoc/nanoflann) registered for path 'third_party/nanoflann'
Submodule 'third_party/nccl' (https://github.com/nvidia/nccl) registered for path 'third_party/nccl'
Submodule 'third_party/phmap' (https://github.com/greg7mdp/parallel-hashmap.git) registered for path 'third_party/phmap'
Submodule 'third_party/tensorpipe' (https://github.com/pytorch/tensorpipe) registered for path 'third_party/tensorpipe'
Submodule 'third_party/thrust' (https://github.com/NVIDIA/thrust.git) registered for path 'third_party/thrust'
Submodule 'third_party/tvm' (https://github.com/apache/incubator-tvm) registered for path 'third_party/tvm'
Submodule 'third_party/xbyak' (https://github.com/herumi/xbyak) registered for path 'third_party/xbyak'
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/METIS'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/dlpack'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/dmlc-core'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/googletest'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/libxsmm'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/nanoflann'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/nccl'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/phmap'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tensorpipe'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/thrust'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tvm'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/xbyak'...
Submodule path 'third_party/METIS': checked out '10603482cc055626ff99d2ac3ab7e73d1119439d'
Submodule 'GKlib' (https://github.com/KarypisLab/GKlib.git) registered for path 'third_party/METIS/GKlib'
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/METIS/GKlib'...
Submodule path 'third_party/METIS/GKlib': checked out '62de20c7f97c719abdc96d7c3e266a1f7cb52bc5'
Submodule path 'third_party/dlpack': checked out 'e2bdd3bee8cb6501558042633fa59144cc8b7f5f'
Submodule path 'third_party/dmlc-core': checked out 'bfad207b448480783a1f428ae3d93d87032d8349'
Submodule path 'third_party/googletest': checked out 'f71fb4f9a912ec945401cc49a287a759b6131026'
Submodule path 'third_party/libxsmm': checked out 'fa687556130b6298430f1c0555a14cf79ab6101c'
Submodule path 'third_party/nanoflann': checked out '4c47ca200209550c5628c89803591f8a753c8181'
Submodule path 'third_party/nccl': checked out 'e11238b3029795d33f958b5868d47c90c4f22628'
Submodule path 'third_party/phmap': checked out '25293cefd8b85491b45600c03fe8edf07647553f'
Submodule path 'third_party/tensorpipe': checked out '6042f1a4cbce8eef997f11ed0012de137b317361'
Submodule 'third_party/googletest' (https://github.com/google/googletest.git) registered for path 'third_party/tensorpipe/third_party/googletest'
Submodule 'third_party/libnop' (https://github.com/google/libnop.git) registered for path 'third_party/tensorpipe/third_party/libnop'
Submodule 'third_party/libuv' (https://github.com/libuv/libuv.git) registered for path 'third_party/tensorpipe/third_party/libuv'
Submodule 'third_party/pybind11' (https://github.com/pybind/pybind11.git) registered for path 'third_party/tensorpipe/third_party/pybind11'
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tensorpipe/third_party/googletest'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tensorpipe/third_party/libnop'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tensorpipe/third_party/libuv'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tensorpipe/third_party/pybind11'...
Submodule path 'third_party/tensorpipe/third_party/googletest': checked out 'aee0f9d9b5b87796ee8a0ab26b7587ec30e8858e'
Submodule path 'third_party/tensorpipe/third_party/libnop': checked out 'aa95422ea8c409e3f078d2ee7708a5f59a8b9fa2'
Submodule path 'third_party/tensorpipe/third_party/libuv': checked out '1dff88e5161cba5c59276d2070d2e304e4dcb242'
Submodule path 'third_party/tensorpipe/third_party/pybind11': checked out 'a23996fce38ff6ccfbcdc09f1e63f2c4be5ea2ef'
Submodule 'tools/clang' (https://github.com/wjakob/clang-cindex-python3) registered for path 'third_party/tensorpipe/third_party/pybind11/tools/clang'
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tensorpipe/third_party/pybind11/tools/clang'...
Submodule path 'third_party/tensorpipe/third_party/pybind11/tools/clang': checked out '6a00cbc4a9b8e68b71caf7f774b3f9c753ae84d5'
Submodule path 'third_party/thrust': checked out '6a3078c64cab0e2f276340fa5dcafa0d758ed890'
Submodule 'cub' (https://github.com/NVIDIA/cub.git) registered for path 'third_party/thrust/dependencies/cub'
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/thrust/dependencies/cub'...
Submodule path 'third_party/thrust/dependencies/cub': checked out 'cdaa9558a85e45d849016e5fe7b6e4ee79113f95'
Submodule path 'third_party/tvm': checked out 'b2e418cb109df4cd1f17a2cf2894a1b396a6b838'
Submodule 'dlpack' (https://github.com/dmlc/dlpack) registered for path 'third_party/tvm/3rdparty/dlpack'
Submodule 'dmlc-core' (https://github.com/dmlc/dmlc-core) registered for path 'third_party/tvm/3rdparty/dmlc-core'
Submodule '3rdparty/rang' (https://github.com/agauniyal/rang) registered for path 'third_party/tvm/3rdparty/rang'
Submodule '3rdparty/vta-hw' (https://github.com/apache/incubator-tvm-vta) registered for path 'third_party/tvm/3rdparty/vta-hw'
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tvm/3rdparty/dlpack'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tvm/3rdparty/dmlc-core'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tvm/3rdparty/rang'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tvm/3rdparty/vta-hw'...
Submodule path 'third_party/tvm/3rdparty/dlpack': checked out '3ec04430e89a6834e5a1b99471f415fa939bf642'
Submodule path 'third_party/tvm/3rdparty/dmlc-core': checked out '6c401e242c59a1f4c913918246591bb13fd714e7'
Submodule path 'third_party/tvm/3rdparty/rang': checked out 'cabe04d6d6b05356fa8f9741704924788f0dd762'
Submodule path 'third_party/tvm/3rdparty/vta-hw': checked out '87ce9acfae550d1a487746e9d06c2e250076e54c'
Submodule path 'third_party/xbyak': checked out '757e4063f6464740b8ff4a2cae9136d2f8458020'
[Pipeline] sh
+ bash tests/scripts/task_lint.sh
Checking code style of C++ codes...
=====113/113 cpp-header files passed check=====
=====176/176 cpp-source files passed check=====
All passed!
Checking code style of python codes...


Report
======
19102 statements analysed.

Statistics by type
------------------

+---------+-------+-----------+-----------+------------+---------+
|type     |number |old number |difference |%documented |%badname |
+=========+=======+===========+===========+============+=========+
|module   |229    |NC         |NC         |100.00      |0.00     |
+---------+-------+-----------+-----------+------------+---------+
|class    |343    |NC         |NC         |99.71       |1.46     |
+---------+-------+-----------+-----------+------------+---------+
|method   |1521   |NC         |NC         |99.61       |0.33     |
+---------+-------+-----------+-----------+------------+---------+
|function |663    |NC         |NC         |100.00      |1.06     |
+---------+-------+-----------+-----------+------------+---------+



External dependencies
---------------------
::

    dgl (dgl.distgnn.tools.tools)
      \-_api_internal (dgl._ffi.object)
      \-_dataloading 
      | \-dataloader (dgl._dataloading.pytorch.dataloader)
      \-_deprecate 
      | \-graph (dgl.transforms.functional)
      | \-nodeflow (dgl.network)
      | \-runtime 
      |   \-ir (dgl.function.message)
      |     \-var (dgl.function.message)
      \-_ffi 
      | \-_ctypes 
      | | \-function (dgl._ffi.function)
      | | \-ndarray (dgl._ffi._ctypes.function)
      | | \-object (dgl._ffi._ctypes.function)
      | | \-types (dgl._ffi._ctypes.ndarray)
      | \-base (dgl._ffi._ctypes.ndarray)
      | \-function (dgl.transforms.functional)
      | \-libinfo (dgl._ffi.base)
      | \-ndarray (dgl.distributed.shared_mem_utils)
      | \-object (dgl.distributed.rpc)
      | \-object_generic (dgl._ffi._ctypes.function)
      | \-runtime_ctypes (dgl._ffi._ctypes.ndarray)
      | \-streams (dgl.heterograph_index)
      \-backend (dgl.transforms.functional)
      | \-pytorch (dgl.nn.pytorch.sparse_emb)
      \-base (dgl.transforms.functional)
      \-batch (dgl.transforms.functional)
      \-convert (dgl.transforms.functional)
      \-core (dgl.heterograph)
      \-data 
      | \-utils (dgl.distgnn.tools.tools)
      \-dataloading 
      | \-base (dgl.dataloading.dataloader)
      \-distributed (dgl._dataloading.pytorch.dataloader)
      | \-constants (dgl.distributed.rpc)
      | \-dist_context (dgl.distributed.dist_tensor)
      | \-dist_tensor (dgl.distributed.optim.pytorch.sparse_optim)
      | \-id_map (dgl.distributed.graph_partition_book)
      | \-kvstore (dgl.distributed.dist_tensor)
      | \-nn 
      | | \-pytorch (dgl.distributed.optim.pytorch.sparse_optim)
      | \-optim 
      | | \-pytorch 
      | |   \-utils (dgl.distributed.optim.pytorch.sparse_optim)
      | \-role (dgl.distributed.dist_tensor)
      | \-rpc (dgl.distributed.dist_tensor)
      | \-shared_mem_utils (dgl.distributed.dist_graph)
      | \-standalone_kvstore (dgl.distributed.kvstore)
      \-frame (dgl.transforms.functional)
      \-function (dgl.transforms.functional)
      | \-base (dgl.function.message)
      \-geometry 
      | \-capi (dgl.geometry.fps)
      \-graph_index (dgl.heterograph)
      \-heterograph (dgl.transforms.functional)
      \-heterograph_index (dgl.transforms.functional)
      \-init (dgl.frame)
      \-mock_sparse 
      | \-diag_matrix (dgl.mock_sparse.elementwise_op_sp)
      | \-sp_matrix (dgl.mock_sparse.elementwise_op_sp)
      \-multiprocessing (dgl.dataloading.dataloader)
      \-ndarray (dgl.transforms.functional)
      \-nn 
      | \-functional (dgl.nn.tensorflow.conv.gatconv)
      | \-mxnet 
      | | \-utils (dgl.nn.mxnet.conv.relgraphconv)
      | \-pytorch (dgl.optim.pytorch.sparse_optim)
      | | \-conv 
      | | | \-graphconv (dgl.nn.pytorch.conv.sgconv)
      | | \-linear (dgl.nn.pytorch.conv.hgtconv)
      | | \-softmax (dgl.nn.pytorch.conv.hgtconv)
      | | \-utils (dgl.nn.pytorch.conv.gatconv)
      | \-tensorflow 
      |   \-utils (dgl.nn.tensorflow.conv.gatconv)
      \-ops (dgl.nn.pytorch.linear)
      \-partition (dgl.transforms.functional)
      \-random (dgl.distributed.partition)
      \-readout (dgl.nn.tensorflow.glob)
      \-sampling (dgl.distributed.graph_services)
      | \-neighbor (dgl.transforms.functional)
      | \-utils (dgl.sampling.neighbor)
      \-sparse (dgl.distgnn.partition.libra_partition)
      \-storages (dgl.dataloading.dataloader)
      | \-base (dgl.storages.numpy)
      \-subgraph (dgl.transforms.functional)
      \-transforms (dgl.distributed.dist_graph)
      | \-functional (dgl.transforms.module)
      \-traversal (dgl.propagate)
      \-udf (dgl.core)
      \-utils (dgl.transforms.functional)
      | \-checks (dgl.utils.data)
      \-view (dgl.heterograph)



Raw metrics
-----------

+----------+-------+------+---------+-----------+
|type      |number |%     |previous |difference |
+==========+=======+======+=========+===========+
|code      |22834  |34.48 |NC       |NC         |
+----------+-------+------+---------+-----------+
|docstring |36999  |55.88 |NC       |NC         |
+----------+-------+------+---------+-----------+
|comment   |1978   |2.99  |NC       |NC         |
+----------+-------+------+---------+-----------+
|empty     |4404   |6.65  |NC       |NC         |
+----------+-------+------+---------+-----------+



Duplication
-----------

+-------------------------+------+---------+-----------+
|                         |now   |previous |difference |
+=========================+======+=========+===========+
|nb duplicated lines      |0     |NC       |NC         |
+-------------------------+------+---------+-----------+
|percent duplicated lines |0.000 |NC       |NC         |
+-------------------------+------+---------+-----------+



Messages by category
--------------------

+-----------+-------+---------+-----------+
|type       |number |previous |difference |
+===========+=======+=========+===========+
|convention |0      |NC       |NC         |
+-----------+-------+---------+-----------+
|refactor   |0      |NC       |NC         |
+-----------+-------+---------+-----------+
|warning    |0      |NC       |NC         |
+-----------+-------+---------+-----------+
|error      |0      |NC       |NC         |
+-----------+-------+---------+-----------+




------------------------------------
Your code has been rated at 10.00/10

Post stage
[Pipeline] cleanWs
[WS-CLEANUP] Deleting project workspace...
[WS-CLEANUP] Deferred wipeout is disabled by the job configuration...
[WS-CLEANUP] done
[Pipeline] }
$ docker stop --time=1 bb5e029d386aa285458fa71b74bfdb1eafdfce2f99026c9cffde6d1346714b02
$ docker rm -f bb5e029d386aa285458fa71b74bfdb1eafdfce2f99026c9cffde6d1346714b02
[Pipeline] // withDockerContainer
[Pipeline] }
[Pipeline] // withEnv
[Pipeline] }
[Pipeline] // node
[Pipeline] }
[Pipeline] // stage
[Pipeline] stage
[Pipeline] { (Build)
[Pipeline] parallel
[Pipeline] { (Branch: CPU Build)
[Pipeline] { (Branch: GPU Build)
[Pipeline] { (Branch: PyTorch Cugraph GPU Build)
[Pipeline] { (Branch: CPU Build (Win64))
[Pipeline] stage
[Pipeline] { (CPU Build)
[Pipeline] stage
[Pipeline] { (GPU Build)
[Pipeline] stage
[Pipeline] { (PyTorch Cugraph GPU Build)
[Pipeline] stage
[Pipeline] { (CPU Build (Win64))
[Pipeline] node
Running on dglci-windows in C:\Jenkins\workspace\dgl_PR-4648
[Pipeline] node
[Pipeline] node
Running on dgl-manual-large-cpu in /root/jenkins/workspace/dgl_PR-4648
Running on dgl-manual-large-cpu in /root/jenkins/workspace/dgl_PR-4648@2
[Pipeline] node
Running on dgl-manual-large-cpu in /root/jenkins/workspace/dgl_PR-4648@3
[Pipeline] {
[Pipeline] {
[Pipeline] {
[Pipeline] {
[Pipeline] checkout
The recommended git tool is: git
using credential 150de63f-189c-4717-bcaf-010460d2f51a
Cloning the remote Git repository
Cloning with configured refspecs honoured and without tags
[Pipeline] checkout
The recommended git tool is: git
[Pipeline] checkout
using credential 150de63f-189c-4717-bcaf-010460d2f51a
The recommended git tool is: git
Cloning the remote Git repository
Cloning with configured refspecs honoured and without tags
[Pipeline] checkout
using credential 150de63f-189c-4717-bcaf-010460d2f51a
The recommended git tool is: git
Cloning the remote Git repository
Cloning with configured refspecs honoured and without tags
using credential 150de63f-189c-4717-bcaf-010460d2f51a
Cloning the remote Git repository
Cloning with configured refspecs honoured and without tags
Cloning repository https://github.com/dmlc/dgl.git
 > git init C:\Jenkins\workspace\dgl_PR-4648 # timeout=10
Fetching upstream changes from https://github.com/dmlc/dgl.git
 > git --version # timeout=10
 > git --version # 'git version 2.20.0.windows.1'
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git fetch --no-tags --force --progress -- https://github.com/dmlc/dgl.git +refs/pull/4648/head:refs/remotes/origin/PR-4648 +refs/heads/master:refs/remotes/origin/master # timeout=10
Cloning repository https://github.com/dmlc/dgl.git
 > git init /root/jenkins/workspace/dgl_PR-4648 # timeout=10
Fetching upstream changes from https://github.com/dmlc/dgl.git
 > git --version # timeout=10
 > git --version # 'git version 2.17.1'
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git fetch --no-tags --progress -- https://github.com/dmlc/dgl.git +refs/pull/4648/head:refs/remotes/origin/PR-4648 +refs/heads/master:refs/remotes/origin/master # timeout=10
Cloning repository https://github.com/dmlc/dgl.git
 > git init /root/jenkins/workspace/dgl_PR-4648@2 # timeout=10
Fetching upstream changes from https://github.com/dmlc/dgl.git
 > git --version # timeout=10
 > git --version # 'git version 2.17.1'
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git fetch --no-tags --progress -- https://github.com/dmlc/dgl.git +refs/pull/4648/head:refs/remotes/origin/PR-4648 +refs/heads/master:refs/remotes/origin/master # timeout=10
Cloning repository https://github.com/dmlc/dgl.git
 > git init /root/jenkins/workspace/dgl_PR-4648@3 # timeout=10
Fetching upstream changes from https://github.com/dmlc/dgl.git
 > git --version # timeout=10
 > git --version # 'git version 2.17.1'
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git fetch --no-tags --progress -- https://github.com/dmlc/dgl.git +refs/pull/4648/head:refs/remotes/origin/PR-4648 +refs/heads/master:refs/remotes/origin/master # timeout=10
Cleaning workspace
Fetching without tags
Cleaning workspace
Fetching without tags
 > git config remote.origin.url https://github.com/dmlc/dgl.git # timeout=10
 > git config --add remote.origin.fetch +refs/pull/4648/head:refs/remotes/origin/PR-4648 # timeout=10
 > git config --add remote.origin.fetch +refs/heads/master:refs/remotes/origin/master # timeout=10
 > git config remote.origin.url https://github.com/dmlc/dgl.git # timeout=10
 > git rev-parse --verify HEAD # timeout=10
No valid HEAD. Skipping the resetting
 > git clean -fdx # timeout=10
Fetching upstream changes from https://github.com/dmlc/dgl.git
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git fetch --no-tags --progress -- https://github.com/dmlc/dgl.git +refs/pull/4648/head:refs/remotes/origin/PR-4648 +refs/heads/master:refs/remotes/origin/master # timeout=10
 > git config remote.origin.url https://github.com/dmlc/dgl.git # timeout=10
 > git config --add remote.origin.fetch +refs/pull/4648/head:refs/remotes/origin/PR-4648 # timeout=10
 > git config --add remote.origin.fetch +refs/heads/master:refs/remotes/origin/master # timeout=10
 > git config remote.origin.url https://github.com/dmlc/dgl.git # timeout=10
 > git rev-parse --verify HEAD # timeout=10
No valid HEAD. Skipping the resetting
 > git clean -fdx # timeout=10
Fetching upstream changes from https://github.com/dmlc/dgl.git
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git fetch --no-tags --progress -- https://github.com/dmlc/dgl.git +refs/pull/4648/head:refs/remotes/origin/PR-4648 +refs/heads/master:refs/remotes/origin/master # timeout=10
Cleaning workspace
Fetching without tags
Merging remotes/origin/master commit d78a3a4baf611b90871a849f58647160c7cd9ab4 into PR head commit 37ed78035f406940ed88a4b60cadd2b7cfde5fea
Merging remotes/origin/master commit d78a3a4baf611b90871a849f58647160c7cd9ab4 into PR head commit 37ed78035f406940ed88a4b60cadd2b7cfde5fea
 > git config remote.origin.url https://github.com/dmlc/dgl.git # timeout=10
 > git config --add remote.origin.fetch +refs/pull/4648/head:refs/remotes/origin/PR-4648 # timeout=10
 > git config --add remote.origin.fetch +refs/heads/master:refs/remotes/origin/master # timeout=10
 > git config remote.origin.url https://github.com/dmlc/dgl.git # timeout=10
Cleaning workspace
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
Fetching without tags
 > git config remote.origin.url https://github.com/dmlc/dgl.git # timeout=10
 > git config --add remote.origin.fetch +refs/pull/4648/head:refs/remotes/origin/PR-4648 # timeout=10
 > git config --add remote.origin.fetch +refs/heads/master:refs/remotes/origin/master # timeout=10
 > git config remote.origin.url https://github.com/dmlc/dgl.git # timeout=10
 > git rev-parse --verify HEAD # timeout=10
No valid HEAD. Skipping the resetting
 > git clean -fdx # timeout=10
Fetching upstream changes from https://github.com/dmlc/dgl.git
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git fetch --no-tags --progress -- https://github.com/dmlc/dgl.git +refs/pull/4648/head:refs/remotes/origin/PR-4648 +refs/heads/master:refs/remotes/origin/master # timeout=10
Merge succeeded, producing 37ed78035f406940ed88a4b60cadd2b7cfde5fea
Checking out Revision 37ed78035f406940ed88a4b60cadd2b7cfde5fea (PR-4648)
Merge succeeded, producing 37ed78035f406940ed88a4b60cadd2b7cfde5fea
Checking out Revision 37ed78035f406940ed88a4b60cadd2b7cfde5fea (PR-4648)
Merging remotes/origin/master commit d78a3a4baf611b90871a849f58647160c7cd9ab4 into PR head commit 37ed78035f406940ed88a4b60cadd2b7cfde5fea
Merging remotes/origin/master commit d78a3a4baf611b90871a849f58647160c7cd9ab4 into PR head commit 37ed78035f406940ed88a4b60cadd2b7cfde5fea
Commit message: "fix for pytorch < 1.12"
Cleaning workspace
Commit message: "fix for pytorch < 1.12"
Cleaning workspace
Merge succeeded, producing 37ed78035f406940ed88a4b60cadd2b7cfde5fea
Checking out Revision 37ed78035f406940ed88a4b60cadd2b7cfde5fea (PR-4648)
Commit message: "fix for pytorch < 1.12"
Cleaning workspace
Merge succeeded, producing 37ed78035f406940ed88a4b60cadd2b7cfde5fea
Checking out Revision 37ed78035f406940ed88a4b60cadd2b7cfde5fea (PR-4648)
 > git rev-parse --verify HEAD # timeout=10
No valid HEAD. Skipping the resetting
 > git clean -fdx # timeout=10
Fetching upstream changes from https://github.com/dmlc/dgl.git
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git fetch --no-tags --force --progress -- https://github.com/dmlc/dgl.git +refs/pull/4648/head:refs/remotes/origin/PR-4648 +refs/heads/master:refs/remotes/origin/master # timeout=10
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git remote # timeout=10
 > git config --get remote.origin.url # timeout=10
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git merge d78a3a4baf611b90871a849f58647160c7cd9ab4 # timeout=10
 > git rev-parse "HEAD^{commit}" # timeout=10
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
[Pipeline] withEnv
[Pipeline] {
 > git remote # timeout=10
 > git config --get remote.origin.url # timeout=10
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git merge d78a3a4baf611b90871a849f58647160c7cd9ab4 # timeout=10
 > git rev-parse HEAD^{commit} # timeout=10
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git branch -a -v --no-abbrev # timeout=10
 > git checkout -b PR-4648 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git rev-parse --verify HEAD # timeout=10
Resetting working tree
 > git reset --hard # timeout=10
 > git clean -fdx # timeout=10
 > git remote # timeout=10
 > git config --get remote.origin.url # timeout=10
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git merge d78a3a4baf611b90871a849f58647160c7cd9ab4 # timeout=10
 > git rev-parse HEAD^{commit} # timeout=10
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git branch -a -v --no-abbrev # timeout=10
 > git checkout -b PR-4648 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git rev-parse --verify HEAD # timeout=10
Resetting working tree
 > git reset --hard # timeout=10
 > git clean -fdx # timeout=10
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git remote # timeout=10
 > git config --get remote.origin.url # timeout=10
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git merge d78a3a4baf611b90871a849f58647160c7cd9ab4 # timeout=10
 > git rev-parse HEAD^{commit} # timeout=10
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git branch -a -v --no-abbrev # timeout=10
 > git checkout -b PR-4648 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git rev-parse --verify HEAD # timeout=10
Resetting working tree
 > git reset --hard # timeout=10
 > git clean -fdx # timeout=10
[Pipeline] isUnix
[Pipeline] withEnv
[Pipeline] {
[Pipeline] withEnv
[Pipeline] {
[Pipeline] sh
[Pipeline] isUnix
[Pipeline] withEnv
[Pipeline] {
[Pipeline] sh
[Pipeline] withEnv
[Pipeline] {
[Pipeline] isUnix
[Pipeline] withEnv
[Pipeline] {
[Pipeline] sh
+ docker pull dgllib/dgl-ci-gpu:cu101_v220816
+ docker pull dgllib/dgl-ci-cpu:v220816
Commit message: "fix for pytorch < 1.12"
Cleaning workspace
+ docker pull rapidsai/cugraph_nightly_torch-cuda:11.5-base-ubuntu18.04-py3.9-pytorch1.12.0-rapids22.10
 > git branch -a -v --no-abbrev # timeout=10
 > git checkout -b PR-4648 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git rev-parse --verify HEAD # timeout=10
Resetting working tree
 > git reset --hard # timeout=10
 > git clean -fdx # timeout=10
cu101_v220816: Pulling from dgllib/dgl-ci-gpu
Digest: sha256:ca40fc52876a2563a4e904d0c271d658c1acc8e6a4f8611b578bb49f8c7fd925
Status: Image is up to date for dgllib/dgl-ci-gpu:cu101_v220816
docker.io/dgllib/dgl-ci-gpu:cu101_v220816
[Pipeline] }
[Pipeline] // withEnv
[Pipeline] isUnix
[Pipeline] withEnv
[Pipeline] {
v220816: Pulling from dgllib/dgl-ci-cpu
Digest: sha256:64b385c33b44dc57cb96ff264a84d8dfb8ced0caa9b30fbc4cec6d5ee511b099
Status: Image is up to date for dgllib/dgl-ci-cpu:v220816
docker.io/dgllib/dgl-ci-cpu:v220816
[Pipeline] }
[Pipeline] // withEnv
[Pipeline] sh
[Pipeline] isUnix
[Pipeline] withEnv
[Pipeline] {
[Pipeline] withEnv
[Pipeline] {
11.5-base-ubuntu18.04-py3.9-pytorch1.12.0-rapids22.10: Pulling from rapidsai/cugraph_nightly_torch-cuda
Digest: sha256:72e8cb2632449beac4895f24b55018dea21f79da47fe39be05ef83a4fd3ddb67
Status: Image is up to date for rapidsai/cugraph_nightly_torch-cuda:11.5-base-ubuntu18.04-py3.9-pytorch1.12.0-rapids22.10
docker.io/rapidsai/cugraph_nightly_torch-cuda:11.5-base-ubuntu18.04-py3.9-pytorch1.12.0-rapids22.10
[Pipeline] sh
[Pipeline] }
[Pipeline] // withEnv
[Pipeline] isUnix
[Pipeline] withEnv
[Pipeline] {
[Pipeline] checkout
The recommended git tool is: git
using credential 150de63f-189c-4717-bcaf-010460d2f51a
[Pipeline] sh
Fetching changes from the remote Git repository
Cleaning workspace
+ docker inspect -f . dgllib/dgl-ci-gpu:cu101_v220816
.
[Pipeline] }
Fetching without tags
[Pipeline] // withEnv
[Pipeline] withDockerContainer
+ docker inspect -f . dgllib/dgl-ci-cpu:v220816
.
dgl-manual-large-cpu does not seem to be running inside a container
+ docker inspect -f . rapidsai/cugraph_nightly_torch-cuda:11.5-base-ubuntu18.04-py3.9-pytorch1.12.0-rapids22.10
.
$ docker run -t -d -u 0:0 -u root -w /root/jenkins/workspace/dgl_PR-4648@2 -v /root/jenkins/workspace/dgl_PR-4648@2:/root/jenkins/workspace/dgl_PR-4648@2:rw,z -v /root/jenkins/workspace/dgl_PR-4648@2@tmp:/root/jenkins/workspace/dgl_PR-4648@2@tmp:rw,z -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** dgllib/dgl-ci-gpu:cu101_v220816 cat
Merging remotes/origin/master commit d78a3a4baf611b90871a849f58647160c7cd9ab4 into PR head commit 37ed78035f406940ed88a4b60cadd2b7cfde5fea
Merge succeeded, producing 37ed78035f406940ed88a4b60cadd2b7cfde5fea
Checking out Revision 37ed78035f406940ed88a4b60cadd2b7cfde5fea (PR-4648)
Commit message: "fix for pytorch < 1.12"
Cleaning workspace
 > git rev-parse --resolve-git-dir C:\Jenkins\workspace\dgl_PR-4648\.git # timeout=10
 > git config remote.origin.url https://github.com/dmlc/dgl.git # timeout=10
 > git rev-parse --verify HEAD # timeout=10
Resetting working tree
 > git reset --hard # timeout=10
 > git clean -fdx # timeout=10
Fetching upstream changes from https://github.com/dmlc/dgl.git
 > git --version # timeout=10
 > git --version # 'git version 2.20.0.windows.1'
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git fetch --no-tags --force --progress -- https://github.com/dmlc/dgl.git +refs/pull/4648/head:refs/remotes/origin/PR-4648 +refs/heads/master:refs/remotes/origin/master # timeout=10
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git remote # timeout=10
 > git config --get remote.origin.url # timeout=10
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git merge d78a3a4baf611b90871a849f58647160c7cd9ab4 # timeout=10
 > git rev-parse "HEAD^{commit}" # timeout=10
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git branch -a -v --no-abbrev # timeout=10
 > git branch -D PR-4648 # timeout=10
 > git checkout -b PR-4648 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git rev-parse --verify HEAD # timeout=10
Resetting working tree
 > git reset --hard # timeout=10
 > git clean -fdx # timeout=10
$ docker top 681e11859cd167f00c1889ac9c7c8c96d552be64f67c250e1cf0b5bc1d83bafe -eo pid,comm
[Pipeline] {
[Pipeline] }
[Pipeline] bat
[Pipeline] }
[Pipeline] // withEnv
[Pipeline] // withEnv
[Pipeline] withDockerContainer
dgl-manual-large-cpu does not seem to be running inside a container
$ docker run -t -d -u 0:0 -u root -w /root/jenkins/workspace/dgl_PR-4648 -v /root/jenkins/workspace/dgl_PR-4648:/root/jenkins/workspace/dgl_PR-4648:rw,z -v /root/jenkins/workspace/dgl_PR-4648@tmp:/root/jenkins/workspace/dgl_PR-4648@tmp:rw,z -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** dgllib/dgl-ci-cpu:v220816 cat

administrator@WIN-O7QS55HVSDB C:\Jenkins\workspace\dgl_PR-4648>git submodule update --recursive --init 
Submodule 'third_party/METIS' (https://github.com/KarypisLab/METIS.git) registered for path 'third_party/METIS'
Submodule 'third_party/dlpack' (https://github.com/dmlc/dlpack.git) registered for path 'third_party/dlpack'
Submodule 'third_party/dmlc-core' (https://github.com/dmlc/dmlc-core.git) registered for path 'third_party/dmlc-core'
Submodule 'third_party/googletest' (https://github.com/google/googletest.git) registered for path 'third_party/googletest'
Submodule 'third_party/libxsmm' (https://github.com/hfp/libxsmm.git) registered for path 'third_party/libxsmm'
Submodule 'third_party/nanoflann' (https://github.com/jlblancoc/nanoflann) registered for path 'third_party/nanoflann'
Submodule 'third_party/nccl' (https://github.com/nvidia/nccl) registered for path 'third_party/nccl'
Submodule 'third_party/phmap' (https://github.com/greg7mdp/parallel-hashmap.git) registered for path 'third_party/phmap'
Submodule 'third_party/tensorpipe' (https://github.com/pytorch/tensorpipe) registered for path 'third_party/tensorpipe'
Submodule 'third_party/thrust' (https://github.com/NVIDIA/thrust.git) registered for path 'third_party/thrust'
Submodule 'third_party/tvm' (https://github.com/apache/incubator-tvm) registered for path 'third_party/tvm'
Submodule 'third_party/xbyak' (https://github.com/herumi/xbyak) registered for path 'third_party/xbyak'
Cloning into 'C:/Jenkins/workspace/dgl_PR-4648/third_party/METIS'...
$ docker top 856b09e8bd127c677672426c8af0a6bfefd91dd73a4c28827d4d41ef0f5d0fcc -eo pid,comm
[Pipeline] {
[Pipeline] withDockerContainer
dgl-manual-large-cpu does not seem to be running inside a container
$ docker run -t -d -u 0:0 -u root -w /root/jenkins/workspace/dgl_PR-4648@3 -v /root/jenkins/workspace/dgl_PR-4648@3:/root/jenkins/workspace/dgl_PR-4648@3:rw,z -v /root/jenkins/workspace/dgl_PR-4648@3@tmp:/root/jenkins/workspace/dgl_PR-4648@3@tmp:rw,z -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** rapidsai/cugraph_nightly_torch-cuda:11.5-base-ubuntu18.04-py3.9-pytorch1.12.0-rapids22.10 cat
$ docker top 4e33dd6e703ed0dd95f590a393973aab145ae4e5b4395a56a04d85836079138d -eo pid,comm
[Pipeline] {
[Pipeline] sh
[Pipeline] sh
[Pipeline] sh
+ rm -rf CMakeLists.txt CONTRIBUTORS.md Jenkinsfile LICENSE NEWS.md README.md apps benchmarks cmake conda dglgo docker docs examples featgraph include pyproject.toml python readthedocs.yml src tensoradapter tests third_party tools tutorials
+ rm -rf CMakeLists.txt CONTRIBUTORS.md Jenkinsfile LICENSE NEWS.md README.md apps benchmarks cmake conda dglgo docker docs examples featgraph include pyproject.toml python readthedocs.yml src tensoradapter tests third_party tools tutorials
+ rm -rf CMakeLists.txt CONTRIBUTORS.md Jenkinsfile LICENSE NEWS.md README.md apps benchmarks cmake conda dglgo docker docs examples featgraph include pyproject.toml python readthedocs.yml src tensoradapter tests third_party tools tutorials
[Pipeline] checkout
The recommended git tool is: git
using credential 150de63f-189c-4717-bcaf-010460d2f51a
Warning: JENKINS-30600: special launcher org.jenkinsci.plugins.docker.workflow.WithContainerStep$Decorator$1@1b0f444b; decorates RemoteLauncher[hudson.remoting.Channel@1d9b9638:dgl-manual-large-cpu] will be ignored (a typical symptom is the Git executable not being run inside a designated container)
Fetching changes from the remote Git repository
Cleaning workspace
[Pipeline] checkout
The recommended git tool is: git
using credential 150de63f-189c-4717-bcaf-010460d2f51a
Warning: JENKINS-30600: special launcher org.jenkinsci.plugins.docker.workflow.WithContainerStep$Decorator$1@106288df; decorates RemoteLauncher[hudson.remoting.Channel@1d9b9638:dgl-manual-large-cpu] will be ignored (a typical symptom is the Git executable not being run inside a designated container)
Fetching changes from the remote Git repository
Cleaning workspace
[Pipeline] checkout
The recommended git tool is: git
using credential 150de63f-189c-4717-bcaf-010460d2f51a
Warning: JENKINS-30600: special launcher org.jenkinsci.plugins.docker.workflow.WithContainerStep$Decorator$1@314ff248; decorates RemoteLauncher[hudson.remoting.Channel@1d9b9638:dgl-manual-large-cpu] will be ignored (a typical symptom is the Git executable not being run inside a designated container)
Fetching changes from the remote Git repository
Cleaning workspace
Fetching without tags
Fetching without tags
Fetching without tags
Merging remotes/origin/master commit d78a3a4baf611b90871a849f58647160c7cd9ab4 into PR head commit 37ed78035f406940ed88a4b60cadd2b7cfde5fea
Merging remotes/origin/master commit d78a3a4baf611b90871a849f58647160c7cd9ab4 into PR head commit 37ed78035f406940ed88a4b60cadd2b7cfde5fea
Merging remotes/origin/master commit d78a3a4baf611b90871a849f58647160c7cd9ab4 into PR head commit 37ed78035f406940ed88a4b60cadd2b7cfde5fea
Merge succeeded, producing 37ed78035f406940ed88a4b60cadd2b7cfde5fea
Checking out Revision 37ed78035f406940ed88a4b60cadd2b7cfde5fea (PR-4648)
Merge succeeded, producing 37ed78035f406940ed88a4b60cadd2b7cfde5fea
Checking out Revision 37ed78035f406940ed88a4b60cadd2b7cfde5fea (PR-4648)
Merge succeeded, producing 37ed78035f406940ed88a4b60cadd2b7cfde5fea
Checking out Revision 37ed78035f406940ed88a4b60cadd2b7cfde5fea (PR-4648)
Commit message: "fix for pytorch < 1.12"
Cleaning workspace
 > git rev-parse --resolve-git-dir /root/jenkins/workspace/dgl_PR-4648@2/.git # timeout=10
 > git config remote.origin.url https://github.com/dmlc/dgl.git # timeout=10
 > git rev-parse --verify HEAD # timeout=10
Resetting working tree
 > git reset --hard # timeout=10
 > git clean -fdx # timeout=10
Fetching upstream changes from https://github.com/dmlc/dgl.git
 > git --version # timeout=10
 > git --version # 'git version 2.17.1'
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git fetch --no-tags --progress -- https://github.com/dmlc/dgl.git +refs/pull/4648/head:refs/remotes/origin/PR-4648 +refs/heads/master:refs/remotes/origin/master # timeout=10
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git remote # timeout=10
 > git config --get remote.origin.url # timeout=10
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git merge d78a3a4baf611b90871a849f58647160c7cd9ab4 # timeout=10
 > git rev-parse HEAD^{commit} # timeout=10
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git branch -a -v --no-abbrev # timeout=10
 > git branch -D PR-4648 # timeout=10
 > git checkout -b PR-4648 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git rev-parse --verify HEAD # timeout=10
Resetting working tree
 > git reset --hard # timeout=10
Commit message: "fix for pytorch < 1.12"
Cleaning workspace
Commit message: "fix for pytorch < 1.12"
Cleaning workspace
 > git rev-parse --resolve-git-dir /root/jenkins/workspace/dgl_PR-4648/.git # timeout=10
 > git config remote.origin.url https://github.com/dmlc/dgl.git # timeout=10
 > git rev-parse --verify HEAD # timeout=10
Resetting working tree
 > git reset --hard # timeout=10
 > git clean -fdx # timeout=10
Fetching upstream changes from https://github.com/dmlc/dgl.git
 > git --version # timeout=10
 > git --version # 'git version 2.17.1'
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git fetch --no-tags --progress -- https://github.com/dmlc/dgl.git +refs/pull/4648/head:refs/remotes/origin/PR-4648 +refs/heads/master:refs/remotes/origin/master # timeout=10
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git remote # timeout=10
 > git config --get remote.origin.url # timeout=10
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git merge d78a3a4baf611b90871a849f58647160c7cd9ab4 # timeout=10
 > git rev-parse HEAD^{commit} # timeout=10
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git branch -a -v --no-abbrev # timeout=10
 > git branch -D PR-4648 # timeout=10
 > git checkout -b PR-4648 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git rev-parse --verify HEAD # timeout=10
Resetting working tree
 > git reset --hard # timeout=10
 > git rev-parse --resolve-git-dir /root/jenkins/workspace/dgl_PR-4648@3/.git # timeout=10
 > git config remote.origin.url https://github.com/dmlc/dgl.git # timeout=10
 > git rev-parse --verify HEAD # timeout=10
Resetting working tree
 > git reset --hard # timeout=10
 > git clean -fdx # timeout=10
Fetching upstream changes from https://github.com/dmlc/dgl.git
 > git --version # timeout=10
 > git --version # 'git version 2.17.1'
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git fetch --no-tags --progress -- https://github.com/dmlc/dgl.git +refs/pull/4648/head:refs/remotes/origin/PR-4648 +refs/heads/master:refs/remotes/origin/master # timeout=10
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git remote # timeout=10
 > git config --get remote.origin.url # timeout=10
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git merge d78a3a4baf611b90871a849f58647160c7cd9ab4 # timeout=10
 > git rev-parse HEAD^{commit} # timeout=10
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git branch -a -v --no-abbrev # timeout=10
 > git branch -D PR-4648 # timeout=10
 > git checkout -b PR-4648 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git rev-parse --verify HEAD # timeout=10
Resetting working tree
 > git reset --hard # timeout=10
 > git clean -fdx # timeout=10
[Pipeline] sh
[Pipeline] sh
[Pipeline] sh
+ git submodule update --recursive --init
+ git submodule update --recursive --init
Submodule 'third_party/METIS' (https://github.com/KarypisLab/METIS.git) registered for path 'third_party/METIS'
Submodule 'third_party/dlpack' (https://github.com/dmlc/dlpack.git) registered for path 'third_party/dlpack'
Submodule 'third_party/dmlc-core' (https://github.com/dmlc/dmlc-core.git) registered for path 'third_party/dmlc-core'
Submodule 'third_party/googletest' (https://github.com/google/googletest.git) registered for path 'third_party/googletest'
Submodule 'third_party/libxsmm' (https://github.com/hfp/libxsmm.git) registered for path 'third_party/libxsmm'
Submodule 'third_party/nanoflann' (https://github.com/jlblancoc/nanoflann) registered for path 'third_party/nanoflann'
Submodule 'third_party/nccl' (https://github.com/nvidia/nccl) registered for path 'third_party/nccl'
Submodule 'third_party/phmap' (https://github.com/greg7mdp/parallel-hashmap.git) registered for path 'third_party/phmap'
Submodule 'third_party/tensorpipe' (https://github.com/pytorch/tensorpipe) registered for path 'third_party/tensorpipe'
Submodule 'third_party/thrust' (https://github.com/NVIDIA/thrust.git) registered for path 'third_party/thrust'
Submodule 'third_party/tvm' (https://github.com/apache/incubator-tvm) registered for path 'third_party/tvm'
Submodule 'third_party/xbyak' (https://github.com/herumi/xbyak) registered for path 'third_party/xbyak'
Cloning into '/root/jenkins/workspace/dgl_PR-4648@2/third_party/METIS'...
 > git clean -fdx # timeout=10
 > git clean -fdx # timeout=10
Submodule 'third_party/METIS' (https://github.com/KarypisLab/METIS.git) registered for path 'third_party/METIS'
Submodule 'third_party/dlpack' (https://github.com/dmlc/dlpack.git) registered for path 'third_party/dlpack'
Submodule 'third_party/dmlc-core' (https://github.com/dmlc/dmlc-core.git) registered for path 'third_party/dmlc-core'
Submodule 'third_party/googletest' (https://github.com/google/googletest.git) registered for path 'third_party/googletest'
Submodule 'third_party/libxsmm' (https://github.com/hfp/libxsmm.git) registered for path 'third_party/libxsmm'
Submodule 'third_party/nanoflann' (https://github.com/jlblancoc/nanoflann) registered for path 'third_party/nanoflann'
Submodule 'third_party/nccl' (https://github.com/nvidia/nccl) registered for path 'third_party/nccl'
Submodule 'third_party/phmap' (https://github.com/greg7mdp/parallel-hashmap.git) registered for path 'third_party/phmap'
Submodule 'third_party/tensorpipe' (https://github.com/pytorch/tensorpipe) registered for path 'third_party/tensorpipe'
Submodule 'third_party/thrust' (https://github.com/NVIDIA/thrust.git) registered for path 'third_party/thrust'
Submodule 'third_party/tvm' (https://github.com/apache/incubator-tvm) registered for path 'third_party/tvm'
Submodule 'third_party/xbyak' (https://github.com/herumi/xbyak) registered for path 'third_party/xbyak'
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/METIS'...
+ git submodule update --recursive --init
Submodule 'third_party/METIS' (https://github.com/KarypisLab/METIS.git) registered for path 'third_party/METIS'
Submodule 'third_party/dlpack' (https://github.com/dmlc/dlpack.git) registered for path 'third_party/dlpack'
Submodule 'third_party/dmlc-core' (https://github.com/dmlc/dmlc-core.git) registered for path 'third_party/dmlc-core'
Submodule 'third_party/googletest' (https://github.com/google/googletest.git) registered for path 'third_party/googletest'
Submodule 'third_party/libxsmm' (https://github.com/hfp/libxsmm.git) registered for path 'third_party/libxsmm'
Submodule 'third_party/nanoflann' (https://github.com/jlblancoc/nanoflann) registered for path 'third_party/nanoflann'
Submodule 'third_party/nccl' (https://github.com/nvidia/nccl) registered for path 'third_party/nccl'
Submodule 'third_party/phmap' (https://github.com/greg7mdp/parallel-hashmap.git) registered for path 'third_party/phmap'
Submodule 'third_party/tensorpipe' (https://github.com/pytorch/tensorpipe) registered for path 'third_party/tensorpipe'
Submodule 'third_party/thrust' (https://github.com/NVIDIA/thrust.git) registered for path 'third_party/thrust'
Submodule 'third_party/tvm' (https://github.com/apache/incubator-tvm) registered for path 'third_party/tvm'
Submodule 'third_party/xbyak' (https://github.com/herumi/xbyak) registered for path 'third_party/xbyak'
Cloning into '/root/jenkins/workspace/dgl_PR-4648@3/third_party/METIS'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648@2/third_party/dlpack'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/dlpack'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648@3/third_party/dlpack'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648@2/third_party/dmlc-core'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/dmlc-core'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648@3/third_party/dmlc-core'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648@2/third_party/googletest'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/googletest'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648@3/third_party/googletest'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/libxsmm'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648@2/third_party/libxsmm'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648@3/third_party/libxsmm'...
Cloning into 'C:/Jenkins/workspace/dgl_PR-4648/third_party/dlpack'...
Cloning into 'C:/Jenkins/workspace/dgl_PR-4648/third_party/dmlc-core'...
Cloning into 'C:/Jenkins/workspace/dgl_PR-4648/third_party/nccl'...
Cloning into 'C:/Jenkins/workspace/dgl_PR-4648/third_party/nanoflann'...
Cloning into 'C:/Jenkins/workspace/dgl_PR-4648/third_party/googletest'...
Cloning into 'C:/Jenkins/workspace/dgl_PR-4648/third_party/phmap'...
Cloning into 'C:/Jenkins/workspace/dgl_PR-4648/third_party/tensorpipe'...
Cloning into 'C:/Jenkins/workspace/dgl_PR-4648/third_party/xbyak'...
Cloning into 'C:/Jenkins/workspace/dgl_PR-4648/third_party/libxsmm'...
Cloning into 'C:/Jenkins/workspace/dgl_PR-4648/third_party/thrust'...
Cloning into 'C:/Jenkins/workspace/dgl_PR-4648/third_party/tvm'...
Submodule path 'third_party/METIS': checked out '10603482cc055626ff99d2ac3ab7e73d1119439d'
Submodule 'GKlib' (https://github.com/KarypisLab/GKlib.git) registered for path 'third_party/METIS/GKlib'
Cloning into 'C:/Jenkins/workspace/dgl_PR-4648/third_party/METIS/GKlib'...
Submodule path 'third_party/METIS/GKlib': checked out '62de20c7f97c719abdc96d7c3e266a1f7cb52bc5'
Submodule path 'third_party/dlpack': checked out 'e2bdd3bee8cb6501558042633fa59144cc8b7f5f'
Submodule path 'third_party/dmlc-core': checked out 'bfad207b448480783a1f428ae3d93d87032d8349'
Submodule path 'third_party/googletest': checked out 'f71fb4f9a912ec945401cc49a287a759b6131026'
Cloning into '/root/jenkins/workspace/dgl_PR-4648@3/third_party/nanoflann'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648@3/third_party/nccl'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648@3/third_party/phmap'...
Submodule path 'third_party/libxsmm': checked out 'fa687556130b6298430f1c0555a14cf79ab6101c'
Submodule path 'third_party/nanoflann': checked out '4c47ca200209550c5628c89803591f8a753c8181'
Cloning into '/root/jenkins/workspace/dgl_PR-4648@2/third_party/nanoflann'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648@2/third_party/nccl'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648@2/third_party/phmap'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648@2/third_party/tensorpipe'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648@3/third_party/tensorpipe'...
Submodule path 'third_party/nccl': checked out 'e11238b3029795d33f958b5868d47c90c4f22628'
Cloning into '/root/jenkins/workspace/dgl_PR-4648@2/third_party/thrust'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648@3/third_party/thrust'...
Submodule path 'third_party/phmap': checked out '25293cefd8b85491b45600c03fe8edf07647553f'
Submodule path 'third_party/tensorpipe': checked out '6042f1a4cbce8eef997f11ed0012de137b317361'
Submodule 'third_party/googletest' (https://github.com/google/googletest.git) registered for path 'third_party/tensorpipe/third_party/googletest'
Submodule 'third_party/libnop' (https://github.com/google/libnop.git) registered for path 'third_party/tensorpipe/third_party/libnop'
Submodule 'third_party/libuv' (https://github.com/libuv/libuv.git) registered for path 'third_party/tensorpipe/third_party/libuv'
Submodule 'third_party/pybind11' (https://github.com/pybind/pybind11.git) registered for path 'third_party/tensorpipe/third_party/pybind11'
Cloning into '/root/jenkins/workspace/dgl_PR-4648@2/third_party/tvm'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648@3/third_party/tvm'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/nanoflann'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/nccl'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/phmap'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tensorpipe'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/thrust'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tvm'...
Cloning into 'C:/Jenkins/workspace/dgl_PR-4648/third_party/tensorpipe/third_party/googletest'...
Cloning into 'C:/Jenkins/workspace/dgl_PR-4648/third_party/tensorpipe/third_party/libnop'...
Cloning into 'C:/Jenkins/workspace/dgl_PR-4648/third_party/tensorpipe/third_party/pybind11'...
Cloning into 'C:/Jenkins/workspace/dgl_PR-4648/third_party/tensorpipe/third_party/libuv'...
Submodule path 'third_party/tensorpipe/third_party/googletest': checked out 'aee0f9d9b5b87796ee8a0ab26b7587ec30e8858e'
Submodule path 'third_party/tensorpipe/third_party/libnop': checked out 'aa95422ea8c409e3f078d2ee7708a5f59a8b9fa2'
Submodule path 'third_party/tensorpipe/third_party/libuv': checked out '1dff88e5161cba5c59276d2070d2e304e4dcb242'
Submodule path 'third_party/tensorpipe/third_party/pybind11': checked out 'a23996fce38ff6ccfbcdc09f1e63f2c4be5ea2ef'
Submodule 'tools/clang' (https://github.com/wjakob/clang-cindex-python3) registered for path 'third_party/tensorpipe/third_party/pybind11/tools/clang'
Cloning into '/root/jenkins/workspace/dgl_PR-4648@2/third_party/xbyak'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648@3/third_party/xbyak'...
Cloning into 'C:/Jenkins/workspace/dgl_PR-4648/third_party/tensorpipe/third_party/pybind11/tools/clang'...
Submodule path 'third_party/METIS': checked out '10603482cc055626ff99d2ac3ab7e73d1119439d'
Submodule 'GKlib' (https://github.com/KarypisLab/GKlib.git) registered for path 'third_party/METIS/GKlib'
Cloning into '/root/jenkins/workspace/dgl_PR-4648@2/third_party/METIS/GKlib'...
Submodule path 'third_party/METIS': checked out '10603482cc055626ff99d2ac3ab7e73d1119439d'
Submodule 'GKlib' (https://github.com/KarypisLab/GKlib.git) registered for path 'third_party/METIS/GKlib'
Cloning into '/root/jenkins/workspace/dgl_PR-4648@3/third_party/METIS/GKlib'...
Submodule path 'third_party/tensorpipe/third_party/pybind11/tools/clang': checked out '6a00cbc4a9b8e68b71caf7f774b3f9c753ae84d5'
Submodule path 'third_party/METIS/GKlib': checked out '62de20c7f97c719abdc96d7c3e266a1f7cb52bc5'
Submodule path 'third_party/dlpack': checked out 'e2bdd3bee8cb6501558042633fa59144cc8b7f5f'
Submodule path 'third_party/METIS/GKlib': checked out '62de20c7f97c719abdc96d7c3e266a1f7cb52bc5'
Submodule path 'third_party/dlpack': checked out 'e2bdd3bee8cb6501558042633fa59144cc8b7f5f'
Submodule path 'third_party/dmlc-core': checked out 'bfad207b448480783a1f428ae3d93d87032d8349'
Submodule path 'third_party/thrust': checked out '6a3078c64cab0e2f276340fa5dcafa0d758ed890'
Submodule 'cub' (https://github.com/NVIDIA/cub.git) registered for path 'third_party/thrust/dependencies/cub'
Submodule path 'third_party/dmlc-core': checked out 'bfad207b448480783a1f428ae3d93d87032d8349'
Submodule path 'third_party/googletest': checked out 'f71fb4f9a912ec945401cc49a287a759b6131026'
Submodule path 'third_party/googletest': checked out 'f71fb4f9a912ec945401cc49a287a759b6131026'
Submodule path 'third_party/libxsmm': checked out 'fa687556130b6298430f1c0555a14cf79ab6101c'
Submodule path 'third_party/libxsmm': checked out 'fa687556130b6298430f1c0555a14cf79ab6101c'
Submodule path 'third_party/nanoflann': checked out '4c47ca200209550c5628c89803591f8a753c8181'
Submodule path 'third_party/nccl': checked out 'e11238b3029795d33f958b5868d47c90c4f22628'
Submodule path 'third_party/nanoflann': checked out '4c47ca200209550c5628c89803591f8a753c8181'
Submodule path 'third_party/nccl': checked out 'e11238b3029795d33f958b5868d47c90c4f22628'
Submodule path 'third_party/phmap': checked out '25293cefd8b85491b45600c03fe8edf07647553f'
Submodule path 'third_party/phmap': checked out '25293cefd8b85491b45600c03fe8edf07647553f'
Submodule path 'third_party/tensorpipe': checked out '6042f1a4cbce8eef997f11ed0012de137b317361'
Submodule 'third_party/googletest' (https://github.com/google/googletest.git) registered for path 'third_party/tensorpipe/third_party/googletest'
Submodule 'third_party/libnop' (https://github.com/google/libnop.git) registered for path 'third_party/tensorpipe/third_party/libnop'
Submodule 'third_party/libuv' (https://github.com/libuv/libuv.git) registered for path 'third_party/tensorpipe/third_party/libuv'
Submodule 'third_party/pybind11' (https://github.com/pybind/pybind11.git) registered for path 'third_party/tensorpipe/third_party/pybind11'
Submodule path 'third_party/tensorpipe': checked out '6042f1a4cbce8eef997f11ed0012de137b317361'
Submodule 'third_party/googletest' (https://github.com/google/googletest.git) registered for path 'third_party/tensorpipe/third_party/googletest'
Submodule 'third_party/libnop' (https://github.com/google/libnop.git) registered for path 'third_party/tensorpipe/third_party/libnop'
Submodule 'third_party/libuv' (https://github.com/libuv/libuv.git) registered for path 'third_party/tensorpipe/third_party/libuv'
Submodule 'third_party/pybind11' (https://github.com/pybind/pybind11.git) registered for path 'third_party/tensorpipe/third_party/pybind11'
Cloning into '/root/jenkins/workspace/dgl_PR-4648@3/third_party/tensorpipe/third_party/googletest'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648@2/third_party/tensorpipe/third_party/googletest'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/xbyak'...
Submodule path 'third_party/METIS': checked out '10603482cc055626ff99d2ac3ab7e73d1119439d'
Submodule 'GKlib' (https://github.com/KarypisLab/GKlib.git) registered for path 'third_party/METIS/GKlib'
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/METIS/GKlib'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648@3/third_party/tensorpipe/third_party/libnop'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648@2/third_party/tensorpipe/third_party/libnop'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648@3/third_party/tensorpipe/third_party/libuv'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648@2/third_party/tensorpipe/third_party/libuv'...
Submodule path 'third_party/METIS/GKlib': checked out '62de20c7f97c719abdc96d7c3e266a1f7cb52bc5'
Submodule path 'third_party/dlpack': checked out 'e2bdd3bee8cb6501558042633fa59144cc8b7f5f'
Submodule path 'third_party/dmlc-core': checked out 'bfad207b448480783a1f428ae3d93d87032d8349'
Submodule path 'third_party/googletest': checked out 'f71fb4f9a912ec945401cc49a287a759b6131026'
Submodule path 'third_party/libxsmm': checked out 'fa687556130b6298430f1c0555a14cf79ab6101c'
Submodule path 'third_party/nanoflann': checked out '4c47ca200209550c5628c89803591f8a753c8181'
Submodule path 'third_party/nccl': checked out 'e11238b3029795d33f958b5868d47c90c4f22628'
Submodule path 'third_party/phmap': checked out '25293cefd8b85491b45600c03fe8edf07647553f'
Cloning into '/root/jenkins/workspace/dgl_PR-4648@3/third_party/tensorpipe/third_party/pybind11'...
Submodule path 'third_party/tensorpipe': checked out '6042f1a4cbce8eef997f11ed0012de137b317361'
Submodule 'third_party/googletest' (https://github.com/google/googletest.git) registered for path 'third_party/tensorpipe/third_party/googletest'
Submodule 'third_party/libnop' (https://github.com/google/libnop.git) registered for path 'third_party/tensorpipe/third_party/libnop'
Submodule 'third_party/libuv' (https://github.com/libuv/libuv.git) registered for path 'third_party/tensorpipe/third_party/libuv'
Submodule 'third_party/pybind11' (https://github.com/pybind/pybind11.git) registered for path 'third_party/tensorpipe/third_party/pybind11'
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tensorpipe/third_party/googletest'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648@2/third_party/tensorpipe/third_party/pybind11'...
Cloning into 'C:/Jenkins/workspace/dgl_PR-4648/third_party/thrust/dependencies/cub'...
Submodule path 'third_party/thrust/dependencies/cub': checked out 'cdaa9558a85e45d849016e5fe7b6e4ee79113f95'
Submodule path 'third_party/tensorpipe/third_party/googletest': checked out 'aee0f9d9b5b87796ee8a0ab26b7587ec30e8858e'
Submodule path 'third_party/tensorpipe/third_party/libnop': checked out 'aa95422ea8c409e3f078d2ee7708a5f59a8b9fa2'
Submodule path 'third_party/tensorpipe/third_party/libuv': checked out '1dff88e5161cba5c59276d2070d2e304e4dcb242'
Submodule path 'third_party/tensorpipe/third_party/pybind11': checked out 'a23996fce38ff6ccfbcdc09f1e63f2c4be5ea2ef'
Submodule 'tools/clang' (https://github.com/wjakob/clang-cindex-python3) registered for path 'third_party/tensorpipe/third_party/pybind11/tools/clang'
Cloning into '/root/jenkins/workspace/dgl_PR-4648@2/third_party/tensorpipe/third_party/pybind11/tools/clang'...
Submodule path 'third_party/tensorpipe/third_party/googletest': checked out 'aee0f9d9b5b87796ee8a0ab26b7587ec30e8858e'
Submodule path 'third_party/tensorpipe/third_party/libnop': checked out 'aa95422ea8c409e3f078d2ee7708a5f59a8b9fa2'
Submodule path 'third_party/tensorpipe/third_party/libuv': checked out '1dff88e5161cba5c59276d2070d2e304e4dcb242'
Submodule path 'third_party/tensorpipe/third_party/pybind11': checked out 'a23996fce38ff6ccfbcdc09f1e63f2c4be5ea2ef'
Submodule 'tools/clang' (https://github.com/wjakob/clang-cindex-python3) registered for path 'third_party/tensorpipe/third_party/pybind11/tools/clang'
Cloning into '/root/jenkins/workspace/dgl_PR-4648@3/third_party/tensorpipe/third_party/pybind11/tools/clang'...
Submodule path 'third_party/tensorpipe/third_party/pybind11/tools/clang': checked out '6a00cbc4a9b8e68b71caf7f774b3f9c753ae84d5'
Submodule path 'third_party/thrust': checked out '6a3078c64cab0e2f276340fa5dcafa0d758ed890'
Submodule 'cub' (https://github.com/NVIDIA/cub.git) registered for path 'third_party/thrust/dependencies/cub'
Cloning into '/root/jenkins/workspace/dgl_PR-4648@2/third_party/thrust/dependencies/cub'...
Submodule path 'third_party/tensorpipe/third_party/pybind11/tools/clang': checked out '6a00cbc4a9b8e68b71caf7f774b3f9c753ae84d5'
Submodule path 'third_party/thrust': checked out '6a3078c64cab0e2f276340fa5dcafa0d758ed890'
Submodule 'cub' (https://github.com/NVIDIA/cub.git) registered for path 'third_party/thrust/dependencies/cub'
Cloning into '/root/jenkins/workspace/dgl_PR-4648@3/third_party/thrust/dependencies/cub'...
Submodule path 'third_party/tvm': checked out 'b2e418cb109df4cd1f17a2cf2894a1b396a6b838'
Submodule 'dlpack' (https://github.com/dmlc/dlpack) registered for path 'third_party/tvm/3rdparty/dlpack'
Submodule 'dmlc-core' (https://github.com/dmlc/dmlc-core) registered for path 'third_party/tvm/3rdparty/dmlc-core'
Submodule '3rdparty/rang' (https://github.com/agauniyal/rang) registered for path 'third_party/tvm/3rdparty/rang'
Submodule '3rdparty/vta-hw' (https://github.com/apache/incubator-tvm-vta) registered for path 'third_party/tvm/3rdparty/vta-hw'
Cloning into 'C:/Jenkins/workspace/dgl_PR-4648/third_party/tvm/3rdparty/dlpack'...
Cloning into 'C:/Jenkins/workspace/dgl_PR-4648/third_party/tvm/3rdparty/rang'...
Cloning into 'C:/Jenkins/workspace/dgl_PR-4648/third_party/tvm/3rdparty/dmlc-core'...
Cloning into 'C:/Jenkins/workspace/dgl_PR-4648/third_party/tvm/3rdparty/vta-hw'...
Submodule path 'third_party/tvm/3rdparty/dlpack': checked out '3ec04430e89a6834e5a1b99471f415fa939bf642'
Submodule path 'third_party/tvm/3rdparty/dmlc-core': checked out '6c401e242c59a1f4c913918246591bb13fd714e7'
Submodule path 'third_party/tvm/3rdparty/rang': checked out 'cabe04d6d6b05356fa8f9741704924788f0dd762'
Submodule path 'third_party/thrust/dependencies/cub': checked out 'cdaa9558a85e45d849016e5fe7b6e4ee79113f95'
Submodule path 'third_party/tvm/3rdparty/vta-hw': checked out '87ce9acfae550d1a487746e9d06c2e250076e54c'
Submodule path 'third_party/xbyak': checked out '757e4063f6464740b8ff4a2cae9136d2f8458020'
[Pipeline] bat
Submodule path 'third_party/thrust/dependencies/cub': checked out 'cdaa9558a85e45d849016e5fe7b6e4ee79113f95'
Submodule path 'third_party/tvm': checked out 'b2e418cb109df4cd1f17a2cf2894a1b396a6b838'
Submodule 'dlpack' (https://github.com/dmlc/dlpack) registered for path 'third_party/tvm/3rdparty/dlpack'
Submodule 'dmlc-core' (https://github.com/dmlc/dmlc-core) registered for path 'third_party/tvm/3rdparty/dmlc-core'
Submodule '3rdparty/rang' (https://github.com/agauniyal/rang) registered for path 'third_party/tvm/3rdparty/rang'
Submodule '3rdparty/vta-hw' (https://github.com/apache/incubator-tvm-vta) registered for path 'third_party/tvm/3rdparty/vta-hw'
Cloning into '/root/jenkins/workspace/dgl_PR-4648@2/third_party/tvm/3rdparty/dlpack'...
Submodule path 'third_party/tvm': checked out 'b2e418cb109df4cd1f17a2cf2894a1b396a6b838'
Submodule 'dlpack' (https://github.com/dmlc/dlpack) registered for path 'third_party/tvm/3rdparty/dlpack'
Submodule 'dmlc-core' (https://github.com/dmlc/dmlc-core) registered for path 'third_party/tvm/3rdparty/dmlc-core'
Submodule '3rdparty/rang' (https://github.com/agauniyal/rang) registered for path 'third_party/tvm/3rdparty/rang'
Submodule '3rdparty/vta-hw' (https://github.com/apache/incubator-tvm-vta) registered for path 'third_party/tvm/3rdparty/vta-hw'
Cloning into '/root/jenkins/workspace/dgl_PR-4648@3/third_party/tvm/3rdparty/dlpack'...

administrator@WIN-O7QS55HVSDB C:\Jenkins\workspace\dgl_PR-4648>CALL tests\scripts\build_dgl.bat 
**********************************************************************
** Visual Studio 2019 Developer Command Prompt v16.5.5
** Copyright (c) 2019 Microsoft Corporation
**********************************************************************
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tensorpipe/third_party/libnop'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tensorpipe/third_party/libuv'...
[vcvarsall.bat] Environment initialized for: 'x64'
Cloning into '/root/jenkins/workspace/dgl_PR-4648@2/third_party/tvm/3rdparty/dmlc-core'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648@3/third_party/tvm/3rdparty/dmlc-core'...
created virtual environment CPython3.6.7.final.0-64 in 345ms
  creator CPython3Windows(dest=C:\Users\Administrator\Envs\jenkins-dgl-PR-4648-2, clear=False, global=True)
  seeder FromAppData(download=False, pip=bundle, setuptools=bundle, wheel=bundle, via=copy, app_data_dir=C:\Users\Administrator\AppData\Local\pypa\virtualenv)
    added seed packages: pip==21.3.1, setuptools==59.6.0, wheel==0.37.1
  activators BashActivator,BatchActivator,FishActivator,PowerShellActivator,PythonActivator,XonshActivator
Deleted file - C:\Jenkins\workspace\dgl_PR-4648\third_party\libxsmm\BUILD
Deleted file - C:\Jenkins\workspace\dgl_PR-4648\third_party\libxsmm\samples\hello\BUILD
Cloning into '/root/jenkins/workspace/dgl_PR-4648@2/third_party/tvm/3rdparty/rang'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648@3/third_party/tvm/3rdparty/rang'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648@2/third_party/tvm/3rdparty/vta-hw'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648@3/third_party/tvm/3rdparty/vta-hw'...
Could Not Find C:\Jenkins\workspace\dgl_PR-4648\_download
-- Selecting Windows SDK version 10.0.18362.0 to target Windows 6.3.9600.
Submodule path 'third_party/tvm/3rdparty/dlpack': checked out '3ec04430e89a6834e5a1b99471f415fa939bf642'
Submodule path 'third_party/tvm/3rdparty/dmlc-core': checked out '6c401e242c59a1f4c913918246591bb13fd714e7'
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tensorpipe/third_party/pybind11'...
Submodule path 'third_party/tvm/3rdparty/dlpack': checked out '3ec04430e89a6834e5a1b99471f415fa939bf642'
Submodule path 'third_party/tvm/3rdparty/dmlc-core': checked out '6c401e242c59a1f4c913918246591bb13fd714e7'
Submodule path 'third_party/tvm/3rdparty/rang': checked out 'cabe04d6d6b05356fa8f9741704924788f0dd762'
Submodule path 'third_party/tvm/3rdparty/vta-hw': checked out '87ce9acfae550d1a487746e9d06c2e250076e54c'
Submodule path 'third_party/tvm/3rdparty/rang': checked out 'cabe04d6d6b05356fa8f9741704924788f0dd762'
Submodule path 'third_party/tvm/3rdparty/vta-hw': checked out '87ce9acfae550d1a487746e9d06c2e250076e54c'
Submodule path 'third_party/xbyak': checked out '757e4063f6464740b8ff4a2cae9136d2f8458020'
Submodule path 'third_party/xbyak': checked out '757e4063f6464740b8ff4a2cae9136d2f8458020'
[Pipeline] sh
[Pipeline] sh
+ bash tests/scripts/build_dgl.sh cugraph
+ bash tests/scripts/build_dgl.sh gpu
~/jenkins/workspace/dgl_PR-4648@2/build ~/jenkins/workspace/dgl_PR-4648@2
-- The C compiler identification is GNU 7.5.0
-- The CXX compiler identification is GNU 7.5.0
-- Check for working C compiler: /usr/bin/cc
-- The C compiler identification is MSVC 19.25.28614.0
~/jenkins/workspace/dgl_PR-4648@3/build ~/jenkins/workspace/dgl_PR-4648@3
-- The C compiler identification is GNU 7.5.0
Submodule path 'third_party/tensorpipe/third_party/googletest': checked out 'aee0f9d9b5b87796ee8a0ab26b7587ec30e8858e'
Submodule path 'third_party/tensorpipe/third_party/libnop': checked out 'aa95422ea8c409e3f078d2ee7708a5f59a8b9fa2'
-- Check for working C compiler: /usr/bin/cc -- works
-- Detecting C compiler ABI info
-- Detecting C compiler ABI info - done
-- The CXX compiler identification is MSVC 19.25.28614.0
-- The CXX compiler identification is GNU 7.5.0
-- Detecting C compiler ABI info
-- Detecting C compiler ABI info - done
-- Check for working C compiler: /usr/bin/cc - skipped
-- Detecting C compile features
-- Detecting C compile features - done
-- Detecting CXX compiler ABI info
Submodule path 'third_party/tensorpipe/third_party/libuv': checked out '1dff88e5161cba5c59276d2070d2e304e4dcb242'
-- Detecting C compile features
-- Detecting C compile features - done
-- Check for working CXX compiler: /usr/bin/c++
-- Check for working CXX compiler: /usr/bin/c++ -- works
-- Detecting CXX compiler ABI info
-- Detecting C compiler ABI info
-- Detecting CXX compiler ABI info - done
-- Check for working CXX compiler: /usr/bin/c++ - skipped
-- Detecting CXX compile features
-- Detecting CXX compile features - done
-- Start configuring project dgl
-- Build with CUDA support
-- Performing Test CMAKE_HAVE_LIBC_PTHREAD
Submodule path 'third_party/tensorpipe/third_party/pybind11': checked out 'a23996fce38ff6ccfbcdc09f1e63f2c4be5ea2ef'
Submodule 'tools/clang' (https://github.com/wjakob/clang-cindex-python3) registered for path 'third_party/tensorpipe/third_party/pybind11/tools/clang'
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tensorpipe/third_party/pybind11/tools/clang'...
-- Detecting CXX compiler ABI info - done
-- Detecting CXX compile features
-- Detecting CXX compile features - done
-- Start configuring project dgl
-- Build with CUDA support
-- Looking for pthread.h
-- Looking for pthread.h - found
-- Performing Test CMAKE_HAVE_LIBC_PTHREAD
-- Performing Test CMAKE_HAVE_LIBC_PTHREAD - Failed
-- Looking for pthread_create in pthreads
-- Looking for pthread_create in pthreads - not found
-- Looking for pthread_create in pthread
-- Performing Test CMAKE_HAVE_LIBC_PTHREAD - Failed
-- Looking for pthread_create in pthreads
-- Looking for pthread_create in pthreads - not found
-- Looking for pthread_create in pthread
-- Detecting C compiler ABI info - done
-- Check for working C compiler: C:/Program Files (x86)/Microsoft Visual Studio/2019/BuildTools/VC/Tools/MSVC/14.25.28610/bin/Hostx64/x64/cl.exe - skipped
-- Detecting C compile features
-- Detecting C compile features - done
-- Detecting CXX compiler ABI info
-- Looking for pthread_create in pthread - found
-- Found Threads: TRUE  
-- Found CUDA: /usr/local/cuda (found version "11.5") 
-- Found CUDA_TOOLKIT_ROOT_DIR=/usr/local/cuda
-- Found CUDA_CUDART_LIBRARY=/usr/local/cuda/lib64/libcudart.so
-- Found CUDA_CUBLAS_LIBRARY=/opt/conda/lib/libcublas.so
-- Found CUDA_CURAND_LIBRARY=/opt/conda/lib/libcurand.so
-- Performing Test SUPPORT_CXX14
-- Looking for pthread_create in pthread - found
-- Found Threads: TRUE  
-- Found CUDA: /usr/local/cuda (found version "10.1") 
-- Found CUDA_TOOLKIT_ROOT_DIR=/usr/local/cuda
-- Found CUDA_CUDART_LIBRARY=/usr/local/cuda/lib64/libcudart.so
-- Found CUDA_CUBLAS_LIBRARY=/usr/lib/x86_64-linux-gnu/libcublas.so
-- Found CUDA_CURAND_LIBRARY=/usr/local/cuda/lib64/libcurand.so
-- Performing Test SUPPORT_CXX14
-- Performing Test SUPPORT_CXX14 - Success
-- Use external CUB/Thrust library for a consistent API and performance.
Submodule path 'third_party/tensorpipe/third_party/pybind11/tools/clang': checked out '6a00cbc4a9b8e68b71caf7f774b3f9c753ae84d5'
Submodule path 'third_party/thrust': checked out '6a3078c64cab0e2f276340fa5dcafa0d758ed890'
Submodule 'cub' (https://github.com/NVIDIA/cub.git) registered for path 'third_party/thrust/dependencies/cub'
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/thrust/dependencies/cub'...
-- Performing Test SUPPORT_CXX14 - Success
-- Use external CUB/Thrust library for a consistent API and performance.
-- Detecting CXX compiler ABI info - done
-- Check for working CXX compiler: C:/Program Files (x86)/Microsoft Visual Studio/2019/BuildTools/VC/Tools/MSVC/14.25.28610/bin/Hostx64/x64/cl.exe - skipped
-- Detecting CXX compile features
-- Detecting CXX compile features - done
-- Start configuring project dgl
-- Found OpenMP_C: -fopenmp (found version "4.5") 
-- Found OpenMP_CXX: -fopenmp (found version "4.5") 
-- Found OpenMP: TRUE (found version "4.5")  
-- Build with OpenMP.
-- Build with LIBXSMM optimization.
-- Looking for sys/epoll.h
-- Found OpenMP_C: -fopenmp (found version "4.5") 
-- Found OpenMP_CXX: -fopenmp (found version "4.5") 
-- Found OpenMP: TRUE (found version "4.5")  
-- Build with OpenMP.
-- Build with LIBXSMM optimization.
-- Looking for sys/epoll.h
-- Looking for sys/epoll.h - found
-- -fopenmp -O2 -Wall -fPIC -std=c++14  -DUSE_AVX -DUSE_LIBXSMM -DDGL_CPU_LLC_SIZE=40000000 -DUSE_EPOLL -DIDXTYPEWIDTH=64 -DREALTYPEWIDTH=32
-- Running GPU architecture autodetection
-- Looking for sys/epoll.h - found
-- -fopenmp -O2 -Wall -fPIC -std=c++14  -DUSE_AVX -DUSE_LIBXSMM -DDGL_CPU_LLC_SIZE=40000000 -DUSE_EPOLL -DIDXTYPEWIDTH=64 -DREALTYPEWIDTH=32
-- Running GPU architecture autodetection
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Submodule path 'third_party/thrust/dependencies/cub': checked out 'cdaa9558a85e45d849016e5fe7b6e4ee79113f95'
CMake Warning at cmake/modules/CUDA.cmake:71 (message):
  Running GPU detection script with nvcc failed:
Call Stack (most recent call first):
  cmake/modules/CUDA.cmake:132 (dgl_detect_installed_gpus)
  cmake/modules/CUDA.cmake:266 (dgl_select_nvcc_arch_flags)
  CMakeLists.txt:170 (dgl_config_cuda)


CMake Warning at cmake/modules/CUDA.cmake:76 (message):
  Automatic GPU detection failed.  Building for all known architectures (35
  50 60 70).
Call Stack (most recent call first):
  cmake/modules/CUDA.cmake:132 (dgl_detect_installed_gpus)
  cmake/modules/CUDA.cmake:266 (dgl_select_nvcc_arch_flags)
  CMakeLists.txt:170 (dgl_config_cuda)


-- CUDA flags: -Xcompiler ,-fopenmp,-O2,-Wall,-fPIC,-std=c++14,,-DUSE_AVX,-DUSE_LIBXSMM,-DDGL_CPU_LLC_SIZE=40000000,-DUSE_EPOLL,-DIDXTYPEWIDTH=64,-DREALTYPEWIDTH=32;--expt-relaxed-constexpr;-gencode;arch=compute_35,code=sm_35;-gencode;arch=compute_50,code=sm_50;-gencode;arch=compute_60,code=sm_60;-gencode;arch=compute_70,code=sm_70;--expt-extended-lambda;-Wno-deprecated-declarations;-std=c++14
-- Found OpenMP_C: -fopenmp (found version "4.5") 
-- Found OpenMP_CXX: -fopenmp (found version "4.5") 
-- Looking for clock_gettime in rt
-- Looking for clock_gettime in rt - found
-- Looking for fopen64
-- Found OpenMP_C: -openmp (found version "2.0") 
CMake Warning at cmake/modules/CUDA.cmake:71 (message):
  Running GPU detection script with nvcc failed:
Call Stack (most recent call first):
  cmake/modules/CUDA.cmake:132 (dgl_detect_installed_gpus)
  cmake/modules/CUDA.cmake:266 (dgl_select_nvcc_arch_flags)
  CMakeLists.txt:170 (dgl_config_cuda)


CMake Warning at cmake/modules/CUDA.cmake:76 (message):
  Automatic GPU detection failed.  Building for all known architectures (35
  50 60 70 80).
Call Stack (most recent call first):
  cmake/modules/CUDA.cmake:132 (dgl_detect_installed_gpus)
  cmake/modules/CUDA.cmake:266 (dgl_select_nvcc_arch_flags)
  CMakeLists.txt:170 (dgl_config_cuda)


-- CUDA flags: -Xcompiler ,-fopenmp,-O2,-Wall,-fPIC,-std=c++14,,-DUSE_AVX,-DUSE_LIBXSMM,-DDGL_CPU_LLC_SIZE=40000000,-DUSE_EPOLL,-DIDXTYPEWIDTH=64,-DREALTYPEWIDTH=32;--expt-relaxed-constexpr;-gencode;arch=compute_35,code=sm_35;-gencode;arch=compute_50,code=sm_50;-gencode;arch=compute_60,code=sm_60;-gencode;arch=compute_70,code=sm_70;-gencode;arch=compute_80,code=sm_80;--expt-extended-lambda;-Wno-deprecated-declarations;-std=c++14
-- Found OpenMP_C: -fopenmp (found version "4.5") 
-- Found OpenMP_CXX: -fopenmp (found version "4.5") 
-- Looking for clock_gettime in rt
-- Looking for clock_gettime in rt - found
-- Looking for fopen64
-- Looking for fopen64 - not found
-- Looking for C++ include cxxabi.h
-- Looking for C++ include cxxabi.h - found
-- Looking for nanosleep
Submodule path 'third_party/tvm': checked out 'b2e418cb109df4cd1f17a2cf2894a1b396a6b838'
Submodule 'dlpack' (https://github.com/dmlc/dlpack) registered for path 'third_party/tvm/3rdparty/dlpack'
Submodule 'dmlc-core' (https://github.com/dmlc/dmlc-core) registered for path 'third_party/tvm/3rdparty/dmlc-core'
Submodule '3rdparty/rang' (https://github.com/agauniyal/rang) registered for path 'third_party/tvm/3rdparty/rang'
Submodule '3rdparty/vta-hw' (https://github.com/apache/incubator-tvm-vta) registered for path 'third_party/tvm/3rdparty/vta-hw'
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tvm/3rdparty/dlpack'...
-- Looking for fopen64 - not found
-- Looking for C++ include cxxabi.h
-- Looking for C++ include cxxabi.h - found
-- Looking for nanosleep
-- Looking for nanosleep - found
-- Looking for backtrace
-- Looking for backtrace - found
-- backtrace facility detected in default set of libraries
-- Found Backtrace: /usr/include  
-- Check if the system is big endian
-- Searching 16 bit integer
-- Looking for sys/types.h
-- Found OpenMP_CXX: -openmp (found version "2.0") 
-- Found OpenMP: TRUE (found version "2.0")  
-- Build with OpenMP.
-- Build with LIBXSMM optimization.
-- Looking for clock_gettime in rt
-- Looking for nanosleep - found
-- Looking for backtrace
-- Looking for sys/types.h - found
-- Looking for stdint.h
-- Looking for stdint.h - found
-- Looking for stddef.h
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tvm/3rdparty/dmlc-core'...
-- Looking for clock_gettime in rt - not found
-- Looking for fopen64
-- Looking for backtrace - found
-- backtrace facility detected in default set of libraries
-- Found Backtrace: /usr/include  
-- /root/jenkins/workspace/dgl_PR-4648@3/third_party/dmlc-core/cmake/build_config.h.in -> include/dmlc/build_config.h
-- Performing Test SUPPORT_MSSE2
-- Performing Test SUPPORT_MSSE2 - Success
-- Looking for execinfo.h
-- Looking for stddef.h - found
-- Check size of unsigned short
-- Check size of unsigned short - done
-- Using unsigned short
-- Looking for execinfo.h - found
-- Looking for getline
-- Looking for getline - found
-- Check if the system is big endian - little endian
-- /root/jenkins/workspace/dgl_PR-4648@2/third_party/dmlc-core/cmake/build_config.h.in -> include/dmlc/build_config.h
-- Performing Test SUPPORT_MSSE2
-- Performing Test SUPPORT_MSSE2 - Success
-- Looking for execinfo.h
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tvm/3rdparty/rang'...
-- Performing Test UV_LINT_W4
-- Performing Test UV_LINT_W4 - Failed
-- Performing Test UV_LINT_NO_UNUSED_PARAMETER_MSVC
-- Performing Test UV_LINT_NO_UNUSED_PARAMETER_MSVC - Failed
-- Performing Test UV_LINT_NO_CONDITIONAL_CONSTANT_MSVC
-- Looking for fopen64 - not found
-- Looking for C++ include cxxabi.h
-- Looking for execinfo.h - found
-- Looking for getline
-- Looking for getline - found
-- Performing Test UV_LINT_NO_CONDITIONAL_CONSTANT_MSVC - Failed
-- Performing Test UV_LINT_NO_NONSTANDARD_MSVC
-- Performing Test UV_LINT_NO_NONSTANDARD_MSVC - Failed
-- Performing Test UV_LINT_NO_NONSTANDARD_EMPTY_TU_MSVC
-- Performing Test UV_LINT_NO_NONSTANDARD_EMPTY_TU_MSVC - Failed
-- Performing Test UV_LINT_NO_NONSTANDARD_FILE_SCOPE_MSVC
-- Performing Test UV_LINT_NO_NONSTANDARD_FILE_SCOPE_MSVC - Failed
-- Performing Test UV_LINT_NO_NONSTANDARD_NONSTATIC_DLIMPORT_MSVC
-- Performing Test UV_LINT_NO_NONSTANDARD_NONSTATIC_DLIMPORT_MSVC - Failed
-- Performing Test UV_LINT_NO_HIDES_LOCAL
-- Looking for C++ include cxxabi.h - not found
-- Looking for nanosleep
-- Performing Test UV_LINT_W4
-- Performing Test UV_LINT_W4 - Failed
-- Performing Test UV_LINT_NO_UNUSED_PARAMETER_MSVC
-- Performing Test UV_LINT_NO_UNUSED_PARAMETER_MSVC - Failed
-- Performing Test UV_LINT_NO_CONDITIONAL_CONSTANT_MSVC
-- Performing Test UV_LINT_NO_CONDITIONAL_CONSTANT_MSVC - Failed
-- Performing Test UV_LINT_NO_NONSTANDARD_MSVC
-- Performing Test UV_LINT_NO_NONSTANDARD_MSVC - Failed
-- Performing Test UV_LINT_NO_NONSTANDARD_EMPTY_TU_MSVC
-- Performing Test UV_LINT_NO_NONSTANDARD_EMPTY_TU_MSVC - Failed
-- Performing Test UV_LINT_NO_NONSTANDARD_FILE_SCOPE_MSVC
-- Performing Test UV_LINT_NO_NONSTANDARD_FILE_SCOPE_MSVC - Failed
-- Performing Test UV_LINT_NO_NONSTANDARD_NONSTATIC_DLIMPORT_MSVC
-- Performing Test UV_LINT_NO_HIDES_LOCAL - Failed
-- Performing Test UV_LINT_NO_HIDES_PARAM
-- Performing Test UV_LINT_NO_HIDES_PARAM - Failed
-- Performing Test UV_LINT_NO_HIDES_GLOBAL
-- Performing Test UV_LINT_NO_HIDES_GLOBAL - Failed
-- Performing Test UV_LINT_NO_CONDITIONAL_ASSIGNMENT_MSVC
-- Performing Test UV_LINT_NO_CONDITIONAL_ASSIGNMENT_MSVC - Failed
-- Performing Test UV_LINT_NO_UNSAFE_MSVC
-- Performing Test UV_LINT_NO_UNSAFE_MSVC - Failed
-- Performing Test UV_LINT_WALL
-- Looking for nanosleep - not found
-- Looking for backtrace
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tvm/3rdparty/vta-hw'...
-- Performing Test UV_LINT_NO_NONSTANDARD_NONSTATIC_DLIMPORT_MSVC - Failed
-- Performing Test UV_LINT_NO_HIDES_LOCAL
-- Performing Test UV_LINT_NO_HIDES_LOCAL - Failed
-- Performing Test UV_LINT_NO_HIDES_PARAM
-- Performing Test UV_LINT_NO_HIDES_PARAM - Failed
-- Performing Test UV_LINT_NO_HIDES_GLOBAL
-- Performing Test UV_LINT_NO_HIDES_GLOBAL - Failed
-- Performing Test UV_LINT_NO_CONDITIONAL_ASSIGNMENT_MSVC
-- Performing Test UV_LINT_NO_CONDITIONAL_ASSIGNMENT_MSVC - Failed
-- Performing Test UV_LINT_NO_UNSAFE_MSVC
-- Performing Test UV_LINT_NO_UNSAFE_MSVC - Failed
-- Performing Test UV_LINT_WALL
-- Performing Test UV_LINT_WALL - Success
-- Performing Test UV_LINT_NO_UNUSED_PARAMETER
-- Performing Test UV_LINT_WALL - Success
-- Performing Test UV_LINT_NO_UNUSED_PARAMETER
-- Performing Test UV_LINT_NO_UNUSED_PARAMETER - Success
-- Performing Test UV_LINT_STRICT_PROTOTYPES
-- Performing Test UV_LINT_NO_UNUSED_PARAMETER - Success
-- Performing Test UV_LINT_STRICT_PROTOTYPES
-- Performing Test UV_LINT_STRICT_PROTOTYPES - Success
-- Performing Test UV_LINT_EXTRA
-- Looking for backtrace - not found
-- Could NOT find Backtrace (missing: Backtrace_LIBRARY Backtrace_INCLUDE_DIR) 
-- Check if the system is big endian
-- Searching 16 bit integer
-- Looking for sys/types.h
Submodule path 'third_party/tvm/3rdparty/dlpack': checked out '3ec04430e89a6834e5a1b99471f415fa939bf642'
-- Performing Test UV_LINT_STRICT_PROTOTYPES - Success
-- Performing Test UV_LINT_EXTRA
-- Performing Test UV_LINT_EXTRA - Success
-- Performing Test UV_LINT_UTF8_MSVC
-- Performing Test UV_LINT_UTF8_MSVC - Failed
-- Performing Test UV_F_STRICT_ALIASING
-- Performing Test UV_LINT_EXTRA - Success
-- Performing Test UV_LINT_UTF8_MSVC
-- Performing Test UV_LINT_UTF8_MSVC - Failed
-- Performing Test UV_F_STRICT_ALIASING
Submodule path 'third_party/tvm/3rdparty/dmlc-core': checked out '6c401e242c59a1f4c913918246591bb13fd714e7'
Submodule path 'third_party/tvm/3rdparty/rang': checked out 'cabe04d6d6b05356fa8f9741704924788f0dd762'
-- Performing Test UV_F_STRICT_ALIASING - Success
-- summary of build options:
    Install prefix:  /usr/local
    Target system:   Linux
    Compiler:
      C compiler:    /usr/bin/cc
      CFLAGS:         -fopenmp -O2  -fPIC  -DUSE_AVX -DUSE_LIBXSMM -DDGL_CPU_LLC_SIZE=40000000 -DUSE_EPOLL -DIDXTYPEWIDTH=64 -DREALTYPEWIDTH=32 -DLINUX -D_FILE_OFFSET_BITS=64 -std=c99 -fno-strict-aliasing -march=native -fPIC -Werror   -Wno-unused-function -Wno-unused-but-set-variable -Wno-unused-variable -Wno-unknown-pragmas -DNDEBUG -DNDEBUG2 -DHAVE_EXECINFO_H -DHAVE_GETLINE -O3

-- Found uv: 1.38.1 (found version "1.38.1") 
-- Build with unittest
-- Found PythonInterp: /usr/bin/python3.6 (found version "3.6.9") 
-- Building dist/rpc tests
-- Configuring done
-- Performing Test UV_F_STRICT_ALIASING - Success
-- summary of build options:
    Install prefix:  /usr/local
    Target system:   Linux
    Compiler:
      C compiler:    /usr/bin/cc
      CFLAGS:         -fopenmp -O2  -fPIC  -DUSE_AVX -DUSE_LIBXSMM -DDGL_CPU_LLC_SIZE=40000000 -DUSE_EPOLL -DIDXTYPEWIDTH=64 -DREALTYPEWIDTH=32 -DLINUX -D_FILE_OFFSET_BITS=64 -std=c99 -fno-strict-aliasing -march=native -fPIC -Werror   -Wno-unused-function -Wno-unused-but-set-variable -Wno-unused-variable -Wno-unknown-pragmas -DNDEBUG -DNDEBUG2 -DHAVE_EXECINFO_H -DHAVE_GETLINE -O3

-- Found uv: 1.38.1 (found version "1.38.1") 
-- Build with unittest
CMake Deprecation Warning at third_party/googletest/CMakeLists.txt:4 (cmake_minimum_required):
  Compatibility with CMake < 2.8.12 will be removed from a future version of
  CMake.

  Update the VERSION argument <min> value or use a ...<max> suffix to tell
  CMake that the project does not need compatibility with older versions.


CMake Deprecation Warning at third_party/googletest/googlemock/CMakeLists.txt:45 (cmake_minimum_required):
  Compatibility with CMake < 2.8.12 will be removed from a future version of
  CMake.

  Update the VERSION argument <min> value or use a ...<max> suffix to tell
  CMake that the project does not need compatibility with older versions.


CMake Deprecation Warning at third_party/googletest/googletest/CMakeLists.txt:56 (cmake_minimum_required):
  Compatibility with CMake < 2.8.12 will be removed from a future version of
  CMake.

  Update the VERSION argument <min> value or use a ...<max> suffix to tell
  CMake that the project does not need compatibility with older versions.


-- Found PythonInterp: /opt/conda/bin/python (found version "3.9.13") 
-- Building dist/rpc tests
-- Configuring done
CMake Warning at /opt/conda/share/cmake-3.24/Modules/FindCUDA.cmake:1985 (add_library):
  Cannot generate a safe runtime search path for target dgl because files in
  some directories may conflict with libraries in implicit directories:

    runtime library [libgomp.so.1] in /usr/lib/gcc/x86_64-linux-gnu/7 may be hidden by files in:
      /opt/conda/lib

  Some of these libraries may not be found correctly.
Call Stack (most recent call first):
  CMakeLists.txt:187 (cuda_add_library)


CMake Warning at /opt/conda/share/cmake-3.24/Modules/FindCUDA.cmake:1985 (add_library):
  Cannot generate a safe runtime search path for target dgl because there is
  a cycle in the constraint graph:

    dir 0 is [/usr/local/cuda/lib64]
      dir 1 must precede it due to runtime library [libcurand.so.10]
    dir 1 is [/opt/conda/lib]
      dir 0 must precede it due to runtime library [libcudart.so.11.0]

  Some of these libraries may not be found correctly.
Call Stack (most recent call first):
  CMakeLists.txt:187 (cuda_add_library)


CMake Warning at CMakeLists.txt:334 (add_executable):
  Cannot generate a safe runtime search path for target runUnitTests because
  files in some directories may conflict with libraries in implicit
  directories:

    runtime library [libgomp.so.1] in /usr/lib/gcc/x86_64-linux-gnu/7 may be hidden by files in:
      /opt/conda/lib

  Some of these libraries may not be found correctly.


CMake Warning at CMakeLists.txt:334 (add_executable):
  Cannot generate a safe runtime search path for target runUnitTests because
  there is a cycle in the constraint graph:

    dir 0 is [/root/jenkins/workspace/dgl_PR-4648@3/build]
    dir 1 is [/usr/local/cuda/lib64]
      dir 2 must precede it due to runtime library [libcurand.so.10]
    dir 2 is [/opt/conda/lib]
      dir 1 must precede it due to runtime library [libcudart.so.11.0]

  Some of these libraries may not be found correctly.


CMake Warning at CMakeLists.txt:342 (add_executable):
  Cannot generate a safe runtime search path for target rpc_client because
  files in some directories may conflict with libraries in implicit
  directories:

    runtime library [libgomp.so.1] in /usr/lib/gcc/x86_64-linux-gnu/7 may be hidden by files in:
      /opt/conda/lib

  Some of these libraries may not be found correctly.


CMake Warning at CMakeLists.txt:342 (add_executable):
  Cannot generate a safe runtime search path for target rpc_client because
  there is a cycle in the constraint graph:

    dir 0 is [/root/jenkins/workspace/dgl_PR-4648@3/build]
    dir 1 is [/usr/local/cuda/lib64]
      dir 2 must precede it due to runtime library [libcurand.so.10]
    dir 2 is [/opt/conda/lib]
      dir 1 must precede it due to runtime library [libcudart.so.11.0]

  Some of these libraries may not be found correctly.


CMake Warning at CMakeLists.txt:345 (add_executable):
  Cannot generate a safe runtime search path for target rpc_server because
  files in some directories may conflict with libraries in implicit
  directories:

    runtime library [libgomp.so.1] in /usr/lib/gcc/x86_64-linux-gnu/7 may be hidden by files in:
      /opt/conda/lib

  Some of these libraries may not be found correctly.


CMake Warning at CMakeLists.txt:345 (add_executable):
  Cannot generate a safe runtime search path for target rpc_server because
  there is a cycle in the constraint graph:

    dir 0 is [/root/jenkins/workspace/dgl_PR-4648@3/build]
    dir 1 is [/usr/local/cuda/lib64]
      dir 2 must precede it due to runtime library [libcurand.so.10]
    dir 2 is [/opt/conda/lib]
      dir 1 must precede it due to runtime library [libcudart.so.11.0]

  Some of these libraries may not be found correctly.


Submodule path 'third_party/tvm/3rdparty/vta-hw': checked out '87ce9acfae550d1a487746e9d06c2e250076e54c'
Submodule path 'third_party/xbyak': checked out '757e4063f6464740b8ff4a2cae9136d2f8458020'
-- Generating done
-- Build files have been written to: /root/jenkins/workspace/dgl_PR-4648@2/build
Scanning dependencies of target nccl_external
Scanning dependencies of target tensoradapter_pytorch
Scanning dependencies of target libxsmm
[  0%] Creating directories for 'nccl_external'
Scanning dependencies of target dmlc
Scanning dependencies of target gtest
[  0%] Building CXX object third_party/dmlc-core/CMakeFiles/dmlc.dir/src/config.cc.o
[  0%] Building CXX object third_party/googletest/googletest/CMakeFiles/gtest.dir/src/gtest-all.cc.o
[  0%] Building CXX object third_party/dmlc-core/CMakeFiles/dmlc.dir/src/data.cc.o
[  0%] Building CXX object third_party/dmlc-core/CMakeFiles/dmlc.dir/src/io.cc.o
[  1%] Building CXX object third_party/dmlc-core/CMakeFiles/dmlc.dir/src/io/indexed_recordio_split.cc.o
[  1%] Building CXX object third_party/dmlc-core/CMakeFiles/dmlc.dir/src/recordio.cc.o
[  2%] Building CXX object third_party/dmlc-core/CMakeFiles/dmlc.dir/src/io/input_split_base.cc.o
[  2%] Building CXX object third_party/dmlc-core/CMakeFiles/dmlc.dir/src/io/line_split.cc.o
[  2%] Building CXX object third_party/dmlc-core/CMakeFiles/dmlc.dir/src/io/recordio_split.cc.o
[  2%] Building CXX object third_party/dmlc-core/CMakeFiles/dmlc.dir/src/io/filesys.cc.o
[  2%] Building CXX object third_party/dmlc-core/CMakeFiles/dmlc.dir/src/io/local_filesys.cc.o
Scanning dependencies of target tensorpipe_uv
Scanning dependencies of target metis
[  2%] No download step for 'nccl_external'
[  3%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/fs-poll.c.o
[  3%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/inet.c.o
[  4%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/strscpy.c.o
[  4%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/random.c.o
[  4%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/idna.c.o
[  4%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/threadpool.c.o
[  4%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/uv-common.c.o
[  4%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/timer.c.o
[  4%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/unix/async.c.o
[  4%] No patch step for 'nccl_external'
[  4%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/unix/core.c.o
[  4%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/uv-data-getter-setters.c.o
[  6%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/version.c.o
[  6%] No update step for 'nccl_external'
[  6%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/__/GKlib/b64.c.o
[  6%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/__/GKlib/error.c.o
[  7%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/unix/fs.c.o
[  7%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/unix/getnameinfo.c.o
[  7%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/unix/getaddrinfo.c.o
[  8%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/unix/loop.c.o
[  9%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/__/GKlib/blas.c.o
[ 10%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/__/GKlib/fkvkselect.c.o
[ 10%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/unix/loop-watcher.c.o
[ 10%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/__/GKlib/csr.c.o
[ 10%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/unix/dl.c.o
[ 12%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/__/GKlib/graph.c.o
[ 12%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/__/GKlib/getopt.c.o
[ 12%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/unix/process.c.o
[ 12%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/unix/pipe.c.o
[ 12%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/__/GKlib/pqueue.c.o
[ 12%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/unix/udp.c.o
[ 12%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/__/GKlib/io.c.o
[ 12%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/__/GKlib/fs.c.o
[ 13%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/checkgraph.c.o
[ 13%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/unix/tcp.c.o
[ 13%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/__/GKlib/evaluate.c.o
[ 13%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/unix/proctitle.c.o
[ 14%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/unix/tty.c.o
[ 15%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/unix/linux-inotify.c.o
[ 15%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/__/GKlib/htable.c.o
[ 15%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/contig.c.o
[ 15%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/__/GKlib/seq.c.o
[ 15%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/kwayfm.c.o
[ 16%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/unix/random-sysctl-linux.c.o
[ 16%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/__/GKlib/gkregex.c.o
[ 16%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/compress.c.o
[ 16%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/__/GKlib/gk_util.c.o
[ 16%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/__/GKlib/sort.c.o
[ 16%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/__/GKlib/memory.c.o
[ 16%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/__/GKlib/random.c.o
[ 16%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/unix/stream.c.o
[ 16%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/__/GKlib/string.c.o
[ 16%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/unix/poll.c.o
[ 16%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/unix/signal.c.o
[ 18%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/unix/linux-core.c.o
[ 18%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/unix/random-devurandom.c.o
[ 18%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/unix/thread.c.o
[ 18%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/__/GKlib/itemsets.c.o
[ 18%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/unix/linux-syscalls.c.o
[ 19%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/unix/procfs-exepath.c.o
[ 19%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/__/GKlib/tokenizer.c.o
[ 20%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/__/GKlib/mcore.c.o
[ 20%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/coarsen.c.o
[ 22%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/debug.c.o
[ 22%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/options.c.o
[ 22%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/kmetis.c.o
[ 24%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/__/GKlib/rw.c.o
[ 24%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/mmd.c.o
[ 24%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/__/GKlib/timers.c.o
[ 25%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/gklib.c.o
[ 25%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/auxapi.c.o
[ 25%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/fm.c.o
[ 25%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/frename.c.o
[ 25%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/unix/random-getrandom.c.o
[ 25%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/balance.c.o
[ 25%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/fortran.c.o
[ 25%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/bucketsort.c.o
[ 26%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/kwayrefine.c.o
[ 26%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/ometis.c.o
[ 26%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/pmetis.c.o
[ 26%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/mesh.c.o
[Pipeline] sh
-- Looking for sys/types.h - found
-- Looking for stdint.h
-- Generating done
-- Build files have been written to: /root/jenkins/workspace/dgl_PR-4648@3/build
[  0%] Creating directories for 'nccl_external'
[  1%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/fs-poll.c.o
[  1%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/idna.c.o
[  1%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/inet.c.o
[  1%] Building CXX object third_party/dmlc-core/CMakeFiles/dmlc.dir/src/config.cc.o
[  1%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/threadpool.c.o
[  1%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/random.c.o
[  1%] Building CXX object third_party/googletest/googletest/CMakeFiles/gtest.dir/src/gtest-all.cc.o
[  1%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/__/GKlib/b64.c.o
[  2%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/strscpy.c.o
[  3%] Building CXX object third_party/dmlc-core/CMakeFiles/dmlc.dir/src/recordio.cc.o
[  3%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/uv-data-getter-setters.c.o
[  3%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/uv-common.c.o
[  4%] Building CXX object third_party/dmlc-core/CMakeFiles/dmlc.dir/src/io/line_split.cc.o
[  4%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/version.c.o
[  4%] Building CXX object third_party/dmlc-core/CMakeFiles/dmlc.dir/src/data.cc.o
[  4%] Building CXX object third_party/dmlc-core/CMakeFiles/dmlc.dir/src/io.cc.o
[  4%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/timer.c.o
[  4%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/unix/async.c.o
[  6%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/__/GKlib/blas.c.o
[ 26%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/graph.c.o
[ 26%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/refine.c.o
[ 26%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/parmetis.c.o
[ 26%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/meshpart.c.o
[ 26%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/srefine.c.o
[ 28%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/mincover.c.o
[ 28%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/sfm.c.o
[ 28%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/minconn.c.o
[ 28%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/initpart.c.o
[ 28%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/mcutil.c.o
[ 28%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/separator.c.o
[ 30%] No configure step for 'nccl_external'
[ 30%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/timing.c.o
[ 30%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/wspace.c.o
[ 30%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/stat.c.o
[ 31%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/util.c.o
[ 31%] Performing build step for 'nccl_external'
-- The C compiler identification is GNU 7.5.0
[  6%] Building CXX object third_party/dmlc-core/CMakeFiles/dmlc.dir/src/io/recordio_split.cc.o
[  6%] Building CXX object third_party/dmlc-core/CMakeFiles/dmlc.dir/src/io/indexed_recordio_split.cc.o
[  6%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/__/GKlib/error.c.o
[  6%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/unix/core.c.o
[  6%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/__/GKlib/evaluate.c.o
[  6%] Building CXX object third_party/dmlc-core/CMakeFiles/dmlc.dir/src/io/filesys.cc.o
[  7%] Building CXX object third_party/dmlc-core/CMakeFiles/dmlc.dir/src/io/input_split_base.cc.o
[  7%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/__/GKlib/fs.c.o
[  7%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/__/GKlib/getopt.c.o
[  8%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/__/GKlib/fkvkselect.c.o
[  8%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/__/GKlib/gk_util.c.o
[  8%] Building CXX object third_party/dmlc-core/CMakeFiles/dmlc.dir/src/io/local_filesys.cc.o
[  8%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/__/GKlib/csr.c.o
[  9%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/unix/fs.c.o
[  9%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/unix/dl.c.o
[  9%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/unix/getnameinfo.c.o
[  9%] No download step for 'nccl_external'
[  9%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/unix/getaddrinfo.c.o
[  9%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/unix/loop-watcher.c.o
[ 10%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/unix/loop.c.o
[ 10%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/__/GKlib/gkregex.c.o
[ 12%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/__/GKlib/graph.c.o
[ 12%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/__/GKlib/htable.c.o
[ 12%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/unix/pipe.c.o
[ 12%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/__/GKlib/io.c.o
[ 13%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/unix/random-devurandom.c.o
[ 13%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/unix/stream.c.o
[ 13%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/unix/poll.c.o
[ 13%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/unix/process.c.o
[ 13%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/unix/thread.c.o
[ 14%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/__/GKlib/mcore.c.o
[ 14%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/unix/tcp.c.o
[ 14%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/__/GKlib/itemsets.c.o
-- The CXX compiler identification is GNU 7.5.0
-- Check for working C compiler: /usr/bin/cc
-- Looking for stdint.h - found
-- Looking for stddef.h
+ bash tests/scripts/build_dgl.sh cpu
~/jenkins/workspace/dgl_PR-4648/build ~/jenkins/workspace/dgl_PR-4648
-- The C compiler identification is GNU 7.5.0
[ 15%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/unix/tty.c.o
[ 15%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/unix/proctitle.c.o
[ 15%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/unix/signal.c.o
[ 15%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/unix/linux-core.c.o
[ 15%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/unix/udp.c.o
[ 15%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/__/GKlib/pqueue.c.o
[ 15%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/__/GKlib/memory.c.o
[ 16%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/unix/linux-inotify.c.o
[ 16%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/__/GKlib/random.c.o
[ 18%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/__/GKlib/rw.c.o
[ 18%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/__/GKlib/sort.c.o
[ 19%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/__/GKlib/tokenizer.c.o
[ 19%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/__/GKlib/timers.c.o
[ 19%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/auxapi.c.o
[ 19%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/__/GKlib/seq.c.o
[ 19%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/unix/linux-syscalls.c.o
[ 19%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/__/GKlib/string.c.o
[ 19%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/unix/procfs-exepath.c.o
[ 20%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/unix/random-sysctl-linux.c.o
[ 20%] No update step for 'nccl_external'
[ 20%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/unix/random-getrandom.c.o
[ 20%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/balance.c.o
[ 21%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/debug.c.o
[ 21%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/fm.c.o
[ 21%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/fortran.c.o
[ 22%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/compress.c.o
[ 22%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/checkgraph.c.o
[ 22%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/contig.c.o
[ 22%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/coarsen.c.o
[ 24%] No patch step for 'nccl_external'
[ 24%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/frename.c.o
[ 24%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/bucketsort.c.o
[ 25%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/gklib.c.o
[ 25%] No configure step for 'nccl_external'
[ 25%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/initpart.c.o
[ 26%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/kwayrefine.c.o
[ 26%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/mcutil.c.o
[ 26%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/mesh.c.o
[ 26%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/meshpart.c.o
[ 26%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/graph.c.o
[ 27%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/minconn.c.o
[ 27%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/kmetis.c.o
[ 27%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/kwayfm.c.o
[ 27%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/mmd.c.o
[ 27%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/ometis.c.o
[ 27%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/mincover.c.o
[ 28%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/options.c.o
[ 28%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/pmetis.c.o
[ 28%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/refine.c.o
[ 30%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/sfm.c.o
[ 30%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/separator.c.o
[ 30%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/srefine.c.o
[ 30%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/parmetis.c.o
[ 30%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/stat.c.o
[ 31%] Performing build step for 'nccl_external'
[ 31%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/timing.c.o
[ 32%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/util.c.o
[ 32%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/wspace.c.o
Grabbing   include/nccl_net.h                  > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/include/nccl_net.h
Generating nccl.h.in                           > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/include/nccl.h
Generating nccl.pc.in                          > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/lib/pkgconfig/nccl.pc
-- Check for working C compiler: /usr/bin/cc -- works
-- Detecting C compiler ABI info
Compiling  init.cc                             > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/init.o
Compiling  channel.cc                          > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/channel.o
Compiling  bootstrap.cc                        > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/bootstrap.o
Compiling  transport.cc                        > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/transport.o
Compiling  enqueue.cc                          > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/enqueue.o
Compiling  group.cc                            > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/group.o
Compiling  debug.cc                            > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/debug.o
-- Detecting C compiler ABI info - done
Compiling  proxy.cc                            > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/proxy.o
Compiling  misc/nvmlwrap.cc                    > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/misc/nvmlwrap.o
-- Detecting C compile features
-- Detecting C compile features - done
-- Check for working CXX compiler: /usr/bin/c++
Compiling  misc/ibvwrap.cc                     > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/misc/ibvwrap.o
Compiling  misc/gdrwrap.cc                     > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/misc/gdrwrap.o
Compiling  misc/utils.cc                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/misc/utils.o
Compiling  misc/argcheck.cc                    > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/misc/argcheck.o
Compiling  transport/p2p.cc                    > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/transport/p2p.o
Compiling  transport/shm.cc                    > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/transport/shm.o
Compiling  transport/net.cc                    > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/transport/net.o
Compiling  transport/net_socket.cc             > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/transport/net_socket.o
Compiling  transport/net_ib.cc                 > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/transport/net_ib.o
Compiling  transport/coll_net.cc               > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/transport/coll_net.o
Compiling  collectives/sendrecv.cc             > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/sendrecv.o
Compiling  collectives/all_reduce.cc           > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/all_reduce.o
-- The CXX compiler identification is GNU 7.5.0
-- Check for working C compiler: /usr/bin/cc
-- Check for working C compiler: /usr/bin/cc -- works
-- Detecting C compiler ABI info
-- Detecting C compiler ABI info - done
-- Detecting C compile features
-- Detecting C compile features - done
-- Check for working CXX compiler: /usr/bin/c++
Generating nccl.h.in                           > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/include/nccl.h
Grabbing   include/nccl_net.h                  > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/include/nccl_net.h
Generating nccl.pc.in                          > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/lib/pkgconfig/nccl.pc
Compiling  init.cc                             > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/init.o
Compiling  channel.cc                          > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/channel.o
Compiling  bootstrap.cc                        > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/bootstrap.o
Compiling  transport.cc                        > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/transport.o
Compiling  enqueue.cc                          > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/enqueue.o
Compiling  group.cc                            > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/group.o
Compiling  debug.cc                            > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/debug.o
Compiling  proxy.cc                            > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/proxy.o
Compiling  misc/nvmlwrap.cc                    > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/misc/nvmlwrap.o
Compiling  misc/ibvwrap.cc                     > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/misc/ibvwrap.o
Compiling  misc/gdrwrap.cc                     > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/misc/gdrwrap.o
Compiling  misc/utils.cc                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/misc/utils.o
Compiling  misc/argcheck.cc                    > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/misc/argcheck.o
Compiling  transport/p2p.cc                    > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/transport/p2p.o
Compiling  transport/shm.cc                    > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/transport/shm.o
Compiling  transport/net.cc                    > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/transport/net.o
Compiling  transport/net_socket.cc             > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/transport/net_socket.o
Compiling  transport/net_ib.cc                 > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/transport/net_ib.o
Compiling  transport/coll_net.cc               > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/transport/coll_net.o
Compiling  collectives/sendrecv.cc             > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/sendrecv.o
Compiling  collectives/all_reduce.cc           > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/all_reduce.o
Compiling  collectives/all_gather.cc           > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/all_gather.o
Compiling  collectives/broadcast.cc            > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/broadcast.o
Compiling  collectives/reduce.cc               > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/reduce.o
Compiling  collectives/all_gather.cc           > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/all_gather.o
-- Check for working CXX compiler: /usr/bin/c++ -- works
Compiling  collectives/broadcast.cc            > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/broadcast.o
-- Detecting CXX compiler ABI info
Compiling  collectives/reduce.cc               > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/reduce.o
Compiling  collectives/reduce_scatter.cc       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/reduce_scatter.o
Compiling  graph/topo.cc                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/graph/topo.o
Compiling  graph/paths.cc                      > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/graph/paths.o
Compiling  graph/search.cc                     > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/graph/search.o
Compiling  graph/connect.cc                    > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/graph/connect.o
Compiling  graph/rings.cc                      > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/graph/rings.o
Compiling  graph/trees.cc                      > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/graph/trees.o
Compiling  graph/tuning.cc                     > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/graph/tuning.o
Compiling  graph/xml.cc                        > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/graph/xml.o
-- Detecting CXX compiler ABI info - done
-- Detecting CXX compile features
[ 31%] Linking C static library libtensorpipe_uv.a
-- Detecting CXX compile features - done
-- Using Python interpreter: /opt/conda/envs/pytorch-ci/bin/python
[ 31%] Built target tensorpipe_uv
Generating rules                               > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/Makefile.rules
-- Check for working CXX compiler: /usr/bin/c++ -- works
-- Detecting CXX compiler ABI info
-- Detecting CXX compiler ABI info - done
-- Detecting CXX compile features
-- Detecting CXX compile features - done
-- Start configuring project dgl
-- Performing Test SUPPORT_CXX14
-- Looking for stddef.h - found
-- Check size of unsigned short
Compiling  collectives/reduce_scatter.cc       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/reduce_scatter.o
Compiling  graph/topo.cc                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/graph/topo.o
Compiling  graph/paths.cc                      > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/graph/paths.o
Compiling  graph/search.cc                     > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/graph/search.o
Compiling  graph/connect.cc                    > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/graph/connect.o
Compiling  graph/rings.cc                      > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/graph/rings.o
Compiling  graph/trees.cc                      > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/graph/trees.o
Compiling  graph/tuning.cc                     > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/graph/tuning.o
Compiling  graph/xml.cc                        > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/graph/xml.o
[ 32%] Linking C static library libtensorpipe_uv.a
Generating rules                               > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/Makefile.rules
[ 32%] Built target tensorpipe_uv
Scanning dependencies of target tensorpipe
[ 31%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/channel/error.cc.o
[ 32%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/channel/helpers.cc.o
[ 32%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/common/allocator.cc.o
[ 32%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/common/address.cc.o
[ 32%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/common/error.cc.o
[ 32%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/common/fd.cc.o
[ 33%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/common/system.cc.o
[ 33%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/core/context.cc.o
[ 33%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/common/socket.cc.o
[ 33%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/core/context_impl.cc.o
[ 34%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/core/error.cc.o
[ 34%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/core/listener.cc.o
[ 34%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/core/pipe_impl.cc.o
[ 34%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/core/pipe.cc.o
[ 36%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/transport/error.cc.o
[ 36%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/core/listener_impl.cc.o
[ 36%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/channel/basic/channel_impl.cc.o
[ 36%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/channel/basic/factory.cc.o
[ 36%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/channel/xth/channel_impl.cc.o
[ 37%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/channel/cma/channel_impl.cc.o
[ 37%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/channel/mpt/factory.cc.o
[ 37%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/channel/cma/factory.cc.o
[ 37%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/channel/cma/context_impl.cc.o
[ 38%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/transport/uv/connection_impl.cc.o
[ 38%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/channel/mpt/channel_impl.cc.o
[ 38%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/transport/uv/factory.cc.o
[ 38%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/channel/basic/context_impl.cc.o
[ 39%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/common/shm_segment.cc.o
[ 39%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/channel/xth/factory.cc.o
[ 40%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/channel/xth/context_impl.cc.o
[ 40%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/transport/uv/context_impl.cc.o
[ 42%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/transport/uv/utility.cc.o
[ 42%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/transport/ibv/connection_impl.cc.o
[ 42%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/transport/shm/reactor.cc.o
[ 43%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/transport/ibv/listener_impl.cc.o
[ 43%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/transport/shm/listener_impl.cc.o
[ 43%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/transport/uv/error.cc.o
[ 43%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/transport/shm/factory.cc.o
[ 43%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/transport/uv/sockaddr.cc.o
[ 43%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/transport/shm/sockaddr.cc.o
[ 43%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/transport/shm/connection_impl.cc.o
[ 43%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/transport/ibv/error.cc.o
[ 44%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/transport/uv/listener_impl.cc.o
[ 44%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/transport/ibv/utility.cc.o
[ 44%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/channel/mpt/context_impl.cc.o
[ 44%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/transport/ibv/context_impl.cc.o
[ 44%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/common/epoll_loop.cc.o
[ 44%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/common/ibv.cc.o
[ 44%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/transport/uv/loop.cc.o
[ 45%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/transport/ibv/reactor.cc.o
[ 45%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/transport/shm/context_impl.cc.o
[ 45%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/transport/ibv/factory.cc.o
[ 45%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/transport/ibv/sockaddr.cc.o
-- Performing Test SUPPORT_CXX14 - Success
[ 32%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/channel/error.cc.o
[ 32%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/channel/helpers.cc.o
[ 33%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/common/address.cc.o
[ 33%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/common/allocator.cc.o
[ 33%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/common/error.cc.o
[ 33%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/common/fd.cc.o
[ 33%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/common/socket.cc.o
[ 34%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/common/system.cc.o
[ 34%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/core/context.cc.o
[ 34%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/core/context_impl.cc.o
[ 34%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/core/error.cc.o
-- Check size of unsigned short - done
-- Searching 16 bit integer - Using unsigned short
[ 36%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/core/listener.cc.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
[ 36%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/core/listener_impl.cc.o
[ 36%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/core/pipe.cc.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
[ 36%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/core/pipe_impl.cc.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
[ 37%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/transport/error.cc.o
In file included from include/core.h:57:0,
                 from include/transport.h:13,
                 from include/comm.h:10,
                 from include/enqueue.h:10,
                 from enqueue.cc:7:
enqueue.cc: In function 'ncclResult_t ncclLaunchCooperativeKernelMultiDevice(cudaLaunchParams*, int*, int, int)':
enqueue.cc:135:97: warning: 'cudaError_t cudaLaunchCooperativeKernelMultiDevice(cudaLaunchParams*, unsigned int, unsigned int)' is deprecated [-Wdeprecated-declarations]
             cudaCooperativeLaunchMultiDeviceNoPreSync|cudaCooperativeLaunchMultiDeviceNoPostSync));
                                                                                                 ^
include/checks.h:14:23: note: in definition of macro 'CUDACHECK'
     cudaError_t err = cmd;                                  \
                       ^~~
In file included from /usr/local/cuda/include/channel_descriptor.h:61:0,
                 from /usr/local/cuda/include/cuda_runtime.h:95,
                 from /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/include/nccl.h:10,
                 from include/devcomm.h:10,
                 from include/transport.h:10,
                 from include/comm.h:10,
                 from include/enqueue.h:10,
                 from enqueue.cc:7:
/usr/local/cuda/include/cuda_runtime_api.h:4198:57: note: declared here
 extern __CUDA_DEPRECATED __host__ cudaError_t CUDARTAPI cudaLaunchCooperativeKernelMultiDevice(struct cudaLaunchParams *launchParamsList, unsigned int numDevices, unsigned int flags  __dv(0));
                                                         ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
[ 37%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/channel/basic/channel_impl.cc.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
[ 37%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/channel/basic/context_impl.cc.o
[ 37%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/channel/basic/factory.cc.o
[ 38%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/channel/xth/context_impl.cc.o
[ 38%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/channel/xth/channel_impl.cc.o
-- Found OpenMP_C: -fopenmp (found version "4.5") 
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
[ 38%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/channel/xth/factory.cc.o
[ 38%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/channel/cma/channel_impl.cc.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
[ 38%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/channel/cma/context_impl.cc.o
[ 39%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/channel/cma/factory.cc.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
[ 39%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/channel/mpt/channel_impl.cc.o
[ 39%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/channel/mpt/factory.cc.o
[ 39%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/channel/mpt/context_impl.cc.o
-- Found OpenMP_CXX: -fopenmp (found version "4.5") 
-- Found OpenMP: TRUE (found version "4.5")  
-- Build with OpenMP.
-- Build with LIBXSMM optimization.
-- Looking for sys/epoll.h
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
[ 39%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/transport/uv/context_impl.cc.o
[ 40%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/transport/uv/connection_impl.cc.o
[ 40%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/transport/uv/error.cc.o
[ 40%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/transport/uv/factory.cc.o
[ 42%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/transport/uv/listener_impl.cc.o
[ 42%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/transport/uv/loop.cc.o
[ 42%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/transport/uv/sockaddr.cc.o
[ 42%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/transport/uv/utility.cc.o
-- Check if the system is big endian - little endian
-- C:/Jenkins/workspace/dgl_PR-4648/third_party/dmlc-core/cmake/build_config.h.in -> include/dmlc/build_config.h
-- Looking for sys/epoll.h - found
-- Found OpenMP_C: -fopenmp (found version "4.5") 
-- Found OpenMP_CXX: -fopenmp (found version "4.5") 
-- Looking for clock_gettime in rt
-- Looking for clock_gettime in rt - found
-- Looking for fopen64
[ 42%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/common/epoll_loop.cc.o
[ 43%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/common/shm_segment.cc.o
[ 43%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/transport/shm/connection_impl.cc.o
[ 43%] Linking C static library libmetis.a
[ 43%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/transport/shm/context_impl.cc.o
[ 43%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/transport/shm/factory.cc.o
[ 44%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/transport/shm/listener_impl.cc.o
[ 44%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/transport/shm/reactor.cc.o
[ 44%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/transport/shm/sockaddr.cc.o
[ 44%] Built target metis
[ 44%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/common/ibv.cc.o
-- Found PythonInterp: C:/Users/Administrator/Envs/jenkins-dgl-PR-4648-2/Scripts/python.exe (found version "3.6.7") 
-- Build with unittest
CMake Deprecation Warning at third_party/googletest/CMakeLists.txt:4 (cmake_minimum_required):
  Compatibility with CMake < 2.8.12 will be removed from a future version of
  CMake.

  Update the VERSION argument <min> value or use a ...<max> suffix to tell
  CMake that the project does not need compatibility with older versions.


CMake Deprecation Warning at third_party/googletest/googlemock/CMakeLists.txt:45 (cmake_minimum_required):
  Compatibility with CMake < 2.8.12 will be removed from a future version of
  CMake.

  Update the VERSION argument <min> value or use a ...<max> suffix to tell
  CMake that the project does not need compatibility with older versions.


CMake Deprecation Warning at third_party/googletest/googletest/CMakeLists.txt:56 (cmake_minimum_required):
  Compatibility with CMake < 2.8.12 will be removed from a future version of
  CMake.

  Update the VERSION argument <min> value or use a ...<max> suffix to tell
  CMake that the project does not need compatibility with older versions.


-- Looking for pthread.h
-- Looking for fopen64 - not found
-- Looking for C++ include cxxabi.h
[ 45%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/transport/ibv/connection_impl.cc.o
[ 45%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/transport/ibv/context_impl.cc.o
[ 45%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/transport/ibv/error.cc.o
[ 45%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/transport/ibv/factory.cc.o
[ 45%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/transport/ibv/listener_impl.cc.o
[ 46%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/transport/ibv/reactor.cc.o
-- Looking for pthread.h - not found
-- Found Threads: TRUE  
-- Configuring done
-- Looking for C++ include cxxabi.h - found
-- Looking for nanosleep
-- Looking for nanosleep - found
-- Looking for backtrace
[ 46%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/transport/ibv/sockaddr.cc.o
[ 46%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/transport/ibv/utility.cc.o
-- Generating done
-- Build files have been written to: C:/Jenkins/workspace/dgl_PR-4648/build
Microsoft (R) Build Engine version 16.5.1+4616136f8 for .NET Framework
Copyright (C) Microsoft Corporation. All rights reserved.

Build started 9/27/2022 4:42:41 AM.
-- find_cmake.py output: /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/torch/share/cmake;1.9.0
-- Configuring for PyTorch 1.9.0
-- Setting directory to /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/torch/share/cmake/Torch
-- Looking for pthread.h
-- Looking for pthread.h - found
-- Performing Test CMAKE_HAVE_LIBC_PTHREAD
-- Looking for backtrace - found
-- backtrace facility detected in default set of libraries
-- Found Backtrace: /usr/include  
-- Check if the system is big endian
-- Searching 16 bit integer
-- Looking for sys/types.h
     1>Project "C:\Jenkins\workspace\dgl_PR-4648\build\dgl.sln" on node 1 (default targets).
     1>ValidateSolutionConfiguration:
         Building solution configuration "Release|x64".
       ValidateProjects:
         The project "INSTALL" is not selected for building in solution configuration "Release|x64".
         The project "RUN_TESTS" is not selected for building in solution configuration "Release|x64".
         The project "dmlc_lint" is not selected for building in solution configuration "Release|x64".
-- Performing Test CMAKE_HAVE_LIBC_PTHREAD - Failed
-- Looking for pthread_create in pthreads
[ 45%] Linking C static library libmetis.a
-- Looking for sys/types.h - found
-- Looking for stdint.h
     1>Project "C:\Jenkins\workspace\dgl_PR-4648\build\dgl.sln" (1) is building "C:\Jenkins\workspace\dgl_PR-4648\build\ALL_BUILD.vcxproj.metaproj" (2) on node 1 (default targets).
     2>Project "C:\Jenkins\workspace\dgl_PR-4648\build\ALL_BUILD.vcxproj.metaproj" (2) is building "C:\Jenkins\workspace\dgl_PR-4648\build\ZERO_CHECK.vcxproj" (11) on node 2 (default targets).
    11>PrepareForBuild:
         Creating directory "x64\Release\ZERO_CHECK\".
         Creating directory "x64\Release\ZERO_CHECK\ZERO_CHECK.tlog\".
       InitializeBuildStatus:
         Creating "x64\Release\ZERO_CHECK\ZERO_CHECK.tlog\unsuccessfulbuild" because "AlwaysCreate" was specified.
       CustomBuild:
         Checking Build System
       FinalizeBuildStatus:
         Deleting file "x64\Release\ZERO_CHECK\ZERO_CHECK.tlog\unsuccessfulbuild".
         Touching "x64\Release\ZERO_CHECK\ZERO_CHECK.tlog\ZERO_CHECK.lastbuildstate".
    11>Done Building Project "C:\Jenkins\workspace\dgl_PR-4648\build\ZERO_CHECK.vcxproj" (default targets).
[ 45%] Built target metis
-- Looking for pthread_create in pthreads - not found
-- Looking for pthread_create in pthread
-- Looking for pthread_create in pthread - found
-- Found Threads: TRUE  
-- Found CUDA: /usr/local/cuda (found version "10.1") 
-- Caffe2: CUDA detected: 10.1
-- Caffe2: CUDA nvcc is: /usr/local/cuda/bin/nvcc
-- Caffe2: CUDA toolkit directory: /usr/local/cuda
     1>Project "C:\Jenkins\workspace\dgl_PR-4648\build\dgl.sln" (1) is building "C:\Jenkins\workspace\dgl_PR-4648\build\tensoradapter_pytorch.vcxproj.metaproj" (10) on node 1 (default targets).
    10>Project "C:\Jenkins\workspace\dgl_PR-4648\build\tensoradapter_pytorch.vcxproj.metaproj" (10) is building "C:\Jenkins\workspace\dgl_PR-4648\build\tensoradapter_pytorch.vcxproj" (12) on node 2 (default targets).
    12>PrepareForBuild:
         Creating directory "x64\Release\tensoradapter_pytorch\".
         Creating directory "x64\Release\tensoradapter_pytorch\tensorad.15EA3878.tlog\".
       InitializeBuildStatus:
         Creating "x64\Release\tensoradapter_pytorch\tensorad.15EA3878.tlog\unsuccessfulbuild" because "AlwaysCreate" was specified.
       CustomBuild:
         
         (jenkins-dgl-PR-4648-2) administrator@WIN-O7QS55HVSDB C:\Jenkins\workspace\dgl_PR-4648\tensoradapter\pytorch>REM Helper script to build tensor adapter libraries for PyTorch 
         Could Not Find C:\Jenkins\workspace\dgl_PR-4648\tensoradapter\pytorch\build
         -- Selecting Windows SDK version 10.0.18362.0 to target Windows 6.3.9600.
     1>Project "C:\Jenkins\workspace\dgl_PR-4648\build\dgl.sln" (1) is building "C:\Jenkins\workspace\dgl_PR-4648\build\third_party\googletest\googletest\gtest.vcxproj.metaproj" (7) on node 1 (default targets).
     7>Project "C:\Jenkins\workspace\dgl_PR-4648\build\third_party\googletest\googletest\gtest.vcxproj.metaproj" (7) is building "C:\Jenkins\workspace\dgl_PR-4648\build\third_party\googletest\googletest\gtest.vcxproj" (13) on node 3 (default targets).
    13>PrepareForBuild:
         Creating directory "gtest.dir\Release\".
         Creating directory "C:\Jenkins\workspace\dgl_PR-4648\build\lib\Release\".
         Creating directory "gtest.dir\Release\gtest.tlog\".
       InitializeBuildStatus:
         Creating "gtest.dir\Release\gtest.tlog\unsuccessfulbuild" because "AlwaysCreate" was specified.
-- Looking for stdint.h - found
-- Looking for stddef.h
-- Looking for stddef.h - found
-- Check size of unsigned short
-- Check size of unsigned short - done
-- Using unsigned short
     1>Project "C:\Jenkins\workspace\dgl_PR-4648\build\dgl.sln" (1) is building "C:\Jenkins\workspace\dgl_PR-4648\build\third_party\googletest\googlemock\gmock_main.vcxproj.metaproj" (6) on node 1 (default targets).
     6>Project "C:\Jenkins\workspace\dgl_PR-4648\build\third_party\googletest\googlemock\gmock_main.vcxproj.metaproj" (6) is building "C:\Jenkins\workspace\dgl_PR-4648\build\third_party\googletest\googlemock\gmock_main.vcxproj" (14) on node 4 (default targets).
    14>PrepareForBuild:
         Creating directory "gmock_main.dir\Release\".
         Creating directory "gmock_main.dir\Release\gmock_main.tlog\".
       InitializeBuildStatus:
         Creating "gmock_main.dir\Release\gmock_main.tlog\unsuccessfulbuild" because "AlwaysCreate" was specified.
    13>CustomBuild:
         Building Custom Rule C:/Jenkins/workspace/dgl_PR-4648/third_party/googletest/googletest/CMakeLists.txt
     1>Project "C:\Jenkins\workspace\dgl_PR-4648\build\dgl.sln" (1) is building "C:\Jenkins\workspace\dgl_PR-4648\build\third_party\googletest\googlemock\gmock.vcxproj.metaproj" (5) on node 1 (default targets).
     5>Project "C:\Jenkins\workspace\dgl_PR-4648\build\third_party\googletest\googlemock\gmock.vcxproj.metaproj" (5) is building "C:\Jenkins\workspace\dgl_PR-4648\build\third_party\googletest\googlemock\gmock.vcxproj" (15) on node 5 (default targets).
    15>PrepareForBuild:
         Creating directory "gmock.dir\Release\".
    13>MakeDirsForCl:
         Creating directory "C:\Jenkins\workspace\dgl_PR-4648\build\bin\Release".
    15>PrepareForBuild:
         Creating directory "gmock.dir\Release\gmock.tlog\".
       InitializeBuildStatus:
         Creating "gmock.dir\Release\gmock.tlog\unsuccessfulbuild" because "AlwaysCreate" was specified.
    14>CustomBuild:
         Building Custom Rule C:/Jenkins/workspace/dgl_PR-4648/third_party/googletest/googlemock/CMakeLists.txt
    13>ClCompile:
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\bin\HostX64\x64\CL.exe /c /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\googletest\googletest\include" /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\googletest\googletest" /Zi /nologo /W4 /WX /diagnostics:column /MP /O2 /Ob2 /D DGL_EXPORTS /D USE_AVX /D USE_LIBXSMM /D DGL_CPU_LLC_SIZE=40000000 /D IDXTYPEWIDTH=64 /D REALTYPEWIDTH=32 /D NDEBUG /D _UNICODE /D UNICODE /D WIN32 /D _WIN32 /D STRICT /D WIN32_LEAN_AND_MEAN /D GTEST_HAS_PTHREAD=0 /D _HAS_EXCEPTIONS=1 /D _CRT_SECURE_NO_WARNINGS /D _SCL_SECURE_NO_WARNINGS /D NOMINMAX /D ENABLE_PARTIAL_FRONTIER=0 /D "CMAKE_INTDIR=\"Release\"" /D _UNICODE /D UNICODE /Gm- /EHsc /MD /GS /fp:precise /Zc:wchar_t /Zc:forScope /Zc:inline /openmp /Fo"gtest.dir\Release\\" /Fd"C:\Jenkins\workspace\dgl_PR-4648\build\bin\Release\gtest.pdb" /Gd /TP /wd4251 /wd4275 /wd4702 /errorReport:queue  -J "C:\Jenkins\workspace\dgl_PR-4648\third_party\googletest\googletest\src\gtest-all.cc"
     1>Project "C:\Jenkins\workspace\dgl_PR-4648\build\dgl.sln" (1) is building "C:\Jenkins\workspace\dgl_PR-4648\build\third_party\dmlc-core\dmlc.vcxproj.metaproj" (4) on node 1 (default targets).
     4>Project "C:\Jenkins\workspace\dgl_PR-4648\build\third_party\dmlc-core\dmlc.vcxproj.metaproj" (4) is building "C:\Jenkins\workspace\dgl_PR-4648\build\third_party\dmlc-core\dmlc.vcxproj" (16) on node 6 (default targets).
    16>PrepareForBuild:
         Creating directory "dmlc.dir\Release\".
         Creating directory "C:\Jenkins\workspace\dgl_PR-4648\build\third_party\dmlc-core\Release\".
         Creating directory "dmlc.dir\Release\dmlc.tlog\".
       InitializeBuildStatus:
         Creating "dmlc.dir\Release\dmlc.tlog\unsuccessfulbuild" because "AlwaysCreate" was specified.
    15>CustomBuild:
         Building Custom Rule C:/Jenkins/workspace/dgl_PR-4648/third_party/googletest/googlemock/CMakeLists.txt
    13>ClCompile:
         gtest-all.cc
    14>ClCompile:
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\bin\HostX64\x64\CL.exe /c /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\googletest\googlemock\include" /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\googletest\googlemock" /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\googletest\googletest\include" /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\googletest\googletest" /Zi /nologo /W4 /WX /diagnostics:column /MP /O2 /Ob2 /D DGL_EXPORTS /D USE_AVX /D USE_LIBXSMM /D DGL_CPU_LLC_SIZE=40000000 /D IDXTYPEWIDTH=64 /D REALTYPEWIDTH=32 /D NDEBUG /D _UNICODE /D UNICODE /D WIN32 /D _WIN32 /D STRICT /D WIN32_LEAN_AND_MEAN /D GTEST_HAS_PTHREAD=0 /D _HAS_EXCEPTIONS=1 /D _CRT_SECURE_NO_WARNINGS /D _SCL_SECURE_NO_WARNINGS /D NOMINMAX /D ENABLE_PARTIAL_FRONTIER=0 /D "CMAKE_INTDIR=\"Release\"" /D _UNICODE /D UNICODE /Gm- /EHsc /MD /GS /fp:precise /Zc:wchar_t /Zc:forScope /Zc:inline /openmp /Fo"gmock_main.dir\Release\\" /Fd"C:\Jenkins\workspace\dgl_PR-4648\build\bin\Release\gmock_main.pdb" /Gd /TP /wd4251 /wd4275 /wd4702 /errorReport:queue  -J "C:\Jenkins\workspace\dgl_PR-4648\third_party\googletest\googletest\src\gtest-all.cc" "C:\Jenkins\workspace\dgl_PR-4648\third_party\googletest\googlemock\src\gmock-all.cc" "C:\Jenkins\workspace\dgl_PR-4648\third_party\googletest\googlemock\src\gmock_main.cc"
    16>CustomBuild:
         Building Custom Rule C:/Jenkins/workspace/dgl_PR-4648/third_party/dmlc-core/CMakeLists.txt
    15>ClCompile:
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\bin\HostX64\x64\CL.exe /c /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\googletest\googlemock\include" /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\googletest\googlemock" /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\googletest\googletest\include" /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\googletest\googletest" /Zi /nologo /W4 /WX /diagnostics:column /MP /O2 /Ob2 /D DGL_EXPORTS /D USE_AVX /D USE_LIBXSMM /D DGL_CPU_LLC_SIZE=40000000 /D IDXTYPEWIDTH=64 /D REALTYPEWIDTH=32 /D NDEBUG /D _UNICODE /D UNICODE /D WIN32 /D _WIN32 /D STRICT /D WIN32_LEAN_AND_MEAN /D GTEST_HAS_PTHREAD=0 /D _HAS_EXCEPTIONS=1 /D _CRT_SECURE_NO_WARNINGS /D _SCL_SECURE_NO_WARNINGS /D NOMINMAX /D ENABLE_PARTIAL_FRONTIER=0 /D "CMAKE_INTDIR=\"Release\"" /D _UNICODE /D UNICODE /Gm- /EHsc /MD /GS /fp:precise /Zc:wchar_t /Zc:forScope /Zc:inline /openmp /Fo"gmock.dir\Release\\" /Fd"C:\Jenkins\workspace\dgl_PR-4648\build\bin\Release\gmock.pdb" /Gd /TP /wd4251 /wd4275 /wd4702 /errorReport:queue  -J "C:\Jenkins\workspace\dgl_PR-4648\third_party\googletest\googletest\src\gtest-all.cc" "C:\Jenkins\workspace\dgl_PR-4648\third_party\googletest\googlemock\src\gmock-all.cc"
    14>ClCompile:
         gtest-all.cc
         gmock-all.cc
         gmock_main.cc
    15>ClCompile:
         gtest-all.cc
         gmock-all.cc
    16>ClCompile:
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\bin\HostX64\x64\CL.exe /c /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\dmlc-core\include" /I"C:\Jenkins\workspace\dgl_PR-4648\build\third_party\dmlc-core\include" /nologo /W1 /WX- /diagnostics:column /MP /O2 /Ob2 /D DGL_EXPORTS /D USE_AVX /D USE_LIBXSMM /D DGL_CPU_LLC_SIZE=40000000 /D IDXTYPEWIDTH=64 /D REALTYPEWIDTH=32 /D NDEBUG /D DMLC_USE_HDFS=0 /D DMLC_USE_S3=0 /D DMLC_USE_AZURE=0 /D _XOPEN_SOURCE=700 /D _POSIX_SOURCE /D _POSIX_C_SOURCE=200809L /D _DARWIN_C_SOURCE /D __USE_XOPEN2K8 /D DMLC_CORE_USE_CMAKE /D DMLC_USE_CXX11=1 /D WIN32_LEAN_AND_MEAN /D _CRT_SECURE_NO_WARNINGS /D _SCL_SECURE_NO_WARNINGS /D NOMINMAX /D ENABLE_PARTIAL_FRONTIER=0 /D "CMAKE_INTDIR=\"Release\"" /D _MBCS /Gm- /EHsc /MD /GS /fp:precise /Zc:wchar_t /Zc:forScope /Zc:inline /openmp /Fo"dmlc.dir\Release\\" /Fd"dmlc.dir\Release\dmlc.pdb" /Gd /TP /errorReport:queue "C:\Jenkins\workspace\dgl_PR-4648\third_party\dmlc-core\src\config.cc" "C:\Jenkins\workspace\dgl_PR-4648\third_party\dmlc-core\src\data.cc" "C:\Jenkins\workspace\dgl_PR-4648\third_party\dmlc-core\src\io.cc" "C:\Jenkins\workspace\dgl_PR-4648\third_party\dmlc-core\src\recordio.cc" "C:\Jenkins\workspace\dgl_PR-4648\third_party\dmlc-core\src\io\line_split.cc" "C:\Jenkins\workspace\dgl_PR-4648\third_party\dmlc-core\src\io\recordio_split.cc" "C:\Jenkins\workspace\dgl_PR-4648\third_party\dmlc-core\src\io\indexed_recordio_split.cc" "C:\Jenkins\workspace\dgl_PR-4648\third_party\dmlc-core\src\io\input_split_base.cc" "C:\Jenkins\workspace\dgl_PR-4648\third_party\dmlc-core\src\io\filesys.cc" "C:\Jenkins\workspace\dgl_PR-4648\third_party\dmlc-core\src\io\local_filesys.cc"
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_reduce_sum_i8.o
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_reduce_sum_u8.o
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_reduce_sum_i32.o
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_reduce_sum_u32.o
-- Check if the system is big endian - little endian
-- /root/jenkins/workspace/dgl_PR-4648/third_party/dmlc-core/cmake/build_config.h.in -> include/dmlc/build_config.h
-- Performing Test SUPPORT_MSSE2
         config.cc
         data.cc
         io.cc
         recordio.cc
         line_split.cc
         recordio_split.cc
         indexed_recordio_split.cc
         input_split_base.cc
         filesys.cc
         local_filesys.cc
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_reduce_sum_i64.o
-- Caffe2: Header version is: 10.1
-- Found CUDNN: /usr/lib/x86_64-linux-gnu/libcudnn.so  
-- Found cuDNN: v7.6.5  (include: /usr/include, library: /usr/lib/x86_64-linux-gnu/libcudnn.so)
CMake Warning at /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/torch/share/cmake/Caffe2/public/cuda.cmake:203 (message):
  Failed to compute shorthash for libnvrtc.so
Call Stack (most recent call first):
  /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/torch/share/cmake/Caffe2/Caffe2Config.cmake:88 (include)
  /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/torch/share/cmake/Torch/TorchConfig.cmake:68 (find_package)
  CMakeLists.txt:26 (find_package)


Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_reduce_sum_u64.o
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_reduce_sum_f16.o
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_reduce_sum_f32.o
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_reduce_sum_f64.o
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_reduce_prod_i8.o
-- Performing Test SUPPORT_MSSE2 - Success
-- Looking for execinfo.h
-- Looking for execinfo.h - found
-- Looking for getline
    16>C:\Jenkins\workspace\dgl_PR-4648\third_party\dmlc-core\include\dmlc/filesystem.h(19,1): warning C4005: 'NOMINMAX': macro redefinition (compiling source file C:\Jenkins\workspace\dgl_PR-4648\third_party\dmlc-core\src\io\indexed_recordio_split.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\third_party\dmlc-core\dmlc.vcxproj]
       C:\Jenkins\workspace\dgl_PR-4648\third_party\dmlc-core\src\io\indexed_recordio_split.cc : message : see previous definition of 'NOMINMAX' [C:\Jenkins\workspace\dgl_PR-4648\build\third_party\dmlc-core\dmlc.vcxproj]
    16>C:\Jenkins\workspace\dgl_PR-4648\third_party\dmlc-core\include\dmlc/filesystem.h(19,1): warning C4005: 'NOMINMAX': macro redefinition (compiling source file C:\Jenkins\workspace\dgl_PR-4648\third_party\dmlc-core\src\io\recordio_split.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\third_party\dmlc-core\dmlc.vcxproj]
       C:\Jenkins\workspace\dgl_PR-4648\third_party\dmlc-core\src\io\recordio_split.cc : message : see previous definition of 'NOMINMAX' [C:\Jenkins\workspace\dgl_PR-4648\build\third_party\dmlc-core\dmlc.vcxproj]
    16>C:\Jenkins\workspace\dgl_PR-4648\third_party\dmlc-core\include\dmlc/filesystem.h(19,1): warning C4005: 'NOMINMAX': macro redefinition (compiling source file C:\Jenkins\workspace\dgl_PR-4648\third_party\dmlc-core\src\io\line_split.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\third_party\dmlc-core\dmlc.vcxproj]
       C:\Jenkins\workspace\dgl_PR-4648\third_party\dmlc-core\src\io\line_split.cc : message : see previous definition of 'NOMINMAX' [C:\Jenkins\workspace\dgl_PR-4648\build\third_party\dmlc-core\dmlc.vcxproj]
    16>C:\Jenkins\workspace\dgl_PR-4648\third_party\dmlc-core\include\dmlc/filesystem.h(19,1): warning C4005: 'NOMINMAX': macro redefinition (compiling source file C:\Jenkins\workspace\dgl_PR-4648\third_party\dmlc-core\src\io\filesys.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\third_party\dmlc-core\dmlc.vcxproj]
       C:\Jenkins\workspace\dgl_PR-4648\third_party\dmlc-core\src\io\filesys.cc : message : see previous definition of 'NOMINMAX' [C:\Jenkins\workspace\dgl_PR-4648\build\third_party\dmlc-core\dmlc.vcxproj]
    16>C:\Jenkins\workspace\dgl_PR-4648\third_party\dmlc-core\include\dmlc/filesystem.h(19,1): warning C4005: 'NOMINMAX': macro redefinition (compiling source file C:\Jenkins\workspace\dgl_PR-4648\third_party\dmlc-core\src\io\input_split_base.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\third_party\dmlc-core\dmlc.vcxproj]
       C:\Jenkins\workspace\dgl_PR-4648\third_party\dmlc-core\src\io\input_split_base.cc : message : see previous definition of 'NOMINMAX' [C:\Jenkins\workspace\dgl_PR-4648\build\third_party\dmlc-core\dmlc.vcxproj]
    16>C:\Jenkins\workspace\dgl_PR-4648\third_party\dmlc-core\include\dmlc/filesystem.h(19,1): warning C4005: 'NOMINMAX': macro redefinition (compiling source file C:\Jenkins\workspace\dgl_PR-4648\third_party\dmlc-core\src\io.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\third_party\dmlc-core\dmlc.vcxproj]
       C:\Jenkins\workspace\dgl_PR-4648\third_party\dmlc-core\src\io.cc : message : see previous definition of 'NOMINMAX' [C:\Jenkins\workspace\dgl_PR-4648\build\third_party\dmlc-core\dmlc.vcxproj]
    12>CustomBuild:
         -- The C compiler identification is MSVC 19.25.28614.0
    16>C:\Jenkins\workspace\dgl_PR-4648\third_party\dmlc-core\include\dmlc/filesystem.h(19,1): warning C4005: 'NOMINMAX': macro redefinition (compiling source file C:\Jenkins\workspace\dgl_PR-4648\third_party\dmlc-core\src\io\local_filesys.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\third_party\dmlc-core\dmlc.vcxproj]
       C:\Jenkins\workspace\dgl_PR-4648\third_party\dmlc-core\src\io\local_filesys.cc : message : see previous definition of 'NOMINMAX' [C:\Jenkins\workspace\dgl_PR-4648\build\third_party\dmlc-core\dmlc.vcxproj]
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_reduce_prod_u8.o
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_reduce_prod_i32.o
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_reduce_prod_u32.o
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_reduce_prod_i64.o
-- Automatic GPU detection failed. Building for common architectures.
-- Autodetected CUDA architecture(s): 3.5;5.0;5.2;6.0;6.1;7.0;7.5;7.5+PTX
-- Added CUDA NVCC flags for: -gencode;arch=compute_35,code=sm_35;-gencode;arch=compute_50,code=sm_50;-gencode;arch=compute_52,code=sm_52;-gencode;arch=compute_60,code=sm_60;-gencode;arch=compute_61,code=sm_61;-gencode;arch=compute_70,code=sm_70;-gencode;arch=compute_75,code=sm_75;-gencode;arch=compute_75,code=compute_75
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_reduce_prod_u64.o
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/sendrecv_sum_i8.o
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/sendrecv_sum_u8.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/sendrecv_sum_i32.o
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/sendrecv_sum_u32.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/sendrecv_sum_i64.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
-- Looking for getline - found
-- Performing Test UV_LINT_W4
-- Performing Test UV_LINT_W4 - Failed
-- Performing Test UV_LINT_NO_UNUSED_PARAMETER_MSVC
-- Performing Test UV_LINT_NO_UNUSED_PARAMETER_MSVC - Failed
-- Performing Test UV_LINT_NO_CONDITIONAL_CONSTANT_MSVC
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_reduce_prod_f16.o
CMake Warning at /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/torch/share/cmake/Torch/TorchConfig.cmake:22 (message):
  static library kineto_LIBRARY-NOTFOUND not found.
Call Stack (most recent call first):
  /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/torch/share/cmake/Torch/TorchConfig.cmake:127 (append_torchlib_if_found)
  CMakeLists.txt:26 (find_package)


-- Found Torch: /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/torch/lib/libtorch.so  
-- tensoradapter found PyTorch includes: /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/torch/include;/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/torch/include/torch/csrc/api/include
-- tensoradapter found PyTorch lib: torch
-- Configured target tensoradapter_pytorch_1.9.0
-- Configuring done
-- Generating done
-- Build files have been written to: /root/jenkins/workspace/dgl_PR-4648@2/tensoradapter/pytorch/build
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_reduce_prod_f32.o
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_reduce_prod_f64.o
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_reduce_min_i8.o
Scanning dependencies of target tensoradapter_pytorch_1.9.0
[ 50%] Building CXX object CMakeFiles/tensoradapter_pytorch_1.9.0.dir/torch.cpp.o
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_reduce_min_u8.o
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/sendrecv_sum_u64.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
-- Performing Test UV_LINT_NO_CONDITIONAL_CONSTANT_MSVC - Failed
-- Performing Test UV_LINT_NO_NONSTANDARD_MSVC
-- Performing Test UV_LINT_NO_NONSTANDARD_MSVC - Failed
-- Performing Test UV_LINT_NO_NONSTANDARD_EMPTY_TU_MSVC
-- Performing Test UV_LINT_NO_NONSTANDARD_EMPTY_TU_MSVC - Failed
-- Performing Test UV_LINT_NO_NONSTANDARD_FILE_SCOPE_MSVC
-- Performing Test UV_LINT_NO_NONSTANDARD_FILE_SCOPE_MSVC - Failed
-- Performing Test UV_LINT_NO_NONSTANDARD_NONSTATIC_DLIMPORT_MSVC
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_reduce_min_i32.o
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_reduce_min_u32.o
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_reduce_min_i64.o
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_reduce_min_u64.o
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_reduce_min_f16.o
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/sendrecv_sum_f16.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/sendrecv_sum_f32.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/sendrecv_sum_f64.o
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/sendrecv_sum_bf16.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
-- Performing Test UV_LINT_NO_NONSTANDARD_NONSTATIC_DLIMPORT_MSVC - Failed
-- Performing Test UV_LINT_NO_HIDES_LOCAL
-- Performing Test UV_LINT_NO_HIDES_LOCAL - Failed
-- Performing Test UV_LINT_NO_HIDES_PARAM
-- Performing Test UV_LINT_NO_HIDES_PARAM - Failed
-- Performing Test UV_LINT_NO_HIDES_GLOBAL
-- Performing Test UV_LINT_NO_HIDES_GLOBAL - Failed
-- Performing Test UV_LINT_NO_CONDITIONAL_ASSIGNMENT_MSVC
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_reduce_min_f32.o
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_reduce_min_f64.o
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_reduce_max_i8.o
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_reduce_max_u8.o
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_reduce_max_i32.o
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_reduce_max_u32.o
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/sendrecv_prod_i8.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/sendrecv_prod_u8.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/sendrecv_prod_i32.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/sendrecv_prod_u32.o
-- Performing Test UV_LINT_NO_CONDITIONAL_ASSIGNMENT_MSVC - Failed
-- Performing Test UV_LINT_NO_UNSAFE_MSVC
-- Performing Test UV_LINT_NO_UNSAFE_MSVC - Failed
-- Performing Test UV_LINT_WALL
-- Performing Test UV_LINT_WALL - Success
-- Performing Test UV_LINT_NO_UNUSED_PARAMETER
       CustomBuild:
         -- The CXX compiler identification is MSVC 19.25.28614.0
         -- Detecting C compiler ABI info
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_reduce_max_i64.o
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_reduce_max_u64.o
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_reduce_max_f16.o
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_reduce_max_f32.o
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_reduce_max_f64.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/sendrecv_prod_i64.o
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/sendrecv_prod_u64.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
-- Performing Test UV_LINT_NO_UNUSED_PARAMETER - Success
-- Performing Test UV_LINT_STRICT_PROTOTYPES
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_reduce_premulsum_i8.o
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_reduce_premulsum_u8.o
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_reduce_premulsum_i32.o
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_reduce_premulsum_u32.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/sendrecv_prod_f16.o
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/sendrecv_prod_f32.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/sendrecv_prod_f64.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/sendrecv_prod_bf16.o
-- Performing Test UV_LINT_STRICT_PROTOTYPES - Success
-- Performing Test UV_LINT_EXTRA
-- Performing Test UV_LINT_EXTRA - Success
-- Performing Test UV_LINT_UTF8_MSVC
-- Performing Test UV_LINT_UTF8_MSVC - Failed
-- Performing Test UV_F_STRICT_ALIASING
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_reduce_premulsum_i64.o
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_reduce_premulsum_u64.o
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_reduce_premulsum_f16.o
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_reduce_premulsum_f32.o
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_reduce_premulsum_f64.o
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_reduce_sumpostdiv_i8.o
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/sendrecv_min_i8.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/sendrecv_min_u8.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/sendrecv_min_i32.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/sendrecv_min_u32.o
-- Performing Test UV_F_STRICT_ALIASING - Success
-- summary of build options:
    Install prefix:  /usr/local
    Target system:   Linux
    Compiler:
      C compiler:    /usr/bin/cc
      CFLAGS:         -fopenmp -O2  -fPIC  -DUSE_AVX -DUSE_LIBXSMM -DDGL_CPU_LLC_SIZE=40000000 -DUSE_EPOLL -DIDXTYPEWIDTH=64 -DREALTYPEWIDTH=32 -DLINUX -D_FILE_OFFSET_BITS=64 -std=c99 -fno-strict-aliasing -march=native -fPIC -Werror   -Wno-unused-function -Wno-unused-but-set-variable -Wno-unused-variable -Wno-unknown-pragmas -DNDEBUG -DNDEBUG2 -DHAVE_EXECINFO_H -DHAVE_GETLINE -O3

-- Found uv: 1.38.1 (found version "1.38.1") 
-- Build with unittest
-- Found PythonInterp: /usr/bin/python3.6 (found version "3.6.9") 
-- Looking for pthread.h
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_reduce_sumpostdiv_u8.o
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_reduce_sumpostdiv_i32.o
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_reduce_sumpostdiv_u32.o
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_reduce_sumpostdiv_i64.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/sendrecv_min_i64.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/sendrecv_min_u64.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/sendrecv_min_f16.o
         -- Detecting C compiler ABI info - done
         -- Check for working C compiler: C:/Program Files (x86)/Microsoft Visual Studio/2019/BuildTools/VC/Tools/MSVC/14.25.28610/bin/Hostx64/x64/cl.exe - skipped
         -- Detecting C compile features
         -- Detecting C compile features - done
         -- Detecting CXX compiler ABI info
-- Looking for pthread.h - found
-- Performing Test CMAKE_HAVE_LIBC_PTHREAD
-- Performing Test CMAKE_HAVE_LIBC_PTHREAD - Success
-- Found Threads: TRUE  
-- Building dist/rpc tests
-- Configuring done
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_reduce_sumpostdiv_u64.o
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_reduce_sumpostdiv_f16.o
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_reduce_sumpostdiv_f32.o
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_reduce_sumpostdiv_f64.o
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_gather_sum_i8.o
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/sendrecv_min_f32.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/sendrecv_min_f64.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/sendrecv_min_bf16.o
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/sendrecv_max_i8.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
-- Generating done
-- Build files have been written to: /root/jenkins/workspace/dgl_PR-4648/build
Scanning dependencies of target tensoradapter_pytorch
Scanning dependencies of target gtest
Scanning dependencies of target libxsmm
Scanning dependencies of target dmlc
[  0%] Building CXX object third_party/googletest/googletest/CMakeFiles/gtest.dir/src/gtest-all.cc.o
[  0%] Building CXX object third_party/dmlc-core/CMakeFiles/dmlc.dir/src/io.cc.o
[  0%] Building CXX object third_party/dmlc-core/CMakeFiles/dmlc.dir/src/config.cc.o
[  1%] Building CXX object third_party/dmlc-core/CMakeFiles/dmlc.dir/src/recordio.cc.o
[  1%] Building CXX object third_party/dmlc-core/CMakeFiles/dmlc.dir/src/io/indexed_recordio_split.cc.o
[  1%] Building CXX object third_party/dmlc-core/CMakeFiles/dmlc.dir/src/io/recordio_split.cc.o
Scanning dependencies of target tensorpipe_uv
[  2%] Building CXX object third_party/dmlc-core/CMakeFiles/dmlc.dir/src/io/input_split_base.cc.o
[  2%] Building CXX object third_party/dmlc-core/CMakeFiles/dmlc.dir/src/data.cc.o
[  2%] Building CXX object third_party/dmlc-core/CMakeFiles/dmlc.dir/src/io/filesys.cc.o
[  2%] Building CXX object third_party/dmlc-core/CMakeFiles/dmlc.dir/src/io/line_split.cc.o
[  2%] Building CXX object third_party/dmlc-core/CMakeFiles/dmlc.dir/src/io/local_filesys.cc.o
Scanning dependencies of target metis
[  3%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/unix/loop-watcher.c.o
[  3%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/unix/signal.c.o
[  3%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/random.c.o
[  3%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/unix/random-devurandom.c.o
[  3%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/strscpy.c.o
[  3%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/unix/fs.c.o
[  4%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/unix/getaddrinfo.c.o
[  4%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/timer.c.o
[  4%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/unix/dl.c.o
[  4%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/unix/async.c.o
[  4%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/unix/core.c.o
[  6%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/threadpool.c.o
[  6%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/unix/procfs-exepath.c.o
[  6%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/inet.c.o
[  6%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/fs-poll.c.o
[  7%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/unix/loop.c.o
[  7%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/unix/process.c.o
[  7%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/uv-data-getter-setters.c.o
[  7%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/unix/linux-core.c.o
[  7%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/unix/getnameinfo.c.o
[  7%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/unix/random-sysctl-linux.c.o
[  7%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/unix/stream.c.o
[  7%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/unix/poll.c.o
[  7%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/unix/udp.c.o
[  7%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/uv-common.c.o
[  7%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/unix/pipe.c.o
[  8%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/unix/proctitle.c.o
[  9%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/idna.c.o
[  9%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/unix/thread.c.o
[ 11%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/version.c.o
[ 12%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/unix/tcp.c.o
[ 13%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/unix/linux-syscalls.c.o
[ 13%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/unix/tty.c.o
[ 13%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/unix/linux-inotify.c.o
[ 13%] Building C object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe_uv.dir/__/third_party/libuv/src/unix/random-getrandom.c.o
[ 13%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/__/GKlib/b64.c.o
[ 13%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/__/GKlib/blas.c.o
[ 13%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/__/GKlib/csr.c.o
[ 14%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/__/GKlib/error.c.o
[ 14%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/__/GKlib/gk_util.c.o
[ 16%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/contig.c.o
[ 16%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/__/GKlib/evaluate.c.o
[ 16%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/__/GKlib/memory.c.o
[ 16%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/__/GKlib/gkregex.c.o
[ 16%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/__/GKlib/tokenizer.c.o
[ 16%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/__/GKlib/fs.c.o
[ 18%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/__/GKlib/timers.c.o
[ 18%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/__/GKlib/graph.c.o
[ 19%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/__/GKlib/getopt.c.o
[ 19%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/__/GKlib/string.c.o
[ 20%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/__/GKlib/mcore.c.o
[ 20%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/bucketsort.c.o
[ 20%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/__/GKlib/fkvkselect.c.o
[ 20%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/kwayfm.c.o
[ 20%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/compress.c.o
[ 20%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/__/GKlib/io.c.o
[ 20%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/fm.c.o
[ 20%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/__/GKlib/seq.c.o
[ 20%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/checkgraph.c.o
[ 22%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/__/GKlib/rw.c.o
[ 22%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/auxapi.c.o
[ 23%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/__/GKlib/htable.c.o
[ 23%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/gklib.c.o
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_gather_sum_u8.o
[100%] Linking CXX shared library libtensoradapter_pytorch_1.9.0.so
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_gather_sum_i32.o
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_gather_sum_u32.o
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_gather_sum_i64.o
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/sendrecv_max_u8.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/sendrecv_max_i32.o
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/sendrecv_max_u32.o
[ 23%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/coarsen.c.o
[ 23%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/graph.c.o
[ 23%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/kwayrefine.c.o
[ 23%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/mesh.c.o
[ 23%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/debug.c.o
[ 23%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/__/GKlib/random.c.o
[ 24%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/mcutil.c.o
[ 24%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/__/GKlib/sort.c.o
[ 24%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/ometis.c.o
[ 24%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/__/GKlib/itemsets.c.o
[ 24%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/options.c.o
[ 24%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/minconn.c.o
[ 25%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/fortran.c.o
[ 25%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/__/GKlib/pqueue.c.o
[ 25%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/frename.c.o
[ 25%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/mmd.c.o
[ 25%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/pmetis.c.o
[ 25%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/meshpart.c.o
[ 27%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/initpart.c.o
[ 27%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/kmetis.c.o
[ 27%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/util.c.o
[ 28%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/separator.c.o
[ 28%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/wspace.c.o
[ 28%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/stat.c.o
[ 29%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/parmetis.c.o
[ 29%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/balance.c.o
[ 30%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/timing.c.o
[ 30%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/sfm.c.o
[ 30%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/refine.c.o
[ 30%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/srefine.c.o
[ 32%] Building C object third_party/METIS/libmetis/CMakeFiles/metis.dir/mincover.c.o
-- The C compiler identification is GNU 7.5.0
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_gather_sum_u64.o
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_gather_sum_f16.o
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_gather_sum_f32.o
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_gather_sum_f64.o
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_gather_prod_i8.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/sendrecv_max_i64.o
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/sendrecv_max_u64.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/sendrecv_max_f16.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
-- The CXX compiler identification is GNU 7.5.0
-- Check for working C compiler: /usr/bin/cc
         -- Detecting CXX compiler ABI info - done
         -- Check for working CXX compiler: C:/Program Files (x86)/Microsoft Visual Studio/2019/BuildTools/VC/Tools/MSVC/14.25.28610/bin/Hostx64/x64/cl.exe - skipped
         -- Detecting CXX compile features
         -- Detecting CXX compile features - done
         -- Using Python interpreter: C:/Program Files/Python36/python.exe
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_gather_prod_u8.o
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_gather_prod_i32.o
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_gather_prod_u32.o
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_gather_prod_i64.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/sendrecv_max_f32.o
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/sendrecv_max_f64.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/sendrecv_max_bf16.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/sendrecv_premulsum_i8.o
-- Check for working C compiler: /usr/bin/cc -- works
-- Detecting C compiler ABI info
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_gather_prod_u64.o
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_gather_prod_f16.o
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_gather_prod_f32.o
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_gather_prod_f64.o
[100%] Built target tensoradapter_pytorch_1.9.0
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_gather_min_i8.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/sendrecv_premulsum_u8.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/sendrecv_premulsum_i32.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
-- Detecting C compiler ABI info - done
-- Detecting C compile features
-- Detecting C compile features - done
-- Check for working CXX compiler: /usr/bin/c++
[ 33%] Linking C static library libtensorpipe_uv.a
         -- find_cmake.py output: C:\Program Files\Python36\lib\site-packages\torch\share\cmake;1.9.0
         -- Configuring for PyTorch 1.9.0
         -- Setting directory to C:\Program Files\Python36\lib\site-packages\torch\share\cmake/Torch
         -- Looking for pthread.h
'libtensoradapter_pytorch_1.9.0.so' -> '/root/jenkins/workspace/dgl_PR-4648@2/build/tensoradapter/pytorch/libtensoradapter_pytorch_1.9.0.so'
[ 45%] Built target tensoradapter_pytorch
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_gather_min_u8.o
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_gather_min_i32.o
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_gather_min_u32.o
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_gather_min_i64.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/sendrecv_premulsum_u32.o
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/sendrecv_premulsum_i64.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/sendrecv_premulsum_u64.o
[ 33%] Built target tensorpipe_uv
-- Check for working CXX compiler: /usr/bin/c++ -- works
-- Detecting CXX compiler ABI info
Scanning dependencies of target tensorpipe
[ 34%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/common/address.cc.o
[ 34%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/channel/helpers.cc.o
[ 34%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/common/allocator.cc.o
[ 35%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/common/socket.cc.o
[ 35%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/channel/error.cc.o
[ 35%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/common/error.cc.o
[ 35%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/common/system.cc.o
[ 35%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/core/context.cc.o
[ 35%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/channel/cma/factory.cc.o
[ 37%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/channel/xth/factory.cc.o
[ 37%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/common/fd.cc.o
[ 39%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/core/pipe.cc.o
[ 39%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/channel/basic/channel_impl.cc.o
[ 39%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/channel/mpt/channel_impl.cc.o
[ 39%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/transport/error.cc.o
[ 39%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/channel/cma/channel_impl.cc.o
[ 39%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/core/listener_impl.cc.o
[ 39%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/channel/mpt/context_impl.cc.o
[ 40%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/transport/uv/error.cc.o
[ 41%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/channel/basic/context_impl.cc.o
[ 41%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/channel/xth/channel_impl.cc.o
[ 41%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/core/context_impl.cc.o
[ 41%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/channel/cma/context_impl.cc.o
[ 41%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/channel/mpt/factory.cc.o
[ 41%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/transport/uv/factory.cc.o
[ 41%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/core/pipe_impl.cc.o
[ 43%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/core/listener.cc.o
[ 43%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/transport/shm/listener_impl.cc.o
[ 43%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/common/ibv.cc.o
[ 43%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/transport/uv/sockaddr.cc.o
[ 44%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/transport/uv/connection_impl.cc.o
[ 44%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/transport/shm/connection_impl.cc.o
[ 44%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/transport/shm/sockaddr.cc.o
[ 45%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/transport/uv/listener_impl.cc.o
[ 45%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/transport/uv/utility.cc.o
[ 45%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/channel/basic/factory.cc.o
[ 45%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/transport/uv/loop.cc.o
[ 45%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/channel/xth/context_impl.cc.o
[ 45%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/transport/ibv/reactor.cc.o
[ 46%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/common/epoll_loop.cc.o
[ 46%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/core/error.cc.o
[ 46%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/transport/shm/context_impl.cc.o
[ 48%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/transport/ibv/error.cc.o
[ 48%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/transport/ibv/connection_impl.cc.o
[ 48%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/transport/ibv/context_impl.cc.o
[ 48%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/transport/uv/context_impl.cc.o
[ 49%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/transport/shm/factory.cc.o
[ 50%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/transport/ibv/sockaddr.cc.o
[ 50%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/transport/ibv/listener_impl.cc.o
[ 50%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/common/shm_segment.cc.o
[ 50%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/transport/shm/reactor.cc.o
[ 50%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/transport/ibv/factory.cc.o
[ 50%] Building CXX object third_party/tensorpipe/tensorpipe/CMakeFiles/tensorpipe.dir/transport/ibv/utility.cc.o
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_gather_min_u64.o
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_gather_min_f16.o
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_gather_min_f32.o
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_gather_min_f64.o
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/sendrecv_premulsum_f16.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/sendrecv_premulsum_f32.o
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/sendrecv_premulsum_f64.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
-- Detecting CXX compiler ABI info - done
-- Detecting CXX compile features
-- Detecting CXX compile features - done
-- Using Python interpreter: /opt/conda/envs/pytorch-ci/bin/python
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_gather_max_i8.o
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_gather_max_u8.o
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_gather_max_i32.o
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_gather_max_u32.o
         -- Looking for pthread.h - not found
         -- Found Threads: TRUE  
         -- Found Torch: C:/Program Files/Python36/Lib/site-packages/torch/lib/torch.lib  
         -- tensoradapter found PyTorch includes: C:/Program Files/Python36/Lib/site-packages/torch/include;C:/Program Files/Python36/Lib/site-packages/torch/include/torch/csrc/api/include
         -- tensoradapter found PyTorch lib: torch
         -- Configured target tensoradapter_pytorch_1.9.0
         -- Configuring done
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/sendrecv_premulsum_bf16.o
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/sendrecv_sumpostdiv_i8.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/sendrecv_sumpostdiv_u8.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_gather_max_i64.o
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_gather_max_u64.o
         -- Generating done
    12>CUSTOMBUILD : CMake warning :  [C:\Jenkins\workspace\dgl_PR-4648\build\tensoradapter_pytorch.vcxproj]
         -- Build files have been written to: C:/Jenkins/workspace/dgl_PR-4648/tensoradapter/pytorch/build
           Manually-specified variables were not used by the project:
         
             CUDA_TOOLKIT_ROOT_DIR
             TORCH_CUDA_ARCH_LIST
         
         
         Microsoft (R) Build Engine version 16.5.1+4616136f8 for .NET Framework
         Copyright (C) Microsoft Corporation. All rights reserved.
         
         Build started 9/27/2022 4:42:46 AM.
              1>Project "C:\Jenkins\workspace\dgl_PR-4648\tensoradapter\pytorch\build\tensoradapter_pytorch.sln" on node 1 (default targets).
              1>ValidateSolutionConfiguration:
                  Building solution configuration "Release|x64".
                ValidateProjects:
                  The project "ALL_BUILD" is not selected for building in solution configuration "Release|x64".
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/sendrecv_sumpostdiv_i32.o
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_gather_max_f16.o
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_gather_max_f32.o
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_gather_max_f64.o
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_gather_premulsum_i8.o
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_gather_premulsum_i32.o
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_gather_premulsum_u8.o
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_gather_premulsum_u32.o
              1>Project "C:\Jenkins\workspace\dgl_PR-4648\tensoradapter\pytorch\build\tensoradapter_pytorch.sln" (1) is building "C:\Jenkins\workspace\dgl_PR-4648\tensoradapter\pytorch\build\ZERO_CHECK.vcxproj" (2) on node 1 (default targets).
              2>PrepareForBuild:
                  Creating directory "x64\Release\ZERO_CHECK\".
                  Creating directory "x64\Release\ZERO_CHECK\ZERO_CHECK.tlog\".
                InitializeBuildStatus:
                  Creating "x64\Release\ZERO_CHECK\ZERO_CHECK.tlog\unsuccessfulbuild" because "AlwaysCreate" was specified.
                CustomBuild:
                  Checking Build System
                FinalizeBuildStatus:
                  Deleting file "x64\Release\ZERO_CHECK\ZERO_CHECK.tlog\unsuccessfulbuild".
                  Touching "x64\Release\ZERO_CHECK\ZERO_CHECK.tlog\ZERO_CHECK.lastbuildstate".
              2>Done Building Project "C:\Jenkins\workspace\dgl_PR-4648\tensoradapter\pytorch\build\ZERO_CHECK.vcxproj" (default targets).
              1>Project "C:\Jenkins\workspace\dgl_PR-4648\tensoradapter\pytorch\build\tensoradapter_pytorch.sln" (1) is building "C:\Jenkins\workspace\dgl_PR-4648\tensoradapter\pytorch\build\tensoradapter_pytorch_1.9.0.vcxproj.metaproj" (3) on node 1 (default targets).
              3>Project "C:\Jenkins\workspace\dgl_PR-4648\tensoradapter\pytorch\build\tensoradapter_pytorch_1.9.0.vcxproj.metaproj" (3) is building "C:\Jenkins\workspace\dgl_PR-4648\tensoradapter\pytorch\build\tensoradapter_pytorch_1.9.0.vcxproj" (4) on node 1 (default targets).
              4>PrepareForBuild:
                  Creating directory "tensoradapter_pytorch_1.9.0.dir\Release\".
                  Creating directory "C:\Jenkins\workspace\dgl_PR-4648\tensoradapter\pytorch\build\Release\".
                  Creating directory "tensoradapter_pytorch_1.9.0.dir\Release\tensorad.D2CCE92A.tlog\".
                InitializeBuildStatus:
                  Creating "tensoradapter_pytorch_1.9.0.dir\Release\tensorad.D2CCE92A.tlog\unsuccessfulbuild" because "AlwaysCreate" was specified.
                CustomBuild:
                  Building Custom Rule C:/Jenkins/workspace/dgl_PR-4648/tensoradapter/pytorch/CMakeLists.txt
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/sendrecv_sumpostdiv_u32.o
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/sendrecv_sumpostdiv_i64.o
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/sendrecv_sumpostdiv_u64.o
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/sendrecv_sumpostdiv_f16.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/sendrecv_sumpostdiv_f32.o
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/sendrecv_sumpostdiv_f64.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/sendrecv_sumpostdiv_bf16.o
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_gather_premulsum_i64.o
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_gather_premulsum_u64.o
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_gather_premulsum_f16.o
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_gather_premulsum_f32.o
                ClCompile:
                  C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\bin\HostX64\x64\CL.exe /c /I"C:\Jenkins\workspace\dgl_PR-4648\tensoradapter\pytorch\..\include" /I"C:\Jenkins\workspace\dgl_PR-4648\tensoradapter\pytorch\..\..\third_party\dlpack\include" /I"C:\Program Files\Python36\Lib\site-packages\torch\include" /I"C:\Program Files\Python36\Lib\site-packages\torch\include\torch\csrc\api\include" /nologo /W3 /WX- /diagnostics:column /O2 /Ob2 /D WIN32 /D _WINDOWS /D NDEBUG /D NOMINMAX /D "CMAKE_INTDIR=\"Release\"" /D tensoradapter_pytorch_1_9_0_EXPORTS /D _WINDLL /D _MBCS /Gm- /EHsc /MD /GS /fp:precise /Zc:wchar_t /Zc:forScope /Zc:inline /GR /std:c++14 /Fo"tensoradapter_pytorch_1.9.0.dir\Release\\" /Fd"tensoradapter_pytorch_1.9.0.dir\Release\vc142.pdb" /Gd /TP /wd4267 /wd4251 /wd4522 /wd4838 /wd4305 /wd4244 /wd4190 /wd4101 /wd4996 /wd4275 /errorReport:queue  /bigobj "C:\Jenkins\workspace\dgl_PR-4648\tensoradapter\pytorch\torch.cpp"
                  torch.cpp
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_reduce_sum_i8.o
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_reduce_sum_u8.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_gather_premulsum_f64.o
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_gather_sumpostdiv_i8.o
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_gather_sumpostdiv_u8.o
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_reduce_sum_i32.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_reduce_sum_u32.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_reduce_sum_i64.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_reduce_sum_u64.o
[ 50%] Linking C static library libmetis.a
[ 50%] Built target metis
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_gather_sumpostdiv_i32.o
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_gather_sumpostdiv_u32.o
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_gather_sumpostdiv_i64.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_reduce_sum_f16.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_reduce_sum_f32.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
                Link:
                  C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\bin\HostX64\x64\link.exe /ERRORREPORT:QUEUE /OUT:"C:\Jenkins\workspace\dgl_PR-4648\tensoradapter\pytorch\build\Release\tensoradapter_pytorch_1.9.0.dll" /INCREMENTAL:NO /NOLOGO "C:\Program Files\Python36\Lib\site-packages\torch\lib\torch.lib" "C:\Program Files\Python36\Lib\site-packages\torch\lib\torch_cpu.lib" "C:\Program Files\Python36\Lib\site-packages\torch\lib\c10.lib" kernel32.lib user32.lib gdi32.lib winspool.lib shell32.lib ole32.lib oleaut32.lib uuid.lib comdlg32.lib advapi32.lib /MANIFEST /MANIFESTUAC:"level='asInvoker' uiAccess='false'" /manifest:embed /PDB:"C:/Jenkins/workspace/dgl_PR-4648/tensoradapter/pytorch/build/Release/tensoradapter_pytorch_1.9.0.pdb" /SUBSYSTEM:CONSOLE /TLBID:1 /DYNAMICBASE /NXCOMPAT /IMPLIB:"C:/Jenkins/workspace/dgl_PR-4648/tensoradapter/pytorch/build/Release/tensoradapter_pytorch_1.9.0.lib" /MACHINE:X64  /machine:x64 /DLL tensoradapter_pytorch_1.9.0.dir\Release\torch.obj
                     Creating library C:/Jenkins/workspace/dgl_PR-4648/tensoradapter/pytorch/build/Release/tensoradapter_pytorch_1.9.0.lib and object C:/Jenkins/workspace/dgl_PR-4648/tensoradapter/pytorch/build/Release/tensoradapter_pytorch_1.9.0.exp
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_reduce_sum_f64.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_reduce_sum_bf16.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
                  tensoradapter_pytorch_1.9.0.vcxproj -> C:\Jenkins\workspace\dgl_PR-4648\tensoradapter\pytorch\build\Release\tensoradapter_pytorch_1.9.0.dll
                FinalizeBuildStatus:
                  Deleting file "tensoradapter_pytorch_1.9.0.dir\Release\tensorad.D2CCE92A.tlog\unsuccessfulbuild".
                  Touching "tensoradapter_pytorch_1.9.0.dir\Release\tensorad.D2CCE92A.tlog\tensoradapter_pytorch_1.9.0.lastbuildstate".
              4>Done Building Project "C:\Jenkins\workspace\dgl_PR-4648\tensoradapter\pytorch\build\tensoradapter_pytorch_1.9.0.vcxproj" (default targets).
              3>Done Building Project "C:\Jenkins\workspace\dgl_PR-4648\tensoradapter\pytorch\build\tensoradapter_pytorch_1.9.0.vcxproj.metaproj" (default targets).
              1>Done Building Project "C:\Jenkins\workspace\dgl_PR-4648\tensoradapter\pytorch\build\tensoradapter_pytorch.sln" (default targets).
         
         Build succeeded.
             0 Warning(s)
             0 Error(s)
         
         Time Elapsed 00:00:01.29
         Release\tensoradapter_pytorch_1.9.0.dll
                 1 file(s) copied.
         Building Custom Rule C:/Jenkins/workspace/dgl_PR-4648/CMakeLists.txt
       FinalizeBuildStatus:
         Deleting file "x64\Release\tensoradapter_pytorch\tensorad.15EA3878.tlog\unsuccessfulbuild".
         Touching "x64\Release\tensoradapter_pytorch\tensorad.15EA3878.tlog\tensoradapter_pytorch.lastbuildstate".
    12>Done Building Project "C:\Jenkins\workspace\dgl_PR-4648\build\tensoradapter_pytorch.vcxproj" (default targets).
    10>Done Building Project "C:\Jenkins\workspace\dgl_PR-4648\build\tensoradapter_pytorch.vcxproj.metaproj" (default targets).
    16>Lib:
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\bin\HostX64\x64\Lib.exe /OUT:"C:\Jenkins\workspace\dgl_PR-4648\build\third_party\dmlc-core\Release\dmlc.lib" /NOLOGO /MACHINE:X64  /machine:x64 dmlc.dir\Release\config.obj
         dmlc.dir\Release\data.obj
         dmlc.dir\Release\io.obj
         dmlc.dir\Release\recordio.obj
         dmlc.dir\Release\line_split.obj
         dmlc.dir\Release\recordio_split.obj
         dmlc.dir\Release\indexed_recordio_split.obj
         dmlc.dir\Release\input_split_base.obj
         dmlc.dir\Release\filesys.obj
         dmlc.dir\Release\local_filesys.obj
    13>Lib:
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\bin\HostX64\x64\Lib.exe /OUT:"C:\Jenkins\workspace\dgl_PR-4648\build\lib\Release\gtest.lib" /NOLOGO /MACHINE:X64  /machine:x64 "gtest.dir\Release\gtest-all.obj"
    15>Lib:
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\bin\HostX64\x64\Lib.exe /OUT:"C:\Jenkins\workspace\dgl_PR-4648\build\lib\Release\gmock.lib" /NOLOGO /MACHINE:X64  /machine:x64 "gmock.dir\Release\gtest-all.obj"
         "gmock.dir\Release\gmock-all.obj"
    14>Lib:
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\bin\HostX64\x64\Lib.exe /OUT:"C:\Jenkins\workspace\dgl_PR-4648\build\lib\Release\gmock_main.lib" /NOLOGO /MACHINE:X64  /machine:x64 "gmock_main.dir\Release\gtest-all.obj"
         "gmock_main.dir\Release\gmock-all.obj"
         gmock_main.dir\Release\gmock_main.obj
    13>Lib:
         gtest.vcxproj -> C:\Jenkins\workspace\dgl_PR-4648\build\lib\Release\gtest.lib
    15>Lib:
         gmock.vcxproj -> C:\Jenkins\workspace\dgl_PR-4648\build\lib\Release\gmock.lib
    14>Lib:
         gmock_main.vcxproj -> C:\Jenkins\workspace\dgl_PR-4648\build\lib\Release\gmock_main.lib
    16>Lib:
         dmlc.vcxproj -> C:\Jenkins\workspace\dgl_PR-4648\build\third_party\dmlc-core\Release\dmlc.lib
    13>FinalizeBuildStatus:
         Deleting file "gtest.dir\Release\gtest.tlog\unsuccessfulbuild".
         Touching "gtest.dir\Release\gtest.tlog\gtest.lastbuildstate".
    13>Done Building Project "C:\Jenkins\workspace\dgl_PR-4648\build\third_party\googletest\googletest\gtest.vcxproj" (default targets).
     7>Done Building Project "C:\Jenkins\workspace\dgl_PR-4648\build\third_party\googletest\googletest\gtest.vcxproj.metaproj" (default targets).
    14>FinalizeBuildStatus:
         Deleting file "gmock_main.dir\Release\gmock_main.tlog\unsuccessfulbuild".
    15>FinalizeBuildStatus:
         Deleting file "gmock.dir\Release\gmock.tlog\unsuccessfulbuild".
    14>FinalizeBuildStatus:
         Touching "gmock_main.dir\Release\gmock_main.tlog\gmock_main.lastbuildstate".
    15>FinalizeBuildStatus:
         Touching "gmock.dir\Release\gmock.tlog\gmock.lastbuildstate".
    14>Done Building Project "C:\Jenkins\workspace\dgl_PR-4648\build\third_party\googletest\googlemock\gmock_main.vcxproj" (default targets).
    15>Done Building Project "C:\Jenkins\workspace\dgl_PR-4648\build\third_party\googletest\googlemock\gmock.vcxproj" (default targets).
    16>FinalizeBuildStatus:
         Deleting file "dmlc.dir\Release\dmlc.tlog\unsuccessfulbuild".
         Touching "dmlc.dir\Release\dmlc.tlog\dmlc.lastbuildstate".
    16>Done Building Project "C:\Jenkins\workspace\dgl_PR-4648\build\third_party\dmlc-core\dmlc.vcxproj" (default targets).
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_gather_sumpostdiv_u64.o
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_gather_sumpostdiv_f16.o
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_gather_sumpostdiv_f32.o
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/all_gather_sumpostdiv_f64.o
-- find_cmake.py output: /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/torch/share/cmake;1.9.0
-- Configuring for PyTorch 1.9.0
-- Setting directory to /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/torch/share/cmake/Torch
-- Looking for pthread.h
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_reduce_prod_i8.o
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_reduce_prod_u8.o
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_reduce_prod_i32.o
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_reduce_prod_u32.o
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_reduce_prod_i64.o
     6>Done Building Project "C:\Jenkins\workspace\dgl_PR-4648\build\third_party\googletest\googlemock\gmock_main.vcxproj.metaproj" (default targets).
     5>Done Building Project "C:\Jenkins\workspace\dgl_PR-4648\build\third_party\googletest\googlemock\gmock.vcxproj.metaproj" (default targets).
     4>Done Building Project "C:\Jenkins\workspace\dgl_PR-4648\build\third_party\dmlc-core\dmlc.vcxproj.metaproj" (default targets).
     1>Project "C:\Jenkins\workspace\dgl_PR-4648\build\dgl.sln" (1) is building "C:\Jenkins\workspace\dgl_PR-4648\build\third_party\googletest\googletest\gtest_main.vcxproj.metaproj" (8) on node 1 (default targets).
     8>Project "C:\Jenkins\workspace\dgl_PR-4648\build\third_party\googletest\googletest\gtest_main.vcxproj.metaproj" (8) is building "C:\Jenkins\workspace\dgl_PR-4648\build\third_party\googletest\googletest\gtest_main.vcxproj" (17) on node 1 (default targets).
    17>PrepareForBuild:
         Creating directory "gtest_main.dir\Release\".
         Creating directory "gtest_main.dir\Release\gtest_main.tlog\".
       InitializeBuildStatus:
         Creating "gtest_main.dir\Release\gtest_main.tlog\unsuccessfulbuild" because "AlwaysCreate" was specified.
       CustomBuild:
         Building Custom Rule C:/Jenkins/workspace/dgl_PR-4648/third_party/googletest/googletest/CMakeLists.txt
       ClCompile:
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\bin\HostX64\x64\CL.exe /c /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\googletest\googletest\include" /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\googletest\googletest" /Zi /nologo /W4 /WX /diagnostics:column /MP /O2 /Ob2 /D DGL_EXPORTS /D USE_AVX /D USE_LIBXSMM /D DGL_CPU_LLC_SIZE=40000000 /D IDXTYPEWIDTH=64 /D REALTYPEWIDTH=32 /D NDEBUG /D _UNICODE /D UNICODE /D WIN32 /D _WIN32 /D STRICT /D WIN32_LEAN_AND_MEAN /D GTEST_HAS_PTHREAD=0 /D _HAS_EXCEPTIONS=1 /D _CRT_SECURE_NO_WARNINGS /D _SCL_SECURE_NO_WARNINGS /D NOMINMAX /D ENABLE_PARTIAL_FRONTIER=0 /D "CMAKE_INTDIR=\"Release\"" /D _UNICODE /D UNICODE /Gm- /EHsc /MD /GS /fp:precise /Zc:wchar_t /Zc:forScope /Zc:inline /openmp /Fo"gtest_main.dir\Release\\" /Fd"C:\Jenkins\workspace\dgl_PR-4648\build\bin\Release\gtest_main.pdb" /Gd /TP /wd4251 /wd4275 /wd4702 /errorReport:queue  -J "C:\Jenkins\workspace\dgl_PR-4648\third_party\googletest\googletest\src\gtest_main.cc"
     1>Project "C:\Jenkins\workspace\dgl_PR-4648\build\dgl.sln" (1) is building "C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj.metaproj" (3) on node 1 (default targets).
     3>Project "C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj.metaproj" (3) is building "C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj" (18) on node 1 (default targets).
    18>PrepareForBuild:
         Creating directory "C:\Jenkins\workspace\dgl_PR-4648\build\Release\".
         Creating directory "dgl.dir\Release\dgl.tlog\".
       InitializeBuildStatus:
         Creating "dgl.dir\Release\dgl.tlog\unsuccessfulbuild" because "AlwaysCreate" was specified.
    17>ClCompile:
         gtest_main.cc
    18>CustomBuild:
         Building Custom Rule C:/Jenkins/workspace/dgl_PR-4648/CMakeLists.txt
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/broadcast_sum_i8.o
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/broadcast_sum_u8.o
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/broadcast_sum_i32.o
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/broadcast_sum_u32.o
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/broadcast_sum_i64.o
-- Looking for pthread.h - found
-- Performing Test CMAKE_HAVE_LIBC_PTHREAD
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_reduce_prod_u64.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_reduce_prod_f16.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
       MakeDirsForCl:
         Creating directory "C:\Jenkins\workspace\dgl_PR-4648\build\dgl.dir\Release\src\array\cpu".
         Creating directory "C:\Jenkins\workspace\dgl_PR-4648\build\dgl.dir\Release\src\graph\transform\cpu".
       ClCompile:
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\bin\HostX64\x64\CL.exe /c /I"C:\Jenkins\workspace\dgl_PR-4648\include" /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\dlpack\include" /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\dmlc-core\include" /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\phmap" /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\xbyak" /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\METIS\include" /I"C:\Jenkins\workspace\dgl_PR-4648\tensoradapter\include" /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\nanoflann\include" /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\libxsmm\include" /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\googletest\googletest\include" /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\googletest\googletest" /I"C:\Jenkins\workspace\dgl_PR-4648\build\third_party\dmlc-core\include" /nologo /W1 /WX- /diagnostics:column /MP /O2 /Ob2 /D DGL_EXPORTS /D USE_AVX /D USE_LIBXSMM /D DGL_CPU_LLC_SIZE=40000000 /D IDXTYPEWIDTH=64 /D REALTYPEWIDTH=32 /D NDEBUG /D WIN32_LEAN_AND_MEAN /D _CRT_SECURE_NO_WARNINGS /D _SCL_SECURE_NO_WARNINGS /D NOMINMAX /D ENABLE_PARTIAL_FRONTIER=0 /D __USE_XOPEN2K8 /D DMLC_CORE_USE_CMAKE /D DMLC_USE_CXX11=1 /D "CMAKE_INTDIR=\"Release\"" /D dgl_EXPORTS /D _WINDLL /D _MBCS /Gm- /EHsc /MD /GS /fp:precise /Zc:wchar_t /Zc:forScope /Zc:inline /openmp /std:c++14 /Fo"dgl.dir\Release\\" /Fd"dgl.dir\Release\vc142.pdb" /Gd /TP /errorReport:queue "C:\Jenkins\workspace\dgl_PR-4648\src\array\array.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\array\array_arith.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\array_cumsum.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\array_index_select.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\array_nonzero.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\array_op_impl.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\array_pack.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\array_repeat.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\array_scatter.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\array_sort.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\coo_coalesce.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\coo_linegraph.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\coo_remove.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\coo_sort.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\csr_get_data.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\csr_mm.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\csr_remove.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\csr_sort.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\csr_sum.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\csr_to_simple.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\csr_union.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\disjoint_union.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\gather_mm.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\negative_sampling.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\rowwise_sampling.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\rowwise_topk.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\sddmm.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\segment_reduce.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\spmat_op_impl_coo.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\spmat_op_impl_csr.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\spmm.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\array\filter.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\array\kernel.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\array\libra_partition.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\array\uvm_array.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\bcast.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\c_api_common.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\geometry\cpu\geometry_op_impl.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\geometry\geometry.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\partition\ndarray_partition.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\random\cpu\choice.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\random\random.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\runtime\c_object_api.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\runtime\c_runtime_api.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\runtime\config.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\runtime\cpu_device_api.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\runtime\dlpack_convert.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\runtime\dso_module.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\runtime\file_util.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\runtime\module.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\runtime\module_util.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\runtime\ndarray.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\runtime\object.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\runtime\registry.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\runtime\resource_manager.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\runtime\semaphore_wrapper.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\runtime\shared_mem.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\runtime\system_lib_module.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\runtime\tensordispatch.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\runtime\thread_pool.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\runtime\threading_backend.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\runtime\utils.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\runtime\workspace_pool.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\api\api_container.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\api\api_test.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\graph\creators.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\graph\gk_ops.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\graph\graph.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\graph\graph_apis.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\graph\graph_op.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\graph\graph_traversal.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\graph\heterograph.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\graph\heterograph_capi.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\graph\immutable_graph.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\graph\metis_partition.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\graph\network.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\graph\nodeflow.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\graph\pickle.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampler.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampling\negative\global_uniform.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampling\neighbor\neighbor.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampling\randomwalks\get_node_types_cpu.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampling\randomwalks\node2vec.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampling\randomwalks\node2vec_cpu.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampling\randomwalks\randomwalk_cpu.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampling\randomwalks\randomwalk_with_restart_cpu.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampling\randomwalks\randomwalks.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\graph\serialize\dglgraph_serialize.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\graph\serialize\graph_serialize.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\graph\serialize\heterograph_serialize.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\graph\serialize\tensor_serialize.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\graph\serialize\zerocopy_serializer.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\graph\shared_mem_manager.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\graph\subgraph.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\compact.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\line_graph.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\metis_partition_hetero.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\partition_hetero.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\remove_edges.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\to_bipartite.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\to_simple.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\graph\unit_graph.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\scheduler\scheduler.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\scheduler\scheduler_apis.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\rpc\network\common.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\rpc\network\msg_queue.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\rpc\network\socket_communicator.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\rpc\network\socket_pool.cc" "C:\Jenkins\workspace\dgl_PR-4648\src\rpc\network\tcp_socket.cc"
         array.cc
         array_arith.cc
         array_cumsum.cc
         array_index_select.cc
         array_nonzero.cc
         array_op_impl.cc
         array_pack.cc
         array_repeat.cc
         array_scatter.cc
         array_sort.cc
         coo_coalesce.cc
         coo_linegraph.cc
         coo_remove.cc
         coo_sort.cc
         csr_get_data.cc
         csr_mm.cc
         csr_remove.cc
         csr_sort.cc
         csr_sum.cc
         csr_to_simple.cc
         csr_union.cc
         disjoint_union.cc
         gather_mm.cc
         negative_sampling.cc
         rowwise_sampling.cc
         rowwise_topk.cc
         sddmm.cc
         segment_reduce.cc
         spmat_op_impl_coo.cc
         spmat_op_impl_csr.cc
         spmm.cc
         filter.cc
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/broadcast_sum_u64.o
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/broadcast_sum_f16.o
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/broadcast_sum_f32.o
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/broadcast_sum_f64.o
-- Performing Test CMAKE_HAVE_LIBC_PTHREAD - Failed
-- Looking for pthread_create in pthreads
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_reduce_prod_f32.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_reduce_prod_f64.o
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/broadcast_prod_i8.o
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/broadcast_prod_u8.o
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/broadcast_prod_i32.o
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/broadcast_prod_u32.o
-- Looking for pthread_create in pthreads - not found
-- Looking for pthread_create in pthread
-- Looking for pthread_create in pthread - found
-- Found Threads: TRUE  
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_reduce_prod_bf16.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_reduce_min_i8.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
    17>Lib:
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\bin\HostX64\x64\Lib.exe /OUT:"C:\Jenkins\workspace\dgl_PR-4648\build\lib\Release\gtest_main.lib" /NOLOGO /MACHINE:X64  /machine:x64 gtest_main.dir\Release\gtest_main.obj
         gtest_main.vcxproj -> C:\Jenkins\workspace\dgl_PR-4648\build\lib\Release\gtest_main.lib
       FinalizeBuildStatus:
         Deleting file "gtest_main.dir\Release\gtest_main.tlog\unsuccessfulbuild".
         Touching "gtest_main.dir\Release\gtest_main.tlog\gtest_main.lastbuildstate".
    17>Done Building Project "C:\Jenkins\workspace\dgl_PR-4648\build\third_party\googletest\googletest\gtest_main.vcxproj" (default targets).
     8>Done Building Project "C:\Jenkins\workspace\dgl_PR-4648\build\third_party\googletest\googletest\gtest_main.vcxproj.metaproj" (default targets).
CMake Warning at /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/torch/share/cmake/Torch/TorchConfig.cmake:22 (message):
  static library kineto_LIBRARY-NOTFOUND not found.
Call Stack (most recent call first):
  /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/torch/share/cmake/Torch/TorchConfig.cmake:127 (append_torchlib_if_found)
  CMakeLists.txt:26 (find_package)


-- Found Torch: /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/torch/lib/libtorch.so  
-- tensoradapter found PyTorch includes: /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/torch/include;/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/torch/include/torch/csrc/api/include
-- tensoradapter found PyTorch lib: torch
-- Configured target tensoradapter_pytorch_1.9.0
-- Configuring done
-- Generating done
CMake Warning:
  Manually-specified variables were not used by the project:

    CUDA_TOOLKIT_ROOT_DIR
    TORCH_CUDA_ARCH_LIST


-- Build files have been written to: /root/jenkins/workspace/dgl_PR-4648/tensoradapter/pytorch/build
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/broadcast_prod_i64.o
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/broadcast_prod_u64.o
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/broadcast_prod_f16.o
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_reduce_min_u8.o
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_reduce_min_i32.o
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_reduce_min_u32.o
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_reduce_min_i64.o
Scanning dependencies of target tensoradapter_pytorch_1.9.0
[ 50%] Building CXX object CMakeFiles/tensoradapter_pytorch_1.9.0.dir/torch.cpp.o
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/broadcast_prod_f32.o
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/broadcast_prod_f64.o
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/broadcast_min_i8.o
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/broadcast_min_u8.o
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/broadcast_min_i32.o
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/broadcast_min_u32.o
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/broadcast_min_i64.o
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_reduce_min_u64.o
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_reduce_min_f16.o
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_reduce_min_f32.o
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_reduce_min_f64.o
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_reduce_min_bf16.o
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_reduce_max_i8.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/broadcast_min_u64.o
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/broadcast_min_f16.o
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/broadcast_min_f32.o
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_reduce_max_u8.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_reduce_max_i32.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_reduce_max_u32.o
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\array_nonzero.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\array_nonzero.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\array_nonzero.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\array_nonzero.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\array_cumsum.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\array_cumsum.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\array_cumsum.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\array_cumsum.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\array_index_select.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\array_index_select.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\array_index_select.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\array_index_select.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\array_scatter.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\array_scatter.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\array_scatter.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\array_scatter.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\coo_coalesce.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\coo_coalesce.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\coo_coalesce.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\coo_coalesce.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\array_sort.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\array_sort.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\array_sort.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\array_sort.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\array_repeat.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\array_repeat.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\array_repeat.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\array_repeat.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\coo_linegraph.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\coo_linegraph.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\coo_linegraph.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\coo_linegraph.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         kernel.cc
         libra_partition.cc
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\array_pack.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\array_pack.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\array_pack.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\array_pack.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\array_op_impl.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\array_op_impl.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\array_op_impl.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\array_op_impl.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\gather_mm.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\gather_mm.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\gather_mm.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\gather_mm.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\coo_sort.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\coo_sort.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\coo_sort.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\coo_sort.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\csr_to_simple.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\csr_to_simple.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\csr_to_simple.cc)
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_reduce_max_i64.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_reduce_max_u64.o
[ 48%] Linking CXX static library ../../../lib/libgtest.a
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/broadcast_min_f64.o
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/broadcast_max_i8.o
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/broadcast_max_u8.o
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/broadcast_max_i32.o
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/broadcast_max_u32.o
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\csr_to_simple.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\csr_union.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\csr_union.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\csr_union.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\csr_union.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         uvm_array.cc
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\disjoint_union.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\disjoint_union.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\disjoint_union.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\disjoint_union.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         bcast.cc
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\csr_sort.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\csr_sort.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\csr_sort.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\csr_sort.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         c_api_common.cc
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\rowwise_topk.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\rowwise_topk.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\rowwise_topk.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\rowwise_topk.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/broadcast_max_i64.o
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/broadcast_max_u64.o
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/broadcast_max_f16.o
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_reduce_max_f16.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
[ 48%] Built target gtest
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_reduce_max_f32.o
[ 49%] Building CXX object third_party/googletest/googlemock/CMakeFiles/gmock.dir/src/gmock-all.cc.o
[100%] Linking CXX shared library libtensoradapter_pytorch_1.9.0.so
         geometry_op_impl.cc
         geometry.cc
         ndarray_partition.cc
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\segment_reduce.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\segment_reduce.cc)
         choice.cc
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\segment_reduce.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\segment_reduce.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\negative_sampling.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\negative_sampling.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\negative_sampling.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\negative_sampling.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         random.cc
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\filter.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\filter.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\filter.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\filter.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/broadcast_max_f32.o
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/broadcast_max_f64.o
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/broadcast_premulsum_i8.o
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/broadcast_premulsum_u8.o
[ 49%] Building CXX object third_party/googletest/googletest/CMakeFiles/gtest_main.dir/src/gtest_main.cc.o
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_reduce_max_f64.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_reduce_max_bf16.o
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_reduce_premulsum_i8.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
================================================================================
LIBXSMM master-1.16.1-1534 (Linux@4e33dd6e703e)
--------------------------------------------------------------------------------
GNU Compiler Collection: gcc 7.5.0, and g++ 7.5.0
C / C++ target: -msse4.2
Fortran Compiler is disabled or missing: no Fortran interface is built!
--------------------------------------------------------------------------------
--- LIBXSMM build log
[100%] Built target tensoradapter_pytorch_1.9.0
'libtensoradapter_pytorch_1.9.0.so' -> '/root/jenkins/workspace/dgl_PR-4648/build/tensoradapter/pytorch/libtensoradapter_pytorch_1.9.0.so'
[ 50%] Built target tensoradapter_pytorch
         c_object_api.cc
         c_runtime_api.cc
         config.cc
         cpu_device_api.cc
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\rowwise_sampling.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\rowwise_sampling.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\rowwise_sampling.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\rowwise_sampling.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         dlpack_convert.cc
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\array_arith.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\array_arith.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\array_arith.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\array_arith.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         dso_module.cc
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/broadcast_premulsum_i32.o
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/broadcast_premulsum_u32.o
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/broadcast_premulsum_i64.o
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_reduce_premulsum_u8.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_reduce_premulsum_i32.o
         file_util.cc
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\csr_get_data.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\csr_get_data.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\csr_get_data.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\csr_get_data.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         module.cc
         module_util.cc
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\coo_remove.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\coo_remove.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\coo_remove.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\coo_remove.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\utility(151,1): warning C4267: 'initializing': conversion from 'size_t' to '_Ty2', possible loss of data [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _Ty2=int32_t
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\coo_remove.cc)
       C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\array_utils.h(57): message : see reference to function template instantiation 'std::pair<K,V>::pair<const IdType&,size_t,0>(_Other1,_Other2 &&) noexcept' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     K=int32_t,
                     V=int32_t,
                     IdType=int32_t,
                     _Other1=const int32_t &,
                     _Other2=size_t
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\coo_remove.cc)
       C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\array_utils.h(57): message : see reference to function template instantiation 'std::pair<K,V>::pair<const IdType&,size_t,0>(_Other1,_Other2 &&) noexcept' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     K=int32_t,
                     V=int32_t,
                     IdType=int32_t,
                     _Other1=const int32_t &,
                     _Other2=size_t
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\coo_remove.cc)
       C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\array_utils.h(50): message : while compiling class template member function 'void dgl::aten::IdHashMap<IdType>::Update(dgl::IdArray)' [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     IdType=int32_t
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\coo_remove.cc)
       C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\array_utils.h(38): message : see reference to function template instantiation 'void dgl::aten::IdHashMap<IdType>::Update(dgl::IdArray)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     IdType=int32_t
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\coo_remove.cc)
       C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\coo_remove.cc(62): message : see reference to class template instantiation 'dgl::aten::IdHashMap<IdType>' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     IdType=int32_t
                 ]
       C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\coo_remove.cc(89): message : see reference to function template instantiation 'void dgl::aten::impl::`anonymous-namespace'::COORemoveShuffled<kDGLCPU,IdType>(dgl::aten::COOMatrix,dgl::IdArray,std::vector<IdType,std::allocator<IdType>> *,std::vector<IdType,std::allocator<IdType>> *,std::vector<IdType,std::allocator<IdType>> *)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     IdType=int32_t
                 ]
       C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\coo_remove.cc(101): message : see reference to function template instantiation 'dgl::aten::COOMatrix dgl::aten::impl::COORemove<kDGLCPU,int32_t>(dgl::aten::COOMatrix,dgl::IdArray)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\csr_remove.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\csr_remove.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\csr_remove.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\csr_remove.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/broadcast_premulsum_u64.o
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/broadcast_premulsum_f16.o
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/broadcast_premulsum_f32.o
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_reduce_premulsum_u32.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_reduce_premulsum_i64.o
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_reduce_premulsum_u64.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\utility(151,1): warning C4267: 'initializing': conversion from 'size_t' to '_Ty2', possible loss of data [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _Ty2=int32_t
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\csr_remove.cc)
       C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\array_utils.h(57): message : see reference to function template instantiation 'std::pair<K,V>::pair<const IdType&,size_t,0>(_Other1,_Other2 &&) noexcept' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     K=int32_t,
                     V=int32_t,
                     IdType=int32_t,
                     _Other1=const int32_t &,
                     _Other2=size_t
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\csr_remove.cc)
       C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\array_utils.h(57): message : see reference to function template instantiation 'std::pair<K,V>::pair<const IdType&,size_t,0>(_Other1,_Other2 &&) noexcept' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     K=int32_t,
                     V=int32_t,
                     IdType=int32_t,
                     _Other1=const int32_t &,
                     _Other2=size_t
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\csr_remove.cc)
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\csr_mm.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\csr_mm.cc)
       C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\array_utils.h(50): message : while compiling class template member function 'void dgl::aten::IdHashMap<IdType>::Update(dgl::IdArray)' [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     IdType=int32_t
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\csr_remove.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\csr_mm.cc)
       C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\array_utils.h(38): message : see reference to function template instantiation 'void dgl::aten::IdHashMap<IdType>::Update(dgl::IdArray)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     IdType=int32_t
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\csr_remove.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\csr_mm.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\csr_remove.cc(63): message : see reference to class template instantiation 'dgl::aten::IdHashMap<IdType>' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     IdType=int32_t
                 ]
       C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\csr_remove.cc(94): message : see reference to function template instantiation 'void dgl::aten::impl::`anonymous-namespace'::CSRRemoveShuffled<kDGLCPU,IdType>(dgl::aten::CSRMatrix,dgl::IdArray,std::vector<IdType,std::allocator<IdType>> *,std::vector<IdType,std::allocator<IdType>> *,std::vector<IdType,std::allocator<IdType>> *)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     IdType=int32_t
                 ]
       C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\csr_remove.cc(106): message : see reference to function template instantiation 'dgl::aten::CSRMatrix dgl::aten::impl::CSRRemove<kDGLCPU,int32_t>(dgl::aten::CSRMatrix,dgl::IdArray)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         ndarray.cc
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(4088,24): warning C4244: '=': conversion from 'const _Ty' to 'IdType', possible loss of data [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _Ty=int64_t
                 ]
                 and
                 [
                     IdType=int32_t
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\spmat_op_impl_csr.cc)
       C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\spmat_op_impl_csr.cc(298): message : see reference to function template instantiation 'void std::fill<IdType*,int64_t>(const _FwdIt,const _FwdIt,const _Ty &)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     IdType=int32_t,
                     _FwdIt=int32_t *,
                     _Ty=int64_t
                 ]
       C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\spmat_op_impl_csr.cc(306): message : see reference to function template instantiation 'dgl::aten::COOMatrix dgl::aten::impl::CSRToCOO<kDGLCPU,int32_t>(dgl::aten::CSRMatrix)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\csr_sum.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\csr_sum.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\csr_sum.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\csr_sum.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         object.cc
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\utility(151,1): warning C4267: 'initializing': conversion from 'size_t' to '_Ty2', possible loss of data [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _Ty2=int32_t
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\spmat_op_impl_csr.cc)
       C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\array_utils.h(57): message : see reference to function template instantiation 'std::pair<K,V>::pair<const IdType&,size_t,0>(_Other1,_Other2 &&) noexcept' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     K=int32_t,
                     V=int32_t,
                     IdType=int32_t,
                     _Other1=const int32_t &,
                     _Other2=size_t
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\spmat_op_impl_csr.cc)
       C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\array_utils.h(57): message : see reference to function template instantiation 'std::pair<K,V>::pair<const IdType&,size_t,0>(_Other1,_Other2 &&) noexcept' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     K=int32_t,
                     V=int32_t,
                     IdType=int32_t,
                     _Other1=const int32_t &,
                     _Other2=size_t
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\spmat_op_impl_csr.cc)
       C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\array_utils.h(50): message : while compiling class template member function 'void dgl::aten::IdHashMap<IdType>::Update(dgl::IdArray)' [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     IdType=int32_t
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\spmat_op_impl_csr.cc)
       C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\array_utils.h(38): message : see reference to function template instantiation 'void dgl::aten::IdHashMap<IdType>::Update(dgl::IdArray)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     IdType=int32_t
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\spmat_op_impl_csr.cc)
       C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\spmat_op_impl_csr.cc(477): message : see reference to class template instantiation 'dgl::aten::IdHashMap<IdType>' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     IdType=int32_t
                 ]
       C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\spmat_op_impl_csr.cc(525): message : see reference to function template instantiation 'dgl::aten::CSRMatrix dgl::aten::impl::CSRSliceMatrix<kDGLCPU,int32_t>(dgl::aten::CSRMatrix,dgl::runtime::NDArray,dgl::runtime::NDArray)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\array.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\array.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\array.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\array.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/broadcast_premulsum_f64.o
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/broadcast_sumpostdiv_i8.o
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/broadcast_sumpostdiv_u8.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_reduce_premulsum_f16.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_reduce_premulsum_f32.o
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\spmat_op_impl_csr.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\spmat_op_impl_csr.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\spmat_op_impl_csr.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\spmat_op_impl_csr.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         registry.cc
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\spmm.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\spmm.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\spmm.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\spmm.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\sddmm.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\sddmm.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\sddmm.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\sddmm.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\utility(151,1): warning C4267: 'initializing': conversion from 'size_t' to '_Ty2', possible loss of data [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _Ty2=int32_t
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\spmat_op_impl_coo.cc)
       C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\array_utils.h(57): message : see reference to function template instantiation 'std::pair<IdType,IdType>::pair<const IdType&,size_t,0>(_Other1,_Other2 &&) noexcept' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     IdType=int32_t,
                     _Other1=const int32_t &,
                     _Other2=size_t
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\spmat_op_impl_coo.cc)
       C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\array_utils.h(57): message : see reference to function template instantiation 'std::pair<IdType,IdType>::pair<const IdType&,size_t,0>(_Other1,_Other2 &&) noexcept' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     IdType=int32_t,
                     _Other1=const int32_t &,
                     _Other2=size_t
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\spmat_op_impl_coo.cc)
       C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\array_utils.h(50): message : while compiling class template member function 'void dgl::aten::IdHashMap<IdType>::Update(dgl::IdArray)' [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     IdType=int32_t
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\spmat_op_impl_coo.cc)
       C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\array_utils.h(38): message : see reference to function template instantiation 'void dgl::aten::IdHashMap<IdType>::Update(dgl::IdArray)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     IdType=int32_t
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\spmat_op_impl_coo.cc)
       C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\spmat_op_impl_coo.cc(684): message : see reference to class template instantiation 'dgl::aten::IdHashMap<IdType>' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     IdType=int32_t
                 ]
       C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\spmat_op_impl_coo.cc(706): message : see reference to function template instantiation 'dgl::aten::COOMatrix dgl::aten::impl::COOSliceRows<kDGLCPU,int32_t>(dgl::aten::COOMatrix,dgl::runtime::NDArray)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_reduce_premulsum_f64.o
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_reduce_premulsum_bf16.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/broadcast_sumpostdiv_i32.o
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/broadcast_sumpostdiv_u32.o
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/broadcast_sumpostdiv_i64.o
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\spmat_op_impl_coo.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\spmat_op_impl_coo.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\spmat_op_impl_coo.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\spmat_op_impl_coo.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\numeric(817,20): warning C4244: '=': conversion from '_Ty' to 'IdType', possible loss of data [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _Ty=int64_t
                 ]
                 and
                 [
                     IdType=int32_t
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\spmat_op_impl_coo.cc)
       C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\spmat_op_impl_coo.cc(383): message : see reference to function template instantiation 'void std::iota<IdType*,int64_t>(_FwdIt,_FwdIt,_Ty)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     IdType=int32_t,
                     _FwdIt=int32_t *,
                     _Ty=int64_t
                 ]
       C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\spmat_op_impl_coo.cc(632): message : see reference to function template instantiation 'dgl::aten::CSRMatrix dgl::aten::impl::`anonymous-namespace'::SortedCOOToCSR<IdType>(const dgl::aten::COOMatrix &)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     IdType=int32_t
                 ]
       C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\spmat_op_impl_coo.cc(635): message : see reference to function template instantiation 'dgl::aten::CSRMatrix dgl::aten::impl::COOToCSR<kDGLCPU,int32_t>(dgl::aten::COOMatrix)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\utility(151,1): warning C4244: 'initializing': conversion from '_Ty' to '_Ty2', possible loss of data [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _Ty=__int64
                 ]
                 and
                 [
                     _Ty2=int32_t
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\spmat_op_impl_coo.cc)
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xmemory(671): message : see reference to function template instantiation 'std::pair<const std::pair<IdType,IdType>,IdType>::pair<_Ty,__int64,0>(_Other1 &&,_Other2 &&) noexcept' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     IdType=int32_t,
                     _Ty=std::pair<int32_t,int32_t>,
                     _Other1=std::pair<int32_t,int32_t>,
                     _Other2=__int64
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\spmat_op_impl_coo.cc)
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xmemory(671): message : see reference to function template instantiation 'std::pair<const std::pair<IdType,IdType>,IdType>::pair<_Ty,__int64,0>(_Other1 &&,_Other2 &&) noexcept' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     IdType=int32_t,
                     _Ty=std::pair<int32_t,int32_t>,
                     _Other1=std::pair<int32_t,int32_t>,
                     _Other2=__int64
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\spmat_op_impl_coo.cc)
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\list(587): message : see reference to function template instantiation 'void std::_Default_allocator_traits<_Alloc>::construct<_Ty,std::pair<IdType,IdType>,__int64>(_Alloc &,_Objty *const ,std::pair<IdType,IdType> &&,__int64 &&)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _Alloc=std::allocator<std::_List_node<std::pair<const std::pair<int32_t,int32_t>,int32_t>,std::_Default_allocator_traits<std::allocator<std::pair<const std::pair<int32_t,int32_t>,int32_t>>>::void_pointer>>,
                     _Ty=std::pair<const std::pair<int32_t,int32_t>,int32_t>,
                     IdType=int32_t,
                     _Objty=std::pair<const std::pair<int32_t,int32_t>,int32_t>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\spmat_op_impl_coo.cc)
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\list(587): message : see reference to function template instantiation 'void std::_Default_allocator_traits<_Alloc>::construct<_Ty,std::pair<IdType,IdType>,__int64>(_Alloc &,_Objty *const ,std::pair<IdType,IdType> &&,__int64 &&)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _Alloc=std::allocator<std::_List_node<std::pair<const std::pair<int32_t,int32_t>,int32_t>,std::_Default_allocator_traits<std::allocator<std::pair<const std::pair<int32_t,int32_t>,int32_t>>>::void_pointer>>,
                     _Ty=std::pair<const std::pair<int32_t,int32_t>,int32_t>,
                     IdType=int32_t,
                     _Objty=std::pair<const std::pair<int32_t,int32_t>,int32_t>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\spmat_op_impl_coo.cc)
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xhash(599): message : see reference to function template instantiation 'std::_List_node_emplace_op2<std::allocator<std::_List_node<_Ty,std::_Default_allocator_traits<_Alloc>::void_pointer>>>::_List_node_emplace_op2<std::pair<IdType,IdType>,__int64>(_Alnode &,std::pair<IdType,IdType> &&,__int64 &&)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _Ty=std::pair<const std::pair<int32_t,int32_t>,int32_t>,
                     _Alloc=std::allocator<std::pair<const std::pair<int32_t,int32_t>,int32_t>>,
                     IdType=int32_t,
                     _Alnode=std::allocator<std::_List_node<std::pair<const std::pair<int32_t,int32_t>,int32_t>,std::_Default_allocator_traits<std::allocator<std::pair<const std::pair<int32_t,int32_t>,int32_t>>>::void_pointer>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\spmat_op_impl_coo.cc)
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xhash(599): message : see reference to function template instantiation 'std::_List_node_emplace_op2<std::allocator<std::_List_node<_Ty,std::_Default_allocator_traits<_Alloc>::void_pointer>>>::_List_node_emplace_op2<std::pair<IdType,IdType>,__int64>(_Alnode &,std::pair<IdType,IdType> &&,__int64 &&)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _Ty=std::pair<const std::pair<int32_t,int32_t>,int32_t>,
                     _Alloc=std::allocator<std::pair<const std::pair<int32_t,int32_t>,int32_t>>,
                     IdType=int32_t,
                     _Alnode=std::allocator<std::_List_node<std::pair<const std::pair<int32_t,int32_t>,int32_t>,std::_Default_allocator_traits<std::allocator<std::pair<const std::pair<int32_t,int32_t>,int32_t>>>::void_pointer>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\spmat_op_impl_coo.cc)
       C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\spmat_op_impl_coo.cc(256): message : see reference to function template instantiation 'std::_List_iterator<std::_List_val<std::_List_simple_types<_Ty>>> std::_Hash<std::_Umap_traits<_Kty,int,std::_Uhash_compare<_Kty,_Hasher,_Keyeq>,_Alloc,true>>::emplace<std::pair<IdType,IdType>,int64_t>(std::pair<IdType,IdType> &&,int64_t &&)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _Ty=std::pair<const std::pair<int32_t,int32_t>,int32_t>,
                     _Kty=std::pair<int32_t,int32_t>,
                     _Hasher=dgl::aten::PairHash,
                     _Keyeq=std::equal_to<std::pair<int32_t,int32_t>>,
                     _Alloc=std::allocator<std::pair<const std::pair<int32_t,int32_t>,int32_t>>,
                     IdType=int32_t
                 ]
       C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\spmat_op_impl_coo.cc(256): message : see reference to function template instantiation 'std::_List_iterator<std::_List_val<std::_List_simple_types<_Ty>>> std::_Hash<std::_Umap_traits<_Kty,int,std::_Uhash_compare<_Kty,_Hasher,_Keyeq>,_Alloc,true>>::emplace<std::pair<IdType,IdType>,int64_t>(std::pair<IdType,IdType> &&,int64_t &&)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _Ty=std::pair<const std::pair<int32_t,int32_t>,int32_t>,
                     _Kty=std::pair<int32_t,int32_t>,
                     _Hasher=dgl::aten::PairHash,
                     _Keyeq=std::equal_to<std::pair<int32_t,int32_t>>,
                     _Alloc=std::allocator<std::pair<const std::pair<int32_t,int32_t>,int32_t>>,
                     IdType=int32_t
                 ]
       C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\spmat_op_impl_coo.cc(290): message : see reference to function template instantiation 'std::vector<dgl::IdArray,std::allocator<dgl::IdArray>> dgl::aten::impl::COOGetDataAndIndices<kDGLCPU,int32_t>(dgl::aten::COOMatrix,dgl::runtime::NDArray,dgl::runtime::NDArray)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_reduce_sumpostdiv_i8.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_reduce_sumpostdiv_u8.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/broadcast_sumpostdiv_u64.o
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/broadcast_sumpostdiv_f16.o
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/broadcast_sumpostdiv_f32.o
         resource_manager.cc
         semaphore_wrapper.cc
         shared_mem.cc
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_reduce_sumpostdiv_i32.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_reduce_sumpostdiv_u32.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/broadcast_sumpostdiv_f64.o
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_sum_i8.o
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_sum_u8.o
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_sum_i32.o
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_reduce_sumpostdiv_i64.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_reduce_sumpostdiv_u64.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_sum_u32.o
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_sum_i64.o
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_sum_u64.o
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_reduce_sumpostdiv_f16.o
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_reduce_sumpostdiv_f32.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_reduce_sumpostdiv_f64.o
[ 49%] Linking CXX static library ../../../lib/libgtest_main.a
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_sum_f16.o
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_sum_f32.o
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_sum_f64.o
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_prod_i8.o
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_prod_u8.o
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_prod_i32.o
================================================================================
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_prod_u32.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  all_reduce.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_reduce_sumpostdiv_bf16.o
[ 49%] Built target gtest_main
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_gather_sum_i8.o
         system_lib_module.cc
         tensordispatch.cc
         thread_pool.cc
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_gather_sum_u8.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_gather_sum_i32.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_prod_i64.o
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_prod_u64.o
LIBXSMM master-1.16.1-1534 (Linux@681e11859cd1)
--------------------------------------------------------------------------------
GNU Compiler Collection: gcc 7.5.0, and g++ 7.5.0
C / C++ target: -msse4.2
Fortran Compiler is disabled or missing: no Fortran interface is built!
--------------------------------------------------------------------------------
--- LIBXSMM build log
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_prod_f16.o
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_prod_f32.o
         threading_backend.cc
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_prod_f64.o
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_min_i8.o
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_min_u8.o
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_gather_sum_u32.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_gather_sum_i64.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_gather_sum_u64.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_gather_sum_f16.o
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_min_i32.o
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_min_u32.o
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_min_i64.o
         utils.cc
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_min_u64.o
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_min_f16.o
[ 46%] Linking CXX static library ../../../lib/libgtest.a
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_min_f32.o
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_min_f64.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_gather_sum_f32.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_gather_sum_f64.o
[ 50%] Linking CXX static library ../../../lib/libgtest.a
[ 50%] Built target gtest
Scanning dependencies of target gtest_main
Scanning dependencies of target gmock
[ 50%] Building CXX object third_party/googletest/googletest/CMakeFiles/gtest_main.dir/src/gtest_main.cc.o
[ 51%] Building CXX object third_party/googletest/googlemock/CMakeFiles/gmock.dir/src/gmock-all.cc.o
         workspace_pool.cc
    18>C:\Jenkins\workspace\dgl_PR-4648\src\array\libra_partition.cc(396,16): warning C4477: 'fscanf' : format string '%ld' requires an argument of type 'long *', but variadic argument 1 has type 'int64_t *' [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Jenkins\workspace\dgl_PR-4648\src\array\libra_partition.cc(396,16): message : consider using '%lld' in the format string [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Jenkins\workspace\dgl_PR-4648\src\array\libra_partition.cc(396,16): message : consider using '%Id' in the format string [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Jenkins\workspace\dgl_PR-4648\src\array\libra_partition.cc(396,16): message : consider using '%I64d' in the format string [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
    18>C:\Jenkins\workspace\dgl_PR-4648\src\array\libra_partition.cc(396,16): warning C4477: 'fscanf' : format string '%ld' requires an argument of type 'long *', but variadic argument 2 has type 'int64_t *' [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Jenkins\workspace\dgl_PR-4648\src\array\libra_partition.cc(396,16): message : consider using '%lld' in the format string [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Jenkins\workspace\dgl_PR-4648\src\array\libra_partition.cc(396,16): message : consider using '%Id' in the format string [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Jenkins\workspace\dgl_PR-4648\src\array\libra_partition.cc(396,16): message : consider using '%I64d' in the format string [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\uvm_array.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\uvm_array.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\uvm_array.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\uvm_array.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
[ 46%] Built target gtest
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_max_i8.o
Scanning dependencies of target gtest_main
Scanning dependencies of target gmock
[ 46%] Building CXX object third_party/googletest/googletest/CMakeFiles/gtest_main.dir/src/gtest_main.cc.o
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_max_u8.o
[ 48%] Building CXX object third_party/googletest/googlemock/CMakeFiles/gmock.dir/src/gmock-all.cc.o
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_max_i32.o
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_gather_sum_bf16.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_gather_prod_i8.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\c_api_common.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\c_api_common.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\c_api_common.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\c_api_common.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         api_container.cc
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\geometry\cpu\geometry_op_impl.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\geometry\cpu\geometry_op_impl.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\geometry\cpu\geometry_op_impl.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\geometry\cpu\geometry_op_impl.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         api_test.cc
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\partition\ndarray_partition.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\partition\ndarray_partition.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\partition\ndarray_partition.cc)
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\geometry\geometry.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\geometry\geometry.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\partition\ndarray_partition.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\geometry\geometry.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\geometry\geometry.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_max_u32.o
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_max_i64.o
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_max_u64.o
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_gather_prod_u8.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_gather_prod_i32.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_gather_prod_u32.o
[ 53%] Linking CXX static library ../../../lib/libgtest_main.a
[ 53%] Built target gtest_main
         creators.cc
    18>C:\Jenkins\workspace\dgl_PR-4648\src\array\libra_partition.cc(273,21): warning C4477: 'fprintf' : format string '%ld' requires an argument of type 'long', but variadic argument 1 has type 'int64_t' [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Jenkins\workspace\dgl_PR-4648\src\array\libra_partition.cc(273,21): message : consider using '%lld' in the format string [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Jenkins\workspace\dgl_PR-4648\src\array\libra_partition.cc(273,21): message : consider using '%Id' in the format string [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Jenkins\workspace\dgl_PR-4648\src\array\libra_partition.cc(273,21): message : consider using '%I64d' in the format string [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Jenkins\workspace\dgl_PR-4648\src\array\libra_partition.cc(317): message : see reference to function template instantiation 'void dgl::aten::LibraVertexCut<IdType,IdType2>(int32_t,dgl::runtime::NDArray,dgl::runtime::NDArray,dgl::runtime::NDArray,dgl::runtime::NDArray,dgl::runtime::NDArray,dgl::runtime::NDArray,dgl::runtime::NDArray,int64_t,int64_t,const std::string &)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
    18>C:\Jenkins\workspace\dgl_PR-4648\src\array\libra_partition.cc(273,21): warning C4477: 'fprintf' : format string '%ld' requires an argument of type 'long', but variadic argument 2 has type 'int64_t' [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Jenkins\workspace\dgl_PR-4648\src\array\libra_partition.cc(273,21): message : consider using '%lld' in the format string [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Jenkins\workspace\dgl_PR-4648\src\array\libra_partition.cc(273,21): message : consider using '%Id' in the format string [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Jenkins\workspace\dgl_PR-4648\src\array\libra_partition.cc(273,21): message : consider using '%I64d' in the format string [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
    18>C:\Jenkins\workspace\dgl_PR-4648\src\array\libra_partition.cc(273,21): warning C4477: 'fprintf' : format string '%ld' requires an argument of type 'long', but variadic argument 3 has type 'int64_t' [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Jenkins\workspace\dgl_PR-4648\src\array\libra_partition.cc(273,21): message : consider using '%lld' in the format string [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Jenkins\workspace\dgl_PR-4648\src\array\libra_partition.cc(273,21): message : consider using '%Id' in the format string [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Jenkins\workspace\dgl_PR-4648\src\array\libra_partition.cc(273,21): message : consider using '%I64d' in the format string [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
    18>C:\Jenkins\workspace\dgl_PR-4648\src\array\libra_partition.cc(284,10): warning C4477: 'printf' : format string '%ld' requires an argument of type 'long', but variadic argument 1 has type 'unsigned __int64' [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Jenkins\workspace\dgl_PR-4648\src\array\libra_partition.cc(284,10): message : consider using '%zd' in the format string [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
    18>C:\Jenkins\workspace\dgl_PR-4648\src\array\libra_partition.cc(287,17): warning C4477: 'fprintf' : format string '%ld' requires an argument of type 'long', but variadic argument 1 has type 'int64_t' [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Jenkins\workspace\dgl_PR-4648\src\array\libra_partition.cc(287,17): message : consider using '%lld' in the format string [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Jenkins\workspace\dgl_PR-4648\src\array\libra_partition.cc(287,17): message : consider using '%Id' in the format string [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Jenkins\workspace\dgl_PR-4648\src\array\libra_partition.cc(287,17): message : consider using '%I64d' in the format string [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
    18>C:\Jenkins\workspace\dgl_PR-4648\src\array\libra_partition.cc(291,12): warning C4477: 'printf' : format string '%ld' requires an argument of type 'long', but variadic argument 1 has type 'int64_t' [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Jenkins\workspace\dgl_PR-4648\src\array\libra_partition.cc(291,12): message : consider using '%lld' in the format string [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Jenkins\workspace\dgl_PR-4648\src\array\libra_partition.cc(291,12): message : consider using '%Id' in the format string [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Jenkins\workspace\dgl_PR-4648\src\array\libra_partition.cc(291,12): message : consider using '%I64d' in the format string [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
    18>C:\Jenkins\workspace\dgl_PR-4648\src\array\libra_partition.cc(296,12): warning C4477: 'printf' : format string '%ld' requires an argument of type 'long', but variadic argument 1 has type 'int64_t' [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Jenkins\workspace\dgl_PR-4648\src\array\libra_partition.cc(296,12): message : consider using '%lld' in the format string [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Jenkins\workspace\dgl_PR-4648\src\array\libra_partition.cc(296,12): message : consider using '%Id' in the format string [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Jenkins\workspace\dgl_PR-4648\src\array\libra_partition.cc(296,12): message : consider using '%I64d' in the format string [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\random\random.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\random\random.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\random\random.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\random\random.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         gk_ops.cc
         graph.cc
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_max_f16.o
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_max_f32.o
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_max_f64.o
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_premulsum_i8.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_gather_prod_i64.o
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_gather_prod_u64.o
/usr/bin/ar: creating lib/libxsmmnoblas.a
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
         graph_apis.cc
         graph_op.cc
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(4088,24): warning C4244: '=': conversion from 'const _Ty' to 'float', possible loss of data [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _Ty=int
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\random\cpu\choice.cc)
       C:\Jenkins\workspace\dgl_PR-4648\src\random\cpu\sample_utils.h(267): message : see reference to function template instantiation 'void std::fill<std::_Vector_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int>(const _FwdIt,const _FwdIt,const int &)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _Ty=ValueType,
                     _FwdIt=std::_Vector_iterator<std::_Vector_val<std::_Simple_types<ValueType>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\random\cpu\choice.cc)
       C:\Jenkins\workspace\dgl_PR-4648\src\random\cpu\sample_utils.h(264): message : while compiling class template member function 'void dgl::utils::TreeSampler<IdxType,FloatType,false>::ResetState(dgl::FloatArray)' [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     IdxType=int64_t,
                     FloatType=float
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\random\cpu\choice.cc)
       C:\Jenkins\workspace\dgl_PR-4648\src\random\cpu\sample_utils.h(281): message : see reference to function template instantiation 'void dgl::utils::TreeSampler<IdxType,FloatType,false>::ResetState(dgl::FloatArray)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     IdxType=int64_t,
                     FloatType=float
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\random\cpu\choice.cc)
       C:\Jenkins\workspace\dgl_PR-4648\src\random\cpu\choice.cc(42): message : see reference to class template instantiation 'dgl::utils::TreeSampler<IdxType,FloatType,false>' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     IdxType=int64_t,
                     FloatType=float
                 ]
       C:\Jenkins\workspace\dgl_PR-4648\src\random\cpu\choice.cc(51): message : see reference to function template instantiation 'void dgl::RandomEngine::Choice<int64_t,float>(int64_t,dgl::FloatArray,int64_t *,bool)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         graph_traversal.cc
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_premulsum_u8.o
[ 48%] Linking CXX static library libtensorpipe.a
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_premulsum_i32.o
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_premulsum_u32.o
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_gather_prod_f16.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_gather_prod_f32.o
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\random\cpu\choice.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\random\cpu\choice.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\random\cpu\choice.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\random\cpu\choice.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         heterograph.cc
         heterograph_capi.cc
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\kernel.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\kernel.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\kernel.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\kernel.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_premulsum_i64.o
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_premulsum_u64.o
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_premulsum_f16.o
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_gather_prod_f64.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_gather_prod_bf16.o
         immutable_graph.cc
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\libra_partition.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\libra_partition.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\libra_partition.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\libra_partition.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         metis_partition.cc
         network.cc
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\stdint.h(51,1): warning C4005: 'INT32_MIN': macro redefinition (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\metis_partition.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Jenkins\workspace\dgl_PR-4648\third_party\METIS\include\metis.h(75): message : see previous definition of 'INT32_MIN' (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\metis_partition.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\stdint.h(52,1): warning C4005: 'INT64_MIN': macro redefinition (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\metis_partition.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Jenkins\workspace\dgl_PR-4648\third_party\METIS\include\metis.h(77): message : see previous definition of 'INT64_MIN' (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\metis_partition.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\stdint.h(55,1): warning C4005: 'INT32_MAX': macro redefinition (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\metis_partition.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Jenkins\workspace\dgl_PR-4648\third_party\METIS\include\metis.h(76): message : see previous definition of 'INT32_MAX' (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\metis_partition.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\stdint.h(56,1): warning C4005: 'INT64_MAX': macro redefinition (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\metis_partition.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Jenkins\workspace\dgl_PR-4648\third_party\METIS\include\metis.h(78): message : see previous definition of 'INT64_MAX' (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\metis_partition.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         nodeflow.cc
         pickle.cc
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_premulsum_f32.o
[ 48%] Built target tensorpipe
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_premulsum_f64.o
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_gather_min_i8.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_gather_min_u8.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_gather_min_i32.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_sumpostdiv_i8.o
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_sumpostdiv_u8.o
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_sumpostdiv_i32.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_gather_min_u32.o
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_gather_min_i64.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
         sampler.cc
         global_uniform.cc
         neighbor.cc
         get_node_types_cpu.cc
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_sumpostdiv_u32.o
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_sumpostdiv_i64.o
[ 48%] Linking CXX static library ../../../lib/libgtest_main.a
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_sumpostdiv_u64.o
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_gather_min_u64.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
================================================================================
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_sumpostdiv_f16.o
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_sumpostdiv_f32.o
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_sumpostdiv_f64.o
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_scatter_sum_i8.o
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_scatter_sum_u8.o
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_gather_min_f16.o
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_gather_min_f32.o
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_gather_min_f64.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_gather_min_bf16.o
LIBXSMM master-1.16.1-1534 (Linux@856b09e8bd12)
--------------------------------------------------------------------------------
GNU Compiler Collection: gcc 7.5.0, and g++ 7.5.0
C / C++ target: -msse4.2
Fortran Compiler is disabled or missing: no Fortran interface is built!
--------------------------------------------------------------------------------
--- LIBXSMM build log
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_scatter_sum_i32.o
[ 48%] Built target gtest_main
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_scatter_sum_u32.o
/usr/bin/ar: creating lib/libxsmmnoblas.a
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_scatter_sum_i64.o
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_gather_max_i8.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_gather_max_u8.o
         node2vec.cc
         node2vec_cpu.cc
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_scatter_sum_u64.o
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_scatter_sum_f16.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_gather_max_i32.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_gather_max_u32.o
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_scatter_sum_f32.o
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_scatter_sum_f64.o
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_scatter_prod_i8.o
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_gather_max_i64.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_gather_max_u64.o
         randomwalk_cpu.cc
         randomwalk_with_restart_cpu.cc
         randomwalks.cc
         dglgraph_serialize.cc
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_scatter_prod_u8.o
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_scatter_prod_i32.o
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_scatter_prod_u32.o
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_scatter_prod_i64.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_gather_max_f16.o
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_gather_max_f32.o
         graph_serialize.cc
         heterograph_serialize.cc
         tensor_serialize.cc
[ 53%] Linking CXX static library ../../../lib/libgmock.a
[ 53%] Built target gmock
/usr/bin/ar: creating lib/libxsmmnoblas.a
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_scatter_prod_u64.o
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_scatter_prod_f16.o
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_scatter_prod_f32.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_gather_max_f64.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_gather_max_bf16.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
         zerocopy_serializer.cc
         shared_mem_manager.cc
Scanning dependencies of target gmock_main
[ 53%] Building CXX object third_party/googletest/googlemock/CMakeFiles/gmock_main.dir/src/gmock_main.cc.o
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_scatter_prod_f64.o
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_scatter_min_i8.o
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_scatter_min_u8.o
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_gather_premulsum_i8.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\runtime\utils.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\runtime\utils.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\aten\./array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\runtime\utils.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl/aten/coo.h(235): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\runtime\utils.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_gather_premulsum_u8.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
[ 49%] Linking CXX static library ../../../lib/libgmock.a
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_gather_premulsum_i32.o
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_gather_premulsum_u32.o
[ 49%] Built target gmock
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_scatter_min_i32.o
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_scatter_min_u32.o
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_scatter_min_i64.o
         subgraph.cc
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\creators.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\creators.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\creators.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\creators.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\gk_ops.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\gk_ops.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\gk_ops.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\gk_ops.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
[ 53%] Linking CXX static library libtensorpipe.a
[ 53%] Built target tensorpipe
         compact.cc
         line_graph.cc
         metis_partition_hetero.cc
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\graph_traversal.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\graph_traversal.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\graph_traversal.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\graph_traversal.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
[ 49%] Building CXX object third_party/googletest/googlemock/CMakeFiles/gmock_main.dir/src/gmock_main.cc.o
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_gather_premulsum_i64.o
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_scatter_min_u64.o
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_scatter_min_f16.o
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_scatter_min_f32.o
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampling\randomwalks\get_node_types_cpu.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampling\randomwalks\get_node_types_cpu.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampling\randomwalks\get_node_types_cpu.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampling\randomwalks\get_node_types_cpu.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\graph.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\graph.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\graph.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\graph.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_gather_premulsum_u64.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_gather_premulsum_f16.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_scatter_min_f64.o
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_scatter_max_i8.o
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\metis_partition.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\metis_partition.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\metis_partition.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\metis_partition.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         partition_hetero.cc
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\nodeflow.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\nodeflow.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\nodeflow.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\nodeflow.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\graph_apis.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\graph_apis.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\graph_apis.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\graph_apis.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         remove_edges.cc
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampling\negative\global_uniform.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampling\negative\global_uniform.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampling\negative\global_uniform.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampling\negative\global_uniform.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         to_bipartite.cc
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\heterograph.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\heterograph.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\heterograph.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\heterograph.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_gather_premulsum_f32.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_gather_premulsum_f64.o
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_scatter_max_u8.o
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_scatter_max_i32.o
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\immutable_graph.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\immutable_graph.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\immutable_graph.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\immutable_graph.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const unsigned __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\heterograph.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const unsigned __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const unsigned __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\heterograph.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=uint64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<uint64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\heterograph.cc)
       C:\Jenkins\workspace\dgl_PR-4648\src\graph\heterograph.cc(511): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<uint64_t>(const std::vector<uint64_t,std::allocator<uint64_t>> &,uint8_t,DGLContext)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Jenkins\workspace\dgl_PR-4648\src\graph\heterograph.cc(414): message : see reference to function template instantiation 'dgl::FlattenedHeteroGraphPtr dgl::HeteroGraph::FlattenImpl<int32_t>(const std::vector<uint64_t,std::allocator<uint64_t>> &) const' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampling\randomwalks\node2vec.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampling\randomwalks\node2vec.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampling\randomwalks\node2vec.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampling\randomwalks\node2vec.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampling\neighbor\neighbor.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampling\neighbor\neighbor.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampling\neighbor\neighbor.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampling\neighbor\neighbor.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         to_simple.cc
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampling\randomwalks\node2vec_cpu.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampling\randomwalks\node2vec_cpu.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampling\randomwalks\node2vec_cpu.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampling\randomwalks\node2vec_cpu.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_gather_premulsum_bf16.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_scatter_max_u32.o
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_scatter_max_i64.o
         unit_graph.cc
    18>C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampler.cc(1526,1): warning C4805: '==': unsafe mix of type 'int64_t' and type 'bool' in operation [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\pickle.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\pickle.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\pickle.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\pickle.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         scheduler.cc
[ 54%] Linking CXX static library ../../../lib/libgmock_main.a
         scheduler_apis.cc
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xstddef(240,54): warning C4018: '<': signed/unsigned mismatch (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampling\randomwalks\node2vec_cpu.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xstddef(239): message : while compiling class template member function 'bool std::less<void>::operator ()<const _Ty&,const IdxType&>(_Ty1,_Ty2) noexcept(<expr>) const' [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _Ty=dgl::dgl_id_t,
                     IdxType=int64_t,
                     _Ty1=const dgl::dgl_id_t &,
                     _Ty2=const int64_t &
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampling\randomwalks\node2vec_cpu.cc)
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\algorithm(2951): message : see reference to function template instantiation 'bool std::binary_search<_FwdIt,_Ty,std::less<void>>(_FwdIt,_FwdIt,const _Ty &,_Pr)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _FwdIt=const int64_t *,
                     _Ty=dgl::dgl_id_t,
                     _Pr=std::less<void>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampling\randomwalks\node2vec_cpu.cc)
       C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampling\randomwalks\node2vec_randomwalk.h(45): message : see reference to function template instantiation 'bool std::binary_search<const IdxType*,dgl::dgl_id_t>(_FwdIt,_FwdIt,const _Ty &)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     IdxType=int64_t,
                     _FwdIt=const int64_t *,
                     _Ty=dgl::dgl_id_t
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampling\randomwalks\node2vec_cpu.cc)
       C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampling\randomwalks\node2vec_randomwalk.h(109): message : see reference to function template instantiation 'bool dgl::sampling::impl::`anonymous-namespace'::has_edge_between<IdxType>(const dgl::aten::CSRMatrix &,dgl::dgl_id_t,dgl::dgl_id_t)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     IdxType=int64_t
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampling\randomwalks\node2vec_cpu.cc)
       C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampling\randomwalks\node2vec_randomwalk.h(163): message : see reference to function template instantiation 'std::tuple<dgl::dgl_id_t,dgl::dgl_id_t,bool> dgl::sampling::impl::`anonymous-namespace'::Node2vecRandomWalkStep<kDGLCPU,IdxType>(IdxType *,dgl::dgl_id_t,dgl::dgl_id_t,const double,const double,int64_t,const dgl::aten::CSRMatrix &,bool,const dgl::FloatArray &,std::function<bool (IdxType *,dgl::dgl_id_t,int64_t)>)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     IdxType=int64_t
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampling\randomwalks\node2vec_cpu.cc)
       C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampling\randomwalks\node2vec_cpu.cc(31): message : see reference to function template instantiation 'std::pair<dgl::runtime::NDArray,dgl::IdArray> dgl::sampling::impl::`anonymous-namespace'::Node2vecRandomWalk<kDGLCPU,IdxType>(const dgl::HeteroGraphPtr,const dgl::IdArray,const double,const double,const int64_t,const dgl::FloatArray &,std::function<bool (IdxType *,dgl::dgl_id_t,int64_t)>)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     IdxType=int64_t
                 ]
       C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampling\randomwalks\node2vec_cpu.cc(45): message : see reference to function template instantiation 'std::pair<dgl::runtime::NDArray,dgl::IdArray> dgl::sampling::impl::Node2vec<kDGLCPU,int64_t>(const dgl::HeteroGraphPtr,const dgl::IdArray,const double,const double,const int64_t,const dgl::FloatArray &)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
    18>C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampler.cc(1751,1): warning C4805: '==': unsafe mix of type 'int64_t' and type 'bool' in operation [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampler.cc(1693): message : while compiling class template member function 'void dgl::WeightedEdgeSamplerObject<float>::Fetch(dgl::runtime::DGLRetValue *)' [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampler.cc(1925): message : see reference to function template instantiation 'void dgl::WeightedEdgeSamplerObject<float>::Fetch(dgl::runtime::DGLRetValue *)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampler.cc(1842): message : see reference to class template instantiation 'dgl::WeightedEdgeSamplerObject<float>' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xstddef(242,43): warning C4018: '<': signed/unsigned mismatch (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampling\randomwalks\node2vec_cpu.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\algorithm(2945): message : see reference to function template instantiation 'bool std::less<void>::operator ()<const _Ty&,const IdxType&>(_Ty1,_Ty2) noexcept const' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _Ty=dgl::dgl_id_t,
                     IdxType=int64_t,
                     _Ty1=const dgl::dgl_id_t &,
                     _Ty2=const int64_t &
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampling\randomwalks\node2vec_cpu.cc)
         common.cc
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_gather_sumpostdiv_i8.o
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_gather_sumpostdiv_u8.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_gather_sumpostdiv_i32.o
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_gather_sumpostdiv_u32.o
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_gather_sumpostdiv_i64.o
[ 54%] Built target gmock_main
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\graph_op.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\graph_op.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\graph_op.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\graph_op.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const unsigned __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\graph_op.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const unsigned __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const unsigned __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\graph_op.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=uint64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<uint64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\graph_op.cc)
       C:\Jenkins\workspace\dgl_PR-4648\src\graph\graph_op.cc(400): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<uint64_t>(const std::vector<uint64_t,std::allocator<uint64_t>> &,uint8_t,DGLContext)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_gather_sumpostdiv_u64.o
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_scatter_max_u64.o
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_scatter_max_f16.o
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_scatter_max_f32.o
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_scatter_max_f64.o
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_scatter_premulsum_i8.o
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_scatter_premulsum_u8.o
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_scatter_premulsum_i32.o
[ 54%] Linking CXX static library libdmlc.a
         msg_queue.cc
         socket_communicator.cc
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_gather_sumpostdiv_f16.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_gather_sumpostdiv_f32.o
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_scatter_premulsum_u32.o
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_scatter_premulsum_i64.o
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_scatter_premulsum_u64.o
[ 54%] Built target dmlc
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampling\randomwalks\randomwalk_with_restart_cpu.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampling\randomwalks\randomwalk_with_restart_cpu.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampling\randomwalks\randomwalk_with_restart_cpu.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampling\randomwalks\randomwalk_with_restart_cpu.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampling\randomwalks\randomwalks.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampling\randomwalks\randomwalks.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampling\randomwalks\randomwalks.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampling\randomwalks\randomwalks.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         socket_pool.cc
         tcp_socket.cc
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\network.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_gather_sumpostdiv_f64.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  all_gather.cu                       > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/all_gather_sumpostdiv_bf16.o
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_scatter_premulsum_f16.o
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_scatter_premulsum_f32.o
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\network.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\network.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\network.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\serialize\tensor_serialize.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\serialize\tensor_serialize.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\serialize\tensor_serialize.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\serialize\tensor_serialize.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\heterograph_capi.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\heterograph_capi.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\heterograph_capi.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\heterograph_capi.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampling\randomwalks\randomwalk_cpu.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampling\randomwalks\randomwalk_cpu.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampling\randomwalks\randomwalk_cpu.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampling\randomwalks\randomwalk_cpu.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\utility(207,16): warning C4244: 'initializing': conversion from '_Ty' to '_Ty1', possible loss of data [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _Ty=std::chrono::system_clock::rep
                 ]
                 and
                 [
                     _Ty1=int32_t
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampling\randomwalks\randomwalk_cpu.cc)
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xmemory(671): message : see reference to function template instantiation 'std::pair<IdxType,IdxType>::pair<__int64,int,0>(std::pair<__int64,int> &&) noexcept' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     IdxType=int32_t
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampling\randomwalks\randomwalk_cpu.cc)
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xmemory(671): message : see reference to function template instantiation 'std::pair<IdxType,IdxType>::pair<__int64,int,0>(std::pair<__int64,int> &&) noexcept' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     IdxType=int32_t
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampling\randomwalks\randomwalk_cpu.cc)
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\vector(687): message : see reference to function template instantiation 'void std::_Default_allocator_traits<_Alloc>::construct<_Ty,std::pair<__int64,int>>(_Alloc &,_Objty *const ,std::pair<__int64,int> &&)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _Alloc=std::allocator<std::pair<int32_t,int32_t>>,
                     _Ty=std::pair<int32_t,int32_t>,
                     _Objty=std::pair<int32_t,int32_t>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampling\randomwalks\randomwalk_cpu.cc)
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\vector(687): message : see reference to function template instantiation 'void std::_Default_allocator_traits<_Alloc>::construct<_Ty,std::pair<__int64,int>>(_Alloc &,_Objty *const ,std::pair<__int64,int> &&)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _Alloc=std::allocator<std::pair<int32_t,int32_t>>,
                     _Ty=std::pair<int32_t,int32_t>,
                     _Objty=std::pair<int32_t,int32_t>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampling\randomwalks\randomwalk_cpu.cc)
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\vector(705): message : see reference to function template instantiation 'void std::vector<std::pair<IdxType,IdxType>,std::allocator<std::pair<IdxType,IdxType>>>::_Emplace_back_with_unused_capacity<_Ty>(_Ty &&)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     IdxType=int32_t,
                     _Ty=std::pair<__int64,int>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampling\randomwalks\randomwalk_cpu.cc)
       C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampling\randomwalks\randomwalk_cpu.cc(61): message : see reference to function template instantiation 'void std::vector<std::pair<IdxType,IdxType>,std::allocator<std::pair<IdxType,IdxType>>>::emplace_back<std::pair<__int64,int>>(std::pair<__int64,int> &&)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     IdxType=int32_t
                 ]
       C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampling\randomwalks\randomwalk_cpu.cc(127): message : see reference to function template instantiation 'std::tuple<dgl::runtime::NDArray,dgl::IdArray,dgl::IdArray> dgl::sampling::impl::SelectPinSageNeighbors<kDGLCPU,int32_t>(const dgl::IdArray,const dgl::IdArray,const int64_t,const int64_t)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/broadcast_sum_i8.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_scatter_premulsum_f64.o
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_scatter_sumpostdiv_i8.o
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/broadcast_sum_u8.o
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/broadcast_sum_i32.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_scatter_sumpostdiv_u8.o
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_scatter_sumpostdiv_i32.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/broadcast_sum_u32.o
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/broadcast_sum_i64.o
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_scatter_sumpostdiv_u32.o
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_scatter_sumpostdiv_i64.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/broadcast_sum_u64.o
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_scatter_sumpostdiv_u64.o
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_scatter_sumpostdiv_f16.o
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\serialize\dglgraph_serialize.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\serialize\dglgraph_serialize.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\aten\./array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\serialize\dglgraph_serialize.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl/aten/coo.h(235): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\serialize\dglgraph_serialize.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampler.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampler.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampler.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampler.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,1): warning C4267: '=': conversion from 'size_t' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampler.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const unsigned __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const unsigned __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampler.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=unsigned __int64,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<unsigned __int64>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampler.cc)
       C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampler.cc(803): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<uint64_t>(const std::vector<uint64_t,std::allocator<uint64_t>> &,uint8_t,DGLContext)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\serialize\graph_serialize.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\serialize\graph_serialize.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\serialize\graph_serialize.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\serialize\graph_serialize.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
/usr/bin/ar: creating lib/libxsmmgen.a
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/broadcast_sum_f16.o
[ 49%] Linking CXX static library ../../../lib/libgmock_main.a
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/broadcast_sum_f32.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_scatter_sumpostdiv_f32.o
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/reduce_scatter_sumpostdiv_f64.o
    18>C:\Jenkins\workspace\dgl_PR-4648\third_party\METIS\include\metis.h(75,1): warning C4005: 'INT32_MIN': macro redefinition (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\metis_partition_hetero.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\stdint.h(51): message : see previous definition of 'INT32_MIN' (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\metis_partition_hetero.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
    18>C:\Jenkins\workspace\dgl_PR-4648\third_party\METIS\include\metis.h(76,1): warning C4005: 'INT32_MAX': macro redefinition (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\metis_partition_hetero.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\stdint.h(55): message : see previous definition of 'INT32_MAX' (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\metis_partition_hetero.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
    18>C:\Jenkins\workspace\dgl_PR-4648\third_party\METIS\include\metis.h(77,1): warning C4005: 'INT64_MIN': macro redefinition (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\metis_partition_hetero.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\stdint.h(52): message : see previous definition of 'INT64_MIN' (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\metis_partition_hetero.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
    18>C:\Jenkins\workspace\dgl_PR-4648\third_party\METIS\include\metis.h(78,1): warning C4005: 'INT64_MAX': macro redefinition (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\metis_partition_hetero.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\stdint.h(56): message : see previous definition of 'INT64_MAX' (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\metis_partition_hetero.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\serialize\heterograph_serialize.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\serialize\heterograph_serialize.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\serialize\heterograph_serialize.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\serialize\heterograph_serialize.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\shared_mem_manager.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\shared_mem_manager.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\shared_mem_manager.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\shared_mem_manager.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
[ 49%] Built target gmock_main
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/broadcast_sum_f64.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/broadcast_sum_bf16.o
Compiling  functions.cu                        > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/functions.o
Compiling  onerank_reduce.cu                   > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/onerank_reduce.o
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\subgraph.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\subgraph.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\subgraph.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\subgraph.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/broadcast_prod_i8.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/sendrecv_sum_i8.o
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/sendrecv_sum_u8.o
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/broadcast_prod_u8.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/broadcast_prod_i32.o
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/sendrecv_sum_i32.o
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/sendrecv_sum_u32.o
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\metis_partition_hetero.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\metis_partition_hetero.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\metis_partition_hetero.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\metis_partition_hetero.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\line_graph.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\line_graph.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\line_graph.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\line_graph.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/broadcast_prod_u32.o
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/broadcast_prod_i64.o
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/sendrecv_sum_i64.o
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/sendrecv_sum_u64.o
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\remove_edges.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\remove_edges.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\remove_edges.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\remove_edges.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\tuple(632,57): warning C4244: '=': conversion from '_Ty' to 'IdType', possible loss of data [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _Ty=dgl::dgl_id_t
                 ]
                 and
                 [
                     IdType=int32_t
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\compact.cc)
       C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\compact.cc(63): message : see reference to function template instantiation 'std::tuple<IdType &,IdType &> &std::tuple<IdType &,IdType &>::operator =<dgl::dgl_id_t,dgl::dgl_id_t,0>(std::pair<dgl::dgl_id_t,dgl::dgl_id_t> &&) noexcept' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     IdType=int32_t
                 ]
       C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\compact.cc(63): message : see reference to function template instantiation 'std::tuple<IdType &,IdType &> &std::tuple<IdType &,IdType &>::operator =<dgl::dgl_id_t,dgl::dgl_id_t,0>(std::pair<dgl::dgl_id_t,dgl::dgl_id_t> &&) noexcept' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     IdType=int32_t
                 ]
       C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\compact.cc(146): message : see reference to function template instantiation 'std::pair<std::vector<dgl::HeteroGraphPtr,std::allocator<dgl::HeteroGraphPtr>>,std::vector<dgl::IdArray,std::allocator<dgl::IdArray>>> dgl::transform::`anonymous-namespace'::CompactGraphsCPU<int32_t>(const std::vector<dgl::HeteroGraphPtr,std::allocator<dgl::HeteroGraphPtr>> &,const std::vector<dgl::IdArray,std::allocator<dgl::IdArray>> &)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\tuple(633,58): warning C4244: '=': conversion from '_Ty' to 'IdType', possible loss of data [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _Ty=dgl::dgl_id_t
                 ]
                 and
                 [
                     IdType=int32_t
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\compact.cc)
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\utility(151,1): warning C4244: 'initializing': conversion from '_Ty' to '_Ty2', possible loss of data [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _Ty=dgl::dgl_id_t
                 ]
                 and
                 [
                     _Ty2=int32_t
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\compact.cc)
       C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\../../array/cpu/array_utils.h(57): message : see reference to function template instantiation 'std::pair<K,V>::pair<const IdType&,size_t,0>(_Other1,_Other2 &&) noexcept' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     K=int32_t,
                     V=int32_t,
                     IdType=int32_t,
                     _Other1=const int32_t &,
                     _Other2=size_t
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\compact.cc)
       C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\../../array/cpu/array_utils.h(57): message : see reference to function template instantiation 'std::pair<K,V>::pair<const IdType&,size_t,0>(_Other1,_Other2 &&) noexcept' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     K=int32_t,
                     V=int32_t,
                     IdType=int32_t,
                     _Other1=const int32_t &,
                     _Other2=size_t
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\compact.cc)
       C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\../../array/cpu/array_utils.h(50): message : while compiling class template member function 'void dgl::aten::IdHashMap<IdType>::Update(dgl::IdArray)' [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     IdType=int32_t
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\compact.cc)
       C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\compact.cc(80): message : see reference to function template instantiation 'void dgl::aten::IdHashMap<IdType>::Update(dgl::IdArray)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     IdType=int32_t
                 ]
       C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\compact.cc(74): message : see reference to class template instantiation 'dgl::aten::IdHashMap<IdType>' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     IdType=int32_t
                 ]
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/broadcast_prod_u64.o
[ 49%] Linking CXX static library libtensorpipe.a
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/broadcast_prod_f16.o
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/sendrecv_sum_f16.o
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/sendrecv_sum_f32.o
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/sendrecv_sum_f64.o
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\compact.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\compact.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\compact.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\compact.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\to_simple.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\to_simple.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\to_simple.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\to_simple.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\partition_hetero.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\partition_hetero.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\partition_hetero.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\partition_hetero.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const unsigned __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\partition_hetero.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const unsigned __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const unsigned __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\partition_hetero.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=uint64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<uint64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\partition_hetero.cc)
       C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\partition_hetero.cc(106): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<uint64_t>(const std::vector<uint64_t,std::allocator<uint64_t>> &,uint8_t,DGLContext)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\scheduler\scheduler_apis.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\scheduler\scheduler_apis.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\scheduler\scheduler_apis.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\scheduler\scheduler_apis.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\unit_graph.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\unit_graph.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\unit_graph.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\unit_graph.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/broadcast_prod_f32.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/broadcast_prod_f64.o
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/sendrecv_prod_i8.o
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/sendrecv_prod_u8.o
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\utility(151,1): warning C4267: 'initializing': conversion from 'size_t' to '_Ty2', possible loss of data [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _Ty2=int32_t
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\to_bipartite.cc)
       C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\../../array/cpu/array_utils.h(57): message : see reference to function template instantiation 'std::pair<K,V>::pair<const IdType&,size_t,0>(_Other1,_Other2 &&) noexcept' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     K=int32_t,
                     V=int32_t,
                     IdType=int32_t,
                     _Other1=const int32_t &,
                     _Other2=size_t
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\to_bipartite.cc)
       C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\../../array/cpu/array_utils.h(57): message : see reference to function template instantiation 'std::pair<K,V>::pair<const IdType&,size_t,0>(_Other1,_Other2 &&) noexcept' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     K=int32_t,
                     V=int32_t,
                     IdType=int32_t,
                     _Other1=const int32_t &,
                     _Other2=size_t
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\to_bipartite.cc)
       C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\../../array/cpu/array_utils.h(50): message : while compiling class template member function 'void dgl::aten::IdHashMap<IdType>::Update(dgl::IdArray)' [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     IdType=int32_t
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\to_bipartite.cc)
       C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\to_bipartite.cc(80): message : see reference to function template instantiation 'void dgl::aten::IdHashMap<IdType>::Update(dgl::IdArray)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     IdType=int32_t
                 ]
       C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\to_bipartite.cc(80): message : see reference to class template instantiation 'dgl::aten::IdHashMap<IdType>' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     IdType=int32_t
                 ]
       C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\to_bipartite.cc(150): message : see reference to function template instantiation 'std::tuple<dgl::HeteroGraphPtr,std::vector<dgl::IdArray,std::allocator<dgl::IdArray>>> dgl::transform::`anonymous-namespace'::ToBlockCPU<int32_t>(dgl::HeteroGraphPtr,const std::vector<dgl::IdArray,std::allocator<dgl::IdArray>> &,bool,std::vector<dgl::IdArray,std::allocator<dgl::IdArray>> *const )' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\to_bipartite.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\to_bipartite.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\to_bipartite.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\to_bipartite.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
[ 49%] Built target tensorpipe
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/broadcast_prod_bf16.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/sendrecv_prod_i32.o
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/sendrecv_prod_u32.o
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/sendrecv_prod_i64.o
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\rpc\network\socket_communicator.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\rpc\network\socket_communicator.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\rpc\network\socket_communicator.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\rpc\network\socket_communicator.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/broadcast_min_i8.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/broadcast_min_u8.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/sendrecv_prod_u64.o
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/sendrecv_prod_f16.o
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/sendrecv_prod_f32.o
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/broadcast_min_i32.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/broadcast_min_u32.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/sendrecv_prod_f64.o
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/sendrecv_min_i8.o
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/sendrecv_min_u8.o
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/broadcast_min_i64.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/broadcast_min_u64.o
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/sendrecv_min_i32.o
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/sendrecv_min_u32.o
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/sendrecv_min_i64.o
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/broadcast_min_f16.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/sendrecv_min_u64.o
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/sendrecv_min_f16.o
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/sendrecv_min_f32.o
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/broadcast_min_f32.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/broadcast_min_f64.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/sendrecv_min_f64.o
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/sendrecv_max_i8.o
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/broadcast_min_bf16.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/broadcast_max_i8.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/sendrecv_max_u8.o
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/sendrecv_max_i32.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/broadcast_max_u8.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/broadcast_max_i32.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/sendrecv_max_u32.o
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/sendrecv_max_i64.o
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\bin\HostX64\x64\CL.exe /c /I"C:\Jenkins\workspace\dgl_PR-4648\include" /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\dlpack\include" /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\dmlc-core\include" /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\phmap" /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\xbyak" /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\METIS\include" /I"C:\Jenkins\workspace\dgl_PR-4648\tensoradapter\include" /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\nanoflann\include" /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\libxsmm\include" /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\googletest\googletest\include" /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\googletest\googletest" /I"C:\Jenkins\workspace\dgl_PR-4648\build\third_party\dmlc-core\include" /nologo /W1 /WX- /diagnostics:column /MP /O2 /Ob2 /D DGL_EXPORTS /D USE_AVX /D USE_LIBXSMM /D DGL_CPU_LLC_SIZE=40000000 /D IDXTYPEWIDTH=64 /D REALTYPEWIDTH=32 /D NDEBUG /D WIN32_LEAN_AND_MEAN /D _CRT_SECURE_NO_WARNINGS /D _SCL_SECURE_NO_WARNINGS /D NOMINMAX /D ENABLE_PARTIAL_FRONTIER=0 /D __USE_XOPEN2K8 /D DMLC_CORE_USE_CMAKE /D DMLC_USE_CXX11=1 /D "CMAKE_INTDIR=\"Release\"" /D dgl_EXPORTS /D _WINDLL /D _MBCS /Gm- /EHsc /MD /GS /fp:precise /Zc:wchar_t /Zc:forScope /Zc:inline /openmp /std:c++14 /Fo"dgl.dir\Release\/src/array/cpu/traversal.cc.obj" /Fd"dgl.dir\Release\vc142.pdb" /Gd /TP /errorReport:queue "C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\traversal.cc"
         traversal.cc
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/broadcast_max_u32.o
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/sendrecv_max_u64.o
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/sendrecv_max_f16.o
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/sendrecv_max_f32.o
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/broadcast_max_i64.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/broadcast_max_u64.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/sendrecv_max_f64.o
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/sendrecv_premulsum_i8.o
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/sendrecv_premulsum_u8.o
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/broadcast_max_f16.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/broadcast_max_f32.o
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ]
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ]
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/sendrecv_premulsum_i32.o
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/sendrecv_premulsum_u32.o
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/broadcast_max_f64.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/broadcast_max_bf16.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/sendrecv_premulsum_i64.o
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/sendrecv_premulsum_u64.o
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/broadcast_premulsum_i8.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/broadcast_premulsum_u8.o
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\bin\HostX64\x64\CL.exe /c /I"C:\Jenkins\workspace\dgl_PR-4648\include" /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\dlpack\include" /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\dmlc-core\include" /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\phmap" /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\xbyak" /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\METIS\include" /I"C:\Jenkins\workspace\dgl_PR-4648\tensoradapter\include" /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\nanoflann\include" /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\libxsmm\include" /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\googletest\googletest\include" /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\googletest\googletest" /I"C:\Jenkins\workspace\dgl_PR-4648\build\third_party\dmlc-core\include" /nologo /W1 /WX- /diagnostics:column /MP /O2 /Ob2 /D DGL_EXPORTS /D USE_AVX /D USE_LIBXSMM /D DGL_CPU_LLC_SIZE=40000000 /D IDXTYPEWIDTH=64 /D REALTYPEWIDTH=32 /D NDEBUG /D WIN32_LEAN_AND_MEAN /D _CRT_SECURE_NO_WARNINGS /D _SCL_SECURE_NO_WARNINGS /D NOMINMAX /D ENABLE_PARTIAL_FRONTIER=0 /D __USE_XOPEN2K8 /D DMLC_CORE_USE_CMAKE /D DMLC_USE_CXX11=1 /D "CMAKE_INTDIR=\"Release\"" /D dgl_EXPORTS /D _WINDLL /D _MBCS /Gm- /EHsc /MD /GS /fp:precise /Zc:wchar_t /Zc:forScope /Zc:inline /openmp /std:c++14 /Fo"dgl.dir\Release\/src/array/union_partition.cc.obj" /Fd"dgl.dir\Release\vc142.pdb" /Gd /TP /errorReport:queue "C:\Jenkins\workspace\dgl_PR-4648\src\array\union_partition.cc"
         union_partition.cc
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/sendrecv_premulsum_f16.o
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/sendrecv_premulsum_f32.o
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/sendrecv_premulsum_f64.o
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/broadcast_premulsum_i32.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/broadcast_premulsum_u32.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/sendrecv_sumpostdiv_i8.o
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/sendrecv_sumpostdiv_u8.o
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/sendrecv_sumpostdiv_i32.o
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/broadcast_premulsum_i64.o
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/broadcast_premulsum_u64.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/sendrecv_sumpostdiv_u32.o
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/sendrecv_sumpostdiv_i64.o
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/sendrecv_sumpostdiv_u64.o
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/broadcast_premulsum_f16.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/sendrecv_sumpostdiv_f16.o
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/sendrecv_sumpostdiv_f32.o
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ]
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ]
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/broadcast_premulsum_f32.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/broadcast_premulsum_f64.o
Compiling  sendrecv.cu                         > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/sendrecv_sumpostdiv_f64.o
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\bin\HostX64\x64\CL.exe /c /I"C:\Jenkins\workspace\dgl_PR-4648\include" /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\dlpack\include" /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\dmlc-core\include" /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\phmap" /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\xbyak" /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\METIS\include" /I"C:\Jenkins\workspace\dgl_PR-4648\tensoradapter\include" /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\nanoflann\include" /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\libxsmm\include" /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\googletest\googletest\include" /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\googletest\googletest" /I"C:\Jenkins\workspace\dgl_PR-4648\build\third_party\dmlc-core\include" /nologo /W1 /WX- /diagnostics:column /MP /O2 /Ob2 /D DGL_EXPORTS /D USE_AVX /D USE_LIBXSMM /D DGL_CPU_LLC_SIZE=40000000 /D IDXTYPEWIDTH=64 /D REALTYPEWIDTH=32 /D NDEBUG /D WIN32_LEAN_AND_MEAN /D _CRT_SECURE_NO_WARNINGS /D _SCL_SECURE_NO_WARNINGS /D NOMINMAX /D ENABLE_PARTIAL_FRONTIER=0 /D __USE_XOPEN2K8 /D DMLC_CORE_USE_CMAKE /D DMLC_USE_CXX11=1 /D "CMAKE_INTDIR=\"Release\"" /D dgl_EXPORTS /D _WINDLL /D _MBCS /Gm- /EHsc /MD /GS /fp:precise /Zc:wchar_t /Zc:forScope /Zc:inline /openmp /std:c++14 /Fo"dgl.dir\Release\/src/graph/transform/cpu/knn.cc.obj" /Fd"dgl.dir\Release\vc142.pdb" /Gd /TP /errorReport:queue "C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\cpu\knn.cc"
         knn.cc
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/broadcast_premulsum_bf16.o
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/broadcast_sumpostdiv_i8.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/broadcast_sumpostdiv_u8.o
/usr/bin/ar: creating lib/libxsmmgen.a
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
/usr/bin/ar: creating lib/libxsmm.a
/usr/bin/ar: creating lib/libxsmmext.a
================================================================================
LIBXSMM master-1.16.1-1534 (Linux@856b09e8bd12)
--------------------------------------------------------------------------------
GNU Compiler Collection: gcc 7.5.0, and g++ 7.5.0
C / C++ target: -msse4.2
Fortran Compiler is disabled or missing: no Fortran interface is built!
--------------------------------------------------------------------------------
BLAS dependency and fallback is removed!
--------------------------------------------------------------------------------
[ 54%] Built target libxsmm
Scanning dependencies of target dgl
[ 54%] Building CXX object CMakeFiles/dgl.dir/src/array/array.cc.o
[ 54%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/coo_remove.cc.o
[ 54%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/array_scatter.cc.o
[ 54%] Building CXX object CMakeFiles/dgl.dir/src/array/array_arith.cc.o
[ 54%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/array_cumsum.cc.o
[ 54%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/csr_remove.cc.o
[ 54%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/array_op_impl.cc.o
[ 54%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/array_pack.cc.o
[ 54%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/coo_sort.cc.o
[ 54%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/array_sort.cc.o
[ 56%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/csr_sum.cc.o
[ 56%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/csr_mm.cc.o
[ 56%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/csr_union.cc.o
[ 58%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/array_index_select.cc.o
[ 58%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/csr_get_data.cc.o
[ 59%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/array_repeat.cc.o
[ 60%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/gather_mm.cc.o
[ 60%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/rowwise_topk.cc.o
[ 60%] Building CXX object CMakeFiles/dgl.dir/src/geometry/cpu/geometry_op_impl.cc.o
[ 60%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/spmat_op_impl_coo.cc.o
[ 60%] Building CXX object CMakeFiles/dgl.dir/src/random/cpu/choice.cc.o
[ 62%] Building CXX object CMakeFiles/dgl.dir/src/runtime/dso_module.cc.o
[ 62%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/sddmm.cc.o
[ 62%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/spmat_op_impl_csr.cc.o
[ 62%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/negative_sampling.cc.o
[ 62%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/rowwise_sampling.cc.o
[ 62%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/segment_reduce.cc.o
[ 62%] Building CXX object CMakeFiles/dgl.dir/src/array/uvm_array.cc.o
[ 62%] Building CXX object CMakeFiles/dgl.dir/src/c_api_common.cc.o
[ 62%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/traversal.cc.o
[ 62%] Building CXX object CMakeFiles/dgl.dir/src/geometry/geometry.cc.o
[ 62%] Building CXX object CMakeFiles/dgl.dir/src/array/union_partition.cc.o
[ 62%] Building CXX object CMakeFiles/dgl.dir/src/runtime/module_util.cc.o
[ 62%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/csr_sort.cc.o
[ 64%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/spmm.cc.o
[ 64%] Building CXX object CMakeFiles/dgl.dir/src/runtime/c_object_api.cc.o
[ 65%] Building CXX object CMakeFiles/dgl.dir/src/runtime/c_runtime_api.cc.o
[ 65%] Building CXX object CMakeFiles/dgl.dir/src/random/random.cc.o
[ 66%] Building CXX object CMakeFiles/dgl.dir/src/runtime/config.cc.o
[ 66%] Building CXX object CMakeFiles/dgl.dir/src/array/libra_partition.cc.o
[ 66%] Building CXX object CMakeFiles/dgl.dir/src/array/filter.cc.o
[ 66%] Building CXX object CMakeFiles/dgl.dir/src/array/kernel.cc.o
[ 67%] Building CXX object CMakeFiles/dgl.dir/src/runtime/registry.cc.o
[ 67%] Building CXX object CMakeFiles/dgl.dir/src/bcast.cc.o
[ 67%] Building CXX object CMakeFiles/dgl.dir/src/runtime/semaphore_wrapper.cc.o
[ 69%] Building CXX object CMakeFiles/dgl.dir/src/runtime/ndarray.cc.o
[ 69%] Building CXX object CMakeFiles/dgl.dir/src/runtime/shared_mem.cc.o
[ 69%] Building CXX object CMakeFiles/dgl.dir/src/runtime/object.cc.o
[ 69%] Building CXX object CMakeFiles/dgl.dir/src/runtime/cpu_device_api.cc.o
[ 69%] Building CXX object CMakeFiles/dgl.dir/src/runtime/file_util.cc.o
[ 69%] Building CXX object CMakeFiles/dgl.dir/src/runtime/dlpack_convert.cc.o
[ 70%] Building CXX object CMakeFiles/dgl.dir/src/partition/ndarray_partition.cc.o
[ 70%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/array_nonzero.cc.o
[ 70%] Building CXX object CMakeFiles/dgl.dir/src/runtime/system_lib_module.cc.o
[ 70%] Building CXX object CMakeFiles/dgl.dir/src/runtime/threading_backend.cc.o
[ 70%] Building CXX object CMakeFiles/dgl.dir/src/runtime/thread_pool.cc.o
[ 70%] Building CXX object CMakeFiles/dgl.dir/src/api/api_container.cc.o
[ 71%] Building CXX object CMakeFiles/dgl.dir/src/runtime/module.cc.o
[ 71%] Building CXX object CMakeFiles/dgl.dir/src/runtime/workspace_pool.cc.o
[ 71%] Building CXX object CMakeFiles/dgl.dir/src/api/api_test.cc.o
[ 71%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/coo_coalesce.cc.o
[ 72%] Building CXX object CMakeFiles/dgl.dir/src/runtime/tensordispatch.cc.o
[ 74%] Building CXX object CMakeFiles/dgl.dir/src/runtime/resource_manager.cc.o
[ 75%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/coo_linegraph.cc.o
[ 75%] Building CXX object CMakeFiles/dgl.dir/src/runtime/utils.cc.o
[ 76%] Building CXX object CMakeFiles/dgl.dir/src/graph/pickle.cc.o
[ 77%] Building CXX object CMakeFiles/dgl.dir/src/graph/metis_partition.cc.o
[ 77%] Building CXX object CMakeFiles/dgl.dir/src/graph/immutable_graph.cc.o
[ 77%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/csr_to_simple.cc.o
[ 77%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/disjoint_union.cc.o
[ 77%] Building CXX object CMakeFiles/dgl.dir/src/graph/network.cc.o
[ 77%] Building CXX object CMakeFiles/dgl.dir/src/graph/heterograph.cc.o
[ 77%] Building CXX object CMakeFiles/dgl.dir/src/graph/graph_apis.cc.o
[ 77%] Building CXX object CMakeFiles/dgl.dir/src/graph/graph.cc.o
[ 77%] Building CXX object CMakeFiles/dgl.dir/src/graph/graph_op.cc.o
[ 79%] Building CXX object CMakeFiles/dgl.dir/src/graph/serialize/heterograph_serialize.cc.o
[ 80%] Building CXX object CMakeFiles/dgl.dir/src/graph/creators.cc.o
[ 80%] Building CXX object CMakeFiles/dgl.dir/src/graph/graph_traversal.cc.o
[ 80%] Building CXX object CMakeFiles/dgl.dir/src/graph/sampling/negative/global_uniform.cc.o
[ 80%] Building CXX object CMakeFiles/dgl.dir/src/graph/nodeflow.cc.o
[ 81%] Building CXX object CMakeFiles/dgl.dir/src/graph/sampling/randomwalks/get_node_types_cpu.cc.o
[ 81%] Building CXX object CMakeFiles/dgl.dir/src/graph/sampler.cc.o
[ 81%] Building CXX object CMakeFiles/dgl.dir/src/graph/sampling/randomwalks/node2vec.cc.o
[ 81%] Building CXX object CMakeFiles/dgl.dir/src/graph/heterograph_capi.cc.o
[ 82%] Building CXX object CMakeFiles/dgl.dir/src/graph/sampling/randomwalks/randomwalk_with_restart_cpu.cc.o
[ 82%] Building CXX object CMakeFiles/dgl.dir/src/graph/sampling/randomwalks/randomwalks.cc.o
[ 82%] Building CXX object CMakeFiles/dgl.dir/src/graph/sampling/randomwalks/randomwalk_cpu.cc.o
[ 82%] Building CXX object CMakeFiles/dgl.dir/src/graph/sampling/randomwalks/node2vec_cpu.cc.o
[ 82%] Building CXX object CMakeFiles/dgl.dir/src/graph/sampling/neighbor/neighbor.cc.o
[ 82%] Building CXX object CMakeFiles/dgl.dir/src/graph/serialize/dglgraph_serialize.cc.o
[ 83%] Building CXX object CMakeFiles/dgl.dir/src/graph/gk_ops.cc.o
[ 83%] Building CXX object CMakeFiles/dgl.dir/src/graph/serialize/zerocopy_serializer.cc.o
[ 83%] Building CXX object CMakeFiles/dgl.dir/src/graph/serialize/tensor_serialize.cc.o
[ 83%] Building CXX object CMakeFiles/dgl.dir/src/graph/transform/compact.cc.o
[ 83%] Building CXX object CMakeFiles/dgl.dir/src/graph/serialize/graph_serialize.cc.o
[ 83%] Building CXX object CMakeFiles/dgl.dir/src/graph/subgraph.cc.o
[ 85%] Building CXX object CMakeFiles/dgl.dir/src/graph/transform/knn.cc.o
[ 86%] Building CXX object CMakeFiles/dgl.dir/src/graph/shared_mem_manager.cc.o
[ 86%] Building CXX object CMakeFiles/dgl.dir/src/graph/transform/cpu/knn.cc.o
[ 86%] Building CXX object CMakeFiles/dgl.dir/src/graph/transform/metis_partition_hetero.cc.o
[ 87%] Building CXX object CMakeFiles/dgl.dir/src/graph/transform/remove_edges.cc.o
[ 87%] Building CXX object CMakeFiles/dgl.dir/src/graph/transform/partition_hetero.cc.o
[ 87%] Building CXX object CMakeFiles/dgl.dir/src/rpc/network/tcp_socket.cc.o
[ 87%] Building CXX object CMakeFiles/dgl.dir/src/graph/transform/union_partition.cc.o
[ 88%] Building CXX object CMakeFiles/dgl.dir/src/rpc/network/common.cc.o
[ 88%] Building CXX object CMakeFiles/dgl.dir/src/scheduler/scheduler_apis.cc.o
[ 88%] Building CXX object CMakeFiles/dgl.dir/src/scheduler/scheduler.cc.o
[ 88%] Building CXX object CMakeFiles/dgl.dir/src/graph/unit_graph.cc.o
[ 88%] Building CXX object CMakeFiles/dgl.dir/src/graph/transform/to_bipartite.cc.o
[ 90%] Building CXX object CMakeFiles/dgl.dir/src/rpc/network/socket_pool.cc.o
[ 90%] Building CXX object CMakeFiles/dgl.dir/src/rpc/network/msg_queue.cc.o
[ 91%] Building CXX object CMakeFiles/dgl.dir/src/graph/traversal.cc.o
[ 91%] Building CXX object CMakeFiles/dgl.dir/src/rpc/network/socket_communicator.cc.o
[ 91%] Building CXX object CMakeFiles/dgl.dir/src/graph/transform/line_graph.cc.o
[ 91%] Building CXX object CMakeFiles/dgl.dir/src/rpc/rpc.cc.o
[ 91%] Building CXX object CMakeFiles/dgl.dir/src/graph/transform/to_simple.cc.o
[ 91%] Building CXX object CMakeFiles/dgl.dir/src/rpc/tensorpipe/tp_communicator.cc.o
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/broadcast_sumpostdiv_i32.o
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/broadcast_sumpostdiv_u32.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/broadcast_sumpostdiv_i64.o
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/broadcast_sumpostdiv_u64.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/broadcast_sumpostdiv_f16.o
/usr/bin/ar: creating lib/libxsmmgen.a
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/broadcast_sumpostdiv_f32.o
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ]
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ]
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
[ 48%] Linking CXX static library ../../../lib/libgmock.a
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/broadcast_sumpostdiv_f64.o
Compiling  broadcast.cu                        > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/broadcast_sumpostdiv_bf16.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
[ 48%] Built target gmock
Scanning dependencies of target gmock_main
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_sum_i8.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_sum_u8.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\bin\HostX64\x64\CL.exe /c /I"C:\Jenkins\workspace\dgl_PR-4648\include" /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\dlpack\include" /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\dmlc-core\include" /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\phmap" /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\xbyak" /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\METIS\include" /I"C:\Jenkins\workspace\dgl_PR-4648\tensoradapter\include" /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\nanoflann\include" /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\libxsmm\include" /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\googletest\googletest\include" /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\googletest\googletest" /I"C:\Jenkins\workspace\dgl_PR-4648\build\third_party\dmlc-core\include" /nologo /W1 /WX- /diagnostics:column /MP /O2 /Ob2 /D DGL_EXPORTS /D USE_AVX /D USE_LIBXSMM /D DGL_CPU_LLC_SIZE=40000000 /D IDXTYPEWIDTH=64 /D REALTYPEWIDTH=32 /D NDEBUG /D WIN32_LEAN_AND_MEAN /D _CRT_SECURE_NO_WARNINGS /D _SCL_SECURE_NO_WARNINGS /D NOMINMAX /D ENABLE_PARTIAL_FRONTIER=0 /D __USE_XOPEN2K8 /D DMLC_CORE_USE_CMAKE /D DMLC_USE_CXX11=1 /D "CMAKE_INTDIR=\"Release\"" /D dgl_EXPORTS /D _WINDLL /D _MBCS /Gm- /EHsc /MD /GS /fp:precise /Zc:wchar_t /Zc:forScope /Zc:inline /openmp /std:c++14 /Fo"dgl.dir\Release\/src/graph/transform/knn.cc.obj" /Fd"dgl.dir\Release\vc142.pdb" /Gd /TP /errorReport:queue "C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\knn.cc"
         knn.cc
[ 48%] Building CXX object third_party/googletest/googlemock/CMakeFiles/gmock_main.dir/src/gmock_main.cc.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_sum_i32.o
[ 49%] Linking CXX static library libdmlc.a
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_sum_u32.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_sum_i64.o
[ 48%] Linking CXX static library libdmlc.a
[ 49%] Built target dmlc
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_sum_u64.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
[ 48%] Built target dmlc
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ]
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ]
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_sum_f16.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_sum_f32.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\bin\HostX64\x64\CL.exe /c /I"C:\Jenkins\workspace\dgl_PR-4648\include" /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\dlpack\include" /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\dmlc-core\include" /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\phmap" /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\xbyak" /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\METIS\include" /I"C:\Jenkins\workspace\dgl_PR-4648\tensoradapter\include" /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\nanoflann\include" /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\libxsmm\include" /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\googletest\googletest\include" /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\googletest\googletest" /I"C:\Jenkins\workspace\dgl_PR-4648\build\third_party\dmlc-core\include" /nologo /W1 /WX- /diagnostics:column /MP /O2 /Ob2 /D DGL_EXPORTS /D USE_AVX /D USE_LIBXSMM /D DGL_CPU_LLC_SIZE=40000000 /D IDXTYPEWIDTH=64 /D REALTYPEWIDTH=32 /D NDEBUG /D WIN32_LEAN_AND_MEAN /D _CRT_SECURE_NO_WARNINGS /D _SCL_SECURE_NO_WARNINGS /D NOMINMAX /D ENABLE_PARTIAL_FRONTIER=0 /D __USE_XOPEN2K8 /D DMLC_CORE_USE_CMAKE /D DMLC_USE_CXX11=1 /D "CMAKE_INTDIR=\"Release\"" /D dgl_EXPORTS /D _WINDLL /D _MBCS /Gm- /EHsc /MD /GS /fp:precise /Zc:wchar_t /Zc:forScope /Zc:inline /openmp /std:c++14 /Fo"dgl.dir\Release\/src/graph/transform/union_partition.cc.obj" /Fd"dgl.dir\Release\vc142.pdb" /Gd /TP /errorReport:queue "C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\union_partition.cc"
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_sum_f64.o
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_sum_bf16.o
         union_partition.cc
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_prod_i8.o
/root/jenkins/workspace/dgl_PR-4648/src/partition/ndarray_partition.cc: In member function 'virtual int64_t dgl::partition::RangePartition::PartSize(int) const':
/root/jenkins/workspace/dgl_PR-4648/src/partition/ndarray_partition.cc:202:3: warning: control reaches end of non-void function [-Wreturn-type]
   }
   ^
In file included from /root/jenkins/workspace/dgl_PR-4648/src/geometry/../c_api_common.h:10:0,
                 from /root/jenkins/workspace/dgl_PR-4648/src/geometry/geometry.cc:9:
/root/jenkins/workspace/dgl_PR-4648/include/dgl/runtime/packed_func.h:502:21: warning: inline function 'TObjectRef dgl::runtime::DGLArgValue::AsObjectRef() const [with TObjectRef = dgl::HeteroGraphRef]' used but never defined
   inline TObjectRef AsObjectRef() const;
                     ^~~~~~~~~~~
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_prod_u8.o
In file included from /root/jenkins/workspace/dgl_PR-4648/third_party/dmlc-core/include/dmlc/logging.h:132:0,
                 from /root/jenkins/workspace/dgl_PR-4648/include/dgl/./runtime/object.h:9,
                 from /root/jenkins/workspace/dgl_PR-4648/include/dgl/graph_interface.h:15,
                 from /root/jenkins/workspace/dgl_PR-4648/include/dgl/sampler.h:13,
                 from /root/jenkins/workspace/dgl_PR-4648/src/graph/sampler.cc:6:
/root/jenkins/workspace/dgl_PR-4648/src/graph/sampler.cc: In member function 'dgl::NegSubgraph dgl::{anonymous}::EdgeSamplerObject::genNegEdgeSubgraph(const dgl::Subgraph&, const string&, int64_t, bool, bool)':
/root/jenkins/workspace/dgl_PR-4648/src/graph/sampler.cc:1189:48: warning: comparison between signed and unsigned integer expressions [-Wsign-compare]
       assert(prev_neg_offset + neg_sample_size == neg_vids.size());
              ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^~~~~~~~~~~~~~~~~~
/root/jenkins/workspace/dgl_PR-4648/src/graph/sampler.cc:1193:48: warning: comparison between signed and unsigned integer expressions [-Wsign-compare]
       assert(prev_neg_offset + neg_sample_size == neg_vids.size());
              ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^~~~~~~~~~~~~~~~~~
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_prod_i32.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ]
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ]
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_prod_u32.o
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_prod_i64.o
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_prod_u64.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\bin\HostX64\x64\CL.exe /c /I"C:\Jenkins\workspace\dgl_PR-4648\include" /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\dlpack\include" /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\dmlc-core\include" /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\phmap" /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\xbyak" /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\METIS\include" /I"C:\Jenkins\workspace\dgl_PR-4648\tensoradapter\include" /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\nanoflann\include" /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\libxsmm\include" /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\googletest\googletest\include" /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\googletest\googletest" /I"C:\Jenkins\workspace\dgl_PR-4648\build\third_party\dmlc-core\include" /nologo /W1 /WX- /diagnostics:column /MP /O2 /Ob2 /D DGL_EXPORTS /D USE_AVX /D USE_LIBXSMM /D DGL_CPU_LLC_SIZE=40000000 /D IDXTYPEWIDTH=64 /D REALTYPEWIDTH=32 /D NDEBUG /D WIN32_LEAN_AND_MEAN /D _CRT_SECURE_NO_WARNINGS /D _SCL_SECURE_NO_WARNINGS /D NOMINMAX /D ENABLE_PARTIAL_FRONTIER=0 /D __USE_XOPEN2K8 /D DMLC_CORE_USE_CMAKE /D DMLC_USE_CXX11=1 /D "CMAKE_INTDIR=\"Release\"" /D dgl_EXPORTS /D _WINDLL /D _MBCS /Gm- /EHsc /MD /GS /fp:precise /Zc:wchar_t /Zc:forScope /Zc:inline /openmp /std:c++14 /Fo"dgl.dir\Release\/src/graph/traversal.cc.obj" /Fd"dgl.dir\Release\vc142.pdb" /Gd /TP /errorReport:queue "C:\Jenkins\workspace\dgl_PR-4648\src\graph\traversal.cc"
         traversal.cc
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_prod_f16.o
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_prod_f32.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
/root/jenkins/workspace/dgl_PR-4648/src/array/libra_partition.cc: In function 'dgl::runtime::List<dgl::runtime::Value> dgl::aten::Libra2dglBuildDict(dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray, int32_t, int32_t, int64_t, const string&)':
/root/jenkins/workspace/dgl_PR-4648/src/array/libra_partition.cc:396:11: warning: ignoring return value of 'int fscanf(FILE*, const char*, ...)', declared with attribute warn_unused_result [-Wunused-result]
     fscanf(fp, "%ld,%ld,%f\n", &u, &v, &w);   // reading an edge - the src and dst global node IDs
     ~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_prod_f64.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
/root/jenkins/workspace/dgl_PR-4648/src/array/libra_partition.cc: In function 'int32_t dgl::aten::Ver2partition(IdType, int64_t*, int32_t) [with IdType = long int]':
/root/jenkins/workspace/dgl_PR-4648/src/array/libra_partition.cc:43:1: warning: control reaches end of non-void function [-Wreturn-type]
 }
 ^
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_prod_bf16.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_min_i8.o
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_min_u8.o
    18>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ]
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ]
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_min_i32.o
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_min_u32.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
    18>PreLinkEvent:
         Auto build dll exports
         setlocal
         cd C:\Jenkins\workspace\dgl_PR-4648\build
         if %errorlevel% neq 0 goto :cmEnd
         C:
         if %errorlevel% neq 0 goto :cmEnd
         "C:\Program Files\CMake\bin\cmake.exe" -E __create_def C:/Jenkins/workspace/dgl_PR-4648/build/dgl.dir/Release/exports.def C:/Jenkins/workspace/dgl_PR-4648/build/dgl.dir/Release//objects.txt
         if %errorlevel% neq 0 goto :cmEnd
         :cmEnd
         endlocal & call :cmErrorLevel %errorlevel% & goto :cmDone
         :cmErrorLevel
         exit /b %1
         :cmDone
         if %errorlevel% neq 0 goto :VCEnd
         :VCEnd
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_min_i64.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_min_u64.o
       Link:
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\bin\HostX64\x64\link.exe /ERRORREPORT:QUEUE /OUT:"C:\Jenkins\workspace\dgl_PR-4648\build\Release\dgl.dll" /INCREMENTAL:NO /NOLOGO "third_party\dmlc-core\Release\dmlc.lib" kernel32.lib user32.lib gdi32.lib winspool.lib shell32.lib ole32.lib oleaut32.lib uuid.lib comdlg32.lib advapi32.lib /DEF:"C:/Jenkins/workspace/dgl_PR-4648/build/dgl.dir/Release/exports.def" /MANIFEST /MANIFESTUAC:"level='asInvoker' uiAccess='false'" /manifest:embed /PDB:"C:/Jenkins/workspace/dgl_PR-4648/build/Release/dgl.pdb" /SUBSYSTEM:CONSOLE /TLBID:1 /DYNAMICBASE /NXCOMPAT /IMPLIB:"C:/Jenkins/workspace/dgl_PR-4648/build/Release/dgl.lib" /MACHINE:X64  /machine:x64 /DLL dgl.dir\Release\array.obj
         dgl.dir\Release\array_arith.obj
         dgl.dir\Release\array_cumsum.obj
         dgl.dir\Release\array_index_select.obj
         dgl.dir\Release\array_nonzero.obj
         dgl.dir\Release\array_op_impl.obj
         dgl.dir\Release\array_pack.obj
         dgl.dir\Release\array_repeat.obj
         dgl.dir\Release\array_scatter.obj
         dgl.dir\Release\array_sort.obj
         dgl.dir\Release\coo_coalesce.obj
         dgl.dir\Release\coo_linegraph.obj
         dgl.dir\Release\coo_remove.obj
         dgl.dir\Release\coo_sort.obj
         dgl.dir\Release\csr_get_data.obj
         dgl.dir\Release\csr_mm.obj
         dgl.dir\Release\csr_remove.obj
         dgl.dir\Release\csr_sort.obj
         dgl.dir\Release\csr_sum.obj
         dgl.dir\Release\csr_to_simple.obj
         dgl.dir\Release\csr_union.obj
         dgl.dir\Release\disjoint_union.obj
         dgl.dir\Release\gather_mm.obj
         dgl.dir\Release\negative_sampling.obj
         dgl.dir\Release\rowwise_sampling.obj
         dgl.dir\Release\rowwise_topk.obj
         dgl.dir\Release\sddmm.obj
         dgl.dir\Release\segment_reduce.obj
         dgl.dir\Release\spmat_op_impl_coo.obj
         dgl.dir\Release\spmat_op_impl_csr.obj
         dgl.dir\Release\spmm.obj
         dgl.dir\Release\filter.obj
         dgl.dir\Release\kernel.obj
         dgl.dir\Release\libra_partition.obj
         dgl.dir\Release\uvm_array.obj
         dgl.dir\Release\bcast.obj
         dgl.dir\Release\c_api_common.obj
         dgl.dir\Release\geometry_op_impl.obj
         dgl.dir\Release\geometry.obj
         dgl.dir\Release\ndarray_partition.obj
         dgl.dir\Release\choice.obj
         dgl.dir\Release\random.obj
         dgl.dir\Release\c_object_api.obj
         dgl.dir\Release\c_runtime_api.obj
         dgl.dir\Release\config.obj
         dgl.dir\Release\cpu_device_api.obj
         dgl.dir\Release\dlpack_convert.obj
         dgl.dir\Release\dso_module.obj
         dgl.dir\Release\file_util.obj
         dgl.dir\Release\module.obj
         dgl.dir\Release\module_util.obj
         dgl.dir\Release\ndarray.obj
         dgl.dir\Release\object.obj
         dgl.dir\Release\registry.obj
         dgl.dir\Release\resource_manager.obj
         dgl.dir\Release\semaphore_wrapper.obj
         dgl.dir\Release\shared_mem.obj
         dgl.dir\Release\system_lib_module.obj
         dgl.dir\Release\tensordispatch.obj
         dgl.dir\Release\thread_pool.obj
         dgl.dir\Release\threading_backend.obj
         dgl.dir\Release\utils.obj
         dgl.dir\Release\workspace_pool.obj
         dgl.dir\Release\api_container.obj
         dgl.dir\Release\api_test.obj
         dgl.dir\Release\creators.obj
         dgl.dir\Release\gk_ops.obj
         dgl.dir\Release\graph.obj
         dgl.dir\Release\graph_apis.obj
         dgl.dir\Release\graph_op.obj
         dgl.dir\Release\graph_traversal.obj
         dgl.dir\Release\heterograph.obj
         dgl.dir\Release\heterograph_capi.obj
         dgl.dir\Release\immutable_graph.obj
         dgl.dir\Release\metis_partition.obj
         dgl.dir\Release\network.obj
         dgl.dir\Release\nodeflow.obj
         dgl.dir\Release\pickle.obj
         dgl.dir\Release\sampler.obj
         dgl.dir\Release\global_uniform.obj
         dgl.dir\Release\neighbor.obj
         dgl.dir\Release\get_node_types_cpu.obj
         dgl.dir\Release\node2vec.obj
         dgl.dir\Release\node2vec_cpu.obj
         dgl.dir\Release\randomwalk_cpu.obj
         dgl.dir\Release\randomwalk_with_restart_cpu.obj
         dgl.dir\Release\randomwalks.obj
         dgl.dir\Release\dglgraph_serialize.obj
         dgl.dir\Release\graph_serialize.obj
         dgl.dir\Release\heterograph_serialize.obj
         dgl.dir\Release\tensor_serialize.obj
         dgl.dir\Release\zerocopy_serializer.obj
         dgl.dir\Release\shared_mem_manager.obj
         dgl.dir\Release\subgraph.obj
         dgl.dir\Release\compact.obj
         dgl.dir\Release\line_graph.obj
         dgl.dir\Release\metis_partition_hetero.obj
         dgl.dir\Release\partition_hetero.obj
         dgl.dir\Release\remove_edges.obj
         dgl.dir\Release\to_bipartite.obj
         dgl.dir\Release\to_simple.obj
         dgl.dir\Release\unit_graph.obj
         dgl.dir\Release\scheduler.obj
         dgl.dir\Release\scheduler_apis.obj
         dgl.dir\Release\common.obj
         dgl.dir\Release\msg_queue.obj
         dgl.dir\Release\socket_communicator.obj
         dgl.dir\Release\socket_pool.obj
         dgl.dir\Release\tcp_socket.obj
         dgl.dir\Release\/src/array/cpu/traversal.cc.obj
         dgl.dir\Release\/src/array/union_partition.cc.obj
         dgl.dir\Release\/src/graph/transform/cpu/knn.cc.obj
         dgl.dir\Release\/src/graph/transform/knn.cc.obj
         dgl.dir\Release\/src/graph/transform/union_partition.cc.obj
         dgl.dir\Release\/src/graph/traversal.cc.obj
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_min_f16.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
            Creating library C:/Jenkins/workspace/dgl_PR-4648/build/Release/dgl.lib and object C:/Jenkins/workspace/dgl_PR-4648/build/Release/dgl.exp
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_min_f32.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_min_f64.o
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_min_bf16.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_max_i8.o
         dgl.vcxproj -> C:\Jenkins\workspace\dgl_PR-4648\build\Release\dgl.dll
       PostBuildEvent:
         setlocal
         "C:\Program Files\CMake\bin\cmake.exe" -E copy C:/Jenkins/workspace/dgl_PR-4648/build/Release/dgl.dll C:/Jenkins/workspace/dgl_PR-4648/build/Release/..
         if %errorlevel% neq 0 goto :cmEnd
         :cmEnd
         endlocal & call :cmErrorLevel %errorlevel% & goto :cmDone
         :cmErrorLevel
         exit /b %1
         :cmDone
         if %errorlevel% neq 0 goto :VCEnd
         :VCEnd
       FinalizeBuildStatus:
         Deleting file "dgl.dir\Release\dgl.tlog\unsuccessfulbuild".
         Touching "dgl.dir\Release\dgl.tlog\dgl.lastbuildstate".
    18>Done Building Project "C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj" (default targets).
     3>Done Building Project "C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj.metaproj" (default targets).
     1>Project "C:\Jenkins\workspace\dgl_PR-4648\build\dgl.sln" (1) is building "C:\Jenkins\workspace\dgl_PR-4648\build\runUnitTests.vcxproj.metaproj" (9) on node 1 (default targets).
     9>Project "C:\Jenkins\workspace\dgl_PR-4648\build\runUnitTests.vcxproj.metaproj" (9) is building "C:\Jenkins\workspace\dgl_PR-4648\build\runUnitTests.vcxproj" (19) on node 3 (default targets).
    19>PrepareForBuild:
         Creating directory "runUnitTests.dir\Release\".
         Creating directory "runUnitTests.dir\Release\runUnitTests.tlog\".
       InitializeBuildStatus:
         Creating "runUnitTests.dir\Release\runUnitTests.tlog\unsuccessfulbuild" because "AlwaysCreate" was specified.
       CustomBuild:
         Building Custom Rule C:/Jenkins/workspace/dgl_PR-4648/CMakeLists.txt
       ClCompile:
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\bin\HostX64\x64\CL.exe /c /I"C:\Jenkins\workspace\dgl_PR-4648\include" /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\dlpack\include" /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\xbyak" /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\dmlc-core\include" /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\phmap" /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\libxsmm\include" /I"C:\Jenkins\workspace\dgl_PR-4648\build\third_party\dmlc-core\include" /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\googletest\googletest\include" /I"C:\Jenkins\workspace\dgl_PR-4648\third_party\googletest\googletest" /nologo /W1 /WX- /diagnostics:column /MP /O2 /Ob2 /D DGL_EXPORTS /D USE_AVX /D USE_LIBXSMM /D DGL_CPU_LLC_SIZE=40000000 /D IDXTYPEWIDTH=64 /D REALTYPEWIDTH=32 /D NDEBUG /D WIN32_LEAN_AND_MEAN /D _CRT_SECURE_NO_WARNINGS /D _SCL_SECURE_NO_WARNINGS /D NOMINMAX /D ENABLE_PARTIAL_FRONTIER=0 /D __USE_XOPEN2K8 /D DMLC_CORE_USE_CMAKE /D DMLC_USE_CXX11=1 /D "CMAKE_INTDIR=\"Release\"" /D _MBCS /Gm- /EHsc /MD /GS /fp:precise /Zc:wchar_t /Zc:forScope /Zc:inline /openmp /Fo"runUnitTests.dir\Release\\" /Fd"runUnitTests.dir\Release\vc142.pdb" /Gd /TP /errorReport:queue "C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\graph_index_test.cc" "C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\message_queue_test.cc" "C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\socket_communicator_test.cc" "C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\string_test.cc" "C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\test_aten.cc" "C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\test_csrmm.cc" "C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\test_partition.cc" "C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\test_rowwise.cc" "C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\test_sampler.cc" "C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\test_serialize.cc" "C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\test_smart_ptr_serialize.cc" "C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\test_spmat_coo.cc" "C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\test_spmat_csr.cc" "C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\test_spmm.cc" "C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\test_unit_graph.cc" "C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\test_zerocopy_serialize.cc"
         graph_index_test.cc
         message_queue_test.cc
         socket_communicator_test.cc
         string_test.cc
         test_aten.cc
         test_csrmm.cc
         test_partition.cc
         test_rowwise.cc
         test_sampler.cc
         test_serialize.cc
         test_smart_ptr_serialize.cc
         test_spmat_coo.cc
         test_spmat_csr.cc
         test_spmm.cc
         test_unit_graph.cc
         test_zerocopy_serialize.cc
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_max_u8.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_max_i32.o
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_max_u32.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_max_i64.o
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_max_u64.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_max_f16.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_max_f32.o
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_max_f64.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
    19>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\graph_index_test.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\runUnitTests.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\runUnitTests.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\graph_index_test.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\runUnitTests.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\graph_index_test.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\graph_index_test.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\runUnitTests.vcxproj]
    19>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\test_spmat_coo.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\runUnitTests.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\runUnitTests.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\test_spmat_coo.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\runUnitTests.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\test_spmat_coo.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\test_spmat_coo.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\runUnitTests.vcxproj]
    19>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\test_partition.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\runUnitTests.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\runUnitTests.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\test_partition.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\runUnitTests.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\test_partition.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\test_partition.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\runUnitTests.vcxproj]
    19>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\test_spmat_csr.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\runUnitTests.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\runUnitTests.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\test_spmat_csr.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\runUnitTests.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\test_spmat_csr.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\test_spmat_csr.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\runUnitTests.vcxproj]
    19>C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\test_unit_graph.cc(137,1): warning C4305: 'argument': truncation from 'int' to 'bool' [C:\Jenkins\workspace\dgl_PR-4648\build\runUnitTests.vcxproj]
       C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\test_unit_graph.cc(400): message : see reference to function template instantiation 'void _TestUnitGraph<int32_t>(DGLContext)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\runUnitTests.vcxproj]
    19>C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\test_unit_graph.cc(139,1): warning C4305: 'argument': truncation from 'int' to 'bool' [C:\Jenkins\workspace\dgl_PR-4648\build\runUnitTests.vcxproj]
    19>C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\test_unit_graph.cc(142,1): warning C4305: 'argument': truncation from 'int' to 'bool' [C:\Jenkins\workspace\dgl_PR-4648\build\runUnitTests.vcxproj]
    19>C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\test_unit_graph.cc(144,1): warning C4305: 'argument': truncation from 'int' to 'bool' [C:\Jenkins\workspace\dgl_PR-4648\build\runUnitTests.vcxproj]
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_max_bf16.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
    19>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\test_zerocopy_serialize.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\runUnitTests.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\runUnitTests.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\test_zerocopy_serialize.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\runUnitTests.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\test_zerocopy_serialize.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\test_zerocopy_serialize.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\runUnitTests.vcxproj]
    19>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\test_serialize.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\runUnitTests.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\runUnitTests.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\test_serialize.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\runUnitTests.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\test_serialize.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\test_serialize.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\runUnitTests.vcxproj]
    19>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\test_unit_graph.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\runUnitTests.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\runUnitTests.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\test_unit_graph.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\runUnitTests.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\test_unit_graph.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\test_unit_graph.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\runUnitTests.vcxproj]
    19>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\test_aten.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\runUnitTests.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\runUnitTests.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\test_aten.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\runUnitTests.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\test_aten.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\test_aten.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\runUnitTests.vcxproj]
    19>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(4088,24): warning C4244: '=': conversion from 'const _Ty' to 'float', possible loss of data [C:\Jenkins\workspace\dgl_PR-4648\build\runUnitTests.vcxproj]
                 with
                 [
                     _Ty=int
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\test_sampler.cc)
       C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\../../src/random/cpu/sample_utils.h(267): message : see reference to function template instantiation 'void std::fill<std::_Vector_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int>(const _FwdIt,const _FwdIt,const int &)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\runUnitTests.vcxproj]
                 with
                 [
                     _Ty=float,
                     _FwdIt=std::_Vector_iterator<std::_Vector_val<std::_Simple_types<float>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\test_sampler.cc)
       C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\../../src/random/cpu/sample_utils.h(264): message : while compiling class template member function 'void dgl::utils::TreeSampler<Idx,DType,false>::ResetState(dgl::FloatArray)' [C:\Jenkins\workspace\dgl_PR-4648\build\runUnitTests.vcxproj]
                 with
                 [
                     Idx=int64_t,
                     DType=float
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\test_sampler.cc)
       C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\../../src/random/cpu/sample_utils.h(281): message : see reference to function template instantiation 'void dgl::utils::TreeSampler<Idx,DType,false>::ResetState(dgl::FloatArray)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\runUnitTests.vcxproj]
                 with
                 [
                     Idx=int64_t,
                     DType=float
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\test_sampler.cc)
       C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\test_sampler.cc(86): message : see reference to class template instantiation 'dgl::utils::TreeSampler<Idx,DType,false>' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\runUnitTests.vcxproj]
                 with
                 [
                     Idx=int64_t,
                     DType=float
                 ]
       C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\test_sampler.cc(99): message : see reference to function template instantiation 'void _TestWithoutReplacementOrder<int64_t,float>(dgl::RandomEngine *)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\runUnitTests.vcxproj]
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_premulsum_i8.o
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_premulsum_u8.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
    19>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\test_sampler.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\runUnitTests.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\runUnitTests.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\test_sampler.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\runUnitTests.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\test_sampler.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\test_sampler.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\runUnitTests.vcxproj]
    19>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\test_csrmm.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\runUnitTests.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\runUnitTests.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\test_csrmm.cc)
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_premulsum_i32.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_premulsum_u32.o
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\runUnitTests.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\test_csrmm.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\test_csrmm.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\runUnitTests.vcxproj]
    19>C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\test_rowwise.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\runUnitTests.vcxproj]
       C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3643): message : see reference to function template instantiation '_OutIt std::_Copy_unchecked<const __int64*,_OutIt>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\runUnitTests.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _InIt=const __int64 *
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\test_rowwise.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/array_ops.h(349): message : see reference to function template instantiation '_OutIt std::copy<std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<_Ty>>>,int32_t*>(_InIt,_InIt,_OutIt)' being compiled [C:\Jenkins\workspace\dgl_PR-4648\build\runUnitTests.vcxproj]
                 with
                 [
                     _OutIt=int32_t *,
                     _Ty=int64_t,
                     _InIt=std::_Vector_const_iterator<std::_Vector_val<std::_Simple_types<int64_t>>>
                 ] (compiling source file C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\test_rowwise.cc)
       C:\Jenkins\workspace\dgl_PR-4648\include\dgl\./aten/csr.h(226): message : see reference to function template instantiation 'dgl::IdArray dgl::aten::VecToIdArray<int64_t>(const std::vector<int64_t,std::allocator<int64_t>> &,uint8_t,DGLContext)' being compiled (compiling source file C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\test_rowwise.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\runUnitTests.vcxproj]
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_premulsum_i64.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_premulsum_u64.o
[ 48%] Linking CXX static library ../../../lib/libgmock_main.a
[ 48%] Built target gmock_main
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_premulsum_f16.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_premulsum_f32.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_premulsum_f64.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_premulsum_bf16.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_sumpostdiv_i8.o
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_sumpostdiv_u8.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_sumpostdiv_i32.o
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_sumpostdiv_u32.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_sumpostdiv_i64.o
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_sumpostdiv_u64.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_sumpostdiv_f16.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_sumpostdiv_f32.o
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_sumpostdiv_f64.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce.cu                           > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_sumpostdiv_bf16.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_scatter_sum_i8.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_scatter_sum_u8.o
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_scatter_sum_i32.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_scatter_sum_u32.o
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_scatter_sum_i64.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_scatter_sum_u64.o
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_scatter_sum_f16.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_scatter_sum_f32.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_scatter_sum_f64.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_scatter_sum_bf16.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_scatter_prod_i8.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_scatter_prod_u8.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_scatter_prod_i32.o
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_scatter_prod_u32.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_scatter_prod_i64.o
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_scatter_prod_u64.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_scatter_prod_f16.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_scatter_prod_f32.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_scatter_prod_f64.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_scatter_prod_bf16.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_scatter_min_i8.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_scatter_min_u8.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_scatter_min_i32.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_scatter_min_u32.o
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_scatter_min_i64.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_scatter_min_u64.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_scatter_min_f16.o
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_scatter_min_f32.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_scatter_min_f64.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_scatter_min_bf16.o
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_scatter_max_i8.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_scatter_max_u8.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_scatter_max_i32.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_scatter_max_u32.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_scatter_max_i64.o
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_scatter_max_u64.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_scatter_max_f16.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_scatter_max_f32.o
       Link:
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\bin\HostX64\x64\link.exe /ERRORREPORT:QUEUE /OUT:"C:\Jenkins\workspace\dgl_PR-4648\build\Release\runUnitTests.exe" /INCREMENTAL:NO /NOLOGO lib\Release\gtest.lib lib\Release\gtest_main.lib Release\dgl.lib lib\Release\gtest.lib "third_party\dmlc-core\Release\dmlc.lib" kernel32.lib user32.lib gdi32.lib winspool.lib shell32.lib ole32.lib oleaut32.lib uuid.lib comdlg32.lib advapi32.lib /MANIFEST /MANIFESTUAC:"level='asInvoker' uiAccess='false'" /manifest:embed /PDB:"C:/Jenkins/workspace/dgl_PR-4648/build/Release/runUnitTests.pdb" /SUBSYSTEM:CONSOLE /TLBID:1 /DYNAMICBASE /NXCOMPAT /IMPLIB:"C:/Jenkins/workspace/dgl_PR-4648/build/Release/runUnitTests.lib" /MACHINE:X64  /machine:x64 runUnitTests.dir\Release\graph_index_test.obj
         runUnitTests.dir\Release\message_queue_test.obj
         runUnitTests.dir\Release\socket_communicator_test.obj
         runUnitTests.dir\Release\string_test.obj
         runUnitTests.dir\Release\test_aten.obj
         runUnitTests.dir\Release\test_csrmm.obj
         runUnitTests.dir\Release\test_partition.obj
         runUnitTests.dir\Release\test_rowwise.obj
         runUnitTests.dir\Release\test_sampler.obj
         runUnitTests.dir\Release\test_serialize.obj
         runUnitTests.dir\Release\test_smart_ptr_serialize.obj
         runUnitTests.dir\Release\test_spmat_coo.obj
         runUnitTests.dir\Release\test_spmat_csr.obj
         runUnitTests.dir\Release\test_spmm.obj
         runUnitTests.dir\Release\test_unit_graph.obj
         runUnitTests.dir\Release\test_zerocopy_serialize.obj
            Creating library C:/Jenkins/workspace/dgl_PR-4648/build/Release/runUnitTests.lib and object C:/Jenkins/workspace/dgl_PR-4648/build/Release/runUnitTests.exp
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_scatter_max_f64.o
         runUnitTests.vcxproj -> C:\Jenkins\workspace\dgl_PR-4648\build\Release\runUnitTests.exe
       FinalizeBuildStatus:
         Deleting file "runUnitTests.dir\Release\runUnitTests.tlog\unsuccessfulbuild".
         Touching "runUnitTests.dir\Release\runUnitTests.tlog\runUnitTests.lastbuildstate".
    19>Done Building Project "C:\Jenkins\workspace\dgl_PR-4648\build\runUnitTests.vcxproj" (default targets).
     9>Done Building Project "C:\Jenkins\workspace\dgl_PR-4648\build\runUnitTests.vcxproj.metaproj" (default targets).
     2>Project "C:\Jenkins\workspace\dgl_PR-4648\build\ALL_BUILD.vcxproj.metaproj" (2) is building "C:\Jenkins\workspace\dgl_PR-4648\build\ALL_BUILD.vcxproj" (20) on node 4 (default targets).
    20>PrepareForBuild:
         Creating directory "x64\Release\ALL_BUILD\".
         Creating directory "x64\Release\ALL_BUILD\ALL_BUILD.tlog\".
       InitializeBuildStatus:
         Creating "x64\Release\ALL_BUILD\ALL_BUILD.tlog\unsuccessfulbuild" because "AlwaysCreate" was specified.
       CustomBuild:
         Building Custom Rule C:/Jenkins/workspace/dgl_PR-4648/CMakeLists.txt
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_scatter_max_bf16.o
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_scatter_premulsum_i8.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
       FinalizeBuildStatus:
         Deleting file "x64\Release\ALL_BUILD\ALL_BUILD.tlog\unsuccessfulbuild".
         Touching "x64\Release\ALL_BUILD\ALL_BUILD.tlog\ALL_BUILD.lastbuildstate".
    20>Done Building Project "C:\Jenkins\workspace\dgl_PR-4648\build\ALL_BUILD.vcxproj" (default targets).
     2>Done Building Project "C:\Jenkins\workspace\dgl_PR-4648\build\ALL_BUILD.vcxproj.metaproj" (default targets).
     1>Done Building Project "C:\Jenkins\workspace\dgl_PR-4648\build\dgl.sln" (default targets).

Build succeeded.

       "C:\Jenkins\workspace\dgl_PR-4648\build\dgl.sln" (default target) (1) ->
       "C:\Jenkins\workspace\dgl_PR-4648\build\third_party\dmlc-core\dmlc.vcxproj.metaproj" (default target) (4) ->
       "C:\Jenkins\workspace\dgl_PR-4648\build\third_party\dmlc-core\dmlc.vcxproj" (default target) (16) ->
       (ClCompile target) -> 
         C:\Jenkins\workspace\dgl_PR-4648\third_party\dmlc-core\include\dmlc/filesystem.h(19,1): warning C4005: 'NOMINMAX': macro redefinition (compiling source file C:\Jenkins\workspace\dgl_PR-4648\third_party\dmlc-core\src\io\indexed_recordio_split.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\third_party\dmlc-core\dmlc.vcxproj]
         C:\Jenkins\workspace\dgl_PR-4648\third_party\dmlc-core\include\dmlc/filesystem.h(19,1): warning C4005: 'NOMINMAX': macro redefinition (compiling source file C:\Jenkins\workspace\dgl_PR-4648\third_party\dmlc-core\src\io\recordio_split.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\third_party\dmlc-core\dmlc.vcxproj]
         C:\Jenkins\workspace\dgl_PR-4648\third_party\dmlc-core\include\dmlc/filesystem.h(19,1): warning C4005: 'NOMINMAX': macro redefinition (compiling source file C:\Jenkins\workspace\dgl_PR-4648\third_party\dmlc-core\src\io\line_split.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\third_party\dmlc-core\dmlc.vcxproj]
         C:\Jenkins\workspace\dgl_PR-4648\third_party\dmlc-core\include\dmlc/filesystem.h(19,1): warning C4005: 'NOMINMAX': macro redefinition (compiling source file C:\Jenkins\workspace\dgl_PR-4648\third_party\dmlc-core\src\io\filesys.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\third_party\dmlc-core\dmlc.vcxproj]
         C:\Jenkins\workspace\dgl_PR-4648\third_party\dmlc-core\include\dmlc/filesystem.h(19,1): warning C4005: 'NOMINMAX': macro redefinition (compiling source file C:\Jenkins\workspace\dgl_PR-4648\third_party\dmlc-core\src\io\input_split_base.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\third_party\dmlc-core\dmlc.vcxproj]
         C:\Jenkins\workspace\dgl_PR-4648\third_party\dmlc-core\include\dmlc/filesystem.h(19,1): warning C4005: 'NOMINMAX': macro redefinition (compiling source file C:\Jenkins\workspace\dgl_PR-4648\third_party\dmlc-core\src\io.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\third_party\dmlc-core\dmlc.vcxproj]
         C:\Jenkins\workspace\dgl_PR-4648\third_party\dmlc-core\include\dmlc/filesystem.h(19,1): warning C4005: 'NOMINMAX': macro redefinition (compiling source file C:\Jenkins\workspace\dgl_PR-4648\third_party\dmlc-core\src\io\local_filesys.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\third_party\dmlc-core\dmlc.vcxproj]


       "C:\Jenkins\workspace\dgl_PR-4648\build\dgl.sln" (default target) (1) ->
       "C:\Jenkins\workspace\dgl_PR-4648\build\tensoradapter_pytorch.vcxproj.metaproj" (default target) (10) ->
       "C:\Jenkins\workspace\dgl_PR-4648\build\tensoradapter_pytorch.vcxproj" (default target) (12) ->
       (CustomBuild target) -> 
         CUSTOMBUILD : CMake warning :  [C:\Jenkins\workspace\dgl_PR-4648\build\tensoradapter_pytorch.vcxproj]


       "C:\Jenkins\workspace\dgl_PR-4648\build\dgl.sln" (default target) (1) ->
       "C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj.metaproj" (default target) (3) ->
       "C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj" (default target) (18) ->
       (ClCompile target) -> 
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\array_nonzero.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\array_cumsum.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\array_index_select.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\array_scatter.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\coo_coalesce.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\array_sort.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\array_repeat.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\coo_linegraph.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\array_pack.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\array_op_impl.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\gather_mm.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\coo_sort.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\csr_to_simple.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\csr_union.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\disjoint_union.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\csr_sort.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\rowwise_topk.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\segment_reduce.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\negative_sampling.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\filter.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\rowwise_sampling.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\array_arith.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\csr_get_data.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\coo_remove.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\utility(151,1): warning C4267: 'initializing': conversion from 'size_t' to '_Ty2', possible loss of data [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\csr_remove.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\utility(151,1): warning C4267: 'initializing': conversion from 'size_t' to '_Ty2', possible loss of data [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\csr_mm.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(4088,24): warning C4244: '=': conversion from 'const _Ty' to 'IdType', possible loss of data [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\csr_sum.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\utility(151,1): warning C4267: 'initializing': conversion from 'size_t' to '_Ty2', possible loss of data [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\array.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\spmat_op_impl_csr.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\spmm.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\sddmm.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\utility(151,1): warning C4267: 'initializing': conversion from 'size_t' to '_Ty2', possible loss of data [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\cpu\spmat_op_impl_coo.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\numeric(817,20): warning C4244: '=': conversion from '_Ty' to 'IdType', possible loss of data [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\utility(151,1): warning C4244: 'initializing': conversion from '_Ty' to '_Ty2', possible loss of data [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Jenkins\workspace\dgl_PR-4648\src\array\libra_partition.cc(396,16): warning C4477: 'fscanf' : format string '%ld' requires an argument of type 'long *', but variadic argument 1 has type 'int64_t *' [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Jenkins\workspace\dgl_PR-4648\src\array\libra_partition.cc(396,16): warning C4477: 'fscanf' : format string '%ld' requires an argument of type 'long *', but variadic argument 2 has type 'int64_t *' [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\uvm_array.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\c_api_common.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\geometry\cpu\geometry_op_impl.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\partition\ndarray_partition.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\geometry\geometry.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Jenkins\workspace\dgl_PR-4648\src\array\libra_partition.cc(273,21): warning C4477: 'fprintf' : format string '%ld' requires an argument of type 'long', but variadic argument 1 has type 'int64_t' [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Jenkins\workspace\dgl_PR-4648\src\array\libra_partition.cc(273,21): warning C4477: 'fprintf' : format string '%ld' requires an argument of type 'long', but variadic argument 2 has type 'int64_t' [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Jenkins\workspace\dgl_PR-4648\src\array\libra_partition.cc(273,21): warning C4477: 'fprintf' : format string '%ld' requires an argument of type 'long', but variadic argument 3 has type 'int64_t' [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Jenkins\workspace\dgl_PR-4648\src\array\libra_partition.cc(284,10): warning C4477: 'printf' : format string '%ld' requires an argument of type 'long', but variadic argument 1 has type 'unsigned __int64' [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Jenkins\workspace\dgl_PR-4648\src\array\libra_partition.cc(287,17): warning C4477: 'fprintf' : format string '%ld' requires an argument of type 'long', but variadic argument 1 has type 'int64_t' [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Jenkins\workspace\dgl_PR-4648\src\array\libra_partition.cc(291,12): warning C4477: 'printf' : format string '%ld' requires an argument of type 'long', but variadic argument 1 has type 'int64_t' [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Jenkins\workspace\dgl_PR-4648\src\array\libra_partition.cc(296,12): warning C4477: 'printf' : format string '%ld' requires an argument of type 'long', but variadic argument 1 has type 'int64_t' [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\random\random.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(4088,24): warning C4244: '=': conversion from 'const _Ty' to 'float', possible loss of data [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\random\cpu\choice.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\kernel.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\array\libra_partition.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\stdint.h(51,1): warning C4005: 'INT32_MIN': macro redefinition (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\metis_partition.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\stdint.h(52,1): warning C4005: 'INT64_MIN': macro redefinition (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\metis_partition.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\stdint.h(55,1): warning C4005: 'INT32_MAX': macro redefinition (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\metis_partition.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\stdint.h(56,1): warning C4005: 'INT64_MAX': macro redefinition (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\metis_partition.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\runtime\utils.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\creators.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\gk_ops.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\graph_traversal.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampling\randomwalks\get_node_types_cpu.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\graph.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\metis_partition.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\nodeflow.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\graph_apis.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampling\negative\global_uniform.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\heterograph.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\immutable_graph.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const unsigned __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\heterograph.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampling\randomwalks\node2vec.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampling\neighbor\neighbor.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampling\randomwalks\node2vec_cpu.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampler.cc(1526,1): warning C4805: '==': unsafe mix of type 'int64_t' and type 'bool' in operation [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\pickle.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xstddef(240,54): warning C4018: '<': signed/unsigned mismatch (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampling\randomwalks\node2vec_cpu.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampler.cc(1751,1): warning C4805: '==': unsafe mix of type 'int64_t' and type 'bool' in operation [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xstddef(242,43): warning C4018: '<': signed/unsigned mismatch (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampling\randomwalks\node2vec_cpu.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\graph_op.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const unsigned __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\graph_op.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampling\randomwalks\randomwalk_with_restart_cpu.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampling\randomwalks\randomwalks.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\network.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\serialize\tensor_serialize.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\heterograph_capi.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampling\randomwalks\randomwalk_cpu.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\utility(207,16): warning C4244: 'initializing': conversion from '_Ty' to '_Ty1', possible loss of data [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\serialize\dglgraph_serialize.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampler.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,1): warning C4267: '=': conversion from 'size_t' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\sampler.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\serialize\graph_serialize.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Jenkins\workspace\dgl_PR-4648\third_party\METIS\include\metis.h(75,1): warning C4005: 'INT32_MIN': macro redefinition (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\metis_partition_hetero.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Jenkins\workspace\dgl_PR-4648\third_party\METIS\include\metis.h(76,1): warning C4005: 'INT32_MAX': macro redefinition (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\metis_partition_hetero.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Jenkins\workspace\dgl_PR-4648\third_party\METIS\include\metis.h(77,1): warning C4005: 'INT64_MIN': macro redefinition (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\metis_partition_hetero.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Jenkins\workspace\dgl_PR-4648\third_party\METIS\include\metis.h(78,1): warning C4005: 'INT64_MAX': macro redefinition (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\metis_partition_hetero.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\serialize\heterograph_serialize.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\shared_mem_manager.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\subgraph.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\metis_partition_hetero.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\line_graph.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\remove_edges.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\tuple(632,57): warning C4244: '=': conversion from '_Ty' to 'IdType', possible loss of data [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\tuple(633,58): warning C4244: '=': conversion from '_Ty' to 'IdType', possible loss of data [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\utility(151,1): warning C4244: 'initializing': conversion from '_Ty' to '_Ty2', possible loss of data [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\compact.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\to_simple.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\partition_hetero.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const unsigned __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\partition_hetero.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\scheduler\scheduler_apis.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\unit_graph.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\utility(151,1): warning C4267: 'initializing': conversion from 'size_t' to '_Ty2', possible loss of data [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\graph\transform\to_bipartite.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\src\rpc\network\socket_communicator.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]


       "C:\Jenkins\workspace\dgl_PR-4648\build\dgl.sln" (default target) (1) ->
       "C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj.metaproj" (default target) (3) ->
       "C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj" (default target) (18) ->
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]


       "C:\Jenkins\workspace\dgl_PR-4648\build\dgl.sln" (default target) (1) ->
       "C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj.metaproj" (default target) (3) ->
       "C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj" (default target) (18) ->
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]


       "C:\Jenkins\workspace\dgl_PR-4648\build\dgl.sln" (default target) (1) ->
       "C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj.metaproj" (default target) (3) ->
       "C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj" (default target) (18) ->
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]


       "C:\Jenkins\workspace\dgl_PR-4648\build\dgl.sln" (default target) (1) ->
       "C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj.metaproj" (default target) (3) ->
       "C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj" (default target) (18) ->
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]


       "C:\Jenkins\workspace\dgl_PR-4648\build\dgl.sln" (default target) (1) ->
       "C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj.metaproj" (default target) (3) ->
       "C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj" (default target) (18) ->
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]


       "C:\Jenkins\workspace\dgl_PR-4648\build\dgl.sln" (default target) (1) ->
       "C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj.metaproj" (default target) (3) ->
       "C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj" (default target) (18) ->
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data [C:\Jenkins\workspace\dgl_PR-4648\build\dgl.vcxproj]


       "C:\Jenkins\workspace\dgl_PR-4648\build\dgl.sln" (default target) (1) ->
       "C:\Jenkins\workspace\dgl_PR-4648\build\runUnitTests.vcxproj.metaproj" (default target) (9) ->
       "C:\Jenkins\workspace\dgl_PR-4648\build\runUnitTests.vcxproj" (default target) (19) ->
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\graph_index_test.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\runUnitTests.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\test_spmat_coo.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\runUnitTests.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\test_partition.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\runUnitTests.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\test_spmat_csr.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\runUnitTests.vcxproj]
         C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\test_unit_graph.cc(137,1): warning C4305: 'argument': truncation from 'int' to 'bool' [C:\Jenkins\workspace\dgl_PR-4648\build\runUnitTests.vcxproj]
         C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\test_unit_graph.cc(139,1): warning C4305: 'argument': truncation from 'int' to 'bool' [C:\Jenkins\workspace\dgl_PR-4648\build\runUnitTests.vcxproj]
         C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\test_unit_graph.cc(142,1): warning C4305: 'argument': truncation from 'int' to 'bool' [C:\Jenkins\workspace\dgl_PR-4648\build\runUnitTests.vcxproj]
         C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\test_unit_graph.cc(144,1): warning C4305: 'argument': truncation from 'int' to 'bool' [C:\Jenkins\workspace\dgl_PR-4648\build\runUnitTests.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\test_zerocopy_serialize.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\runUnitTests.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\test_serialize.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\runUnitTests.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\test_unit_graph.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\runUnitTests.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\test_aten.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\runUnitTests.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(4088,24): warning C4244: '=': conversion from 'const _Ty' to 'float', possible loss of data [C:\Jenkins\workspace\dgl_PR-4648\build\runUnitTests.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\test_sampler.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\runUnitTests.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\test_csrmm.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\runUnitTests.vcxproj]
         C:\Program Files (x86)\Microsoft Visual Studio\2019\BuildTools\VC\Tools\MSVC\14.25.28610\include\xutility(3606,22): warning C4244: '=': conversion from 'const __int64' to 'int32_t', possible loss of data (compiling source file C:\Jenkins\workspace\dgl_PR-4648\tests\cpp\test_rowwise.cc) [C:\Jenkins\workspace\dgl_PR-4648\build\runUnitTests.vcxproj]

    148 Warning(s)
    0 Error(s)

Time Elapsed 00:00:44.54
        1 file(s) copied.
Could Not Find C:\Jenkins\workspace\dgl_PR-4648\python\build
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_scatter_premulsum_u8.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_scatter_premulsum_i32.o
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_scatter_premulsum_u32.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_scatter_premulsum_i64.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_scatter_premulsum_u64.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_scatter_premulsum_f16.o
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_scatter_premulsum_f32.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Obtaining file:///C:/Jenkins/workspace/dgl_PR-4648/python
  Preparing metadata (setup.py): started
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_scatter_premulsum_f64.o
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_scatter_premulsum_bf16.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_scatter_sumpostdiv_i8.o
  Preparing metadata (setup.py): finished with status 'done'
Requirement already satisfied: numpy>=1.14.0 in c:\program files\python36\lib\site-packages (from dgl==0.9) (1.19.5)
Requirement already satisfied: scipy>=1.1.0 in c:\program files\python36\lib\site-packages (from dgl==0.9) (1.2.0)
Requirement already satisfied: networkx>=2.1 in c:\program files\python36\lib\site-packages (from dgl==0.9) (2.2)
Requirement already satisfied: requests>=2.19.0 in c:\program files\python36\lib\site-packages (from dgl==0.9) (2.24.0)
Requirement already satisfied: tqdm in c:\program files\python36\lib\site-packages (from dgl==0.9) (4.58.0)
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_scatter_sumpostdiv_u8.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_scatter_sumpostdiv_i32.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_scatter_sumpostdiv_u32.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_scatter_sumpostdiv_i64.o
Collecting psutil>=5.8.0
  Using cached psutil-5.9.2-cp36-cp36m-win_amd64.whl (248 kB)
Requirement already satisfied: decorator>=4.3.0 in c:\program files\python36\lib\site-packages (from networkx>=2.1->dgl==0.9) (4.4.2)
Requirement already satisfied: chardet<4,>=3.0.2 in c:\program files\python36\lib\site-packages (from requests>=2.19.0->dgl==0.9) (3.0.4)
Requirement already satisfied: urllib3!=1.25.0,!=1.25.1,<1.26,>=1.21.1 in c:\program files\python36\lib\site-packages (from requests>=2.19.0->dgl==0.9) (1.25.10)
Requirement already satisfied: idna<3,>=2.5 in c:\program files\python36\lib\site-packages (from requests>=2.19.0->dgl==0.9) (2.10)
Requirement already satisfied: certifi>=2017.4.17 in c:\program files\python36\lib\site-packages (from requests>=2.19.0->dgl==0.9) (2020.6.20)
Installing collected packages: psutil, dgl
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_scatter_sumpostdiv_u64.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_scatter_sumpostdiv_f16.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
  Running setup.py develop for dgl
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_scatter_sumpostdiv_f32.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_scatter_sumpostdiv_f64.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Compiling  reduce_scatter.cu                   > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/reduce_scatter_sumpostdiv_bf16.o
Compiling  functions.cu                        > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/functions.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Successfully installed dgl psutil-5.9.2
[Pipeline] echo
Packing build\dgl.dll, build\runUnitTests.exe, build\tensoradapter\pytorch\*.dll into dgl-cpu-win64
Compiling  onerank_reduce.cu                   > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/onerank_reduce.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
[Pipeline] stash
Stashed 3 file(s)
Post stage
[Pipeline] cleanWs
[WS-CLEANUP] Deleting project workspace...
[WS-CLEANUP] Deferred wipeout is disabled by the job configuration...
/usr/bin/ar: creating lib/libxsmm.a
/usr/bin/ar: creating lib/libxsmmext.a
/usr/bin/ar: creating lib/libxsmm.a
/usr/bin/ar: creating lib/libxsmmext.a
================================================================================
LIBXSMM master-1.16.1-1534 (Linux@4e33dd6e703e)
--------------------------------------------------------------------------------
GNU Compiler Collection: gcc 7.5.0, and g++ 7.5.0
C / C++ target: -msse4.2
Fortran Compiler is disabled or missing: no Fortran interface is built!
--------------------------------------------------------------------------------
BLAS dependency and fallback is removed!
--------------------------------------------------------------------------------
[ 49%] Built target libxsmm
================================================================================
LIBXSMM master-1.16.1-1534 (Linux@681e11859cd1)
--------------------------------------------------------------------------------
GNU Compiler Collection: gcc 7.5.0, and g++ 7.5.0
C / C++ target: -msse4.2
Fortran Compiler is disabled or missing: no Fortran interface is built!
--------------------------------------------------------------------------------
BLAS dependency and fallback is removed!
--------------------------------------------------------------------------------
[ 48%] Built target libxsmm
[WS-CLEANUP] done
[Pipeline] }
[Pipeline] // withEnv
[Pipeline] }
[Pipeline] // node
[Pipeline] }
[Pipeline] // stage
[Pipeline] }
[ 92%] Linking CXX shared library libdgl.so
[ 92%] Built target dgl
Scanning dependencies of target rpc_client
Scanning dependencies of target rpc_server
[ 93%] Building CXX object CMakeFiles/rpc_client.dir/tests/dist/cpp/rpc_client.cc.o
[ 93%] Building CXX object CMakeFiles/rpc_server.dir/tests/dist/cpp/rpc_server.cc.o
Scanning dependencies of target runUnitTests
[ 95%] Building CXX object CMakeFiles/runUnitTests.dir/tests/cpp/graph_index_test.cc.o
[ 95%] Building CXX object CMakeFiles/runUnitTests.dir/tests/cpp/test_csrmm.cc.o
[ 96%] Building CXX object CMakeFiles/runUnitTests.dir/tests/cpp/test_partition.cc.o
[ 97%] Building CXX object CMakeFiles/runUnitTests.dir/tests/cpp/test_aten.cc.o
[ 96%] Building CXX object CMakeFiles/runUnitTests.dir/tests/cpp/test_spmat_coo.cc.o
[ 98%] Building CXX object CMakeFiles/runUnitTests.dir/tests/cpp/test_rowwise.cc.o
[ 97%] Building CXX object CMakeFiles/runUnitTests.dir/tests/cpp/test_spmm.cc.o
[ 97%] Building CXX object CMakeFiles/runUnitTests.dir/tests/cpp/socket_communicator_test.cc.o
[ 97%] Building CXX object CMakeFiles/runUnitTests.dir/tests/cpp/test_spmat_csr.cc.o
[100%] Building CXX object CMakeFiles/runUnitTests.dir/tests/cpp/test_zerocopy_serialize.cc.o
[100%] Building CXX object CMakeFiles/runUnitTests.dir/tests/cpp/test_serialize.cc.o
[ 98%] Building CXX object CMakeFiles/runUnitTests.dir/tests/cpp/message_queue_test.cc.o
[ 98%] Building CXX object CMakeFiles/runUnitTests.dir/tests/cpp/test_unit_graph.cc.o
[ 98%] Building CXX object CMakeFiles/runUnitTests.dir/tests/cpp/string_test.cc.o
[ 98%] Building CXX object CMakeFiles/runUnitTests.dir/tests/cpp/test_sampler.cc.o
[ 98%] Building CXX object CMakeFiles/runUnitTests.dir/tests/cpp/test_smart_ptr_serialize.cc.o
/root/jenkins/workspace/dgl_PR-4648/tests/dist/cpp/rpc_client.cc: In member function 'void RPCClient::StartClient()':
/root/jenkins/workspace/dgl_PR-4648/tests/dist/cpp/rpc_client.cc:50:15: warning: unused variable 'num_machines' [-Wunused-variable]
     const int num_machines = ips_.size();
               ^~~~~~~~~~~~
[100%] Linking CXX executable rpc_server
[100%] Linking CXX executable rpc_client
[100%] Built target rpc_server
[100%] Built target rpc_client
[100%] Linking CXX executable runUnitTests
[100%] Built target runUnitTests
~/jenkins/workspace/dgl_PR-4648
~/jenkins/workspace/dgl_PR-4648/python ~/jenkins/workspace/dgl_PR-4648
WARNING: Skipping dgl as it is not installed.
WARNING: Running pip as the 'root' user can result in broken permissions and conflicting behaviour with the system package manager. It is recommended to use a virtual environment instead: https://pip.pypa.io/warnings/venv
[1/1] Cythonizing dgl/_ffi/_cython/core.pyx
running install
running bdist_egg
running egg_info
creating dgl.egg-info
writing dgl.egg-info/PKG-INFO
writing dependency_links to dgl.egg-info/dependency_links.txt
writing requirements to dgl.egg-info/requires.txt
writing top-level names to dgl.egg-info/top_level.txt
writing manifest file 'dgl.egg-info/SOURCES.txt'
reading manifest file 'dgl.egg-info/SOURCES.txt'
writing manifest file 'dgl.egg-info/SOURCES.txt'
installing library code to build/bdist.linux-x86_64/egg
running install_lib
running build_py
creating build
creating build/lib.linux-x86_64-3.7
creating build/lib.linux-x86_64-3.7/dgl
copying dgl/partition.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/core.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/subgraph.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/traversal.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/base.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/heterograph_index.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/sparse.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/global_config.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/_api_internal.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/heterograph.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/network.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/logging.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/graph_index.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/udf.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/ndarray.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/init.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/view.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/generators.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/convert.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/container.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/merge.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/readout.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/__init__.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/propagate.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/batch.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/frame.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/random.py -> build/lib.linux-x86_64-3.7/dgl
creating build/lib.linux-x86_64-3.7/dgl/mock_sparse
copying dgl/mock_sparse/sp_matrix.py -> build/lib.linux-x86_64-3.7/dgl/mock_sparse
copying dgl/mock_sparse/diag_matrix.py -> build/lib.linux-x86_64-3.7/dgl/mock_sparse
copying dgl/mock_sparse/reduction.py -> build/lib.linux-x86_64-3.7/dgl/mock_sparse
copying dgl/mock_sparse/__init__.py -> build/lib.linux-x86_64-3.7/dgl/mock_sparse
copying dgl/mock_sparse/elementwise_op_sp.py -> build/lib.linux-x86_64-3.7/dgl/mock_sparse
creating build/lib.linux-x86_64-3.7/dgl/geometry
copying dgl/geometry/edge_coarsening.py -> build/lib.linux-x86_64-3.7/dgl/geometry
copying dgl/geometry/fps.py -> build/lib.linux-x86_64-3.7/dgl/geometry
copying dgl/geometry/capi.py -> build/lib.linux-x86_64-3.7/dgl/geometry
copying dgl/geometry/__init__.py -> build/lib.linux-x86_64-3.7/dgl/geometry
creating build/lib.linux-x86_64-3.7/dgl/function
copying dgl/function/base.py -> build/lib.linux-x86_64-3.7/dgl/function
copying dgl/function/message.py -> build/lib.linux-x86_64-3.7/dgl/function
copying dgl/function/reducer.py -> build/lib.linux-x86_64-3.7/dgl/function
copying dgl/function/__init__.py -> build/lib.linux-x86_64-3.7/dgl/function
creating build/lib.linux-x86_64-3.7/dgl/nn
copying dgl/nn/__init__.py -> build/lib.linux-x86_64-3.7/dgl/nn
creating build/lib.linux-x86_64-3.7/dgl/contrib
copying dgl/contrib/dis_kvstore.py -> build/lib.linux-x86_64-3.7/dgl/contrib
copying dgl/contrib/unified_tensor.py -> build/lib.linux-x86_64-3.7/dgl/contrib
copying dgl/contrib/graph_store.py -> build/lib.linux-x86_64-3.7/dgl/contrib
copying dgl/contrib/__init__.py -> build/lib.linux-x86_64-3.7/dgl/contrib
creating build/lib.linux-x86_64-3.7/dgl/utils
copying dgl/utils/checks.py -> build/lib.linux-x86_64-3.7/dgl/utils
copying dgl/utils/exception.py -> build/lib.linux-x86_64-3.7/dgl/utils
copying dgl/utils/pin_memory.py -> build/lib.linux-x86_64-3.7/dgl/utils
copying dgl/utils/internal.py -> build/lib.linux-x86_64-3.7/dgl/utils
copying dgl/utils/filter.py -> build/lib.linux-x86_64-3.7/dgl/utils
copying dgl/utils/data.py -> build/lib.linux-x86_64-3.7/dgl/utils
copying dgl/utils/__init__.py -> build/lib.linux-x86_64-3.7/dgl/utils
copying dgl/utils/shared_mem.py -> build/lib.linux-x86_64-3.7/dgl/utils
creating build/lib.linux-x86_64-3.7/dgl/multiprocessing
copying dgl/multiprocessing/pytorch.py -> build/lib.linux-x86_64-3.7/dgl/multiprocessing
copying dgl/multiprocessing/__init__.py -> build/lib.linux-x86_64-3.7/dgl/multiprocessing
creating build/lib.linux-x86_64-3.7/dgl/optim
copying dgl/optim/__init__.py -> build/lib.linux-x86_64-3.7/dgl/optim
creating build/lib.linux-x86_64-3.7/dgl/dataloading
copying dgl/dataloading/base.py -> build/lib.linux-x86_64-3.7/dgl/dataloading
copying dgl/dataloading/dataloader.py -> build/lib.linux-x86_64-3.7/dgl/dataloading
copying dgl/dataloading/graphsaint.py -> build/lib.linux-x86_64-3.7/dgl/dataloading
copying dgl/dataloading/shadow.py -> build/lib.linux-x86_64-3.7/dgl/dataloading
copying dgl/dataloading/negative_sampler.py -> build/lib.linux-x86_64-3.7/dgl/dataloading
copying dgl/dataloading/__init__.py -> build/lib.linux-x86_64-3.7/dgl/dataloading
copying dgl/dataloading/cluster_gcn.py -> build/lib.linux-x86_64-3.7/dgl/dataloading
copying dgl/dataloading/neighbor_sampler.py -> build/lib.linux-x86_64-3.7/dgl/dataloading
copying dgl/dataloading/dist_dataloader.py -> build/lib.linux-x86_64-3.7/dgl/dataloading
creating build/lib.linux-x86_64-3.7/dgl/_dataloading
copying dgl/_dataloading/neighbor.py -> build/lib.linux-x86_64-3.7/dgl/_dataloading
copying dgl/_dataloading/dataloader.py -> build/lib.linux-x86_64-3.7/dgl/_dataloading
copying dgl/_dataloading/shadow.py -> build/lib.linux-x86_64-3.7/dgl/_dataloading
copying dgl/_dataloading/negative_sampler.py -> build/lib.linux-x86_64-3.7/dgl/_dataloading
copying dgl/_dataloading/__init__.py -> build/lib.linux-x86_64-3.7/dgl/_dataloading
copying dgl/_dataloading/cluster_gcn.py -> build/lib.linux-x86_64-3.7/dgl/_dataloading
creating build/lib.linux-x86_64-3.7/dgl/backend
copying dgl/backend/backend.py -> build/lib.linux-x86_64-3.7/dgl/backend
copying dgl/backend/__init__.py -> build/lib.linux-x86_64-3.7/dgl/backend
copying dgl/backend/set_default_backend.py -> build/lib.linux-x86_64-3.7/dgl/backend
creating build/lib.linux-x86_64-3.7/dgl/ops
copying dgl/ops/edge_softmax.py -> build/lib.linux-x86_64-3.7/dgl/ops
copying dgl/ops/sddmm.py -> build/lib.linux-x86_64-3.7/dgl/ops
copying dgl/ops/gather_mm.py -> build/lib.linux-x86_64-3.7/dgl/ops
copying dgl/ops/segment.py -> build/lib.linux-x86_64-3.7/dgl/ops
copying dgl/ops/__init__.py -> build/lib.linux-x86_64-3.7/dgl/ops
copying dgl/ops/spmm.py -> build/lib.linux-x86_64-3.7/dgl/ops
creating build/lib.linux-x86_64-3.7/dgl/cuda
copying dgl/cuda/nccl.py -> build/lib.linux-x86_64-3.7/dgl/cuda
copying dgl/cuda/__init__.py -> build/lib.linux-x86_64-3.7/dgl/cuda
creating build/lib.linux-x86_64-3.7/dgl/storages
copying dgl/storages/pytorch_tensor.py -> build/lib.linux-x86_64-3.7/dgl/storages
copying dgl/storages/base.py -> build/lib.linux-x86_64-3.7/dgl/storages
copying dgl/storages/tensor.py -> build/lib.linux-x86_64-3.7/dgl/storages
copying dgl/storages/__init__.py -> build/lib.linux-x86_64-3.7/dgl/storages
copying dgl/storages/numpy.py -> build/lib.linux-x86_64-3.7/dgl/storages
creating build/lib.linux-x86_64-3.7/dgl/sampling
copying dgl/sampling/neighbor.py -> build/lib.linux-x86_64-3.7/dgl/sampling
copying dgl/sampling/pinsage.py -> build/lib.linux-x86_64-3.7/dgl/sampling
copying dgl/sampling/negative.py -> build/lib.linux-x86_64-3.7/dgl/sampling
copying dgl/sampling/randomwalks.py -> build/lib.linux-x86_64-3.7/dgl/sampling
copying dgl/sampling/utils.py -> build/lib.linux-x86_64-3.7/dgl/sampling
copying dgl/sampling/node2vec_randomwalk.py -> build/lib.linux-x86_64-3.7/dgl/sampling
copying dgl/sampling/__init__.py -> build/lib.linux-x86_64-3.7/dgl/sampling
creating build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/qm7b.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/graph_serialize.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/dgl_dataset.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/csv_dataset_base.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/knowledge_graph.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/qm9_edge.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/gnn_benchmark.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/qm9.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/icews18.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/gdelt.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/utils.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/sbm.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/csv_dataset.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/karate.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/citation_graph.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/rdf.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/bitcoinotc.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/ppi.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/tensor_serialize.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/flickr.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/wikics.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/heterograph_serialize.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/tu.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/__init__.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/gindt.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/tree.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/synthetic.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/adapter.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/fakenews.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/fraud.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/minigc.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/yelp.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/reddit.py -> build/lib.linux-x86_64-3.7/dgl/data
creating build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/dist_graph.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/graph_partition_book.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/partition.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/standalone_kvstore.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/graph_services.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/role.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/constants.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/rpc_server.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/kvstore.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/dist_context.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/rpc.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/rpc_client.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/id_map.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/server_state.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/__init__.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/shared_mem_utils.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/dist_tensor.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/dist_dataloader.py -> build/lib.linux-x86_64-3.7/dgl/distributed
creating build/lib.linux-x86_64-3.7/dgl/_deprecate
copying dgl/_deprecate/nodeflow.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate
copying dgl/_deprecate/udf.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate
copying dgl/_deprecate/view.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate
copying dgl/_deprecate/kernel.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate
copying dgl/_deprecate/__init__.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate
copying dgl/_deprecate/graph.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate
copying dgl/_deprecate/frame.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate
creating build/lib.linux-x86_64-3.7/dgl/transforms
copying dgl/transforms/functional.py -> build/lib.linux-x86_64-3.7/dgl/transforms
copying dgl/transforms/__init__.py -> build/lib.linux-x86_64-3.7/dgl/transforms
copying dgl/transforms/module.py -> build/lib.linux-x86_64-3.7/dgl/transforms
creating build/lib.linux-x86_64-3.7/dgl/distgnn
copying dgl/distgnn/__init__.py -> build/lib.linux-x86_64-3.7/dgl/distgnn
creating build/lib.linux-x86_64-3.7/dgl/_ffi
copying dgl/_ffi/runtime_ctypes.py -> build/lib.linux-x86_64-3.7/dgl/_ffi
copying dgl/_ffi/function.py -> build/lib.linux-x86_64-3.7/dgl/_ffi
copying dgl/_ffi/libinfo.py -> build/lib.linux-x86_64-3.7/dgl/_ffi
copying dgl/_ffi/streams.py -> build/lib.linux-x86_64-3.7/dgl/_ffi
copying dgl/_ffi/base.py -> build/lib.linux-x86_64-3.7/dgl/_ffi
copying dgl/_ffi/ndarray.py -> build/lib.linux-x86_64-3.7/dgl/_ffi
copying dgl/_ffi/object_generic.py -> build/lib.linux-x86_64-3.7/dgl/_ffi
copying dgl/_ffi/__init__.py -> build/lib.linux-x86_64-3.7/dgl/_ffi
copying dgl/_ffi/object.py -> build/lib.linux-x86_64-3.7/dgl/_ffi
creating build/lib.linux-x86_64-3.7/dgl/nn/pytorch
copying dgl/nn/pytorch/glob.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch
copying dgl/nn/pytorch/hetero.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch
copying dgl/nn/pytorch/sparse_emb.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch
copying dgl/nn/pytorch/linear.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch
copying dgl/nn/pytorch/utils.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch
copying dgl/nn/pytorch/__init__.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch
copying dgl/nn/pytorch/factory.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch
copying dgl/nn/pytorch/softmax.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch
creating build/lib.linux-x86_64-3.7/dgl/nn/mxnet
copying dgl/nn/mxnet/glob.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet
copying dgl/nn/mxnet/hetero.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet
copying dgl/nn/mxnet/utils.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet
copying dgl/nn/mxnet/__init__.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet
copying dgl/nn/mxnet/softmax.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet
creating build/lib.linux-x86_64-3.7/dgl/nn/tensorflow
copying dgl/nn/tensorflow/glob.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow
copying dgl/nn/tensorflow/hetero.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow
copying dgl/nn/tensorflow/utils.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow
copying dgl/nn/tensorflow/__init__.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow
copying dgl/nn/tensorflow/softmax.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow
creating build/lib.linux-x86_64-3.7/dgl/nn/functional
copying dgl/nn/functional/__init__.py -> build/lib.linux-x86_64-3.7/dgl/nn/functional
creating build/lib.linux-x86_64-3.7/dgl/nn/pytorch/link
copying dgl/nn/pytorch/link/transe.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/link
copying dgl/nn/pytorch/link/edgepred.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/link
copying dgl/nn/pytorch/link/transr.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/link
copying dgl/nn/pytorch/link/__init__.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/link
creating build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/agnnconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/relgraphconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/appnpconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/dgnconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/grouprevres.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/twirlsconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/egatconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/graphconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/densechebconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/gcn2conv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/ginconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/dotgatconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/cfconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/gmmconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/sageconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/densesageconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/gineconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/edgeconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/nnconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/__init__.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/gatv2conv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/hgtconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/pnaconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/gatedgraphconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/atomicconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/tagconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/densegraphconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/egnnconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/chebconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/sgconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/gatconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
creating build/lib.linux-x86_64-3.7/dgl/nn/pytorch/explain
copying dgl/nn/pytorch/explain/gnnexplainer.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/explain
copying dgl/nn/pytorch/explain/__init__.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/explain
creating build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/agnnconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/relgraphconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/appnpconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/graphconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/densechebconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/ginconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/gmmconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/sageconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/densesageconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/edgeconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/nnconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/__init__.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/gatedgraphconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/tagconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/densegraphconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/chebconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/sgconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/gatconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
creating build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv
copying dgl/nn/tensorflow/conv/relgraphconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv
copying dgl/nn/tensorflow/conv/appnpconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv
copying dgl/nn/tensorflow/conv/graphconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv
copying dgl/nn/tensorflow/conv/densechebconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv
copying dgl/nn/tensorflow/conv/ginconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv
copying dgl/nn/tensorflow/conv/sageconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv
copying dgl/nn/tensorflow/conv/edgeconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv
copying dgl/nn/tensorflow/conv/__init__.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv
copying dgl/nn/tensorflow/conv/chebconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv
copying dgl/nn/tensorflow/conv/sgconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv
copying dgl/nn/tensorflow/conv/gatconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv
creating build/lib.linux-x86_64-3.7/dgl/contrib/sampling
copying dgl/contrib/sampling/sampler.py -> build/lib.linux-x86_64-3.7/dgl/contrib/sampling
copying dgl/contrib/sampling/dis_sampler.py -> build/lib.linux-x86_64-3.7/dgl/contrib/sampling
copying dgl/contrib/sampling/__init__.py -> build/lib.linux-x86_64-3.7/dgl/contrib/sampling
creating build/lib.linux-x86_64-3.7/dgl/contrib/data
copying dgl/contrib/data/knowledge_graph.py -> build/lib.linux-x86_64-3.7/dgl/contrib/data
copying dgl/contrib/data/__init__.py -> build/lib.linux-x86_64-3.7/dgl/contrib/data
creating build/lib.linux-x86_64-3.7/dgl/optim/pytorch
copying dgl/optim/pytorch/__init__.py -> build/lib.linux-x86_64-3.7/dgl/optim/pytorch
copying dgl/optim/pytorch/sparse_optim.py -> build/lib.linux-x86_64-3.7/dgl/optim/pytorch
creating build/lib.linux-x86_64-3.7/dgl/optim/mxnet
copying dgl/optim/mxnet/__init__.py -> build/lib.linux-x86_64-3.7/dgl/optim/mxnet
creating build/lib.linux-x86_64-3.7/dgl/optim/tensorflow
copying dgl/optim/tensorflow/__init__.py -> build/lib.linux-x86_64-3.7/dgl/optim/tensorflow
creating build/lib.linux-x86_64-3.7/dgl/_dataloading/pytorch
copying dgl/_dataloading/pytorch/dataloader.py -> build/lib.linux-x86_64-3.7/dgl/_dataloading/pytorch
copying dgl/_dataloading/pytorch/__init__.py -> build/lib.linux-x86_64-3.7/dgl/_dataloading/pytorch
creating build/lib.linux-x86_64-3.7/dgl/backend/pytorch
copying dgl/backend/pytorch/sparse.py -> build/lib.linux-x86_64-3.7/dgl/backend/pytorch
copying dgl/backend/pytorch/tensor.py -> build/lib.linux-x86_64-3.7/dgl/backend/pytorch
copying dgl/backend/pytorch/__init__.py -> build/lib.linux-x86_64-3.7/dgl/backend/pytorch
creating build/lib.linux-x86_64-3.7/dgl/backend/mxnet
copying dgl/backend/mxnet/sparse.py -> build/lib.linux-x86_64-3.7/dgl/backend/mxnet
copying dgl/backend/mxnet/tensor.py -> build/lib.linux-x86_64-3.7/dgl/backend/mxnet
copying dgl/backend/mxnet/__init__.py -> build/lib.linux-x86_64-3.7/dgl/backend/mxnet
copying dgl/backend/mxnet/sparse_optim.py -> build/lib.linux-x86_64-3.7/dgl/backend/mxnet
creating build/lib.linux-x86_64-3.7/dgl/backend/tensorflow
copying dgl/backend/tensorflow/sparse.py -> build/lib.linux-x86_64-3.7/dgl/backend/tensorflow
copying dgl/backend/tensorflow/tensor.py -> build/lib.linux-x86_64-3.7/dgl/backend/tensorflow
copying dgl/backend/tensorflow/__init__.py -> build/lib.linux-x86_64-3.7/dgl/backend/tensorflow
copying dgl/backend/tensorflow/sparse_optim.py -> build/lib.linux-x86_64-3.7/dgl/backend/tensorflow
creating build/lib.linux-x86_64-3.7/dgl/distributed/nn
copying dgl/distributed/nn/__init__.py -> build/lib.linux-x86_64-3.7/dgl/distributed/nn
creating build/lib.linux-x86_64-3.7/dgl/distributed/optim
copying dgl/distributed/optim/__init__.py -> build/lib.linux-x86_64-3.7/dgl/distributed/optim
creating build/lib.linux-x86_64-3.7/dgl/distributed/nn/pytorch
copying dgl/distributed/nn/pytorch/sparse_emb.py -> build/lib.linux-x86_64-3.7/dgl/distributed/nn/pytorch
copying dgl/distributed/nn/pytorch/__init__.py -> build/lib.linux-x86_64-3.7/dgl/distributed/nn/pytorch
creating build/lib.linux-x86_64-3.7/dgl/distributed/nn/mxnet
copying dgl/distributed/nn/mxnet/__init__.py -> build/lib.linux-x86_64-3.7/dgl/distributed/nn/mxnet
creating build/lib.linux-x86_64-3.7/dgl/distributed/nn/tensorflow
copying dgl/distributed/nn/tensorflow/__init__.py -> build/lib.linux-x86_64-3.7/dgl/distributed/nn/tensorflow
creating build/lib.linux-x86_64-3.7/dgl/distributed/optim/pytorch
copying dgl/distributed/optim/pytorch/utils.py -> build/lib.linux-x86_64-3.7/dgl/distributed/optim/pytorch
copying dgl/distributed/optim/pytorch/__init__.py -> build/lib.linux-x86_64-3.7/dgl/distributed/optim/pytorch
copying dgl/distributed/optim/pytorch/sparse_optim.py -> build/lib.linux-x86_64-3.7/dgl/distributed/optim/pytorch
creating build/lib.linux-x86_64-3.7/dgl/distributed/optim/mxnet
copying dgl/distributed/optim/mxnet/__init__.py -> build/lib.linux-x86_64-3.7/dgl/distributed/optim/mxnet
creating build/lib.linux-x86_64-3.7/dgl/distributed/optim/tensorflow
copying dgl/distributed/optim/tensorflow/__init__.py -> build/lib.linux-x86_64-3.7/dgl/distributed/optim/tensorflow
creating build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime
copying dgl/_deprecate/runtime/spmv.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime
copying dgl/_deprecate/runtime/degree_bucketing.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime
copying dgl/_deprecate/runtime/__init__.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime
copying dgl/_deprecate/runtime/scheduler.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime
copying dgl/_deprecate/runtime/adapter.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime
copying dgl/_deprecate/runtime/runtime.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime
creating build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/ir
copying dgl/_deprecate/runtime/ir/program.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/ir
copying dgl/_deprecate/runtime/ir/executor.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/ir
copying dgl/_deprecate/runtime/ir/var.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/ir
copying dgl/_deprecate/runtime/ir/registry.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/ir
copying dgl/_deprecate/runtime/ir/__init__.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/ir
creating build/lib.linux-x86_64-3.7/dgl/distgnn/partition
copying dgl/distgnn/partition/libra_partition.py -> build/lib.linux-x86_64-3.7/dgl/distgnn/partition
copying dgl/distgnn/partition/__init__.py -> build/lib.linux-x86_64-3.7/dgl/distgnn/partition
creating build/lib.linux-x86_64-3.7/dgl/distgnn/tools
copying dgl/distgnn/tools/__init__.py -> build/lib.linux-x86_64-3.7/dgl/distgnn/tools
copying dgl/distgnn/tools/tools.py -> build/lib.linux-x86_64-3.7/dgl/distgnn/tools
creating build/lib.linux-x86_64-3.7/dgl/_ffi/_ctypes
copying dgl/_ffi/_ctypes/function.py -> build/lib.linux-x86_64-3.7/dgl/_ffi/_ctypes
copying dgl/_ffi/_ctypes/types.py -> build/lib.linux-x86_64-3.7/dgl/_ffi/_ctypes
copying dgl/_ffi/_ctypes/ndarray.py -> build/lib.linux-x86_64-3.7/dgl/_ffi/_ctypes
copying dgl/_ffi/_ctypes/__init__.py -> build/lib.linux-x86_64-3.7/dgl/_ffi/_ctypes
copying dgl/_ffi/_ctypes/object.py -> build/lib.linux-x86_64-3.7/dgl/_ffi/_ctypes
creating build/lib.linux-x86_64-3.7/dgl/_ffi/_cy3
copying dgl/_ffi/_cy3/__init__.py -> build/lib.linux-x86_64-3.7/dgl/_ffi/_cy3
creating build/lib.linux-x86_64-3.7/dgl/_ffi/_cy2
copying dgl/_ffi/_cy2/__init__.py -> build/lib.linux-x86_64-3.7/dgl/_ffi/_cy2
creating build/lib.linux-x86_64-3.7/dgl/_ffi/_cython
copying dgl/_ffi/_cython/core.cpp -> build/lib.linux-x86_64-3.7/dgl/_ffi/_cython
running build_ext
building 'dgl._ffi._cy3.core' extension
creating build/temp.linux-x86_64-3.7
creating build/temp.linux-x86_64-3.7/dgl
creating build/temp.linux-x86_64-3.7/dgl/_ffi
creating build/temp.linux-x86_64-3.7/dgl/_ffi/_cython
gcc -pthread -B /opt/conda/envs/pytorch-ci/compiler_compat -Wl,--sysroot=/ -Wsign-compare -DNDEBUG -g -fwrapv -O3 -Wall -Wstrict-prototypes -fPIC -I../include/ -I../third_party/dmlc-core/include -I../third_party/dlpack/include -I/opt/conda/envs/pytorch-ci/include/python3.7m -c dgl/_ffi/_cython/core.cpp -o build/temp.linux-x86_64-3.7/dgl/_ffi/_cython/core.o
cc1plus: warning: command line option ‘-Wstrict-prototypes’ is valid for C/ObjC but not for C++
g++ -pthread -B /opt/conda/envs/pytorch-ci/compiler_compat -Wl,--sysroot=/ -pthread -shared -B /opt/conda/envs/pytorch-ci/compiler_compat -L/opt/conda/envs/pytorch-ci/lib -Wl,-rpath=/opt/conda/envs/pytorch-ci/lib -Wl,--no-as-needed -Wl,--sysroot=/ build/temp.linux-x86_64-3.7/dgl/_ffi/_cython/core.o -Ldgl -L../build/Release -L../build -ldgl -o build/lib.linux-x86_64-3.7/dgl/_ffi/_cy3/core.cpython-37m-x86_64-linux-gnu.so
/opt/conda/envs/pytorch-ci/compiler_compat/ld: warning: /opt/conda/envs/pytorch-ci/lib/libstdc++.so: unsupported GNU_PROPERTY_TYPE (5) type: 0xc0010001
/opt/conda/envs/pytorch-ci/compiler_compat/ld: warning: /opt/conda/envs/pytorch-ci/lib/libstdc++.so: unsupported GNU_PROPERTY_TYPE (5) type: 0xc0010002
/opt/conda/envs/pytorch-ci/compiler_compat/ld: warning: /opt/conda/envs/pytorch-ci/lib/libgcc_s.so.1: unsupported GNU_PROPERTY_TYPE (5) type: 0xc0010001
/opt/conda/envs/pytorch-ci/compiler_compat/ld: warning: /opt/conda/envs/pytorch-ci/lib/libgcc_s.so.1: unsupported GNU_PROPERTY_TYPE (5) type: 0xc0010002
/opt/conda/envs/pytorch-ci/compiler_compat/ld: warning: /opt/conda/envs/pytorch-ci/lib/libgcc_s.so.1: unsupported GNU_PROPERTY_TYPE (5) type: 0xc0010001
/opt/conda/envs/pytorch-ci/compiler_compat/ld: warning: /opt/conda/envs/pytorch-ci/lib/libgcc_s.so.1: unsupported GNU_PROPERTY_TYPE (5) type: 0xc0010002
creating build/bdist.linux-x86_64
creating build/bdist.linux-x86_64/egg
creating build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/mock_sparse
copying build/lib.linux-x86_64-3.7/dgl/mock_sparse/sp_matrix.py -> build/bdist.linux-x86_64/egg/dgl/mock_sparse
copying build/lib.linux-x86_64-3.7/dgl/mock_sparse/diag_matrix.py -> build/bdist.linux-x86_64/egg/dgl/mock_sparse
copying build/lib.linux-x86_64-3.7/dgl/mock_sparse/reduction.py -> build/bdist.linux-x86_64/egg/dgl/mock_sparse
copying build/lib.linux-x86_64-3.7/dgl/mock_sparse/__init__.py -> build/bdist.linux-x86_64/egg/dgl/mock_sparse
copying build/lib.linux-x86_64-3.7/dgl/mock_sparse/elementwise_op_sp.py -> build/bdist.linux-x86_64/egg/dgl/mock_sparse
creating build/bdist.linux-x86_64/egg/dgl/geometry
copying build/lib.linux-x86_64-3.7/dgl/geometry/edge_coarsening.py -> build/bdist.linux-x86_64/egg/dgl/geometry
copying build/lib.linux-x86_64-3.7/dgl/geometry/fps.py -> build/bdist.linux-x86_64/egg/dgl/geometry
copying build/lib.linux-x86_64-3.7/dgl/geometry/capi.py -> build/bdist.linux-x86_64/egg/dgl/geometry
copying build/lib.linux-x86_64-3.7/dgl/geometry/__init__.py -> build/bdist.linux-x86_64/egg/dgl/geometry
creating build/bdist.linux-x86_64/egg/dgl/function
copying build/lib.linux-x86_64-3.7/dgl/function/base.py -> build/bdist.linux-x86_64/egg/dgl/function
copying build/lib.linux-x86_64-3.7/dgl/function/message.py -> build/bdist.linux-x86_64/egg/dgl/function
copying build/lib.linux-x86_64-3.7/dgl/function/reducer.py -> build/bdist.linux-x86_64/egg/dgl/function
copying build/lib.linux-x86_64-3.7/dgl/function/__init__.py -> build/bdist.linux-x86_64/egg/dgl/function
copying build/lib.linux-x86_64-3.7/dgl/partition.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/nn
creating build/bdist.linux-x86_64/egg/dgl/nn/pytorch
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/glob.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch
creating build/bdist.linux-x86_64/egg/dgl/nn/pytorch/link
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/link/transe.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/link
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/link/edgepred.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/link
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/link/transr.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/link
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/link/__init__.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/link
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/hetero.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/sparse_emb.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/linear.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/utils.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch
creating build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/agnnconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/relgraphconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/appnpconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/dgnconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/grouprevres.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/twirlsconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/egatconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/graphconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/densechebconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/gcn2conv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/ginconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/dotgatconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/cfconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/gmmconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/sageconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/densesageconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/gineconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/edgeconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/nnconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/__init__.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/gatv2conv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/hgtconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/pnaconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/gatedgraphconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/atomicconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/tagconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/densegraphconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/egnnconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/chebconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/sgconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/gatconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/__init__.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch
creating build/bdist.linux-x86_64/egg/dgl/nn/pytorch/explain
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/explain/gnnexplainer.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/explain
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/explain/__init__.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/explain
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/factory.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/softmax.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch
creating build/bdist.linux-x86_64/egg/dgl/nn/mxnet
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/glob.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/hetero.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/utils.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet
creating build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/agnnconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/relgraphconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/appnpconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/graphconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/densechebconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/ginconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/gmmconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/sageconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/densesageconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/edgeconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/nnconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/__init__.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/gatedgraphconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/tagconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/densegraphconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/chebconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/sgconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/gatconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/__init__.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/softmax.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet
copying build/lib.linux-x86_64-3.7/dgl/nn/__init__.py -> build/bdist.linux-x86_64/egg/dgl/nn
creating build/bdist.linux-x86_64/egg/dgl/nn/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/glob.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/hetero.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/utils.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow
creating build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv/relgraphconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv/appnpconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv/graphconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv/densechebconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv/ginconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv/sageconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv/edgeconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv/__init__.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv/chebconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv/sgconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv/gatconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/__init__.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/softmax.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow
creating build/bdist.linux-x86_64/egg/dgl/nn/functional
copying build/lib.linux-x86_64-3.7/dgl/nn/functional/__init__.py -> build/bdist.linux-x86_64/egg/dgl/nn/functional
copying build/lib.linux-x86_64-3.7/dgl/core.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-3.7/dgl/subgraph.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/contrib
copying build/lib.linux-x86_64-3.7/dgl/contrib/dis_kvstore.py -> build/bdist.linux-x86_64/egg/dgl/contrib
copying build/lib.linux-x86_64-3.7/dgl/contrib/unified_tensor.py -> build/bdist.linux-x86_64/egg/dgl/contrib
copying build/lib.linux-x86_64-3.7/dgl/contrib/graph_store.py -> build/bdist.linux-x86_64/egg/dgl/contrib
creating build/bdist.linux-x86_64/egg/dgl/contrib/sampling
copying build/lib.linux-x86_64-3.7/dgl/contrib/sampling/sampler.py -> build/bdist.linux-x86_64/egg/dgl/contrib/sampling
copying build/lib.linux-x86_64-3.7/dgl/contrib/sampling/dis_sampler.py -> build/bdist.linux-x86_64/egg/dgl/contrib/sampling
copying build/lib.linux-x86_64-3.7/dgl/contrib/sampling/__init__.py -> build/bdist.linux-x86_64/egg/dgl/contrib/sampling
creating build/bdist.linux-x86_64/egg/dgl/contrib/data
copying build/lib.linux-x86_64-3.7/dgl/contrib/data/knowledge_graph.py -> build/bdist.linux-x86_64/egg/dgl/contrib/data
copying build/lib.linux-x86_64-3.7/dgl/contrib/data/__init__.py -> build/bdist.linux-x86_64/egg/dgl/contrib/data
copying build/lib.linux-x86_64-3.7/dgl/contrib/__init__.py -> build/bdist.linux-x86_64/egg/dgl/contrib
copying build/lib.linux-x86_64-3.7/dgl/traversal.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/utils
copying build/lib.linux-x86_64-3.7/dgl/utils/checks.py -> build/bdist.linux-x86_64/egg/dgl/utils
copying build/lib.linux-x86_64-3.7/dgl/utils/exception.py -> build/bdist.linux-x86_64/egg/dgl/utils
copying build/lib.linux-x86_64-3.7/dgl/utils/pin_memory.py -> build/bdist.linux-x86_64/egg/dgl/utils
copying build/lib.linux-x86_64-3.7/dgl/utils/internal.py -> build/bdist.linux-x86_64/egg/dgl/utils
copying build/lib.linux-x86_64-3.7/dgl/utils/filter.py -> build/bdist.linux-x86_64/egg/dgl/utils
copying build/lib.linux-x86_64-3.7/dgl/utils/data.py -> build/bdist.linux-x86_64/egg/dgl/utils
copying build/lib.linux-x86_64-3.7/dgl/utils/__init__.py -> build/bdist.linux-x86_64/egg/dgl/utils
copying build/lib.linux-x86_64-3.7/dgl/utils/shared_mem.py -> build/bdist.linux-x86_64/egg/dgl/utils
creating build/bdist.linux-x86_64/egg/dgl/multiprocessing
copying build/lib.linux-x86_64-3.7/dgl/multiprocessing/pytorch.py -> build/bdist.linux-x86_64/egg/dgl/multiprocessing
copying build/lib.linux-x86_64-3.7/dgl/multiprocessing/__init__.py -> build/bdist.linux-x86_64/egg/dgl/multiprocessing
copying build/lib.linux-x86_64-3.7/dgl/base.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/optim
creating build/bdist.linux-x86_64/egg/dgl/optim/pytorch
copying build/lib.linux-x86_64-3.7/dgl/optim/pytorch/__init__.py -> build/bdist.linux-x86_64/egg/dgl/optim/pytorch
copying build/lib.linux-x86_64-3.7/dgl/optim/pytorch/sparse_optim.py -> build/bdist.linux-x86_64/egg/dgl/optim/pytorch
creating build/bdist.linux-x86_64/egg/dgl/optim/mxnet
copying build/lib.linux-x86_64-3.7/dgl/optim/mxnet/__init__.py -> build/bdist.linux-x86_64/egg/dgl/optim/mxnet
copying build/lib.linux-x86_64-3.7/dgl/optim/__init__.py -> build/bdist.linux-x86_64/egg/dgl/optim
creating build/bdist.linux-x86_64/egg/dgl/optim/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/optim/tensorflow/__init__.py -> build/bdist.linux-x86_64/egg/dgl/optim/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/heterograph_index.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-3.7/dgl/sparse.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-3.7/dgl/global_config.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/dataloading
copying build/lib.linux-x86_64-3.7/dgl/dataloading/base.py -> build/bdist.linux-x86_64/egg/dgl/dataloading
copying build/lib.linux-x86_64-3.7/dgl/dataloading/dataloader.py -> build/bdist.linux-x86_64/egg/dgl/dataloading
copying build/lib.linux-x86_64-3.7/dgl/dataloading/graphsaint.py -> build/bdist.linux-x86_64/egg/dgl/dataloading
copying build/lib.linux-x86_64-3.7/dgl/dataloading/shadow.py -> build/bdist.linux-x86_64/egg/dgl/dataloading
copying build/lib.linux-x86_64-3.7/dgl/dataloading/negative_sampler.py -> build/bdist.linux-x86_64/egg/dgl/dataloading
copying build/lib.linux-x86_64-3.7/dgl/dataloading/__init__.py -> build/bdist.linux-x86_64/egg/dgl/dataloading
copying build/lib.linux-x86_64-3.7/dgl/dataloading/cluster_gcn.py -> build/bdist.linux-x86_64/egg/dgl/dataloading
copying build/lib.linux-x86_64-3.7/dgl/dataloading/neighbor_sampler.py -> build/bdist.linux-x86_64/egg/dgl/dataloading
copying build/lib.linux-x86_64-3.7/dgl/dataloading/dist_dataloader.py -> build/bdist.linux-x86_64/egg/dgl/dataloading
creating build/bdist.linux-x86_64/egg/dgl/_dataloading
copying build/lib.linux-x86_64-3.7/dgl/_dataloading/neighbor.py -> build/bdist.linux-x86_64/egg/dgl/_dataloading
copying build/lib.linux-x86_64-3.7/dgl/_dataloading/dataloader.py -> build/bdist.linux-x86_64/egg/dgl/_dataloading
creating build/bdist.linux-x86_64/egg/dgl/_dataloading/pytorch
copying build/lib.linux-x86_64-3.7/dgl/_dataloading/pytorch/dataloader.py -> build/bdist.linux-x86_64/egg/dgl/_dataloading/pytorch
copying build/lib.linux-x86_64-3.7/dgl/_dataloading/pytorch/__init__.py -> build/bdist.linux-x86_64/egg/dgl/_dataloading/pytorch
copying build/lib.linux-x86_64-3.7/dgl/_dataloading/shadow.py -> build/bdist.linux-x86_64/egg/dgl/_dataloading
copying build/lib.linux-x86_64-3.7/dgl/_dataloading/negative_sampler.py -> build/bdist.linux-x86_64/egg/dgl/_dataloading
copying build/lib.linux-x86_64-3.7/dgl/_dataloading/__init__.py -> build/bdist.linux-x86_64/egg/dgl/_dataloading
copying build/lib.linux-x86_64-3.7/dgl/_dataloading/cluster_gcn.py -> build/bdist.linux-x86_64/egg/dgl/_dataloading
copying build/lib.linux-x86_64-3.7/dgl/_api_internal.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-3.7/dgl/heterograph.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/backend
copying build/lib.linux-x86_64-3.7/dgl/backend/backend.py -> build/bdist.linux-x86_64/egg/dgl/backend
creating build/bdist.linux-x86_64/egg/dgl/backend/pytorch
copying build/lib.linux-x86_64-3.7/dgl/backend/pytorch/sparse.py -> build/bdist.linux-x86_64/egg/dgl/backend/pytorch
copying build/lib.linux-x86_64-3.7/dgl/backend/pytorch/tensor.py -> build/bdist.linux-x86_64/egg/dgl/backend/pytorch
copying build/lib.linux-x86_64-3.7/dgl/backend/pytorch/__init__.py -> build/bdist.linux-x86_64/egg/dgl/backend/pytorch
creating build/bdist.linux-x86_64/egg/dgl/backend/mxnet
copying build/lib.linux-x86_64-3.7/dgl/backend/mxnet/sparse.py -> build/bdist.linux-x86_64/egg/dgl/backend/mxnet
copying build/lib.linux-x86_64-3.7/dgl/backend/mxnet/tensor.py -> build/bdist.linux-x86_64/egg/dgl/backend/mxnet
copying build/lib.linux-x86_64-3.7/dgl/backend/mxnet/__init__.py -> build/bdist.linux-x86_64/egg/dgl/backend/mxnet
copying build/lib.linux-x86_64-3.7/dgl/backend/mxnet/sparse_optim.py -> build/bdist.linux-x86_64/egg/dgl/backend/mxnet
copying build/lib.linux-x86_64-3.7/dgl/backend/__init__.py -> build/bdist.linux-x86_64/egg/dgl/backend
creating build/bdist.linux-x86_64/egg/dgl/backend/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/backend/tensorflow/sparse.py -> build/bdist.linux-x86_64/egg/dgl/backend/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/backend/tensorflow/tensor.py -> build/bdist.linux-x86_64/egg/dgl/backend/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/backend/tensorflow/__init__.py -> build/bdist.linux-x86_64/egg/dgl/backend/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/backend/tensorflow/sparse_optim.py -> build/bdist.linux-x86_64/egg/dgl/backend/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/backend/set_default_backend.py -> build/bdist.linux-x86_64/egg/dgl/backend
copying build/lib.linux-x86_64-3.7/dgl/network.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/ops
copying build/lib.linux-x86_64-3.7/dgl/ops/edge_softmax.py -> build/bdist.linux-x86_64/egg/dgl/ops
copying build/lib.linux-x86_64-3.7/dgl/ops/sddmm.py -> build/bdist.linux-x86_64/egg/dgl/ops
copying build/lib.linux-x86_64-3.7/dgl/ops/gather_mm.py -> build/bdist.linux-x86_64/egg/dgl/ops
copying build/lib.linux-x86_64-3.7/dgl/ops/segment.py -> build/bdist.linux-x86_64/egg/dgl/ops
copying build/lib.linux-x86_64-3.7/dgl/ops/__init__.py -> build/bdist.linux-x86_64/egg/dgl/ops
copying build/lib.linux-x86_64-3.7/dgl/ops/spmm.py -> build/bdist.linux-x86_64/egg/dgl/ops
copying build/lib.linux-x86_64-3.7/dgl/logging.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-3.7/dgl/graph_index.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/cuda
copying build/lib.linux-x86_64-3.7/dgl/cuda/nccl.py -> build/bdist.linux-x86_64/egg/dgl/cuda
copying build/lib.linux-x86_64-3.7/dgl/cuda/__init__.py -> build/bdist.linux-x86_64/egg/dgl/cuda
copying build/lib.linux-x86_64-3.7/dgl/udf.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/storages
copying build/lib.linux-x86_64-3.7/dgl/storages/pytorch_tensor.py -> build/bdist.linux-x86_64/egg/dgl/storages
copying build/lib.linux-x86_64-3.7/dgl/storages/base.py -> build/bdist.linux-x86_64/egg/dgl/storages
copying build/lib.linux-x86_64-3.7/dgl/storages/tensor.py -> build/bdist.linux-x86_64/egg/dgl/storages
copying build/lib.linux-x86_64-3.7/dgl/storages/__init__.py -> build/bdist.linux-x86_64/egg/dgl/storages
copying build/lib.linux-x86_64-3.7/dgl/storages/numpy.py -> build/bdist.linux-x86_64/egg/dgl/storages
copying build/lib.linux-x86_64-3.7/dgl/ndarray.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-3.7/dgl/init.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-3.7/dgl/view.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/sampling
copying build/lib.linux-x86_64-3.7/dgl/sampling/neighbor.py -> build/bdist.linux-x86_64/egg/dgl/sampling
copying build/lib.linux-x86_64-3.7/dgl/sampling/pinsage.py -> build/bdist.linux-x86_64/egg/dgl/sampling
copying build/lib.linux-x86_64-3.7/dgl/sampling/negative.py -> build/bdist.linux-x86_64/egg/dgl/sampling
copying build/lib.linux-x86_64-3.7/dgl/sampling/randomwalks.py -> build/bdist.linux-x86_64/egg/dgl/sampling
copying build/lib.linux-x86_64-3.7/dgl/sampling/utils.py -> build/bdist.linux-x86_64/egg/dgl/sampling
copying build/lib.linux-x86_64-3.7/dgl/sampling/node2vec_randomwalk.py -> build/bdist.linux-x86_64/egg/dgl/sampling
copying build/lib.linux-x86_64-3.7/dgl/sampling/__init__.py -> build/bdist.linux-x86_64/egg/dgl/sampling
copying build/lib.linux-x86_64-3.7/dgl/generators.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/qm7b.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/graph_serialize.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/dgl_dataset.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/csv_dataset_base.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/knowledge_graph.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/qm9_edge.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/gnn_benchmark.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/qm9.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/icews18.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/gdelt.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/utils.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/sbm.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/csv_dataset.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/karate.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/citation_graph.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/rdf.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/bitcoinotc.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/ppi.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/tensor_serialize.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/flickr.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/wikics.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/heterograph_serialize.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/tu.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/__init__.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/gindt.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/tree.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/synthetic.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/adapter.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/fakenews.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/fraud.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/minigc.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/yelp.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/reddit.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/convert.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-3.7/dgl/container.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-3.7/dgl/merge.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-3.7/dgl/readout.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-3.7/dgl/__init__.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/dist_graph.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/graph_partition_book.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/partition.py -> build/bdist.linux-x86_64/egg/dgl/distributed
creating build/bdist.linux-x86_64/egg/dgl/distributed/nn
creating build/bdist.linux-x86_64/egg/dgl/distributed/nn/pytorch
copying build/lib.linux-x86_64-3.7/dgl/distributed/nn/pytorch/sparse_emb.py -> build/bdist.linux-x86_64/egg/dgl/distributed/nn/pytorch
copying build/lib.linux-x86_64-3.7/dgl/distributed/nn/pytorch/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distributed/nn/pytorch
creating build/bdist.linux-x86_64/egg/dgl/distributed/nn/mxnet
copying build/lib.linux-x86_64-3.7/dgl/distributed/nn/mxnet/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distributed/nn/mxnet
copying build/lib.linux-x86_64-3.7/dgl/distributed/nn/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distributed/nn
creating build/bdist.linux-x86_64/egg/dgl/distributed/nn/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/distributed/nn/tensorflow/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distributed/nn/tensorflow
creating build/bdist.linux-x86_64/egg/dgl/distributed/optim
creating build/bdist.linux-x86_64/egg/dgl/distributed/optim/pytorch
copying build/lib.linux-x86_64-3.7/dgl/distributed/optim/pytorch/utils.py -> build/bdist.linux-x86_64/egg/dgl/distributed/optim/pytorch
copying build/lib.linux-x86_64-3.7/dgl/distributed/optim/pytorch/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distributed/optim/pytorch
copying build/lib.linux-x86_64-3.7/dgl/distributed/optim/pytorch/sparse_optim.py -> build/bdist.linux-x86_64/egg/dgl/distributed/optim/pytorch
creating build/bdist.linux-x86_64/egg/dgl/distributed/optim/mxnet
copying build/lib.linux-x86_64-3.7/dgl/distributed/optim/mxnet/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distributed/optim/mxnet
copying build/lib.linux-x86_64-3.7/dgl/distributed/optim/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distributed/optim
creating build/bdist.linux-x86_64/egg/dgl/distributed/optim/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/distributed/optim/tensorflow/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distributed/optim/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/distributed/standalone_kvstore.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/graph_services.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/role.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/constants.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/rpc_server.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/kvstore.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/dist_context.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/rpc.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/rpc_client.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/id_map.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/server_state.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/shared_mem_utils.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/dist_tensor.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/dist_dataloader.py -> build/bdist.linux-x86_64/egg/dgl/distributed
creating build/bdist.linux-x86_64/egg/dgl/_deprecate
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/nodeflow.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/udf.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/view.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/kernel.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/__init__.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate
creating build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/spmv.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/degree_bucketing.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/__init__.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/scheduler.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/adapter.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/runtime.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime
creating build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/ir
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/ir/program.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/ir
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/ir/executor.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/ir
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/ir/var.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/ir
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/ir/registry.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/ir
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/ir/__init__.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/ir
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/graph.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/frame.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate
copying build/lib.linux-x86_64-3.7/dgl/propagate.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-3.7/dgl/batch.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/transforms
copying build/lib.linux-x86_64-3.7/dgl/transforms/functional.py -> build/bdist.linux-x86_64/egg/dgl/transforms
copying build/lib.linux-x86_64-3.7/dgl/transforms/__init__.py -> build/bdist.linux-x86_64/egg/dgl/transforms
copying build/lib.linux-x86_64-3.7/dgl/transforms/module.py -> build/bdist.linux-x86_64/egg/dgl/transforms
creating build/bdist.linux-x86_64/egg/dgl/distgnn
creating build/bdist.linux-x86_64/egg/dgl/distgnn/partition
copying build/lib.linux-x86_64-3.7/dgl/distgnn/partition/libra_partition.py -> build/bdist.linux-x86_64/egg/dgl/distgnn/partition
copying build/lib.linux-x86_64-3.7/dgl/distgnn/partition/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distgnn/partition
creating build/bdist.linux-x86_64/egg/dgl/distgnn/tools
copying build/lib.linux-x86_64-3.7/dgl/distgnn/tools/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distgnn/tools
copying build/lib.linux-x86_64-3.7/dgl/distgnn/tools/tools.py -> build/bdist.linux-x86_64/egg/dgl/distgnn/tools
copying build/lib.linux-x86_64-3.7/dgl/distgnn/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distgnn
creating build/bdist.linux-x86_64/egg/dgl/_ffi
copying build/lib.linux-x86_64-3.7/dgl/_ffi/runtime_ctypes.py -> build/bdist.linux-x86_64/egg/dgl/_ffi
copying build/lib.linux-x86_64-3.7/dgl/_ffi/function.py -> build/bdist.linux-x86_64/egg/dgl/_ffi
copying build/lib.linux-x86_64-3.7/dgl/_ffi/libinfo.py -> build/bdist.linux-x86_64/egg/dgl/_ffi
copying build/lib.linux-x86_64-3.7/dgl/_ffi/streams.py -> build/bdist.linux-x86_64/egg/dgl/_ffi
copying build/lib.linux-x86_64-3.7/dgl/_ffi/base.py -> build/bdist.linux-x86_64/egg/dgl/_ffi
creating build/bdist.linux-x86_64/egg/dgl/_ffi/_ctypes
copying build/lib.linux-x86_64-3.7/dgl/_ffi/_ctypes/function.py -> build/bdist.linux-x86_64/egg/dgl/_ffi/_ctypes
copying build/lib.linux-x86_64-3.7/dgl/_ffi/_ctypes/types.py -> build/bdist.linux-x86_64/egg/dgl/_ffi/_ctypes
copying build/lib.linux-x86_64-3.7/dgl/_ffi/_ctypes/ndarray.py -> build/bdist.linux-x86_64/egg/dgl/_ffi/_ctypes
copying build/lib.linux-x86_64-3.7/dgl/_ffi/_ctypes/__init__.py -> build/bdist.linux-x86_64/egg/dgl/_ffi/_ctypes
copying build/lib.linux-x86_64-3.7/dgl/_ffi/_ctypes/object.py -> build/bdist.linux-x86_64/egg/dgl/_ffi/_ctypes
copying build/lib.linux-x86_64-3.7/dgl/_ffi/ndarray.py -> build/bdist.linux-x86_64/egg/dgl/_ffi
copying build/lib.linux-x86_64-3.7/dgl/_ffi/object_generic.py -> build/bdist.linux-x86_64/egg/dgl/_ffi
copying build/lib.linux-x86_64-3.7/dgl/_ffi/__init__.py -> build/bdist.linux-x86_64/egg/dgl/_ffi
copying build/lib.linux-x86_64-3.7/dgl/_ffi/object.py -> build/bdist.linux-x86_64/egg/dgl/_ffi
creating build/bdist.linux-x86_64/egg/dgl/_ffi/_cy3
copying build/lib.linux-x86_64-3.7/dgl/_ffi/_cy3/__init__.py -> build/bdist.linux-x86_64/egg/dgl/_ffi/_cy3
copying build/lib.linux-x86_64-3.7/dgl/_ffi/_cy3/core.cpython-37m-x86_64-linux-gnu.so -> build/bdist.linux-x86_64/egg/dgl/_ffi/_cy3
creating build/bdist.linux-x86_64/egg/dgl/_ffi/_cy2
copying build/lib.linux-x86_64-3.7/dgl/_ffi/_cy2/__init__.py -> build/bdist.linux-x86_64/egg/dgl/_ffi/_cy2
creating build/bdist.linux-x86_64/egg/dgl/_ffi/_cython
copying build/lib.linux-x86_64-3.7/dgl/_ffi/_cython/core.cpp -> build/bdist.linux-x86_64/egg/dgl/_ffi/_cython
copying build/lib.linux-x86_64-3.7/dgl/frame.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-3.7/dgl/random.py -> build/bdist.linux-x86_64/egg/dgl
byte-compiling build/bdist.linux-x86_64/egg/dgl/mock_sparse/sp_matrix.py to sp_matrix.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/mock_sparse/diag_matrix.py to diag_matrix.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/mock_sparse/reduction.py to reduction.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/mock_sparse/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/mock_sparse/elementwise_op_sp.py to elementwise_op_sp.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/geometry/edge_coarsening.py to edge_coarsening.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/geometry/fps.py to fps.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/geometry/capi.py to capi.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/geometry/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/function/base.py to base.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/function/message.py to message.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/function/reducer.py to reducer.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/function/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/partition.py to partition.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/glob.py to glob.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/link/transe.py to transe.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/link/edgepred.py to edgepred.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/link/transr.py to transr.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/link/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/hetero.py to hetero.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/sparse_emb.py to sparse_emb.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/linear.py to linear.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/utils.py to utils.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/agnnconv.py to agnnconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/relgraphconv.py to relgraphconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/appnpconv.py to appnpconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/dgnconv.py to dgnconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/grouprevres.py to grouprevres.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/twirlsconv.py to twirlsconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/egatconv.py to egatconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/graphconv.py to graphconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/densechebconv.py to densechebconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/gcn2conv.py to gcn2conv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/ginconv.py to ginconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/dotgatconv.py to dotgatconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/cfconv.py to cfconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/gmmconv.py to gmmconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/sageconv.py to sageconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/densesageconv.py to densesageconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/gineconv.py to gineconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/edgeconv.py to edgeconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/nnconv.py to nnconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/gatv2conv.py to gatv2conv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/hgtconv.py to hgtconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/pnaconv.py to pnaconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/gatedgraphconv.py to gatedgraphconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/atomicconv.py to atomicconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/tagconv.py to tagconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/densegraphconv.py to densegraphconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/egnnconv.py to egnnconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/chebconv.py to chebconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/sgconv.py to sgconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/gatconv.py to gatconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/explain/gnnexplainer.py to gnnexplainer.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/explain/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/factory.py to factory.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/softmax.py to softmax.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/glob.py to glob.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/hetero.py to hetero.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/utils.py to utils.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/agnnconv.py to agnnconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/relgraphconv.py to relgraphconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/appnpconv.py to appnpconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/graphconv.py to graphconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/densechebconv.py to densechebconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/ginconv.py to ginconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/gmmconv.py to gmmconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/sageconv.py to sageconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/densesageconv.py to densesageconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/edgeconv.py to edgeconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/nnconv.py to nnconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/gatedgraphconv.py to gatedgraphconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/tagconv.py to tagconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/densegraphconv.py to densegraphconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/chebconv.py to chebconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/sgconv.py to sgconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/gatconv.py to gatconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/softmax.py to softmax.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/glob.py to glob.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/hetero.py to hetero.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/utils.py to utils.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv/relgraphconv.py to relgraphconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv/appnpconv.py to appnpconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv/graphconv.py to graphconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv/densechebconv.py to densechebconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv/ginconv.py to ginconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv/sageconv.py to sageconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv/edgeconv.py to edgeconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv/chebconv.py to chebconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv/sgconv.py to sgconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv/gatconv.py to gatconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/softmax.py to softmax.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/functional/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/core.py to core.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/subgraph.py to subgraph.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/contrib/dis_kvstore.py to dis_kvstore.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/contrib/unified_tensor.py to unified_tensor.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/contrib/graph_store.py to graph_store.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/contrib/sampling/sampler.py to sampler.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/contrib/sampling/dis_sampler.py to dis_sampler.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/contrib/sampling/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/contrib/data/knowledge_graph.py to knowledge_graph.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/contrib/data/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/contrib/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/traversal.py to traversal.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/utils/checks.py to checks.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/utils/exception.py to exception.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/utils/pin_memory.py to pin_memory.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/utils/internal.py to internal.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/utils/filter.py to filter.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/utils/data.py to data.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/utils/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/utils/shared_mem.py to shared_mem.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/multiprocessing/pytorch.py to pytorch.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/multiprocessing/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/base.py to base.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/optim/pytorch/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/optim/pytorch/sparse_optim.py to sparse_optim.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/optim/mxnet/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/optim/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/optim/tensorflow/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/heterograph_index.py to heterograph_index.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/sparse.py to sparse.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/global_config.py to global_config.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/dataloading/base.py to base.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/dataloading/dataloader.py to dataloader.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/dataloading/graphsaint.py to graphsaint.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/dataloading/shadow.py to shadow.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/dataloading/negative_sampler.py to negative_sampler.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/dataloading/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/dataloading/cluster_gcn.py to cluster_gcn.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/dataloading/neighbor_sampler.py to neighbor_sampler.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/dataloading/dist_dataloader.py to dist_dataloader.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_dataloading/neighbor.py to neighbor.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_dataloading/dataloader.py to dataloader.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_dataloading/pytorch/dataloader.py to dataloader.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_dataloading/pytorch/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_dataloading/shadow.py to shadow.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_dataloading/negative_sampler.py to negative_sampler.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_dataloading/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_dataloading/cluster_gcn.py to cluster_gcn.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_api_internal.py to _api_internal.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/heterograph.py to heterograph.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/backend.py to backend.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/pytorch/sparse.py to sparse.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/pytorch/tensor.py to tensor.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/pytorch/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/mxnet/sparse.py to sparse.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/mxnet/tensor.py to tensor.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/mxnet/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/mxnet/sparse_optim.py to sparse_optim.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/tensorflow/sparse.py to sparse.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/tensorflow/tensor.py to tensor.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/tensorflow/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/tensorflow/sparse_optim.py to sparse_optim.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/set_default_backend.py to set_default_backend.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/network.py to network.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/ops/edge_softmax.py to edge_softmax.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/ops/sddmm.py to sddmm.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/ops/gather_mm.py to gather_mm.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/ops/segment.py to segment.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/ops/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/ops/spmm.py to spmm.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/logging.py to logging.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/graph_index.py to graph_index.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/cuda/nccl.py to nccl.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/cuda/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/udf.py to udf.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/storages/pytorch_tensor.py to pytorch_tensor.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/storages/base.py to base.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/storages/tensor.py to tensor.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/storages/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/storages/numpy.py to numpy.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/ndarray.py to ndarray.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/init.py to init.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/view.py to view.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/sampling/neighbor.py to neighbor.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/sampling/pinsage.py to pinsage.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/sampling/negative.py to negative.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/sampling/randomwalks.py to randomwalks.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/sampling/utils.py to utils.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/sampling/node2vec_randomwalk.py to node2vec_randomwalk.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/sampling/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/generators.py to generators.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/qm7b.py to qm7b.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/graph_serialize.py to graph_serialize.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/dgl_dataset.py to dgl_dataset.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/csv_dataset_base.py to csv_dataset_base.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/knowledge_graph.py to knowledge_graph.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/qm9_edge.py to qm9_edge.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/gnn_benchmark.py to gnn_benchmark.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/qm9.py to qm9.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/icews18.py to icews18.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/gdelt.py to gdelt.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/utils.py to utils.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/sbm.py to sbm.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/csv_dataset.py to csv_dataset.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/karate.py to karate.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/citation_graph.py to citation_graph.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/rdf.py to rdf.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/bitcoinotc.py to bitcoinotc.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/ppi.py to ppi.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/tensor_serialize.py to tensor_serialize.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/flickr.py to flickr.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/wikics.py to wikics.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/heterograph_serialize.py to heterograph_serialize.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/tu.py to tu.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/gindt.py to gindt.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/tree.py to tree.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/synthetic.py to synthetic.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/adapter.py to adapter.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/fakenews.py to fakenews.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/fraud.py to fraud.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/minigc.py to minigc.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/yelp.py to yelp.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/reddit.py to reddit.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/convert.py to convert.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/container.py to container.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/merge.py to merge.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/readout.py to readout.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/dist_graph.py to dist_graph.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/graph_partition_book.py to graph_partition_book.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/partition.py to partition.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/nn/pytorch/sparse_emb.py to sparse_emb.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/nn/pytorch/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/nn/mxnet/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/nn/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/nn/tensorflow/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/optim/pytorch/utils.py to utils.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/optim/pytorch/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/optim/pytorch/sparse_optim.py to sparse_optim.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/optim/mxnet/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/optim/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/optim/tensorflow/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/standalone_kvstore.py to standalone_kvstore.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/graph_services.py to graph_services.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/role.py to role.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/constants.py to constants.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/rpc_server.py to rpc_server.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/kvstore.py to kvstore.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/dist_context.py to dist_context.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/rpc.py to rpc.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/rpc_client.py to rpc_client.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/id_map.py to id_map.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/server_state.py to server_state.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/shared_mem_utils.py to shared_mem_utils.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/dist_tensor.py to dist_tensor.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/dist_dataloader.py to dist_dataloader.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/nodeflow.py to nodeflow.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/udf.py to udf.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/view.py to view.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/kernel.py to kernel.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/spmv.py to spmv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/degree_bucketing.py to degree_bucketing.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/scheduler.py to scheduler.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/adapter.py to adapter.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/runtime.py to runtime.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/ir/program.py to program.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/ir/executor.py to executor.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/ir/var.py to var.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/ir/registry.py to registry.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/ir/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/graph.py to graph.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/frame.py to frame.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/propagate.py to propagate.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/batch.py to batch.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/transforms/functional.py to functional.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/transforms/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/transforms/module.py to module.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distgnn/partition/libra_partition.py to libra_partition.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distgnn/partition/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distgnn/tools/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distgnn/tools/tools.py to tools.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distgnn/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/runtime_ctypes.py to runtime_ctypes.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/function.py to function.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/libinfo.py to libinfo.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/streams.py to streams.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/base.py to base.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/_ctypes/function.py to function.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/_ctypes/types.py to types.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/_ctypes/ndarray.py to ndarray.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/_ctypes/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/_ctypes/object.py to object.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/ndarray.py to ndarray.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/object_generic.py to object_generic.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/object.py to object.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/_cy3/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/_cy2/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/frame.py to frame.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/random.py to random.cpython-37.pyc
creating stub loader for dgl/_ffi/_cy3/core.cpython-37m-x86_64-linux-gnu.so
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/_cy3/core.py to core.cpython-37.pyc
installing package data to build/bdist.linux-x86_64/egg
running install_data
copying ../build/libdgl.so -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/tensoradapter
creating build/bdist.linux-x86_64/egg/dgl/tensoradapter/pytorch
copying ../build/tensoradapter/pytorch/libtensoradapter_pytorch_1.9.0.so -> build/bdist.linux-x86_64/egg/dgl/tensoradapter/pytorch
creating build/bdist.linux-x86_64/egg/EGG-INFO
copying dgl.egg-info/PKG-INFO -> build/bdist.linux-x86_64/egg/EGG-INFO
copying dgl.egg-info/SOURCES.txt -> build/bdist.linux-x86_64/egg/EGG-INFO
copying dgl.egg-info/dependency_links.txt -> build/bdist.linux-x86_64/egg/EGG-INFO
copying dgl.egg-info/not-zip-safe -> build/bdist.linux-x86_64/egg/EGG-INFO
copying dgl.egg-info/requires.txt -> build/bdist.linux-x86_64/egg/EGG-INFO
copying dgl.egg-info/top_level.txt -> build/bdist.linux-x86_64/egg/EGG-INFO
writing build/bdist.linux-x86_64/egg/EGG-INFO/native_libs.txt
creating dist
creating 'dist/dgl-0.9-py3.7-linux-x86_64.egg' and adding 'build/bdist.linux-x86_64/egg' to it
removing 'build/bdist.linux-x86_64/egg' (and everything under it)
Processing dgl-0.9-py3.7-linux-x86_64.egg
creating /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/dgl-0.9-py3.7-linux-x86_64.egg
Extracting dgl-0.9-py3.7-linux-x86_64.egg to /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages
Adding dgl 0.9 to easy-install.pth file

Installed /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/dgl-0.9-py3.7-linux-x86_64.egg
Processing dependencies for dgl==0.9
Searching for psutil>=5.8.0
Reading https://pypi.org/simple/psutil/
Downloading https://files.pythonhosted.org/packages/3d/73/d8c87b5612c58d1e6c6d91997c1590771d34e4ee27d9c11eb1e64ecbf365/psutil-5.9.2-cp37-cp37m-manylinux_2_12_x86_64.manylinux2010_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl#sha256=4fb54941aac044a61db9d8eb56fc5bee207db3bc58645d657249030e15ba3727
Best match: psutil 5.9.2
Processing psutil-5.9.2-cp37-cp37m-manylinux_2_12_x86_64.manylinux2010_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl
Installing psutil-5.9.2-cp37-cp37m-manylinux_2_12_x86_64.manylinux2010_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl to /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages
Adding psutil 5.9.2 to easy-install.pth file

Installed /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/psutil-5.9.2-py3.7-linux-x86_64.egg
Searching for tqdm==4.64.0
Best match: tqdm 4.64.0
Adding tqdm 4.64.0 to easy-install.pth file
Installing tqdm script to /opt/conda/envs/pytorch-ci/bin

Using /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages
Searching for requests==2.28.1
Best match: requests 2.28.1
Adding requests 2.28.1 to easy-install.pth file

Using /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages
Searching for networkx==2.6.3
Best match: networkx 2.6.3
Adding networkx 2.6.3 to easy-install.pth file

Using /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages
Searching for scipy==1.7.3
Best match: scipy 1.7.3
Adding scipy 1.7.3 to easy-install.pth file

Using /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages
Searching for numpy==1.21.6
Best match: numpy 1.21.6
Adding numpy 1.21.6 to easy-install.pth file
Installing f2py script to /opt/conda/envs/pytorch-ci/bin
Installing f2py3 script to /opt/conda/envs/pytorch-ci/bin
Installing f2py3.7 script to /opt/conda/envs/pytorch-ci/bin

Using /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages
Searching for certifi==2022.6.15
Best match: certifi 2022.6.15
Adding certifi 2022.6.15 to easy-install.pth file

Using /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages
Searching for urllib3==1.26.11
Best match: urllib3 1.26.11
Adding urllib3 1.26.11 to easy-install.pth file

Using /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages
Searching for idna==3.3
Best match: idna 3.3
Adding idna 3.3 to easy-install.pth file

Using /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages
Searching for charset-normalizer==2.1.0
Best match: charset-normalizer 2.1.0
Adding charset-normalizer 2.1.0 to easy-install.pth file
Installing normalizer script to /opt/conda/envs/pytorch-ci/bin

Using /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages
Finished processing dependencies for dgl==0.9
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/Cython/Compiler/Main.py:369: FutureWarning: Cython directive 'language_level' not set, using 2 for now (Py2). This will change in a later release! File: /root/jenkins/workspace/dgl_PR-4648/python/dgl/_ffi/_cython/core.pyx
  tree = Parsing.p_module(s, pxd, full_module_name)
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/setuptools/command/install.py:37: SetuptoolsDeprecationWarning: setup.py install is deprecated. Use build and pip and other standards-based tools.
  setuptools.SetuptoolsDeprecationWarning,
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/setuptools/command/easy_install.py:147: EasyInstallDeprecationWarning: easy_install command is deprecated. Use build and pip and other standards-based tools.
  EasyInstallDeprecationWarning,
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/pkg_resources/__init__.py:125: PkgResourcesDeprecationWarning:  is an invalid version and will not be supported in a future release
  PkgResourcesDeprecationWarning,
[1/1] Cythonizing dgl/_ffi/_cython/core.pyx
running build_ext
building 'dgl._ffi._cy3.core' extension
gcc -pthread -B /opt/conda/envs/pytorch-ci/compiler_compat -Wl,--sysroot=/ -Wsign-compare -DNDEBUG -g -fwrapv -O3 -Wall -Wstrict-prototypes -fPIC -I../include/ -I../third_party/dmlc-core/include -I../third_party/dlpack/include -I/opt/conda/envs/pytorch-ci/include/python3.7m -c dgl/_ffi/_cython/core.cpp -o build/temp.linux-x86_64-3.7/dgl/_ffi/_cython/core.o
cc1plus: warning: command line option ‘-Wstrict-prototypes’ is valid for C/ObjC but not for C++
g++ -pthread -B /opt/conda/envs/pytorch-ci/compiler_compat -Wl,--sysroot=/ -pthread -shared -B /opt/conda/envs/pytorch-ci/compiler_compat -L/opt/conda/envs/pytorch-ci/lib -Wl,-rpath=/opt/conda/envs/pytorch-ci/lib -Wl,--no-as-needed -Wl,--sysroot=/ build/temp.linux-x86_64-3.7/dgl/_ffi/_cython/core.o -Ldgl -L../build/Release -L../build -ldgl -o build/lib.linux-x86_64-3.7/dgl/_ffi/_cy3/core.cpython-37m-x86_64-linux-gnu.so
/opt/conda/envs/pytorch-ci/compiler_compat/ld: warning: /opt/conda/envs/pytorch-ci/lib/libstdc++.so: unsupported GNU_PROPERTY_TYPE (5) type: 0xc0010001
/opt/conda/envs/pytorch-ci/compiler_compat/ld: warning: /opt/conda/envs/pytorch-ci/lib/libstdc++.so: unsupported GNU_PROPERTY_TYPE (5) type: 0xc0010002
/opt/conda/envs/pytorch-ci/compiler_compat/ld: warning: /opt/conda/envs/pytorch-ci/lib/libgcc_s.so.1: unsupported GNU_PROPERTY_TYPE (5) type: 0xc0010001
/opt/conda/envs/pytorch-ci/compiler_compat/ld: warning: /opt/conda/envs/pytorch-ci/lib/libgcc_s.so.1: unsupported GNU_PROPERTY_TYPE (5) type: 0xc0010002
/opt/conda/envs/pytorch-ci/compiler_compat/ld: warning: /opt/conda/envs/pytorch-ci/lib/libgcc_s.so.1: unsupported GNU_PROPERTY_TYPE (5) type: 0xc0010001
/opt/conda/envs/pytorch-ci/compiler_compat/ld: warning: /opt/conda/envs/pytorch-ci/lib/libgcc_s.so.1: unsupported GNU_PROPERTY_TYPE (5) type: 0xc0010002
copying build/lib.linux-x86_64-3.7/dgl/_ffi/_cy3/core.cpython-37m-x86_64-linux-gnu.so -> dgl/_ffi/_cy3
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/Cython/Compiler/Main.py:369: FutureWarning: Cython directive 'language_level' not set, using 2 for now (Py2). This will change in a later release! File: /root/jenkins/workspace/dgl_PR-4648/python/dgl/_ffi/_cython/core.pyx
  tree = Parsing.p_module(s, pxd, full_module_name)
WARNING: Skipping dgl as it is not installed.
WARNING: Running pip as the 'root' user can result in broken permissions and conflicting behaviour with the system package manager. It is recommended to use a virtual environment instead: https://pip.pypa.io/warnings/venv
[1/1] Cythonizing dgl/_ffi/_cython/core.pyx
running install
running bdist_egg
running egg_info
creating dgl.egg-info
writing dgl.egg-info/PKG-INFO
writing dependency_links to dgl.egg-info/dependency_links.txt
writing requirements to dgl.egg-info/requires.txt
writing top-level names to dgl.egg-info/top_level.txt
writing manifest file 'dgl.egg-info/SOURCES.txt'
reading manifest file 'dgl.egg-info/SOURCES.txt'
writing manifest file 'dgl.egg-info/SOURCES.txt'
installing library code to build/bdist.linux-x86_64/egg
running install_lib
running build_py
creating build
creating build/lib.linux-x86_64-3.7
creating build/lib.linux-x86_64-3.7/dgl
copying dgl/partition.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/core.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/subgraph.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/traversal.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/base.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/heterograph_index.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/sparse.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/global_config.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/_api_internal.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/heterograph.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/network.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/logging.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/graph_index.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/udf.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/ndarray.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/init.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/view.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/generators.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/convert.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/container.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/merge.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/readout.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/__init__.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/propagate.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/batch.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/frame.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/random.py -> build/lib.linux-x86_64-3.7/dgl
creating build/lib.linux-x86_64-3.7/dgl/mock_sparse
copying dgl/mock_sparse/sp_matrix.py -> build/lib.linux-x86_64-3.7/dgl/mock_sparse
copying dgl/mock_sparse/diag_matrix.py -> build/lib.linux-x86_64-3.7/dgl/mock_sparse
copying dgl/mock_sparse/reduction.py -> build/lib.linux-x86_64-3.7/dgl/mock_sparse
copying dgl/mock_sparse/__init__.py -> build/lib.linux-x86_64-3.7/dgl/mock_sparse
copying dgl/mock_sparse/elementwise_op_sp.py -> build/lib.linux-x86_64-3.7/dgl/mock_sparse
creating build/lib.linux-x86_64-3.7/dgl/geometry
copying dgl/geometry/edge_coarsening.py -> build/lib.linux-x86_64-3.7/dgl/geometry
copying dgl/geometry/fps.py -> build/lib.linux-x86_64-3.7/dgl/geometry
copying dgl/geometry/capi.py -> build/lib.linux-x86_64-3.7/dgl/geometry
copying dgl/geometry/__init__.py -> build/lib.linux-x86_64-3.7/dgl/geometry
creating build/lib.linux-x86_64-3.7/dgl/function
copying dgl/function/base.py -> build/lib.linux-x86_64-3.7/dgl/function
copying dgl/function/message.py -> build/lib.linux-x86_64-3.7/dgl/function
copying dgl/function/reducer.py -> build/lib.linux-x86_64-3.7/dgl/function
copying dgl/function/__init__.py -> build/lib.linux-x86_64-3.7/dgl/function
creating build/lib.linux-x86_64-3.7/dgl/nn
copying dgl/nn/__init__.py -> build/lib.linux-x86_64-3.7/dgl/nn
creating build/lib.linux-x86_64-3.7/dgl/contrib
copying dgl/contrib/dis_kvstore.py -> build/lib.linux-x86_64-3.7/dgl/contrib
copying dgl/contrib/unified_tensor.py -> build/lib.linux-x86_64-3.7/dgl/contrib
copying dgl/contrib/graph_store.py -> build/lib.linux-x86_64-3.7/dgl/contrib
copying dgl/contrib/__init__.py -> build/lib.linux-x86_64-3.7/dgl/contrib
creating build/lib.linux-x86_64-3.7/dgl/utils
copying dgl/utils/checks.py -> build/lib.linux-x86_64-3.7/dgl/utils
copying dgl/utils/exception.py -> build/lib.linux-x86_64-3.7/dgl/utils
copying dgl/utils/pin_memory.py -> build/lib.linux-x86_64-3.7/dgl/utils
copying dgl/utils/internal.py -> build/lib.linux-x86_64-3.7/dgl/utils
copying dgl/utils/filter.py -> build/lib.linux-x86_64-3.7/dgl/utils
copying dgl/utils/data.py -> build/lib.linux-x86_64-3.7/dgl/utils
copying dgl/utils/__init__.py -> build/lib.linux-x86_64-3.7/dgl/utils
copying dgl/utils/shared_mem.py -> build/lib.linux-x86_64-3.7/dgl/utils
creating build/lib.linux-x86_64-3.7/dgl/multiprocessing
copying dgl/multiprocessing/pytorch.py -> build/lib.linux-x86_64-3.7/dgl/multiprocessing
copying dgl/multiprocessing/__init__.py -> build/lib.linux-x86_64-3.7/dgl/multiprocessing
creating build/lib.linux-x86_64-3.7/dgl/optim
copying dgl/optim/__init__.py -> build/lib.linux-x86_64-3.7/dgl/optim
creating build/lib.linux-x86_64-3.7/dgl/dataloading
copying dgl/dataloading/base.py -> build/lib.linux-x86_64-3.7/dgl/dataloading
copying dgl/dataloading/dataloader.py -> build/lib.linux-x86_64-3.7/dgl/dataloading
copying dgl/dataloading/graphsaint.py -> build/lib.linux-x86_64-3.7/dgl/dataloading
copying dgl/dataloading/shadow.py -> build/lib.linux-x86_64-3.7/dgl/dataloading
copying dgl/dataloading/negative_sampler.py -> build/lib.linux-x86_64-3.7/dgl/dataloading
copying dgl/dataloading/__init__.py -> build/lib.linux-x86_64-3.7/dgl/dataloading
copying dgl/dataloading/cluster_gcn.py -> build/lib.linux-x86_64-3.7/dgl/dataloading
copying dgl/dataloading/neighbor_sampler.py -> build/lib.linux-x86_64-3.7/dgl/dataloading
copying dgl/dataloading/dist_dataloader.py -> build/lib.linux-x86_64-3.7/dgl/dataloading
creating build/lib.linux-x86_64-3.7/dgl/_dataloading
copying dgl/_dataloading/neighbor.py -> build/lib.linux-x86_64-3.7/dgl/_dataloading
copying dgl/_dataloading/dataloader.py -> build/lib.linux-x86_64-3.7/dgl/_dataloading
copying dgl/_dataloading/shadow.py -> build/lib.linux-x86_64-3.7/dgl/_dataloading
copying dgl/_dataloading/negative_sampler.py -> build/lib.linux-x86_64-3.7/dgl/_dataloading
copying dgl/_dataloading/__init__.py -> build/lib.linux-x86_64-3.7/dgl/_dataloading
copying dgl/_dataloading/cluster_gcn.py -> build/lib.linux-x86_64-3.7/dgl/_dataloading
creating build/lib.linux-x86_64-3.7/dgl/backend
copying dgl/backend/backend.py -> build/lib.linux-x86_64-3.7/dgl/backend
copying dgl/backend/__init__.py -> build/lib.linux-x86_64-3.7/dgl/backend
copying dgl/backend/set_default_backend.py -> build/lib.linux-x86_64-3.7/dgl/backend
creating build/lib.linux-x86_64-3.7/dgl/ops
copying dgl/ops/edge_softmax.py -> build/lib.linux-x86_64-3.7/dgl/ops
copying dgl/ops/sddmm.py -> build/lib.linux-x86_64-3.7/dgl/ops
copying dgl/ops/gather_mm.py -> build/lib.linux-x86_64-3.7/dgl/ops
copying dgl/ops/segment.py -> build/lib.linux-x86_64-3.7/dgl/ops
copying dgl/ops/__init__.py -> build/lib.linux-x86_64-3.7/dgl/ops
copying dgl/ops/spmm.py -> build/lib.linux-x86_64-3.7/dgl/ops
creating build/lib.linux-x86_64-3.7/dgl/cuda
copying dgl/cuda/nccl.py -> build/lib.linux-x86_64-3.7/dgl/cuda
copying dgl/cuda/__init__.py -> build/lib.linux-x86_64-3.7/dgl/cuda
creating build/lib.linux-x86_64-3.7/dgl/storages
copying dgl/storages/pytorch_tensor.py -> build/lib.linux-x86_64-3.7/dgl/storages
copying dgl/storages/base.py -> build/lib.linux-x86_64-3.7/dgl/storages
copying dgl/storages/tensor.py -> build/lib.linux-x86_64-3.7/dgl/storages
copying dgl/storages/__init__.py -> build/lib.linux-x86_64-3.7/dgl/storages
copying dgl/storages/numpy.py -> build/lib.linux-x86_64-3.7/dgl/storages
creating build/lib.linux-x86_64-3.7/dgl/sampling
copying dgl/sampling/neighbor.py -> build/lib.linux-x86_64-3.7/dgl/sampling
copying dgl/sampling/pinsage.py -> build/lib.linux-x86_64-3.7/dgl/sampling
copying dgl/sampling/negative.py -> build/lib.linux-x86_64-3.7/dgl/sampling
copying dgl/sampling/randomwalks.py -> build/lib.linux-x86_64-3.7/dgl/sampling
copying dgl/sampling/utils.py -> build/lib.linux-x86_64-3.7/dgl/sampling
copying dgl/sampling/node2vec_randomwalk.py -> build/lib.linux-x86_64-3.7/dgl/sampling
copying dgl/sampling/__init__.py -> build/lib.linux-x86_64-3.7/dgl/sampling
creating build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/qm7b.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/graph_serialize.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/dgl_dataset.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/csv_dataset_base.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/knowledge_graph.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/qm9_edge.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/gnn_benchmark.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/qm9.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/icews18.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/gdelt.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/utils.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/sbm.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/csv_dataset.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/karate.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/citation_graph.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/rdf.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/bitcoinotc.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/ppi.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/tensor_serialize.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/flickr.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/wikics.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/heterograph_serialize.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/tu.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/__init__.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/gindt.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/tree.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/synthetic.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/adapter.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/fakenews.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/fraud.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/minigc.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/yelp.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/reddit.py -> build/lib.linux-x86_64-3.7/dgl/data
creating build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/dist_graph.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/graph_partition_book.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/partition.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/standalone_kvstore.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/graph_services.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/role.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/constants.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/rpc_server.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/kvstore.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/dist_context.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/rpc.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/rpc_client.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/id_map.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/server_state.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/__init__.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/shared_mem_utils.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/dist_tensor.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/dist_dataloader.py -> build/lib.linux-x86_64-3.7/dgl/distributed
creating build/lib.linux-x86_64-3.7/dgl/_deprecate
copying dgl/_deprecate/nodeflow.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate
copying dgl/_deprecate/udf.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate
copying dgl/_deprecate/view.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate
copying dgl/_deprecate/kernel.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate
copying dgl/_deprecate/__init__.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate
copying dgl/_deprecate/graph.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate
copying dgl/_deprecate/frame.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate
creating build/lib.linux-x86_64-3.7/dgl/transforms
copying dgl/transforms/functional.py -> build/lib.linux-x86_64-3.7/dgl/transforms
copying dgl/transforms/__init__.py -> build/lib.linux-x86_64-3.7/dgl/transforms
copying dgl/transforms/module.py -> build/lib.linux-x86_64-3.7/dgl/transforms
creating build/lib.linux-x86_64-3.7/dgl/distgnn
copying dgl/distgnn/__init__.py -> build/lib.linux-x86_64-3.7/dgl/distgnn
creating build/lib.linux-x86_64-3.7/dgl/_ffi
copying dgl/_ffi/runtime_ctypes.py -> build/lib.linux-x86_64-3.7/dgl/_ffi
copying dgl/_ffi/function.py -> build/lib.linux-x86_64-3.7/dgl/_ffi
copying dgl/_ffi/libinfo.py -> build/lib.linux-x86_64-3.7/dgl/_ffi
copying dgl/_ffi/streams.py -> build/lib.linux-x86_64-3.7/dgl/_ffi
copying dgl/_ffi/base.py -> build/lib.linux-x86_64-3.7/dgl/_ffi
copying dgl/_ffi/ndarray.py -> build/lib.linux-x86_64-3.7/dgl/_ffi
copying dgl/_ffi/object_generic.py -> build/lib.linux-x86_64-3.7/dgl/_ffi
copying dgl/_ffi/__init__.py -> build/lib.linux-x86_64-3.7/dgl/_ffi
copying dgl/_ffi/object.py -> build/lib.linux-x86_64-3.7/dgl/_ffi
creating build/lib.linux-x86_64-3.7/dgl/nn/pytorch
copying dgl/nn/pytorch/glob.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch
copying dgl/nn/pytorch/hetero.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch
copying dgl/nn/pytorch/sparse_emb.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch
copying dgl/nn/pytorch/linear.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch
copying dgl/nn/pytorch/utils.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch
copying dgl/nn/pytorch/__init__.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch
copying dgl/nn/pytorch/factory.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch
copying dgl/nn/pytorch/softmax.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch
creating build/lib.linux-x86_64-3.7/dgl/nn/mxnet
copying dgl/nn/mxnet/glob.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet
copying dgl/nn/mxnet/hetero.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet
copying dgl/nn/mxnet/utils.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet
copying dgl/nn/mxnet/__init__.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet
copying dgl/nn/mxnet/softmax.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet
creating build/lib.linux-x86_64-3.7/dgl/nn/tensorflow
copying dgl/nn/tensorflow/glob.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow
copying dgl/nn/tensorflow/hetero.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow
copying dgl/nn/tensorflow/utils.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow
copying dgl/nn/tensorflow/__init__.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow
copying dgl/nn/tensorflow/softmax.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow
creating build/lib.linux-x86_64-3.7/dgl/nn/functional
copying dgl/nn/functional/__init__.py -> build/lib.linux-x86_64-3.7/dgl/nn/functional
creating build/lib.linux-x86_64-3.7/dgl/nn/pytorch/link
copying dgl/nn/pytorch/link/transe.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/link
copying dgl/nn/pytorch/link/edgepred.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/link
copying dgl/nn/pytorch/link/transr.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/link
copying dgl/nn/pytorch/link/__init__.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/link
creating build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/agnnconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/relgraphconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/appnpconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/dgnconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/grouprevres.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/twirlsconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/egatconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/graphconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/densechebconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/gcn2conv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/ginconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/dotgatconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/cfconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/gmmconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/sageconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/densesageconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/gineconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/edgeconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/nnconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/__init__.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/gatv2conv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/hgtconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/pnaconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/gatedgraphconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/atomicconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/tagconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/densegraphconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/egnnconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/chebconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/sgconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/gatconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
creating build/lib.linux-x86_64-3.7/dgl/nn/pytorch/explain
copying dgl/nn/pytorch/explain/gnnexplainer.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/explain
copying dgl/nn/pytorch/explain/__init__.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/explain
creating build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/agnnconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/relgraphconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/appnpconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/graphconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/densechebconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/ginconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/gmmconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/sageconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/densesageconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/edgeconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/nnconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/__init__.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/gatedgraphconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/tagconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/densegraphconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/chebconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/sgconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/gatconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
creating build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv
copying dgl/nn/tensorflow/conv/relgraphconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv
copying dgl/nn/tensorflow/conv/appnpconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv
copying dgl/nn/tensorflow/conv/graphconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv
copying dgl/nn/tensorflow/conv/densechebconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv
copying dgl/nn/tensorflow/conv/ginconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv
copying dgl/nn/tensorflow/conv/sageconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv
copying dgl/nn/tensorflow/conv/edgeconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv
copying dgl/nn/tensorflow/conv/__init__.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv
copying dgl/nn/tensorflow/conv/chebconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv
copying dgl/nn/tensorflow/conv/sgconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv
copying dgl/nn/tensorflow/conv/gatconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv
creating build/lib.linux-x86_64-3.7/dgl/contrib/sampling
copying dgl/contrib/sampling/sampler.py -> build/lib.linux-x86_64-3.7/dgl/contrib/sampling
copying dgl/contrib/sampling/dis_sampler.py -> build/lib.linux-x86_64-3.7/dgl/contrib/sampling
copying dgl/contrib/sampling/__init__.py -> build/lib.linux-x86_64-3.7/dgl/contrib/sampling
creating build/lib.linux-x86_64-3.7/dgl/contrib/data
copying dgl/contrib/data/knowledge_graph.py -> build/lib.linux-x86_64-3.7/dgl/contrib/data
copying dgl/contrib/data/__init__.py -> build/lib.linux-x86_64-3.7/dgl/contrib/data
creating build/lib.linux-x86_64-3.7/dgl/optim/pytorch
copying dgl/optim/pytorch/__init__.py -> build/lib.linux-x86_64-3.7/dgl/optim/pytorch
copying dgl/optim/pytorch/sparse_optim.py -> build/lib.linux-x86_64-3.7/dgl/optim/pytorch
creating build/lib.linux-x86_64-3.7/dgl/optim/mxnet
copying dgl/optim/mxnet/__init__.py -> build/lib.linux-x86_64-3.7/dgl/optim/mxnet
creating build/lib.linux-x86_64-3.7/dgl/optim/tensorflow
copying dgl/optim/tensorflow/__init__.py -> build/lib.linux-x86_64-3.7/dgl/optim/tensorflow
creating build/lib.linux-x86_64-3.7/dgl/_dataloading/pytorch
copying dgl/_dataloading/pytorch/dataloader.py -> build/lib.linux-x86_64-3.7/dgl/_dataloading/pytorch
copying dgl/_dataloading/pytorch/__init__.py -> build/lib.linux-x86_64-3.7/dgl/_dataloading/pytorch
creating build/lib.linux-x86_64-3.7/dgl/backend/pytorch
copying dgl/backend/pytorch/sparse.py -> build/lib.linux-x86_64-3.7/dgl/backend/pytorch
copying dgl/backend/pytorch/tensor.py -> build/lib.linux-x86_64-3.7/dgl/backend/pytorch
copying dgl/backend/pytorch/__init__.py -> build/lib.linux-x86_64-3.7/dgl/backend/pytorch
creating build/lib.linux-x86_64-3.7/dgl/backend/mxnet
copying dgl/backend/mxnet/sparse.py -> build/lib.linux-x86_64-3.7/dgl/backend/mxnet
copying dgl/backend/mxnet/tensor.py -> build/lib.linux-x86_64-3.7/dgl/backend/mxnet
copying dgl/backend/mxnet/__init__.py -> build/lib.linux-x86_64-3.7/dgl/backend/mxnet
copying dgl/backend/mxnet/sparse_optim.py -> build/lib.linux-x86_64-3.7/dgl/backend/mxnet
creating build/lib.linux-x86_64-3.7/dgl/backend/tensorflow
copying dgl/backend/tensorflow/sparse.py -> build/lib.linux-x86_64-3.7/dgl/backend/tensorflow
copying dgl/backend/tensorflow/tensor.py -> build/lib.linux-x86_64-3.7/dgl/backend/tensorflow
copying dgl/backend/tensorflow/__init__.py -> build/lib.linux-x86_64-3.7/dgl/backend/tensorflow
copying dgl/backend/tensorflow/sparse_optim.py -> build/lib.linux-x86_64-3.7/dgl/backend/tensorflow
creating build/lib.linux-x86_64-3.7/dgl/distributed/nn
copying dgl/distributed/nn/__init__.py -> build/lib.linux-x86_64-3.7/dgl/distributed/nn
creating build/lib.linux-x86_64-3.7/dgl/distributed/optim
copying dgl/distributed/optim/__init__.py -> build/lib.linux-x86_64-3.7/dgl/distributed/optim
creating build/lib.linux-x86_64-3.7/dgl/distributed/nn/pytorch
copying dgl/distributed/nn/pytorch/sparse_emb.py -> build/lib.linux-x86_64-3.7/dgl/distributed/nn/pytorch
copying dgl/distributed/nn/pytorch/__init__.py -> build/lib.linux-x86_64-3.7/dgl/distributed/nn/pytorch
creating build/lib.linux-x86_64-3.7/dgl/distributed/nn/mxnet
copying dgl/distributed/nn/mxnet/__init__.py -> build/lib.linux-x86_64-3.7/dgl/distributed/nn/mxnet
creating build/lib.linux-x86_64-3.7/dgl/distributed/nn/tensorflow
copying dgl/distributed/nn/tensorflow/__init__.py -> build/lib.linux-x86_64-3.7/dgl/distributed/nn/tensorflow
creating build/lib.linux-x86_64-3.7/dgl/distributed/optim/pytorch
copying dgl/distributed/optim/pytorch/utils.py -> build/lib.linux-x86_64-3.7/dgl/distributed/optim/pytorch
copying dgl/distributed/optim/pytorch/__init__.py -> build/lib.linux-x86_64-3.7/dgl/distributed/optim/pytorch
copying dgl/distributed/optim/pytorch/sparse_optim.py -> build/lib.linux-x86_64-3.7/dgl/distributed/optim/pytorch
creating build/lib.linux-x86_64-3.7/dgl/distributed/optim/mxnet
copying dgl/distributed/optim/mxnet/__init__.py -> build/lib.linux-x86_64-3.7/dgl/distributed/optim/mxnet
creating build/lib.linux-x86_64-3.7/dgl/distributed/optim/tensorflow
copying dgl/distributed/optim/tensorflow/__init__.py -> build/lib.linux-x86_64-3.7/dgl/distributed/optim/tensorflow
creating build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime
copying dgl/_deprecate/runtime/spmv.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime
copying dgl/_deprecate/runtime/degree_bucketing.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime
copying dgl/_deprecate/runtime/__init__.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime
copying dgl/_deprecate/runtime/scheduler.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime
copying dgl/_deprecate/runtime/adapter.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime
copying dgl/_deprecate/runtime/runtime.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime
creating build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/ir
copying dgl/_deprecate/runtime/ir/program.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/ir
copying dgl/_deprecate/runtime/ir/executor.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/ir
copying dgl/_deprecate/runtime/ir/var.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/ir
copying dgl/_deprecate/runtime/ir/registry.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/ir
copying dgl/_deprecate/runtime/ir/__init__.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/ir
creating build/lib.linux-x86_64-3.7/dgl/distgnn/partition
copying dgl/distgnn/partition/libra_partition.py -> build/lib.linux-x86_64-3.7/dgl/distgnn/partition
copying dgl/distgnn/partition/__init__.py -> build/lib.linux-x86_64-3.7/dgl/distgnn/partition
creating build/lib.linux-x86_64-3.7/dgl/distgnn/tools
copying dgl/distgnn/tools/__init__.py -> build/lib.linux-x86_64-3.7/dgl/distgnn/tools
copying dgl/distgnn/tools/tools.py -> build/lib.linux-x86_64-3.7/dgl/distgnn/tools
creating build/lib.linux-x86_64-3.7/dgl/_ffi/_ctypes
copying dgl/_ffi/_ctypes/function.py -> build/lib.linux-x86_64-3.7/dgl/_ffi/_ctypes
copying dgl/_ffi/_ctypes/types.py -> build/lib.linux-x86_64-3.7/dgl/_ffi/_ctypes
copying dgl/_ffi/_ctypes/ndarray.py -> build/lib.linux-x86_64-3.7/dgl/_ffi/_ctypes
copying dgl/_ffi/_ctypes/__init__.py -> build/lib.linux-x86_64-3.7/dgl/_ffi/_ctypes
copying dgl/_ffi/_ctypes/object.py -> build/lib.linux-x86_64-3.7/dgl/_ffi/_ctypes
creating build/lib.linux-x86_64-3.7/dgl/_ffi/_cy3
copying dgl/_ffi/_cy3/__init__.py -> build/lib.linux-x86_64-3.7/dgl/_ffi/_cy3
creating build/lib.linux-x86_64-3.7/dgl/_ffi/_cy2
copying dgl/_ffi/_cy2/__init__.py -> build/lib.linux-x86_64-3.7/dgl/_ffi/_cy2
creating build/lib.linux-x86_64-3.7/dgl/_ffi/_cython
copying dgl/_ffi/_cython/core.cpp -> build/lib.linux-x86_64-3.7/dgl/_ffi/_cython
running build_ext
building 'dgl._ffi._cy3.core' extension
creating build/temp.linux-x86_64-3.7
creating build/temp.linux-x86_64-3.7/dgl
creating build/temp.linux-x86_64-3.7/dgl/_ffi
creating build/temp.linux-x86_64-3.7/dgl/_ffi/_cython
gcc -pthread -B /opt/conda/envs/mxnet-ci/compiler_compat -Wl,--sysroot=/ -Wsign-compare -DNDEBUG -g -fwrapv -O3 -Wall -Wstrict-prototypes -fPIC -I../include/ -I../third_party/dmlc-core/include -I../third_party/dlpack/include -I/opt/conda/envs/mxnet-ci/include/python3.7m -c dgl/_ffi/_cython/core.cpp -o build/temp.linux-x86_64-3.7/dgl/_ffi/_cython/core.o
cc1plus: warning: command line option ‘-Wstrict-prototypes’ is valid for C/ObjC but not for C++
g++ -pthread -B /opt/conda/envs/mxnet-ci/compiler_compat -Wl,--sysroot=/ -pthread -shared -B /opt/conda/envs/mxnet-ci/compiler_compat -L/opt/conda/envs/mxnet-ci/lib -Wl,-rpath=/opt/conda/envs/mxnet-ci/lib -Wl,--no-as-needed -Wl,--sysroot=/ build/temp.linux-x86_64-3.7/dgl/_ffi/_cython/core.o -Ldgl -L../build/Release -L../build -ldgl -o build/lib.linux-x86_64-3.7/dgl/_ffi/_cy3/core.cpython-37m-x86_64-linux-gnu.so
/opt/conda/envs/mxnet-ci/compiler_compat/ld: warning: /opt/conda/envs/mxnet-ci/lib/libstdc++.so: unsupported GNU_PROPERTY_TYPE (5) type: 0xc0010001
/opt/conda/envs/mxnet-ci/compiler_compat/ld: warning: /opt/conda/envs/mxnet-ci/lib/libstdc++.so: unsupported GNU_PROPERTY_TYPE (5) type: 0xc0010002
/opt/conda/envs/mxnet-ci/compiler_compat/ld: warning: /opt/conda/envs/mxnet-ci/lib/libgcc_s.so.1: unsupported GNU_PROPERTY_TYPE (5) type: 0xc0010001
/opt/conda/envs/mxnet-ci/compiler_compat/ld: warning: /opt/conda/envs/mxnet-ci/lib/libgcc_s.so.1: unsupported GNU_PROPERTY_TYPE (5) type: 0xc0010002
/opt/conda/envs/mxnet-ci/compiler_compat/ld: warning: /opt/conda/envs/mxnet-ci/lib/libgcc_s.so.1: unsupported GNU_PROPERTY_TYPE (5) type: 0xc0010001
/opt/conda/envs/mxnet-ci/compiler_compat/ld: warning: /opt/conda/envs/mxnet-ci/lib/libgcc_s.so.1: unsupported GNU_PROPERTY_TYPE (5) type: 0xc0010002
creating build/bdist.linux-x86_64
creating build/bdist.linux-x86_64/egg
creating build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/mock_sparse
copying build/lib.linux-x86_64-3.7/dgl/mock_sparse/sp_matrix.py -> build/bdist.linux-x86_64/egg/dgl/mock_sparse
copying build/lib.linux-x86_64-3.7/dgl/mock_sparse/diag_matrix.py -> build/bdist.linux-x86_64/egg/dgl/mock_sparse
copying build/lib.linux-x86_64-3.7/dgl/mock_sparse/reduction.py -> build/bdist.linux-x86_64/egg/dgl/mock_sparse
copying build/lib.linux-x86_64-3.7/dgl/mock_sparse/__init__.py -> build/bdist.linux-x86_64/egg/dgl/mock_sparse
copying build/lib.linux-x86_64-3.7/dgl/mock_sparse/elementwise_op_sp.py -> build/bdist.linux-x86_64/egg/dgl/mock_sparse
creating build/bdist.linux-x86_64/egg/dgl/geometry
copying build/lib.linux-x86_64-3.7/dgl/geometry/edge_coarsening.py -> build/bdist.linux-x86_64/egg/dgl/geometry
copying build/lib.linux-x86_64-3.7/dgl/geometry/fps.py -> build/bdist.linux-x86_64/egg/dgl/geometry
copying build/lib.linux-x86_64-3.7/dgl/geometry/capi.py -> build/bdist.linux-x86_64/egg/dgl/geometry
copying build/lib.linux-x86_64-3.7/dgl/geometry/__init__.py -> build/bdist.linux-x86_64/egg/dgl/geometry
creating build/bdist.linux-x86_64/egg/dgl/function
copying build/lib.linux-x86_64-3.7/dgl/function/base.py -> build/bdist.linux-x86_64/egg/dgl/function
copying build/lib.linux-x86_64-3.7/dgl/function/message.py -> build/bdist.linux-x86_64/egg/dgl/function
copying build/lib.linux-x86_64-3.7/dgl/function/reducer.py -> build/bdist.linux-x86_64/egg/dgl/function
copying build/lib.linux-x86_64-3.7/dgl/function/__init__.py -> build/bdist.linux-x86_64/egg/dgl/function
copying build/lib.linux-x86_64-3.7/dgl/partition.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/nn
creating build/bdist.linux-x86_64/egg/dgl/nn/pytorch
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/glob.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch
creating build/bdist.linux-x86_64/egg/dgl/nn/pytorch/link
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/link/transe.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/link
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/link/edgepred.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/link
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/link/transr.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/link
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/link/__init__.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/link
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/hetero.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/sparse_emb.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/linear.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/utils.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch
creating build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/agnnconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/relgraphconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/appnpconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/dgnconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/grouprevres.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/twirlsconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/egatconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/graphconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/densechebconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/gcn2conv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/ginconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/dotgatconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/cfconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/gmmconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/sageconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/densesageconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/gineconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/edgeconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/nnconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/__init__.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/gatv2conv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/hgtconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/pnaconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/gatedgraphconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/atomicconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/tagconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/densegraphconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/egnnconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/chebconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/sgconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/gatconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/__init__.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch
creating build/bdist.linux-x86_64/egg/dgl/nn/pytorch/explain
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/explain/gnnexplainer.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/explain
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/explain/__init__.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/explain
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/factory.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/softmax.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch
creating build/bdist.linux-x86_64/egg/dgl/nn/mxnet
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/glob.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/hetero.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/utils.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet
creating build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/agnnconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/relgraphconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/appnpconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/graphconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/densechebconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/ginconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/gmmconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/sageconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/densesageconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/edgeconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/nnconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/__init__.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/gatedgraphconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/tagconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/densegraphconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/chebconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/sgconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/gatconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/__init__.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/softmax.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet
copying build/lib.linux-x86_64-3.7/dgl/nn/__init__.py -> build/bdist.linux-x86_64/egg/dgl/nn
creating build/bdist.linux-x86_64/egg/dgl/nn/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/glob.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/hetero.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/utils.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow
creating build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv/relgraphconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv/appnpconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv/graphconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv/densechebconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv/ginconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv/sageconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv/edgeconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv/__init__.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv/chebconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv/sgconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv/gatconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/__init__.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/softmax.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow
creating build/bdist.linux-x86_64/egg/dgl/nn/functional
copying build/lib.linux-x86_64-3.7/dgl/nn/functional/__init__.py -> build/bdist.linux-x86_64/egg/dgl/nn/functional
copying build/lib.linux-x86_64-3.7/dgl/core.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-3.7/dgl/subgraph.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/contrib
copying build/lib.linux-x86_64-3.7/dgl/contrib/dis_kvstore.py -> build/bdist.linux-x86_64/egg/dgl/contrib
copying build/lib.linux-x86_64-3.7/dgl/contrib/unified_tensor.py -> build/bdist.linux-x86_64/egg/dgl/contrib
copying build/lib.linux-x86_64-3.7/dgl/contrib/graph_store.py -> build/bdist.linux-x86_64/egg/dgl/contrib
creating build/bdist.linux-x86_64/egg/dgl/contrib/sampling
copying build/lib.linux-x86_64-3.7/dgl/contrib/sampling/sampler.py -> build/bdist.linux-x86_64/egg/dgl/contrib/sampling
copying build/lib.linux-x86_64-3.7/dgl/contrib/sampling/dis_sampler.py -> build/bdist.linux-x86_64/egg/dgl/contrib/sampling
copying build/lib.linux-x86_64-3.7/dgl/contrib/sampling/__init__.py -> build/bdist.linux-x86_64/egg/dgl/contrib/sampling
creating build/bdist.linux-x86_64/egg/dgl/contrib/data
copying build/lib.linux-x86_64-3.7/dgl/contrib/data/knowledge_graph.py -> build/bdist.linux-x86_64/egg/dgl/contrib/data
copying build/lib.linux-x86_64-3.7/dgl/contrib/data/__init__.py -> build/bdist.linux-x86_64/egg/dgl/contrib/data
copying build/lib.linux-x86_64-3.7/dgl/contrib/__init__.py -> build/bdist.linux-x86_64/egg/dgl/contrib
copying build/lib.linux-x86_64-3.7/dgl/traversal.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/utils
copying build/lib.linux-x86_64-3.7/dgl/utils/checks.py -> build/bdist.linux-x86_64/egg/dgl/utils
copying build/lib.linux-x86_64-3.7/dgl/utils/exception.py -> build/bdist.linux-x86_64/egg/dgl/utils
copying build/lib.linux-x86_64-3.7/dgl/utils/pin_memory.py -> build/bdist.linux-x86_64/egg/dgl/utils
copying build/lib.linux-x86_64-3.7/dgl/utils/internal.py -> build/bdist.linux-x86_64/egg/dgl/utils
copying build/lib.linux-x86_64-3.7/dgl/utils/filter.py -> build/bdist.linux-x86_64/egg/dgl/utils
copying build/lib.linux-x86_64-3.7/dgl/utils/data.py -> build/bdist.linux-x86_64/egg/dgl/utils
copying build/lib.linux-x86_64-3.7/dgl/utils/__init__.py -> build/bdist.linux-x86_64/egg/dgl/utils
copying build/lib.linux-x86_64-3.7/dgl/utils/shared_mem.py -> build/bdist.linux-x86_64/egg/dgl/utils
creating build/bdist.linux-x86_64/egg/dgl/multiprocessing
copying build/lib.linux-x86_64-3.7/dgl/multiprocessing/pytorch.py -> build/bdist.linux-x86_64/egg/dgl/multiprocessing
copying build/lib.linux-x86_64-3.7/dgl/multiprocessing/__init__.py -> build/bdist.linux-x86_64/egg/dgl/multiprocessing
copying build/lib.linux-x86_64-3.7/dgl/base.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/optim
creating build/bdist.linux-x86_64/egg/dgl/optim/pytorch
copying build/lib.linux-x86_64-3.7/dgl/optim/pytorch/__init__.py -> build/bdist.linux-x86_64/egg/dgl/optim/pytorch
copying build/lib.linux-x86_64-3.7/dgl/optim/pytorch/sparse_optim.py -> build/bdist.linux-x86_64/egg/dgl/optim/pytorch
creating build/bdist.linux-x86_64/egg/dgl/optim/mxnet
copying build/lib.linux-x86_64-3.7/dgl/optim/mxnet/__init__.py -> build/bdist.linux-x86_64/egg/dgl/optim/mxnet
copying build/lib.linux-x86_64-3.7/dgl/optim/__init__.py -> build/bdist.linux-x86_64/egg/dgl/optim
creating build/bdist.linux-x86_64/egg/dgl/optim/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/optim/tensorflow/__init__.py -> build/bdist.linux-x86_64/egg/dgl/optim/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/heterograph_index.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-3.7/dgl/sparse.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-3.7/dgl/global_config.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/dataloading
copying build/lib.linux-x86_64-3.7/dgl/dataloading/base.py -> build/bdist.linux-x86_64/egg/dgl/dataloading
copying build/lib.linux-x86_64-3.7/dgl/dataloading/dataloader.py -> build/bdist.linux-x86_64/egg/dgl/dataloading
copying build/lib.linux-x86_64-3.7/dgl/dataloading/graphsaint.py -> build/bdist.linux-x86_64/egg/dgl/dataloading
copying build/lib.linux-x86_64-3.7/dgl/dataloading/shadow.py -> build/bdist.linux-x86_64/egg/dgl/dataloading
copying build/lib.linux-x86_64-3.7/dgl/dataloading/negative_sampler.py -> build/bdist.linux-x86_64/egg/dgl/dataloading
copying build/lib.linux-x86_64-3.7/dgl/dataloading/__init__.py -> build/bdist.linux-x86_64/egg/dgl/dataloading
copying build/lib.linux-x86_64-3.7/dgl/dataloading/cluster_gcn.py -> build/bdist.linux-x86_64/egg/dgl/dataloading
copying build/lib.linux-x86_64-3.7/dgl/dataloading/neighbor_sampler.py -> build/bdist.linux-x86_64/egg/dgl/dataloading
copying build/lib.linux-x86_64-3.7/dgl/dataloading/dist_dataloader.py -> build/bdist.linux-x86_64/egg/dgl/dataloading
creating build/bdist.linux-x86_64/egg/dgl/_dataloading
copying build/lib.linux-x86_64-3.7/dgl/_dataloading/neighbor.py -> build/bdist.linux-x86_64/egg/dgl/_dataloading
copying build/lib.linux-x86_64-3.7/dgl/_dataloading/dataloader.py -> build/bdist.linux-x86_64/egg/dgl/_dataloading
creating build/bdist.linux-x86_64/egg/dgl/_dataloading/pytorch
copying build/lib.linux-x86_64-3.7/dgl/_dataloading/pytorch/dataloader.py -> build/bdist.linux-x86_64/egg/dgl/_dataloading/pytorch
copying build/lib.linux-x86_64-3.7/dgl/_dataloading/pytorch/__init__.py -> build/bdist.linux-x86_64/egg/dgl/_dataloading/pytorch
copying build/lib.linux-x86_64-3.7/dgl/_dataloading/shadow.py -> build/bdist.linux-x86_64/egg/dgl/_dataloading
copying build/lib.linux-x86_64-3.7/dgl/_dataloading/negative_sampler.py -> build/bdist.linux-x86_64/egg/dgl/_dataloading
copying build/lib.linux-x86_64-3.7/dgl/_dataloading/__init__.py -> build/bdist.linux-x86_64/egg/dgl/_dataloading
copying build/lib.linux-x86_64-3.7/dgl/_dataloading/cluster_gcn.py -> build/bdist.linux-x86_64/egg/dgl/_dataloading
copying build/lib.linux-x86_64-3.7/dgl/_api_internal.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-3.7/dgl/heterograph.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/backend
copying build/lib.linux-x86_64-3.7/dgl/backend/backend.py -> build/bdist.linux-x86_64/egg/dgl/backend
creating build/bdist.linux-x86_64/egg/dgl/backend/pytorch
copying build/lib.linux-x86_64-3.7/dgl/backend/pytorch/sparse.py -> build/bdist.linux-x86_64/egg/dgl/backend/pytorch
copying build/lib.linux-x86_64-3.7/dgl/backend/pytorch/tensor.py -> build/bdist.linux-x86_64/egg/dgl/backend/pytorch
copying build/lib.linux-x86_64-3.7/dgl/backend/pytorch/__init__.py -> build/bdist.linux-x86_64/egg/dgl/backend/pytorch
creating build/bdist.linux-x86_64/egg/dgl/backend/mxnet
copying build/lib.linux-x86_64-3.7/dgl/backend/mxnet/sparse.py -> build/bdist.linux-x86_64/egg/dgl/backend/mxnet
copying build/lib.linux-x86_64-3.7/dgl/backend/mxnet/tensor.py -> build/bdist.linux-x86_64/egg/dgl/backend/mxnet
copying build/lib.linux-x86_64-3.7/dgl/backend/mxnet/__init__.py -> build/bdist.linux-x86_64/egg/dgl/backend/mxnet
copying build/lib.linux-x86_64-3.7/dgl/backend/mxnet/sparse_optim.py -> build/bdist.linux-x86_64/egg/dgl/backend/mxnet
copying build/lib.linux-x86_64-3.7/dgl/backend/__init__.py -> build/bdist.linux-x86_64/egg/dgl/backend
creating build/bdist.linux-x86_64/egg/dgl/backend/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/backend/tensorflow/sparse.py -> build/bdist.linux-x86_64/egg/dgl/backend/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/backend/tensorflow/tensor.py -> build/bdist.linux-x86_64/egg/dgl/backend/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/backend/tensorflow/__init__.py -> build/bdist.linux-x86_64/egg/dgl/backend/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/backend/tensorflow/sparse_optim.py -> build/bdist.linux-x86_64/egg/dgl/backend/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/backend/set_default_backend.py -> build/bdist.linux-x86_64/egg/dgl/backend
copying build/lib.linux-x86_64-3.7/dgl/network.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/ops
copying build/lib.linux-x86_64-3.7/dgl/ops/edge_softmax.py -> build/bdist.linux-x86_64/egg/dgl/ops
copying build/lib.linux-x86_64-3.7/dgl/ops/sddmm.py -> build/bdist.linux-x86_64/egg/dgl/ops
copying build/lib.linux-x86_64-3.7/dgl/ops/gather_mm.py -> build/bdist.linux-x86_64/egg/dgl/ops
copying build/lib.linux-x86_64-3.7/dgl/ops/segment.py -> build/bdist.linux-x86_64/egg/dgl/ops
copying build/lib.linux-x86_64-3.7/dgl/ops/__init__.py -> build/bdist.linux-x86_64/egg/dgl/ops
copying build/lib.linux-x86_64-3.7/dgl/ops/spmm.py -> build/bdist.linux-x86_64/egg/dgl/ops
copying build/lib.linux-x86_64-3.7/dgl/logging.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-3.7/dgl/graph_index.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/cuda
copying build/lib.linux-x86_64-3.7/dgl/cuda/nccl.py -> build/bdist.linux-x86_64/egg/dgl/cuda
copying build/lib.linux-x86_64-3.7/dgl/cuda/__init__.py -> build/bdist.linux-x86_64/egg/dgl/cuda
copying build/lib.linux-x86_64-3.7/dgl/udf.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/storages
copying build/lib.linux-x86_64-3.7/dgl/storages/pytorch_tensor.py -> build/bdist.linux-x86_64/egg/dgl/storages
copying build/lib.linux-x86_64-3.7/dgl/storages/base.py -> build/bdist.linux-x86_64/egg/dgl/storages
copying build/lib.linux-x86_64-3.7/dgl/storages/tensor.py -> build/bdist.linux-x86_64/egg/dgl/storages
copying build/lib.linux-x86_64-3.7/dgl/storages/__init__.py -> build/bdist.linux-x86_64/egg/dgl/storages
copying build/lib.linux-x86_64-3.7/dgl/storages/numpy.py -> build/bdist.linux-x86_64/egg/dgl/storages
copying build/lib.linux-x86_64-3.7/dgl/ndarray.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-3.7/dgl/init.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-3.7/dgl/view.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/sampling
copying build/lib.linux-x86_64-3.7/dgl/sampling/neighbor.py -> build/bdist.linux-x86_64/egg/dgl/sampling
copying build/lib.linux-x86_64-3.7/dgl/sampling/pinsage.py -> build/bdist.linux-x86_64/egg/dgl/sampling
copying build/lib.linux-x86_64-3.7/dgl/sampling/negative.py -> build/bdist.linux-x86_64/egg/dgl/sampling
copying build/lib.linux-x86_64-3.7/dgl/sampling/randomwalks.py -> build/bdist.linux-x86_64/egg/dgl/sampling
copying build/lib.linux-x86_64-3.7/dgl/sampling/utils.py -> build/bdist.linux-x86_64/egg/dgl/sampling
copying build/lib.linux-x86_64-3.7/dgl/sampling/node2vec_randomwalk.py -> build/bdist.linux-x86_64/egg/dgl/sampling
copying build/lib.linux-x86_64-3.7/dgl/sampling/__init__.py -> build/bdist.linux-x86_64/egg/dgl/sampling
copying build/lib.linux-x86_64-3.7/dgl/generators.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/qm7b.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/graph_serialize.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/dgl_dataset.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/csv_dataset_base.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/knowledge_graph.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/qm9_edge.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/gnn_benchmark.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/qm9.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/icews18.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/gdelt.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/utils.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/sbm.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/csv_dataset.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/karate.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/citation_graph.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/rdf.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/bitcoinotc.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/ppi.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/tensor_serialize.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/flickr.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/wikics.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/heterograph_serialize.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/tu.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/__init__.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/gindt.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/tree.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/synthetic.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/adapter.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/fakenews.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/fraud.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/minigc.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/yelp.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/reddit.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/convert.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-3.7/dgl/container.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-3.7/dgl/merge.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-3.7/dgl/readout.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-3.7/dgl/__init__.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/dist_graph.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/graph_partition_book.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/partition.py -> build/bdist.linux-x86_64/egg/dgl/distributed
creating build/bdist.linux-x86_64/egg/dgl/distributed/nn
creating build/bdist.linux-x86_64/egg/dgl/distributed/nn/pytorch
copying build/lib.linux-x86_64-3.7/dgl/distributed/nn/pytorch/sparse_emb.py -> build/bdist.linux-x86_64/egg/dgl/distributed/nn/pytorch
copying build/lib.linux-x86_64-3.7/dgl/distributed/nn/pytorch/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distributed/nn/pytorch
creating build/bdist.linux-x86_64/egg/dgl/distributed/nn/mxnet
copying build/lib.linux-x86_64-3.7/dgl/distributed/nn/mxnet/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distributed/nn/mxnet
copying build/lib.linux-x86_64-3.7/dgl/distributed/nn/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distributed/nn
creating build/bdist.linux-x86_64/egg/dgl/distributed/nn/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/distributed/nn/tensorflow/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distributed/nn/tensorflow
creating build/bdist.linux-x86_64/egg/dgl/distributed/optim
creating build/bdist.linux-x86_64/egg/dgl/distributed/optim/pytorch
copying build/lib.linux-x86_64-3.7/dgl/distributed/optim/pytorch/utils.py -> build/bdist.linux-x86_64/egg/dgl/distributed/optim/pytorch
copying build/lib.linux-x86_64-3.7/dgl/distributed/optim/pytorch/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distributed/optim/pytorch
copying build/lib.linux-x86_64-3.7/dgl/distributed/optim/pytorch/sparse_optim.py -> build/bdist.linux-x86_64/egg/dgl/distributed/optim/pytorch
creating build/bdist.linux-x86_64/egg/dgl/distributed/optim/mxnet
copying build/lib.linux-x86_64-3.7/dgl/distributed/optim/mxnet/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distributed/optim/mxnet
copying build/lib.linux-x86_64-3.7/dgl/distributed/optim/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distributed/optim
creating build/bdist.linux-x86_64/egg/dgl/distributed/optim/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/distributed/optim/tensorflow/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distributed/optim/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/distributed/standalone_kvstore.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/graph_services.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/role.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/constants.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/rpc_server.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/kvstore.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/dist_context.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/rpc.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/rpc_client.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/id_map.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/server_state.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/shared_mem_utils.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/dist_tensor.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/dist_dataloader.py -> build/bdist.linux-x86_64/egg/dgl/distributed
creating build/bdist.linux-x86_64/egg/dgl/_deprecate
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/nodeflow.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/udf.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/view.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/kernel.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/__init__.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate
creating build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/spmv.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/degree_bucketing.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/__init__.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/scheduler.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/adapter.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/runtime.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime
creating build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/ir
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/ir/program.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/ir
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/ir/executor.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/ir
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/ir/var.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/ir
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/ir/registry.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/ir
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/ir/__init__.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/ir
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/graph.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/frame.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate
copying build/lib.linux-x86_64-3.7/dgl/propagate.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-3.7/dgl/batch.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/transforms
copying build/lib.linux-x86_64-3.7/dgl/transforms/functional.py -> build/bdist.linux-x86_64/egg/dgl/transforms
copying build/lib.linux-x86_64-3.7/dgl/transforms/__init__.py -> build/bdist.linux-x86_64/egg/dgl/transforms
copying build/lib.linux-x86_64-3.7/dgl/transforms/module.py -> build/bdist.linux-x86_64/egg/dgl/transforms
creating build/bdist.linux-x86_64/egg/dgl/distgnn
creating build/bdist.linux-x86_64/egg/dgl/distgnn/partition
copying build/lib.linux-x86_64-3.7/dgl/distgnn/partition/libra_partition.py -> build/bdist.linux-x86_64/egg/dgl/distgnn/partition
copying build/lib.linux-x86_64-3.7/dgl/distgnn/partition/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distgnn/partition
creating build/bdist.linux-x86_64/egg/dgl/distgnn/tools
copying build/lib.linux-x86_64-3.7/dgl/distgnn/tools/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distgnn/tools
copying build/lib.linux-x86_64-3.7/dgl/distgnn/tools/tools.py -> build/bdist.linux-x86_64/egg/dgl/distgnn/tools
copying build/lib.linux-x86_64-3.7/dgl/distgnn/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distgnn
creating build/bdist.linux-x86_64/egg/dgl/_ffi
copying build/lib.linux-x86_64-3.7/dgl/_ffi/runtime_ctypes.py -> build/bdist.linux-x86_64/egg/dgl/_ffi
copying build/lib.linux-x86_64-3.7/dgl/_ffi/function.py -> build/bdist.linux-x86_64/egg/dgl/_ffi
copying build/lib.linux-x86_64-3.7/dgl/_ffi/libinfo.py -> build/bdist.linux-x86_64/egg/dgl/_ffi
copying build/lib.linux-x86_64-3.7/dgl/_ffi/streams.py -> build/bdist.linux-x86_64/egg/dgl/_ffi
copying build/lib.linux-x86_64-3.7/dgl/_ffi/base.py -> build/bdist.linux-x86_64/egg/dgl/_ffi
creating build/bdist.linux-x86_64/egg/dgl/_ffi/_ctypes
copying build/lib.linux-x86_64-3.7/dgl/_ffi/_ctypes/function.py -> build/bdist.linux-x86_64/egg/dgl/_ffi/_ctypes
copying build/lib.linux-x86_64-3.7/dgl/_ffi/_ctypes/types.py -> build/bdist.linux-x86_64/egg/dgl/_ffi/_ctypes
copying build/lib.linux-x86_64-3.7/dgl/_ffi/_ctypes/ndarray.py -> build/bdist.linux-x86_64/egg/dgl/_ffi/_ctypes
copying build/lib.linux-x86_64-3.7/dgl/_ffi/_ctypes/__init__.py -> build/bdist.linux-x86_64/egg/dgl/_ffi/_ctypes
copying build/lib.linux-x86_64-3.7/dgl/_ffi/_ctypes/object.py -> build/bdist.linux-x86_64/egg/dgl/_ffi/_ctypes
copying build/lib.linux-x86_64-3.7/dgl/_ffi/ndarray.py -> build/bdist.linux-x86_64/egg/dgl/_ffi
copying build/lib.linux-x86_64-3.7/dgl/_ffi/object_generic.py -> build/bdist.linux-x86_64/egg/dgl/_ffi
copying build/lib.linux-x86_64-3.7/dgl/_ffi/__init__.py -> build/bdist.linux-x86_64/egg/dgl/_ffi
copying build/lib.linux-x86_64-3.7/dgl/_ffi/object.py -> build/bdist.linux-x86_64/egg/dgl/_ffi
creating build/bdist.linux-x86_64/egg/dgl/_ffi/_cy3
copying build/lib.linux-x86_64-3.7/dgl/_ffi/_cy3/__init__.py -> build/bdist.linux-x86_64/egg/dgl/_ffi/_cy3
copying build/lib.linux-x86_64-3.7/dgl/_ffi/_cy3/core.cpython-37m-x86_64-linux-gnu.so -> build/bdist.linux-x86_64/egg/dgl/_ffi/_cy3
creating build/bdist.linux-x86_64/egg/dgl/_ffi/_cy2
copying build/lib.linux-x86_64-3.7/dgl/_ffi/_cy2/__init__.py -> build/bdist.linux-x86_64/egg/dgl/_ffi/_cy2
creating build/bdist.linux-x86_64/egg/dgl/_ffi/_cython
copying build/lib.linux-x86_64-3.7/dgl/_ffi/_cython/core.cpp -> build/bdist.linux-x86_64/egg/dgl/_ffi/_cython
copying build/lib.linux-x86_64-3.7/dgl/frame.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-3.7/dgl/random.py -> build/bdist.linux-x86_64/egg/dgl
byte-compiling build/bdist.linux-x86_64/egg/dgl/mock_sparse/sp_matrix.py to sp_matrix.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/mock_sparse/diag_matrix.py to diag_matrix.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/mock_sparse/reduction.py to reduction.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/mock_sparse/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/mock_sparse/elementwise_op_sp.py to elementwise_op_sp.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/geometry/edge_coarsening.py to edge_coarsening.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/geometry/fps.py to fps.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/geometry/capi.py to capi.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/geometry/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/function/base.py to base.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/function/message.py to message.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/function/reducer.py to reducer.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/function/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/partition.py to partition.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/glob.py to glob.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/link/transe.py to transe.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/link/edgepred.py to edgepred.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/link/transr.py to transr.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/link/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/hetero.py to hetero.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/sparse_emb.py to sparse_emb.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/linear.py to linear.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/utils.py to utils.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/agnnconv.py to agnnconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/relgraphconv.py to relgraphconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/appnpconv.py to appnpconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/dgnconv.py to dgnconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/grouprevres.py to grouprevres.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/twirlsconv.py to twirlsconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/egatconv.py to egatconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/graphconv.py to graphconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/densechebconv.py to densechebconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/gcn2conv.py to gcn2conv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/ginconv.py to ginconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/dotgatconv.py to dotgatconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/cfconv.py to cfconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/gmmconv.py to gmmconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/sageconv.py to sageconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/densesageconv.py to densesageconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/gineconv.py to gineconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/edgeconv.py to edgeconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/nnconv.py to nnconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/gatv2conv.py to gatv2conv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/hgtconv.py to hgtconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/pnaconv.py to pnaconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/gatedgraphconv.py to gatedgraphconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/atomicconv.py to atomicconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/tagconv.py to tagconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/densegraphconv.py to densegraphconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/egnnconv.py to egnnconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/chebconv.py to chebconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/sgconv.py to sgconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/gatconv.py to gatconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/explain/gnnexplainer.py to gnnexplainer.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/explain/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/factory.py to factory.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/softmax.py to softmax.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/glob.py to glob.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/hetero.py to hetero.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/utils.py to utils.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/agnnconv.py to agnnconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/relgraphconv.py to relgraphconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/appnpconv.py to appnpconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/graphconv.py to graphconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/densechebconv.py to densechebconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/ginconv.py to ginconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/gmmconv.py to gmmconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/sageconv.py to sageconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/densesageconv.py to densesageconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/edgeconv.py to edgeconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/nnconv.py to nnconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/gatedgraphconv.py to gatedgraphconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/tagconv.py to tagconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/densegraphconv.py to densegraphconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/chebconv.py to chebconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/sgconv.py to sgconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/gatconv.py to gatconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/softmax.py to softmax.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/glob.py to glob.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/hetero.py to hetero.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/utils.py to utils.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv/relgraphconv.py to relgraphconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv/appnpconv.py to appnpconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv/graphconv.py to graphconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv/densechebconv.py to densechebconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv/ginconv.py to ginconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv/sageconv.py to sageconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv/edgeconv.py to edgeconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv/chebconv.py to chebconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv/sgconv.py to sgconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv/gatconv.py to gatconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/softmax.py to softmax.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/functional/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/core.py to core.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/subgraph.py to subgraph.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/contrib/dis_kvstore.py to dis_kvstore.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/contrib/unified_tensor.py to unified_tensor.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/contrib/graph_store.py to graph_store.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/contrib/sampling/sampler.py to sampler.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/contrib/sampling/dis_sampler.py to dis_sampler.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/contrib/sampling/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/contrib/data/knowledge_graph.py to knowledge_graph.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/contrib/data/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/contrib/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/traversal.py to traversal.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/utils/checks.py to checks.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/utils/exception.py to exception.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/utils/pin_memory.py to pin_memory.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/utils/internal.py to internal.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/utils/filter.py to filter.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/utils/data.py to data.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/utils/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/utils/shared_mem.py to shared_mem.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/multiprocessing/pytorch.py to pytorch.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/multiprocessing/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/base.py to base.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/optim/pytorch/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/optim/pytorch/sparse_optim.py to sparse_optim.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/optim/mxnet/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/optim/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/optim/tensorflow/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/heterograph_index.py to heterograph_index.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/sparse.py to sparse.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/global_config.py to global_config.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/dataloading/base.py to base.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/dataloading/dataloader.py to dataloader.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/dataloading/graphsaint.py to graphsaint.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/dataloading/shadow.py to shadow.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/dataloading/negative_sampler.py to negative_sampler.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/dataloading/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/dataloading/cluster_gcn.py to cluster_gcn.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/dataloading/neighbor_sampler.py to neighbor_sampler.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/dataloading/dist_dataloader.py to dist_dataloader.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_dataloading/neighbor.py to neighbor.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_dataloading/dataloader.py to dataloader.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_dataloading/pytorch/dataloader.py to dataloader.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_dataloading/pytorch/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_dataloading/shadow.py to shadow.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_dataloading/negative_sampler.py to negative_sampler.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_dataloading/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_dataloading/cluster_gcn.py to cluster_gcn.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_api_internal.py to _api_internal.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/heterograph.py to heterograph.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/backend.py to backend.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/pytorch/sparse.py to sparse.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/pytorch/tensor.py to tensor.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/pytorch/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/mxnet/sparse.py to sparse.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/mxnet/tensor.py to tensor.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/mxnet/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/mxnet/sparse_optim.py to sparse_optim.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/tensorflow/sparse.py to sparse.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/tensorflow/tensor.py to tensor.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/tensorflow/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/tensorflow/sparse_optim.py to sparse_optim.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/set_default_backend.py to set_default_backend.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/network.py to network.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/ops/edge_softmax.py to edge_softmax.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/ops/sddmm.py to sddmm.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/ops/gather_mm.py to gather_mm.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/ops/segment.py to segment.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/ops/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/ops/spmm.py to spmm.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/logging.py to logging.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/graph_index.py to graph_index.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/cuda/nccl.py to nccl.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/cuda/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/udf.py to udf.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/storages/pytorch_tensor.py to pytorch_tensor.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/storages/base.py to base.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/storages/tensor.py to tensor.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/storages/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/storages/numpy.py to numpy.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/ndarray.py to ndarray.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/init.py to init.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/view.py to view.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/sampling/neighbor.py to neighbor.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/sampling/pinsage.py to pinsage.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/sampling/negative.py to negative.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/sampling/randomwalks.py to randomwalks.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/sampling/utils.py to utils.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/sampling/node2vec_randomwalk.py to node2vec_randomwalk.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/sampling/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/generators.py to generators.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/qm7b.py to qm7b.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/graph_serialize.py to graph_serialize.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/dgl_dataset.py to dgl_dataset.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/csv_dataset_base.py to csv_dataset_base.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/knowledge_graph.py to knowledge_graph.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/qm9_edge.py to qm9_edge.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/gnn_benchmark.py to gnn_benchmark.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/qm9.py to qm9.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/icews18.py to icews18.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/gdelt.py to gdelt.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/utils.py to utils.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/sbm.py to sbm.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/csv_dataset.py to csv_dataset.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/karate.py to karate.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/citation_graph.py to citation_graph.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/rdf.py to rdf.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/bitcoinotc.py to bitcoinotc.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/ppi.py to ppi.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/tensor_serialize.py to tensor_serialize.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/flickr.py to flickr.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/wikics.py to wikics.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/heterograph_serialize.py to heterograph_serialize.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/tu.py to tu.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/gindt.py to gindt.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/tree.py to tree.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/synthetic.py to synthetic.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/adapter.py to adapter.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/fakenews.py to fakenews.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/fraud.py to fraud.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/minigc.py to minigc.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/yelp.py to yelp.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/reddit.py to reddit.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/convert.py to convert.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/container.py to container.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/merge.py to merge.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/readout.py to readout.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/dist_graph.py to dist_graph.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/graph_partition_book.py to graph_partition_book.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/partition.py to partition.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/nn/pytorch/sparse_emb.py to sparse_emb.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/nn/pytorch/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/nn/mxnet/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/nn/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/nn/tensorflow/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/optim/pytorch/utils.py to utils.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/optim/pytorch/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/optim/pytorch/sparse_optim.py to sparse_optim.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/optim/mxnet/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/optim/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/optim/tensorflow/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/standalone_kvstore.py to standalone_kvstore.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/graph_services.py to graph_services.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/role.py to role.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/constants.py to constants.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/rpc_server.py to rpc_server.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/kvstore.py to kvstore.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/dist_context.py to dist_context.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/rpc.py to rpc.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/rpc_client.py to rpc_client.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/id_map.py to id_map.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/server_state.py to server_state.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/shared_mem_utils.py to shared_mem_utils.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/dist_tensor.py to dist_tensor.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/dist_dataloader.py to dist_dataloader.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/nodeflow.py to nodeflow.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/udf.py to udf.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/view.py to view.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/kernel.py to kernel.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/spmv.py to spmv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/degree_bucketing.py to degree_bucketing.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/scheduler.py to scheduler.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/adapter.py to adapter.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/runtime.py to runtime.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/ir/program.py to program.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/ir/executor.py to executor.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/ir/var.py to var.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/ir/registry.py to registry.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/ir/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/graph.py to graph.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/frame.py to frame.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/propagate.py to propagate.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/batch.py to batch.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/transforms/functional.py to functional.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/transforms/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/transforms/module.py to module.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distgnn/partition/libra_partition.py to libra_partition.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distgnn/partition/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distgnn/tools/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distgnn/tools/tools.py to tools.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distgnn/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/runtime_ctypes.py to runtime_ctypes.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/function.py to function.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/libinfo.py to libinfo.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/streams.py to streams.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/base.py to base.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/_ctypes/function.py to function.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/_ctypes/types.py to types.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/_ctypes/ndarray.py to ndarray.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/_ctypes/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/_ctypes/object.py to object.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/ndarray.py to ndarray.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/object_generic.py to object_generic.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/object.py to object.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/_cy3/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/_cy2/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/frame.py to frame.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/random.py to random.cpython-37.pyc
creating stub loader for dgl/_ffi/_cy3/core.cpython-37m-x86_64-linux-gnu.so
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/_cy3/core.py to core.cpython-37.pyc
installing package data to build/bdist.linux-x86_64/egg
running install_data
copying ../build/libdgl.so -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/tensoradapter
creating build/bdist.linux-x86_64/egg/dgl/tensoradapter/pytorch
copying ../build/tensoradapter/pytorch/libtensoradapter_pytorch_1.9.0.so -> build/bdist.linux-x86_64/egg/dgl/tensoradapter/pytorch
creating build/bdist.linux-x86_64/egg/EGG-INFO
copying dgl.egg-info/PKG-INFO -> build/bdist.linux-x86_64/egg/EGG-INFO
copying dgl.egg-info/SOURCES.txt -> build/bdist.linux-x86_64/egg/EGG-INFO
copying dgl.egg-info/dependency_links.txt -> build/bdist.linux-x86_64/egg/EGG-INFO
copying dgl.egg-info/not-zip-safe -> build/bdist.linux-x86_64/egg/EGG-INFO
copying dgl.egg-info/requires.txt -> build/bdist.linux-x86_64/egg/EGG-INFO
copying dgl.egg-info/top_level.txt -> build/bdist.linux-x86_64/egg/EGG-INFO
writing build/bdist.linux-x86_64/egg/EGG-INFO/native_libs.txt
creating dist
creating 'dist/dgl-0.9-py3.7-linux-x86_64.egg' and adding 'build/bdist.linux-x86_64/egg' to it
removing 'build/bdist.linux-x86_64/egg' (and everything under it)
Processing dgl-0.9-py3.7-linux-x86_64.egg
creating /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages/dgl-0.9-py3.7-linux-x86_64.egg
Extracting dgl-0.9-py3.7-linux-x86_64.egg to /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages
Adding dgl 0.9 to easy-install.pth file

Installed /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages/dgl-0.9-py3.7-linux-x86_64.egg
Processing dependencies for dgl==0.9
Searching for psutil>=5.8.0
Reading https://pypi.org/simple/psutil/
Downloading https://files.pythonhosted.org/packages/3d/73/d8c87b5612c58d1e6c6d91997c1590771d34e4ee27d9c11eb1e64ecbf365/psutil-5.9.2-cp37-cp37m-manylinux_2_12_x86_64.manylinux2010_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl#sha256=4fb54941aac044a61db9d8eb56fc5bee207db3bc58645d657249030e15ba3727
Best match: psutil 5.9.2
Processing psutil-5.9.2-cp37-cp37m-manylinux_2_12_x86_64.manylinux2010_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl
Installing psutil-5.9.2-cp37-cp37m-manylinux_2_12_x86_64.manylinux2010_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl to /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages
Adding psutil 5.9.2 to easy-install.pth file

Installed /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages/psutil-5.9.2-py3.7-linux-x86_64.egg
Searching for tqdm==4.64.0
Best match: tqdm 4.64.0
Adding tqdm 4.64.0 to easy-install.pth file
Installing tqdm script to /opt/conda/envs/mxnet-ci/bin

Using /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages
Searching for requests==2.28.1
Best match: requests 2.28.1
Adding requests 2.28.1 to easy-install.pth file

Using /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages
Searching for networkx==2.6.3
Best match: networkx 2.6.3
Adding networkx 2.6.3 to easy-install.pth file

Using /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages
Searching for scipy==1.7.3
Best match: scipy 1.7.3
Adding scipy 1.7.3 to easy-install.pth file

Using /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages
Searching for numpy==1.21.6
Best match: numpy 1.21.6
Adding numpy 1.21.6 to easy-install.pth file
Installing f2py script to /opt/conda/envs/mxnet-ci/bin
Installing f2py3 script to /opt/conda/envs/mxnet-ci/bin
Installing f2py3.7 script to /opt/conda/envs/mxnet-ci/bin

Using /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages
Searching for certifi==2022.6.15
Best match: certifi 2022.6.15
Adding certifi 2022.6.15 to easy-install.pth file

Using /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages
Searching for urllib3==1.26.11
Best match: urllib3 1.26.11
Adding urllib3 1.26.11 to easy-install.pth file

Using /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages
Searching for idna==3.3
Best match: idna 3.3
Adding idna 3.3 to easy-install.pth file

Using /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages
Searching for charset-normalizer==2.1.0
Best match: charset-normalizer 2.1.0
Adding charset-normalizer 2.1.0 to easy-install.pth file
Installing normalizer script to /opt/conda/envs/mxnet-ci/bin

Using /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages
Finished processing dependencies for dgl==0.9
/opt/conda/envs/mxnet-ci/lib/python3.7/site-packages/Cython/Compiler/Main.py:369: FutureWarning: Cython directive 'language_level' not set, using 2 for now (Py2). This will change in a later release! File: /root/jenkins/workspace/dgl_PR-4648/python/dgl/_ffi/_cython/core.pyx
  tree = Parsing.p_module(s, pxd, full_module_name)
/opt/conda/envs/mxnet-ci/lib/python3.7/site-packages/setuptools/command/install.py:37: SetuptoolsDeprecationWarning: setup.py install is deprecated. Use build and pip and other standards-based tools.
  setuptools.SetuptoolsDeprecationWarning,
/opt/conda/envs/mxnet-ci/lib/python3.7/site-packages/setuptools/command/easy_install.py:147: EasyInstallDeprecationWarning: easy_install command is deprecated. Use build and pip and other standards-based tools.
  EasyInstallDeprecationWarning,
/opt/conda/envs/mxnet-ci/lib/python3.7/site-packages/pkg_resources/__init__.py:125: PkgResourcesDeprecationWarning:  is an invalid version and will not be supported in a future release
  PkgResourcesDeprecationWarning,
[1/1] Cythonizing dgl/_ffi/_cython/core.pyx
running build_ext
building 'dgl._ffi._cy3.core' extension
gcc -pthread -B /opt/conda/envs/mxnet-ci/compiler_compat -Wl,--sysroot=/ -Wsign-compare -DNDEBUG -g -fwrapv -O3 -Wall -Wstrict-prototypes -fPIC -I../include/ -I../third_party/dmlc-core/include -I../third_party/dlpack/include -I/opt/conda/envs/mxnet-ci/include/python3.7m -c dgl/_ffi/_cython/core.cpp -o build/temp.linux-x86_64-3.7/dgl/_ffi/_cython/core.o
cc1plus: warning: command line option ‘-Wstrict-prototypes’ is valid for C/ObjC but not for C++
g++ -pthread -B /opt/conda/envs/mxnet-ci/compiler_compat -Wl,--sysroot=/ -pthread -shared -B /opt/conda/envs/mxnet-ci/compiler_compat -L/opt/conda/envs/mxnet-ci/lib -Wl,-rpath=/opt/conda/envs/mxnet-ci/lib -Wl,--no-as-needed -Wl,--sysroot=/ build/temp.linux-x86_64-3.7/dgl/_ffi/_cython/core.o -Ldgl -L../build/Release -L../build -ldgl -o build/lib.linux-x86_64-3.7/dgl/_ffi/_cy3/core.cpython-37m-x86_64-linux-gnu.so
/opt/conda/envs/mxnet-ci/compiler_compat/ld: warning: /opt/conda/envs/mxnet-ci/lib/libstdc++.so: unsupported GNU_PROPERTY_TYPE (5) type: 0xc0010001
/opt/conda/envs/mxnet-ci/compiler_compat/ld: warning: /opt/conda/envs/mxnet-ci/lib/libstdc++.so: unsupported GNU_PROPERTY_TYPE (5) type: 0xc0010002
/opt/conda/envs/mxnet-ci/compiler_compat/ld: warning: /opt/conda/envs/mxnet-ci/lib/libgcc_s.so.1: unsupported GNU_PROPERTY_TYPE (5) type: 0xc0010001
/opt/conda/envs/mxnet-ci/compiler_compat/ld: warning: /opt/conda/envs/mxnet-ci/lib/libgcc_s.so.1: unsupported GNU_PROPERTY_TYPE (5) type: 0xc0010002
/opt/conda/envs/mxnet-ci/compiler_compat/ld: warning: /opt/conda/envs/mxnet-ci/lib/libgcc_s.so.1: unsupported GNU_PROPERTY_TYPE (5) type: 0xc0010001
/opt/conda/envs/mxnet-ci/compiler_compat/ld: warning: /opt/conda/envs/mxnet-ci/lib/libgcc_s.so.1: unsupported GNU_PROPERTY_TYPE (5) type: 0xc0010002
copying build/lib.linux-x86_64-3.7/dgl/_ffi/_cy3/core.cpython-37m-x86_64-linux-gnu.so -> dgl/_ffi/_cy3
/opt/conda/envs/mxnet-ci/lib/python3.7/site-packages/Cython/Compiler/Main.py:369: FutureWarning: Cython directive 'language_level' not set, using 2 for now (Py2). This will change in a later release! File: /root/jenkins/workspace/dgl_PR-4648/python/dgl/_ffi/_cython/core.pyx
  tree = Parsing.p_module(s, pxd, full_module_name)
WARNING: Skipping dgl as it is not installed.
WARNING: Running pip as the 'root' user can result in broken permissions and conflicting behaviour with the system package manager. It is recommended to use a virtual environment instead: https://pip.pypa.io/warnings/venv
[1/1] Cythonizing dgl/_ffi/_cython/core.pyx
running install
running bdist_egg
running egg_info
creating dgl.egg-info
writing dgl.egg-info/PKG-INFO
writing dependency_links to dgl.egg-info/dependency_links.txt
writing requirements to dgl.egg-info/requires.txt
writing top-level names to dgl.egg-info/top_level.txt
writing manifest file 'dgl.egg-info/SOURCES.txt'
reading manifest file 'dgl.egg-info/SOURCES.txt'
writing manifest file 'dgl.egg-info/SOURCES.txt'
installing library code to build/bdist.linux-x86_64/egg
running install_lib
running build_py
creating build
creating build/lib.linux-x86_64-3.7
creating build/lib.linux-x86_64-3.7/dgl
copying dgl/partition.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/core.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/subgraph.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/traversal.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/base.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/heterograph_index.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/sparse.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/global_config.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/_api_internal.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/heterograph.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/network.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/logging.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/graph_index.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/udf.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/ndarray.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/init.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/view.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/generators.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/convert.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/container.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/merge.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/readout.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/__init__.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/propagate.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/batch.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/frame.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/random.py -> build/lib.linux-x86_64-3.7/dgl
creating build/lib.linux-x86_64-3.7/dgl/mock_sparse
copying dgl/mock_sparse/sp_matrix.py -> build/lib.linux-x86_64-3.7/dgl/mock_sparse
copying dgl/mock_sparse/diag_matrix.py -> build/lib.linux-x86_64-3.7/dgl/mock_sparse
copying dgl/mock_sparse/reduction.py -> build/lib.linux-x86_64-3.7/dgl/mock_sparse
copying dgl/mock_sparse/__init__.py -> build/lib.linux-x86_64-3.7/dgl/mock_sparse
copying dgl/mock_sparse/elementwise_op_sp.py -> build/lib.linux-x86_64-3.7/dgl/mock_sparse
creating build/lib.linux-x86_64-3.7/dgl/geometry
copying dgl/geometry/edge_coarsening.py -> build/lib.linux-x86_64-3.7/dgl/geometry
copying dgl/geometry/fps.py -> build/lib.linux-x86_64-3.7/dgl/geometry
copying dgl/geometry/capi.py -> build/lib.linux-x86_64-3.7/dgl/geometry
copying dgl/geometry/__init__.py -> build/lib.linux-x86_64-3.7/dgl/geometry
creating build/lib.linux-x86_64-3.7/dgl/function
copying dgl/function/base.py -> build/lib.linux-x86_64-3.7/dgl/function
copying dgl/function/message.py -> build/lib.linux-x86_64-3.7/dgl/function
copying dgl/function/reducer.py -> build/lib.linux-x86_64-3.7/dgl/function
copying dgl/function/__init__.py -> build/lib.linux-x86_64-3.7/dgl/function
creating build/lib.linux-x86_64-3.7/dgl/nn
copying dgl/nn/__init__.py -> build/lib.linux-x86_64-3.7/dgl/nn
creating build/lib.linux-x86_64-3.7/dgl/contrib
copying dgl/contrib/dis_kvstore.py -> build/lib.linux-x86_64-3.7/dgl/contrib
copying dgl/contrib/unified_tensor.py -> build/lib.linux-x86_64-3.7/dgl/contrib
copying dgl/contrib/graph_store.py -> build/lib.linux-x86_64-3.7/dgl/contrib
copying dgl/contrib/__init__.py -> build/lib.linux-x86_64-3.7/dgl/contrib
creating build/lib.linux-x86_64-3.7/dgl/utils
copying dgl/utils/checks.py -> build/lib.linux-x86_64-3.7/dgl/utils
copying dgl/utils/exception.py -> build/lib.linux-x86_64-3.7/dgl/utils
copying dgl/utils/pin_memory.py -> build/lib.linux-x86_64-3.7/dgl/utils
copying dgl/utils/internal.py -> build/lib.linux-x86_64-3.7/dgl/utils
copying dgl/utils/filter.py -> build/lib.linux-x86_64-3.7/dgl/utils
copying dgl/utils/data.py -> build/lib.linux-x86_64-3.7/dgl/utils
copying dgl/utils/__init__.py -> build/lib.linux-x86_64-3.7/dgl/utils
copying dgl/utils/shared_mem.py -> build/lib.linux-x86_64-3.7/dgl/utils
creating build/lib.linux-x86_64-3.7/dgl/multiprocessing
copying dgl/multiprocessing/pytorch.py -> build/lib.linux-x86_64-3.7/dgl/multiprocessing
copying dgl/multiprocessing/__init__.py -> build/lib.linux-x86_64-3.7/dgl/multiprocessing
creating build/lib.linux-x86_64-3.7/dgl/optim
copying dgl/optim/__init__.py -> build/lib.linux-x86_64-3.7/dgl/optim
creating build/lib.linux-x86_64-3.7/dgl/dataloading
copying dgl/dataloading/base.py -> build/lib.linux-x86_64-3.7/dgl/dataloading
copying dgl/dataloading/dataloader.py -> build/lib.linux-x86_64-3.7/dgl/dataloading
copying dgl/dataloading/graphsaint.py -> build/lib.linux-x86_64-3.7/dgl/dataloading
copying dgl/dataloading/shadow.py -> build/lib.linux-x86_64-3.7/dgl/dataloading
copying dgl/dataloading/negative_sampler.py -> build/lib.linux-x86_64-3.7/dgl/dataloading
copying dgl/dataloading/__init__.py -> build/lib.linux-x86_64-3.7/dgl/dataloading
copying dgl/dataloading/cluster_gcn.py -> build/lib.linux-x86_64-3.7/dgl/dataloading
copying dgl/dataloading/neighbor_sampler.py -> build/lib.linux-x86_64-3.7/dgl/dataloading
copying dgl/dataloading/dist_dataloader.py -> build/lib.linux-x86_64-3.7/dgl/dataloading
creating build/lib.linux-x86_64-3.7/dgl/_dataloading
copying dgl/_dataloading/neighbor.py -> build/lib.linux-x86_64-3.7/dgl/_dataloading
copying dgl/_dataloading/dataloader.py -> build/lib.linux-x86_64-3.7/dgl/_dataloading
copying dgl/_dataloading/shadow.py -> build/lib.linux-x86_64-3.7/dgl/_dataloading
copying dgl/_dataloading/negative_sampler.py -> build/lib.linux-x86_64-3.7/dgl/_dataloading
copying dgl/_dataloading/__init__.py -> build/lib.linux-x86_64-3.7/dgl/_dataloading
copying dgl/_dataloading/cluster_gcn.py -> build/lib.linux-x86_64-3.7/dgl/_dataloading
creating build/lib.linux-x86_64-3.7/dgl/backend
copying dgl/backend/backend.py -> build/lib.linux-x86_64-3.7/dgl/backend
copying dgl/backend/__init__.py -> build/lib.linux-x86_64-3.7/dgl/backend
copying dgl/backend/set_default_backend.py -> build/lib.linux-x86_64-3.7/dgl/backend
creating build/lib.linux-x86_64-3.7/dgl/ops
copying dgl/ops/edge_softmax.py -> build/lib.linux-x86_64-3.7/dgl/ops
copying dgl/ops/sddmm.py -> build/lib.linux-x86_64-3.7/dgl/ops
copying dgl/ops/gather_mm.py -> build/lib.linux-x86_64-3.7/dgl/ops
copying dgl/ops/segment.py -> build/lib.linux-x86_64-3.7/dgl/ops
copying dgl/ops/__init__.py -> build/lib.linux-x86_64-3.7/dgl/ops
copying dgl/ops/spmm.py -> build/lib.linux-x86_64-3.7/dgl/ops
creating build/lib.linux-x86_64-3.7/dgl/cuda
copying dgl/cuda/nccl.py -> build/lib.linux-x86_64-3.7/dgl/cuda
copying dgl/cuda/__init__.py -> build/lib.linux-x86_64-3.7/dgl/cuda
creating build/lib.linux-x86_64-3.7/dgl/storages
copying dgl/storages/pytorch_tensor.py -> build/lib.linux-x86_64-3.7/dgl/storages
copying dgl/storages/base.py -> build/lib.linux-x86_64-3.7/dgl/storages
copying dgl/storages/tensor.py -> build/lib.linux-x86_64-3.7/dgl/storages
copying dgl/storages/__init__.py -> build/lib.linux-x86_64-3.7/dgl/storages
copying dgl/storages/numpy.py -> build/lib.linux-x86_64-3.7/dgl/storages
creating build/lib.linux-x86_64-3.7/dgl/sampling
copying dgl/sampling/neighbor.py -> build/lib.linux-x86_64-3.7/dgl/sampling
copying dgl/sampling/pinsage.py -> build/lib.linux-x86_64-3.7/dgl/sampling
copying dgl/sampling/negative.py -> build/lib.linux-x86_64-3.7/dgl/sampling
copying dgl/sampling/randomwalks.py -> build/lib.linux-x86_64-3.7/dgl/sampling
copying dgl/sampling/utils.py -> build/lib.linux-x86_64-3.7/dgl/sampling
copying dgl/sampling/node2vec_randomwalk.py -> build/lib.linux-x86_64-3.7/dgl/sampling
copying dgl/sampling/__init__.py -> build/lib.linux-x86_64-3.7/dgl/sampling
creating build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/qm7b.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/graph_serialize.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/dgl_dataset.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/csv_dataset_base.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/knowledge_graph.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/qm9_edge.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/gnn_benchmark.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/qm9.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/icews18.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/gdelt.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/utils.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/sbm.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/csv_dataset.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/karate.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/citation_graph.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/rdf.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/bitcoinotc.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/ppi.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/tensor_serialize.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/flickr.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/wikics.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/heterograph_serialize.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/tu.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/__init__.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/gindt.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/tree.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/synthetic.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/adapter.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/fakenews.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/fraud.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/minigc.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/yelp.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/reddit.py -> build/lib.linux-x86_64-3.7/dgl/data
creating build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/dist_graph.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/graph_partition_book.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/partition.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/standalone_kvstore.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/graph_services.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/role.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/constants.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/rpc_server.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/kvstore.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/dist_context.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/rpc.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/rpc_client.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/id_map.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/server_state.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/__init__.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/shared_mem_utils.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/dist_tensor.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/dist_dataloader.py -> build/lib.linux-x86_64-3.7/dgl/distributed
creating build/lib.linux-x86_64-3.7/dgl/_deprecate
copying dgl/_deprecate/nodeflow.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate
copying dgl/_deprecate/udf.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate
copying dgl/_deprecate/view.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate
copying dgl/_deprecate/kernel.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate
copying dgl/_deprecate/__init__.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate
copying dgl/_deprecate/graph.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate
copying dgl/_deprecate/frame.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate
creating build/lib.linux-x86_64-3.7/dgl/transforms
copying dgl/transforms/functional.py -> build/lib.linux-x86_64-3.7/dgl/transforms
copying dgl/transforms/__init__.py -> build/lib.linux-x86_64-3.7/dgl/transforms
copying dgl/transforms/module.py -> build/lib.linux-x86_64-3.7/dgl/transforms
creating build/lib.linux-x86_64-3.7/dgl/distgnn
copying dgl/distgnn/__init__.py -> build/lib.linux-x86_64-3.7/dgl/distgnn
creating build/lib.linux-x86_64-3.7/dgl/_ffi
copying dgl/_ffi/runtime_ctypes.py -> build/lib.linux-x86_64-3.7/dgl/_ffi
copying dgl/_ffi/function.py -> build/lib.linux-x86_64-3.7/dgl/_ffi
copying dgl/_ffi/libinfo.py -> build/lib.linux-x86_64-3.7/dgl/_ffi
copying dgl/_ffi/streams.py -> build/lib.linux-x86_64-3.7/dgl/_ffi
copying dgl/_ffi/base.py -> build/lib.linux-x86_64-3.7/dgl/_ffi
copying dgl/_ffi/ndarray.py -> build/lib.linux-x86_64-3.7/dgl/_ffi
copying dgl/_ffi/object_generic.py -> build/lib.linux-x86_64-3.7/dgl/_ffi
copying dgl/_ffi/__init__.py -> build/lib.linux-x86_64-3.7/dgl/_ffi
copying dgl/_ffi/object.py -> build/lib.linux-x86_64-3.7/dgl/_ffi
creating build/lib.linux-x86_64-3.7/dgl/nn/pytorch
copying dgl/nn/pytorch/glob.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch
copying dgl/nn/pytorch/hetero.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch
copying dgl/nn/pytorch/sparse_emb.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch
copying dgl/nn/pytorch/linear.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch
copying dgl/nn/pytorch/utils.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch
copying dgl/nn/pytorch/__init__.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch
copying dgl/nn/pytorch/factory.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch
copying dgl/nn/pytorch/softmax.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch
creating build/lib.linux-x86_64-3.7/dgl/nn/mxnet
copying dgl/nn/mxnet/glob.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet
copying dgl/nn/mxnet/hetero.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet
copying dgl/nn/mxnet/utils.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet
copying dgl/nn/mxnet/__init__.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet
copying dgl/nn/mxnet/softmax.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet
creating build/lib.linux-x86_64-3.7/dgl/nn/tensorflow
copying dgl/nn/tensorflow/glob.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow
copying dgl/nn/tensorflow/hetero.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow
copying dgl/nn/tensorflow/utils.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow
copying dgl/nn/tensorflow/__init__.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow
copying dgl/nn/tensorflow/softmax.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow
creating build/lib.linux-x86_64-3.7/dgl/nn/functional
copying dgl/nn/functional/__init__.py -> build/lib.linux-x86_64-3.7/dgl/nn/functional
creating build/lib.linux-x86_64-3.7/dgl/nn/pytorch/link
copying dgl/nn/pytorch/link/transe.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/link
copying dgl/nn/pytorch/link/edgepred.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/link
copying dgl/nn/pytorch/link/transr.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/link
copying dgl/nn/pytorch/link/__init__.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/link
creating build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/agnnconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/relgraphconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/appnpconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/dgnconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/grouprevres.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/twirlsconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/egatconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/graphconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/densechebconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/gcn2conv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/ginconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/dotgatconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/cfconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/gmmconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/sageconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/densesageconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/gineconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/edgeconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/nnconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/__init__.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/gatv2conv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/hgtconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/pnaconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/gatedgraphconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/atomicconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/tagconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/densegraphconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/egnnconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/chebconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/sgconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/gatconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
creating build/lib.linux-x86_64-3.7/dgl/nn/pytorch/explain
copying dgl/nn/pytorch/explain/gnnexplainer.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/explain
copying dgl/nn/pytorch/explain/__init__.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/explain
creating build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/agnnconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/relgraphconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/appnpconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/graphconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/densechebconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/ginconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/gmmconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/sageconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/densesageconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/edgeconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/nnconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/__init__.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/gatedgraphconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/tagconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/densegraphconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/chebconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/sgconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/gatconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
creating build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv
copying dgl/nn/tensorflow/conv/relgraphconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv
copying dgl/nn/tensorflow/conv/appnpconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv
copying dgl/nn/tensorflow/conv/graphconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv
copying dgl/nn/tensorflow/conv/densechebconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv
copying dgl/nn/tensorflow/conv/ginconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv
copying dgl/nn/tensorflow/conv/sageconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv
copying dgl/nn/tensorflow/conv/edgeconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv
copying dgl/nn/tensorflow/conv/__init__.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv
copying dgl/nn/tensorflow/conv/chebconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv
copying dgl/nn/tensorflow/conv/sgconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv
copying dgl/nn/tensorflow/conv/gatconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv
creating build/lib.linux-x86_64-3.7/dgl/contrib/sampling
copying dgl/contrib/sampling/sampler.py -> build/lib.linux-x86_64-3.7/dgl/contrib/sampling
copying dgl/contrib/sampling/dis_sampler.py -> build/lib.linux-x86_64-3.7/dgl/contrib/sampling
copying dgl/contrib/sampling/__init__.py -> build/lib.linux-x86_64-3.7/dgl/contrib/sampling
creating build/lib.linux-x86_64-3.7/dgl/contrib/data
copying dgl/contrib/data/knowledge_graph.py -> build/lib.linux-x86_64-3.7/dgl/contrib/data
copying dgl/contrib/data/__init__.py -> build/lib.linux-x86_64-3.7/dgl/contrib/data
creating build/lib.linux-x86_64-3.7/dgl/optim/pytorch
copying dgl/optim/pytorch/__init__.py -> build/lib.linux-x86_64-3.7/dgl/optim/pytorch
copying dgl/optim/pytorch/sparse_optim.py -> build/lib.linux-x86_64-3.7/dgl/optim/pytorch
creating build/lib.linux-x86_64-3.7/dgl/optim/mxnet
copying dgl/optim/mxnet/__init__.py -> build/lib.linux-x86_64-3.7/dgl/optim/mxnet
creating build/lib.linux-x86_64-3.7/dgl/optim/tensorflow
copying dgl/optim/tensorflow/__init__.py -> build/lib.linux-x86_64-3.7/dgl/optim/tensorflow
creating build/lib.linux-x86_64-3.7/dgl/_dataloading/pytorch
copying dgl/_dataloading/pytorch/dataloader.py -> build/lib.linux-x86_64-3.7/dgl/_dataloading/pytorch
copying dgl/_dataloading/pytorch/__init__.py -> build/lib.linux-x86_64-3.7/dgl/_dataloading/pytorch
creating build/lib.linux-x86_64-3.7/dgl/backend/pytorch
copying dgl/backend/pytorch/sparse.py -> build/lib.linux-x86_64-3.7/dgl/backend/pytorch
copying dgl/backend/pytorch/tensor.py -> build/lib.linux-x86_64-3.7/dgl/backend/pytorch
copying dgl/backend/pytorch/__init__.py -> build/lib.linux-x86_64-3.7/dgl/backend/pytorch
creating build/lib.linux-x86_64-3.7/dgl/backend/mxnet
copying dgl/backend/mxnet/sparse.py -> build/lib.linux-x86_64-3.7/dgl/backend/mxnet
copying dgl/backend/mxnet/tensor.py -> build/lib.linux-x86_64-3.7/dgl/backend/mxnet
copying dgl/backend/mxnet/__init__.py -> build/lib.linux-x86_64-3.7/dgl/backend/mxnet
copying dgl/backend/mxnet/sparse_optim.py -> build/lib.linux-x86_64-3.7/dgl/backend/mxnet
creating build/lib.linux-x86_64-3.7/dgl/backend/tensorflow
copying dgl/backend/tensorflow/sparse.py -> build/lib.linux-x86_64-3.7/dgl/backend/tensorflow
copying dgl/backend/tensorflow/tensor.py -> build/lib.linux-x86_64-3.7/dgl/backend/tensorflow
copying dgl/backend/tensorflow/__init__.py -> build/lib.linux-x86_64-3.7/dgl/backend/tensorflow
copying dgl/backend/tensorflow/sparse_optim.py -> build/lib.linux-x86_64-3.7/dgl/backend/tensorflow
creating build/lib.linux-x86_64-3.7/dgl/distributed/nn
copying dgl/distributed/nn/__init__.py -> build/lib.linux-x86_64-3.7/dgl/distributed/nn
creating build/lib.linux-x86_64-3.7/dgl/distributed/optim
copying dgl/distributed/optim/__init__.py -> build/lib.linux-x86_64-3.7/dgl/distributed/optim
creating build/lib.linux-x86_64-3.7/dgl/distributed/nn/pytorch
copying dgl/distributed/nn/pytorch/sparse_emb.py -> build/lib.linux-x86_64-3.7/dgl/distributed/nn/pytorch
copying dgl/distributed/nn/pytorch/__init__.py -> build/lib.linux-x86_64-3.7/dgl/distributed/nn/pytorch
creating build/lib.linux-x86_64-3.7/dgl/distributed/nn/mxnet
copying dgl/distributed/nn/mxnet/__init__.py -> build/lib.linux-x86_64-3.7/dgl/distributed/nn/mxnet
creating build/lib.linux-x86_64-3.7/dgl/distributed/nn/tensorflow
copying dgl/distributed/nn/tensorflow/__init__.py -> build/lib.linux-x86_64-3.7/dgl/distributed/nn/tensorflow
creating build/lib.linux-x86_64-3.7/dgl/distributed/optim/pytorch
copying dgl/distributed/optim/pytorch/utils.py -> build/lib.linux-x86_64-3.7/dgl/distributed/optim/pytorch
copying dgl/distributed/optim/pytorch/__init__.py -> build/lib.linux-x86_64-3.7/dgl/distributed/optim/pytorch
copying dgl/distributed/optim/pytorch/sparse_optim.py -> build/lib.linux-x86_64-3.7/dgl/distributed/optim/pytorch
creating build/lib.linux-x86_64-3.7/dgl/distributed/optim/mxnet
copying dgl/distributed/optim/mxnet/__init__.py -> build/lib.linux-x86_64-3.7/dgl/distributed/optim/mxnet
creating build/lib.linux-x86_64-3.7/dgl/distributed/optim/tensorflow
copying dgl/distributed/optim/tensorflow/__init__.py -> build/lib.linux-x86_64-3.7/dgl/distributed/optim/tensorflow
creating build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime
copying dgl/_deprecate/runtime/spmv.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime
copying dgl/_deprecate/runtime/degree_bucketing.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime
copying dgl/_deprecate/runtime/__init__.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime
copying dgl/_deprecate/runtime/scheduler.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime
copying dgl/_deprecate/runtime/adapter.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime
copying dgl/_deprecate/runtime/runtime.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime
creating build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/ir
copying dgl/_deprecate/runtime/ir/program.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/ir
copying dgl/_deprecate/runtime/ir/executor.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/ir
copying dgl/_deprecate/runtime/ir/var.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/ir
copying dgl/_deprecate/runtime/ir/registry.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/ir
copying dgl/_deprecate/runtime/ir/__init__.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/ir
creating build/lib.linux-x86_64-3.7/dgl/distgnn/partition
copying dgl/distgnn/partition/libra_partition.py -> build/lib.linux-x86_64-3.7/dgl/distgnn/partition
copying dgl/distgnn/partition/__init__.py -> build/lib.linux-x86_64-3.7/dgl/distgnn/partition
creating build/lib.linux-x86_64-3.7/dgl/distgnn/tools
copying dgl/distgnn/tools/__init__.py -> build/lib.linux-x86_64-3.7/dgl/distgnn/tools
copying dgl/distgnn/tools/tools.py -> build/lib.linux-x86_64-3.7/dgl/distgnn/tools
creating build/lib.linux-x86_64-3.7/dgl/_ffi/_ctypes
copying dgl/_ffi/_ctypes/function.py -> build/lib.linux-x86_64-3.7/dgl/_ffi/_ctypes
copying dgl/_ffi/_ctypes/types.py -> build/lib.linux-x86_64-3.7/dgl/_ffi/_ctypes
copying dgl/_ffi/_ctypes/ndarray.py -> build/lib.linux-x86_64-3.7/dgl/_ffi/_ctypes
copying dgl/_ffi/_ctypes/__init__.py -> build/lib.linux-x86_64-3.7/dgl/_ffi/_ctypes
copying dgl/_ffi/_ctypes/object.py -> build/lib.linux-x86_64-3.7/dgl/_ffi/_ctypes
creating build/lib.linux-x86_64-3.7/dgl/_ffi/_cy3
copying dgl/_ffi/_cy3/__init__.py -> build/lib.linux-x86_64-3.7/dgl/_ffi/_cy3
creating build/lib.linux-x86_64-3.7/dgl/_ffi/_cy2
copying dgl/_ffi/_cy2/__init__.py -> build/lib.linux-x86_64-3.7/dgl/_ffi/_cy2
creating build/lib.linux-x86_64-3.7/dgl/_ffi/_cython
copying dgl/_ffi/_cython/core.cpp -> build/lib.linux-x86_64-3.7/dgl/_ffi/_cython
running build_ext
building 'dgl._ffi._cy3.core' extension
creating build/temp.linux-x86_64-3.7
creating build/temp.linux-x86_64-3.7/dgl
creating build/temp.linux-x86_64-3.7/dgl/_ffi
creating build/temp.linux-x86_64-3.7/dgl/_ffi/_cython
gcc -pthread -B /opt/conda/envs/tensorflow-ci/compiler_compat -Wl,--sysroot=/ -Wsign-compare -DNDEBUG -g -fwrapv -O3 -Wall -Wstrict-prototypes -fPIC -I../include/ -I../third_party/dmlc-core/include -I../third_party/dlpack/include -I/opt/conda/envs/tensorflow-ci/include/python3.7m -c dgl/_ffi/_cython/core.cpp -o build/temp.linux-x86_64-3.7/dgl/_ffi/_cython/core.o
cc1plus: warning: command line option ‘-Wstrict-prototypes’ is valid for C/ObjC but not for C++
g++ -pthread -B /opt/conda/envs/tensorflow-ci/compiler_compat -Wl,--sysroot=/ -pthread -shared -B /opt/conda/envs/tensorflow-ci/compiler_compat -L/opt/conda/envs/tensorflow-ci/lib -Wl,-rpath=/opt/conda/envs/tensorflow-ci/lib -Wl,--no-as-needed -Wl,--sysroot=/ build/temp.linux-x86_64-3.7/dgl/_ffi/_cython/core.o -Ldgl -L../build/Release -L../build -ldgl -o build/lib.linux-x86_64-3.7/dgl/_ffi/_cy3/core.cpython-37m-x86_64-linux-gnu.so
/opt/conda/envs/tensorflow-ci/compiler_compat/ld: warning: /opt/conda/envs/tensorflow-ci/lib/libstdc++.so: unsupported GNU_PROPERTY_TYPE (5) type: 0xc0010001
/opt/conda/envs/tensorflow-ci/compiler_compat/ld: warning: /opt/conda/envs/tensorflow-ci/lib/libstdc++.so: unsupported GNU_PROPERTY_TYPE (5) type: 0xc0010002
/opt/conda/envs/tensorflow-ci/compiler_compat/ld: warning: /opt/conda/envs/tensorflow-ci/lib/libgcc_s.so.1: unsupported GNU_PROPERTY_TYPE (5) type: 0xc0010001
/opt/conda/envs/tensorflow-ci/compiler_compat/ld: warning: /opt/conda/envs/tensorflow-ci/lib/libgcc_s.so.1: unsupported GNU_PROPERTY_TYPE (5) type: 0xc0010002
/opt/conda/envs/tensorflow-ci/compiler_compat/ld: warning: /opt/conda/envs/tensorflow-ci/lib/libgcc_s.so.1: unsupported GNU_PROPERTY_TYPE (5) type: 0xc0010001
/opt/conda/envs/tensorflow-ci/compiler_compat/ld: warning: /opt/conda/envs/tensorflow-ci/lib/libgcc_s.so.1: unsupported GNU_PROPERTY_TYPE (5) type: 0xc0010002
creating build/bdist.linux-x86_64
creating build/bdist.linux-x86_64/egg
creating build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/mock_sparse
copying build/lib.linux-x86_64-3.7/dgl/mock_sparse/sp_matrix.py -> build/bdist.linux-x86_64/egg/dgl/mock_sparse
copying build/lib.linux-x86_64-3.7/dgl/mock_sparse/diag_matrix.py -> build/bdist.linux-x86_64/egg/dgl/mock_sparse
copying build/lib.linux-x86_64-3.7/dgl/mock_sparse/reduction.py -> build/bdist.linux-x86_64/egg/dgl/mock_sparse
copying build/lib.linux-x86_64-3.7/dgl/mock_sparse/__init__.py -> build/bdist.linux-x86_64/egg/dgl/mock_sparse
copying build/lib.linux-x86_64-3.7/dgl/mock_sparse/elementwise_op_sp.py -> build/bdist.linux-x86_64/egg/dgl/mock_sparse
creating build/bdist.linux-x86_64/egg/dgl/geometry
copying build/lib.linux-x86_64-3.7/dgl/geometry/edge_coarsening.py -> build/bdist.linux-x86_64/egg/dgl/geometry
copying build/lib.linux-x86_64-3.7/dgl/geometry/fps.py -> build/bdist.linux-x86_64/egg/dgl/geometry
copying build/lib.linux-x86_64-3.7/dgl/geometry/capi.py -> build/bdist.linux-x86_64/egg/dgl/geometry
copying build/lib.linux-x86_64-3.7/dgl/geometry/__init__.py -> build/bdist.linux-x86_64/egg/dgl/geometry
creating build/bdist.linux-x86_64/egg/dgl/function
copying build/lib.linux-x86_64-3.7/dgl/function/base.py -> build/bdist.linux-x86_64/egg/dgl/function
copying build/lib.linux-x86_64-3.7/dgl/function/message.py -> build/bdist.linux-x86_64/egg/dgl/function
copying build/lib.linux-x86_64-3.7/dgl/function/reducer.py -> build/bdist.linux-x86_64/egg/dgl/function
copying build/lib.linux-x86_64-3.7/dgl/function/__init__.py -> build/bdist.linux-x86_64/egg/dgl/function
copying build/lib.linux-x86_64-3.7/dgl/partition.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/nn
creating build/bdist.linux-x86_64/egg/dgl/nn/pytorch
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/glob.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch
creating build/bdist.linux-x86_64/egg/dgl/nn/pytorch/link
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/link/transe.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/link
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/link/edgepred.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/link
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/link/transr.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/link
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/link/__init__.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/link
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/hetero.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/sparse_emb.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/linear.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/utils.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch
creating build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/agnnconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/relgraphconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/appnpconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/dgnconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/grouprevres.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/twirlsconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/egatconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/graphconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/densechebconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/gcn2conv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/ginconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/dotgatconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/cfconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/gmmconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/sageconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/densesageconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/gineconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/edgeconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/nnconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/__init__.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/gatv2conv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/hgtconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/pnaconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/gatedgraphconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/atomicconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/tagconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/densegraphconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/egnnconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/chebconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/sgconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/gatconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/__init__.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch
creating build/bdist.linux-x86_64/egg/dgl/nn/pytorch/explain
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/explain/gnnexplainer.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/explain
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/explain/__init__.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/explain
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/factory.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/softmax.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch
creating build/bdist.linux-x86_64/egg/dgl/nn/mxnet
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/glob.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/hetero.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/utils.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet
creating build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/agnnconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/relgraphconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/appnpconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/graphconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/densechebconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/ginconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/gmmconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/sageconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/densesageconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/edgeconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/nnconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/__init__.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/gatedgraphconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/tagconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/densegraphconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/chebconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/sgconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/gatconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/__init__.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/softmax.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet
copying build/lib.linux-x86_64-3.7/dgl/nn/__init__.py -> build/bdist.linux-x86_64/egg/dgl/nn
creating build/bdist.linux-x86_64/egg/dgl/nn/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/glob.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/hetero.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/utils.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow
creating build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv/relgraphconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv/appnpconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv/graphconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv/densechebconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv/ginconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv/sageconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv/edgeconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv/__init__.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv/chebconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv/sgconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv/gatconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/__init__.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/softmax.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow
creating build/bdist.linux-x86_64/egg/dgl/nn/functional
copying build/lib.linux-x86_64-3.7/dgl/nn/functional/__init__.py -> build/bdist.linux-x86_64/egg/dgl/nn/functional
copying build/lib.linux-x86_64-3.7/dgl/core.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-3.7/dgl/subgraph.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/contrib
copying build/lib.linux-x86_64-3.7/dgl/contrib/dis_kvstore.py -> build/bdist.linux-x86_64/egg/dgl/contrib
copying build/lib.linux-x86_64-3.7/dgl/contrib/unified_tensor.py -> build/bdist.linux-x86_64/egg/dgl/contrib
copying build/lib.linux-x86_64-3.7/dgl/contrib/graph_store.py -> build/bdist.linux-x86_64/egg/dgl/contrib
creating build/bdist.linux-x86_64/egg/dgl/contrib/sampling
copying build/lib.linux-x86_64-3.7/dgl/contrib/sampling/sampler.py -> build/bdist.linux-x86_64/egg/dgl/contrib/sampling
copying build/lib.linux-x86_64-3.7/dgl/contrib/sampling/dis_sampler.py -> build/bdist.linux-x86_64/egg/dgl/contrib/sampling
copying build/lib.linux-x86_64-3.7/dgl/contrib/sampling/__init__.py -> build/bdist.linux-x86_64/egg/dgl/contrib/sampling
creating build/bdist.linux-x86_64/egg/dgl/contrib/data
copying build/lib.linux-x86_64-3.7/dgl/contrib/data/knowledge_graph.py -> build/bdist.linux-x86_64/egg/dgl/contrib/data
copying build/lib.linux-x86_64-3.7/dgl/contrib/data/__init__.py -> build/bdist.linux-x86_64/egg/dgl/contrib/data
copying build/lib.linux-x86_64-3.7/dgl/contrib/__init__.py -> build/bdist.linux-x86_64/egg/dgl/contrib
copying build/lib.linux-x86_64-3.7/dgl/traversal.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/utils
copying build/lib.linux-x86_64-3.7/dgl/utils/checks.py -> build/bdist.linux-x86_64/egg/dgl/utils
copying build/lib.linux-x86_64-3.7/dgl/utils/exception.py -> build/bdist.linux-x86_64/egg/dgl/utils
copying build/lib.linux-x86_64-3.7/dgl/utils/pin_memory.py -> build/bdist.linux-x86_64/egg/dgl/utils
copying build/lib.linux-x86_64-3.7/dgl/utils/internal.py -> build/bdist.linux-x86_64/egg/dgl/utils
copying build/lib.linux-x86_64-3.7/dgl/utils/filter.py -> build/bdist.linux-x86_64/egg/dgl/utils
copying build/lib.linux-x86_64-3.7/dgl/utils/data.py -> build/bdist.linux-x86_64/egg/dgl/utils
copying build/lib.linux-x86_64-3.7/dgl/utils/__init__.py -> build/bdist.linux-x86_64/egg/dgl/utils
copying build/lib.linux-x86_64-3.7/dgl/utils/shared_mem.py -> build/bdist.linux-x86_64/egg/dgl/utils
creating build/bdist.linux-x86_64/egg/dgl/multiprocessing
copying build/lib.linux-x86_64-3.7/dgl/multiprocessing/pytorch.py -> build/bdist.linux-x86_64/egg/dgl/multiprocessing
copying build/lib.linux-x86_64-3.7/dgl/multiprocessing/__init__.py -> build/bdist.linux-x86_64/egg/dgl/multiprocessing
copying build/lib.linux-x86_64-3.7/dgl/base.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/optim
creating build/bdist.linux-x86_64/egg/dgl/optim/pytorch
copying build/lib.linux-x86_64-3.7/dgl/optim/pytorch/__init__.py -> build/bdist.linux-x86_64/egg/dgl/optim/pytorch
copying build/lib.linux-x86_64-3.7/dgl/optim/pytorch/sparse_optim.py -> build/bdist.linux-x86_64/egg/dgl/optim/pytorch
creating build/bdist.linux-x86_64/egg/dgl/optim/mxnet
copying build/lib.linux-x86_64-3.7/dgl/optim/mxnet/__init__.py -> build/bdist.linux-x86_64/egg/dgl/optim/mxnet
copying build/lib.linux-x86_64-3.7/dgl/optim/__init__.py -> build/bdist.linux-x86_64/egg/dgl/optim
creating build/bdist.linux-x86_64/egg/dgl/optim/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/optim/tensorflow/__init__.py -> build/bdist.linux-x86_64/egg/dgl/optim/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/heterograph_index.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-3.7/dgl/sparse.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-3.7/dgl/global_config.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/dataloading
copying build/lib.linux-x86_64-3.7/dgl/dataloading/base.py -> build/bdist.linux-x86_64/egg/dgl/dataloading
copying build/lib.linux-x86_64-3.7/dgl/dataloading/dataloader.py -> build/bdist.linux-x86_64/egg/dgl/dataloading
copying build/lib.linux-x86_64-3.7/dgl/dataloading/graphsaint.py -> build/bdist.linux-x86_64/egg/dgl/dataloading
copying build/lib.linux-x86_64-3.7/dgl/dataloading/shadow.py -> build/bdist.linux-x86_64/egg/dgl/dataloading
copying build/lib.linux-x86_64-3.7/dgl/dataloading/negative_sampler.py -> build/bdist.linux-x86_64/egg/dgl/dataloading
copying build/lib.linux-x86_64-3.7/dgl/dataloading/__init__.py -> build/bdist.linux-x86_64/egg/dgl/dataloading
copying build/lib.linux-x86_64-3.7/dgl/dataloading/cluster_gcn.py -> build/bdist.linux-x86_64/egg/dgl/dataloading
copying build/lib.linux-x86_64-3.7/dgl/dataloading/neighbor_sampler.py -> build/bdist.linux-x86_64/egg/dgl/dataloading
copying build/lib.linux-x86_64-3.7/dgl/dataloading/dist_dataloader.py -> build/bdist.linux-x86_64/egg/dgl/dataloading
creating build/bdist.linux-x86_64/egg/dgl/_dataloading
copying build/lib.linux-x86_64-3.7/dgl/_dataloading/neighbor.py -> build/bdist.linux-x86_64/egg/dgl/_dataloading
copying build/lib.linux-x86_64-3.7/dgl/_dataloading/dataloader.py -> build/bdist.linux-x86_64/egg/dgl/_dataloading
creating build/bdist.linux-x86_64/egg/dgl/_dataloading/pytorch
copying build/lib.linux-x86_64-3.7/dgl/_dataloading/pytorch/dataloader.py -> build/bdist.linux-x86_64/egg/dgl/_dataloading/pytorch
copying build/lib.linux-x86_64-3.7/dgl/_dataloading/pytorch/__init__.py -> build/bdist.linux-x86_64/egg/dgl/_dataloading/pytorch
copying build/lib.linux-x86_64-3.7/dgl/_dataloading/shadow.py -> build/bdist.linux-x86_64/egg/dgl/_dataloading
copying build/lib.linux-x86_64-3.7/dgl/_dataloading/negative_sampler.py -> build/bdist.linux-x86_64/egg/dgl/_dataloading
copying build/lib.linux-x86_64-3.7/dgl/_dataloading/__init__.py -> build/bdist.linux-x86_64/egg/dgl/_dataloading
copying build/lib.linux-x86_64-3.7/dgl/_dataloading/cluster_gcn.py -> build/bdist.linux-x86_64/egg/dgl/_dataloading
copying build/lib.linux-x86_64-3.7/dgl/_api_internal.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-3.7/dgl/heterograph.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/backend
copying build/lib.linux-x86_64-3.7/dgl/backend/backend.py -> build/bdist.linux-x86_64/egg/dgl/backend
creating build/bdist.linux-x86_64/egg/dgl/backend/pytorch
copying build/lib.linux-x86_64-3.7/dgl/backend/pytorch/sparse.py -> build/bdist.linux-x86_64/egg/dgl/backend/pytorch
copying build/lib.linux-x86_64-3.7/dgl/backend/pytorch/tensor.py -> build/bdist.linux-x86_64/egg/dgl/backend/pytorch
copying build/lib.linux-x86_64-3.7/dgl/backend/pytorch/__init__.py -> build/bdist.linux-x86_64/egg/dgl/backend/pytorch
creating build/bdist.linux-x86_64/egg/dgl/backend/mxnet
copying build/lib.linux-x86_64-3.7/dgl/backend/mxnet/sparse.py -> build/bdist.linux-x86_64/egg/dgl/backend/mxnet
copying build/lib.linux-x86_64-3.7/dgl/backend/mxnet/tensor.py -> build/bdist.linux-x86_64/egg/dgl/backend/mxnet
copying build/lib.linux-x86_64-3.7/dgl/backend/mxnet/__init__.py -> build/bdist.linux-x86_64/egg/dgl/backend/mxnet
copying build/lib.linux-x86_64-3.7/dgl/backend/mxnet/sparse_optim.py -> build/bdist.linux-x86_64/egg/dgl/backend/mxnet
copying build/lib.linux-x86_64-3.7/dgl/backend/__init__.py -> build/bdist.linux-x86_64/egg/dgl/backend
creating build/bdist.linux-x86_64/egg/dgl/backend/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/backend/tensorflow/sparse.py -> build/bdist.linux-x86_64/egg/dgl/backend/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/backend/tensorflow/tensor.py -> build/bdist.linux-x86_64/egg/dgl/backend/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/backend/tensorflow/__init__.py -> build/bdist.linux-x86_64/egg/dgl/backend/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/backend/tensorflow/sparse_optim.py -> build/bdist.linux-x86_64/egg/dgl/backend/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/backend/set_default_backend.py -> build/bdist.linux-x86_64/egg/dgl/backend
copying build/lib.linux-x86_64-3.7/dgl/network.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/ops
copying build/lib.linux-x86_64-3.7/dgl/ops/edge_softmax.py -> build/bdist.linux-x86_64/egg/dgl/ops
copying build/lib.linux-x86_64-3.7/dgl/ops/sddmm.py -> build/bdist.linux-x86_64/egg/dgl/ops
copying build/lib.linux-x86_64-3.7/dgl/ops/gather_mm.py -> build/bdist.linux-x86_64/egg/dgl/ops
copying build/lib.linux-x86_64-3.7/dgl/ops/segment.py -> build/bdist.linux-x86_64/egg/dgl/ops
copying build/lib.linux-x86_64-3.7/dgl/ops/__init__.py -> build/bdist.linux-x86_64/egg/dgl/ops
copying build/lib.linux-x86_64-3.7/dgl/ops/spmm.py -> build/bdist.linux-x86_64/egg/dgl/ops
copying build/lib.linux-x86_64-3.7/dgl/logging.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-3.7/dgl/graph_index.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/cuda
copying build/lib.linux-x86_64-3.7/dgl/cuda/nccl.py -> build/bdist.linux-x86_64/egg/dgl/cuda
copying build/lib.linux-x86_64-3.7/dgl/cuda/__init__.py -> build/bdist.linux-x86_64/egg/dgl/cuda
copying build/lib.linux-x86_64-3.7/dgl/udf.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/storages
copying build/lib.linux-x86_64-3.7/dgl/storages/pytorch_tensor.py -> build/bdist.linux-x86_64/egg/dgl/storages
copying build/lib.linux-x86_64-3.7/dgl/storages/base.py -> build/bdist.linux-x86_64/egg/dgl/storages
copying build/lib.linux-x86_64-3.7/dgl/storages/tensor.py -> build/bdist.linux-x86_64/egg/dgl/storages
copying build/lib.linux-x86_64-3.7/dgl/storages/__init__.py -> build/bdist.linux-x86_64/egg/dgl/storages
copying build/lib.linux-x86_64-3.7/dgl/storages/numpy.py -> build/bdist.linux-x86_64/egg/dgl/storages
copying build/lib.linux-x86_64-3.7/dgl/ndarray.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-3.7/dgl/init.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-3.7/dgl/view.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/sampling
copying build/lib.linux-x86_64-3.7/dgl/sampling/neighbor.py -> build/bdist.linux-x86_64/egg/dgl/sampling
copying build/lib.linux-x86_64-3.7/dgl/sampling/pinsage.py -> build/bdist.linux-x86_64/egg/dgl/sampling
copying build/lib.linux-x86_64-3.7/dgl/sampling/negative.py -> build/bdist.linux-x86_64/egg/dgl/sampling
copying build/lib.linux-x86_64-3.7/dgl/sampling/randomwalks.py -> build/bdist.linux-x86_64/egg/dgl/sampling
copying build/lib.linux-x86_64-3.7/dgl/sampling/utils.py -> build/bdist.linux-x86_64/egg/dgl/sampling
copying build/lib.linux-x86_64-3.7/dgl/sampling/node2vec_randomwalk.py -> build/bdist.linux-x86_64/egg/dgl/sampling
copying build/lib.linux-x86_64-3.7/dgl/sampling/__init__.py -> build/bdist.linux-x86_64/egg/dgl/sampling
copying build/lib.linux-x86_64-3.7/dgl/generators.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/qm7b.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/graph_serialize.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/dgl_dataset.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/csv_dataset_base.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/knowledge_graph.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/qm9_edge.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/gnn_benchmark.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/qm9.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/icews18.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/gdelt.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/utils.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/sbm.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/csv_dataset.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/karate.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/citation_graph.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/rdf.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/bitcoinotc.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/ppi.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/tensor_serialize.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/flickr.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/wikics.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/heterograph_serialize.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/tu.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/__init__.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/gindt.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/tree.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/synthetic.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/adapter.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/fakenews.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/fraud.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/minigc.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/yelp.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/reddit.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/convert.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-3.7/dgl/container.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-3.7/dgl/merge.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-3.7/dgl/readout.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-3.7/dgl/__init__.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/dist_graph.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/graph_partition_book.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/partition.py -> build/bdist.linux-x86_64/egg/dgl/distributed
creating build/bdist.linux-x86_64/egg/dgl/distributed/nn
creating build/bdist.linux-x86_64/egg/dgl/distributed/nn/pytorch
copying build/lib.linux-x86_64-3.7/dgl/distributed/nn/pytorch/sparse_emb.py -> build/bdist.linux-x86_64/egg/dgl/distributed/nn/pytorch
copying build/lib.linux-x86_64-3.7/dgl/distributed/nn/pytorch/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distributed/nn/pytorch
creating build/bdist.linux-x86_64/egg/dgl/distributed/nn/mxnet
copying build/lib.linux-x86_64-3.7/dgl/distributed/nn/mxnet/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distributed/nn/mxnet
copying build/lib.linux-x86_64-3.7/dgl/distributed/nn/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distributed/nn
creating build/bdist.linux-x86_64/egg/dgl/distributed/nn/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/distributed/nn/tensorflow/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distributed/nn/tensorflow
creating build/bdist.linux-x86_64/egg/dgl/distributed/optim
creating build/bdist.linux-x86_64/egg/dgl/distributed/optim/pytorch
copying build/lib.linux-x86_64-3.7/dgl/distributed/optim/pytorch/utils.py -> build/bdist.linux-x86_64/egg/dgl/distributed/optim/pytorch
copying build/lib.linux-x86_64-3.7/dgl/distributed/optim/pytorch/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distributed/optim/pytorch
copying build/lib.linux-x86_64-3.7/dgl/distributed/optim/pytorch/sparse_optim.py -> build/bdist.linux-x86_64/egg/dgl/distributed/optim/pytorch
creating build/bdist.linux-x86_64/egg/dgl/distributed/optim/mxnet
copying build/lib.linux-x86_64-3.7/dgl/distributed/optim/mxnet/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distributed/optim/mxnet
copying build/lib.linux-x86_64-3.7/dgl/distributed/optim/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distributed/optim
creating build/bdist.linux-x86_64/egg/dgl/distributed/optim/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/distributed/optim/tensorflow/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distributed/optim/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/distributed/standalone_kvstore.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/graph_services.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/role.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/constants.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/rpc_server.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/kvstore.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/dist_context.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/rpc.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/rpc_client.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/id_map.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/server_state.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/shared_mem_utils.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/dist_tensor.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/dist_dataloader.py -> build/bdist.linux-x86_64/egg/dgl/distributed
creating build/bdist.linux-x86_64/egg/dgl/_deprecate
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/nodeflow.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/udf.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/view.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/kernel.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/__init__.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate
creating build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/spmv.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/degree_bucketing.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/__init__.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/scheduler.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/adapter.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/runtime.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime
creating build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/ir
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/ir/program.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/ir
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/ir/executor.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/ir
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/ir/var.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/ir
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/ir/registry.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/ir
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/ir/__init__.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/ir
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/graph.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/frame.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate
copying build/lib.linux-x86_64-3.7/dgl/propagate.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-3.7/dgl/batch.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/transforms
copying build/lib.linux-x86_64-3.7/dgl/transforms/functional.py -> build/bdist.linux-x86_64/egg/dgl/transforms
copying build/lib.linux-x86_64-3.7/dgl/transforms/__init__.py -> build/bdist.linux-x86_64/egg/dgl/transforms
copying build/lib.linux-x86_64-3.7/dgl/transforms/module.py -> build/bdist.linux-x86_64/egg/dgl/transforms
creating build/bdist.linux-x86_64/egg/dgl/distgnn
creating build/bdist.linux-x86_64/egg/dgl/distgnn/partition
copying build/lib.linux-x86_64-3.7/dgl/distgnn/partition/libra_partition.py -> build/bdist.linux-x86_64/egg/dgl/distgnn/partition
copying build/lib.linux-x86_64-3.7/dgl/distgnn/partition/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distgnn/partition
creating build/bdist.linux-x86_64/egg/dgl/distgnn/tools
copying build/lib.linux-x86_64-3.7/dgl/distgnn/tools/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distgnn/tools
copying build/lib.linux-x86_64-3.7/dgl/distgnn/tools/tools.py -> build/bdist.linux-x86_64/egg/dgl/distgnn/tools
copying build/lib.linux-x86_64-3.7/dgl/distgnn/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distgnn
creating build/bdist.linux-x86_64/egg/dgl/_ffi
copying build/lib.linux-x86_64-3.7/dgl/_ffi/runtime_ctypes.py -> build/bdist.linux-x86_64/egg/dgl/_ffi
copying build/lib.linux-x86_64-3.7/dgl/_ffi/function.py -> build/bdist.linux-x86_64/egg/dgl/_ffi
copying build/lib.linux-x86_64-3.7/dgl/_ffi/libinfo.py -> build/bdist.linux-x86_64/egg/dgl/_ffi
copying build/lib.linux-x86_64-3.7/dgl/_ffi/streams.py -> build/bdist.linux-x86_64/egg/dgl/_ffi
copying build/lib.linux-x86_64-3.7/dgl/_ffi/base.py -> build/bdist.linux-x86_64/egg/dgl/_ffi
creating build/bdist.linux-x86_64/egg/dgl/_ffi/_ctypes
copying build/lib.linux-x86_64-3.7/dgl/_ffi/_ctypes/function.py -> build/bdist.linux-x86_64/egg/dgl/_ffi/_ctypes
copying build/lib.linux-x86_64-3.7/dgl/_ffi/_ctypes/types.py -> build/bdist.linux-x86_64/egg/dgl/_ffi/_ctypes
copying build/lib.linux-x86_64-3.7/dgl/_ffi/_ctypes/ndarray.py -> build/bdist.linux-x86_64/egg/dgl/_ffi/_ctypes
copying build/lib.linux-x86_64-3.7/dgl/_ffi/_ctypes/__init__.py -> build/bdist.linux-x86_64/egg/dgl/_ffi/_ctypes
copying build/lib.linux-x86_64-3.7/dgl/_ffi/_ctypes/object.py -> build/bdist.linux-x86_64/egg/dgl/_ffi/_ctypes
copying build/lib.linux-x86_64-3.7/dgl/_ffi/ndarray.py -> build/bdist.linux-x86_64/egg/dgl/_ffi
copying build/lib.linux-x86_64-3.7/dgl/_ffi/object_generic.py -> build/bdist.linux-x86_64/egg/dgl/_ffi
copying build/lib.linux-x86_64-3.7/dgl/_ffi/__init__.py -> build/bdist.linux-x86_64/egg/dgl/_ffi
copying build/lib.linux-x86_64-3.7/dgl/_ffi/object.py -> build/bdist.linux-x86_64/egg/dgl/_ffi
creating build/bdist.linux-x86_64/egg/dgl/_ffi/_cy3
copying build/lib.linux-x86_64-3.7/dgl/_ffi/_cy3/__init__.py -> build/bdist.linux-x86_64/egg/dgl/_ffi/_cy3
copying build/lib.linux-x86_64-3.7/dgl/_ffi/_cy3/core.cpython-37m-x86_64-linux-gnu.so -> build/bdist.linux-x86_64/egg/dgl/_ffi/_cy3
creating build/bdist.linux-x86_64/egg/dgl/_ffi/_cy2
copying build/lib.linux-x86_64-3.7/dgl/_ffi/_cy2/__init__.py -> build/bdist.linux-x86_64/egg/dgl/_ffi/_cy2
creating build/bdist.linux-x86_64/egg/dgl/_ffi/_cython
copying build/lib.linux-x86_64-3.7/dgl/_ffi/_cython/core.cpp -> build/bdist.linux-x86_64/egg/dgl/_ffi/_cython
copying build/lib.linux-x86_64-3.7/dgl/frame.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-3.7/dgl/random.py -> build/bdist.linux-x86_64/egg/dgl
byte-compiling build/bdist.linux-x86_64/egg/dgl/mock_sparse/sp_matrix.py to sp_matrix.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/mock_sparse/diag_matrix.py to diag_matrix.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/mock_sparse/reduction.py to reduction.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/mock_sparse/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/mock_sparse/elementwise_op_sp.py to elementwise_op_sp.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/geometry/edge_coarsening.py to edge_coarsening.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/geometry/fps.py to fps.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/geometry/capi.py to capi.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/geometry/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/function/base.py to base.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/function/message.py to message.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/function/reducer.py to reducer.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/function/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/partition.py to partition.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/glob.py to glob.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/link/transe.py to transe.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/link/edgepred.py to edgepred.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/link/transr.py to transr.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/link/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/hetero.py to hetero.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/sparse_emb.py to sparse_emb.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/linear.py to linear.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/utils.py to utils.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/agnnconv.py to agnnconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/relgraphconv.py to relgraphconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/appnpconv.py to appnpconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/dgnconv.py to dgnconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/grouprevres.py to grouprevres.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/twirlsconv.py to twirlsconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/egatconv.py to egatconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/graphconv.py to graphconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/densechebconv.py to densechebconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/gcn2conv.py to gcn2conv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/ginconv.py to ginconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/dotgatconv.py to dotgatconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/cfconv.py to cfconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/gmmconv.py to gmmconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/sageconv.py to sageconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/densesageconv.py to densesageconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/gineconv.py to gineconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/edgeconv.py to edgeconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/nnconv.py to nnconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/gatv2conv.py to gatv2conv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/hgtconv.py to hgtconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/pnaconv.py to pnaconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/gatedgraphconv.py to gatedgraphconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/atomicconv.py to atomicconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/tagconv.py to tagconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/densegraphconv.py to densegraphconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/egnnconv.py to egnnconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/chebconv.py to chebconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/sgconv.py to sgconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/gatconv.py to gatconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/explain/gnnexplainer.py to gnnexplainer.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/explain/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/factory.py to factory.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/softmax.py to softmax.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/glob.py to glob.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/hetero.py to hetero.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/utils.py to utils.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/agnnconv.py to agnnconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/relgraphconv.py to relgraphconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/appnpconv.py to appnpconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/graphconv.py to graphconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/densechebconv.py to densechebconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/ginconv.py to ginconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/gmmconv.py to gmmconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/sageconv.py to sageconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/densesageconv.py to densesageconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/edgeconv.py to edgeconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/nnconv.py to nnconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/gatedgraphconv.py to gatedgraphconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/tagconv.py to tagconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/densegraphconv.py to densegraphconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/chebconv.py to chebconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/sgconv.py to sgconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/gatconv.py to gatconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/softmax.py to softmax.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/glob.py to glob.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/hetero.py to hetero.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/utils.py to utils.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv/relgraphconv.py to relgraphconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv/appnpconv.py to appnpconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv/graphconv.py to graphconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv/densechebconv.py to densechebconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv/ginconv.py to ginconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv/sageconv.py to sageconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv/edgeconv.py to edgeconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv/chebconv.py to chebconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv/sgconv.py to sgconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv/gatconv.py to gatconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/softmax.py to softmax.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/functional/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/core.py to core.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/subgraph.py to subgraph.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/contrib/dis_kvstore.py to dis_kvstore.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/contrib/unified_tensor.py to unified_tensor.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/contrib/graph_store.py to graph_store.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/contrib/sampling/sampler.py to sampler.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/contrib/sampling/dis_sampler.py to dis_sampler.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/contrib/sampling/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/contrib/data/knowledge_graph.py to knowledge_graph.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/contrib/data/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/contrib/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/traversal.py to traversal.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/utils/checks.py to checks.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/utils/exception.py to exception.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/utils/pin_memory.py to pin_memory.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/utils/internal.py to internal.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/utils/filter.py to filter.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/utils/data.py to data.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/utils/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/utils/shared_mem.py to shared_mem.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/multiprocessing/pytorch.py to pytorch.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/multiprocessing/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/base.py to base.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/optim/pytorch/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/optim/pytorch/sparse_optim.py to sparse_optim.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/optim/mxnet/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/optim/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/optim/tensorflow/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/heterograph_index.py to heterograph_index.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/sparse.py to sparse.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/global_config.py to global_config.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/dataloading/base.py to base.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/dataloading/dataloader.py to dataloader.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/dataloading/graphsaint.py to graphsaint.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/dataloading/shadow.py to shadow.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/dataloading/negative_sampler.py to negative_sampler.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/dataloading/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/dataloading/cluster_gcn.py to cluster_gcn.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/dataloading/neighbor_sampler.py to neighbor_sampler.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/dataloading/dist_dataloader.py to dist_dataloader.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_dataloading/neighbor.py to neighbor.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_dataloading/dataloader.py to dataloader.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_dataloading/pytorch/dataloader.py to dataloader.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_dataloading/pytorch/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_dataloading/shadow.py to shadow.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_dataloading/negative_sampler.py to negative_sampler.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_dataloading/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_dataloading/cluster_gcn.py to cluster_gcn.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_api_internal.py to _api_internal.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/heterograph.py to heterograph.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/backend.py to backend.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/pytorch/sparse.py to sparse.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/pytorch/tensor.py to tensor.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/pytorch/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/mxnet/sparse.py to sparse.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/mxnet/tensor.py to tensor.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/mxnet/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/mxnet/sparse_optim.py to sparse_optim.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/tensorflow/sparse.py to sparse.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/tensorflow/tensor.py to tensor.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/tensorflow/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/tensorflow/sparse_optim.py to sparse_optim.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/set_default_backend.py to set_default_backend.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/network.py to network.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/ops/edge_softmax.py to edge_softmax.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/ops/sddmm.py to sddmm.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/ops/gather_mm.py to gather_mm.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/ops/segment.py to segment.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/ops/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/ops/spmm.py to spmm.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/logging.py to logging.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/graph_index.py to graph_index.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/cuda/nccl.py to nccl.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/cuda/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/udf.py to udf.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/storages/pytorch_tensor.py to pytorch_tensor.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/storages/base.py to base.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/storages/tensor.py to tensor.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/storages/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/storages/numpy.py to numpy.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/ndarray.py to ndarray.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/init.py to init.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/view.py to view.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/sampling/neighbor.py to neighbor.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/sampling/pinsage.py to pinsage.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/sampling/negative.py to negative.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/sampling/randomwalks.py to randomwalks.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/sampling/utils.py to utils.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/sampling/node2vec_randomwalk.py to node2vec_randomwalk.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/sampling/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/generators.py to generators.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/qm7b.py to qm7b.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/graph_serialize.py to graph_serialize.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/dgl_dataset.py to dgl_dataset.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/csv_dataset_base.py to csv_dataset_base.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/knowledge_graph.py to knowledge_graph.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/qm9_edge.py to qm9_edge.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/gnn_benchmark.py to gnn_benchmark.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/qm9.py to qm9.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/icews18.py to icews18.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/gdelt.py to gdelt.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/utils.py to utils.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/sbm.py to sbm.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/csv_dataset.py to csv_dataset.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/karate.py to karate.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/citation_graph.py to citation_graph.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/rdf.py to rdf.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/bitcoinotc.py to bitcoinotc.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/ppi.py to ppi.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/tensor_serialize.py to tensor_serialize.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/flickr.py to flickr.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/wikics.py to wikics.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/heterograph_serialize.py to heterograph_serialize.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/tu.py to tu.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/gindt.py to gindt.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/tree.py to tree.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/synthetic.py to synthetic.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/adapter.py to adapter.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/fakenews.py to fakenews.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/fraud.py to fraud.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/minigc.py to minigc.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/yelp.py to yelp.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/reddit.py to reddit.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/convert.py to convert.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/container.py to container.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/merge.py to merge.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/readout.py to readout.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/dist_graph.py to dist_graph.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/graph_partition_book.py to graph_partition_book.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/partition.py to partition.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/nn/pytorch/sparse_emb.py to sparse_emb.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/nn/pytorch/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/nn/mxnet/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/nn/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/nn/tensorflow/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/optim/pytorch/utils.py to utils.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/optim/pytorch/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/optim/pytorch/sparse_optim.py to sparse_optim.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/optim/mxnet/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/optim/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/optim/tensorflow/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/standalone_kvstore.py to standalone_kvstore.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/graph_services.py to graph_services.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/role.py to role.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/constants.py to constants.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/rpc_server.py to rpc_server.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/kvstore.py to kvstore.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/dist_context.py to dist_context.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/rpc.py to rpc.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/rpc_client.py to rpc_client.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/id_map.py to id_map.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/server_state.py to server_state.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/shared_mem_utils.py to shared_mem_utils.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/dist_tensor.py to dist_tensor.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/dist_dataloader.py to dist_dataloader.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/nodeflow.py to nodeflow.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/udf.py to udf.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/view.py to view.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/kernel.py to kernel.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/spmv.py to spmv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/degree_bucketing.py to degree_bucketing.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/scheduler.py to scheduler.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/adapter.py to adapter.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/runtime.py to runtime.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/ir/program.py to program.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/ir/executor.py to executor.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/ir/var.py to var.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/ir/registry.py to registry.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/ir/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/graph.py to graph.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/frame.py to frame.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/propagate.py to propagate.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/batch.py to batch.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/transforms/functional.py to functional.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/transforms/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/transforms/module.py to module.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distgnn/partition/libra_partition.py to libra_partition.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distgnn/partition/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distgnn/tools/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distgnn/tools/tools.py to tools.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distgnn/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/runtime_ctypes.py to runtime_ctypes.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/function.py to function.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/libinfo.py to libinfo.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/streams.py to streams.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/base.py to base.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/_ctypes/function.py to function.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/_ctypes/types.py to types.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/_ctypes/ndarray.py to ndarray.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/_ctypes/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/_ctypes/object.py to object.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/ndarray.py to ndarray.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/object_generic.py to object_generic.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/object.py to object.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/_cy3/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/_cy2/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/frame.py to frame.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/random.py to random.cpython-37.pyc
creating stub loader for dgl/_ffi/_cy3/core.cpython-37m-x86_64-linux-gnu.so
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/_cy3/core.py to core.cpython-37.pyc
installing package data to build/bdist.linux-x86_64/egg
running install_data
copying ../build/libdgl.so -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/tensoradapter
creating build/bdist.linux-x86_64/egg/dgl/tensoradapter/pytorch
copying ../build/tensoradapter/pytorch/libtensoradapter_pytorch_1.9.0.so -> build/bdist.linux-x86_64/egg/dgl/tensoradapter/pytorch
creating build/bdist.linux-x86_64/egg/EGG-INFO
copying dgl.egg-info/PKG-INFO -> build/bdist.linux-x86_64/egg/EGG-INFO
copying dgl.egg-info/SOURCES.txt -> build/bdist.linux-x86_64/egg/EGG-INFO
copying dgl.egg-info/dependency_links.txt -> build/bdist.linux-x86_64/egg/EGG-INFO
copying dgl.egg-info/not-zip-safe -> build/bdist.linux-x86_64/egg/EGG-INFO
copying dgl.egg-info/requires.txt -> build/bdist.linux-x86_64/egg/EGG-INFO
copying dgl.egg-info/top_level.txt -> build/bdist.linux-x86_64/egg/EGG-INFO
writing build/bdist.linux-x86_64/egg/EGG-INFO/native_libs.txt
creating dist
creating 'dist/dgl-0.9-py3.7-linux-x86_64.egg' and adding 'build/bdist.linux-x86_64/egg' to it
removing 'build/bdist.linux-x86_64/egg' (and everything under it)
Processing dgl-0.9-py3.7-linux-x86_64.egg
creating /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/dgl-0.9-py3.7-linux-x86_64.egg
Extracting dgl-0.9-py3.7-linux-x86_64.egg to /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages
Adding dgl 0.9 to easy-install.pth file

Installed /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/dgl-0.9-py3.7-linux-x86_64.egg
Processing dependencies for dgl==0.9
Searching for psutil>=5.8.0
Reading https://pypi.org/simple/psutil/
Downloading https://files.pythonhosted.org/packages/3d/73/d8c87b5612c58d1e6c6d91997c1590771d34e4ee27d9c11eb1e64ecbf365/psutil-5.9.2-cp37-cp37m-manylinux_2_12_x86_64.manylinux2010_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl#sha256=4fb54941aac044a61db9d8eb56fc5bee207db3bc58645d657249030e15ba3727
Best match: psutil 5.9.2
Processing psutil-5.9.2-cp37-cp37m-manylinux_2_12_x86_64.manylinux2010_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl
Installing psutil-5.9.2-cp37-cp37m-manylinux_2_12_x86_64.manylinux2010_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl to /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages
Adding psutil 5.9.2 to easy-install.pth file

Installed /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/psutil-5.9.2-py3.7-linux-x86_64.egg
Searching for tqdm==4.64.0
Best match: tqdm 4.64.0
Adding tqdm 4.64.0 to easy-install.pth file
Installing tqdm script to /opt/conda/envs/tensorflow-ci/bin

Using /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages
Searching for requests==2.28.1
Best match: requests 2.28.1
Adding requests 2.28.1 to easy-install.pth file

Using /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages
Searching for networkx==2.6.3
Best match: networkx 2.6.3
Adding networkx 2.6.3 to easy-install.pth file

Using /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages
Searching for scipy==1.4.1
Best match: scipy 1.4.1
Adding scipy 1.4.1 to easy-install.pth file

Using /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages
Searching for numpy==1.18.5
Best match: numpy 1.18.5
Adding numpy 1.18.5 to easy-install.pth file
Installing f2py script to /opt/conda/envs/tensorflow-ci/bin
Installing f2py3 script to /opt/conda/envs/tensorflow-ci/bin
Installing f2py3.7 script to /opt/conda/envs/tensorflow-ci/bin

Using /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages
Searching for certifi==2022.6.15
Best match: certifi 2022.6.15
Adding certifi 2022.6.15 to easy-install.pth file

Using /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages
Searching for urllib3==1.26.11
Best match: urllib3 1.26.11
Adding urllib3 1.26.11 to easy-install.pth file

Using /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages
Searching for idna==3.3
Best match: idna 3.3
Adding idna 3.3 to easy-install.pth file

Using /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages
Searching for charset-normalizer==2.1.0
Best match: charset-normalizer 2.1.0
Adding charset-normalizer 2.1.0 to easy-install.pth file
Installing normalizer script to /opt/conda/envs/tensorflow-ci/bin

Using /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages
Finished processing dependencies for dgl==0.9
/opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/Cython/Compiler/Main.py:369: FutureWarning: Cython directive 'language_level' not set, using 2 for now (Py2). This will change in a later release! File: /root/jenkins/workspace/dgl_PR-4648/python/dgl/_ffi/_cython/core.pyx
  tree = Parsing.p_module(s, pxd, full_module_name)
/opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/setuptools/command/install.py:37: SetuptoolsDeprecationWarning: setup.py install is deprecated. Use build and pip and other standards-based tools.
  setuptools.SetuptoolsDeprecationWarning,
/opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/setuptools/command/easy_install.py:147: EasyInstallDeprecationWarning: easy_install command is deprecated. Use build and pip and other standards-based tools.
  EasyInstallDeprecationWarning,
/opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/pkg_resources/__init__.py:125: PkgResourcesDeprecationWarning:  is an invalid version and will not be supported in a future release
  PkgResourcesDeprecationWarning,
[1/1] Cythonizing dgl/_ffi/_cython/core.pyx
running build_ext
building 'dgl._ffi._cy3.core' extension
gcc -pthread -B /opt/conda/envs/tensorflow-ci/compiler_compat -Wl,--sysroot=/ -Wsign-compare -DNDEBUG -g -fwrapv -O3 -Wall -Wstrict-prototypes -fPIC -I../include/ -I../third_party/dmlc-core/include -I../third_party/dlpack/include -I/opt/conda/envs/tensorflow-ci/include/python3.7m -c dgl/_ffi/_cython/core.cpp -o build/temp.linux-x86_64-3.7/dgl/_ffi/_cython/core.o
cc1plus: warning: command line option ‘-Wstrict-prototypes’ is valid for C/ObjC but not for C++
g++ -pthread -B /opt/conda/envs/tensorflow-ci/compiler_compat -Wl,--sysroot=/ -pthread -shared -B /opt/conda/envs/tensorflow-ci/compiler_compat -L/opt/conda/envs/tensorflow-ci/lib -Wl,-rpath=/opt/conda/envs/tensorflow-ci/lib -Wl,--no-as-needed -Wl,--sysroot=/ build/temp.linux-x86_64-3.7/dgl/_ffi/_cython/core.o -Ldgl -L../build/Release -L../build -ldgl -o build/lib.linux-x86_64-3.7/dgl/_ffi/_cy3/core.cpython-37m-x86_64-linux-gnu.so
/opt/conda/envs/tensorflow-ci/compiler_compat/ld: warning: /opt/conda/envs/tensorflow-ci/lib/libstdc++.so: unsupported GNU_PROPERTY_TYPE (5) type: 0xc0010001
/opt/conda/envs/tensorflow-ci/compiler_compat/ld: warning: /opt/conda/envs/tensorflow-ci/lib/libstdc++.so: unsupported GNU_PROPERTY_TYPE (5) type: 0xc0010002
/opt/conda/envs/tensorflow-ci/compiler_compat/ld: warning: /opt/conda/envs/tensorflow-ci/lib/libgcc_s.so.1: unsupported GNU_PROPERTY_TYPE (5) type: 0xc0010001
/opt/conda/envs/tensorflow-ci/compiler_compat/ld: warning: /opt/conda/envs/tensorflow-ci/lib/libgcc_s.so.1: unsupported GNU_PROPERTY_TYPE (5) type: 0xc0010002
/opt/conda/envs/tensorflow-ci/compiler_compat/ld: warning: /opt/conda/envs/tensorflow-ci/lib/libgcc_s.so.1: unsupported GNU_PROPERTY_TYPE (5) type: 0xc0010001
/opt/conda/envs/tensorflow-ci/compiler_compat/ld: warning: /opt/conda/envs/tensorflow-ci/lib/libgcc_s.so.1: unsupported GNU_PROPERTY_TYPE (5) type: 0xc0010002
copying build/lib.linux-x86_64-3.7/dgl/_ffi/_cy3/core.cpython-37m-x86_64-linux-gnu.so -> dgl/_ffi/_cy3
/opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/Cython/Compiler/Main.py:369: FutureWarning: Cython directive 'language_level' not set, using 2 for now (Py2). This will change in a later release! File: /root/jenkins/workspace/dgl_PR-4648/python/dgl/_ffi/_cython/core.pyx
  tree = Parsing.p_module(s, pxd, full_module_name)
~/jenkins/workspace/dgl_PR-4648
[Pipeline] sh
+ ls -lh /usr/lib/x86_64-linux-gnu/
total 654M
-rw-r--r--  1 root root   496 May  3 10:19 Mcrt1.o
-rw-r--r--  1 root root  1.8K May  3 10:19 Scrt1.o
drwxr-xr-x  2 root root  4.0K Aug  1 13:21 audit
drwxr-xr-x  2 root root  4.0K Aug 16 08:41 avahi
drwxr-xr-x  2 root root  4.0K Aug  1 13:21 coreutils
-rw-r--r--  1 root root  1.9K May  3 10:19 crt1.o
-rw-r--r--  1 root root  1.2K May  3 10:19 crti.o
-rw-r--r--  1 root root   648 May  3 10:19 crtn.o
drwxr-xr-x  2 root root  4.0K Aug 16 08:41 dri
drwxr-xr-x  2 root root  4.0K Aug 16 08:40 engines-1.1
drwxr-xr-x  2 root root   12K Aug  1 13:21 gconv
-rw-r--r--  1 root root  2.5K May  3 10:19 gcrt1.o
drwxr-xr-x  2 root root  4.0K Aug 16 08:41 gdcm-2.8
drwxr-xr-x  3 root root  4.0K Aug 16 08:40 gdk-pixbuf-2.0
drwxr-xr-x  3 root root  4.0K Aug 16 08:40 gio
drwxr-xr-x  2 root root  4.0K Aug 16 08:41 girepository-1.0
drwxr-xr-x  3 root root  4.0K Aug 16 08:41 glib-2.0
drwxr-xr-x  2 root root  4.0K Aug 16 08:41 graphviz
-rw-r--r--  1 root root  2.3K May  3 10:19 grcrt1.o
drwxr-xr-x  3 root root  4.0K Aug 16 08:41 gtk-2.0
drwxr-xr-x  3 root root  4.0K Aug 16 08:41 gtk-3.0
drwxr-xr-x  2 root root  4.0K Aug 16 08:41 hwloc
drwxr-xr-x  3 root root  4.0K Aug 16 08:41 icu
drwxr-xr-x  3 root root  4.0K Aug 16 08:40 krb5
drwxr-xr-x  2 root root  4.0K Aug 16 08:40 ldscripts
-rw-r--r--  1 root root  2.0K May  3 10:19 libBrokenLocale.a
lrwxrwxrwx  1 root root    42 May  3 10:19 libBrokenLocale.so -> /lib/x86_64-linux-gnu/libBrokenLocale.so.1
lrwxrwxrwx  1 root root    16 May 23  2016 libCharLS.so.1 -> libCharLS.so.1.0
-rw-r--r--  1 root root  219K May 23  2016 libCharLS.so.1.0
lrwxrwxrwx  1 root root    14 May 10  2019 libGL.so.1 -> libGL.so.1.0.0
-rw-r--r--  1 root root  555K May 10  2019 libGL.so.1.0.0
lrwxrwxrwx  1 root root    15 May 10  2019 libGLX.so.0 -> libGLX.so.0.0.0
-rw-r--r--  1 root root   67K May 10  2019 libGLX.so.0.0.0
lrwxrwxrwx  1 root root    16 Jun 12  2020 libGLX_indirect.so.0 -> libGLX_mesa.so.0
lrwxrwxrwx  1 root root    20 Jun 12  2020 libGLX_mesa.so.0 -> libGLX_mesa.so.0.0.0
-rw-r--r--  1 root root  477K Jun 12  2020 libGLX_mesa.so.0.0.0
lrwxrwxrwx  1 root root    22 May 10  2019 libGLdispatch.so.0 -> libGLdispatch.so.0.0.0
-rw-r--r--  1 root root  599K May 10  2019 libGLdispatch.so.0.0.0
lrwxrwxrwx  1 root root    17 Apr 16  2016 libHalf.so -> libHalf.so.12.0.0
lrwxrwxrwx  1 root root    17 Apr 16  2016 libHalf.so.12 -> libHalf.so.12.0.0
-rw-r--r--  1 root root  267K Apr 16  2016 libHalf.so.12.0.0
lrwxrwxrwx  1 root root    15 Feb 28  2017 libICE.so.6 -> libICE.so.6.3.0
-rw-r--r--  1 root root   96K Feb 28  2017 libICE.so.6.3.0
lrwxrwxrwx  1 root root    20 Apr 16  2016 libIex-2_2.so.12 -> libIex-2_2.so.12.0.0
-rw-r--r--  1 root root  119K Apr 16  2016 libIex-2_2.so.12.0.0
lrwxrwxrwx  1 root root    20 Apr 16  2016 libIex.so -> libIex-2_2.so.12.0.0
lrwxrwxrwx  1 root root    24 Apr 16  2016 libIexMath-2_2.so.12 -> libIexMath-2_2.so.12.0.0
-rw-r--r--  1 root root   15K Apr 16  2016 libIexMath-2_2.so.12.0.0
lrwxrwxrwx  1 root root    24 Apr 16  2016 libIexMath.so -> libIexMath-2_2.so.12.0.0
lrwxrwxrwx  1 root root    23 Nov 16  2021 libIlmImf-2_2.so.22 -> libIlmImf-2_2.so.22.0.0
-rw-r--r--  1 root root  2.8M Nov 16  2021 libIlmImf-2_2.so.22.0.0
-rw-r--r--  1 root root  4.4M Nov 16  2021 libIlmImf.a
lrwxrwxrwx  1 root root    23 Nov 16  2021 libIlmImf.so -> libIlmImf-2_2.so.22.0.0
lrwxrwxrwx  1 root root    27 Nov 16  2021 libIlmImfUtil-2_2.so.22 -> libIlmImfUtil-2_2.so.22.0.0
-rw-r--r--  1 root root  140K Nov 16  2021 libIlmImfUtil-2_2.so.22.0.0
-rw-r--r--  1 root root  319K Nov 16  2021 libIlmImfUtil.a
lrwxrwxrwx  1 root root    27 Nov 16  2021 libIlmImfUtil.so -> libIlmImfUtil-2_2.so.22.0.0
lrwxrwxrwx  1 root root    26 Apr 16  2016 libIlmThread-2_2.so.12 -> libIlmThread-2_2.so.12.0.0
-rw-r--r--  1 root root   27K Apr 16  2016 libIlmThread-2_2.so.12.0.0
lrwxrwxrwx  1 root root    26 Apr 16  2016 libIlmThread.so -> libIlmThread-2_2.so.12.0.0
lrwxrwxrwx  1 root root    22 Apr 16  2016 libImath-2_2.so.12 -> libImath-2_2.so.12.0.0
-rw-r--r--  1 root root   71K Apr 16  2016 libImath-2_2.so.12.0.0
lrwxrwxrwx  1 root root    22 Apr 16  2016 libImath.so -> libImath-2_2.so.12.0.0
lrwxrwxrwx  1 root root    15 Jul 14  2020 libLLVM-10.so -> libLLVM-10.so.1
-rw-r--r--  1 root root   71M Jul 14  2020 libLLVM-10.so.1
lrwxrwxrwx  1 root root    18 Apr  5  2017 libOpenCL.so.1 -> libOpenCL.so.1.0.0
-rw-r--r--  1 root root   43K Apr  5  2017 libOpenCL.so.1.0.0
lrwxrwxrwx  1 root root    14 Jul 13  2014 libSM.so.6 -> libSM.so.6.0.1
-rw-r--r--  1 root root   31K Jul 13  2014 libSM.so.6.0.1
lrwxrwxrwx  1 root root    19 May 19  2021 libX11-xcb.so.1 -> libX11-xcb.so.1.0.0
-rw-r--r--  1 root root  5.7K May 19  2021 libX11-xcb.so.1.0.0
lrwxrwxrwx  1 root root    15 May 19  2021 libX11.so.6 -> libX11.so.6.3.0
-rw-r--r--  1 root root  1.3M May 19  2021 libX11.so.6.3.0
lrwxrwxrwx  1 root root    15 Apr 21  2020 libXau.so.6 -> libXau.so.6.0.0
-rw-r--r--  1 root root   15K Apr 21  2020 libXau.so.6.0.0
lrwxrwxrwx  1 root root    12 Aug  3  2015 libXaw.so.7 -> libXaw7.so.7
lrwxrwxrwx  1 root root    16 Aug  3  2015 libXaw7.so.7 -> libXaw7.so.7.0.0
-rw-r--r--  1 root root  461K Aug  3  2015 libXaw7.so.7.0.0
lrwxrwxrwx  1 root root    22 Jan 20  2017 libXcomposite.so.1 -> libXcomposite.so.1.0.0
-rw-r--r--  1 root root   11K Jan 20  2017 libXcomposite.so.1.0.0
lrwxrwxrwx  1 root root    19 Dec 19  2017 libXcursor.so.1 -> libXcursor.so.1.0.2
-rw-r--r--  1 root root   39K Dec 19  2017 libXcursor.so.1.0.2
lrwxrwxrwx  1 root root    19 Aug 20  2017 libXdamage.so.1 -> libXdamage.so.1.1.0
-rw-r--r--  1 root root   11K Aug 20  2017 libXdamage.so.1.1.0
lrwxrwxrwx  1 root root    17 Mar  2  2017 libXdmcp.so.6 -> libXdmcp.so.6.0.0
-rw-r--r--  1 root root   23K Mar  2  2017 libXdmcp.so.6.0.0
lrwxrwxrwx  1 root root    16 Oct 25  2014 libXext.so.6 -> libXext.so.6.4.0
-rw-r--r--  1 root root   72K Oct 25  2014 libXext.so.6.4.0
lrwxrwxrwx  1 root root    18 Dec  5  2016 libXfixes.so.3 -> libXfixes.so.3.1.0
-rw-r--r--  1 root root   23K Dec  5  2016 libXfixes.so.3.1.0
lrwxrwxrwx  1 root root    15 Jul 13  2014 libXft.so.2 -> libXft.so.2.3.2
-rw-r--r--  1 root root   84K Jul 13  2014 libXft.so.2.3.2
lrwxrwxrwx  1 root root    14 Jan 23  2017 libXi.so.6 -> libXi.so.6.1.0
-rw-r--r--  1 root root   63K Jan 23  2017 libXi.so.6.1.0
lrwxrwxrwx  1 root root    20 Jun 30  2013 libXinerama.so.1 -> libXinerama.so.1.0.0
-rw-r--r--  1 root root   11K Jun 30  2013 libXinerama.so.1.0.0
lrwxrwxrwx  1 root root    15 Dec 10  2015 libXmu.so.6 -> libXmu.so.6.2.0
-rw-r--r--  1 root root  101K Dec 10  2015 libXmu.so.6.2.0
lrwxrwxrwx  1 root root    16 Dec 22  2016 libXpm.so.4 -> libXpm.so.4.11.0
-rw-r--r--  1 root root   71K Dec 22  2016 libXpm.so.4.11.0
lrwxrwxrwx  1 root root    18 Dec  6  2016 libXrandr.so.2 -> libXrandr.so.2.2.0
-rw-r--r--  1 root root   43K Dec  6  2016 libXrandr.so.2.2.0
lrwxrwxrwx  1 root root    19 Dec  5  2016 libXrender.so.1 -> libXrender.so.1.3.0
-rw-r--r--  1 root root   39K Dec  5  2016 libXrender.so.1.3.0
lrwxrwxrwx  1 root root    15 May  9  2012 libXss.so.1 -> libXss.so.1.0.0
-rw-r--r--  1 root root   15K May  9  2012 libXss.so.1.0.0
lrwxrwxrwx  1 root root    14 May 24  2016 libXt.so.6 -> libXt.so.6.0.0
-rw-r--r--  1 root root  416K May 24  2016 libXt.so.6.0.0
lrwxrwxrwx  1 root root    19 May  6  2015 libXxf86vm.so.1 -> libXxf86vm.so.1.0.0
-rw-r--r--  1 root root   23K May  6  2015 libXxf86vm.so.1.0.0
lrwxrwxrwx  1 root root    15 Jul 28  2017 libaec.so.0 -> libaec.so.0.0.3
-rw-r--r--  1 root root   30K Jul 28  2017 libaec.so.0.0.3
-rw-r--r--  1 root root   23K May  3 10:19 libanl.a
lrwxrwxrwx  1 root root    33 May  3 10:19 libanl.so -> /lib/x86_64-linux-gnu/libanl.so.1
lrwxrwxrwx  1 root root    20 Jun 15  2021 libapt-inst.so.2.0 -> libapt-inst.so.2.0.0
-rw-r--r--  1 root root   51K Jun 15  2021 libapt-inst.so.2.0.0
lrwxrwxrwx  1 root root    19 Jun 15  2021 libapt-pkg.so.5.0 -> libapt-pkg.so.5.0.2
-rw-r--r--  1 root root  1.8M Jun 15  2021 libapt-pkg.so.5.0.2
lrwxrwxrwx  1 root root    23 Jun 15  2021 libapt-private.so.0.0 -> libapt-private.so.0.0.0
-rw-r--r--  1 root root  415K Jun 15  2021 libapt-private.so.0.0.0
lrwxrwxrwx  1 root root    18 Jan 24  2018 libarpack.so.2 -> libarpack.so.2.0.0
-rw-r--r--  1 root root  295K Jan 24  2018 libarpack.so.2.0.0
lrwxrwxrwx  1 root root    16 Dec  4  2019 libasan.so.4 -> libasan.so.4.0.0
-rw-r--r--  1 root root  1.4M Dec  4  2019 libasan.so.4.0.0
lrwxrwxrwx  1 root root    16 Dec 15  2017 libasn1.so.8 -> libasn1.so.8.0.0
-rw-r--r--  1 root root  647K Dec 15  2017 libasn1.so.8.0.0
lrwxrwxrwx  1 root root    23 Mar 13  2018 libatk-1.0.so.0 -> libatk-1.0.so.0.22810.1
-rw-r--r--  1 root root  150K Mar 13  2018 libatk-1.0.so.0.22810.1
lrwxrwxrwx  1 root root    26 Mar 13  2018 libatk-bridge-2.0.so.0 -> libatk-bridge-2.0.so.0.0.0
-rw-r--r--  1 root root  194K Mar 13  2018 libatk-bridge-2.0.so.0.0.0
lrwxrwxrwx  1 root root    18 Mar 10  2020 libatomic.so.1 -> libatomic.so.1.2.0
-rw-r--r--  1 root root   27K Mar 10  2020 libatomic.so.1.2.0
lrwxrwxrwx  1 root root    17 Mar 13  2018 libatspi.so.0 -> libatspi.so.0.0.1
-rw-r--r--  1 root root  191K Mar 13  2018 libatspi.so.0.0.1
lrwxrwxrwx  1 root root    24 Jul  6  2021 libavahi-client.so.3 -> libavahi-client.so.3.2.9
-rw-r--r--  1 root root   67K Jul  6  2021 libavahi-client.so.3.2.9
lrwxrwxrwx  1 root root    24 Jul  6  2021 libavahi-common.so.3 -> libavahi-common.so.3.5.3
-rw-r--r--  1 root root   47K Jul  6  2021 libavahi-common.so.3.5.3
-rw-r--r--  1 root root   19M May 18 20:01 libavcodec.a
lrwxrwxrwx  1 root root    24 May 18 20:01 libavcodec.so -> libavcodec.so.57.107.100
lrwxrwxrwx  1 root root    24 May 18 20:01 libavcodec.so.57 -> libavcodec.so.57.107.100
-rw-r--r--  1 root root   14M May 18 20:01 libavcodec.so.57.107.100
-rw-r--r--  1 root root  4.7M May 18 20:01 libavformat.a
lrwxrwxrwx  1 root root    24 May 18 20:01 libavformat.so -> libavformat.so.57.83.100
lrwxrwxrwx  1 root root    24 May 18 20:01 libavformat.so.57 -> libavformat.so.57.83.100
-rw-r--r--  1 root root  2.4M May 18 20:01 libavformat.so.57.83.100
-rw-r--r--  1 root root  214K May 18 20:01 libavresample.a
lrwxrwxrwx  1 root root    22 May 18 20:01 libavresample.so -> libavresample.so.3.7.0
lrwxrwxrwx  1 root root    22 May 18 20:01 libavresample.so.3 -> libavresample.so.3.7.0
-rw-r--r--  1 root root  130K May 18 20:01 libavresample.so.3.7.0
-rw-r--r--  1 root root  755K May 18 20:01 libavutil.a
lrwxrwxrwx  1 root root    22 May 18 20:01 libavutil.so -> libavutil.so.55.78.100
lrwxrwxrwx  1 root root    22 May 18 20:01 libavutil.so.55 -> libavutil.so.55.78.100
-rw-r--r--  1 root root  487K May 18 20:01 libavutil.so.55.78.100
-rw-r--r--  1 root root  1.3M Oct 20  2021 libbfd-2.30-system.so
lrwxrwxrwx  1 root root    44 Aug 16 08:41 libblas.a -> /etc/alternatives/libblas.a-x86_64-linux-gnu
lrwxrwxrwx  1 root root    45 Aug 16 08:41 libblas.so -> /etc/alternatives/libblas.so-x86_64-linux-gnu
lrwxrwxrwx  1 root root    47 Aug 16 08:41 libblas.so.3 -> /etc/alternatives/libblas.so.3-x86_64-linux-gnu
lrwxrwxrwx  1 root root    18 Apr  3  2018 libbluray.so.2 -> libbluray.so.2.0.2
-rw-r--r--  1 root root  319K Apr  3  2018 libbluray.so.2.0.2
-rw-r--r--  1 root root  5.3M May  3 10:19 libc.a
-rw-r--r--  1 root root   298 May  3 10:19 libc.so
-rw-r--r--  1 root root   20K May  3 10:19 libc_nonshared.a
lrwxrwxrwx  1 root root    29 Jan 22  2019 libcairo-gobject.so.2 -> libcairo-gobject.so.2.11510.0
-rw-r--r--  1 root root   34K Jan 22  2019 libcairo-gobject.so.2.11510.0
lrwxrwxrwx  1 root root    21 Jan 22  2019 libcairo.so.2 -> libcairo.so.2.11510.0
-rw-r--r--  1 root root  1.2M Jan 22  2019 libcairo.so.2.11510.0
lrwxrwxrwx  1 root root    15 Mar 10  2020 libcc1.so.0 -> libcc1.so.0.0.0
-rw-r--r--  1 root root  116K Mar 10  2020 libcc1.so.0.0.0
lrwxrwxrwx  1 root root    15 Mar 24  2018 libcdt.so -> libcdt.so.5.0.0
lrwxrwxrwx  1 root root    15 Mar 24  2018 libcdt.so.5 -> libcdt.so.5.0.0
-rw-r--r--  1 root root   27K Mar 24  2018 libcdt.so.5.0.0
lrwxrwxrwx  1 root root    18 Mar 24  2018 libcgraph.so -> libcgraph.so.6.0.0
lrwxrwxrwx  1 root root    18 Mar 24  2018 libcgraph.so.6 -> libcgraph.so.6.0.0
-rw-r--r--  1 root root   88K Mar 24  2018 libcgraph.so.6.0.0
lrwxrwxrwx  1 root root    23 Feb  7  2018 libchromaprint.so.1 -> libchromaprint.so.1.4.3
-rw-r--r--  1 root root   75K Feb  7  2018 libchromaprint.so.1.4.3
lrwxrwxrwx  1 root root    34 May  3 10:19 libcidn.so -> /lib/x86_64-linux-gnu/libcidn.so.1
lrwxrwxrwx  1 root root    19 Dec  4  2019 libcilkrts.so.5 -> libcilkrts.so.5.0.0
-rw-r--r--  1 root root  117K Dec  4  2019 libcilkrts.so.5.0.0
lrwxrwxrwx  1 root root    18 Jul 23  2017 libcolord.so.2 -> libcolord.so.2.0.5
-rw-r--r--  1 root root  126K Jul 23  2017 libcolord.so.2.0.5
lrwxrwxrwx  1 root root    25 Jul 23  2017 libcolordprivate.so.2 -> libcolordprivate.so.2.0.5
-rw-r--r--  1 root root  188K Jul 23  2017 libcolordprivate.so.2.0.5
lrwxrwxrwx  1 root root    21 Dec 16  2017 libcroco-0.6.so.3 -> libcroco-0.6.so.3.0.1
-rw-r--r--  1 root root  235K Dec 16  2017 libcroco-0.6.so.3.0.1
-rw-r--r--  1 root root   61K May  3 10:19 libcrypt.a
lrwxrwxrwx  1 root root    35 May  3 10:19 libcrypt.so -> /lib/x86_64-linux-gnu/libcrypt.so.1
-rw-r--r--  1 root root  2.8M Jul  4 11:25 libcrypto.so.1.1
lrwxrwxrwx  1 root root    19 Mar 13  2017 libcrystalhd.so.3 -> libcrystalhd.so.3.6
-rw-r--r--  1 root root  109K Mar 13  2017 libcrystalhd.so.3.6
-rw-r--r--  1 root root  558K May 27 15:03 libcups.so.2
lrwxrwxrwx  1 root root    19 Jun 22 16:00 libcurl-gnutls.so.3 -> libcurl-gnutls.so.4
lrwxrwxrwx  1 root root    23 Jun 22 16:00 libcurl-gnutls.so.4 -> libcurl-gnutls.so.4.5.0
-rw-r--r--  1 root root  515K Jun 22 16:00 libcurl-gnutls.so.4.5.0
lrwxrwxrwx  1 root root    16 Jun 22 16:00 libcurl.so.4 -> libcurl.so.4.5.0
-rw-r--r--  1 root root  519K Jun 22 16:00 libcurl.so.4.5.0
lrwxrwxrwx  1 root root    16 Feb 20  2018 libdap.so.25 -> libdap.so.25.0.1
-rw-r--r--  1 root root  1.6M Feb 20  2018 libdap.so.25.0.1
lrwxrwxrwx  1 root root    21 Feb 20  2018 libdapclient.so.6 -> libdapclient.so.6.1.7
-rw-r--r--  1 root root  259K Feb 20  2018 libdapclient.so.6.1.7
lrwxrwxrwx  1 root root    18 Feb 14  2018 libdatrie.so.1 -> libdatrie.so.1.3.3
-rw-r--r--  1 root root   26K Feb 14  2018 libdatrie.so.1.3.3
-rw-r--r--  1 root root  1.7M Jun  3  2019 libdb-5.3.so
-rw-r--r--  1 root root  365K Jan 15  2017 libdc1394.a
lrwxrwxrwx  1 root root    19 Jan 15  2017 libdc1394.so -> libdc1394.so.22.2.1
lrwxrwxrwx  1 root root    19 Jan 15  2017 libdc1394.so.22 -> libdc1394.so.22.2.1
-rw-r--r--  1 root root  216K Jan 15  2017 libdc1394.so.22.2.1
lrwxrwxrwx  1 root root    17 Mar 29  2018 libdconf.so.1 -> libdconf.so.1.0.0
-rw-r--r--  1 root root   56K Mar 29  2018 libdconf.so.1.0.0
lrwxrwxrwx  1 root root    25 Jun  8  2016 libdebconfclient.so.0 -> libdebconfclient.so.0.0.0
-rw-r--r--  1 root root   11K Jun  8  2016 libdebconfclient.so.0.0.0
-rw-r--r--  1 root root   12K May  3 10:19 libdl.a
lrwxrwxrwx  1 root root    32 May  3 10:19 libdl.so -> /lib/x86_64-linux-gnu/libdl.so.2
lrwxrwxrwx  1 root root    15 May 11  2020 libdrm.so.2 -> libdrm.so.2.4.0
-rw-r--r--  1 root root   67K May 11  2020 libdrm.so.2.4.0
lrwxrwxrwx  1 root root    22 May 11  2020 libdrm_amdgpu.so.1 -> libdrm_amdgpu.so.1.0.0
-rw-r--r--  1 root root   39K May 11  2020 libdrm_amdgpu.so.1.0.0
lrwxrwxrwx  1 root root    21 May 11  2020 libdrm_intel.so.1 -> libdrm_intel.so.1.0.0
-rw-r--r--  1 root root  140K May 11  2020 libdrm_intel.so.1.0.0
lrwxrwxrwx  1 root root    23 May 11  2020 libdrm_nouveau.so.2 -> libdrm_nouveau.so.2.0.0
-rw-r--r--  1 root root   31K May 11  2020 libdrm_nouveau.so.2.0.0
lrwxrwxrwx  1 root root    22 May 11  2020 libdrm_radeon.so.1 -> libdrm_radeon.so.1.0.1
-rw-r--r--  1 root root   47K May 11  2020 libdrm_radeon.so.1.0.1
lrwxrwxrwx  1 root root    17 Jun 19  2017 libedit.so.2 -> libedit.so.2.0.56
-rw-r--r--  1 root root  204K Jun 19  2017 libedit.so.2.0.56
-rw-r--r--  1 root root  103K Jun  7  2019 libelf-0.170.so
lrwxrwxrwx  1 root root    15 Jun  7  2019 libelf.so.1 -> libelf-0.170.so
lrwxrwxrwx  1 root root    17 Nov 13  2017 libepoxy.so.0 -> libepoxy.so.0.0.0
-rw-r--r--  1 root root  1.1M Nov 13  2017 libepoxy.so.0.0.0
lrwxrwxrwx  1 root root    19 Apr 27  2016 libepsilon.so.1 -> libepsilon.so.1.0.0
-rw-r--r--  1 root root   97K Apr 27  2016 libepsilon.so.1.0.0
-rw-r--r--  1 root root  400K Nov  6  2020 libexif.a
lrwxrwxrwx  1 root root    17 Nov  6  2020 libexif.so -> libexif.so.12.3.3
lrwxrwxrwx  1 root root    17 Nov  6  2020 libexif.so.12 -> libexif.so.12.3.3
-rw-r--r--  1 root root  274K Nov  6  2020 libexif.so.12.3.3
lrwxrwxrwx  1 root root    18 Mar  8  2022 libexpatw.so.1 -> libexpatw.so.1.6.7
-rw-r--r--  1 root root  203K Mar  8  2022 libexpatw.so.1.6.7
lrwxrwxrwx  1 root root    18 Jan  4  2018 libfabric.so.1 -> libfabric.so.1.9.3
-rw-r--r--  1 root root  676K Jan  4  2018 libfabric.so.1.9.3
lrwxrwxrwx  1 root root    15 Jan  7  2018 libffi.so.6 -> libffi.so.6.0.4
-rw-r--r--  1 root root   31K Jan  7  2018 libffi.so.6.0.4
lrwxrwxrwx  1 root root    23 Apr  5  2018 libfontconfig.so.1 -> libfontconfig.so.1.10.1
-rw-r--r--  1 root root  276K Apr  5  2018 libfontconfig.so.1.10.1
lrwxrwxrwx  1 root root    14 May 23  2018 libform.so.5 -> libform.so.5.9
-rw-r--r--  1 root root   60K May 23  2018 libform.so.5.9
lrwxrwxrwx  1 root root    15 May 23  2018 libformw.so.5 -> libformw.so.5.9
-rw-r--r--  1 root root   68K May 23  2018 libformw.so.5.9
lrwxrwxrwx  1 root root    18 Jul  6 11:25 libfreebl3.chk -> nss/libfreebl3.chk
lrwxrwxrwx  1 root root    17 Jul  6 11:25 libfreebl3.so -> nss/libfreebl3.so
lrwxrwxrwx  1 root root    22 Jul  6 11:25 libfreeblpriv3.chk -> nss/libfreeblpriv3.chk
lrwxrwxrwx  1 root root    21 Jul  6 11:25 libfreeblpriv3.so -> nss/libfreeblpriv3.so
lrwxrwxrwx  1 root root    21 Jul 19 16:39 libfreetype.so.6 -> libfreetype.so.6.15.0
-rw-r--r--  1 root root  719K Jul 19 16:39 libfreetype.so.6.15.0
lrwxrwxrwx  1 root root    18 Feb 22  2018 libfreexl.so.1 -> libfreexl.so.1.1.0
-rw-r--r--  1 root root   34K Feb 22  2018 libfreexl.so.1.1.0
lrwxrwxrwx  1 root root    16 Aug 19  2016 libfyba.so.0 -> libfyba.so.0.0.0
-rw-r--r--  1 root root  215K Aug 19  2016 libfyba.so.0.0.0
lrwxrwxrwx  1 root root    16 Aug 19  2016 libfygm.so.0 -> libfygm.so.0.0.0
-rw-r--r--  1 root root   27K Aug 19  2016 libfygm.so.0.0.0
lrwxrwxrwx  1 root root    16 Aug 19  2016 libfyut.so.0 -> libfyut.so.0.0.0
-rw-r--r--  1 root root   35K Aug 19  2016 libfyut.so.0.0.0
-rw-r--r--  1 root root  1.2K May  3 10:19 libg.a
lrwxrwxrwx  1 root root    14 Aug 30  2021 libgd.so.3 -> libgd.so.3.0.5
-rw-r--r--  1 root root  397K Aug 30  2021 libgd.so.3.0.5
lrwxrwxrwx  1 root root    16 Mar 14  2018 libgdbm.so.5 -> libgdbm.so.5.0.0
-rw-r--r--  1 root root   51K Mar 14  2018 libgdbm.so.5.0.0
lrwxrwxrwx  1 root root    23 Mar 14  2018 libgdbm_compat.so.4 -> libgdbm_compat.so.4.0.0
-rw-r--r--  1 root root   14K Mar 14  2018 libgdbm_compat.so.4.0.0
lrwxrwxrwx  1 root root    20 Feb  5  2018 libgdcmCommon.so -> libgdcmCommon.so.2.8
lrwxrwxrwx  1 root root    22 Feb  5  2018 libgdcmCommon.so.2.8 -> libgdcmCommon.so.2.8.4
-rw-r--r--  1 root root  155K Feb  5  2018 libgdcmCommon.so.2.8.4
lrwxrwxrwx  1 root root    18 Feb  5  2018 libgdcmDICT.so -> libgdcmDICT.so.2.8
lrwxrwxrwx  1 root root    20 Feb  5  2018 libgdcmDICT.so.2.8 -> libgdcmDICT.so.2.8.4
-rw-r--r--  1 root root  2.4M Feb  5  2018 libgdcmDICT.so.2.8.4
lrwxrwxrwx  1 root root    18 Feb  5  2018 libgdcmDSED.so -> libgdcmDSED.so.2.8
lrwxrwxrwx  1 root root    20 Feb  5  2018 libgdcmDSED.so.2.8 -> libgdcmDSED.so.2.8.4
-rw-r--r--  1 root root  1.1M Feb  5  2018 libgdcmDSED.so.2.8.4
lrwxrwxrwx  1 root root    17 Feb  5  2018 libgdcmIOD.so -> libgdcmIOD.so.2.8
lrwxrwxrwx  1 root root    19 Feb  5  2018 libgdcmIOD.so.2.8 -> libgdcmIOD.so.2.8.4
-rw-r--r--  1 root root   87K Feb  5  2018 libgdcmIOD.so.2.8.4
lrwxrwxrwx  1 root root    18 Feb  5  2018 libgdcmMEXD.so -> libgdcmMEXD.so.2.8
lrwxrwxrwx  1 root root    20 Feb  5  2018 libgdcmMEXD.so.2.8 -> libgdcmMEXD.so.2.8.4
-rw-r--r--  1 root root  1.1M Feb  5  2018 libgdcmMEXD.so.2.8.4
lrwxrwxrwx  1 root root    18 Feb  5  2018 libgdcmMSFF.so -> libgdcmMSFF.so.2.8
lrwxrwxrwx  1 root root    20 Feb  5  2018 libgdcmMSFF.so.2.8 -> libgdcmMSFF.so.2.8.4
-rw-r--r--  1 root root  2.6M Feb  5  2018 libgdcmMSFF.so.2.8.4
lrwxrwxrwx  1 root root    20 Feb  5  2018 libgdcmjpeg12.so -> libgdcmjpeg12.so.2.8
lrwxrwxrwx  1 root root    22 Feb  5  2018 libgdcmjpeg12.so.2.8 -> libgdcmjpeg12.so.2.8.4
-rw-r--r--  1 root root  158K Feb  5  2018 libgdcmjpeg12.so.2.8.4
lrwxrwxrwx  1 root root    20 Feb  5  2018 libgdcmjpeg16.so -> libgdcmjpeg16.so.2.8
lrwxrwxrwx  1 root root    22 Feb  5  2018 libgdcmjpeg16.so.2.8 -> libgdcmjpeg16.so.2.8.4
-rw-r--r--  1 root root  158K Feb  5  2018 libgdcmjpeg16.so.2.8.4
lrwxrwxrwx  1 root root    19 Feb  5  2018 libgdcmjpeg8.so -> libgdcmjpeg8.so.2.8
lrwxrwxrwx  1 root root    21 Feb  5  2018 libgdcmjpeg8.so.2.8 -> libgdcmjpeg8.so.2.8.4
-rw-r--r--  1 root root  158K Feb  5  2018 libgdcmjpeg8.so.2.8.4
lrwxrwxrwx  1 root root    21 Jun  5  2019 libgdk-3.so.0 -> libgdk-3.so.0.2200.30
-rw-r--r--  1 root root  981K Jun  5  2019 libgdk-3.so.0.2200.30
lrwxrwxrwx  1 root root    27 Mar 24  2018 libgdk-x11-2.0.so.0 -> libgdk-x11-2.0.so.0.2400.32
-rw-r--r--  1 root root  724K Mar 24  2018 libgdk-x11-2.0.so.0.2400.32
lrwxrwxrwx  1 root root    29 Mar 16  2018 libgdk_pixbuf-2.0.so.0 -> libgdk_pixbuf-2.0.so.0.3611.0
-rw-r--r--  1 root root  144K Mar 16  2018 libgdk_pixbuf-2.0.so.0.3611.0
lrwxrwxrwx  1 root root    34 Mar 16  2018 libgdk_pixbuf_xlib-2.0.so.0 -> libgdk_pixbuf_xlib-2.0.so.0.3611.0
-rw-r--r--  1 root root   67K Mar 16  2018 libgdk_pixbuf_xlib-2.0.so.0.3611.0
-rw-r--r--  1 root root  1.6M Mar  1  2018 libgeos-3.6.2.so
lrwxrwxrwx  1 root root    19 Mar  1  2018 libgeos_c.so.1 -> libgeos_c.so.1.10.2
-rw-r--r--  1 root root  187K Mar  1  2018 libgeos_c.so.1.10.2
lrwxrwxrwx  1 root root    19 Nov 17  2016 libgeotiff.so.2 -> libgeotiff.so.2.1.2
-rw-r--r--  1 root root  215K Nov 17  2016 libgeotiff.so.2.1.2
lrwxrwxrwx  1 root root    20 Dec  4  2019 libgfortran.so.4 -> libgfortran.so.4.0.0
-rw-r--r--  1 root root  1.9M Dec  4  2019 libgfortran.so.4.0.0
lrwxrwxrwx  1 root root    15 Aug 19  2019 libgif.so.7 -> libgif.so.7.0.0
-rw-r--r--  1 root root   34K Aug 19  2019 libgif.so.7.0.0
-rw-r--r--  1 root root  4.0M Nov 29  2021 libgio-2.0.a
lrwxrwxrwx  1 root root    22 Nov 29  2021 libgio-2.0.so -> libgio-2.0.so.0.5600.4
lrwxrwxrwx  1 root root    22 Nov 29  2021 libgio-2.0.so.0 -> libgio-2.0.so.0.5600.4
-rw-r--r--  1 root root  1.7M Nov 29  2021 libgio-2.0.so.0.5600.4
lrwxrwxrwx  1 root root    28 Apr 15  2018 libgirepository-1.0.so.1 -> libgirepository-1.0.so.1.0.0
-rw-r--r--  1 root root  208K Apr 15  2018 libgirepository-1.0.so.1.0.0
lrwxrwxrwx  1 root root    17 Sep 27  2017 libgl2ps.so.1.4 -> libgl2ps.so.1.4.0
-rw-r--r--  1 root root   79K Sep 27  2017 libgl2ps.so.1.4.0
lrwxrwxrwx  1 root root    17 Jun 12  2020 libglapi.so.0 -> libglapi.so.0.0.0
-rw-r--r--  1 root root  215K Jun 12  2020 libglapi.so.0.0.0
-rw-r--r--  1 root root  2.0M Nov 29  2021 libglib-2.0.a
lrwxrwxrwx  1 root root    23 Nov 29  2021 libglib-2.0.so -> libglib-2.0.so.0.5600.4
lrwxrwxrwx  1 root root    23 Nov 29  2021 libglib-2.0.so.0 -> libglib-2.0.so.0.5600.4
-rw-r--r--  1 root root  1.1M Nov 29  2021 libglib-2.0.so.0.5600.4
lrwxrwxrwx  1 root root    15 Dec  7  2017 libgme.so.0 -> libgme.so.0.6.2
-rw-r--r--  1 root root  303K Dec  7  2017 libgme.so.0.6.2
-rw-r--r--  1 root root   16K Nov 29  2021 libgmodule-2.0.a
lrwxrwxrwx  1 root root    26 Nov 29  2021 libgmodule-2.0.so -> libgmodule-2.0.so.0.5600.4
lrwxrwxrwx  1 root root    26 Nov 29  2021 libgmodule-2.0.so.0 -> libgmodule-2.0.so.0.5600.4
-rw-r--r--  1 root root   14K Nov 29  2021 libgmodule-2.0.so.0.5600.4
lrwxrwxrwx  1 root root    16 Jan 24  2018 libgmp.so.10 -> libgmp.so.10.3.2
-rw-r--r--  1 root root  515K Jan 24  2018 libgmp.so.10.3.2
lrwxrwxrwx  1 root root    21 Aug 25  2021 libgnutls.so.30 -> libgnutls.so.30.14.10
-rw-r--r--  1 root root  1.4M Aug 25  2021 libgnutls.so.30.14.10
-rw-r--r--  1 root root  667K Nov 29  2021 libgobject-2.0.a
lrwxrwxrwx  1 root root    26 Nov 29  2021 libgobject-2.0.so -> libgobject-2.0.so.0.5600.4
lrwxrwxrwx  1 root root    26 Nov 29  2021 libgobject-2.0.so.0 -> libgobject-2.0.so.0.5600.4
-rw-r--r--  1 root root  335K Nov 29  2021 libgobject-2.0.so.0.5600.4
lrwxrwxrwx  1 root root    16 Mar 10  2020 libgomp.so.1 -> libgomp.so.1.0.0
-rw-r--r--  1 root root  188K Mar 10  2020 libgomp.so.1.0.0
drwxr-xr-x  3 root root  4.0K Aug 16 08:41 libgphoto2
drwxr-xr-x  3 root root  4.0K Aug 16 08:41 libgphoto2-dev
-rw-r--r--  1 root root  262K Dec 23  2017 libgphoto2.a
lrwxrwxrwx  1 root root    19 Dec 23  2017 libgphoto2.so -> libgphoto2.so.6.0.0
lrwxrwxrwx  1 root root    19 Dec 23  2017 libgphoto2.so.6 -> libgphoto2.so.6.0.0
-rw-r--r--  1 root root  143K Dec 23  2017 libgphoto2.so.6.0.0
drwxr-xr-x  3 root root  4.0K Aug 16 08:41 libgphoto2_port
-rw-r--r--  1 root root   79K Dec 23  2017 libgphoto2_port.a
lrwxrwxrwx  1 root root    25 Dec 23  2017 libgphoto2_port.so -> libgphoto2_port.so.12.0.0
lrwxrwxrwx  1 root root    25 Dec 23  2017 libgphoto2_port.so.12 -> libgphoto2_port.so.12.0.0
-rw-r--r--  1 root root   43K Dec 23  2017 libgphoto2_port.so.12.0.0
lrwxrwxrwx  1 root root    17 Mar 11  2018 libgraphite2.so -> libgraphite2.so.3
lrwxrwxrwx  1 root root    17 Mar 11  2018 libgraphite2.so.2.0.0 -> libgraphite2.so.3
lrwxrwxrwx  1 root root    21 Mar 11  2018 libgraphite2.so.3 -> libgraphite2.so.3.0.1
-rw-r--r--  1 root root  179K Mar 11  2018 libgraphite2.so.3.0.1
lrwxrwxrwx  1 root root    16 Apr  3  2018 libgsm.so.1 -> libgsm.so.1.0.12
-rw-r--r--  1 root root   51K Apr  3  2018 libgsm.so.1.0.12
lrwxrwxrwx  1 root root    18 Dec 15  2017 libgssapi.so.3 -> libgssapi.so.3.0.0
-rw-r--r--  1 root root  260K Dec 15  2017 libgssapi.so.3.0.0
lrwxrwxrwx  1 root root    21 Nov 11  2020 libgssapi_krb5.so.2 -> libgssapi_krb5.so.2.2
-rw-r--r--  1 root root  299K Nov 11  2020 libgssapi_krb5.so.2.2
-rw-r--r--  1 root root  2.7K Nov 29  2021 libgthread-2.0.a
lrwxrwxrwx  1 root root    26 Nov 29  2021 libgthread-2.0.so -> libgthread-2.0.so.0.5600.4
lrwxrwxrwx  1 root root    26 Nov 29  2021 libgthread-2.0.so.0 -> libgthread-2.0.so.0.5600.4
-rw-r--r--  1 root root  5.9K Nov 29  2021 libgthread-2.0.so.0.5600.4
drwxr-xr-x  2 root root  4.0K Aug 16 08:41 libgtk-3-0
lrwxrwxrwx  1 root root    21 Jun  5  2019 libgtk-3.so.0 -> libgtk-3.so.0.2200.30
-rw-r--r--  1 root root  7.1M Jun  5  2019 libgtk-3.so.0.2200.30
lrwxrwxrwx  1 root root    27 Mar 24  2018 libgtk-x11-2.0.so.0 -> libgtk-x11-2.0.so.0.2400.32
-rw-r--r--  1 root root  4.3M Mar 24  2018 libgtk-x11-2.0.so.0.2400.32
drwxr-xr-x  2 root root  4.0K Aug 16 08:41 libgtk2.0-0
lrwxrwxrwx  1 root root    19 Feb  3  2017 libgts-0.7.so.5 -> libgts-0.7.so.5.0.1
-rw-r--r--  1 root root  371K Feb  3  2017 libgts-0.7.so.5.0.1
lrwxrwxrwx  1 root root    15 Mar 24  2018 libgvc.so -> libgvc.so.6.0.0
lrwxrwxrwx  1 root root    15 Mar 24  2018 libgvc.so.6 -> libgvc.so.6.0.0
-rw-r--r--  1 root root  613K Mar 24  2018 libgvc.so.6.0.0
lrwxrwxrwx  1 root root    16 Mar 24  2018 libgvpr.so -> libgvpr.so.2.0.0
lrwxrwxrwx  1 root root    16 Mar 24  2018 libgvpr.so.2 -> libgvpr.so.2.0.0
-rw-r--r--  1 root root  480K Mar 24  2018 libgvpr.so.2.0.0
-rw-r--r--  1 root root   67K Apr 13  2018 libharfbuzz-gobject.a
lrwxrwxrwx  1 root root    32 Apr 13  2018 libharfbuzz-gobject.so -> libharfbuzz-gobject.so.0.10702.0
lrwxrwxrwx  1 root root    32 Apr 13  2018 libharfbuzz-gobject.so.0 -> libharfbuzz-gobject.so.0.10702.0
-rw-r--r--  1 root root   55K Apr 13  2018 libharfbuzz-gobject.so.0.10702.0
-rw-r--r--  1 root root  7.0K Apr 13  2018 libharfbuzz-icu.a
lrwxrwxrwx  1 root root    28 Apr 13  2018 libharfbuzz-icu.so -> libharfbuzz-icu.so.0.10702.0
lrwxrwxrwx  1 root root    28 Apr 13  2018 libharfbuzz-icu.so.0 -> libharfbuzz-icu.so.0.10702.0
-rw-r--r--  1 root root   11K Apr 13  2018 libharfbuzz-icu.so.0.10702.0
-rw-r--r--  1 root root  965K Apr 13  2018 libharfbuzz.a
lrwxrwxrwx  1 root root    24 Apr 13  2018 libharfbuzz.so -> libharfbuzz.so.0.10702.0
lrwxrwxrwx  1 root root    24 Apr 13  2018 libharfbuzz.so.0 -> libharfbuzz.so.0.10702.0
-rw-r--r--  1 root root  631K Apr 13  2018 libharfbuzz.so.0.10702.0
lrwxrwxrwx  1 root root    19 Dec 15  2017 libhcrypto.so.4 -> libhcrypto.so.4.1.0
-rw-r--r--  1 root root  213K Dec 15  2017 libhcrypto.so.4.1.0
lrwxrwxrwx  1 root root    26 Aug 13  2017 libhdf5_openmpi.so.100 -> libhdf5_openmpi.so.100.0.1
-rw-r--r--  1 root root  3.5M Aug 13  2017 libhdf5_openmpi.so.100.0.1
lrwxrwxrwx  1 root root    34 Aug 13  2017 libhdf5_openmpi_fortran.so.100 -> libhdf5_openmpi_fortran.so.100.0.1
-rw-r--r--  1 root root  249K Aug 13  2017 libhdf5_openmpi_fortran.so.100.0.1
lrwxrwxrwx  1 root root    29 Aug 13  2017 libhdf5_openmpi_hl.so.100 -> libhdf5_openmpi_hl.so.100.0.0
-rw-r--r--  1 root root  136K Aug 13  2017 libhdf5_openmpi_hl.so.100.0.0
lrwxrwxrwx  1 root root    36 Aug 13  2017 libhdf5_openmpihl_fortran.so.100 -> libhdf5_openmpihl_fortran.so.100.0.0
-rw-r--r--  1 root root  119K Aug 13  2017 libhdf5_openmpihl_fortran.so.100.0.0
lrwxrwxrwx  1 root root    25 Aug 13  2017 libhdf5_serial.so.100 -> libhdf5_serial.so.100.0.1
-rw-r--r--  1 root root  3.4M Aug 13  2017 libhdf5_serial.so.100.0.1
lrwxrwxrwx  1 root root    33 Aug 13  2017 libhdf5_serial_fortran.so.100 -> libhdf5_serial_fortran.so.100.0.1
-rw-r--r--  1 root root  245K Aug 13  2017 libhdf5_serial_fortran.so.100.0.1
lrwxrwxrwx  1 root root    28 Aug 13  2017 libhdf5_serial_hl.so.100 -> libhdf5_serial_hl.so.100.0.0
-rw-r--r--  1 root root  136K Aug 13  2017 libhdf5_serial_hl.so.100.0.0
lrwxrwxrwx  1 root root    35 Aug 13  2017 libhdf5_serialhl_fortran.so.100 -> libhdf5_serialhl_fortran.so.100.0.0
-rw-r--r--  1 root root  119K Aug 13  2017 libhdf5_serialhl_fortran.so.100.0.0
lrwxrwxrwx  1 root root    20 Dec 15  2017 libheimbase.so.1 -> libheimbase.so.1.0.0
-rw-r--r--  1 root root   59K Dec 15  2017 libheimbase.so.1.0.0
lrwxrwxrwx  1 root root    20 Dec 15  2017 libheimntlm.so.0 -> libheimntlm.so.0.1.0
-rw-r--r--  1 root root   35K Dec 15  2017 libheimntlm.so.0.1.0
lrwxrwxrwx  1 root root    17 Jun 14  2021 libhogweed.so.4 -> libhogweed.so.4.5
-rw-r--r--  1 root root  215K Jun 14  2021 libhogweed.so.4.5
lrwxrwxrwx  1 root root    13 Jan 19  2018 libhwloc.so.0 -> libhwloc.so.5
lrwxrwxrwx  1 root root    13 Jan 19  2018 libhwloc.so.1 -> libhwloc.so.5
lrwxrwxrwx  1 root root    13 Jan 19  2018 libhwloc.so.2 -> libhwloc.so.5
lrwxrwxrwx  1 root root    13 Jan 19  2018 libhwloc.so.3 -> libhwloc.so.5
lrwxrwxrwx  1 root root    13 Jan 19  2018 libhwloc.so.4 -> libhwloc.so.5
lrwxrwxrwx  1 root root    17 Jan 19  2018 libhwloc.so.5 -> libhwloc.so.5.7.6
-rw-r--r--  1 root root  244K Jan 19  2018 libhwloc.so.5.7.6
lrwxrwxrwx  1 root root    17 Dec 15  2017 libhx509.so.5 -> libhx509.so.5.0.0
-rw-r--r--  1 root root  294K Dec 15  2017 libhx509.so.5.0.0
lrwxrwxrwx  1 root root    22 Aug  5  2019 libibverbs.so.1 -> libibverbs.so.1.1.17.1
-rw-r--r--  1 root root   87K Aug  5  2019 libibverbs.so.1.1.17.1
-rw-r--r--  1 root root   55K Dec  9  2017 libicu-le-hb.a
lrwxrwxrwx  1 root root    21 Dec  9  2017 libicu-le-hb.so -> libicu-le-hb.so.0.0.0
lrwxrwxrwx  1 root root    21 Dec  9  2017 libicu-le-hb.so.0 -> libicu-le-hb.so.0.0.0
-rw-r--r--  1 root root   35K Dec  9  2017 libicu-le-hb.so.0.0.0
-rw-r--r--  1 root root   26M Oct 19  2021 libicudata.a
lrwxrwxrwx  1 root root    18 Oct 19  2021 libicudata.so -> libicudata.so.60.2
lrwxrwxrwx  1 root root    18 Oct 19  2021 libicudata.so.60 -> libicudata.so.60.2
-rw-r--r--  1 root root   26M Oct 19  2021 libicudata.so.60.2
-rw-r--r--  1 root root  5.9M Oct 19  2021 libicui18n.a
lrwxrwxrwx  1 root root    18 Oct 19  2021 libicui18n.so -> libicui18n.so.60.2
lrwxrwxrwx  1 root root    18 Oct 19  2021 libicui18n.so.60 -> libicui18n.so.60.2
-rw-r--r--  1 root root  2.7M Oct 19  2021 libicui18n.so.60.2
-rw-r--r--  1 root root   85K Oct 19  2021 libicuio.a
lrwxrwxrwx  1 root root    16 Oct 19  2021 libicuio.so -> libicuio.so.60.2
lrwxrwxrwx  1 root root    16 Oct 19  2021 libicuio.so.60 -> libicuio.so.60.2
-rw-r--r--  1 root root   55K Oct 19  2021 libicuio.so.60.2
-rw-r--r--  1 root root   75K Oct 19  2021 libiculx.a
lrwxrwxrwx  1 root root    16 Oct 19  2021 libiculx.so -> libiculx.so.60.2
lrwxrwxrwx  1 root root    16 Oct 19  2021 libiculx.so.60 -> libiculx.so.60.2
-rw-r--r--  1 root root   47K Oct 19  2021 libiculx.so.60.2
-rw-r--r--  1 root root  109K Oct 19  2021 libicutest.a
lrwxrwxrwx  1 root root    18 Oct 19  2021 libicutest.so -> libicutest.so.60.2
lrwxrwxrwx  1 root root    18 Oct 19  2021 libicutest.so.60 -> libicutest.so.60.2
-rw-r--r--  1 root root   64K Oct 19  2021 libicutest.so.60.2
-rw-r--r--  1 root root  359K Oct 19  2021 libicutu.a
lrwxrwxrwx  1 root root    16 Oct 19  2021 libicutu.so -> libicutu.so.60.2
lrwxrwxrwx  1 root root    16 Oct 19  2021 libicutu.so.60 -> libicutu.so.60.2
-rw-r--r--  1 root root  199K Oct 19  2021 libicutu.so.60.2
-rw-r--r--  1 root root  3.1M Oct 19  2021 libicuuc.a
lrwxrwxrwx  1 root root    16 Oct 19  2021 libicuuc.so -> libicuuc.so.60.2
lrwxrwxrwx  1 root root    16 Oct 19  2021 libicuuc.so.60 -> libicuuc.so.60.2
-rw-r--r--  1 root root  1.8M Oct 19  2021 libicuuc.so.60.2
lrwxrwxrwx  1 root root    16 Oct 24  2019 libidn2.so.0 -> libidn2.so.0.3.3
-rw-r--r--  1 root root  114K Oct 24  2019 libidn2.so.0.3.3
lrwxrwxrwx  1 root root    20 Jan 14  2018 libinfinipath.so.4 -> libinfinipath.so.4.0
-rw-r--r--  1 root root   59K Jan 14  2018 libinfinipath.so.4.0
lrwxrwxrwx  1 root root    16 Mar  9  2018 libisl.so.19 -> libisl.so.19.0.0
-rw-r--r--  1 root root  1.6M Mar  9  2018 libisl.so.19.0.0
lrwxrwxrwx  1 root root    15 Mar 10  2020 libitm.so.1 -> libitm.so.1.0.0
-rw-r--r--  1 root root  111K Mar 10  2020 libitm.so.1.0.0
-rw-r--r--  1 root root   51K Apr  3  2018 libjbig.a
lrwxrwxrwx  1 root root    12 Apr  3  2018 libjbig.so -> libjbig.so.0
-rw-r--r--  1 root root   57K Apr  3  2018 libjbig.so.0
-rw-r--r--  1 root root  579K Jun  4  2020 libjpeg.a
lrwxrwxrwx  1 root root    16 Jun  4  2020 libjpeg.so -> libjpeg.so.8.1.2
lrwxrwxrwx  1 root root    16 Jun  4  2020 libjpeg.so.8 -> libjpeg.so.8.1.2
-rw-r--r--  1 root root  415K Jun  4  2020 libjpeg.so.8.1.2
lrwxrwxrwx  1 root root    27 Apr  6  2020 libjson-glib-1.0.so.0 -> libjson-glib-1.0.so.0.400.2
-rw-r--r--  1 root root  154K Apr  6  2020 libjson-glib-1.0.so.0.400.2
lrwxrwxrwx  1 root root    19 Aug 23  2016 libjsoncpp.so.1 -> libjsoncpp.so.1.7.4
-rw-r--r--  1 root root  199K Aug 23  2016 libjsoncpp.so.1.7.4
lrwxrwxrwx  1 root root    18 Nov 11  2020 libk5crypto.so.3 -> libk5crypto.so.3.1
-rw-r--r--  1 root root  195K Nov 11  2020 libk5crypto.so.3.1
lrwxrwxrwx  1 root root    19 Sep 25  2017 libkmlbase.so.1 -> libkmlbase.so.1.3.0
-rw-r--r--  1 root root  107K Sep 25  2017 libkmlbase.so.1.3.0
lrwxrwxrwx  1 root root    18 Sep 25  2017 libkmldom.so.1 -> libkmldom.so.1.3.0
-rw-r--r--  1 root root  731K Sep 25  2017 libkmldom.so.1.3.0
lrwxrwxrwx  1 root root    21 Sep 25  2017 libkmlengine.so.1 -> libkmlengine.so.1.3.0
-rw-r--r--  1 root root  223K Sep 25  2017 libkmlengine.so.1.3.0
lrwxrwxrwx  1 root root    17 Dec 15  2017 libkrb5.so.26 -> libkrb5.so.26.0.0
-rw-r--r--  1 root root  561K Dec 15  2017 libkrb5.so.26.0.0
lrwxrwxrwx  1 root root    14 Nov 11  2020 libkrb5.so.3 -> libkrb5.so.3.3
-rw-r--r--  1 root root  857K Nov 11  2020 libkrb5.so.3.3
lrwxrwxrwx  1 root root    21 Nov 11  2020 libkrb5support.so.0 -> libkrb5support.so.0.1
-rw-r--r--  1 root root   43K Nov 11  2020 libkrb5support.so.0.1
lrwxrwxrwx  1 root root    21 Mar 24  2018 liblab_gamut.so -> liblab_gamut.so.1.0.0
lrwxrwxrwx  1 root root    21 Mar 24  2018 liblab_gamut.so.1 -> liblab_gamut.so.1.0.0
-rw-r--r--  1 root root  2.4M Mar 24  2018 liblab_gamut.so.1.0.0
lrwxrwxrwx  1 root root    46 Aug 16 08:41 liblapack.a -> /etc/alternatives/liblapack.a-x86_64-linux-gnu
lrwxrwxrwx  1 root root    47 Aug 16 08:41 liblapack.so -> /etc/alternatives/liblapack.so-x86_64-linux-gnu
lrwxrwxrwx  1 root root    49 Aug 16 08:41 liblapack.so.3 -> /etc/alternatives/liblapack.so.3-x86_64-linux-gnu
lrwxrwxrwx  1 root root    21 May 12 13:52 liblber-2.4.so.2 -> liblber-2.4.so.2.10.8
-rw-r--r--  1 root root   55K May 12 13:52 liblber-2.4.so.2.10.8
lrwxrwxrwx  1 root root    17 Sep 20  2018 liblcms2.so.2 -> liblcms2.so.2.0.8
-rw-r--r--  1 root root  349K Sep 20  2018 liblcms2.so.2.0.8
lrwxrwxrwx  1 root root    18 May 12 13:52 libldap-2.4.so.2 -> libldap_r-2.4.so.2
lrwxrwxrwx  1 root root    23 May 12 13:52 libldap_r-2.4.so.2 -> libldap_r-2.4.so.2.10.8
-rw-r--r--  1 root root  320K May 12 13:52 libldap_r-2.4.so.2.10.8
lrwxrwxrwx  1 root root    16 Mar  2  2018 liblept.so.5 -> liblept.so.5.0.2
-rw-r--r--  1 root root  2.4M Mar  2  2018 liblept.so.5.0.2
lrwxrwxrwx  1 root root    16 Mar 10  2020 liblsan.so.0 -> liblsan.so.0.0.0
-rw-r--r--  1 root root  338K Mar 10  2020 liblsan.so.0.0.0
lrwxrwxrwx  1 root root    16 Aug 20  2016 libltdl.so.7 -> libltdl.so.7.3.1
-rw-r--r--  1 root root   39K Aug 20  2016 libltdl.so.7.3.1
lrwxrwxrwx  1 root root    15 May 20  2021 liblz4.so.1 -> liblz4.so.1.7.1
-rw-r--r--  1 root root  111K May 20  2021 liblz4.so.1.7.1
-rw-r--r--  1 root root  261K Apr  8 12:56 liblzma.a
lrwxrwxrwx  1 root root    38 Apr  8 12:56 liblzma.so -> /lib/x86_64-linux-gnu/liblzma.so.5.2.2
-rw-r--r--  1 root root  3.5M May  3 10:19 libm-2.27.a
-rw-r--r--  1 root root   132 May  3 10:19 libm.a
-rw-r--r--  1 root root   186 May  3 10:19 libm.so
lrwxrwxrwx  1 root root    34 Feb  5  2018 libmca_common_libfabric.so.20 -> libmca_common_libfabric.so.20.10.0
lrwxrwxrwx  1 root root    46 Feb  5  2018 libmca_common_libfabric.so.20.10.0 -> openmpi/lib/libmca_common_libfabric.so.20.10.0
lrwxrwxrwx  1 root root    27 Feb  5  2018 libmca_common_sm.so.20 -> libmca_common_sm.so.20.10.1
lrwxrwxrwx  1 root root    39 Feb  5  2018 libmca_common_sm.so.20.10.1 -> openmpi/lib/libmca_common_sm.so.20.10.1
lrwxrwxrwx  1 root root    30 Feb  5  2018 libmca_common_verbs.so.20 -> libmca_common_verbs.so.20.10.0
lrwxrwxrwx  1 root root    42 Feb  5  2018 libmca_common_verbs.so.20.10.0 -> openmpi/lib/libmca_common_verbs.so.20.10.0
-rw-r--r--  1 root root  1.5K May  3 10:19 libmcheck.a
lrwxrwxrwx  1 root root    14 May 23  2018 libmenu.so.5 -> libmenu.so.5.9
-rw-r--r--  1 root root   35K May 23  2018 libmenu.so.5.9
lrwxrwxrwx  1 root root    15 May 23  2018 libmenuw.so.5 -> libmenuw.so.5.9
-rw-r--r--  1 root root   35K May 23  2018 libmenuw.so.5.9
lrwxrwxrwx  1 root root    19 Apr  3  2018 libminizip.so.1 -> libminizip.so.1.0.0
-rw-r--r--  1 root root   43K Apr  3  2018 libminizip.so.1.0.0
lrwxrwxrwx  1 root root    19 Oct 21  2017 libmp3lame.so.0 -> libmp3lame.so.0.0.0
-rw-r--r--  1 root root  294K Oct 21  2017 libmp3lame.so.0.0.0
lrwxrwxrwx  1 root root    15 Jan 23  2018 libmpc.so.3 -> libmpc.so.3.1.0
-rw-r--r--  1 root root   96K Jan 23  2018 libmpc.so.3.1.0
lrwxrwxrwx  1 root root    17 Apr 23  2018 libmpdec.so.2 -> libmpdec.so.2.4.2
-rw-r--r--  1 root root  223K Apr 23  2018 libmpdec.so.2.4.2
lrwxrwxrwx  1 root root    16 Feb  7  2018 libmpfr.so.6 -> libmpfr.so.6.0.1
-rw-r--r--  1 root root  512K Feb  7  2018 libmpfr.so.6.0.1
lrwxrwxrwx  1 root root    19 Mar 10  2018 libmpg123.so.0 -> libmpg123.so.0.44.8
-rw-r--r--  1 root root  316K Mar 10  2018 libmpg123.so.0.44.8
lrwxrwxrwx  1 root root    17 Feb  5  2018 libmpi.so.20 -> libmpi.so.20.10.1
lrwxrwxrwx  1 root root    29 Feb  5  2018 libmpi.so.20.10.1 -> openmpi/lib/libmpi.so.20.10.1
lrwxrwxrwx  1 root root    21 Feb  5  2018 libmpi_cxx.so.20 -> libmpi_cxx.so.20.10.0
lrwxrwxrwx  1 root root    33 Feb  5  2018 libmpi_cxx.so.20.10.0 -> openmpi/lib/libmpi_cxx.so.20.10.0
lrwxrwxrwx  1 root root    22 Feb  5  2018 libmpi_java.so.20 -> libmpi_java.so.20.10.0
lrwxrwxrwx  1 root root    34 Feb  5  2018 libmpi_java.so.20.10.0 -> openmpi/lib/libmpi_java.so.20.10.0
lrwxrwxrwx  1 root root    23 Feb  5  2018 libmpi_mpifh.so.20 -> libmpi_mpifh.so.20.11.0
lrwxrwxrwx  1 root root    35 Feb  5  2018 libmpi_mpifh.so.20.11.0 -> openmpi/lib/libmpi_mpifh.so.20.11.0
lrwxrwxrwx  1 root root    35 Feb  5  2018 libmpi_usempi_ignore_tkr.so.20 -> libmpi_usempi_ignore_tkr.so.20.10.0
lrwxrwxrwx  1 root root    47 Feb  5  2018 libmpi_usempi_ignore_tkr.so.20.10.0 -> openmpi/lib/libmpi_usempi_ignore_tkr.so.20.10.0
lrwxrwxrwx  1 root root    27 Feb  5  2018 libmpi_usempif08.so.20 -> libmpi_usempif08.so.20.10.0
lrwxrwxrwx  1 root root    39 Feb  5  2018 libmpi_usempif08.so.20.10.0 -> openmpi/lib/libmpi_usempif08.so.20.10.0
lrwxrwxrwx  1 root root    15 Mar 10  2020 libmpx.so.2 -> libmpx.so.2.0.1
-rw-r--r--  1 root root   19K Mar 10  2020 libmpx.so.2.0.1
lrwxrwxrwx  1 root root    23 Mar 10  2020 libmpxwrappers.so.2 -> libmpxwrappers.so.2.0.1
-rw-r--r--  1 root root   15K Mar 10  2020 libmpxwrappers.so.2.0.1
-rw-r--r--  1 root root  337K May  3 10:19 libmvec.a
lrwxrwxrwx  1 root root    34 May  3 10:19 libmvec.so -> /lib/x86_64-linux-gnu/libmvec.so.1
-rw-r--r--  1 root root  5.9K May  3 10:19 libmvec_nonshared.a
lrwxrwxrwx  1 root root    25 Jul 26 11:28 libmysqlclient.so.20 -> libmysqlclient.so.20.3.26
-rw-r--r--  1 root root  3.8M Jul 26 11:28 libmysqlclient.so.20.3.26
-rw-r--r--  1 root root  1.1M Feb  9  2018 libnetcdf.so.13
lrwxrwxrwx  1 root root    22 Nov  9  2017 libnetcdf_c++.so.4 -> libnetcdf_c++.so.4.2.0
-rw-r--r--  1 root root  119K Nov  9  2017 libnetcdf_c++.so.4.2.0
lrwxrwxrwx  1 root root    16 Jun 14  2021 libnettle.so.6 -> libnettle.so.6.5
-rw-r--r--  1 root root  215K Jun 14  2021 libnettle.so.6.5
lrwxrwxrwx  1 root root    21 Apr 10  2018 libnghttp2.so.14 -> libnghttp2.so.14.15.2
-rw-r--r--  1 root root  150K Apr 10  2018 libnghttp2.so.14.15.2
lrwxrwxrwx  1 root root    25 Jun  2  2017 libnl-route-3.so.200 -> libnl-route-3.so.200.24.0
-rw-r--r--  1 root root  458K Jun  2  2017 libnl-route-3.so.200.24.0
-rw-r--r--  1 root root  201K May  3 10:19 libnsl.a
lrwxrwxrwx  1 root root    33 May  3 10:19 libnsl.so -> /lib/x86_64-linux-gnu/libnsl.so.1
-rw-r--r--  1 root root  234K Feb 22  2018 libnspr4.so
-rw-r--r--  1 root root  1.3M Jul  6 11:25 libnss3.so
lrwxrwxrwx  1 root root    40 May  3 10:19 libnss_compat.so -> /lib/x86_64-linux-gnu/libnss_compat.so.2
lrwxrwxrwx  1 root root    37 May  3 10:19 libnss_dns.so -> /lib/x86_64-linux-gnu/libnss_dns.so.2
lrwxrwxrwx  1 root root    39 May  3 10:19 libnss_files.so -> /lib/x86_64-linux-gnu/libnss_files.so.2
lrwxrwxrwx  1 root root    40 May  3 10:19 libnss_hesiod.so -> /lib/x86_64-linux-gnu/libnss_hesiod.so.2
lrwxrwxrwx  1 root root    37 May  3 10:19 libnss_nis.so -> /lib/x86_64-linux-gnu/libnss_nis.so.2
lrwxrwxrwx  1 root root    41 May  3 10:19 libnss_nisplus.so -> /lib/x86_64-linux-gnu/libnss_nisplus.so.2
-rw-r--r--  1 root root  187K Jul  6 11:25 libnssutil3.so
lrwxrwxrwx  1 root root    16 Jun 20  2018 libnuma.so.1 -> libnuma.so.1.0.0
-rw-r--r--  1 root root   43K Jun 20  2018 libnuma.so.1.0.0
lrwxrwxrwx  1 root root    12 Mar 26  2018 libodbc.so.1 -> libodbc.so.2
lrwxrwxrwx  1 root root    16 Mar 26  2018 libodbc.so.2 -> libodbc.so.2.0.0
-rw-r--r--  1 root root  422K Mar 26  2018 libodbc.so.2.0.0
lrwxrwxrwx  1 root root    14 Mar 26  2018 libodbccr.so.1 -> libodbccr.so.2
lrwxrwxrwx  1 root root    18 Mar 26  2018 libodbccr.so.2 -> libodbccr.so.2.0.0
-rw-r--r--  1 root root   43K Mar 26  2018 libodbccr.so.2.0.0
lrwxrwxrwx  1 root root    16 Mar 26  2018 libodbcinst.so.1 -> libodbcinst.so.2
lrwxrwxrwx  1 root root    20 Mar 26  2018 libodbcinst.so.2 -> libodbcinst.so.2.0.0
-rw-r--r--  1 root root   71K Mar 26  2018 libodbcinst.so.2.0.0
lrwxrwxrwx  1 root root    15 Jun  2  2014 libogg.so.0 -> libogg.so.0.8.2
-rw-r--r--  1 root root   34K Jun  2  2014 libogg.so.0.8.2
lrwxrwxrwx  1 root root    23 Feb  5  2018 libompitrace.so.20 -> libompitrace.so.20.10.0
lrwxrwxrwx  1 root root    35 Feb  5  2018 libompitrace.so.20.10.0 -> openmpi/lib/libompitrace.so.20.10.0
-rw-r--r--  1 root root  1.7M Oct 20  2021 libopcodes-2.30-system.so
lrwxrwxrwx  1 root root    22 Feb  5  2018 libopen-pal.so.20 -> libopen-pal.so.20.10.1
lrwxrwxrwx  1 root root    34 Feb  5  2018 libopen-pal.so.20.10.1 -> openmpi/lib/libopen-pal.so.20.10.1
lrwxrwxrwx  1 root root    22 Feb  5  2018 libopen-rte.so.20 -> libopen-rte.so.20.10.1
lrwxrwxrwx  1 root root    34 Feb  5  2018 libopen-rte.so.20.10.1 -> openmpi/lib/libopen-rte.so.20.10.1
lrwxrwxrwx  1 root root    22 Sep 19  2017 libopenblas.a -> libopenblasp-r0.2.20.a
lrwxrwxrwx  1 root root    23 Sep 19  2017 libopenblas.so -> libopenblasp-r0.2.20.so
lrwxrwxrwx  1 root root    23 Sep 19  2017 libopenblas.so.0 -> libopenblasp-r0.2.20.so
-rw-r--r--  1 root root   51M Sep 19  2017 libopenblasp-r0.2.20.a
-rw-r--r--  1 root root   33M Sep 19  2017 libopenblasp-r0.2.20.so
-rw-r--r--  1 root root  483K Sep 20  2018 libopencv_aruco.a
lrwxrwxrwx  1 root root    22 Sep 20  2018 libopencv_aruco.so -> libopencv_aruco.so.3.2
lrwxrwxrwx  1 root root    24 Sep 20  2018 libopencv_aruco.so.3.2 -> libopencv_aruco.so.3.2.0
-rw-r--r--  1 root root  285K Sep 20  2018 libopencv_aruco.so.3.2.0
-rw-r--r--  1 root root  116K Sep 20  2018 libopencv_bgsegm.a
lrwxrwxrwx  1 root root    23 Sep 20  2018 libopencv_bgsegm.so -> libopencv_bgsegm.so.3.2
lrwxrwxrwx  1 root root    25 Sep 20  2018 libopencv_bgsegm.so.3.2 -> libopencv_bgsegm.so.3.2.0
-rw-r--r--  1 root root   51K Sep 20  2018 libopencv_bgsegm.so.3.2.0
-rw-r--r--  1 root root  449K Sep 20  2018 libopencv_bioinspired.a
lrwxrwxrwx  1 root root    28 Sep 20  2018 libopencv_bioinspired.so -> libopencv_bioinspired.so.3.2
lrwxrwxrwx  1 root root    30 Sep 20  2018 libopencv_bioinspired.so.3.2 -> libopencv_bioinspired.so.3.2.0
-rw-r--r--  1 root root  171K Sep 20  2018 libopencv_bioinspired.so.3.2.0
-rw-r--r--  1 root root  2.3M Sep 20  2018 libopencv_calib3d.a
lrwxrwxrwx  1 root root    24 Sep 20  2018 libopencv_calib3d.so -> libopencv_calib3d.so.3.2
lrwxrwxrwx  1 root root    26 Sep 20  2018 libopencv_calib3d.so.3.2 -> libopencv_calib3d.so.3.2.0
-rw-r--r--  1 root root  1.4M Sep 20  2018 libopencv_calib3d.so.3.2.0
-rw-r--r--  1 root root  633K Sep 20  2018 libopencv_ccalib.a
lrwxrwxrwx  1 root root    23 Sep 20  2018 libopencv_ccalib.so -> libopencv_ccalib.so.3.2
lrwxrwxrwx  1 root root    25 Sep 20  2018 libopencv_ccalib.so.3.2 -> libopencv_ccalib.so.3.2.0
-rw-r--r--  1 root root  363K Sep 20  2018 libopencv_ccalib.so.3.2.0
-rw-r--r--  1 root root  4.6M Sep 20  2018 libopencv_core.a
lrwxrwxrwx  1 root root    21 Sep 20  2018 libopencv_core.so -> libopencv_core.so.3.2
lrwxrwxrwx  1 root root    23 Sep 20  2018 libopencv_core.so.3.2 -> libopencv_core.so.3.2.0
-rw-r--r--  1 root root  2.3M Sep 20  2018 libopencv_core.so.3.2.0
-rw-r--r--  1 root root  1.4M Sep 20  2018 libopencv_datasets.a
lrwxrwxrwx  1 root root    25 Sep 20  2018 libopencv_datasets.so -> libopencv_datasets.so.3.2
lrwxrwxrwx  1 root root    27 Sep 20  2018 libopencv_datasets.so.3.2 -> libopencv_datasets.so.3.2.0
-rw-r--r--  1 root root  399K Sep 20  2018 libopencv_datasets.so.3.2.0
-rw-r--r--  1 root root  206K Sep 20  2018 libopencv_dpm.a
lrwxrwxrwx  1 root root    20 Sep 20  2018 libopencv_dpm.so -> libopencv_dpm.so.3.2
lrwxrwxrwx  1 root root    22 Sep 20  2018 libopencv_dpm.so.3.2 -> libopencv_dpm.so.3.2.0
-rw-r--r--  1 root root   99K Sep 20  2018 libopencv_dpm.so.3.2.0
-rw-r--r--  1 root root  331K Sep 20  2018 libopencv_face.a
lrwxrwxrwx  1 root root    21 Sep 20  2018 libopencv_face.so -> libopencv_face.so.3.2
lrwxrwxrwx  1 root root    23 Sep 20  2018 libopencv_face.so.3.2 -> libopencv_face.so.3.2.0
-rw-r--r--  1 root root  143K Sep 20  2018 libopencv_face.so.3.2.0
-rw-r--r--  1 root root  1.4M Sep 20  2018 libopencv_features2d.a
lrwxrwxrwx  1 root root    27 Sep 20  2018 libopencv_features2d.so -> libopencv_features2d.so.3.2
lrwxrwxrwx  1 root root    29 Sep 20  2018 libopencv_features2d.so.3.2 -> libopencv_features2d.so.3.2.0
-rw-r--r--  1 root root  723K Sep 20  2018 libopencv_features2d.so.3.2.0
-rw-r--r--  1 root root 1007K Sep 20  2018 libopencv_flann.a
lrwxrwxrwx  1 root root    22 Sep 20  2018 libopencv_flann.so -> libopencv_flann.so.3.2
lrwxrwxrwx  1 root root    24 Sep 20  2018 libopencv_flann.so.3.2 -> libopencv_flann.so.3.2.0
-rw-r--r--  1 root root  319K Sep 20  2018 libopencv_flann.so.3.2.0
-rw-r--r--  1 root root   43K Sep 20  2018 libopencv_freetype.a
lrwxrwxrwx  1 root root    25 Sep 20  2018 libopencv_freetype.so -> libopencv_freetype.so.3.2
lrwxrwxrwx  1 root root    27 Sep 20  2018 libopencv_freetype.so.3.2 -> libopencv_freetype.so.3.2.0
-rw-r--r--  1 root root   27K Sep 20  2018 libopencv_freetype.so.3.2.0
-rw-r--r--  1 root root   73K Sep 20  2018 libopencv_fuzzy.a
lrwxrwxrwx  1 root root    22 Sep 20  2018 libopencv_fuzzy.so -> libopencv_fuzzy.so.3.2
lrwxrwxrwx  1 root root    24 Sep 20  2018 libopencv_fuzzy.so.3.2 -> libopencv_fuzzy.so.3.2.0
-rw-r--r--  1 root root   54K Sep 20  2018 libopencv_fuzzy.so.3.2.0
-rw-r--r--  1 root root   75K Sep 20  2018 libopencv_hdf.a
lrwxrwxrwx  1 root root    20 Sep 20  2018 libopencv_hdf.so -> libopencv_hdf.so.3.2
lrwxrwxrwx  1 root root    22 Sep 20  2018 libopencv_hdf.so.3.2 -> libopencv_hdf.so.3.2.0
-rw-r--r--  1 root root   39K Sep 20  2018 libopencv_hdf.so.3.2.0
-rw-r--r--  1 root root   94K Sep 20  2018 libopencv_highgui.a
lrwxrwxrwx  1 root root    24 Sep 20  2018 libopencv_highgui.so -> libopencv_highgui.so.3.2
lrwxrwxrwx  1 root root    26 Sep 20  2018 libopencv_highgui.so.3.2 -> libopencv_highgui.so.3.2.0
-rw-r--r--  1 root root   47K Sep 20  2018 libopencv_highgui.so.3.2.0
-rw-r--r--  1 root root  756K Sep 20  2018 libopencv_imgcodecs.a
lrwxrwxrwx  1 root root    26 Sep 20  2018 libopencv_imgcodecs.so -> libopencv_imgcodecs.so.3.2
lrwxrwxrwx  1 root root    28 Sep 20  2018 libopencv_imgcodecs.so.3.2 -> libopencv_imgcodecs.so.3.2.0
-rw-r--r--  1 root root  227K Sep 20  2018 libopencv_imgcodecs.so.3.2.0
-rw-r--r--  1 root root  5.5M Sep 20  2018 libopencv_imgproc.a
lrwxrwxrwx  1 root root    24 Sep 20  2018 libopencv_imgproc.so -> libopencv_imgproc.so.3.2
lrwxrwxrwx  1 root root    26 Sep 20  2018 libopencv_imgproc.so.3.2 -> libopencv_imgproc.so.3.2.0
-rw-r--r--  1 root root  2.8M Sep 20  2018 libopencv_imgproc.so.3.2.0
-rw-r--r--  1 root root  312K Sep 20  2018 libopencv_line_descriptor.a
lrwxrwxrwx  1 root root    32 Sep 20  2018 libopencv_line_descriptor.so -> libopencv_line_descriptor.so.3.2
lrwxrwxrwx  1 root root    34 Sep 20  2018 libopencv_line_descriptor.so.3.2 -> libopencv_line_descriptor.so.3.2.0
-rw-r--r--  1 root root  159K Sep 20  2018 libopencv_line_descriptor.so.3.2.0
-rw-r--r--  1 root root  1.5M Sep 20  2018 libopencv_ml.a
lrwxrwxrwx  1 root root    19 Sep 20  2018 libopencv_ml.so -> libopencv_ml.so.3.2
lrwxrwxrwx  1 root root    21 Sep 20  2018 libopencv_ml.so.3.2 -> libopencv_ml.so.3.2.0
-rw-r--r--  1 root root  679K Sep 20  2018 libopencv_ml.so.3.2.0
-rw-r--r--  1 root root  689K Sep 20  2018 libopencv_objdetect.a
lrwxrwxrwx  1 root root    26 Sep 20  2018 libopencv_objdetect.so -> libopencv_objdetect.so.3.2
lrwxrwxrwx  1 root root    28 Sep 20  2018 libopencv_objdetect.so.3.2 -> libopencv_objdetect.so.3.2.0
-rw-r--r--  1 root root  343K Sep 20  2018 libopencv_objdetect.so.3.2.0
-rw-r--r--  1 root root  648K Sep 20  2018 libopencv_optflow.a
lrwxrwxrwx  1 root root    24 Sep 20  2018 libopencv_optflow.so -> libopencv_optflow.so.3.2
lrwxrwxrwx  1 root root    26 Sep 20  2018 libopencv_optflow.so.3.2 -> libopencv_optflow.so.3.2.0
-rw-r--r--  1 root root  327K Sep 20  2018 libopencv_optflow.so.3.2.0
-rw-r--r--  1 root root   62K Sep 20  2018 libopencv_phase_unwrapping.a
lrwxrwxrwx  1 root root    33 Sep 20  2018 libopencv_phase_unwrapping.so -> libopencv_phase_unwrapping.so.3.2
lrwxrwxrwx  1 root root    35 Sep 20  2018 libopencv_phase_unwrapping.so.3.2 -> libopencv_phase_unwrapping.so.3.2.0
-rw-r--r--  1 root root   35K Sep 20  2018 libopencv_phase_unwrapping.so.3.2.0
-rw-r--r--  1 root root  1.5M Sep 20  2018 libopencv_photo.a
lrwxrwxrwx  1 root root    22 Sep 20  2018 libopencv_photo.so -> libopencv_photo.so.3.2
lrwxrwxrwx  1 root root    24 Sep 20  2018 libopencv_photo.so.3.2 -> libopencv_photo.so.3.2.0
-rw-r--r--  1 root root  771K Sep 20  2018 libopencv_photo.so.3.2.0
-rw-r--r--  1 root root   66K Sep 20  2018 libopencv_plot.a
lrwxrwxrwx  1 root root    21 Sep 20  2018 libopencv_plot.so -> libopencv_plot.so.3.2
lrwxrwxrwx  1 root root    23 Sep 20  2018 libopencv_plot.so.3.2 -> libopencv_plot.so.3.2.0
-rw-r--r--  1 root root   43K Sep 20  2018 libopencv_plot.so.3.2.0
-rw-r--r--  1 root root  268K Sep 20  2018 libopencv_reg.a
lrwxrwxrwx  1 root root    20 Sep 20  2018 libopencv_reg.so -> libopencv_reg.so.3.2
lrwxrwxrwx  1 root root    22 Sep 20  2018 libopencv_reg.so.3.2 -> libopencv_reg.so.3.2.0
-rw-r--r--  1 root root  127K Sep 20  2018 libopencv_reg.so.3.2.0
-rw-r--r--  1 root root  982K Sep 20  2018 libopencv_rgbd.a
lrwxrwxrwx  1 root root    21 Sep 20  2018 libopencv_rgbd.so -> libopencv_rgbd.so.3.2
lrwxrwxrwx  1 root root    23 Sep 20  2018 libopencv_rgbd.so.3.2 -> libopencv_rgbd.so.3.2.0
-rw-r--r--  1 root root  535K Sep 20  2018 libopencv_rgbd.so.3.2.0
-rw-r--r--  1 root root  353K Sep 20  2018 libopencv_saliency.a
lrwxrwxrwx  1 root root    25 Sep 20  2018 libopencv_saliency.so -> libopencv_saliency.so.3.2
lrwxrwxrwx  1 root root    27 Sep 20  2018 libopencv_saliency.so.3.2 -> libopencv_saliency.so.3.2.0
-rw-r--r--  1 root root  179K Sep 20  2018 libopencv_saliency.so.3.2.0
-rw-r--r--  1 root root  428K Sep 20  2018 libopencv_shape.a
lrwxrwxrwx  1 root root    22 Sep 20  2018 libopencv_shape.so -> libopencv_shape.so.3.2
lrwxrwxrwx  1 root root    24 Sep 20  2018 libopencv_shape.so.3.2 -> libopencv_shape.so.3.2.0
-rw-r--r--  1 root root  187K Sep 20  2018 libopencv_shape.so.3.2.0
-rw-r--r--  1 root root  327K Sep 20  2018 libopencv_stereo.a
lrwxrwxrwx  1 root root    23 Sep 20  2018 libopencv_stereo.so -> libopencv_stereo.so.3.2
lrwxrwxrwx  1 root root    25 Sep 20  2018 libopencv_stereo.so.3.2 -> libopencv_stereo.so.3.2.0
-rw-r--r--  1 root root  123K Sep 20  2018 libopencv_stereo.so.3.2.0
-rw-r--r--  1 root root  1.1M Sep 20  2018 libopencv_stitching.a
lrwxrwxrwx  1 root root    26 Sep 20  2018 libopencv_stitching.so -> libopencv_stitching.so.3.2
lrwxrwxrwx  1 root root    28 Sep 20  2018 libopencv_stitching.so.3.2 -> libopencv_stitching.so.3.2.0
-rw-r--r--  1 root root  519K Sep 20  2018 libopencv_stitching.so.3.2.0
-rw-r--r--  1 root root  156K Sep 20  2018 libopencv_structured_light.a
lrwxrwxrwx  1 root root    33 Sep 20  2018 libopencv_structured_light.so -> libopencv_structured_light.so.3.2
lrwxrwxrwx  1 root root    35 Sep 20  2018 libopencv_structured_light.so.3.2 -> libopencv_structured_light.so.3.2.0
-rw-r--r--  1 root root   91K Sep 20  2018 libopencv_structured_light.so.3.2.0
-rw-r--r--  1 root root  295K Sep 20  2018 libopencv_superres.a
lrwxrwxrwx  1 root root    25 Sep 20  2018 libopencv_superres.so -> libopencv_superres.so.3.2
lrwxrwxrwx  1 root root    27 Sep 20  2018 libopencv_superres.so.3.2 -> libopencv_superres.so.3.2.0
-rw-r--r--  1 root root  151K Sep 20  2018 libopencv_superres.so.3.2.0
-rw-r--r--  1 root root  1.1M Sep 20  2018 libopencv_surface_matching.a
lrwxrwxrwx  1 root root    33 Sep 20  2018 libopencv_surface_matching.so -> libopencv_surface_matching.so.3.2
lrwxrwxrwx  1 root root    35 Sep 20  2018 libopencv_surface_matching.so.3.2 -> libopencv_surface_matching.so.3.2.0
-rw-r--r--  1 root root  259K Sep 20  2018 libopencv_surface_matching.so.3.2.0
-rw-r--r--  1 root root  761K Sep 20  2018 libopencv_text.a
lrwxrwxrwx  1 root root    21 Sep 20  2018 libopencv_text.so -> libopencv_text.so.3.2
lrwxrwxrwx  1 root root    23 Sep 20  2018 libopencv_text.so.3.2 -> libopencv_text.so.3.2.0
-rw-r--r--  1 root root  367K Sep 20  2018 libopencv_text.so.3.2.0
-rw-r--r--  1 root root  1.6M Sep 20  2018 libopencv_ts.a
-rw-r--r--  1 root root  700K Sep 20  2018 libopencv_video.a
lrwxrwxrwx  1 root root    22 Sep 20  2018 libopencv_video.so -> libopencv_video.so.3.2
lrwxrwxrwx  1 root root    24 Sep 20  2018 libopencv_video.so.3.2 -> libopencv_video.so.3.2.0
-rw-r--r--  1 root root  395K Sep 20  2018 libopencv_video.so.3.2.0
-rw-r--r--  1 root root  488K Sep 20  2018 libopencv_videoio.a
lrwxrwxrwx  1 root root    24 Sep 20  2018 libopencv_videoio.so -> libopencv_videoio.so.3.2
lrwxrwxrwx  1 root root    26 Sep 20  2018 libopencv_videoio.so.3.2 -> libopencv_videoio.so.3.2.0
-rw-r--r--  1 root root  211K Sep 20  2018 libopencv_videoio.so.3.2.0
-rw-r--r--  1 root root  688K Sep 20  2018 libopencv_videostab.a
lrwxrwxrwx  1 root root    26 Sep 20  2018 libopencv_videostab.so -> libopencv_videostab.so.3.2
lrwxrwxrwx  1 root root    28 Sep 20  2018 libopencv_videostab.so.3.2 -> libopencv_videostab.so.3.2.0
-rw-r--r--  1 root root  347K Sep 20  2018 libopencv_videostab.so.3.2.0
-rw-r--r--  1 root root 1019K Sep 20  2018 libopencv_viz.a
lrwxrwxrwx  1 root root    20 Sep 20  2018 libopencv_viz.so -> libopencv_viz.so.3.2
lrwxrwxrwx  1 root root    22 Sep 20  2018 libopencv_viz.so.3.2 -> libopencv_viz.so.3.2.0
-rw-r--r--  1 root root  363K Sep 20  2018 libopencv_viz.so.3.2.0
-rw-r--r--  1 root root  2.3M Sep 20  2018 libopencv_ximgproc.a
lrwxrwxrwx  1 root root    25 Sep 20  2018 libopencv_ximgproc.so -> libopencv_ximgproc.so.3.2
lrwxrwxrwx  1 root root    27 Sep 20  2018 libopencv_ximgproc.so.3.2 -> libopencv_ximgproc.so.3.2.0
-rw-r--r--  1 root root  983K Sep 20  2018 libopencv_ximgproc.so.3.2.0
-rw-r--r--  1 root root  186K Sep 20  2018 libopencv_xobjdetect.a
lrwxrwxrwx  1 root root    27 Sep 20  2018 libopencv_xobjdetect.so -> libopencv_xobjdetect.so.3.2
lrwxrwxrwx  1 root root    29 Sep 20  2018 libopencv_xobjdetect.so.3.2 -> libopencv_xobjdetect.so.3.2.0
-rw-r--r--  1 root root   87K Sep 20  2018 libopencv_xobjdetect.so.3.2.0
-rw-r--r--  1 root root  441K Sep 20  2018 libopencv_xphoto.a
lrwxrwxrwx  1 root root    23 Sep 20  2018 libopencv_xphoto.so -> libopencv_xphoto.so.3.2
lrwxrwxrwx  1 root root    25 Sep 20  2018 libopencv_xphoto.so.3.2 -> libopencv_xphoto.so.3.2.0
-rw-r--r--  1 root root  220K Sep 20  2018 libopencv_xphoto.so.3.2.0
-rw-r--r--  1 root root  343K Aug 20  2019 libopenjp2.so.2.3.0
lrwxrwxrwx  1 root root    19 Aug 20  2019 libopenjp2.so.7 -> libopenjp2.so.2.3.0
lrwxrwxrwx  1 root root    19 Feb  4  2018 libopenmpt.so.0 -> libopenmpt.so.0.1.1
-rw-r--r--  1 root root  1.5M Feb  4  2018 libopenmpt.so.0.1.1
lrwxrwxrwx  1 root root    16 Feb 17  2016 libopus.so.0 -> libopus.so.0.5.2
-rw-r--r--  1 root root  295K Feb 17  2016 libopus.so.0.5.2
lrwxrwxrwx  1 root root    32 Feb  5  2018 liboshmem.so.20 -> openmpi/lib/liboshmem.so.20.10.1
lrwxrwxrwx  1 root root    19 Jan  4  2021 libp11-kit.so.0 -> libp11-kit.so.0.3.0
-rw-r--r--  1 root root  1.2M Jan  4  2021 libp11-kit.so.0.3.0
lrwxrwxrwx  1 root root    15 May 23  2018 libpanel.so.5 -> libpanel.so.5.9
-rw-r--r--  1 root root   14K May 23  2018 libpanel.so.5.9
lrwxrwxrwx  1 root root    16 May 23  2018 libpanelw.so.5 -> libpanelw.so.5.9
-rw-r--r--  1 root root   14K May 23  2018 libpanelw.so.5.9
lrwxrwxrwx  1 root root    25 Aug 21  2018 libpango-1.0.so.0 -> libpango-1.0.so.0.4000.14
-rw-r--r--  1 root root  306K Aug 21  2018 libpango-1.0.so.0.4000.14
lrwxrwxrwx  1 root root    30 Aug 21  2018 libpangocairo-1.0.so.0 -> libpangocairo-1.0.so.0.4000.14
-rw-r--r--  1 root root   50K Aug 21  2018 libpangocairo-1.0.so.0.4000.14
lrwxrwxrwx  1 root root    28 Aug 21  2018 libpangoft2-1.0.so.0 -> libpangoft2-1.0.so.0.4000.14
-rw-r--r--  1 root root   86K Aug 21  2018 libpangoft2-1.0.so.0.4000.14
lrwxrwxrwx  1 root root    20 Mar 24  2018 libpathplan.so -> libpathplan.so.4.0.0
lrwxrwxrwx  1 root root    20 Mar 24  2018 libpathplan.so.4 -> libpathplan.so.4.0.0
-rw-r--r--  1 root root   31K Mar 24  2018 libpathplan.so.4.0.0
lrwxrwxrwx  1 root root    22 Mar 18  2018 libpciaccess.so.0 -> libpciaccess.so.0.11.1
-rw-r--r--  1 root root   35K Mar 18  2018 libpciaccess.so.0.11.1
-rw-r--r--  1 root root  615K May 17 07:42 libpcre.a
lrwxrwxrwx  1 root root    34 May 17 07:42 libpcre.so -> /lib/x86_64-linux-gnu/libpcre.so.3
-rw-r--r--  1 root root  576K May 17 07:42 libpcre16.a
lrwxrwxrwx  1 root root    19 May 17 07:42 libpcre16.so -> libpcre16.so.3.13.3
lrwxrwxrwx  1 root root    19 May 17 07:42 libpcre16.so.3 -> libpcre16.so.3.13.3
-rw-r--r--  1 root root  410K May 17 07:42 libpcre16.so.3.13.3
-rw-r--r--  1 root root  558K May 17 07:42 libpcre32.a
lrwxrwxrwx  1 root root    19 May 17 07:42 libpcre32.so -> libpcre32.so.3.13.3
lrwxrwxrwx  1 root root    19 May 17 07:42 libpcre32.so.3 -> libpcre32.so.3.13.3
-rw-r--r--  1 root root  390K May 17 07:42 libpcre32.so.3.13.3
-rw-r--r--  1 root root   51K May 17 07:42 libpcrecpp.a
lrwxrwxrwx  1 root root    19 May 17 07:42 libpcrecpp.so -> libpcrecpp.so.0.0.1
lrwxrwxrwx  1 root root    19 May 17 07:42 libpcrecpp.so.0 -> libpcrecpp.so.0.0.1
-rw-r--r--  1 root root   34K May 17 07:42 libpcrecpp.so.0.0.1
-rw-r--r--  1 root root  6.6K May 17 07:42 libpcreposix.a
lrwxrwxrwx  1 root root    22 May 17 07:42 libpcreposix.so -> libpcreposix.so.3.13.3
lrwxrwxrwx  1 root root    22 May 17 07:42 libpcreposix.so.3 -> libpcreposix.so.3.13.3
-rw-r--r--  1 root root  9.9K May 17 07:42 libpcreposix.so.3.13.3
lrwxrwxrwx  1 root root    17 Oct 19  2020 libperl.so.5.26 -> libperl.so.5.26.1
-rw-r--r--  1 root root  2.0M Oct 19  2020 libperl.so.5.26.1
lrwxrwxrwx  1 root root    21 Dec 17  2017 libpixman-1.so.0 -> libpixman-1.so.0.34.0
-rw-r--r--  1 root root  659K Dec 17  2017 libpixman-1.so.0.34.0
-rw-r--r--  1 root root   19K Feb 22  2018 libplc4.so
-rw-r--r--  1 root root   15K Feb 22  2018 libplds4.so
lrwxrwxrwx  1 root root    10 Apr 30  2019 libpng.a -> libpng16.a
lrwxrwxrwx  1 root root    11 Apr 30  2019 libpng.so -> libpng16.so
-rw-r--r--  1 root root  328K Apr 30  2019 libpng16.a
lrwxrwxrwx  1 root root    19 Apr 30  2019 libpng16.so -> libpng16.so.16.34.0
lrwxrwxrwx  1 root root    19 Apr 30  2019 libpng16.so.16 -> libpng16.so.16.34.0
-rw-r--r--  1 root root  198K Apr 30  2019 libpng16.so.16.34.0
lrwxrwxrwx  1 root root    20 Nov 26  2020 libpoppler.so.73 -> libpoppler.so.73.0.0
-rw-r--r--  1 root root  2.6M Nov 26  2020 libpoppler.so.73.0.0
lrwxrwxrwx  1 root root    13 May 18 01:58 libpq.so.5 -> libpq.so.5.10
-rw-r--r--  1 root root  290K May 18 01:58 libpq.so.5.10
lrwxrwxrwx  1 root root    17 Jun 18  2017 libproj.so.12 -> libproj.so.12.0.0
-rw-r--r--  1 root root  420K Jun 18  2017 libproj.so.12.0.0
lrwxrwxrwx  1 root root    17 Dec  8  2020 libproxy.so.1 -> libproxy.so.1.0.0
-rw-r--r--  1 root root  123K Dec  8  2020 libproxy.so.1.0.0
lrwxrwxrwx  1 root root    15 Mar  4  2018 libpsl.so.5 -> libpsl.so.5.2.0
-rw-r--r--  1 root root   54K Mar  4  2018 libpsl.so.5.2.0
lrwxrwxrwx  1 root root    40 Aug 16 08:41 libpsm_infinipath.so.1 -> /etc/alternatives/libpsm_infinipath.so.1
-rw-r--r--  1 root root  6.0M May  3 10:19 libpthread.a
-rw-r--r--  1 root root   252 May  3 10:19 libpthread.so
-rw-r--r--  1 root root   29K May  3 10:19 libpthread_nonshared.a
lrwxrwxrwx  1 root root    19 Jul  1 15:56 libpython2.7.so.1 -> libpython2.7.so.1.0
-rw-r--r--  1 root root  3.4M Jul  1 15:56 libpython2.7.so.1.0
lrwxrwxrwx  1 root root    17 Dec  9  2017 libqhull.so.7 -> libqhull.so.7.2.0
-rw-r--r--  1 root root  358K Dec  9  2017 libqhull.so.7.2.0
lrwxrwxrwx  1 root root    20 Mar 10  2020 libquadmath.so.0 -> libquadmath.so.0.0.0
-rw-r--r--  1 root root  256K Mar 10  2020 libquadmath.so.0.0.0
-rw-r--r--  1 root root   91K Apr 26  2016 libraw1394.a
-rw-r--r--  1 root root   969 Apr 26  2016 libraw1394.la
lrwxrwxrwx  1 root root    20 Apr 26  2016 libraw1394.so -> libraw1394.so.11.1.0
lrwxrwxrwx  1 root root    20 Apr 26  2016 libraw1394.so.11 -> libraw1394.so.11.1.0
-rw-r--r--  1 root root   59K Apr 26  2016 libraw1394.so.11.1.0
lrwxrwxrwx  1 root root    21 Aug  5  2019 librdmacm.so.1 -> librdmacm.so.1.1.17.1
-rw-r--r--  1 root root   87K Aug  5  2019 librdmacm.so.1.1.17.1
-rw-r--r--  1 root root  136K May  3 10:19 libresolv.a
lrwxrwxrwx  1 root root    36 May  3 10:19 libresolv.so -> /lib/x86_64-linux-gnu/libresolv.so.2
lrwxrwxrwx  1 root root    20 Oct 24  2016 librest-0.7.so.0 -> librest-0.7.so.0.0.0
-rw-r--r--  1 root root   88K Oct 24  2016 librest-0.7.so.0.0.0
lrwxrwxrwx  1 root root    18 Dec 15  2017 libroken.so.18 -> libroken.so.18.1.0
-rw-r--r--  1 root root   87K Dec 15  2017 libroken.so.18.1.0
-rw-r--r--  1 root root   52K May  3 10:19 librpcsvc.a
lrwxrwxrwx  1 root root    20 Jul 28  2020 librsvg-2.so.2 -> librsvg-2.so.2.40.20
-rw-r--r--  1 root root  223K Jul 28  2020 librsvg-2.so.2.40.20
-rw-r--r--  1 root root   76K May  3 10:19 librt.a
lrwxrwxrwx  1 root root    32 May  3 10:19 librt.so -> /lib/x86_64-linux-gnu/librt.so.1
-rw-r--r--  1 root root  111K Apr 27  2016 librtmp.so.1
lrwxrwxrwx  1 root root    18 Feb 15  2022 libsasl2.so.2 -> libsasl2.so.2.0.25
-rw-r--r--  1 root root  107K Feb 15  2022 libsasl2.so.2.0.25
-rw-r--r--  1 root root  251K Mar  1  2018 libsemanage.so.1
lrwxrwxrwx  1 root root    19 Mar 31 20:53 libsensors.so.4 -> libsensors.so.4.4.0
-rw-r--r--  1 root root   59K Mar 31 20:53 libsensors.so.4.4.0
lrwxrwxrwx  1 root root    17 Aug  6  2017 libshine.so.3 -> libshine.so.3.0.1
-rw-r--r--  1 root root   43K Aug  6  2017 libshine.so.3.0.1
-rw-r--r--  1 root root  181K Jul  6 11:25 libsmime3.so
lrwxrwxrwx  1 root root    18 Sep  5  2017 libsnappy.so.1 -> libsnappy.so.1.1.7
-rw-r--r--  1 root root   31K Sep  5  2017 libsnappy.so.1.1.7
lrwxrwxrwx  1 root root    20 Oct  5  2017 libsocket++.so.1 -> libsocket++.so.1.0.2
-rw-r--r--  1 root root  132K Oct  5  2017 libsocket++.so.1.0.2
lrwxrwxrwx  1 root root    20 Oct  9  2019 libsoup-2.4.so.1 -> libsoup-2.4.so.1.8.0
-rw-r--r--  1 root root  971K Oct  9  2019 libsoup-2.4.so.1.8.0
lrwxrwxrwx  1 root root    26 Oct  9  2019 libsoup-gnome-2.4.so.1 -> libsoup-gnome-2.4.so.1.8.0
-rw-r--r--  1 root root  9.9K Oct  9  2019 libsoup-gnome-2.4.so.1.8.0
lrwxrwxrwx  1 root root    16 Nov  4  2017 libsoxr.so.0 -> libsoxr.so.0.1.1
-rw-r--r--  1 root root  183K Nov  4  2017 libsoxr.so.0.1.1
lrwxrwxrwx  1 root root    22 Nov 17  2016 libspatialite.so.7 -> libspatialite.so.7.1.0
-rw-r--r--  1 root root  5.6M Nov 17  2016 libspatialite.so.7.1.0
lrwxrwxrwx  1 root root    17 Feb  8  2022 libspeex.so.1 -> libspeex.so.1.5.0
-rw-r--r--  1 root root  102K Feb  8  2022 libspeex.so.1.5.0
lrwxrwxrwx  1 root root    19 Apr 28 14:04 libsqlite3.so.0 -> libsqlite3.so.0.8.6
-rw-r--r--  1 root root  1.1M Apr 28 14:04 libsqlite3.so.0.8.6
lrwxrwxrwx  1 root root    22 Jul 31  2020 libssh-gcrypt.so.4 -> libssh-gcrypt.so.4.5.0
-rw-r--r--  1 root root  465K Jul 31  2020 libssh-gcrypt.so.4.5.0
lrwxrwxrwx  1 root root    30 Jul 31  2020 libssh-gcrypt_threads.so.4 -> libssh-gcrypt_threads.so.4.5.0
-rw-r--r--  1 root root  6.0K Jul 31  2020 libssh-gcrypt_threads.so.4.5.0
-rw-r--r--  1 root root  564K Jul  4 11:25 libssl.so.1.1
-rw-r--r--  1 root root  321K Jul  6 11:25 libssl3.so
lrwxrwxrwx  1 root root    19 Mar 10  2020 libstdc++.so.6 -> libstdc++.so.6.0.25
-rw-r--r--  1 root root  1.6M Mar 10  2020 libstdc++.so.6.0.25
lrwxrwxrwx  1 root root    19 Nov  6  2017 libsuperlu.so.5 -> libsuperlu.so.5.2.1
-rw-r--r--  1 root root  447K Nov  6  2017 libsuperlu.so.5.2.1
-rw-r--r--  1 root root  220K May 18 20:01 libswresample.a
lrwxrwxrwx  1 root root    24 May 18 20:01 libswresample.so -> libswresample.so.2.9.100
lrwxrwxrwx  1 root root    24 May 18 20:01 libswresample.so.2 -> libswresample.so.2.9.100
-rw-r--r--  1 root root  122K May 18 20:01 libswresample.so.2.9.100
-rw-r--r--  1 root root  780K May 18 20:01 libswscale.a
lrwxrwxrwx  1 root root    21 May 18 20:01 libswscale.so -> libswscale.so.4.8.100
lrwxrwxrwx  1 root root    21 May 18 20:01 libswscale.so.4 -> libswscale.so.4.8.100
-rw-r--r--  1 root root  538K May 18 20:01 libswscale.so.4.8.100
lrwxrwxrwx  1 root root    14 Jul 28  2017 libsz.so.2 -> libsz.so.2.0.1
-rw-r--r--  1 root root  9.9K Jul 28  2017 libsz.so.2.0.1
lrwxrwxrwx  1 root root    17 Jan 21  2018 libtasn1.so.6 -> libtasn1.so.6.5.5
-rw-r--r--  1 root root   74K Jan 21  2018 libtasn1.so.6.5.5
lrwxrwxrwx  1 root root    11 Oct  4  2017 libtbb.so -> libtbb.so.2
-rw-r--r--  1 root root  230K Oct  4  2017 libtbb.so.2
lrwxrwxrwx  1 root root    17 Oct  4  2017 libtbbmalloc.so -> libtbbmalloc.so.2
-rw-r--r--  1 root root  110K Oct  4  2017 libtbbmalloc.so.2
lrwxrwxrwx  1 root root    23 Oct  4  2017 libtbbmalloc_proxy.so -> libtbbmalloc_proxy.so.2
-rw-r--r--  1 root root   11K Oct  4  2017 libtbbmalloc_proxy.so.2
-rw-r--r--  1 root root  1.7M Feb 22  2018 libtcl8.6.so
lrwxrwxrwx  1 root root    12 Feb 22  2018 libtcl8.6.so.0 -> libtcl8.6.so
lrwxrwxrwx  1 root root    21 Apr  7  2018 libtesseract.so.4 -> libtesseract.so.4.0.0
-rw-r--r--  1 root root  3.1M Apr  7  2018 libtesseract.so.4.0.0
lrwxrwxrwx  1 root root    16 Feb 16  2018 libthai.so.0 -> libthai.so.0.3.0
-rw-r--r--  1 root root   37K Feb 16  2018 libthai.so.0.3.0
lrwxrwxrwx  1 root root    19 May 27  2016 libtheora.so.0 -> libtheora.so.0.3.10
-rw-r--r--  1 root root  327K May 27  2016 libtheora.so.0.3.10
lrwxrwxrwx  1 root root    21 May 27  2016 libtheoradec.so.1 -> libtheoradec.so.1.1.4
-rw-r--r--  1 root root  119K May 27  2016 libtheoradec.so.1.1.4
lrwxrwxrwx  1 root root    21 May 27  2016 libtheoraenc.so.1 -> libtheoraenc.so.1.1.2
-rw-r--r--  1 root root  251K May 27  2016 libtheoraenc.so.1.1.2
lrwxrwxrwx  1 root root    39 May  3 10:19 libthread_db.so -> /lib/x86_64-linux-gnu/libthread_db.so.1
lrwxrwxrwx  1 root root    13 May 23  2018 libtic.so.5 -> libtic.so.5.9
-rw-r--r--  1 root root   62K May 23  2018 libtic.so.5.9
-rw-r--r--  1 root root  752K May 11 15:09 libtiff.a
lrwxrwxrwx  1 root root    16 May 11 15:09 libtiff.so -> libtiff.so.5.3.0
lrwxrwxrwx  1 root root    16 May 11 15:09 libtiff.so.5 -> libtiff.so.5.3.0
-rw-r--r--  1 root root  479K May 11 15:09 libtiff.so.5.3.0
-rw-r--r--  1 root root  8.5K May 11 15:09 libtiffxx.a
lrwxrwxrwx  1 root root    18 May 11 15:09 libtiffxx.so -> libtiffxx.so.5.3.0
lrwxrwxrwx  1 root root    18 May 11 15:09 libtiffxx.so.5 -> libtiffxx.so.5.3.0
-rw-r--r--  1 root root   10K May 11 15:09 libtiffxx.so.5.3.0
-rw-r--r--  1 root root  1.4M Mar 24  2018 libtk8.6.so
lrwxrwxrwx  1 root root    11 Mar 24  2018 libtk8.6.so.0 -> libtk8.6.so
lrwxrwxrwx  1 root root    16 Mar 10  2020 libtsan.so.0 -> libtsan.so.0.0.0
-rw-r--r--  1 root root  942K Mar 10  2020 libtsan.so.0.0.0
-rw-r--r--  1 root root  3.6K Mar 10  2020 libtsan_preinit.o
lrwxrwxrwx  1 root root    19 Aug 29  2017 libtwolame.so.0 -> libtwolame.so.0.0.0
-rw-r--r--  1 root root  122K Aug 29  2017 libtwolame.so.0.0.0
lrwxrwxrwx  1 root root    17 Dec  4  2019 libubsan.so.0 -> libubsan.so.0.0.0
-rw-r--r--  1 root root  322K Dec  4  2019 libubsan.so.0.0.0
lrwxrwxrwx  1 root root    21 Mar 21  2019 libunistring.so.2 -> libunistring.so.2.1.0
-rw-r--r--  1 root root  1.5M Mar 21  2019 libunistring.so.2.1.0
lrwxrwxrwx  1 root root    22 Jul 12 13:52 liburiparser.so.1 -> liburiparser.so.1.0.20
-rw-r--r--  1 root root  110K Jul 12 13:52 liburiparser.so.1.0.20
-rw-r--r--  1 root root   15K May  3 10:19 libutil.a
lrwxrwxrwx  1 root root    34 May  3 10:19 libutil.so -> /lib/x86_64-linux-gnu/libutil.so.1
lrwxrwxrwx  1 root root    20 Apr  3  2018 libva-drm.so.2 -> libva-drm.so.2.100.0
-rw-r--r--  1 root root   11K Apr  3  2018 libva-drm.so.2.100.0
lrwxrwxrwx  1 root root    20 Apr  3  2018 libva-x11.so.2 -> libva-x11.so.2.100.0
-rw-r--r--  1 root root   23K Apr  3  2018 libva-x11.so.2.100.0
lrwxrwxrwx  1 root root    16 Apr  3  2018 libva.so.2 -> libva.so.2.100.0
-rw-r--r--  1 root root  131K Apr  3  2018 libva.so.2.100.0
lrwxrwxrwx  1 root root    17 Feb 19  2016 libvdpau.so.1 -> libvdpau.so.1.0.0
-rw-r--r--  1 root root   15K Feb 19  2016 libvdpau.so.1.0.0
lrwxrwxrwx  1 root root    18 Mar 16  2018 libvorbis.so.0 -> libvorbis.so.0.4.8
-rw-r--r--  1 root root  171K Mar 16  2018 libvorbis.so.0.4.8
lrwxrwxrwx  1 root root    22 Mar 16  2018 libvorbisenc.so.2 -> libvorbisenc.so.2.0.11
-rw-r--r--  1 root root  675K Mar 16  2018 libvorbisenc.so.2.0.11
lrwxrwxrwx  1 root root    22 Mar 16  2018 libvorbisfile.so.3 -> libvorbisfile.so.3.3.7
-rw-r--r--  1 root root   31K Mar 16  2018 libvorbisfile.so.3.3.7
lrwxrwxrwx  1 root root    15 Nov 19  2019 libvpx.so.5 -> libvpx.so.5.0.0
lrwxrwxrwx  1 root root    15 Nov 19  2019 libvpx.so.5.0 -> libvpx.so.5.0.0
-rw-r--r--  1 root root  2.3M Nov 19  2019 libvpx.so.5.0.0
lrwxrwxrwx  1 root root    29 Dec 17  2017 libvtkChartsCore-6.3.so.6.3 -> libvtkChartsCore-6.3.so.6.3.0
-rw-r--r--  1 root root  1.4M Dec 17  2017 libvtkChartsCore-6.3.so.6.3.0
lrwxrwxrwx  1 root root    38 Dec 17  2017 libvtkChartsCorePython27D-6.3.so.6.3 -> libvtkChartsCorePython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  939K Dec 17  2017 libvtkChartsCorePython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    32 Dec 17  2017 libvtkChartsCoreTCL-6.3.so.6.3 -> libvtkChartsCoreTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  573K Dec 17  2017 libvtkChartsCoreTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    30 Dec 17  2017 libvtkCommonColor-6.3.so.6.3 -> libvtkCommonColor-6.3.so.6.3.0
-rw-r--r--  1 root root  106K Dec 17  2017 libvtkCommonColor-6.3.so.6.3.0
lrwxrwxrwx  1 root root    39 Dec 17  2017 libvtkCommonColorPython27D-6.3.so.6.3 -> libvtkCommonColorPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   70K Dec 17  2017 libvtkCommonColorPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    33 Dec 17  2017 libvtkCommonColorTCL-6.3.so.6.3 -> libvtkCommonColorTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   27K Dec 17  2017 libvtkCommonColorTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    46 Dec 17  2017 libvtkCommonComputationalGeometry-6.3.so.6.3 -> libvtkCommonComputationalGeometry-6.3.so.6.3.0
-rw-r--r--  1 root root  240K Dec 17  2017 libvtkCommonComputationalGeometry-6.3.so.6.3.0
lrwxrwxrwx  1 root root    55 Dec 17  2017 libvtkCommonComputationalGeometryPython27D-6.3.so.6.3 -> libvtkCommonComputationalGeometryPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  423K Dec 17  2017 libvtkCommonComputationalGeometryPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    49 Dec 17  2017 libvtkCommonComputationalGeometryTCL-6.3.so.6.3 -> libvtkCommonComputationalGeometryTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  259K Dec 17  2017 libvtkCommonComputationalGeometryTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    29 Dec 17  2017 libvtkCommonCore-6.3.so.6.3 -> libvtkCommonCore-6.3.so.6.3.0
-rw-r--r--  1 root root  3.2M Dec 17  2017 libvtkCommonCore-6.3.so.6.3.0
lrwxrwxrwx  1 root root    38 Dec 17  2017 libvtkCommonCorePython27D-6.3.so.6.3 -> libvtkCommonCorePython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  3.0M Dec 17  2017 libvtkCommonCorePython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    32 Dec 17  2017 libvtkCommonCoreTCL-6.3.so.6.3 -> libvtkCommonCoreTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  900K Dec 17  2017 libvtkCommonCoreTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    34 Dec 17  2017 libvtkCommonDataModel-6.3.so.6.3 -> libvtkCommonDataModel-6.3.so.6.3.0
-rw-r--r--  1 root root  3.4M Dec 17  2017 libvtkCommonDataModel-6.3.so.6.3.0
lrwxrwxrwx  1 root root    43 Dec 17  2017 libvtkCommonDataModelPython27D-6.3.so.6.3 -> libvtkCommonDataModelPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  4.2M Dec 17  2017 libvtkCommonDataModelPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    37 Dec 17  2017 libvtkCommonDataModelTCL-6.3.so.6.3 -> libvtkCommonDataModelTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  2.2M Dec 17  2017 libvtkCommonDataModelTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    39 Dec 17  2017 libvtkCommonExecutionModel-6.3.so.6.3 -> libvtkCommonExecutionModel-6.3.so.6.3.0
-rw-r--r--  1 root root  669K Dec 17  2017 libvtkCommonExecutionModel-6.3.so.6.3.0
lrwxrwxrwx  1 root root    48 Dec 17  2017 libvtkCommonExecutionModelPython27D-6.3.so.6.3 -> libvtkCommonExecutionModelPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  621K Dec 17  2017 libvtkCommonExecutionModelPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    42 Dec 17  2017 libvtkCommonExecutionModelTCL-6.3.so.6.3 -> libvtkCommonExecutionModelTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  474K Dec 17  2017 libvtkCommonExecutionModelTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    29 Dec 17  2017 libvtkCommonMath-6.3.so.6.3 -> libvtkCommonMath-6.3.so.6.3.0
-rw-r--r--  1 root root  128K Dec 17  2017 libvtkCommonMath-6.3.so.6.3.0
lrwxrwxrwx  1 root root    38 Dec 17  2017 libvtkCommonMathPython27D-6.3.so.6.3 -> libvtkCommonMathPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  347K Dec 17  2017 libvtkCommonMathPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    32 Dec 17  2017 libvtkCommonMathTCL-6.3.so.6.3 -> libvtkCommonMathTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  103K Dec 17  2017 libvtkCommonMathTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    29 Dec 17  2017 libvtkCommonMisc-6.3.so.6.3 -> libvtkCommonMisc-6.3.so.6.3.0
-rw-r--r--  1 root root   88K Dec 17  2017 libvtkCommonMisc-6.3.so.6.3.0
lrwxrwxrwx  1 root root    38 Dec 17  2017 libvtkCommonMiscPython27D-6.3.so.6.3 -> libvtkCommonMiscPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   88K Dec 17  2017 libvtkCommonMiscPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    32 Dec 17  2017 libvtkCommonMiscTCL-6.3.so.6.3 -> libvtkCommonMiscTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   47K Dec 17  2017 libvtkCommonMiscTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    31 Dec 17  2017 libvtkCommonSystem-6.3.so.6.3 -> libvtkCommonSystem-6.3.so.6.3.0
-rw-r--r--  1 root root   76K Dec 17  2017 libvtkCommonSystem-6.3.so.6.3.0
lrwxrwxrwx  1 root root    40 Dec 17  2017 libvtkCommonSystemPython27D-6.3.so.6.3 -> libvtkCommonSystemPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   71K Dec 17  2017 libvtkCommonSystemPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    34 Dec 17  2017 libvtkCommonSystemTCL-6.3.so.6.3 -> libvtkCommonSystemTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   67K Dec 17  2017 libvtkCommonSystemTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    35 Dec 17  2017 libvtkCommonTransforms-6.3.so.6.3 -> libvtkCommonTransforms-6.3.so.6.3.0
-rw-r--r--  1 root root  184K Dec 17  2017 libvtkCommonTransforms-6.3.so.6.3.0
lrwxrwxrwx  1 root root    44 Dec 17  2017 libvtkCommonTransformsPython27D-6.3.so.6.3 -> libvtkCommonTransformsPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  263K Dec 17  2017 libvtkCommonTransformsPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    38 Dec 17  2017 libvtkCommonTransformsTCL-6.3.so.6.3 -> libvtkCommonTransformsTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  200K Dec 17  2017 libvtkCommonTransformsTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    30 Dec 17  2017 libvtkDICOMParser-6.3.so.6.3 -> libvtkDICOMParser-6.3.so.6.3.0
-rw-r--r--  1 root root   96K Dec 17  2017 libvtkDICOMParser-6.3.so.6.3.0
lrwxrwxrwx  1 root root    35 Dec 17  2017 libvtkDomainsChemistry-6.3.so.6.3 -> libvtkDomainsChemistry-6.3.so.6.3.0
-rw-r--r--  1 root root  293K Dec 17  2017 libvtkDomainsChemistry-6.3.so.6.3.0
lrwxrwxrwx  1 root root    44 Dec 17  2017 libvtkDomainsChemistryPython27D-6.3.so.6.3 -> libvtkDomainsChemistryPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  181K Dec 17  2017 libvtkDomainsChemistryPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    38 Dec 17  2017 libvtkDomainsChemistryTCL-6.3.so.6.3 -> libvtkDomainsChemistryTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  127K Dec 17  2017 libvtkDomainsChemistryTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    29 Dec 17  2017 libvtkFiltersAMR-6.3.so.6.3 -> libvtkFiltersAMR-6.3.so.6.3.0
-rw-r--r--  1 root root  273K Dec 17  2017 libvtkFiltersAMR-6.3.so.6.3.0
lrwxrwxrwx  1 root root    38 Dec 17  2017 libvtkFiltersAMRPython27D-6.3.so.6.3 -> libvtkFiltersAMRPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  183K Dec 17  2017 libvtkFiltersAMRPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    32 Dec 17  2017 libvtkFiltersAMRTCL-6.3.so.6.3 -> libvtkFiltersAMRTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  139K Dec 17  2017 libvtkFiltersAMRTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    30 Dec 17  2017 libvtkFiltersCore-6.3.so.6.3 -> libvtkFiltersCore-6.3.so.6.3.0
-rw-r--r--  1 root root  4.9M Dec 17  2017 libvtkFiltersCore-6.3.so.6.3.0
lrwxrwxrwx  1 root root    39 Dec 17  2017 libvtkFiltersCorePython27D-6.3.so.6.3 -> libvtkFiltersCorePython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  2.6M Dec 17  2017 libvtkFiltersCorePython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    33 Dec 17  2017 libvtkFiltersCoreTCL-6.3.so.6.3 -> libvtkFiltersCoreTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  1.5M Dec 17  2017 libvtkFiltersCoreTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    36 Dec 17  2017 libvtkFiltersExtraction-6.3.so.6.3 -> libvtkFiltersExtraction-6.3.so.6.3.0
-rw-r--r--  1 root root  1.1M Dec 17  2017 libvtkFiltersExtraction-6.3.so.6.3.0
lrwxrwxrwx  1 root root    45 Dec 17  2017 libvtkFiltersExtractionPython27D-6.3.so.6.3 -> libvtkFiltersExtractionPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  497K Dec 17  2017 libvtkFiltersExtractionPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    39 Dec 17  2017 libvtkFiltersExtractionTCL-6.3.so.6.3 -> libvtkFiltersExtractionTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  360K Dec 17  2017 libvtkFiltersExtractionTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    35 Dec 17  2017 libvtkFiltersFlowPaths-6.3.so.6.3 -> libvtkFiltersFlowPaths-6.3.so.6.3.0
-rw-r--r--  1 root root  473K Dec 17  2017 libvtkFiltersFlowPaths-6.3.so.6.3.0
lrwxrwxrwx  1 root root    44 Dec 17  2017 libvtkFiltersFlowPathsPython27D-6.3.so.6.3 -> libvtkFiltersFlowPathsPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  495K Dec 17  2017 libvtkFiltersFlowPathsPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    38 Dec 17  2017 libvtkFiltersFlowPathsTCL-6.3.so.6.3 -> libvtkFiltersFlowPathsTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  264K Dec 17  2017 libvtkFiltersFlowPathsTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    33 Dec 17  2017 libvtkFiltersGeneral-6.3.so.6.3 -> libvtkFiltersGeneral-6.3.so.6.3.0
-rw-r--r--  1 root root  3.2M Dec 17  2017 libvtkFiltersGeneral-6.3.so.6.3.0
lrwxrwxrwx  1 root root    42 Dec 17  2017 libvtkFiltersGeneralPython27D-6.3.so.6.3 -> libvtkFiltersGeneralPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  2.2M Dec 17  2017 libvtkFiltersGeneralPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    36 Dec 17  2017 libvtkFiltersGeneralTCL-6.3.so.6.3 -> libvtkFiltersGeneralTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  1.5M Dec 17  2017 libvtkFiltersGeneralTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    33 Dec 17  2017 libvtkFiltersGeneric-6.3.so.6.3 -> libvtkFiltersGeneric-6.3.so.6.3.0
-rw-r--r--  1 root root  296K Dec 17  2017 libvtkFiltersGeneric-6.3.so.6.3.0
lrwxrwxrwx  1 root root    42 Dec 17  2017 libvtkFiltersGenericPython27D-6.3.so.6.3 -> libvtkFiltersGenericPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  391K Dec 17  2017 libvtkFiltersGenericPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    36 Dec 17  2017 libvtkFiltersGenericTCL-6.3.so.6.3 -> libvtkFiltersGenericTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  220K Dec 17  2017 libvtkFiltersGenericTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    34 Dec 17  2017 libvtkFiltersGeometry-6.3.so.6.3 -> libvtkFiltersGeometry-6.3.so.6.3.0
-rw-r--r--  1 root root  546K Dec 17  2017 libvtkFiltersGeometry-6.3.so.6.3.0
lrwxrwxrwx  1 root root    43 Dec 17  2017 libvtkFiltersGeometryPython27D-6.3.so.6.3 -> libvtkFiltersGeometryPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  423K Dec 17  2017 libvtkFiltersGeometryPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    37 Dec 17  2017 libvtkFiltersGeometryTCL-6.3.so.6.3 -> libvtkFiltersGeometryTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  280K Dec 17  2017 libvtkFiltersGeometryTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    32 Dec 17  2017 libvtkFiltersHybrid-6.3.so.6.3 -> libvtkFiltersHybrid-6.3.so.6.3.0
-rw-r--r--  1 root root  867K Dec 17  2017 libvtkFiltersHybrid-6.3.so.6.3.0
lrwxrwxrwx  1 root root    41 Dec 17  2017 libvtkFiltersHybridPython27D-6.3.so.6.3 -> libvtkFiltersHybridPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  701K Dec 17  2017 libvtkFiltersHybridPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    35 Dec 17  2017 libvtkFiltersHybridTCL-6.3.so.6.3 -> libvtkFiltersHybridTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  432K Dec 17  2017 libvtkFiltersHybridTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    35 Dec 17  2017 libvtkFiltersHyperTree-6.3.so.6.3 -> libvtkFiltersHyperTree-6.3.so.6.3.0
-rw-r--r--  1 root root  297K Dec 17  2017 libvtkFiltersHyperTree-6.3.so.6.3.0
lrwxrwxrwx  1 root root    44 Dec 17  2017 libvtkFiltersHyperTreePython27D-6.3.so.6.3 -> libvtkFiltersHyperTreePython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  218K Dec 17  2017 libvtkFiltersHyperTreePython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    38 Dec 17  2017 libvtkFiltersHyperTreeTCL-6.3.so.6.3 -> libvtkFiltersHyperTreeTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  163K Dec 17  2017 libvtkFiltersHyperTreeTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    33 Dec 17  2017 libvtkFiltersImaging-6.3.so.6.3 -> libvtkFiltersImaging-6.3.so.6.3.0
-rw-r--r--  1 root root  136K Dec 17  2017 libvtkFiltersImaging-6.3.so.6.3.0
lrwxrwxrwx  1 root root    42 Dec 17  2017 libvtkFiltersImagingPython27D-6.3.so.6.3 -> libvtkFiltersImagingPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  111K Dec 17  2017 libvtkFiltersImagingPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    36 Dec 17  2017 libvtkFiltersImagingTCL-6.3.so.6.3 -> libvtkFiltersImagingTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   71K Dec 17  2017 libvtkFiltersImagingTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    34 Dec 17  2017 libvtkFiltersModeling-6.3.so.6.3 -> libvtkFiltersModeling-6.3.so.6.3.0
-rw-r--r--  1 root root  481K Dec 17  2017 libvtkFiltersModeling-6.3.so.6.3.0
lrwxrwxrwx  1 root root    43 Dec 17  2017 libvtkFiltersModelingPython27D-6.3.so.6.3 -> libvtkFiltersModelingPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  607K Dec 17  2017 libvtkFiltersModelingPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    37 Dec 17  2017 libvtkFiltersModelingTCL-6.3.so.6.3 -> libvtkFiltersModelingTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  388K Dec 17  2017 libvtkFiltersModelingTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    34 Dec 17  2017 libvtkFiltersParallel-6.3.so.6.3 -> libvtkFiltersParallel-6.3.so.6.3.0
-rw-r--r--  1 root root  860K Dec 17  2017 libvtkFiltersParallel-6.3.so.6.3.0
lrwxrwxrwx  1 root root    43 Dec 17  2017 libvtkFiltersParallelFlowPaths-6.3.so.6.3 -> libvtkFiltersParallelFlowPaths-6.3.so.6.3.0
-rw-r--r--  1 root root  232K Dec 17  2017 libvtkFiltersParallelFlowPaths-6.3.so.6.3.0
lrwxrwxrwx  1 root root    52 Dec 17  2017 libvtkFiltersParallelFlowPathsPython27D-6.3.so.6.3 -> libvtkFiltersParallelFlowPathsPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   40K Dec 17  2017 libvtkFiltersParallelFlowPathsPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    46 Dec 17  2017 libvtkFiltersParallelFlowPathsTCL-6.3.so.6.3 -> libvtkFiltersParallelFlowPathsTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   43K Dec 17  2017 libvtkFiltersParallelFlowPathsTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    42 Dec 17  2017 libvtkFiltersParallelGeometry-6.3.so.6.3 -> libvtkFiltersParallelGeometry-6.3.so.6.3.0
-rw-r--r--  1 root root  265K Dec 17  2017 libvtkFiltersParallelGeometry-6.3.so.6.3.0
lrwxrwxrwx  1 root root    51 Dec 17  2017 libvtkFiltersParallelGeometryPython27D-6.3.so.6.3 -> libvtkFiltersParallelGeometryPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   65K Dec 17  2017 libvtkFiltersParallelGeometryPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    45 Dec 17  2017 libvtkFiltersParallelGeometryTCL-6.3.so.6.3 -> libvtkFiltersParallelGeometryTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   55K Dec 17  2017 libvtkFiltersParallelGeometryTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    41 Dec 17  2017 libvtkFiltersParallelImaging-6.3.so.6.3 -> libvtkFiltersParallelImaging-6.3.so.6.3.0
-rw-r--r--  1 root root  140K Dec 17  2017 libvtkFiltersParallelImaging-6.3.so.6.3.0
lrwxrwxrwx  1 root root    50 Dec 17  2017 libvtkFiltersParallelImagingPython27D-6.3.so.6.3 -> libvtkFiltersParallelImagingPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   45K Dec 17  2017 libvtkFiltersParallelImagingPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    44 Dec 17  2017 libvtkFiltersParallelImagingTCL-6.3.so.6.3 -> libvtkFiltersParallelImagingTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   47K Dec 17  2017 libvtkFiltersParallelImagingTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    37 Dec 17  2017 libvtkFiltersParallelMPI-6.3.so.6.3 -> libvtkFiltersParallelMPI-6.3.so.6.3.0
-rw-r--r--  1 root root  237K Dec 17  2017 libvtkFiltersParallelMPI-6.3.so.6.3.0
lrwxrwxrwx  1 root root    46 Dec 17  2017 libvtkFiltersParallelMPIPython27D-6.3.so.6.3 -> libvtkFiltersParallelMPIPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   94K Dec 17  2017 libvtkFiltersParallelMPIPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    40 Dec 17  2017 libvtkFiltersParallelMPITCL-6.3.so.6.3 -> libvtkFiltersParallelMPITCL-6.3.so.6.3.0
-rw-r--r--  1 root root   67K Dec 17  2017 libvtkFiltersParallelMPITCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    43 Dec 17  2017 libvtkFiltersParallelPython27D-6.3.so.6.3 -> libvtkFiltersParallelPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  484K Dec 17  2017 libvtkFiltersParallelPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    44 Dec 17  2017 libvtkFiltersParallelStatistics-6.3.so.6.3 -> libvtkFiltersParallelStatistics-6.3.so.6.3.0
-rw-r--r--  1 root root  220K Dec 17  2017 libvtkFiltersParallelStatistics-6.3.so.6.3.0
lrwxrwxrwx  1 root root    53 Dec 17  2017 libvtkFiltersParallelStatisticsPython27D-6.3.so.6.3 -> libvtkFiltersParallelStatisticsPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   74K Dec 17  2017 libvtkFiltersParallelStatisticsPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    47 Dec 17  2017 libvtkFiltersParallelStatisticsTCL-6.3.so.6.3 -> libvtkFiltersParallelStatisticsTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   71K Dec 17  2017 libvtkFiltersParallelStatisticsTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    37 Dec 17  2017 libvtkFiltersParallelTCL-6.3.so.6.3 -> libvtkFiltersParallelTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  380K Dec 17  2017 libvtkFiltersParallelTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    38 Dec 17  2017 libvtkFiltersProgrammable-6.3.so.6.3 -> libvtkFiltersProgrammable-6.3.so.6.3.0
-rw-r--r--  1 root root   63K Dec 17  2017 libvtkFiltersProgrammable-6.3.so.6.3.0
lrwxrwxrwx  1 root root    47 Dec 17  2017 libvtkFiltersProgrammablePython27D-6.3.so.6.3 -> libvtkFiltersProgrammablePython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   61K Dec 17  2017 libvtkFiltersProgrammablePython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    41 Dec 17  2017 libvtkFiltersProgrammableTCL-6.3.so.6.3 -> libvtkFiltersProgrammableTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   47K Dec 17  2017 libvtkFiltersProgrammableTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    32 Dec 17  2017 libvtkFiltersPython-6.3.so.6.3 -> libvtkFiltersPython-6.3.so.6.3.0
-rw-r--r--  1 root root   35K Dec 17  2017 libvtkFiltersPython-6.3.so.6.3.0
lrwxrwxrwx  1 root root    41 Dec 17  2017 libvtkFiltersPythonPython27D-6.3.so.6.3 -> libvtkFiltersPythonPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   15K Dec 17  2017 libvtkFiltersPythonPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    35 Dec 17  2017 libvtkFiltersReebGraph-6.3.so.6.3 -> libvtkFiltersReebGraph-6.3.so.6.3.0
-rw-r--r--  1 root root  108K Dec 17  2017 libvtkFiltersReebGraph-6.3.so.6.3.0
lrwxrwxrwx  1 root root    44 Dec 17  2017 libvtkFiltersReebGraphPython27D-6.3.so.6.3 -> libvtkFiltersReebGraphPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   73K Dec 17  2017 libvtkFiltersReebGraphPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    38 Dec 17  2017 libvtkFiltersReebGraphTCL-6.3.so.6.3 -> libvtkFiltersReebGraphTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   63K Dec 17  2017 libvtkFiltersReebGraphTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    29 Dec 17  2017 libvtkFiltersSMP-6.3.so.6.3 -> libvtkFiltersSMP-6.3.so.6.3.0
-rw-r--r--  1 root root  407K Dec 17  2017 libvtkFiltersSMP-6.3.so.6.3.0
lrwxrwxrwx  1 root root    38 Dec 17  2017 libvtkFiltersSMPPython27D-6.3.so.6.3 -> libvtkFiltersSMPPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  110K Dec 17  2017 libvtkFiltersSMPPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    32 Dec 17  2017 libvtkFiltersSMPTCL-6.3.so.6.3 -> libvtkFiltersSMPTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   87K Dec 17  2017 libvtkFiltersSMPTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    35 Dec 17  2017 libvtkFiltersSelection-6.3.so.6.3 -> libvtkFiltersSelection-6.3.so.6.3.0
-rw-r--r--  1 root root  100K Dec 17  2017 libvtkFiltersSelection-6.3.so.6.3.0
lrwxrwxrwx  1 root root    44 Dec 17  2017 libvtkFiltersSelectionPython27D-6.3.so.6.3 -> libvtkFiltersSelectionPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  109K Dec 17  2017 libvtkFiltersSelectionPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    38 Dec 17  2017 libvtkFiltersSelectionTCL-6.3.so.6.3 -> libvtkFiltersSelectionTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   83K Dec 17  2017 libvtkFiltersSelectionTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    33 Dec 17  2017 libvtkFiltersSources-6.3.so.6.3 -> libvtkFiltersSources-6.3.so.6.3.0
-rw-r--r--  1 root root  773K Dec 17  2017 libvtkFiltersSources-6.3.so.6.3.0
lrwxrwxrwx  1 root root    42 Dec 17  2017 libvtkFiltersSourcesPython27D-6.3.so.6.3 -> libvtkFiltersSourcesPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  1.2M Dec 17  2017 libvtkFiltersSourcesPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    36 Dec 17  2017 libvtkFiltersSourcesTCL-6.3.so.6.3 -> libvtkFiltersSourcesTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  760K Dec 17  2017 libvtkFiltersSourcesTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    36 Dec 17  2017 libvtkFiltersStatistics-6.3.so.6.3 -> libvtkFiltersStatistics-6.3.so.6.3.0
-rw-r--r--  1 root root  549K Dec 17  2017 libvtkFiltersStatistics-6.3.so.6.3.0
lrwxrwxrwx  1 root root    45 Dec 17  2017 libvtkFiltersStatisticsPython27D-6.3.so.6.3 -> libvtkFiltersStatisticsPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  341K Dec 17  2017 libvtkFiltersStatisticsPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    39 Dec 17  2017 libvtkFiltersStatisticsTCL-6.3.so.6.3 -> libvtkFiltersStatisticsTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  224K Dec 17  2017 libvtkFiltersStatisticsTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    33 Dec 17  2017 libvtkFiltersTexture-6.3.so.6.3 -> libvtkFiltersTexture-6.3.so.6.3.0
-rw-r--r--  1 root root  151K Dec 17  2017 libvtkFiltersTexture-6.3.so.6.3.0
lrwxrwxrwx  1 root root    42 Dec 17  2017 libvtkFiltersTexturePython27D-6.3.so.6.3 -> libvtkFiltersTexturePython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  219K Dec 17  2017 libvtkFiltersTexturePython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    36 Dec 17  2017 libvtkFiltersTextureTCL-6.3.so.6.3 -> libvtkFiltersTextureTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  167K Dec 17  2017 libvtkFiltersTextureTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    33 Dec 17  2017 libvtkFiltersVerdict-6.3.so.6.3 -> libvtkFiltersVerdict-6.3.so.6.3.0
-rw-r--r--  1 root root  124K Dec 17  2017 libvtkFiltersVerdict-6.3.so.6.3.0
lrwxrwxrwx  1 root root    42 Dec 17  2017 libvtkFiltersVerdictPython27D-6.3.so.6.3 -> libvtkFiltersVerdictPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  267K Dec 17  2017 libvtkFiltersVerdictPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    36 Dec 17  2017 libvtkFiltersVerdictTCL-6.3.so.6.3 -> libvtkFiltersVerdictTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  156K Dec 17  2017 libvtkFiltersVerdictTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    29 Dec 17  2017 libvtkGeovisCore-6.3.so.6.3 -> libvtkGeovisCore-6.3.so.6.3.0
-rw-r--r--  1 root root  707K Dec 17  2017 libvtkGeovisCore-6.3.so.6.3.0
lrwxrwxrwx  1 root root    38 Dec 17  2017 libvtkGeovisCorePython27D-6.3.so.6.3 -> libvtkGeovisCorePython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  555K Dec 17  2017 libvtkGeovisCorePython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    32 Dec 17  2017 libvtkGeovisCoreTCL-6.3.so.6.3 -> libvtkGeovisCoreTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  392K Dec 17  2017 libvtkGeovisCoreTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    24 Dec 17  2017 libvtkIOAMR-6.3.so.6.3 -> libvtkIOAMR-6.3.so.6.3.0
-rw-r--r--  1 root root  282K Dec 17  2017 libvtkIOAMR-6.3.so.6.3.0
lrwxrwxrwx  1 root root    33 Dec 17  2017 libvtkIOAMRPython27D-6.3.so.6.3 -> libvtkIOAMRPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  107K Dec 17  2017 libvtkIOAMRPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    27 Dec 17  2017 libvtkIOAMRTCL-6.3.so.6.3 -> libvtkIOAMRTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   91K Dec 17  2017 libvtkIOAMRTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    25 Dec 17  2017 libvtkIOCore-6.3.so.6.3 -> libvtkIOCore-6.3.so.6.3.0
-rw-r--r--  1 root root  458K Dec 17  2017 libvtkIOCore-6.3.so.6.3.0
lrwxrwxrwx  1 root root    34 Dec 17  2017 libvtkIOCorePython27D-6.3.so.6.3 -> libvtkIOCorePython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  322K Dec 17  2017 libvtkIOCorePython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    28 Dec 17  2017 libvtkIOCoreTCL-6.3.so.6.3 -> libvtkIOCoreTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  240K Dec 17  2017 libvtkIOCoreTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    28 Dec 17  2017 libvtkIOEnSight-6.3.so.6.3 -> libvtkIOEnSight-6.3.so.6.3.0
-rw-r--r--  1 root root  549K Dec 17  2017 libvtkIOEnSight-6.3.so.6.3.0
lrwxrwxrwx  1 root root    37 Dec 17  2017 libvtkIOEnSightPython27D-6.3.so.6.3 -> libvtkIOEnSightPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  141K Dec 17  2017 libvtkIOEnSightPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    31 Dec 17  2017 libvtkIOEnSightTCL-6.3.so.6.3 -> libvtkIOEnSightTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   95K Dec 17  2017 libvtkIOEnSightTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    27 Dec 17  2017 libvtkIOExodus-6.3.so.6.3 -> libvtkIOExodus-6.3.so.6.3.0
-rw-r--r--  1 root root  660K Dec 17  2017 libvtkIOExodus-6.3.so.6.3.0
lrwxrwxrwx  1 root root    36 Dec 17  2017 libvtkIOExodusPython27D-6.3.so.6.3 -> libvtkIOExodusPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  385K Dec 17  2017 libvtkIOExodusPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    30 Dec 17  2017 libvtkIOExodusTCL-6.3.so.6.3 -> libvtkIOExodusTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  220K Dec 17  2017 libvtkIOExodusTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    27 Dec 17  2017 libvtkIOExport-6.3.so.6.3 -> libvtkIOExport-6.3.so.6.3.0
-rw-r--r--  1 root root  513K Dec 17  2017 libvtkIOExport-6.3.so.6.3.0
lrwxrwxrwx  1 root root    36 Dec 17  2017 libvtkIOExportPython27D-6.3.so.6.3 -> libvtkIOExportPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  288K Dec 17  2017 libvtkIOExportPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    30 Dec 17  2017 libvtkIOExportTCL-6.3.so.6.3 -> libvtkIOExportTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  183K Dec 17  2017 libvtkIOExportTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    27 Dec 17  2017 libvtkIOFFMPEG-6.3.so.6.3 -> libvtkIOFFMPEG-6.3.so.6.3.0
-rw-r--r--  1 root root   55K Dec 17  2017 libvtkIOFFMPEG-6.3.so.6.3.0
lrwxrwxrwx  1 root root    36 Dec 17  2017 libvtkIOFFMPEGPython27D-6.3.so.6.3 -> libvtkIOFFMPEGPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   44K Dec 17  2017 libvtkIOFFMPEGPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    30 Dec 17  2017 libvtkIOFFMPEGTCL-6.3.so.6.3 -> libvtkIOFFMPEGTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   31K Dec 17  2017 libvtkIOFFMPEGTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    25 Dec 17  2017 libvtkIOGDAL-6.3.so.6.3 -> libvtkIOGDAL-6.3.so.6.3.0
-rw-r--r--  1 root root  152K Dec 17  2017 libvtkIOGDAL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    34 Dec 17  2017 libvtkIOGDALPython27D-6.3.so.6.3 -> libvtkIOGDALPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   60K Dec 17  2017 libvtkIOGDALPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    28 Dec 17  2017 libvtkIOGDALTCL-6.3.so.6.3 -> libvtkIOGDALTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   47K Dec 17  2017 libvtkIOGDALTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    28 Dec 17  2017 libvtkIOGeoJSON-6.3.so.6.3 -> libvtkIOGeoJSON-6.3.so.6.3.0
-rw-r--r--  1 root root   63K Dec 17  2017 libvtkIOGeoJSON-6.3.so.6.3.0
lrwxrwxrwx  1 root root    37 Dec 17  2017 libvtkIOGeoJSONPython27D-6.3.so.6.3 -> libvtkIOGeoJSONPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   40K Dec 17  2017 libvtkIOGeoJSONPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    31 Dec 17  2017 libvtkIOGeoJSONTCL-6.3.so.6.3 -> libvtkIOGeoJSONTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   35K Dec 17  2017 libvtkIOGeoJSONTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    29 Dec 17  2017 libvtkIOGeometry-6.3.so.6.3 -> libvtkIOGeometry-6.3.so.6.3.0
-rw-r--r--  1 root root  1.1M Dec 17  2017 libvtkIOGeometry-6.3.so.6.3.0
lrwxrwxrwx  1 root root    38 Dec 17  2017 libvtkIOGeometryPython27D-6.3.so.6.3 -> libvtkIOGeometryPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  621K Dec 17  2017 libvtkIOGeometryPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    32 Dec 17  2017 libvtkIOGeometryTCL-6.3.so.6.3 -> libvtkIOGeometryTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  400K Dec 17  2017 libvtkIOGeometryTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    26 Dec 17  2017 libvtkIOImage-6.3.so.6.3 -> libvtkIOImage-6.3.so.6.3.0
-rw-r--r--  1 root root  1.6M Dec 17  2017 libvtkIOImage-6.3.so.6.3.0
lrwxrwxrwx  1 root root    35 Dec 17  2017 libvtkIOImagePython27D-6.3.so.6.3 -> libvtkIOImagePython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  900K Dec 17  2017 libvtkIOImagePython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    29 Dec 17  2017 libvtkIOImageTCL-6.3.so.6.3 -> libvtkIOImageTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  540K Dec 17  2017 libvtkIOImageTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    27 Dec 17  2017 libvtkIOImport-6.3.so.6.3 -> libvtkIOImport-6.3.so.6.3.0
-rw-r--r--  1 root root  276K Dec 17  2017 libvtkIOImport-6.3.so.6.3.0
lrwxrwxrwx  1 root root    36 Dec 17  2017 libvtkIOImportPython27D-6.3.so.6.3 -> libvtkIOImportPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   57K Dec 17  2017 libvtkIOImportPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    30 Dec 17  2017 libvtkIOImportTCL-6.3.so.6.3 -> libvtkIOImportTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   47K Dec 17  2017 libvtkIOImportTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    28 Dec 17  2017 libvtkIOInfovis-6.3.so.6.3 -> libvtkIOInfovis-6.3.so.6.3.0
-rw-r--r--  1 root root  562K Dec 17  2017 libvtkIOInfovis-6.3.so.6.3.0
lrwxrwxrwx  1 root root    37 Dec 17  2017 libvtkIOInfovisPython27D-6.3.so.6.3 -> libvtkIOInfovisPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  320K Dec 17  2017 libvtkIOInfovisPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    31 Dec 17  2017 libvtkIOInfovisTCL-6.3.so.6.3 -> libvtkIOInfovisTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  212K Dec 17  2017 libvtkIOInfovisTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    27 Dec 17  2017 libvtkIOLSDyna-6.3.so.6.3 -> libvtkIOLSDyna-6.3.so.6.3.0
-rw-r--r--  1 root root  313K Dec 17  2017 libvtkIOLSDyna-6.3.so.6.3.0
lrwxrwxrwx  1 root root    36 Dec 17  2017 libvtkIOLSDynaPython27D-6.3.so.6.3 -> libvtkIOLSDynaPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  130K Dec 17  2017 libvtkIOLSDynaPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    30 Dec 17  2017 libvtkIOLSDynaTCL-6.3.so.6.3 -> libvtkIOLSDynaTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   71K Dec 17  2017 libvtkIOLSDynaTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    27 Dec 17  2017 libvtkIOLegacy-6.3.so.6.3 -> libvtkIOLegacy-6.3.so.6.3.0
-rw-r--r--  1 root root  655K Dec 17  2017 libvtkIOLegacy-6.3.so.6.3.0
lrwxrwxrwx  1 root root    36 Dec 17  2017 libvtkIOLegacyPython27D-6.3.so.6.3 -> libvtkIOLegacyPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  354K Dec 17  2017 libvtkIOLegacyPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    30 Dec 17  2017 libvtkIOLegacyTCL-6.3.so.6.3 -> libvtkIOLegacyTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  240K Dec 17  2017 libvtkIOLegacyTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    25 Dec 17  2017 libvtkIOMINC-6.3.so.6.3 -> libvtkIOMINC-6.3.so.6.3.0
-rw-r--r--  1 root root  450K Dec 17  2017 libvtkIOMINC-6.3.so.6.3.0
lrwxrwxrwx  1 root root    34 Dec 17  2017 libvtkIOMINCPython27D-6.3.so.6.3 -> libvtkIOMINCPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  214K Dec 17  2017 libvtkIOMINCPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    28 Dec 17  2017 libvtkIOMINCTCL-6.3.so.6.3 -> libvtkIOMINCTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  151K Dec 17  2017 libvtkIOMINCTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    29 Dec 17  2017 libvtkIOMPIImage-6.3.so.6.3 -> libvtkIOMPIImage-6.3.so.6.3.0
-rw-r--r--  1 root root  161K Dec 17  2017 libvtkIOMPIImage-6.3.so.6.3.0
lrwxrwxrwx  1 root root    38 Dec 17  2017 libvtkIOMPIImagePython27D-6.3.so.6.3 -> libvtkIOMPIImagePython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   28K Dec 17  2017 libvtkIOMPIImagePython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    32 Dec 17  2017 libvtkIOMPIImageTCL-6.3.so.6.3 -> libvtkIOMPIImageTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   23K Dec 17  2017 libvtkIOMPIImageTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    32 Dec 17  2017 libvtkIOMPIParallel-6.3.so.6.3 -> libvtkIOMPIParallel-6.3.so.6.3.0
-rw-r--r--  1 root root  217K Dec 17  2017 libvtkIOMPIParallel-6.3.so.6.3.0
lrwxrwxrwx  1 root root    41 Dec 17  2017 libvtkIOMPIParallelPython27D-6.3.so.6.3 -> libvtkIOMPIParallelPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   27K Dec 17  2017 libvtkIOMPIParallelPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    35 Dec 17  2017 libvtkIOMPIParallelTCL-6.3.so.6.3 -> libvtkIOMPIParallelTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   27K Dec 17  2017 libvtkIOMPIParallelTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    26 Dec 17  2017 libvtkIOMovie-6.3.so.6.3 -> libvtkIOMovie-6.3.so.6.3.0
-rw-r--r--  1 root root   59K Dec 17  2017 libvtkIOMovie-6.3.so.6.3.0
lrwxrwxrwx  1 root root    35 Dec 17  2017 libvtkIOMoviePython27D-6.3.so.6.3 -> libvtkIOMoviePython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   49K Dec 17  2017 libvtkIOMoviePython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    29 Dec 17  2017 libvtkIOMovieTCL-6.3.so.6.3 -> libvtkIOMovieTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   39K Dec 17  2017 libvtkIOMovieTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    26 Dec 17  2017 libvtkIOMySQL-6.3.so.6.3 -> libvtkIOMySQL-6.3.so.6.3.0
-rw-r--r--  1 root root  144K Dec 17  2017 libvtkIOMySQL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    35 Dec 17  2017 libvtkIOMySQLPython27D-6.3.so.6.3 -> libvtkIOMySQLPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   86K Dec 17  2017 libvtkIOMySQLPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    29 Dec 17  2017 libvtkIOMySQLTCL-6.3.so.6.3 -> libvtkIOMySQLTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   63K Dec 17  2017 libvtkIOMySQLTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    27 Dec 17  2017 libvtkIONetCDF-6.3.so.6.3 -> libvtkIONetCDF-6.3.so.6.3.0
-rw-r--r--  1 root root  410K Dec 17  2017 libvtkIONetCDF-6.3.so.6.3.0
lrwxrwxrwx  1 root root    36 Dec 17  2017 libvtkIONetCDFPython27D-6.3.so.6.3 -> libvtkIONetCDFPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  201K Dec 17  2017 libvtkIONetCDFPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    30 Dec 17  2017 libvtkIONetCDFTCL-6.3.so.6.3 -> libvtkIONetCDFTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  139K Dec 17  2017 libvtkIONetCDFTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    25 Dec 17  2017 libvtkIOODBC-6.3.so.6.3 -> libvtkIOODBC-6.3.so.6.3.0
-rw-r--r--  1 root root  156K Dec 17  2017 libvtkIOODBC-6.3.so.6.3.0
lrwxrwxrwx  1 root root    34 Dec 17  2017 libvtkIOODBCPython27D-6.3.so.6.3 -> libvtkIOODBCPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   65K Dec 17  2017 libvtkIOODBCPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    28 Dec 17  2017 libvtkIOODBCTCL-6.3.so.6.3 -> libvtkIOODBCTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   47K Dec 17  2017 libvtkIOODBCTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    24 Dec 17  2017 libvtkIOPLY-6.3.so.6.3 -> libvtkIOPLY-6.3.so.6.3.0
-rw-r--r--  1 root root   97K Dec 17  2017 libvtkIOPLY-6.3.so.6.3.0
lrwxrwxrwx  1 root root    33 Dec 17  2017 libvtkIOPLYPython27D-6.3.so.6.3 -> libvtkIOPLYPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   73K Dec 17  2017 libvtkIOPLYPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    27 Dec 17  2017 libvtkIOPLYTCL-6.3.so.6.3 -> libvtkIOPLYTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   43K Dec 17  2017 libvtkIOPLYTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    29 Dec 17  2017 libvtkIOParallel-6.3.so.6.3 -> libvtkIOParallel-6.3.so.6.3.0
-rw-r--r--  1 root root  526K Dec 17  2017 libvtkIOParallel-6.3.so.6.3.0
lrwxrwxrwx  1 root root    35 Dec 17  2017 libvtkIOParallelExodus-6.3.so.6.3 -> libvtkIOParallelExodus-6.3.so.6.3.0
-rw-r--r--  1 root root  148K Dec 17  2017 libvtkIOParallelExodus-6.3.so.6.3.0
lrwxrwxrwx  1 root root    44 Dec 17  2017 libvtkIOParallelExodusPython27D-6.3.so.6.3 -> libvtkIOParallelExodusPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   48K Dec 17  2017 libvtkIOParallelExodusPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    38 Dec 17  2017 libvtkIOParallelExodusTCL-6.3.so.6.3 -> libvtkIOParallelExodusTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   35K Dec 17  2017 libvtkIOParallelExodusTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    35 Dec 17  2017 libvtkIOParallelLSDyna-6.3.so.6.3 -> libvtkIOParallelLSDyna-6.3.so.6.3.0
-rw-r--r--  1 root root   55K Dec 17  2017 libvtkIOParallelLSDyna-6.3.so.6.3.0
lrwxrwxrwx  1 root root    44 Dec 17  2017 libvtkIOParallelLSDynaPython27D-6.3.so.6.3 -> libvtkIOParallelLSDynaPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   19K Dec 17  2017 libvtkIOParallelLSDynaPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    38 Dec 17  2017 libvtkIOParallelLSDynaTCL-6.3.so.6.3 -> libvtkIOParallelLSDynaTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   19K Dec 17  2017 libvtkIOParallelLSDynaTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    35 Dec 17  2017 libvtkIOParallelNetCDF-6.3.so.6.3 -> libvtkIOParallelNetCDF-6.3.so.6.3.0
-rw-r--r--  1 root root  113K Dec 17  2017 libvtkIOParallelNetCDF-6.3.so.6.3.0
lrwxrwxrwx  1 root root    44 Dec 17  2017 libvtkIOParallelNetCDFPython27D-6.3.so.6.3 -> libvtkIOParallelNetCDFPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   32K Dec 17  2017 libvtkIOParallelNetCDFPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    38 Dec 17  2017 libvtkIOParallelNetCDFTCL-6.3.so.6.3 -> libvtkIOParallelNetCDFTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   27K Dec 17  2017 libvtkIOParallelNetCDFTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    38 Dec 17  2017 libvtkIOParallelPython27D-6.3.so.6.3 -> libvtkIOParallelPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  221K Dec 17  2017 libvtkIOParallelPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    32 Dec 17  2017 libvtkIOParallelTCL-6.3.so.6.3 -> libvtkIOParallelTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  139K Dec 17  2017 libvtkIOParallelTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    32 Dec 17  2017 libvtkIOParallelXML-6.3.so.6.3 -> libvtkIOParallelXML-6.3.so.6.3.0
-rw-r--r--  1 root root  204K Dec 17  2017 libvtkIOParallelXML-6.3.so.6.3.0
lrwxrwxrwx  1 root root    41 Dec 17  2017 libvtkIOParallelXMLPython27D-6.3.so.6.3 -> libvtkIOParallelXMLPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   95K Dec 17  2017 libvtkIOParallelXMLPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    35 Dec 17  2017 libvtkIOParallelXMLTCL-6.3.so.6.3 -> libvtkIOParallelXMLTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   87K Dec 17  2017 libvtkIOParallelXMLTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    31 Dec 17  2017 libvtkIOPostgreSQL-6.3.so.6.3 -> libvtkIOPostgreSQL-6.3.so.6.3.0
-rw-r--r--  1 root root  160K Dec 17  2017 libvtkIOPostgreSQL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    40 Dec 17  2017 libvtkIOPostgreSQLPython27D-6.3.so.6.3 -> libvtkIOPostgreSQLPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   73K Dec 17  2017 libvtkIOPostgreSQLPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    34 Dec 17  2017 libvtkIOPostgreSQLTCL-6.3.so.6.3 -> libvtkIOPostgreSQLTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   59K Dec 17  2017 libvtkIOPostgreSQLTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    24 Dec 17  2017 libvtkIOSQL-6.3.so.6.3 -> libvtkIOSQL-6.3.so.6.3.0
-rw-r--r--  1 root root  277K Dec 17  2017 libvtkIOSQL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    33 Dec 17  2017 libvtkIOSQLPython27D-6.3.so.6.3 -> libvtkIOSQLPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  202K Dec 17  2017 libvtkIOSQLPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    27 Dec 17  2017 libvtkIOSQLTCL-6.3.so.6.3 -> libvtkIOSQLTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  148K Dec 17  2017 libvtkIOSQLTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    25 Dec 17  2017 libvtkIOVPIC-6.3.so.6.3 -> libvtkIOVPIC-6.3.so.6.3.0
-rw-r--r--  1 root root   63K Dec 17  2017 libvtkIOVPIC-6.3.so.6.3.0
lrwxrwxrwx  1 root root    34 Dec 17  2017 libvtkIOVPICPython27D-6.3.so.6.3 -> libvtkIOVPICPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   44K Dec 17  2017 libvtkIOVPICPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    28 Dec 17  2017 libvtkIOVPICTCL-6.3.so.6.3 -> libvtkIOVPICTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   39K Dec 17  2017 libvtkIOVPICTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    26 Dec 17  2017 libvtkIOVideo-6.3.so.6.3 -> libvtkIOVideo-6.3.so.6.3.0
-rw-r--r--  1 root root   71K Dec 17  2017 libvtkIOVideo-6.3.so.6.3.0
lrwxrwxrwx  1 root root    35 Dec 17  2017 libvtkIOVideoPython27D-6.3.so.6.3 -> libvtkIOVideoPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   85K Dec 17  2017 libvtkIOVideoPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    29 Dec 17  2017 libvtkIOVideoTCL-6.3.so.6.3 -> libvtkIOVideoTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   43K Dec 17  2017 libvtkIOVideoTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    24 Dec 17  2017 libvtkIOXML-6.3.so.6.3 -> libvtkIOXML-6.3.so.6.3.0
-rw-r--r--  1 root root  945K Dec 17  2017 libvtkIOXML-6.3.so.6.3.0
lrwxrwxrwx  1 root root    30 Dec 17  2017 libvtkIOXMLParser-6.3.so.6.3 -> libvtkIOXMLParser-6.3.so.6.3.0
-rw-r--r--  1 root root  112K Dec 17  2017 libvtkIOXMLParser-6.3.so.6.3.0
lrwxrwxrwx  1 root root    39 Dec 17  2017 libvtkIOXMLParserPython27D-6.3.so.6.3 -> libvtkIOXMLParserPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   69K Dec 17  2017 libvtkIOXMLParserPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    33 Dec 17  2017 libvtkIOXMLParserTCL-6.3.so.6.3 -> libvtkIOXMLParserTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   51K Dec 17  2017 libvtkIOXMLParserTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    33 Dec 17  2017 libvtkIOXMLPython27D-6.3.so.6.3 -> libvtkIOXMLPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  375K Dec 17  2017 libvtkIOXMLPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    27 Dec 17  2017 libvtkIOXMLTCL-6.3.so.6.3 -> libvtkIOXMLTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  304K Dec 17  2017 libvtkIOXMLTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    26 Dec 17  2017 libvtkIOXdmf2-6.3.so.6.3 -> libvtkIOXdmf2-6.3.so.6.3.0
-rw-r--r--  1 root root  335K Dec 17  2017 libvtkIOXdmf2-6.3.so.6.3.0
lrwxrwxrwx  1 root root    35 Dec 17  2017 libvtkIOXdmf2Python27D-6.3.so.6.3 -> libvtkIOXdmf2Python27D-6.3.so.6.3.0
-rw-r--r--  1 root root   93K Dec 17  2017 libvtkIOXdmf2Python27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    30 Dec 17  2017 libvtkIOXdmfIITCL-6.3.so.6.3 -> libvtkIOXdmfIITCL-6.3.so.6.3.0
-rw-r--r--  1 root root   67K Dec 17  2017 libvtkIOXdmfIITCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    31 Dec 17  2017 libvtkImagingColor-6.3.so.6.3 -> libvtkImagingColor-6.3.so.6.3.0
-rw-r--r--  1 root root  280K Dec 17  2017 libvtkImagingColor-6.3.so.6.3.0
lrwxrwxrwx  1 root root    40 Dec 17  2017 libvtkImagingColorPython27D-6.3.so.6.3 -> libvtkImagingColorPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   85K Dec 17  2017 libvtkImagingColorPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    34 Dec 17  2017 libvtkImagingColorTCL-6.3.so.6.3 -> libvtkImagingColorTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   75K Dec 17  2017 libvtkImagingColorTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    30 Dec 17  2017 libvtkImagingCore-6.3.so.6.3 -> libvtkImagingCore-6.3.so.6.3.0
-rw-r--r--  1 root root  1.9M Dec 17  2017 libvtkImagingCore-6.3.so.6.3.0
lrwxrwxrwx  1 root root    39 Dec 17  2017 libvtkImagingCorePython27D-6.3.so.6.3 -> libvtkImagingCorePython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  934K Dec 17  2017 libvtkImagingCorePython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    33 Dec 17  2017 libvtkImagingCoreTCL-6.3.so.6.3 -> libvtkImagingCoreTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  625K Dec 17  2017 libvtkImagingCoreTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    33 Dec 17  2017 libvtkImagingFourier-6.3.so.6.3 -> libvtkImagingFourier-6.3.so.6.3.0
-rw-r--r--  1 root root  187K Dec 17  2017 libvtkImagingFourier-6.3.so.6.3.0
lrwxrwxrwx  1 root root    42 Dec 17  2017 libvtkImagingFourierPython27D-6.3.so.6.3 -> libvtkImagingFourierPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  103K Dec 17  2017 libvtkImagingFourierPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    36 Dec 17  2017 libvtkImagingFourierTCL-6.3.so.6.3 -> libvtkImagingFourierTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   87K Dec 17  2017 libvtkImagingFourierTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    33 Dec 17  2017 libvtkImagingGeneral-6.3.so.6.3 -> libvtkImagingGeneral-6.3.so.6.3.0
-rw-r--r--  1 root root  1.1M Dec 17  2017 libvtkImagingGeneral-6.3.so.6.3.0
lrwxrwxrwx  1 root root    42 Dec 17  2017 libvtkImagingGeneralPython27D-6.3.so.6.3 -> libvtkImagingGeneralPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  378K Dec 17  2017 libvtkImagingGeneralPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    36 Dec 17  2017 libvtkImagingGeneralTCL-6.3.so.6.3 -> libvtkImagingGeneralTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  348K Dec 17  2017 libvtkImagingGeneralTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    32 Dec 17  2017 libvtkImagingHybrid-6.3.so.6.3 -> libvtkImagingHybrid-6.3.so.6.3.0
-rw-r--r--  1 root root  505K Dec 17  2017 libvtkImagingHybrid-6.3.so.6.3.0
lrwxrwxrwx  1 root root    41 Dec 17  2017 libvtkImagingHybridPython27D-6.3.so.6.3 -> libvtkImagingHybridPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  539K Dec 17  2017 libvtkImagingHybridPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    35 Dec 17  2017 libvtkImagingHybridTCL-6.3.so.6.3 -> libvtkImagingHybridTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  315K Dec 17  2017 libvtkImagingHybridTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    30 Dec 17  2017 libvtkImagingMath-6.3.so.6.3 -> libvtkImagingMath-6.3.so.6.3.0
-rw-r--r--  1 root root  296K Dec 17  2017 libvtkImagingMath-6.3.so.6.3.0
lrwxrwxrwx  1 root root    39 Dec 17  2017 libvtkImagingMathPython27D-6.3.so.6.3 -> libvtkImagingMathPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  143K Dec 17  2017 libvtkImagingMathPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    33 Dec 17  2017 libvtkImagingMathTCL-6.3.so.6.3 -> libvtkImagingMathTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   99K Dec 17  2017 libvtkImagingMathTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    39 Dec 17  2017 libvtkImagingMorphological-6.3.so.6.3 -> libvtkImagingMorphological-6.3.so.6.3.0
-rw-r--r--  1 root root  536K Dec 17  2017 libvtkImagingMorphological-6.3.so.6.3.0
lrwxrwxrwx  1 root root    48 Dec 17  2017 libvtkImagingMorphologicalPython27D-6.3.so.6.3 -> libvtkImagingMorphologicalPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  205K Dec 17  2017 libvtkImagingMorphologicalPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    42 Dec 17  2017 libvtkImagingMorphologicalTCL-6.3.so.6.3 -> libvtkImagingMorphologicalTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  143K Dec 17  2017 libvtkImagingMorphologicalTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    33 Dec 17  2017 libvtkImagingSources-6.3.so.6.3 -> libvtkImagingSources-6.3.so.6.3.0
-rw-r--r--  1 root root  331K Dec 17  2017 libvtkImagingSources-6.3.so.6.3.0
lrwxrwxrwx  1 root root    42 Dec 17  2017 libvtkImagingSourcesPython27D-6.3.so.6.3 -> libvtkImagingSourcesPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  253K Dec 17  2017 libvtkImagingSourcesPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    36 Dec 17  2017 libvtkImagingSourcesTCL-6.3.so.6.3 -> libvtkImagingSourcesTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  183K Dec 17  2017 libvtkImagingSourcesTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    36 Dec 17  2017 libvtkImagingStatistics-6.3.so.6.3 -> libvtkImagingStatistics-6.3.so.6.3.0
-rw-r--r--  1 root root  156K Dec 17  2017 libvtkImagingStatistics-6.3.so.6.3.0
lrwxrwxrwx  1 root root    45 Dec 17  2017 libvtkImagingStatisticsPython27D-6.3.so.6.3 -> libvtkImagingStatisticsPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  138K Dec 17  2017 libvtkImagingStatisticsPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    39 Dec 17  2017 libvtkImagingStatisticsTCL-6.3.so.6.3 -> libvtkImagingStatisticsTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   99K Dec 17  2017 libvtkImagingStatisticsTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    33 Dec 17  2017 libvtkImagingStencil-6.3.so.6.3 -> libvtkImagingStencil-6.3.so.6.3.0
-rw-r--r--  1 root root  212K Dec 17  2017 libvtkImagingStencil-6.3.so.6.3.0
lrwxrwxrwx  1 root root    42 Dec 17  2017 libvtkImagingStencilPython27D-6.3.so.6.3 -> libvtkImagingStencilPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  147K Dec 17  2017 libvtkImagingStencilPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    36 Dec 17  2017 libvtkImagingStencilTCL-6.3.so.6.3 -> libvtkImagingStencilTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  107K Dec 17  2017 libvtkImagingStencilTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    46 Dec 17  2017 libvtkInfovisBoostGraphAlgorithms-6.3.so.6.3 -> libvtkInfovisBoostGraphAlgorithms-6.3.so.6.3.0
-rw-r--r--  1 root root  477K Dec 17  2017 libvtkInfovisBoostGraphAlgorithms-6.3.so.6.3.0
lrwxrwxrwx  1 root root    55 Dec 17  2017 libvtkInfovisBoostGraphAlgorithmsPython27D-6.3.so.6.3 -> libvtkInfovisBoostGraphAlgorithmsPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  205K Dec 17  2017 libvtkInfovisBoostGraphAlgorithmsPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    49 Dec 17  2017 libvtkInfovisBoostGraphAlgorithmsTCL-6.3.so.6.3 -> libvtkInfovisBoostGraphAlgorithmsTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  163K Dec 17  2017 libvtkInfovisBoostGraphAlgorithmsTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    30 Dec 17  2017 libvtkInfovisCore-6.3.so.6.3 -> libvtkInfovisCore-6.3.so.6.3.0
-rw-r--r--  1 root root 1008K Dec 17  2017 libvtkInfovisCore-6.3.so.6.3.0
lrwxrwxrwx  1 root root    39 Dec 17  2017 libvtkInfovisCorePython27D-6.3.so.6.3 -> libvtkInfovisCorePython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  752K Dec 17  2017 libvtkInfovisCorePython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    33 Dec 17  2017 libvtkInfovisCoreTCL-6.3.so.6.3 -> libvtkInfovisCoreTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  536K Dec 17  2017 libvtkInfovisCoreTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    32 Dec 17  2017 libvtkInfovisLayout-6.3.so.6.3 -> libvtkInfovisLayout-6.3.so.6.3.0
-rw-r--r--  1 root root  670K Dec 17  2017 libvtkInfovisLayout-6.3.so.6.3.0
lrwxrwxrwx  1 root root    41 Dec 17  2017 libvtkInfovisLayoutPython27D-6.3.so.6.3 -> libvtkInfovisLayoutPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  897K Dec 17  2017 libvtkInfovisLayoutPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    35 Dec 17  2017 libvtkInfovisLayoutTCL-6.3.so.6.3 -> libvtkInfovisLayoutTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  600K Dec 17  2017 libvtkInfovisLayoutTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    35 Dec 17  2017 libvtkInteractionImage-6.3.so.6.3 -> libvtkInteractionImage-6.3.so.6.3.0
-rw-r--r--  1 root root  152K Dec 17  2017 libvtkInteractionImage-6.3.so.6.3.0
lrwxrwxrwx  1 root root    44 Dec 17  2017 libvtkInteractionImagePython27D-6.3.so.6.3 -> libvtkInteractionImagePython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  144K Dec 17  2017 libvtkInteractionImagePython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    38 Dec 17  2017 libvtkInteractionImageTCL-6.3.so.6.3 -> libvtkInteractionImageTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   91K Dec 17  2017 libvtkInteractionImageTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    35 Dec 17  2017 libvtkInteractionStyle-6.3.so.6.3 -> libvtkInteractionStyle-6.3.so.6.3.0
-rw-r--r--  1 root root  537K Dec 17  2017 libvtkInteractionStyle-6.3.so.6.3.0
lrwxrwxrwx  1 root root    44 Dec 17  2017 libvtkInteractionStylePython27D-6.3.so.6.3 -> libvtkInteractionStylePython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  349K Dec 17  2017 libvtkInteractionStylePython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    38 Dec 17  2017 libvtkInteractionStyleTCL-6.3.so.6.3 -> libvtkInteractionStyleTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  219K Dec 17  2017 libvtkInteractionStyleTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    37 Dec 17  2017 libvtkInteractionWidgets-6.3.so.6.3 -> libvtkInteractionWidgets-6.3.so.6.3.0
-rw-r--r--  1 root root  3.2M Dec 17  2017 libvtkInteractionWidgets-6.3.so.6.3.0
lrwxrwxrwx  1 root root    46 Dec 17  2017 libvtkInteractionWidgetsPython27D-6.3.so.6.3 -> libvtkInteractionWidgetsPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  3.6M Dec 17  2017 libvtkInteractionWidgetsPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    40 Dec 17  2017 libvtkInteractionWidgetsTCL-6.3.so.6.3 -> libvtkInteractionWidgetsTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  2.1M Dec 17  2017 libvtkInteractionWidgetsTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    31 Dec 17  2017 libvtkLocalExample-6.3.so.6.3 -> libvtkLocalExample-6.3.so.6.3.0
-rw-r--r--  1 root root   15K Dec 17  2017 libvtkLocalExample-6.3.so.6.3.0
lrwxrwxrwx  1 root root    34 Dec 17  2017 libvtkLocalExampleTCL-6.3.so.6.3 -> libvtkLocalExampleTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   15K Dec 17  2017 libvtkLocalExampleTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    31 Dec 17  2017 libvtkParallelCore-6.3.so.6.3 -> libvtkParallelCore-6.3.so.6.3.0
-rw-r--r--  1 root root  305K Dec 17  2017 libvtkParallelCore-6.3.so.6.3.0
lrwxrwxrwx  1 root root    40 Dec 17  2017 libvtkParallelCorePython27D-6.3.so.6.3 -> libvtkParallelCorePython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  376K Dec 17  2017 libvtkParallelCorePython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    34 Dec 17  2017 libvtkParallelCoreTCL-6.3.so.6.3 -> libvtkParallelCoreTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  152K Dec 17  2017 libvtkParallelCoreTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    30 Dec 17  2017 libvtkParallelMPI-6.3.so.6.3 -> libvtkParallelMPI-6.3.so.6.3.0
-rw-r--r--  1 root root  149K Dec 17  2017 libvtkParallelMPI-6.3.so.6.3.0
lrwxrwxrwx  1 root root    33 Dec 17  2017 libvtkParallelMPI4Py-6.3.so.6.3 -> libvtkParallelMPI4Py-6.3.so.6.3.0
-rw-r--r--  1 root root   68K Dec 17  2017 libvtkParallelMPI4Py-6.3.so.6.3.0
lrwxrwxrwx  1 root root    42 Dec 17  2017 libvtkParallelMPI4PyPython27D-6.3.so.6.3 -> libvtkParallelMPI4PyPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   15K Dec 17  2017 libvtkParallelMPI4PyPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    39 Dec 17  2017 libvtkParallelMPIPython27D-6.3.so.6.3 -> libvtkParallelMPIPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   73K Dec 17  2017 libvtkParallelMPIPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    33 Dec 17  2017 libvtkParallelMPITCL-6.3.so.6.3 -> libvtkParallelMPITCL-6.3.so.6.3.0
-rw-r--r--  1 root root   35K Dec 17  2017 libvtkParallelMPITCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    36 Dec 17  2017 libvtkPythonInterpreter-6.3.so.6.3 -> libvtkPythonInterpreter-6.3.so.6.3.0
-rw-r--r--  1 root root   40K Dec 17  2017 libvtkPythonInterpreter-6.3.so.6.3.0
lrwxrwxrwx  1 root root    45 Dec 17  2017 libvtkPythonInterpreterPython27D-6.3.so.6.3 -> libvtkPythonInterpreterPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   28K Dec 17  2017 libvtkPythonInterpreterPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    39 Dec 17  2017 libvtkPythonInterpreterTCL-6.3.so.6.3 -> libvtkPythonInterpreterTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   27K Dec 17  2017 libvtkPythonInterpreterTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    38 Dec 17  2017 libvtkRenderingAnnotation-6.3.so.6.3 -> libvtkRenderingAnnotation-6.3.so.6.3.0
-rw-r--r--  1 root root  1.3M Dec 17  2017 libvtkRenderingAnnotation-6.3.so.6.3.0
lrwxrwxrwx  1 root root    47 Dec 17  2017 libvtkRenderingAnnotationPython27D-6.3.so.6.3 -> libvtkRenderingAnnotationPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  1.9M Dec 17  2017 libvtkRenderingAnnotationPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    41 Dec 17  2017 libvtkRenderingAnnotationTCL-6.3.so.6.3 -> libvtkRenderingAnnotationTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  945K Dec 17  2017 libvtkRenderingAnnotationTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    37 Dec 17  2017 libvtkRenderingContext2D-6.3.so.6.3 -> libvtkRenderingContext2D-6.3.so.6.3.0
-rw-r--r--  1 root root  252K Dec 17  2017 libvtkRenderingContext2D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    46 Dec 17  2017 libvtkRenderingContext2DPython27D-6.3.so.6.3 -> libvtkRenderingContext2DPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  320K Dec 17  2017 libvtkRenderingContext2DPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    41 Dec 17  2017 libvtkRenderingContextIIDTCL-6.3.so.6.3 -> libvtkRenderingContextIIDTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  212K Dec 17  2017 libvtkRenderingContextIIDTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    41 Dec 17  2017 libvtkRenderingContextOpenGL-6.3.so.6.3 -> libvtkRenderingContextOpenGL-6.3.so.6.3.0
-rw-r--r--  1 root root  196K Dec 17  2017 libvtkRenderingContextOpenGL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    50 Dec 17  2017 libvtkRenderingContextOpenGLPython27D-6.3.so.6.3 -> libvtkRenderingContextOpenGLPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   15K Dec 17  2017 libvtkRenderingContextOpenGLPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    44 Dec 17  2017 libvtkRenderingContextOpenGLTCL-6.3.so.6.3 -> libvtkRenderingContextOpenGLTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   15K Dec 17  2017 libvtkRenderingContextOpenGLTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    32 Dec 17  2017 libvtkRenderingCore-6.3.so.6.3 -> libvtkRenderingCore-6.3.so.6.3.0
-rw-r--r--  1 root root  2.5M Dec 17  2017 libvtkRenderingCore-6.3.so.6.3.0
lrwxrwxrwx  1 root root    41 Dec 17  2017 libvtkRenderingCorePython27D-6.3.so.6.3 -> libvtkRenderingCorePython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  3.6M Dec 17  2017 libvtkRenderingCorePython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    35 Dec 17  2017 libvtkRenderingCoreTCL-6.3.so.6.3 -> libvtkRenderingCoreTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  2.2M Dec 17  2017 libvtkRenderingCoreTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    36 Dec 17  2017 libvtkRenderingExternal-6.3.so.6.3 -> libvtkRenderingExternal-6.3.so.6.3.0
-rw-r--r--  1 root root  236K Dec 17  2017 libvtkRenderingExternal-6.3.so.6.3.0
lrwxrwxrwx  1 root root    45 Dec 17  2017 libvtkRenderingExternalPython27D-6.3.so.6.3 -> libvtkRenderingExternalPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   86K Dec 17  2017 libvtkRenderingExternalPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    39 Dec 17  2017 libvtkRenderingExternalTCL-6.3.so.6.3 -> libvtkRenderingExternalTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   71K Dec 17  2017 libvtkRenderingExternalTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    36 Dec 17  2017 libvtkRenderingFreeType-6.3.so.6.3 -> libvtkRenderingFreeType-6.3.so.6.3.0
-rw-r--r--  1 root root  781K Dec 17  2017 libvtkRenderingFreeType-6.3.so.6.3.0
lrwxrwxrwx  1 root root    46 Dec 17  2017 libvtkRenderingFreeTypeFontConfig-6.3.so.6.3 -> libvtkRenderingFreeTypeFontConfig-6.3.so.6.3.0
-rw-r--r--  1 root root   39K Dec 17  2017 libvtkRenderingFreeTypeFontConfig-6.3.so.6.3.0
lrwxrwxrwx  1 root root    45 Dec 17  2017 libvtkRenderingFreeTypePython27D-6.3.so.6.3 -> libvtkRenderingFreeTypePython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   61K Dec 17  2017 libvtkRenderingFreeTypePython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    39 Dec 17  2017 libvtkRenderingFreeTypeTCL-6.3.so.6.3 -> libvtkRenderingFreeTypeTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   47K Dec 17  2017 libvtkRenderingFreeTypeTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    33 Dec 17  2017 libvtkRenderingGL2PS-6.3.so.6.3 -> libvtkRenderingGL2PS-6.3.so.6.3.0
-rw-r--r--  1 root root  112K Dec 17  2017 libvtkRenderingGL2PS-6.3.so.6.3.0
lrwxrwxrwx  1 root root    42 Dec 17  2017 libvtkRenderingGL2PSPython27D-6.3.so.6.3 -> libvtkRenderingGL2PSPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   23K Dec 17  2017 libvtkRenderingGL2PSPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    37 Dec 17  2017 libvtkRenderingGLtoPSTCL-6.3.so.6.3 -> libvtkRenderingGLtoPSTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   19K Dec 17  2017 libvtkRenderingGLtoPSTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    33 Dec 17  2017 libvtkRenderingImage-6.3.so.6.3 -> libvtkRenderingImage-6.3.so.6.3.0
-rw-r--r--  1 root root  136K Dec 17  2017 libvtkRenderingImage-6.3.so.6.3.0
lrwxrwxrwx  1 root root    42 Dec 17  2017 libvtkRenderingImagePython27D-6.3.so.6.3 -> libvtkRenderingImagePython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  110K Dec 17  2017 libvtkRenderingImagePython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    36 Dec 17  2017 libvtkRenderingImageTCL-6.3.so.6.3 -> libvtkRenderingImageTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   63K Dec 17  2017 libvtkRenderingImageTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    31 Dec 17  2017 libvtkRenderingLIC-6.3.so.6.3 -> libvtkRenderingLIC-6.3.so.6.3.0
-rw-r--r--  1 root root  647K Dec 17  2017 libvtkRenderingLIC-6.3.so.6.3.0
lrwxrwxrwx  1 root root    40 Dec 17  2017 libvtkRenderingLICPython27D-6.3.so.6.3 -> libvtkRenderingLICPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  164K Dec 17  2017 libvtkRenderingLICPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    34 Dec 17  2017 libvtkRenderingLICTCL-6.3.so.6.3 -> libvtkRenderingLICTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   99K Dec 17  2017 libvtkRenderingLICTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    31 Dec 17  2017 libvtkRenderingLOD-6.3.so.6.3 -> libvtkRenderingLOD-6.3.so.6.3.0
-rw-r--r--  1 root root   99K Dec 17  2017 libvtkRenderingLOD-6.3.so.6.3.0
lrwxrwxrwx  1 root root    40 Dec 17  2017 libvtkRenderingLODPython27D-6.3.so.6.3 -> libvtkRenderingLODPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  106K Dec 17  2017 libvtkRenderingLODPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    34 Dec 17  2017 libvtkRenderingLODTCL-6.3.so.6.3 -> libvtkRenderingLODTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   59K Dec 17  2017 libvtkRenderingLODTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    33 Dec 17  2017 libvtkRenderingLabel-6.3.so.6.3 -> libvtkRenderingLabel-6.3.so.6.3.0
-rw-r--r--  1 root root  689K Dec 17  2017 libvtkRenderingLabel-6.3.so.6.3.0
lrwxrwxrwx  1 root root    42 Dec 17  2017 libvtkRenderingLabelPython27D-6.3.so.6.3 -> libvtkRenderingLabelPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  434K Dec 17  2017 libvtkRenderingLabelPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    36 Dec 17  2017 libvtkRenderingLabelTCL-6.3.so.6.3 -> libvtkRenderingLabelTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  239K Dec 17  2017 libvtkRenderingLabelTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    38 Dec 17  2017 libvtkRenderingMatplotlib-6.3.so.6.3 -> libvtkRenderingMatplotlib-6.3.so.6.3.0
-rw-r--r--  1 root root   67K Dec 17  2017 libvtkRenderingMatplotlib-6.3.so.6.3.0
lrwxrwxrwx  1 root root    47 Dec 17  2017 libvtkRenderingMatplotlibPython27D-6.3.so.6.3 -> libvtkRenderingMatplotlibPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   23K Dec 17  2017 libvtkRenderingMatplotlibPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    41 Dec 17  2017 libvtkRenderingMatplotlibTCL-6.3.so.6.3 -> libvtkRenderingMatplotlibTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   23K Dec 17  2017 libvtkRenderingMatplotlibTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    34 Dec 17  2017 libvtkRenderingOpenGL-6.3.so.6.3 -> libvtkRenderingOpenGL-6.3.so.6.3.0
-rw-r--r--  1 root root  2.7M Dec 17  2017 libvtkRenderingOpenGL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    43 Dec 17  2017 libvtkRenderingOpenGLPython27D-6.3.so.6.3 -> libvtkRenderingOpenGLPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  983K Dec 17  2017 libvtkRenderingOpenGLPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    37 Dec 17  2017 libvtkRenderingOpenGLTCL-6.3.so.6.3 -> libvtkRenderingOpenGLTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  711K Dec 17  2017 libvtkRenderingOpenGLTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    36 Dec 17  2017 libvtkRenderingParallel-6.3.so.6.3 -> libvtkRenderingParallel-6.3.so.6.3.0
-rw-r--r--  1 root root  341K Dec 17  2017 libvtkRenderingParallel-6.3.so.6.3.0
lrwxrwxrwx  1 root root    39 Dec 17  2017 libvtkRenderingParallelLIC-6.3.so.6.3 -> libvtkRenderingParallelLIC-6.3.so.6.3.0
-rw-r--r--  1 root root  621K Dec 17  2017 libvtkRenderingParallelLIC-6.3.so.6.3.0
lrwxrwxrwx  1 root root    48 Dec 17  2017 libvtkRenderingParallelLICPython27D-6.3.so.6.3 -> libvtkRenderingParallelLICPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   48K Dec 17  2017 libvtkRenderingParallelLICPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    42 Dec 17  2017 libvtkRenderingParallelLICTCL-6.3.so.6.3 -> libvtkRenderingParallelLICTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   39K Dec 17  2017 libvtkRenderingParallelLICTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    45 Dec 17  2017 libvtkRenderingParallelPython27D-6.3.so.6.3 -> libvtkRenderingParallelPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  284K Dec 17  2017 libvtkRenderingParallelPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    39 Dec 17  2017 libvtkRenderingParallelTCL-6.3.so.6.3 -> libvtkRenderingParallelTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  179K Dec 17  2017 libvtkRenderingParallelTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    33 Dec 17  2017 libvtkRenderingTkTCL-6.3.so.6.3 -> libvtkRenderingTkTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  104K Dec 17  2017 libvtkRenderingTkTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    34 Dec 17  2017 libvtkRenderingVolume-6.3.so.6.3 -> libvtkRenderingVolume-6.3.so.6.3.0
-rw-r--r--  1 root root  4.2M Dec 17  2017 libvtkRenderingVolume-6.3.so.6.3.0
lrwxrwxrwx  1 root root    37 Dec 17  2017 libvtkRenderingVolumeAMR-6.3.so.6.3 -> libvtkRenderingVolumeAMR-6.3.so.6.3.0
-rw-r--r--  1 root root   87K Dec 17  2017 libvtkRenderingVolumeAMR-6.3.so.6.3.0
lrwxrwxrwx  1 root root    46 Dec 17  2017 libvtkRenderingVolumeAMRPython27D-6.3.so.6.3 -> libvtkRenderingVolumeAMRPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   69K Dec 17  2017 libvtkRenderingVolumeAMRPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    40 Dec 17  2017 libvtkRenderingVolumeAMRTCL-6.3.so.6.3 -> libvtkRenderingVolumeAMRTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   39K Dec 17  2017 libvtkRenderingVolumeAMRTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    40 Dec 17  2017 libvtkRenderingVolumeOpenGL-6.3.so.6.3 -> libvtkRenderingVolumeOpenGL-6.3.so.6.3.0
-rw-r--r--  1 root root  670K Dec 17  2017 libvtkRenderingVolumeOpenGL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    49 Dec 17  2017 libvtkRenderingVolumeOpenGLPython27D-6.3.so.6.3 -> libvtkRenderingVolumeOpenGLPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  127K Dec 17  2017 libvtkRenderingVolumeOpenGLPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    43 Dec 17  2017 libvtkRenderingVolumeOpenGLTCL-6.3.so.6.3 -> libvtkRenderingVolumeOpenGLTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   91K Dec 17  2017 libvtkRenderingVolumeOpenGLTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    43 Dec 17  2017 libvtkRenderingVolumePython27D-6.3.so.6.3 -> libvtkRenderingVolumePython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  1.1M Dec 17  2017 libvtkRenderingVolumePython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    37 Dec 17  2017 libvtkRenderingVolumeTCL-6.3.so.6.3 -> libvtkRenderingVolumeTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  620K Dec 17  2017 libvtkRenderingVolumeTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    39 Dec 17  2017 libvtkTestingGenericBridge-6.3.so.6.3 -> libvtkTestingGenericBridge-6.3.so.6.3.0
-rw-r--r--  1 root root  132K Dec 17  2017 libvtkTestingGenericBridge-6.3.so.6.3.0
lrwxrwxrwx  1 root root    31 Dec 17  2017 libvtkTestingIOSQL-6.3.so.6.3 -> libvtkTestingIOSQL-6.3.so.6.3.0
-rw-r--r--  1 root root   15K Dec 17  2017 libvtkTestingIOSQL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    35 Dec 17  2017 libvtkTestingRendering-6.3.so.6.3 -> libvtkTestingRendering-6.3.so.6.3.0
-rw-r--r--  1 root root  136K Dec 17  2017 libvtkTestingRendering-6.3.so.6.3.0
lrwxrwxrwx  1 root root    44 Dec 17  2017 libvtkTestingRenderingPython27D-6.3.so.6.3 -> libvtkTestingRenderingPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   61K Dec 17  2017 libvtkTestingRenderingPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    38 Dec 17  2017 libvtkTestingRenderingTCL-6.3.so.6.3 -> libvtkTestingRenderingTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   43K Dec 17  2017 libvtkTestingRenderingTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    23 Dec 17  2017 libvtkVPIC-6.3.so.6.3 -> libvtkVPIC-6.3.so.6.3.0
-rw-r--r--  1 root root   95K Dec 17  2017 libvtkVPIC-6.3.so.6.3.0
lrwxrwxrwx  1 root root    33 Dec 17  2017 libvtkViewsContext2D-6.3.so.6.3 -> libvtkViewsContext2D-6.3.so.6.3.0
-rw-r--r--  1 root root   75K Dec 17  2017 libvtkViewsContext2D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    42 Dec 17  2017 libvtkViewsContext2DPython27D-6.3.so.6.3 -> libvtkViewsContext2DPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   32K Dec 17  2017 libvtkViewsContext2DPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    37 Dec 17  2017 libvtkViewsContextIIDTCL-6.3.so.6.3 -> libvtkViewsContextIIDTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   27K Dec 17  2017 libvtkViewsContextIIDTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    28 Dec 17  2017 libvtkViewsCore-6.3.so.6.3 -> libvtkViewsCore-6.3.so.6.3.0
-rw-r--r--  1 root root  172K Dec 17  2017 libvtkViewsCore-6.3.so.6.3.0
lrwxrwxrwx  1 root root    37 Dec 17  2017 libvtkViewsCorePython27D-6.3.so.6.3 -> libvtkViewsCorePython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  181K Dec 17  2017 libvtkViewsCorePython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    31 Dec 17  2017 libvtkViewsCoreTCL-6.3.so.6.3 -> libvtkViewsCoreTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  119K Dec 17  2017 libvtkViewsCoreTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    30 Dec 17  2017 libvtkViewsGeovis-6.3.so.6.3 -> libvtkViewsGeovis-6.3.so.6.3.0
-rw-r--r--  1 root root   71K Dec 17  2017 libvtkViewsGeovis-6.3.so.6.3.0
lrwxrwxrwx  1 root root    39 Dec 17  2017 libvtkViewsGeovisPython27D-6.3.so.6.3 -> libvtkViewsGeovisPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   32K Dec 17  2017 libvtkViewsGeovisPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    33 Dec 17  2017 libvtkViewsGeovisTCL-6.3.so.6.3 -> libvtkViewsGeovisTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   27K Dec 17  2017 libvtkViewsGeovisTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    31 Dec 17  2017 libvtkViewsInfovis-6.3.so.6.3 -> libvtkViewsInfovis-6.3.so.6.3.0
-rw-r--r--  1 root root  1.1M Dec 17  2017 libvtkViewsInfovis-6.3.so.6.3.0
lrwxrwxrwx  1 root root    40 Dec 17  2017 libvtkViewsInfovisPython27D-6.3.so.6.3 -> libvtkViewsInfovisPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  927K Dec 17  2017 libvtkViewsInfovisPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    34 Dec 17  2017 libvtkViewsInfovisTCL-6.3.so.6.3 -> libvtkViewsInfovisTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  565K Dec 17  2017 libvtkViewsInfovisTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    31 Dec 17  2017 libvtkWrappingJava-6.3.so.6.3 -> libvtkWrappingJava-6.3.so.6.3.0
-rw-r--r--  1 root root   19K Dec 17  2017 libvtkWrappingJava-6.3.so.6.3.0
lrwxrwxrwx  1 root root    39 Dec 17  2017 libvtkWrappingPython27Core-6.3.so.6.3 -> libvtkWrappingPython27Core-6.3.so.6.3.0
-rw-r--r--  1 root root  180K Dec 17  2017 libvtkWrappingPython27Core-6.3.so.6.3.0
lrwxrwxrwx  1 root root    25 Dec 17  2017 libvtkalglib-6.3.so.6.3 -> libvtkalglib-6.3.so.6.3.0
-rw-r--r--  1 root root  139K Dec 17  2017 libvtkalglib-6.3.so.6.3.0
lrwxrwxrwx  1 root root    25 Dec 17  2017 libvtkexoIIc-6.3.so.6.3 -> libvtkexoIIc-6.3.so.6.3.0
-rw-r--r--  1 root root  288K Dec 17  2017 libvtkexoIIc-6.3.so.6.3.0
lrwxrwxrwx  1 root root    23 Dec 17  2017 libvtkftgl-6.3.so.6.3 -> libvtkftgl-6.3.so.6.3.0
-rw-r--r--  1 root root   43K Dec 17  2017 libvtkftgl-6.3.so.6.3.0
lrwxrwxrwx  1 root root    25 Dec 17  2017 libvtkmetaio-6.3.so.6.3 -> libvtkmetaio-6.3.so.6.3.0
-rw-r--r--  1 root root  591K Dec 17  2017 libvtkmetaio-6.3.so.6.3.0
lrwxrwxrwx  1 root root    22 Dec 17  2017 libvtksys-6.3.so.6.3 -> libvtksys-6.3.so.6.3.0
-rw-r--r--  1 root root  278K Dec 17  2017 libvtksys-6.3.so.6.3.0
lrwxrwxrwx  1 root root    26 Dec 17  2017 libvtkverdict-6.3.so.6.3 -> libvtkverdict-6.3.so.6.3.0
-rw-r--r--  1 root root  195K Dec 17  2017 libvtkverdict-6.3.so.6.3.0
lrwxrwxrwx  1 root root    24 Dec 17  2017 libvtkxdmf2-6.3.so.6.3 -> libvtkxdmf2-6.3.so.6.3.0
-rw-r--r--  1 root root  472K Dec 17  2017 libvtkxdmf2-6.3.so.6.3.0
lrwxrwxrwx  1 root root    19 Jan  5  2021 libwavpack.so.1 -> libwavpack.so.1.2.0
-rw-r--r--  1 root root  167K Jan  5  2021 libwavpack.so.1.2.0
lrwxrwxrwx  1 root root    26 Feb  1  2020 libwayland-client.so.0 -> libwayland-client.so.0.3.0
-rw-r--r--  1 root root   60K Feb  1  2020 libwayland-client.so.0.3.0
lrwxrwxrwx  1 root root    26 Feb  1  2020 libwayland-cursor.so.0 -> libwayland-cursor.so.0.0.0
-rw-r--r--  1 root root   31K Feb  1  2020 libwayland-cursor.so.0.0.0
lrwxrwxrwx  1 root root    23 Feb  1  2020 libwayland-egl.so.1 -> libwayland-egl.so.1.0.0
-rw-r--r--  1 root root  5.9K Feb  1  2020 libwayland-egl.so.1.0.0
lrwxrwxrwx  1 root root    16 May 20  2021 libwebp.so.6 -> libwebp.so.6.0.2
-rw-r--r--  1 root root  411K May 20  2021 libwebp.so.6.0.2
lrwxrwxrwx  1 root root    19 May 20  2021 libwebpmux.so.3 -> libwebpmux.so.3.0.1
-rw-r--r--  1 root root   39K May 20  2021 libwebpmux.so.3.0.1
lrwxrwxrwx  1 root root    16 Dec 15  2017 libwind.so.0 -> libwind.so.0.0.0
-rw-r--r--  1 root root  162K Dec 15  2017 libwind.so.0.0.0
-rw-r--r--  1 root root  1.2M Jan 19  2018 libx264.so.152
-rw-r--r--  1 root root   11M Dec 31  2017 libx265.so.146
lrwxrwxrwx  1 root root    20 Jun 21  2018 libxcb-dri2.so.0 -> libxcb-dri2.so.0.0.0
-rw-r--r--  1 root root   19K Jun 21  2018 libxcb-dri2.so.0.0.0
lrwxrwxrwx  1 root root    20 Jun 21  2018 libxcb-dri3.so.0 -> libxcb-dri3.so.0.0.0
-rw-r--r--  1 root root   15K Jun 21  2018 libxcb-dri3.so.0.0.0
lrwxrwxrwx  1 root root    19 Jun 21  2018 libxcb-glx.so.0 -> libxcb-glx.so.0.0.0
-rw-r--r--  1 root root  107K Jun 21  2018 libxcb-glx.so.0.0.0
lrwxrwxrwx  1 root root    23 Jun 21  2018 libxcb-present.so.0 -> libxcb-present.so.0.0.0
-rw-r--r--  1 root root   11K Jun 21  2018 libxcb-present.so.0.0.0
lrwxrwxrwx  1 root root    22 Jun 21  2018 libxcb-render.so.0 -> libxcb-render.so.0.0.0
-rw-r--r--  1 root root   51K Jun 21  2018 libxcb-render.so.0.0.0
lrwxrwxrwx  1 root root    19 Jun 21  2018 libxcb-shm.so.0 -> libxcb-shm.so.0.0.0
-rw-r--r--  1 root root   11K Jun 21  2018 libxcb-shm.so.0.0.0
lrwxrwxrwx  1 root root    20 Jun 21  2018 libxcb-sync.so.1 -> libxcb-sync.so.1.0.0
-rw-r--r--  1 root root   27K Jun 21  2018 libxcb-sync.so.1.0.0
lrwxrwxrwx  1 root root    15 Jun 21  2018 libxcb.so.1 -> libxcb.so.1.1.0
-rw-r--r--  1 root root  159K Jun 21  2018 libxcb.so.1.1.0
lrwxrwxrwx  1 root root    16 Mar 24  2018 libxdot.so -> libxdot.so.4.0.0
lrwxrwxrwx  1 root root    16 Mar 24  2018 libxdot.so.4 -> libxdot.so.4.0.0
-rw-r--r--  1 root root   23K Mar 24  2018 libxdot.so.4.0.0
-rw-r--r--  1 root root  3.5M Nov 10  2017 libxerces-c-3.2.so
lrwxrwxrwx  1 root root    21 Mar  5  2019 libxkbcommon.so.0 -> libxkbcommon.so.0.0.0
-rw-r--r--  1 root root  251K Mar  5  2019 libxkbcommon.so.0.0.0
-rw-r--r--  1 root root  2.9M Aug  1 14:25 libxml2.a
lrwxrwxrwx  1 root root    16 Aug  1 14:25 libxml2.so -> libxml2.so.2.9.4
lrwxrwxrwx  1 root root    16 Aug  1 14:25 libxml2.so.2 -> libxml2.so.2.9.4
-rw-r--r--  1 root root  1.8M Aug  1 14:25 libxml2.so.2.9.4
lrwxrwxrwx  1 root root    21 Mar 18  2018 libxshmfence.so.1 -> libxshmfence.so.1.0.0
-rw-r--r--  1 root root  6.1K Mar 18  2018 libxshmfence.so.1.0.0
lrwxrwxrwx  1 root root    18 Jan 28  2018 libxvidcore.so.4 -> libxvidcore.so.4.3
-rw-r--r--  1 root root  672K Jan 28  2018 libxvidcore.so.4.3
-rw-r--r--  1 root root  162K Mar 26  2022 libz.a
lrwxrwxrwx  1 root root    36 Mar 26  2022 libz.so -> /lib/x86_64-linux-gnu/libz.so.1.2.11
lrwxrwxrwx  1 root root    16 Mar  3  2021 libzstd.so.1 -> libzstd.so.1.3.3
-rw-r--r--  1 root root  490K Mar  3  2021 libzstd.so.1.3.3
lrwxrwxrwx  1 root root    23 Nov 17  2016 libzvbi-chains.so.0 -> libzvbi-chains.so.0.0.0
-rw-r--r--  1 root root   58K Nov 17  2016 libzvbi-chains.so.0.0.0
lrwxrwxrwx  1 root root    17 Nov 17  2016 libzvbi.so.0 -> libzvbi.so.0.13.2
-rw-r--r--  1 root root  556K Nov 17  2016 libzvbi.so.0.13.2
drwxr-xr-x  2 root root  4.0K Aug 16 08:41 nss
drwxr-xr-x  2 root root  4.0K Aug 16 08:41 odbc
drwxr-xr-x  2 root root  4.0K Aug 16 08:41 openblas
drwxr-xr-x  3 root root  4.0K Aug 16 08:41 openmpi
drwxr-xr-x  1 root root  4.0K Aug 16 08:40 perl
drwxr-xr-x 16 root root  4.0K Aug  1 13:21 perl-base
drwxr-xr-x  2 root root  4.0K Aug 16 08:41 pkgconfig
-rw-r--r--  1 root root  1.8K May  3 10:19 rcrt1.o
drwxr-xr-x  2 root root  4.0K Aug 16 08:41 rsocket
drwxr-xr-x  2 root root  4.0K Aug 16 08:41 sasl2
drwxr-xr-x  2 root root  4.0K Aug 16 08:41 vdpau
drwxr-xr-x  2 root root  4.0K Aug 16 08:41 x264-10bit
-rw-r--r--  1 root root   204 Aug  1 14:25 xml2Conf.sh
[Pipeline] echo
Packing build/libdgl.so, build/runUnitTests, python/dgl/_ffi/_cy3/core.cpython-*-x86_64-linux-gnu.so, build/tensoradapter/pytorch/*.so into dgl-cpu-linux
[Pipeline] stash
Stashed 4 file(s)
Post stage
[Pipeline] cleanWs
[WS-CLEANUP] Deleting project workspace...
[WS-CLEANUP] Deferred wipeout is disabled by the job configuration...
[WS-CLEANUP] done
[Pipeline] }
$ docker stop --time=1 856b09e8bd127c677672426c8af0a6bfefd91dd73a4c28827d4d41ef0f5d0fcc
$ docker rm -f 856b09e8bd127c677672426c8af0a6bfefd91dd73a4c28827d4d41ef0f5d0fcc
[Pipeline] // withDockerContainer
[Pipeline] }
[Pipeline] // withEnv
[Pipeline] }
[Pipeline] // node
[Pipeline] }
[Pipeline] // stage
[Pipeline] }
Archiving  objects                             > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/obj/collectives/device/colldevice.a
Linking    libnccl.so.2.11.4                   > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/lib/libnccl.so.2.11.4
Archiving  libnccl_static.a                    > /root/jenkins/workspace/dgl_PR-4648@2/build/nccl/lib/libnccl_static.a
/root/jenkins/workspace/dgl_PR-4648@2/third_party/nccl/src
[ 49%] No install step for 'nccl_external'
[ 49%] Completed 'nccl_external'
[ 49%] Built target nccl_external
[ 49%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/runtime/cuda/dgl_generated_nccl_api.cu.o
[ 49%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/array/cuda/dgl_generated_array_cumsum.cu.o
[ 49%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/array/cuda/dgl_generated_array_index_select.cu.o
[ 49%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/array/cuda/dgl_generated_array_nonzero.cu.o
[ 49%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/array/cuda/dgl_generated_array_op_impl.cu.o
[ 50%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/array/cuda/dgl_generated_array_scatter.cu.o
[ 50%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/array/cuda/dgl_generated_csr_get_data.cu.o
[ 50%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/array/cuda/dgl_generated_array_sort.cu.o
[ 50%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/array/cuda/dgl_generated_csr_mm.cu.o
[ 50%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/array/cuda/dgl_generated_coo2csr.cu.o
[ 50%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/array/cuda/dgl_generated_coo_sort.cu.o
[ 51%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/array/cuda/dgl_generated_csr2coo.cu.o
[ 51%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/array/cuda/dgl_generated_csr_sort.cu.o
[ 53%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/array/cuda/dgl_generated_csr_sum.cu.o
[ 53%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/array/cuda/dgl_generated_cuda_filter.cu.o
[ 53%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/array/cuda/dgl_generated_disjoint_union.cu.o
[ 53%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/array/cuda/dgl_generated_gather_mm.cu.o
[ 53%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/array/cuda/dgl_generated_negative_sampling.cu.o
[ 53%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/array/cuda/dgl_generated_rowwise_sampling_prob.cu.o
[ 54%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/array/cuda/dgl_generated_rowwise_sampling.cu.o
[ 54%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/array/cuda/dgl_generated_sddmm_hetero_coo.cu.o
[ 54%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/array/cuda/dgl_generated_sddmm.cu.o
[ 54%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/array/cuda/dgl_generated_spmat_op_impl_csr.cu.o
[ 54%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/array/cuda/dgl_generated_segment_reduce.cu.o
[ 54%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/array/cuda/dgl_generated_spmat_op_impl_coo.cu.o
[ 55%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/array/cuda/dgl_generated_sddmm_hetero_csr.cu.o
[ 56%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/geometry/cuda/dgl_generated_geometry_op_impl.cu.o
[ 57%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/array/cuda/dgl_generated_spmm.cu.o
[ 57%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/array/cuda/dgl_generated_spmm_hetero.cu.o
[ 57%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/array/cuda/dgl_generated_utils.cu.o
[ 57%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/array/cuda/uvm/dgl_generated_array_index_select_uvm.cu.o
[ 57%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/graph/sampling/randomwalks/dgl_generated_frequency_hashmap.cu.o
[ 57%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/geometry/cuda/dgl_generated_edge_coarsening_impl.cu.o
[ 57%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/graph/sampling/randomwalks/dgl_generated_get_node_types_gpu.cu.o
[ 57%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/graph/sampling/randomwalks/dgl_generated_randomwalk_gpu.cu.o
[ 59%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/graph/transform/cuda/dgl_generated_cuda_compact_graph.cu.o
[ 59%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/graph/transform/cuda/dgl_generated_cuda_to_block.cu.o
[ 59%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/partition/cuda/dgl_generated_partition_op.cu.o
[ 59%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/graph/transform/cuda/dgl_generated_knn.cu.o
[ 60%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/runtime/cuda/dgl_generated_cuda_hashtable.cu.o
/root/jenkins/workspace/dgl_PR-4648@2/src/array/cuda/gather_mm.cu(181): warning: variable "out_reg" was declared but never referenced
          detected during:
            instantiation of "void dgl::aten::cuda::GatherMMScatterKernel2(const DType *, const DType *, DType *, const Idx *, const Idx *, const Idx *, int64_t, int64_t, int64_t) [with Idx=int32_t, DType=float]" 
(391): here
            instantiation of "void dgl::aten::GatherMMScatter<XPU,IdType,DType>(dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray) [with XPU=2, IdType=int32_t, DType=float]" 
(444): here

/root/jenkins/workspace/dgl_PR-4648@2/src/array/cuda/gather_mm.cu(181): warning: variable "out_reg" was declared but never referenced
          detected during:
            instantiation of "void dgl::aten::cuda::GatherMMScatterKernel2(const DType *, const DType *, DType *, const Idx *, const Idx *, const Idx *, int64_t, int64_t, int64_t) [with Idx=int64_t, DType=float]" 
(391): here
            instantiation of "void dgl::aten::GatherMMScatter<XPU,IdType,DType>(dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray) [with XPU=2, IdType=int64_t, DType=float]" 
(447): here

/root/jenkins/workspace/dgl_PR-4648@2/src/array/cuda/gather_mm.cu(181): warning: variable "out_reg" was declared but never referenced
          detected during:
            instantiation of "void dgl::aten::cuda::GatherMMScatterKernel2(const DType *, const DType *, DType *, const Idx *, const Idx *, const Idx *, int64_t, int64_t, int64_t) [with Idx=int32_t, DType=double]" 
(391): here
            instantiation of "void dgl::aten::GatherMMScatter<XPU,IdType,DType>(dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray) [with XPU=2, IdType=int32_t, DType=double]" 
(450): here

/root/jenkins/workspace/dgl_PR-4648@2/src/array/cuda/gather_mm.cu(181): warning: variable "out_reg" was declared but never referenced
          detected during:
            instantiation of "void dgl::aten::cuda::GatherMMScatterKernel2(const DType *, const DType *, DType *, const Idx *, const Idx *, const Idx *, int64_t, int64_t, int64_t) [with Idx=int64_t, DType=double]" 
(391): here
            instantiation of "void dgl::aten::GatherMMScatter<XPU,IdType,DType>(dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray) [with XPU=2, IdType=int64_t, DType=double]" 
(453): here

/root/jenkins/workspace/dgl_PR-4648@2/src/array/cuda/gather_mm.cu(181): warning: variable "out_reg" was declared but never referenced
          detected during:
            instantiation of "void dgl::aten::cuda::GatherMMScatterKernel2(const DType *, const DType *, DType *, const Idx *, const Idx *, const Idx *, int64_t, int64_t, int64_t) [with Idx=int32_t, DType=float]" 
(391): here
            instantiation of "void dgl::aten::GatherMMScatter<XPU,IdType,DType>(dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray) [with XPU=2, IdType=int32_t, DType=float]" 
(444): here

/root/jenkins/workspace/dgl_PR-4648@2/src/array/cuda/gather_mm.cu(181): warning: variable "out_reg" was declared but never referenced
          detected during:
            instantiation of "void dgl::aten::cuda::GatherMMScatterKernel2(const DType *, const DType *, DType *, const Idx *, const Idx *, const Idx *, int64_t, int64_t, int64_t) [with Idx=int64_t, DType=float]" 
(391): here
            instantiation of "void dgl::aten::GatherMMScatter<XPU,IdType,DType>(dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray) [with XPU=2, IdType=int64_t, DType=float]" 
(447): here

/root/jenkins/workspace/dgl_PR-4648@2/src/array/cuda/gather_mm.cu(181): warning: variable "out_reg" was declared but never referenced
          detected during:
            instantiation of "void dgl::aten::cuda::GatherMMScatterKernel2(const DType *, const DType *, DType *, const Idx *, const Idx *, const Idx *, int64_t, int64_t, int64_t) [with Idx=int32_t, DType=double]" 
(391): here
            instantiation of "void dgl::aten::GatherMMScatter<XPU,IdType,DType>(dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray) [with XPU=2, IdType=int32_t, DType=double]" 
(450): here

/root/jenkins/workspace/dgl_PR-4648@2/src/array/cuda/gather_mm.cu(181): warning: variable "out_reg" was declared but never referenced
          detected during:
            instantiation of "void dgl::aten::cuda::GatherMMScatterKernel2(const DType *, const DType *, DType *, const Idx *, const Idx *, const Idx *, int64_t, int64_t, int64_t) [with Idx=int64_t, DType=double]" 
(391): here
            instantiation of "void dgl::aten::GatherMMScatter<XPU,IdType,DType>(dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray) [with XPU=2, IdType=int64_t, DType=double]" 
(453): here

/root/jenkins/workspace/dgl_PR-4648@2/src/array/cuda/gather_mm.cu(181): warning: variable "out_reg" was declared but never referenced
          detected during:
            instantiation of "void dgl::aten::cuda::GatherMMScatterKernel2(const DType *, const DType *, DType *, const Idx *, const Idx *, const Idx *, int64_t, int64_t, int64_t) [with Idx=int32_t, DType=float]" 
(391): here
            instantiation of "void dgl::aten::GatherMMScatter<XPU,IdType,DType>(dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray) [with XPU=2, IdType=int32_t, DType=float]" 
(444): here

/root/jenkins/workspace/dgl_PR-4648@2/src/array/cuda/gather_mm.cu(181): warning: variable "out_reg" was declared but never referenced
          detected during:
            instantiation of "void dgl::aten::cuda::GatherMMScatterKernel2(const DType *, const DType *, DType *, const Idx *, const Idx *, const Idx *, int64_t, int64_t, int64_t) [with Idx=int64_t, DType=float]" 
(391): here
            instantiation of "void dgl::aten::GatherMMScatter<XPU,IdType,DType>(dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray) [with XPU=2, IdType=int64_t, DType=float]" 
(447): here

/root/jenkins/workspace/dgl_PR-4648@2/src/array/cuda/gather_mm.cu(181): warning: variable "out_reg" was declared but never referenced
          detected during:
            instantiation of "void dgl::aten::cuda::GatherMMScatterKernel2(const DType *, const DType *, DType *, const Idx *, const Idx *, const Idx *, int64_t, int64_t, int64_t) [with Idx=int32_t, DType=double]" 
(391): here
            instantiation of "void dgl::aten::GatherMMScatter<XPU,IdType,DType>(dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray) [with XPU=2, IdType=int32_t, DType=double]" 
(450): here

/root/jenkins/workspace/dgl_PR-4648@2/src/array/cuda/gather_mm.cu(181): warning: variable "out_reg" was declared but never referenced
          detected during:
            instantiation of "void dgl::aten::cuda::GatherMMScatterKernel2(const DType *, const DType *, DType *, const Idx *, const Idx *, const Idx *, int64_t, int64_t, int64_t) [with Idx=int64_t, DType=double]" 
(391): here
            instantiation of "void dgl::aten::GatherMMScatter<XPU,IdType,DType>(dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray) [with XPU=2, IdType=int64_t, DType=double]" 
(453): here

/root/jenkins/workspace/dgl_PR-4648@2/src/array/cuda/gather_mm.cu(181): warning: variable "out_reg" was declared but never referenced
          detected during:
            instantiation of "void dgl::aten::cuda::GatherMMScatterKernel2(const DType *, const DType *, DType *, const Idx *, const Idx *, const Idx *, int64_t, int64_t, int64_t) [with Idx=int32_t, DType=float]" 
(391): here
            instantiation of "void dgl::aten::GatherMMScatter<XPU,IdType,DType>(dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray) [with XPU=2, IdType=int32_t, DType=float]" 
(444): here

/root/jenkins/workspace/dgl_PR-4648@2/src/array/cuda/gather_mm.cu(181): warning: variable "out_reg" was declared but never referenced
          detected during:
            instantiation of "void dgl::aten::cuda::GatherMMScatterKernel2(const DType *, const DType *, DType *, const Idx *, const Idx *, const Idx *, int64_t, int64_t, int64_t) [with Idx=int64_t, DType=float]" 
(391): here
            instantiation of "void dgl::aten::GatherMMScatter<XPU,IdType,DType>(dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray) [with XPU=2, IdType=int64_t, DType=float]" 
(447): here

/root/jenkins/workspace/dgl_PR-4648@2/src/array/cuda/gather_mm.cu(181): warning: variable "out_reg" was declared but never referenced
          detected during:
            instantiation of "void dgl::aten::cuda::GatherMMScatterKernel2(const DType *, const DType *, DType *, const Idx *, const Idx *, const Idx *, int64_t, int64_t, int64_t) [with Idx=int32_t, DType=double]" 
(391): here
            instantiation of "void dgl::aten::GatherMMScatter<XPU,IdType,DType>(dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray) [with XPU=2, IdType=int32_t, DType=double]" 
(450): here

/root/jenkins/workspace/dgl_PR-4648@2/src/array/cuda/gather_mm.cu(181): warning: variable "out_reg" was declared but never referenced
          detected during:
            instantiation of "void dgl::aten::cuda::GatherMMScatterKernel2(const DType *, const DType *, DType *, const Idx *, const Idx *, const Idx *, int64_t, int64_t, int64_t) [with Idx=int64_t, DType=double]" 
(391): here
            instantiation of "void dgl::aten::GatherMMScatter<XPU,IdType,DType>(dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray) [with XPU=2, IdType=int64_t, DType=double]" 
(453): here

nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
Archiving  objects                             > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/obj/collectives/device/colldevice.a
Linking    libnccl.so.2.11.4                   > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/lib/libnccl.so.2.11.4
Archiving  libnccl_static.a                    > /root/jenkins/workspace/dgl_PR-4648@3/build/nccl/lib/libnccl_static.a
/root/jenkins/workspace/dgl_PR-4648@3/third_party/nccl/src
[ 49%] No install step for 'nccl_external'
[ 49%] Completed 'nccl_external'
[ 49%] Built target nccl_external
[ 50%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/array/cuda/dgl_generated_array_index_select.cu.o
[ 50%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/array/cuda/dgl_generated_array_nonzero.cu.o
[ 50%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/array/cuda/dgl_generated_array_scatter.cu.o
[ 50%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/array/cuda/dgl_generated_array_sort.cu.o
[ 50%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/array/cuda/dgl_generated_coo2csr.cu.o
[ 50%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/array/cuda/dgl_generated_array_op_impl.cu.o
[ 50%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/array/cuda/dgl_generated_coo_sort.cu.o
[ 51%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/array/cuda/dgl_generated_csr2coo.cu.o
[ 51%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/array/cuda/dgl_generated_csr_get_data.cu.o
[ 51%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/array/cuda/dgl_generated_array_cumsum.cu.o
[ 51%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/array/cuda/dgl_generated_cuda_filter.cu.o
[ 51%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/array/cuda/dgl_generated_rowwise_sampling_prob.cu.o
[ 51%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/array/cuda/dgl_generated_gather_mm.cu.o
[ 51%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/array/cuda/dgl_generated_sddmm.cu.o
[ 51%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/array/cuda/dgl_generated_sddmm_hetero_coo.cu.o
[ 53%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/array/cuda/dgl_generated_sddmm_hetero_csr.cu.o
[ 53%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/array/cuda/dgl_generated_segment_reduce.cu.o
[ 53%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/array/cuda/dgl_generated_csr_sort.cu.o
[ 53%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/runtime/cuda/dgl_generated_nccl_api.cu.o
[ 53%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/array/cuda/dgl_generated_csr_mm.cu.o
[ 53%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/array/cuda/dgl_generated_spmat_op_impl_csr.cu.o
[ 54%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/array/cuda/dgl_generated_spmm.cu.o
[ 54%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/array/cuda/dgl_generated_spmm_hetero.cu.o
[ 54%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/array/cuda/dgl_generated_disjoint_union.cu.o
[ 54%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/array/cuda/dgl_generated_utils.cu.o
[ 54%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/array/cuda/dgl_generated_negative_sampling.cu.o
[ 55%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/array/cuda/uvm/dgl_generated_array_index_select_uvm.cu.o
[ 55%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/array/cuda/dgl_generated_csr_sum.cu.o
[ 56%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/array/cuda/dgl_generated_rowwise_sampling.cu.o
[ 56%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/graph/sampling/randomwalks/dgl_generated_frequency_hashmap.cu.o
[ 56%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/geometry/cuda/dgl_generated_edge_coarsening_impl.cu.o
[ 57%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/geometry/cuda/dgl_generated_geometry_op_impl.cu.o
[ 57%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/graph/sampling/randomwalks/dgl_generated_get_node_types_gpu.cu.o
[ 57%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/graph/sampling/randomwalks/dgl_generated_randomwalk_gpu.cu.o
[ 57%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/array/cuda/dgl_generated_spmat_op_impl_coo.cu.o
[ 59%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/graph/transform/cuda/dgl_generated_cuda_compact_graph.cu.o
[ 59%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/graph/transform/cuda/dgl_generated_knn.cu.o
[ 59%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/graph/transform/cuda/dgl_generated_cuda_to_block.cu.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
[ 59%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/partition/cuda/dgl_generated_partition_op.cu.o
[ 60%] Building NVCC (Device) object CMakeFiles/dgl.dir/src/runtime/cuda/dgl_generated_cuda_hashtable.cu.o
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(56): warning #177-D: parameter "handle" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(56): warning #177-D: parameter "transa" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(57): warning #177-D: parameter "transb" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(57): warning #177-D: parameter "m" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(57): warning #177-D: parameter "n" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(58): warning #177-D: parameter "alpha" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(58): warning #177-D: parameter "A" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(58): warning #177-D: parameter "lda" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(59): warning #177-D: parameter "beta" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(59): warning #177-D: parameter "B" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(59): warning #177-D: parameter "ldb" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(60): warning #177-D: parameter "C" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(60): warning #177-D: parameter "ldc" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(69): warning #177-D: parameter "handle" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(69): warning #177-D: parameter "transa" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(70): warning #177-D: parameter "transb" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(70): warning #177-D: parameter "m" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(70): warning #177-D: parameter "n" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(71): warning #177-D: parameter "alpha" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(71): warning #177-D: parameter "A" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(71): warning #177-D: parameter "lda" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(72): warning #177-D: parameter "beta" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(72): warning #177-D: parameter "B" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(72): warning #177-D: parameter "ldb" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(73): warning #177-D: parameter "C" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(73): warning #177-D: parameter "ldc" was declared but never referenced

nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(56): warning #177-D: parameter "handle" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(56): warning #177-D: parameter "transa" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(57): warning #177-D: parameter "transb" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(57): warning #177-D: parameter "m" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(57): warning #177-D: parameter "n" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(58): warning #177-D: parameter "alpha" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(58): warning #177-D: parameter "A" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(58): warning #177-D: parameter "lda" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(59): warning #177-D: parameter "beta" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(59): warning #177-D: parameter "B" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(59): warning #177-D: parameter "ldb" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(60): warning #177-D: parameter "C" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(60): warning #177-D: parameter "ldc" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(69): warning #177-D: parameter "handle" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(69): warning #177-D: parameter "transa" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(70): warning #177-D: parameter "transb" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(70): warning #177-D: parameter "m" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(70): warning #177-D: parameter "n" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(71): warning #177-D: parameter "alpha" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(71): warning #177-D: parameter "A" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(71): warning #177-D: parameter "lda" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(72): warning #177-D: parameter "beta" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(72): warning #177-D: parameter "B" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(72): warning #177-D: parameter "ldb" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(73): warning #177-D: parameter "C" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(73): warning #177-D: parameter "ldc" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(203): warning #177-D: variable "arr_len" was declared but never referenced
          detected during instantiation of "void dgl::aten::SpMMCsr<XPU,IdType,DType>(const std::__cxx11::string &, const std::__cxx11::string &, const dgl::BcastOff &, const dgl::aten::CSRMatrix &, dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray, std::vector<dgl::IdArray, std::allocator<dgl::IdArray>>) [with XPU=2, IdType=int32_t, DType=__half]" 
/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/spmm.cu(114): here

nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
/root/jenkins/workspace/dgl_PR-4648@3/src/graph/transform/cuda/cuda_to_block.cu(123): warning #177-D: variable "num_ntypes" was declared but never referenced
          detected during instantiation of "std::tuple<dgl::HeteroGraphPtr, std::vector<dgl::IdArray, std::allocator<dgl::IdArray>>> dgl::transform::<unnamed>::ToBlockGPU<IdType>(dgl::HeteroGraphPtr, const std::vector<dgl::IdArray, std::allocator<dgl::IdArray>> &, __nv_bool, std::vector<dgl::IdArray, std::allocator<dgl::IdArray>> *) [with IdType=int32_t]" 
(392): here

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(203): warning #177-D: variable "arr_len" was declared but never referenced
          detected during instantiation of "void dgl::aten::SpMMCsrHetero<XPU,IdType,DType>(const std::__cxx11::string &, const std::__cxx11::string &, const dgl::BcastOff &, const std::vector<dgl::aten::CSRMatrix, std::allocator<dgl::aten::CSRMatrix>> &, const std::vector<dgl::IdArray, std::allocator<dgl::IdArray>> &, const std::vector<dgl::IdArray, std::allocator<dgl::IdArray>> &, std::vector<dgl::IdArray, std::allocator<dgl::IdArray>> *, std::vector<std::vector<dgl::IdArray, std::allocator<dgl::IdArray>>, std::allocator<std::vector<dgl::IdArray, std::allocator<dgl::IdArray>>>> *, const std::vector<uint64_t, std::allocator<std::_Bit_type>> &, const std::vector<uint64_t, std::allocator<std::_Bit_type>> &) [with XPU=2, IdType=int32_t, DType=__half]" 
/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/spmm_hetero.cu(182): here

nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
/root/jenkins/workspace/dgl_PR-4648@3/include/dgl/random.h(28): warning #177-D: variable "guard" was declared but never referenced

nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
nvcc warning : The 'compute_35', 'compute_37', 'compute_50', 'sm_35', 'sm_37' and 'sm_50' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
/root/jenkins/workspace/dgl_PR-4648@3/include/dgl/random.h(28): warning #177-D: variable "guard" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/include/dgl/random.h(28): warning #177-D: variable "guard" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/graph/sampling/randomwalks/frequency_hashmap.cu(105): warning #177-D: parameter "dst_data" was declared but never referenced
          detected during instantiation of "std::tuple<dgl::runtime::NDArray, dgl::IdArray, dgl::IdArray> dgl::sampling::impl::FrequencyHashmap<IdxType>::Topk(const IdxType *, const IdxType *, DGLDataType, int64_t, int64_t, int64_t) [with IdxType=int64_t]" 
(435): here

/root/jenkins/workspace/dgl_PR-4648@3/include/dgl/random.h(28): warning #177-D: variable "guard" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/include/dgl/random.h(28): warning #177-D: variable "guard" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/include/dgl/random.h(28): warning #177-D: variable "guard" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/graph/transform/cuda/cuda_to_block.cu(123): warning #177-D: variable "num_ntypes" was declared but never referenced
          detected during instantiation of "std::tuple<dgl::HeteroGraphPtr, std::vector<dgl::IdArray, std::allocator<dgl::IdArray>>> dgl::transform::<unnamed>::ToBlockGPU<IdType>(dgl::HeteroGraphPtr, const std::vector<dgl::IdArray, std::allocator<dgl::IdArray>> &, __nv_bool, std::vector<dgl::IdArray, std::allocator<dgl::IdArray>> *) [with IdType=int32_t]" 
(392): here

/root/jenkins/workspace/dgl_PR-4648@3/include/dgl/random.h(28): warning #177-D: variable "guard" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/include/dgl/random.h(28): warning #177-D: variable "guard" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/graph/transform/cuda/cuda_to_block.cu(123): warning #177-D: variable "num_ntypes" was declared but never referenced
          detected during instantiation of "std::tuple<dgl::HeteroGraphPtr, std::vector<dgl::IdArray, std::allocator<dgl::IdArray>>> dgl::transform::<unnamed>::ToBlockGPU<IdType>(dgl::HeteroGraphPtr, const std::vector<dgl::IdArray, std::allocator<dgl::IdArray>> &, __nv_bool, std::vector<dgl::IdArray, std::allocator<dgl::IdArray>> *) [with IdType=int32_t]" 
(392): here

/root/jenkins/workspace/dgl_PR-4648@3/include/dgl/random.h(28): warning #177-D: variable "guard" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/include/dgl/random.h(28): warning #177-D: variable "guard" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/graph/transform/cuda/cuda_to_block.cu(123): warning #177-D: variable "num_ntypes" was declared but never referenced
          detected during instantiation of "std::tuple<dgl::HeteroGraphPtr, std::vector<dgl::IdArray, std::allocator<dgl::IdArray>>> dgl::transform::<unnamed>::ToBlockGPU<IdType>(dgl::HeteroGraphPtr, const std::vector<dgl::IdArray, std::allocator<dgl::IdArray>> &, __nv_bool, std::vector<dgl::IdArray, std::allocator<dgl::IdArray>> *) [with IdType=int32_t]" 
(392): here

/root/jenkins/workspace/dgl_PR-4648@3/include/dgl/random.h(28): warning #177-D: variable "guard" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/include/dgl/random.h(28): warning #177-D: variable "guard" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/graph/transform/cuda/cuda_to_block.cu(123): warning #177-D: variable "num_ntypes" was declared but never referenced
          detected during instantiation of "std::tuple<dgl::HeteroGraphPtr, std::vector<dgl::IdArray, std::allocator<dgl::IdArray>>> dgl::transform::<unnamed>::ToBlockGPU<IdType>(dgl::HeteroGraphPtr, const std::vector<dgl::IdArray, std::allocator<dgl::IdArray>> &, __nv_bool, std::vector<dgl::IdArray, std::allocator<dgl::IdArray>> *) [with IdType=int32_t]" 
(392): here

/root/jenkins/workspace/dgl_PR-4648@3/include/dgl/random.h(28): warning #177-D: variable "guard" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/include/dgl/random.h(28): warning #177-D: variable "guard" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/include/dgl/random.h(28): warning #177-D: variable "guard" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/include/dgl/random.h(28): warning #177-D: variable "guard" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(56): warning #177-D: parameter "handle" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(56): warning #177-D: parameter "transa" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(57): warning #177-D: parameter "transb" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(57): warning #177-D: parameter "m" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(57): warning #177-D: parameter "n" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(58): warning #177-D: parameter "alpha" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(58): warning #177-D: parameter "A" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(58): warning #177-D: parameter "lda" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(59): warning #177-D: parameter "beta" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(59): warning #177-D: parameter "B" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(59): warning #177-D: parameter "ldb" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(60): warning #177-D: parameter "C" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(60): warning #177-D: parameter "ldc" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(69): warning #177-D: parameter "handle" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(69): warning #177-D: parameter "transa" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(70): warning #177-D: parameter "transb" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(70): warning #177-D: parameter "m" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(70): warning #177-D: parameter "n" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(71): warning #177-D: parameter "alpha" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(71): warning #177-D: parameter "A" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(71): warning #177-D: parameter "lda" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(72): warning #177-D: parameter "beta" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(72): warning #177-D: parameter "B" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(72): warning #177-D: parameter "ldb" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(73): warning #177-D: parameter "C" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(73): warning #177-D: parameter "ldc" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/include/dgl/random.h(28): warning #177-D: variable "guard" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(203): warning #177-D: variable "arr_len" was declared but never referenced
          detected during instantiation of "void dgl::aten::SpMMCsrHetero<XPU,IdType,DType>(const std::__cxx11::string &, const std::__cxx11::string &, const dgl::BcastOff &, const std::vector<dgl::aten::CSRMatrix, std::allocator<dgl::aten::CSRMatrix>> &, const std::vector<dgl::IdArray, std::allocator<dgl::IdArray>> &, const std::vector<dgl::IdArray, std::allocator<dgl::IdArray>> &, std::vector<dgl::IdArray, std::allocator<dgl::IdArray>> *, std::vector<std::vector<dgl::IdArray, std::allocator<dgl::IdArray>>, std::allocator<std::vector<dgl::IdArray, std::allocator<dgl::IdArray>>>> *, const std::vector<uint64_t, std::allocator<std::_Bit_type>> &, const std::vector<uint64_t, std::allocator<std::_Bit_type>> &) [with XPU=2, IdType=int32_t, DType=__half]" 
/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/spmm_hetero.cu(182): here

/root/jenkins/workspace/dgl_PR-4648@3/src/graph/sampling/randomwalks/frequency_hashmap.cu(105): warning #177-D: parameter "dst_data" was declared but never referenced
          detected during instantiation of "std::tuple<dgl::runtime::NDArray, dgl::IdArray, dgl::IdArray> dgl::sampling::impl::FrequencyHashmap<IdxType>::Topk(const IdxType *, const IdxType *, DGLDataType, int64_t, int64_t, int64_t) [with IdxType=int64_t]" 
(435): here

/root/jenkins/workspace/dgl_PR-4648@3/include/dgl/random.h(28): warning #177-D: variable "guard" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/include/dgl/random.h(28): warning #177-D: variable "guard" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/include/dgl/random.h(28): warning #177-D: variable "guard" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/include/dgl/random.h(28): warning #177-D: variable "guard" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/include/dgl/random.h(28): warning #177-D: variable "guard" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/include/dgl/random.h(28): warning #177-D: variable "guard" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/graph/sampling/randomwalks/frequency_hashmap.cu(105): warning #177-D: parameter "dst_data" was declared but never referenced
          detected during instantiation of "std::tuple<dgl::runtime::NDArray, dgl::IdArray, dgl::IdArray> dgl::sampling::impl::FrequencyHashmap<IdxType>::Topk(const IdxType *, const IdxType *, DGLDataType, int64_t, int64_t, int64_t) [with IdxType=int64_t]" 
(435): here

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(56): warning #177-D: parameter "handle" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(56): warning #177-D: parameter "transa" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(57): warning #177-D: parameter "transb" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(57): warning #177-D: parameter "m" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(57): warning #177-D: parameter "n" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(58): warning #177-D: parameter "alpha" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(58): warning #177-D: parameter "A" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(58): warning #177-D: parameter "lda" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(59): warning #177-D: parameter "beta" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(59): warning #177-D: parameter "B" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(59): warning #177-D: parameter "ldb" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(60): warning #177-D: parameter "C" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(60): warning #177-D: parameter "ldc" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(69): warning #177-D: parameter "handle" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(69): warning #177-D: parameter "transa" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(70): warning #177-D: parameter "transb" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(70): warning #177-D: parameter "m" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(70): warning #177-D: parameter "n" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(71): warning #177-D: parameter "alpha" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(71): warning #177-D: parameter "A" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(71): warning #177-D: parameter "lda" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(72): warning #177-D: parameter "beta" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(72): warning #177-D: parameter "B" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(72): warning #177-D: parameter "ldb" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(73): warning #177-D: parameter "C" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(73): warning #177-D: parameter "ldc" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(203): warning #177-D: variable "arr_len" was declared but never referenced
          detected during instantiation of "void dgl::aten::SpMMCsrHetero<XPU,IdType,DType>(const std::__cxx11::string &, const std::__cxx11::string &, const dgl::BcastOff &, const std::vector<dgl::aten::CSRMatrix, std::allocator<dgl::aten::CSRMatrix>> &, const std::vector<dgl::IdArray, std::allocator<dgl::IdArray>> &, const std::vector<dgl::IdArray, std::allocator<dgl::IdArray>> &, std::vector<dgl::IdArray, std::allocator<dgl::IdArray>> *, std::vector<std::vector<dgl::IdArray, std::allocator<dgl::IdArray>>, std::allocator<std::vector<dgl::IdArray, std::allocator<dgl::IdArray>>>> *, const std::vector<uint64_t, std::allocator<std::_Bit_type>> &, const std::vector<uint64_t, std::allocator<std::_Bit_type>> &) [with XPU=2, IdType=int32_t, DType=__half]" 
/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/spmm_hetero.cu(182): here

/root/jenkins/workspace/dgl_PR-4648@3/include/dgl/random.h(28): warning #177-D: variable "guard" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/include/dgl/random.h(28): warning #177-D: variable "guard" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(56): warning #177-D: parameter "handle" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(56): warning #177-D: parameter "transa" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(57): warning #177-D: parameter "transb" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(57): warning #177-D: parameter "m" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(57): warning #177-D: parameter "n" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(58): warning #177-D: parameter "alpha" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(58): warning #177-D: parameter "A" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(58): warning #177-D: parameter "lda" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(59): warning #177-D: parameter "beta" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(59): warning #177-D: parameter "B" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(59): warning #177-D: parameter "ldb" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(60): warning #177-D: parameter "C" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(60): warning #177-D: parameter "ldc" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(69): warning #177-D: parameter "handle" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(69): warning #177-D: parameter "transa" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(70): warning #177-D: parameter "transb" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(70): warning #177-D: parameter "m" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(70): warning #177-D: parameter "n" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(71): warning #177-D: parameter "alpha" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(71): warning #177-D: parameter "A" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(71): warning #177-D: parameter "lda" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(72): warning #177-D: parameter "beta" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(72): warning #177-D: parameter "B" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(72): warning #177-D: parameter "ldb" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(73): warning #177-D: parameter "C" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(73): warning #177-D: parameter "ldc" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(203): warning #177-D: variable "arr_len" was declared but never referenced
          detected during instantiation of "void dgl::aten::SpMMCsr<XPU,IdType,DType>(const std::__cxx11::string &, const std::__cxx11::string &, const dgl::BcastOff &, const dgl::aten::CSRMatrix &, dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray, std::vector<dgl::IdArray, std::allocator<dgl::IdArray>>) [with XPU=2, IdType=int32_t, DType=__half]" 
/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/spmm.cu(114): here

/root/jenkins/workspace/dgl_PR-4648@3/include/dgl/random.h(28): warning #177-D: variable "guard" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/graph/sampling/randomwalks/frequency_hashmap.cu(105): warning #177-D: parameter "dst_data" was declared but never referenced
          detected during instantiation of "std::tuple<dgl::runtime::NDArray, dgl::IdArray, dgl::IdArray> dgl::sampling::impl::FrequencyHashmap<IdxType>::Topk(const IdxType *, const IdxType *, DGLDataType, int64_t, int64_t, int64_t) [with IdxType=int64_t]" 
(435): here

/root/jenkins/workspace/dgl_PR-4648@3/include/dgl/random.h(28): warning #177-D: variable "guard" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(56): warning #177-D: parameter "handle" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(56): warning #177-D: parameter "transa" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(57): warning #177-D: parameter "transb" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(57): warning #177-D: parameter "m" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(57): warning #177-D: parameter "n" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(58): warning #177-D: parameter "alpha" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(58): warning #177-D: parameter "A" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(58): warning #177-D: parameter "lda" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(59): warning #177-D: parameter "beta" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(59): warning #177-D: parameter "B" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(59): warning #177-D: parameter "ldb" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(60): warning #177-D: parameter "C" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(60): warning #177-D: parameter "ldc" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(69): warning #177-D: parameter "handle" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(69): warning #177-D: parameter "transa" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(70): warning #177-D: parameter "transb" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(70): warning #177-D: parameter "m" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(70): warning #177-D: parameter "n" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(71): warning #177-D: parameter "alpha" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(71): warning #177-D: parameter "A" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(71): warning #177-D: parameter "lda" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(72): warning #177-D: parameter "beta" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(72): warning #177-D: parameter "B" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(72): warning #177-D: parameter "ldb" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(73): warning #177-D: parameter "C" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(73): warning #177-D: parameter "ldc" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(203): warning #177-D: variable "arr_len" was declared but never referenced
          detected during instantiation of "void dgl::aten::SpMMCsrHetero<XPU,IdType,DType>(const std::__cxx11::string &, const std::__cxx11::string &, const dgl::BcastOff &, const std::vector<dgl::aten::CSRMatrix, std::allocator<dgl::aten::CSRMatrix>> &, const std::vector<dgl::IdArray, std::allocator<dgl::IdArray>> &, const std::vector<dgl::IdArray, std::allocator<dgl::IdArray>> &, std::vector<dgl::IdArray, std::allocator<dgl::IdArray>> *, std::vector<std::vector<dgl::IdArray, std::allocator<dgl::IdArray>>, std::allocator<std::vector<dgl::IdArray, std::allocator<dgl::IdArray>>>> *, const std::vector<uint64_t, std::allocator<std::_Bit_type>> &, const std::vector<uint64_t, std::allocator<std::_Bit_type>> &) [with XPU=2, IdType=int32_t, DType=__half]" 
/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/spmm_hetero.cu(182): here

/root/jenkins/workspace/dgl_PR-4648@3/include/dgl/random.h(28): warning #177-D: variable "guard" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/graph/sampling/randomwalks/frequency_hashmap.cu(105): warning #177-D: parameter "dst_data" was declared but never referenced
          detected during instantiation of "std::tuple<dgl::runtime::NDArray, dgl::IdArray, dgl::IdArray> dgl::sampling::impl::FrequencyHashmap<IdxType>::Topk(const IdxType *, const IdxType *, DGLDataType, int64_t, int64_t, int64_t) [with IdxType=int64_t]" 
(435): here

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(56): warning #177-D: parameter "handle" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(56): warning #177-D: parameter "transa" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(57): warning #177-D: parameter "transb" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(57): warning #177-D: parameter "m" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(57): warning #177-D: parameter "n" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(58): warning #177-D: parameter "alpha" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(58): warning #177-D: parameter "A" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(58): warning #177-D: parameter "lda" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(59): warning #177-D: parameter "beta" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(59): warning #177-D: parameter "B" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(59): warning #177-D: parameter "ldb" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(60): warning #177-D: parameter "C" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(60): warning #177-D: parameter "ldc" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(69): warning #177-D: parameter "handle" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(69): warning #177-D: parameter "transa" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(70): warning #177-D: parameter "transb" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(70): warning #177-D: parameter "m" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(70): warning #177-D: parameter "n" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(71): warning #177-D: parameter "alpha" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(71): warning #177-D: parameter "A" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(71): warning #177-D: parameter "lda" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(72): warning #177-D: parameter "beta" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(72): warning #177-D: parameter "B" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(72): warning #177-D: parameter "ldb" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(73): warning #177-D: parameter "C" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(73): warning #177-D: parameter "ldc" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(203): warning #177-D: variable "arr_len" was declared but never referenced
          detected during instantiation of "void dgl::aten::SpMMCsrHetero<XPU,IdType,DType>(const std::__cxx11::string &, const std::__cxx11::string &, const dgl::BcastOff &, const std::vector<dgl::aten::CSRMatrix, std::allocator<dgl::aten::CSRMatrix>> &, const std::vector<dgl::IdArray, std::allocator<dgl::IdArray>> &, const std::vector<dgl::IdArray, std::allocator<dgl::IdArray>> &, std::vector<dgl::IdArray, std::allocator<dgl::IdArray>> *, std::vector<std::vector<dgl::IdArray, std::allocator<dgl::IdArray>>, std::allocator<std::vector<dgl::IdArray, std::allocator<dgl::IdArray>>>> *, const std::vector<uint64_t, std::allocator<std::_Bit_type>> &, const std::vector<uint64_t, std::allocator<std::_Bit_type>> &) [with XPU=2, IdType=int32_t, DType=__half]" 
/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/spmm_hetero.cu(182): here

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(56): warning #177-D: parameter "handle" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(56): warning #177-D: parameter "transa" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(57): warning #177-D: parameter "transb" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(57): warning #177-D: parameter "m" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(57): warning #177-D: parameter "n" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(58): warning #177-D: parameter "alpha" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(58): warning #177-D: parameter "A" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(58): warning #177-D: parameter "lda" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(59): warning #177-D: parameter "beta" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(59): warning #177-D: parameter "B" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(59): warning #177-D: parameter "ldb" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(60): warning #177-D: parameter "C" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(60): warning #177-D: parameter "ldc" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(69): warning #177-D: parameter "handle" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(69): warning #177-D: parameter "transa" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(70): warning #177-D: parameter "transb" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(70): warning #177-D: parameter "m" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(70): warning #177-D: parameter "n" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(71): warning #177-D: parameter "alpha" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(71): warning #177-D: parameter "A" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(71): warning #177-D: parameter "lda" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(72): warning #177-D: parameter "beta" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(72): warning #177-D: parameter "B" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(72): warning #177-D: parameter "ldb" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(73): warning #177-D: parameter "C" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(73): warning #177-D: parameter "ldc" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(203): warning #177-D: variable "arr_len" was declared but never referenced
          detected during instantiation of "void dgl::aten::SpMMCsr<XPU,IdType,DType>(const std::__cxx11::string &, const std::__cxx11::string &, const dgl::BcastOff &, const dgl::aten::CSRMatrix &, dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray, std::vector<dgl::IdArray, std::allocator<dgl::IdArray>>) [with XPU=2, IdType=int32_t, DType=__half]" 
/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/spmm.cu(114): here

/root/jenkins/workspace/dgl_PR-4648@3/include/dgl/random.h(28): warning #177-D: variable "guard" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/include/dgl/random.h(28): warning #177-D: variable "guard" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(56): warning #177-D: parameter "handle" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(56): warning #177-D: parameter "transa" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(57): warning #177-D: parameter "transb" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(57): warning #177-D: parameter "m" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(57): warning #177-D: parameter "n" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(58): warning #177-D: parameter "alpha" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(58): warning #177-D: parameter "A" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(58): warning #177-D: parameter "lda" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(59): warning #177-D: parameter "beta" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(59): warning #177-D: parameter "B" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(59): warning #177-D: parameter "ldb" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(60): warning #177-D: parameter "C" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(60): warning #177-D: parameter "ldc" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(69): warning #177-D: parameter "handle" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(69): warning #177-D: parameter "transa" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(70): warning #177-D: parameter "transb" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(70): warning #177-D: parameter "m" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(70): warning #177-D: parameter "n" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(71): warning #177-D: parameter "alpha" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(71): warning #177-D: parameter "A" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(71): warning #177-D: parameter "lda" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(72): warning #177-D: parameter "beta" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(72): warning #177-D: parameter "B" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(72): warning #177-D: parameter "ldb" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(73): warning #177-D: parameter "C" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(73): warning #177-D: parameter "ldc" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(203): warning #177-D: variable "arr_len" was declared but never referenced
          detected during instantiation of "void dgl::aten::SpMMCsr<XPU,IdType,DType>(const std::__cxx11::string &, const std::__cxx11::string &, const dgl::BcastOff &, const dgl::aten::CSRMatrix &, dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray, std::vector<dgl::IdArray, std::allocator<dgl::IdArray>>) [with XPU=2, IdType=int32_t, DType=__half]" 
/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/spmm.cu(114): here

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(56): warning #177-D: parameter "handle" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(56): warning #177-D: parameter "transa" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(57): warning #177-D: parameter "transb" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(57): warning #177-D: parameter "m" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(57): warning #177-D: parameter "n" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(58): warning #177-D: parameter "alpha" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(58): warning #177-D: parameter "A" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(58): warning #177-D: parameter "lda" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(59): warning #177-D: parameter "beta" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(59): warning #177-D: parameter "B" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(59): warning #177-D: parameter "ldb" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(60): warning #177-D: parameter "C" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(60): warning #177-D: parameter "ldc" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(69): warning #177-D: parameter "handle" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(69): warning #177-D: parameter "transa" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(70): warning #177-D: parameter "transb" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(70): warning #177-D: parameter "m" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(70): warning #177-D: parameter "n" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(71): warning #177-D: parameter "alpha" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(71): warning #177-D: parameter "A" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(71): warning #177-D: parameter "lda" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(72): warning #177-D: parameter "beta" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(72): warning #177-D: parameter "B" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(72): warning #177-D: parameter "ldb" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(73): warning #177-D: parameter "C" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(73): warning #177-D: parameter "ldc" was declared but never referenced

/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/./spmm.cuh(203): warning #177-D: variable "arr_len" was declared but never referenced
          detected during instantiation of "void dgl::aten::SpMMCsr<XPU,IdType,DType>(const std::__cxx11::string &, const std::__cxx11::string &, const dgl::BcastOff &, const dgl::aten::CSRMatrix &, dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray, std::vector<dgl::IdArray, std::allocator<dgl::IdArray>>) [with XPU=2, IdType=int32_t, DType=__half]" 
/root/jenkins/workspace/dgl_PR-4648@3/src/array/cuda/spmm.cu(114): here

Scanning dependencies of target dgl
[ 60%] Building CXX object CMakeFiles/dgl.dir/src/array/array.cc.o
[ 60%] Building CXX object CMakeFiles/dgl.dir/src/array/array_arith.cc.o
[ 61%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/array_cumsum.cc.o
[ 61%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/array_sort.cc.o
[ 61%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/array_index_select.cc.o
[ 61%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/array_nonzero.cc.o
[ 61%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/array_pack.cc.o
[ 61%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/array_op_impl.cc.o
[ 61%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/array_scatter.cc.o
[ 61%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/csr_remove.cc.o
[ 61%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/csr_sort.cc.o
[ 62%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/array_repeat.cc.o
[ 62%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/coo_coalesce.cc.o
[ 63%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/csr_union.cc.o
[ 63%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/coo_remove.cc.o
[ 65%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/csr_get_data.cc.o
[ 65%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/coo_linegraph.cc.o
[ 65%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/coo_sort.cc.o
[ 66%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/csr_mm.cc.o
[ 66%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/negative_sampling.cc.o
[ 66%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/csr_sum.cc.o
[ 66%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/gather_mm.cc.o
[ 66%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/csr_to_simple.cc.o
[ 67%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/spmat_op_impl_coo.cc.o
[ 67%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/rowwise_topk.cc.o
[ 67%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/disjoint_union.cc.o
[ 68%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/rowwise_sampling.cc.o
[ 68%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/segment_reduce.cc.o
[ 69%] Building CXX object CMakeFiles/dgl.dir/src/array/kernel.cc.o
[ 69%] Building CXX object CMakeFiles/dgl.dir/src/array/filter.cc.o
[ 69%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/spmat_op_impl_csr.cc.o
[ 69%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/spmm.cc.o
[ 69%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/sddmm.cc.o
[ 69%] Building CXX object CMakeFiles/dgl.dir/src/random/cpu/choice.cc.o
[ 69%] Building CXX object CMakeFiles/dgl.dir/src/geometry/geometry.cc.o
[ 71%] Building CXX object CMakeFiles/dgl.dir/src/partition/ndarray_partition.cc.o
[ 71%] Building CXX object CMakeFiles/dgl.dir/src/array/libra_partition.cc.o
[ 71%] Building CXX object CMakeFiles/dgl.dir/src/array/uvm_array.cc.o
[ 71%] Building CXX object CMakeFiles/dgl.dir/src/random/random.cc.o
[ 71%] Building CXX object CMakeFiles/dgl.dir/src/runtime/module_util.cc.o
[ 71%] Building CXX object CMakeFiles/dgl.dir/src/geometry/cpu/geometry_op_impl.cc.o
[ 72%] Building CXX object CMakeFiles/dgl.dir/src/runtime/object.cc.o
[ 72%] Building CXX object CMakeFiles/dgl.dir/src/runtime/module.cc.o
[ 73%] Building CXX object CMakeFiles/dgl.dir/src/bcast.cc.o
[ 73%] Building CXX object CMakeFiles/dgl.dir/src/array/union_partition.cc.o
[ 73%] Building CXX object CMakeFiles/dgl.dir/src/runtime/c_object_api.cc.o
[ 73%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/traversal.cc.o
[ 73%] Building CXX object CMakeFiles/dgl.dir/src/graph/gk_ops.cc.o
[ 73%] Building CXX object CMakeFiles/dgl.dir/src/runtime/dlpack_convert.cc.o
[ 73%] Building CXX object CMakeFiles/dgl.dir/src/runtime/config.cc.o
[ 73%] Building CXX object CMakeFiles/dgl.dir/src/runtime/dso_module.cc.o
[ 73%] Building CXX object CMakeFiles/dgl.dir/src/runtime/ndarray.cc.o
[ 73%] Building CXX object CMakeFiles/dgl.dir/src/runtime/cpu_device_api.cc.o
[ 73%] Building CXX object CMakeFiles/dgl.dir/src/runtime/thread_pool.cc.o
[ 73%] Building CXX object CMakeFiles/dgl.dir/src/c_api_common.cc.o
[ 74%] Building CXX object CMakeFiles/dgl.dir/src/runtime/c_runtime_api.cc.o
[ 74%] Building CXX object CMakeFiles/dgl.dir/src/graph/graph.cc.o
[ 75%] Building CXX object CMakeFiles/dgl.dir/src/runtime/file_util.cc.o
[ 75%] Building CXX object CMakeFiles/dgl.dir/src/runtime/semaphore_wrapper.cc.o
[ 78%] Building CXX object CMakeFiles/dgl.dir/src/runtime/shared_mem.cc.o
[ 78%] Building CXX object CMakeFiles/dgl.dir/src/runtime/utils.cc.o
[ 78%] Building CXX object CMakeFiles/dgl.dir/src/runtime/threading_backend.cc.o
[ 78%] Building CXX object CMakeFiles/dgl.dir/src/runtime/registry.cc.o
[ 78%] Building CXX object CMakeFiles/dgl.dir/src/api/api_test.cc.o
[ 79%] Building CXX object CMakeFiles/dgl.dir/src/runtime/resource_manager.cc.o
[ 79%] Building CXX object CMakeFiles/dgl.dir/src/graph/creators.cc.o
[ 79%] Building CXX object CMakeFiles/dgl.dir/src/runtime/system_lib_module.cc.o
[ 79%] Building CXX object CMakeFiles/dgl.dir/src/graph/network.cc.o
[ 79%] Building CXX object CMakeFiles/dgl.dir/src/graph/graph_apis.cc.o
[ 79%] Building CXX object CMakeFiles/dgl.dir/src/graph/sampling/randomwalks/get_node_types_cpu.cc.o
[ 79%] Building CXX object CMakeFiles/dgl.dir/src/graph/heterograph.cc.o
[ 80%] Building CXX object CMakeFiles/dgl.dir/src/graph/sampler.cc.o
[ 80%] Building CXX object CMakeFiles/dgl.dir/src/runtime/tensordispatch.cc.o
[ 81%] Building CXX object CMakeFiles/dgl.dir/src/graph/sampling/randomwalks/node2vec.cc.o
[ 81%] Building CXX object CMakeFiles/dgl.dir/src/graph/heterograph_capi.cc.o
[ 81%] Building CXX object CMakeFiles/dgl.dir/src/runtime/workspace_pool.cc.o
[ 83%] Building CXX object CMakeFiles/dgl.dir/src/graph/nodeflow.cc.o
[ 83%] Building CXX object CMakeFiles/dgl.dir/src/graph/metis_partition.cc.o
[ 83%] Building CXX object CMakeFiles/dgl.dir/src/graph/graph_traversal.cc.o
[ 83%] Building CXX object CMakeFiles/dgl.dir/src/api/api_container.cc.o
[ 84%] Building CXX object CMakeFiles/dgl.dir/src/graph/graph_op.cc.o
[ 84%] Building CXX object CMakeFiles/dgl.dir/src/graph/immutable_graph.cc.o
[ 84%] Building CXX object CMakeFiles/dgl.dir/src/graph/sampling/negative/global_uniform.cc.o
[ 84%] Building CXX object CMakeFiles/dgl.dir/src/graph/pickle.cc.o
[ 84%] Building CXX object CMakeFiles/dgl.dir/src/graph/sampling/randomwalks/node2vec_cpu.cc.o
[ 84%] Building CXX object CMakeFiles/dgl.dir/src/graph/sampling/neighbor/neighbor.cc.o
[ 85%] Building CXX object CMakeFiles/dgl.dir/src/graph/transform/cpu/knn.cc.o
[ 85%] Building CXX object CMakeFiles/dgl.dir/src/graph/serialize/heterograph_serialize.cc.o
[ 85%] Building CXX object CMakeFiles/dgl.dir/src/graph/sampling/randomwalks/randomwalk_with_restart_cpu.cc.o
[ 85%] Building CXX object CMakeFiles/dgl.dir/src/graph/sampling/randomwalks/randomwalk_cpu.cc.o
[ 85%] Building CXX object CMakeFiles/dgl.dir/src/graph/serialize/tensor_serialize.cc.o
[ 86%] Building CXX object CMakeFiles/dgl.dir/src/graph/serialize/zerocopy_serializer.cc.o
[ 86%] Building CXX object CMakeFiles/dgl.dir/src/graph/transform/line_graph.cc.o
[ 86%] Building CXX object CMakeFiles/dgl.dir/src/graph/transform/compact.cc.o
[ 87%] Building CXX object CMakeFiles/dgl.dir/src/graph/sampling/randomwalks/randomwalks.cc.o
[ 87%] Building CXX object CMakeFiles/dgl.dir/src/graph/serialize/dglgraph_serialize.cc.o
[ 87%] Building CXX object CMakeFiles/dgl.dir/src/graph/serialize/graph_serialize.cc.o
[ 87%] Building CXX object CMakeFiles/dgl.dir/src/graph/transform/knn.cc.o
[ 87%] Building CXX object CMakeFiles/dgl.dir/src/graph/subgraph.cc.o
[ 87%] Building CXX object CMakeFiles/dgl.dir/src/graph/transform/remove_edges.cc.o
[ 87%] Building CXX object CMakeFiles/dgl.dir/src/graph/shared_mem_manager.cc.o
[ 87%] Building CXX object CMakeFiles/dgl.dir/src/graph/transform/to_simple.cc.o
[ 87%] Building CXX object CMakeFiles/dgl.dir/src/graph/transform/metis_partition_hetero.cc.o
[ 87%] Building CXX object CMakeFiles/dgl.dir/src/graph/transform/union_partition.cc.o
[ 87%] Building CXX object CMakeFiles/dgl.dir/src/rpc/network/socket_pool.cc.o
[ 87%] Building CXX object CMakeFiles/dgl.dir/src/rpc/network/msg_queue.cc.o
[ 89%] Building CXX object CMakeFiles/dgl.dir/src/graph/transform/partition_hetero.cc.o
[ 89%] Building CXX object CMakeFiles/dgl.dir/src/rpc/network/socket_communicator.cc.o
[ 89%] Building CXX object CMakeFiles/dgl.dir/src/graph/transform/to_bipartite.cc.o
[ 89%] Building CXX object CMakeFiles/dgl.dir/src/graph/unit_graph.cc.o
[ 89%] Building CXX object CMakeFiles/dgl.dir/src/array/cuda/csr_transpose.cc.o
[ 89%] Building CXX object CMakeFiles/dgl.dir/src/scheduler/scheduler_apis.cc.o
[ 90%] Building CXX object CMakeFiles/dgl.dir/src/graph/traversal.cc.o
[ 91%] Building CXX object CMakeFiles/dgl.dir/src/rpc/network/common.cc.o
[ 91%] Building CXX object CMakeFiles/dgl.dir/src/rpc/rpc.cc.o
[ 91%] Building CXX object CMakeFiles/dgl.dir/src/scheduler/scheduler.cc.o
[ 91%] Building CXX object CMakeFiles/dgl.dir/src/runtime/cuda/cuda_device_api.cc.o
[ 92%] Building CXX object CMakeFiles/dgl.dir/src/rpc/network/tcp_socket.cc.o
[ 92%] Building CXX object CMakeFiles/dgl.dir/src/rpc/tensorpipe/tp_communicator.cc.o
In file included from /root/jenkins/workspace/dgl_PR-4648@2/src/geometry/../c_api_common.h:10:0,
                 from /root/jenkins/workspace/dgl_PR-4648@2/src/geometry/geometry.cc:9:
/root/jenkins/workspace/dgl_PR-4648@2/include/dgl/runtime/packed_func.h:502:21: warning: inline function 'TObjectRef dgl::runtime::DGLArgValue::AsObjectRef() const [with TObjectRef = dgl::HeteroGraphRef]' used but never defined
   inline TObjectRef AsObjectRef() const;
                     ^~~~~~~~~~~
/root/jenkins/workspace/dgl_PR-4648@2/src/partition/ndarray_partition.cc: In member function 'virtual int64_t dgl::partition::RangePartition::PartSize(int) const':
/root/jenkins/workspace/dgl_PR-4648@2/src/partition/ndarray_partition.cc:202:3: warning: control reaches end of non-void function [-Wreturn-type]
   }
   ^
In file included from /root/jenkins/workspace/dgl_PR-4648@2/third_party/dmlc-core/include/dmlc/logging.h:132:0,
                 from /root/jenkins/workspace/dgl_PR-4648@2/include/dgl/./runtime/object.h:9,
                 from /root/jenkins/workspace/dgl_PR-4648@2/include/dgl/graph_interface.h:15,
                 from /root/jenkins/workspace/dgl_PR-4648@2/include/dgl/sampler.h:13,
                 from /root/jenkins/workspace/dgl_PR-4648@2/src/graph/sampler.cc:6:
/root/jenkins/workspace/dgl_PR-4648@2/src/graph/sampler.cc: In member function 'dgl::NegSubgraph dgl::{anonymous}::EdgeSamplerObject::genNegEdgeSubgraph(const dgl::Subgraph&, const string&, int64_t, bool, bool)':
/root/jenkins/workspace/dgl_PR-4648@2/src/graph/sampler.cc:1189:48: warning: comparison between signed and unsigned integer expressions [-Wsign-compare]
       assert(prev_neg_offset + neg_sample_size == neg_vids.size());
              ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^~~~~~~~~~~~~~~~~~
/root/jenkins/workspace/dgl_PR-4648@2/src/graph/sampler.cc:1193:48: warning: comparison between signed and unsigned integer expressions [-Wsign-compare]
       assert(prev_neg_offset + neg_sample_size == neg_vids.size());
              ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^~~~~~~~~~~~~~~~~~
/root/jenkins/workspace/dgl_PR-4648@2/src/array/libra_partition.cc: In function 'dgl::runtime::List<dgl::runtime::Value> dgl::aten::Libra2dglBuildDict(dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray, int32_t, int32_t, int64_t, const string&)':
/root/jenkins/workspace/dgl_PR-4648@2/src/array/libra_partition.cc:396:11: warning: ignoring return value of 'int fscanf(FILE*, const char*, ...)', declared with attribute warn_unused_result [-Wunused-result]
     fscanf(fp, "%ld,%ld,%f\n", &u, &v, &w);   // reading an edge - the src and dst global node IDs
     ~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
/root/jenkins/workspace/dgl_PR-4648@2/src/array/libra_partition.cc: In function 'int32_t dgl::aten::Ver2partition(IdType, int64_t*, int32_t) [with IdType = long int]':
/root/jenkins/workspace/dgl_PR-4648@2/src/array/libra_partition.cc:43:1: warning: control reaches end of non-void function [-Wreturn-type]
 }
 ^
[ 93%] Linking CXX shared library libdgl.so
[ 93%] Built target dgl
Scanning dependencies of target rpc_server
Scanning dependencies of target rpc_client
[ 95%] Building CXX object CMakeFiles/rpc_server.dir/tests/dist/cpp/rpc_server.cc.o
[ 95%] Building CXX object CMakeFiles/rpc_client.dir/tests/dist/cpp/rpc_client.cc.o
Scanning dependencies of target runUnitTests
[ 95%] Building CXX object CMakeFiles/runUnitTests.dir/tests/cpp/graph_index_test.cc.o
[ 95%] Building CXX object CMakeFiles/runUnitTests.dir/tests/cpp/message_queue_test.cc.o
[ 96%] Building CXX object CMakeFiles/runUnitTests.dir/tests/cpp/socket_communicator_test.cc.o
[ 96%] Building CXX object CMakeFiles/runUnitTests.dir/tests/cpp/string_test.cc.o
[ 96%] Building CXX object CMakeFiles/runUnitTests.dir/tests/cpp/test_aten.cc.o
[ 96%] Building CXX object CMakeFiles/runUnitTests.dir/tests/cpp/test_sampler.cc.o
[ 96%] Building CXX object CMakeFiles/runUnitTests.dir/tests/cpp/test_csrmm.cc.o
[ 97%] Building CXX object CMakeFiles/runUnitTests.dir/tests/cpp/test_serialize.cc.o
[ 97%] Building CXX object CMakeFiles/runUnitTests.dir/tests/cpp/test_partition.cc.o
[ 97%] Building CXX object CMakeFiles/runUnitTests.dir/tests/cpp/test_rowwise.cc.o
[ 97%] Building CXX object CMakeFiles/runUnitTests.dir/tests/cpp/test_spmm.cc.o
[ 97%] Building CXX object CMakeFiles/runUnitTests.dir/tests/cpp/test_spmat_csr.cc.o
[ 97%] Building CXX object CMakeFiles/runUnitTests.dir/tests/cpp/test_smart_ptr_serialize.cc.o
[ 98%] Building CXX object CMakeFiles/runUnitTests.dir/tests/cpp/test_spmat_coo.cc.o
[ 98%] Building CXX object CMakeFiles/runUnitTests.dir/tests/cpp/test_unit_graph.cc.o
[100%] Building CXX object CMakeFiles/runUnitTests.dir/tests/cpp/test_zerocopy_serialize.cc.o
/root/jenkins/workspace/dgl_PR-4648@2/tests/dist/cpp/rpc_client.cc: In member function 'void RPCClient::StartClient()':
/root/jenkins/workspace/dgl_PR-4648@2/tests/dist/cpp/rpc_client.cc:50:15: warning: unused variable 'num_machines' [-Wunused-variable]
     const int num_machines = ips_.size();
               ^~~~~~~~~~~~
/root/jenkins/workspace/dgl_PR-4648@2/tests/cpp/test_partition.cc: In instantiation of 'void _TestRemainder_MapToX() [with DGLDeviceType XPU = (DGLDeviceType)2; IdType = int]':
/root/jenkins/workspace/dgl_PR-4648@2/tests/cpp/test_partition.cc:86:44:   required from here
/root/jenkins/workspace/dgl_PR-4648@2/tests/cpp/test_partition.cc:66:26: warning: comparison between signed and unsigned integer expressions [-Wsign-compare]
     for (size_t i = 0; i < global->shape[0]; ++i) {
/root/jenkins/workspace/dgl_PR-4648@2/tests/cpp/test_partition.cc:74:26: warning: comparison between signed and unsigned integer expressions [-Wsign-compare]
     for (size_t i = 0; i < act_local->shape[0]; ++i) {
/root/jenkins/workspace/dgl_PR-4648@2/tests/cpp/test_partition.cc: In instantiation of 'void _TestRemainder_MapToX() [with DGLDeviceType XPU = (DGLDeviceType)2; IdType = long int]':
/root/jenkins/workspace/dgl_PR-4648@2/tests/cpp/test_partition.cc:87:44:   required from here
/root/jenkins/workspace/dgl_PR-4648@2/tests/cpp/test_partition.cc:66:26: warning: comparison between signed and unsigned integer expressions [-Wsign-compare]
     for (size_t i = 0; i < global->shape[0]; ++i) {
/root/jenkins/workspace/dgl_PR-4648@2/tests/cpp/test_partition.cc:74:26: warning: comparison between signed and unsigned integer expressions [-Wsign-compare]
     for (size_t i = 0; i < act_local->shape[0]; ++i) {
/root/jenkins/workspace/dgl_PR-4648@2/tests/cpp/test_partition.cc: In instantiation of 'void _TestRange_MapToX() [with DGLDeviceType XPU = (DGLDeviceType)2; IdType = int]':
/root/jenkins/workspace/dgl_PR-4648@2/tests/cpp/test_partition.cc:195:40:   required from here
/root/jenkins/workspace/dgl_PR-4648@2/tests/cpp/test_partition.cc:176:26: warning: comparison between signed and unsigned integer expressions [-Wsign-compare]
     for (size_t i = 0; i < global->shape[0]; ++i) {
/root/jenkins/workspace/dgl_PR-4648@2/tests/cpp/test_partition.cc:184:26: warning: comparison between signed and unsigned integer expressions [-Wsign-compare]
     for (size_t i = 0; i < act_local->shape[0]; ++i) {
/root/jenkins/workspace/dgl_PR-4648@2/tests/cpp/test_partition.cc: In instantiation of 'void _TestRange_MapToX() [with DGLDeviceType XPU = (DGLDeviceType)2; IdType = long int]':
/root/jenkins/workspace/dgl_PR-4648@2/tests/cpp/test_partition.cc:196:40:   required from here
/root/jenkins/workspace/dgl_PR-4648@2/tests/cpp/test_partition.cc:176:26: warning: comparison between signed and unsigned integer expressions [-Wsign-compare]
     for (size_t i = 0; i < global->shape[0]; ++i) {
/root/jenkins/workspace/dgl_PR-4648@2/tests/cpp/test_partition.cc:184:26: warning: comparison between signed and unsigned integer expressions [-Wsign-compare]
     for (size_t i = 0; i < act_local->shape[0]; ++i) {
[100%] Linking CXX executable rpc_server
[100%] Linking CXX executable rpc_client
[100%] Built target rpc_server
[100%] Built target rpc_client
[100%] Linking CXX executable runUnitTests
[100%] Built target runUnitTests
~/jenkins/workspace/dgl_PR-4648@2
~/jenkins/workspace/dgl_PR-4648@2/python ~/jenkins/workspace/dgl_PR-4648@2
WARNING: Skipping dgl as it is not installed.
WARNING: Running pip as the 'root' user can result in broken permissions and conflicting behaviour with the system package manager. It is recommended to use a virtual environment instead: https://pip.pypa.io/warnings/venv
[1/1] Cythonizing dgl/_ffi/_cython/core.pyx
running install
running bdist_egg
running egg_info
creating dgl.egg-info
writing dgl.egg-info/PKG-INFO
writing dependency_links to dgl.egg-info/dependency_links.txt
writing requirements to dgl.egg-info/requires.txt
writing top-level names to dgl.egg-info/top_level.txt
writing manifest file 'dgl.egg-info/SOURCES.txt'
reading manifest file 'dgl.egg-info/SOURCES.txt'
writing manifest file 'dgl.egg-info/SOURCES.txt'
installing library code to build/bdist.linux-x86_64/egg
running install_lib
running build_py
creating build
creating build/lib.linux-x86_64-3.7
creating build/lib.linux-x86_64-3.7/dgl
copying dgl/partition.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/core.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/subgraph.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/traversal.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/base.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/heterograph_index.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/sparse.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/global_config.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/_api_internal.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/heterograph.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/network.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/logging.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/graph_index.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/udf.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/ndarray.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/init.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/view.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/generators.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/convert.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/container.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/merge.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/readout.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/__init__.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/propagate.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/batch.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/frame.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/random.py -> build/lib.linux-x86_64-3.7/dgl
creating build/lib.linux-x86_64-3.7/dgl/mock_sparse
copying dgl/mock_sparse/sp_matrix.py -> build/lib.linux-x86_64-3.7/dgl/mock_sparse
copying dgl/mock_sparse/diag_matrix.py -> build/lib.linux-x86_64-3.7/dgl/mock_sparse
copying dgl/mock_sparse/reduction.py -> build/lib.linux-x86_64-3.7/dgl/mock_sparse
copying dgl/mock_sparse/__init__.py -> build/lib.linux-x86_64-3.7/dgl/mock_sparse
copying dgl/mock_sparse/elementwise_op_sp.py -> build/lib.linux-x86_64-3.7/dgl/mock_sparse
creating build/lib.linux-x86_64-3.7/dgl/geometry
copying dgl/geometry/edge_coarsening.py -> build/lib.linux-x86_64-3.7/dgl/geometry
copying dgl/geometry/fps.py -> build/lib.linux-x86_64-3.7/dgl/geometry
copying dgl/geometry/capi.py -> build/lib.linux-x86_64-3.7/dgl/geometry
copying dgl/geometry/__init__.py -> build/lib.linux-x86_64-3.7/dgl/geometry
creating build/lib.linux-x86_64-3.7/dgl/function
copying dgl/function/base.py -> build/lib.linux-x86_64-3.7/dgl/function
copying dgl/function/message.py -> build/lib.linux-x86_64-3.7/dgl/function
copying dgl/function/reducer.py -> build/lib.linux-x86_64-3.7/dgl/function
copying dgl/function/__init__.py -> build/lib.linux-x86_64-3.7/dgl/function
creating build/lib.linux-x86_64-3.7/dgl/nn
copying dgl/nn/__init__.py -> build/lib.linux-x86_64-3.7/dgl/nn
creating build/lib.linux-x86_64-3.7/dgl/contrib
copying dgl/contrib/dis_kvstore.py -> build/lib.linux-x86_64-3.7/dgl/contrib
copying dgl/contrib/unified_tensor.py -> build/lib.linux-x86_64-3.7/dgl/contrib
copying dgl/contrib/graph_store.py -> build/lib.linux-x86_64-3.7/dgl/contrib
copying dgl/contrib/__init__.py -> build/lib.linux-x86_64-3.7/dgl/contrib
creating build/lib.linux-x86_64-3.7/dgl/utils
copying dgl/utils/checks.py -> build/lib.linux-x86_64-3.7/dgl/utils
copying dgl/utils/exception.py -> build/lib.linux-x86_64-3.7/dgl/utils
copying dgl/utils/pin_memory.py -> build/lib.linux-x86_64-3.7/dgl/utils
copying dgl/utils/internal.py -> build/lib.linux-x86_64-3.7/dgl/utils
copying dgl/utils/filter.py -> build/lib.linux-x86_64-3.7/dgl/utils
copying dgl/utils/data.py -> build/lib.linux-x86_64-3.7/dgl/utils
copying dgl/utils/__init__.py -> build/lib.linux-x86_64-3.7/dgl/utils
copying dgl/utils/shared_mem.py -> build/lib.linux-x86_64-3.7/dgl/utils
creating build/lib.linux-x86_64-3.7/dgl/multiprocessing
copying dgl/multiprocessing/pytorch.py -> build/lib.linux-x86_64-3.7/dgl/multiprocessing
copying dgl/multiprocessing/__init__.py -> build/lib.linux-x86_64-3.7/dgl/multiprocessing
creating build/lib.linux-x86_64-3.7/dgl/optim
copying dgl/optim/__init__.py -> build/lib.linux-x86_64-3.7/dgl/optim
creating build/lib.linux-x86_64-3.7/dgl/dataloading
copying dgl/dataloading/base.py -> build/lib.linux-x86_64-3.7/dgl/dataloading
copying dgl/dataloading/dataloader.py -> build/lib.linux-x86_64-3.7/dgl/dataloading
copying dgl/dataloading/graphsaint.py -> build/lib.linux-x86_64-3.7/dgl/dataloading
copying dgl/dataloading/shadow.py -> build/lib.linux-x86_64-3.7/dgl/dataloading
copying dgl/dataloading/negative_sampler.py -> build/lib.linux-x86_64-3.7/dgl/dataloading
copying dgl/dataloading/__init__.py -> build/lib.linux-x86_64-3.7/dgl/dataloading
copying dgl/dataloading/cluster_gcn.py -> build/lib.linux-x86_64-3.7/dgl/dataloading
copying dgl/dataloading/neighbor_sampler.py -> build/lib.linux-x86_64-3.7/dgl/dataloading
copying dgl/dataloading/dist_dataloader.py -> build/lib.linux-x86_64-3.7/dgl/dataloading
creating build/lib.linux-x86_64-3.7/dgl/_dataloading
copying dgl/_dataloading/neighbor.py -> build/lib.linux-x86_64-3.7/dgl/_dataloading
copying dgl/_dataloading/dataloader.py -> build/lib.linux-x86_64-3.7/dgl/_dataloading
copying dgl/_dataloading/shadow.py -> build/lib.linux-x86_64-3.7/dgl/_dataloading
copying dgl/_dataloading/negative_sampler.py -> build/lib.linux-x86_64-3.7/dgl/_dataloading
copying dgl/_dataloading/__init__.py -> build/lib.linux-x86_64-3.7/dgl/_dataloading
copying dgl/_dataloading/cluster_gcn.py -> build/lib.linux-x86_64-3.7/dgl/_dataloading
creating build/lib.linux-x86_64-3.7/dgl/backend
copying dgl/backend/backend.py -> build/lib.linux-x86_64-3.7/dgl/backend
copying dgl/backend/__init__.py -> build/lib.linux-x86_64-3.7/dgl/backend
copying dgl/backend/set_default_backend.py -> build/lib.linux-x86_64-3.7/dgl/backend
creating build/lib.linux-x86_64-3.7/dgl/ops
copying dgl/ops/edge_softmax.py -> build/lib.linux-x86_64-3.7/dgl/ops
copying dgl/ops/sddmm.py -> build/lib.linux-x86_64-3.7/dgl/ops
copying dgl/ops/gather_mm.py -> build/lib.linux-x86_64-3.7/dgl/ops
copying dgl/ops/segment.py -> build/lib.linux-x86_64-3.7/dgl/ops
copying dgl/ops/__init__.py -> build/lib.linux-x86_64-3.7/dgl/ops
copying dgl/ops/spmm.py -> build/lib.linux-x86_64-3.7/dgl/ops
creating build/lib.linux-x86_64-3.7/dgl/cuda
copying dgl/cuda/nccl.py -> build/lib.linux-x86_64-3.7/dgl/cuda
copying dgl/cuda/__init__.py -> build/lib.linux-x86_64-3.7/dgl/cuda
creating build/lib.linux-x86_64-3.7/dgl/storages
copying dgl/storages/pytorch_tensor.py -> build/lib.linux-x86_64-3.7/dgl/storages
copying dgl/storages/base.py -> build/lib.linux-x86_64-3.7/dgl/storages
copying dgl/storages/tensor.py -> build/lib.linux-x86_64-3.7/dgl/storages
copying dgl/storages/__init__.py -> build/lib.linux-x86_64-3.7/dgl/storages
copying dgl/storages/numpy.py -> build/lib.linux-x86_64-3.7/dgl/storages
creating build/lib.linux-x86_64-3.7/dgl/sampling
copying dgl/sampling/neighbor.py -> build/lib.linux-x86_64-3.7/dgl/sampling
copying dgl/sampling/pinsage.py -> build/lib.linux-x86_64-3.7/dgl/sampling
copying dgl/sampling/negative.py -> build/lib.linux-x86_64-3.7/dgl/sampling
copying dgl/sampling/randomwalks.py -> build/lib.linux-x86_64-3.7/dgl/sampling
copying dgl/sampling/utils.py -> build/lib.linux-x86_64-3.7/dgl/sampling
copying dgl/sampling/node2vec_randomwalk.py -> build/lib.linux-x86_64-3.7/dgl/sampling
copying dgl/sampling/__init__.py -> build/lib.linux-x86_64-3.7/dgl/sampling
creating build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/qm7b.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/graph_serialize.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/dgl_dataset.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/csv_dataset_base.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/knowledge_graph.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/qm9_edge.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/gnn_benchmark.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/qm9.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/icews18.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/gdelt.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/utils.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/sbm.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/csv_dataset.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/karate.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/citation_graph.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/rdf.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/bitcoinotc.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/ppi.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/tensor_serialize.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/flickr.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/wikics.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/heterograph_serialize.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/tu.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/__init__.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/gindt.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/tree.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/synthetic.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/adapter.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/fakenews.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/fraud.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/minigc.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/yelp.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/reddit.py -> build/lib.linux-x86_64-3.7/dgl/data
creating build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/dist_graph.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/graph_partition_book.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/partition.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/standalone_kvstore.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/graph_services.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/role.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/constants.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/rpc_server.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/kvstore.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/dist_context.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/rpc.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/rpc_client.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/id_map.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/server_state.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/__init__.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/shared_mem_utils.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/dist_tensor.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/dist_dataloader.py -> build/lib.linux-x86_64-3.7/dgl/distributed
creating build/lib.linux-x86_64-3.7/dgl/_deprecate
copying dgl/_deprecate/nodeflow.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate
copying dgl/_deprecate/udf.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate
copying dgl/_deprecate/view.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate
copying dgl/_deprecate/kernel.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate
copying dgl/_deprecate/__init__.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate
copying dgl/_deprecate/graph.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate
copying dgl/_deprecate/frame.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate
creating build/lib.linux-x86_64-3.7/dgl/transforms
copying dgl/transforms/functional.py -> build/lib.linux-x86_64-3.7/dgl/transforms
copying dgl/transforms/__init__.py -> build/lib.linux-x86_64-3.7/dgl/transforms
copying dgl/transforms/module.py -> build/lib.linux-x86_64-3.7/dgl/transforms
creating build/lib.linux-x86_64-3.7/dgl/distgnn
copying dgl/distgnn/__init__.py -> build/lib.linux-x86_64-3.7/dgl/distgnn
creating build/lib.linux-x86_64-3.7/dgl/_ffi
copying dgl/_ffi/runtime_ctypes.py -> build/lib.linux-x86_64-3.7/dgl/_ffi
copying dgl/_ffi/function.py -> build/lib.linux-x86_64-3.7/dgl/_ffi
copying dgl/_ffi/libinfo.py -> build/lib.linux-x86_64-3.7/dgl/_ffi
copying dgl/_ffi/streams.py -> build/lib.linux-x86_64-3.7/dgl/_ffi
copying dgl/_ffi/base.py -> build/lib.linux-x86_64-3.7/dgl/_ffi
copying dgl/_ffi/ndarray.py -> build/lib.linux-x86_64-3.7/dgl/_ffi
copying dgl/_ffi/object_generic.py -> build/lib.linux-x86_64-3.7/dgl/_ffi
copying dgl/_ffi/__init__.py -> build/lib.linux-x86_64-3.7/dgl/_ffi
copying dgl/_ffi/object.py -> build/lib.linux-x86_64-3.7/dgl/_ffi
creating build/lib.linux-x86_64-3.7/dgl/nn/pytorch
copying dgl/nn/pytorch/glob.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch
copying dgl/nn/pytorch/hetero.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch
copying dgl/nn/pytorch/sparse_emb.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch
copying dgl/nn/pytorch/linear.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch
copying dgl/nn/pytorch/utils.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch
copying dgl/nn/pytorch/__init__.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch
copying dgl/nn/pytorch/factory.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch
copying dgl/nn/pytorch/softmax.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch
creating build/lib.linux-x86_64-3.7/dgl/nn/mxnet
copying dgl/nn/mxnet/glob.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet
copying dgl/nn/mxnet/hetero.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet
copying dgl/nn/mxnet/utils.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet
copying dgl/nn/mxnet/__init__.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet
copying dgl/nn/mxnet/softmax.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet
creating build/lib.linux-x86_64-3.7/dgl/nn/tensorflow
copying dgl/nn/tensorflow/glob.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow
copying dgl/nn/tensorflow/hetero.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow
copying dgl/nn/tensorflow/utils.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow
copying dgl/nn/tensorflow/__init__.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow
copying dgl/nn/tensorflow/softmax.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow
creating build/lib.linux-x86_64-3.7/dgl/nn/functional
copying dgl/nn/functional/__init__.py -> build/lib.linux-x86_64-3.7/dgl/nn/functional
creating build/lib.linux-x86_64-3.7/dgl/nn/pytorch/link
copying dgl/nn/pytorch/link/transe.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/link
copying dgl/nn/pytorch/link/edgepred.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/link
copying dgl/nn/pytorch/link/transr.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/link
copying dgl/nn/pytorch/link/__init__.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/link
creating build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/agnnconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/relgraphconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/appnpconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/dgnconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/grouprevres.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/twirlsconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/egatconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/graphconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/densechebconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/gcn2conv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/ginconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/dotgatconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/cfconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/gmmconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/sageconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/densesageconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/gineconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/edgeconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/nnconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/__init__.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/gatv2conv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/hgtconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/pnaconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/gatedgraphconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/atomicconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/tagconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/densegraphconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/egnnconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/chebconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/sgconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/gatconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
creating build/lib.linux-x86_64-3.7/dgl/nn/pytorch/explain
copying dgl/nn/pytorch/explain/gnnexplainer.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/explain
copying dgl/nn/pytorch/explain/__init__.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/explain
creating build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/agnnconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/relgraphconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/appnpconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/graphconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/densechebconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/ginconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/gmmconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/sageconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/densesageconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/edgeconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/nnconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/__init__.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/gatedgraphconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/tagconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/densegraphconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/chebconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/sgconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/gatconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
creating build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv
copying dgl/nn/tensorflow/conv/relgraphconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv
copying dgl/nn/tensorflow/conv/appnpconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv
copying dgl/nn/tensorflow/conv/graphconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv
copying dgl/nn/tensorflow/conv/densechebconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv
copying dgl/nn/tensorflow/conv/ginconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv
copying dgl/nn/tensorflow/conv/sageconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv
copying dgl/nn/tensorflow/conv/edgeconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv
copying dgl/nn/tensorflow/conv/__init__.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv
copying dgl/nn/tensorflow/conv/chebconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv
copying dgl/nn/tensorflow/conv/sgconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv
copying dgl/nn/tensorflow/conv/gatconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv
creating build/lib.linux-x86_64-3.7/dgl/contrib/sampling
copying dgl/contrib/sampling/sampler.py -> build/lib.linux-x86_64-3.7/dgl/contrib/sampling
copying dgl/contrib/sampling/dis_sampler.py -> build/lib.linux-x86_64-3.7/dgl/contrib/sampling
copying dgl/contrib/sampling/__init__.py -> build/lib.linux-x86_64-3.7/dgl/contrib/sampling
creating build/lib.linux-x86_64-3.7/dgl/contrib/data
copying dgl/contrib/data/knowledge_graph.py -> build/lib.linux-x86_64-3.7/dgl/contrib/data
copying dgl/contrib/data/__init__.py -> build/lib.linux-x86_64-3.7/dgl/contrib/data
creating build/lib.linux-x86_64-3.7/dgl/optim/pytorch
copying dgl/optim/pytorch/__init__.py -> build/lib.linux-x86_64-3.7/dgl/optim/pytorch
copying dgl/optim/pytorch/sparse_optim.py -> build/lib.linux-x86_64-3.7/dgl/optim/pytorch
creating build/lib.linux-x86_64-3.7/dgl/optim/mxnet
copying dgl/optim/mxnet/__init__.py -> build/lib.linux-x86_64-3.7/dgl/optim/mxnet
creating build/lib.linux-x86_64-3.7/dgl/optim/tensorflow
copying dgl/optim/tensorflow/__init__.py -> build/lib.linux-x86_64-3.7/dgl/optim/tensorflow
creating build/lib.linux-x86_64-3.7/dgl/_dataloading/pytorch
copying dgl/_dataloading/pytorch/dataloader.py -> build/lib.linux-x86_64-3.7/dgl/_dataloading/pytorch
copying dgl/_dataloading/pytorch/__init__.py -> build/lib.linux-x86_64-3.7/dgl/_dataloading/pytorch
creating build/lib.linux-x86_64-3.7/dgl/backend/pytorch
copying dgl/backend/pytorch/sparse.py -> build/lib.linux-x86_64-3.7/dgl/backend/pytorch
copying dgl/backend/pytorch/tensor.py -> build/lib.linux-x86_64-3.7/dgl/backend/pytorch
copying dgl/backend/pytorch/__init__.py -> build/lib.linux-x86_64-3.7/dgl/backend/pytorch
creating build/lib.linux-x86_64-3.7/dgl/backend/mxnet
copying dgl/backend/mxnet/sparse.py -> build/lib.linux-x86_64-3.7/dgl/backend/mxnet
copying dgl/backend/mxnet/tensor.py -> build/lib.linux-x86_64-3.7/dgl/backend/mxnet
copying dgl/backend/mxnet/__init__.py -> build/lib.linux-x86_64-3.7/dgl/backend/mxnet
copying dgl/backend/mxnet/sparse_optim.py -> build/lib.linux-x86_64-3.7/dgl/backend/mxnet
creating build/lib.linux-x86_64-3.7/dgl/backend/tensorflow
copying dgl/backend/tensorflow/sparse.py -> build/lib.linux-x86_64-3.7/dgl/backend/tensorflow
copying dgl/backend/tensorflow/tensor.py -> build/lib.linux-x86_64-3.7/dgl/backend/tensorflow
copying dgl/backend/tensorflow/__init__.py -> build/lib.linux-x86_64-3.7/dgl/backend/tensorflow
copying dgl/backend/tensorflow/sparse_optim.py -> build/lib.linux-x86_64-3.7/dgl/backend/tensorflow
creating build/lib.linux-x86_64-3.7/dgl/distributed/nn
copying dgl/distributed/nn/__init__.py -> build/lib.linux-x86_64-3.7/dgl/distributed/nn
creating build/lib.linux-x86_64-3.7/dgl/distributed/optim
copying dgl/distributed/optim/__init__.py -> build/lib.linux-x86_64-3.7/dgl/distributed/optim
creating build/lib.linux-x86_64-3.7/dgl/distributed/nn/pytorch
copying dgl/distributed/nn/pytorch/sparse_emb.py -> build/lib.linux-x86_64-3.7/dgl/distributed/nn/pytorch
copying dgl/distributed/nn/pytorch/__init__.py -> build/lib.linux-x86_64-3.7/dgl/distributed/nn/pytorch
creating build/lib.linux-x86_64-3.7/dgl/distributed/nn/mxnet
copying dgl/distributed/nn/mxnet/__init__.py -> build/lib.linux-x86_64-3.7/dgl/distributed/nn/mxnet
creating build/lib.linux-x86_64-3.7/dgl/distributed/nn/tensorflow
copying dgl/distributed/nn/tensorflow/__init__.py -> build/lib.linux-x86_64-3.7/dgl/distributed/nn/tensorflow
creating build/lib.linux-x86_64-3.7/dgl/distributed/optim/pytorch
copying dgl/distributed/optim/pytorch/utils.py -> build/lib.linux-x86_64-3.7/dgl/distributed/optim/pytorch
copying dgl/distributed/optim/pytorch/__init__.py -> build/lib.linux-x86_64-3.7/dgl/distributed/optim/pytorch
copying dgl/distributed/optim/pytorch/sparse_optim.py -> build/lib.linux-x86_64-3.7/dgl/distributed/optim/pytorch
creating build/lib.linux-x86_64-3.7/dgl/distributed/optim/mxnet
copying dgl/distributed/optim/mxnet/__init__.py -> build/lib.linux-x86_64-3.7/dgl/distributed/optim/mxnet
creating build/lib.linux-x86_64-3.7/dgl/distributed/optim/tensorflow
copying dgl/distributed/optim/tensorflow/__init__.py -> build/lib.linux-x86_64-3.7/dgl/distributed/optim/tensorflow
creating build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime
copying dgl/_deprecate/runtime/spmv.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime
copying dgl/_deprecate/runtime/degree_bucketing.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime
copying dgl/_deprecate/runtime/__init__.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime
copying dgl/_deprecate/runtime/scheduler.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime
copying dgl/_deprecate/runtime/adapter.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime
copying dgl/_deprecate/runtime/runtime.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime
creating build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/ir
copying dgl/_deprecate/runtime/ir/program.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/ir
copying dgl/_deprecate/runtime/ir/executor.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/ir
copying dgl/_deprecate/runtime/ir/var.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/ir
copying dgl/_deprecate/runtime/ir/registry.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/ir
copying dgl/_deprecate/runtime/ir/__init__.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/ir
creating build/lib.linux-x86_64-3.7/dgl/distgnn/partition
copying dgl/distgnn/partition/libra_partition.py -> build/lib.linux-x86_64-3.7/dgl/distgnn/partition
copying dgl/distgnn/partition/__init__.py -> build/lib.linux-x86_64-3.7/dgl/distgnn/partition
creating build/lib.linux-x86_64-3.7/dgl/distgnn/tools
copying dgl/distgnn/tools/__init__.py -> build/lib.linux-x86_64-3.7/dgl/distgnn/tools
copying dgl/distgnn/tools/tools.py -> build/lib.linux-x86_64-3.7/dgl/distgnn/tools
creating build/lib.linux-x86_64-3.7/dgl/_ffi/_ctypes
copying dgl/_ffi/_ctypes/function.py -> build/lib.linux-x86_64-3.7/dgl/_ffi/_ctypes
copying dgl/_ffi/_ctypes/types.py -> build/lib.linux-x86_64-3.7/dgl/_ffi/_ctypes
copying dgl/_ffi/_ctypes/ndarray.py -> build/lib.linux-x86_64-3.7/dgl/_ffi/_ctypes
copying dgl/_ffi/_ctypes/__init__.py -> build/lib.linux-x86_64-3.7/dgl/_ffi/_ctypes
copying dgl/_ffi/_ctypes/object.py -> build/lib.linux-x86_64-3.7/dgl/_ffi/_ctypes
creating build/lib.linux-x86_64-3.7/dgl/_ffi/_cy3
copying dgl/_ffi/_cy3/__init__.py -> build/lib.linux-x86_64-3.7/dgl/_ffi/_cy3
creating build/lib.linux-x86_64-3.7/dgl/_ffi/_cy2
copying dgl/_ffi/_cy2/__init__.py -> build/lib.linux-x86_64-3.7/dgl/_ffi/_cy2
creating build/lib.linux-x86_64-3.7/dgl/_ffi/_cython
copying dgl/_ffi/_cython/core.cpp -> build/lib.linux-x86_64-3.7/dgl/_ffi/_cython
running build_ext
building 'dgl._ffi._cy3.core' extension
creating build/temp.linux-x86_64-3.7
creating build/temp.linux-x86_64-3.7/dgl
creating build/temp.linux-x86_64-3.7/dgl/_ffi
creating build/temp.linux-x86_64-3.7/dgl/_ffi/_cython
gcc -pthread -B /opt/conda/envs/pytorch-ci/compiler_compat -Wl,--sysroot=/ -Wsign-compare -DNDEBUG -g -fwrapv -O3 -Wall -Wstrict-prototypes -fPIC -I../include/ -I../third_party/dmlc-core/include -I../third_party/dlpack/include -I/opt/conda/envs/pytorch-ci/include/python3.7m -c dgl/_ffi/_cython/core.cpp -o build/temp.linux-x86_64-3.7/dgl/_ffi/_cython/core.o
cc1plus: warning: command line option ‘-Wstrict-prototypes’ is valid for C/ObjC but not for C++
g++ -pthread -B /opt/conda/envs/pytorch-ci/compiler_compat -Wl,--sysroot=/ -pthread -shared -B /opt/conda/envs/pytorch-ci/compiler_compat -L/opt/conda/envs/pytorch-ci/lib -Wl,-rpath=/opt/conda/envs/pytorch-ci/lib -Wl,--no-as-needed -Wl,--sysroot=/ build/temp.linux-x86_64-3.7/dgl/_ffi/_cython/core.o -Ldgl -L../build/Release -L../build -ldgl -o build/lib.linux-x86_64-3.7/dgl/_ffi/_cy3/core.cpython-37m-x86_64-linux-gnu.so
/opt/conda/envs/pytorch-ci/compiler_compat/ld: warning: /opt/conda/envs/pytorch-ci/lib/libstdc++.so: unsupported GNU_PROPERTY_TYPE (5) type: 0xc0010001
/opt/conda/envs/pytorch-ci/compiler_compat/ld: warning: /opt/conda/envs/pytorch-ci/lib/libstdc++.so: unsupported GNU_PROPERTY_TYPE (5) type: 0xc0010002
/opt/conda/envs/pytorch-ci/compiler_compat/ld: warning: /opt/conda/envs/pytorch-ci/lib/libgcc_s.so.1: unsupported GNU_PROPERTY_TYPE (5) type: 0xc0010001
/opt/conda/envs/pytorch-ci/compiler_compat/ld: warning: /opt/conda/envs/pytorch-ci/lib/libgcc_s.so.1: unsupported GNU_PROPERTY_TYPE (5) type: 0xc0010002
/opt/conda/envs/pytorch-ci/compiler_compat/ld: warning: /opt/conda/envs/pytorch-ci/lib/libgcc_s.so.1: unsupported GNU_PROPERTY_TYPE (5) type: 0xc0010001
/opt/conda/envs/pytorch-ci/compiler_compat/ld: warning: /opt/conda/envs/pytorch-ci/lib/libgcc_s.so.1: unsupported GNU_PROPERTY_TYPE (5) type: 0xc0010002
creating build/bdist.linux-x86_64
creating build/bdist.linux-x86_64/egg
creating build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/mock_sparse
copying build/lib.linux-x86_64-3.7/dgl/mock_sparse/sp_matrix.py -> build/bdist.linux-x86_64/egg/dgl/mock_sparse
copying build/lib.linux-x86_64-3.7/dgl/mock_sparse/diag_matrix.py -> build/bdist.linux-x86_64/egg/dgl/mock_sparse
copying build/lib.linux-x86_64-3.7/dgl/mock_sparse/reduction.py -> build/bdist.linux-x86_64/egg/dgl/mock_sparse
copying build/lib.linux-x86_64-3.7/dgl/mock_sparse/__init__.py -> build/bdist.linux-x86_64/egg/dgl/mock_sparse
copying build/lib.linux-x86_64-3.7/dgl/mock_sparse/elementwise_op_sp.py -> build/bdist.linux-x86_64/egg/dgl/mock_sparse
creating build/bdist.linux-x86_64/egg/dgl/geometry
copying build/lib.linux-x86_64-3.7/dgl/geometry/edge_coarsening.py -> build/bdist.linux-x86_64/egg/dgl/geometry
copying build/lib.linux-x86_64-3.7/dgl/geometry/fps.py -> build/bdist.linux-x86_64/egg/dgl/geometry
copying build/lib.linux-x86_64-3.7/dgl/geometry/capi.py -> build/bdist.linux-x86_64/egg/dgl/geometry
copying build/lib.linux-x86_64-3.7/dgl/geometry/__init__.py -> build/bdist.linux-x86_64/egg/dgl/geometry
creating build/bdist.linux-x86_64/egg/dgl/function
copying build/lib.linux-x86_64-3.7/dgl/function/base.py -> build/bdist.linux-x86_64/egg/dgl/function
copying build/lib.linux-x86_64-3.7/dgl/function/message.py -> build/bdist.linux-x86_64/egg/dgl/function
copying build/lib.linux-x86_64-3.7/dgl/function/reducer.py -> build/bdist.linux-x86_64/egg/dgl/function
copying build/lib.linux-x86_64-3.7/dgl/function/__init__.py -> build/bdist.linux-x86_64/egg/dgl/function
copying build/lib.linux-x86_64-3.7/dgl/partition.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/nn
creating build/bdist.linux-x86_64/egg/dgl/nn/pytorch
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/glob.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch
creating build/bdist.linux-x86_64/egg/dgl/nn/pytorch/link
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/link/transe.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/link
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/link/edgepred.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/link
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/link/transr.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/link
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/link/__init__.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/link
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/hetero.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/sparse_emb.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/linear.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/utils.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch
creating build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/agnnconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/relgraphconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/appnpconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/dgnconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/grouprevres.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/twirlsconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/egatconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/graphconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/densechebconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/gcn2conv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/ginconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/dotgatconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/cfconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/gmmconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/sageconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/densesageconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/gineconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/edgeconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/nnconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/__init__.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/gatv2conv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/hgtconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/pnaconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/gatedgraphconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/atomicconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/tagconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/densegraphconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/egnnconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/chebconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/sgconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/gatconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/__init__.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch
creating build/bdist.linux-x86_64/egg/dgl/nn/pytorch/explain
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/explain/gnnexplainer.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/explain
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/explain/__init__.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/explain
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/factory.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/softmax.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch
creating build/bdist.linux-x86_64/egg/dgl/nn/mxnet
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/glob.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/hetero.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/utils.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet
creating build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/agnnconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/relgraphconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/appnpconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/graphconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/densechebconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/ginconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/gmmconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/sageconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/densesageconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/edgeconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/nnconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/__init__.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/gatedgraphconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/tagconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/densegraphconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/chebconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/sgconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/gatconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/__init__.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/softmax.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet
copying build/lib.linux-x86_64-3.7/dgl/nn/__init__.py -> build/bdist.linux-x86_64/egg/dgl/nn
creating build/bdist.linux-x86_64/egg/dgl/nn/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/glob.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/hetero.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/utils.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow
creating build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv/relgraphconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv/appnpconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv/graphconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv/densechebconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv/ginconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv/sageconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv/edgeconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv/__init__.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv/chebconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv/sgconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv/gatconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/__init__.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/softmax.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow
creating build/bdist.linux-x86_64/egg/dgl/nn/functional
copying build/lib.linux-x86_64-3.7/dgl/nn/functional/__init__.py -> build/bdist.linux-x86_64/egg/dgl/nn/functional
copying build/lib.linux-x86_64-3.7/dgl/core.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-3.7/dgl/subgraph.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/contrib
copying build/lib.linux-x86_64-3.7/dgl/contrib/dis_kvstore.py -> build/bdist.linux-x86_64/egg/dgl/contrib
copying build/lib.linux-x86_64-3.7/dgl/contrib/unified_tensor.py -> build/bdist.linux-x86_64/egg/dgl/contrib
copying build/lib.linux-x86_64-3.7/dgl/contrib/graph_store.py -> build/bdist.linux-x86_64/egg/dgl/contrib
creating build/bdist.linux-x86_64/egg/dgl/contrib/sampling
copying build/lib.linux-x86_64-3.7/dgl/contrib/sampling/sampler.py -> build/bdist.linux-x86_64/egg/dgl/contrib/sampling
copying build/lib.linux-x86_64-3.7/dgl/contrib/sampling/dis_sampler.py -> build/bdist.linux-x86_64/egg/dgl/contrib/sampling
copying build/lib.linux-x86_64-3.7/dgl/contrib/sampling/__init__.py -> build/bdist.linux-x86_64/egg/dgl/contrib/sampling
creating build/bdist.linux-x86_64/egg/dgl/contrib/data
copying build/lib.linux-x86_64-3.7/dgl/contrib/data/knowledge_graph.py -> build/bdist.linux-x86_64/egg/dgl/contrib/data
copying build/lib.linux-x86_64-3.7/dgl/contrib/data/__init__.py -> build/bdist.linux-x86_64/egg/dgl/contrib/data
copying build/lib.linux-x86_64-3.7/dgl/contrib/__init__.py -> build/bdist.linux-x86_64/egg/dgl/contrib
copying build/lib.linux-x86_64-3.7/dgl/traversal.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/utils
copying build/lib.linux-x86_64-3.7/dgl/utils/checks.py -> build/bdist.linux-x86_64/egg/dgl/utils
copying build/lib.linux-x86_64-3.7/dgl/utils/exception.py -> build/bdist.linux-x86_64/egg/dgl/utils
copying build/lib.linux-x86_64-3.7/dgl/utils/pin_memory.py -> build/bdist.linux-x86_64/egg/dgl/utils
copying build/lib.linux-x86_64-3.7/dgl/utils/internal.py -> build/bdist.linux-x86_64/egg/dgl/utils
copying build/lib.linux-x86_64-3.7/dgl/utils/filter.py -> build/bdist.linux-x86_64/egg/dgl/utils
copying build/lib.linux-x86_64-3.7/dgl/utils/data.py -> build/bdist.linux-x86_64/egg/dgl/utils
copying build/lib.linux-x86_64-3.7/dgl/utils/__init__.py -> build/bdist.linux-x86_64/egg/dgl/utils
copying build/lib.linux-x86_64-3.7/dgl/utils/shared_mem.py -> build/bdist.linux-x86_64/egg/dgl/utils
creating build/bdist.linux-x86_64/egg/dgl/multiprocessing
copying build/lib.linux-x86_64-3.7/dgl/multiprocessing/pytorch.py -> build/bdist.linux-x86_64/egg/dgl/multiprocessing
copying build/lib.linux-x86_64-3.7/dgl/multiprocessing/__init__.py -> build/bdist.linux-x86_64/egg/dgl/multiprocessing
copying build/lib.linux-x86_64-3.7/dgl/base.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/optim
creating build/bdist.linux-x86_64/egg/dgl/optim/pytorch
copying build/lib.linux-x86_64-3.7/dgl/optim/pytorch/__init__.py -> build/bdist.linux-x86_64/egg/dgl/optim/pytorch
copying build/lib.linux-x86_64-3.7/dgl/optim/pytorch/sparse_optim.py -> build/bdist.linux-x86_64/egg/dgl/optim/pytorch
creating build/bdist.linux-x86_64/egg/dgl/optim/mxnet
copying build/lib.linux-x86_64-3.7/dgl/optim/mxnet/__init__.py -> build/bdist.linux-x86_64/egg/dgl/optim/mxnet
copying build/lib.linux-x86_64-3.7/dgl/optim/__init__.py -> build/bdist.linux-x86_64/egg/dgl/optim
creating build/bdist.linux-x86_64/egg/dgl/optim/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/optim/tensorflow/__init__.py -> build/bdist.linux-x86_64/egg/dgl/optim/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/heterograph_index.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-3.7/dgl/sparse.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-3.7/dgl/global_config.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/dataloading
copying build/lib.linux-x86_64-3.7/dgl/dataloading/base.py -> build/bdist.linux-x86_64/egg/dgl/dataloading
copying build/lib.linux-x86_64-3.7/dgl/dataloading/dataloader.py -> build/bdist.linux-x86_64/egg/dgl/dataloading
copying build/lib.linux-x86_64-3.7/dgl/dataloading/graphsaint.py -> build/bdist.linux-x86_64/egg/dgl/dataloading
copying build/lib.linux-x86_64-3.7/dgl/dataloading/shadow.py -> build/bdist.linux-x86_64/egg/dgl/dataloading
copying build/lib.linux-x86_64-3.7/dgl/dataloading/negative_sampler.py -> build/bdist.linux-x86_64/egg/dgl/dataloading
copying build/lib.linux-x86_64-3.7/dgl/dataloading/__init__.py -> build/bdist.linux-x86_64/egg/dgl/dataloading
copying build/lib.linux-x86_64-3.7/dgl/dataloading/cluster_gcn.py -> build/bdist.linux-x86_64/egg/dgl/dataloading
copying build/lib.linux-x86_64-3.7/dgl/dataloading/neighbor_sampler.py -> build/bdist.linux-x86_64/egg/dgl/dataloading
copying build/lib.linux-x86_64-3.7/dgl/dataloading/dist_dataloader.py -> build/bdist.linux-x86_64/egg/dgl/dataloading
creating build/bdist.linux-x86_64/egg/dgl/_dataloading
copying build/lib.linux-x86_64-3.7/dgl/_dataloading/neighbor.py -> build/bdist.linux-x86_64/egg/dgl/_dataloading
copying build/lib.linux-x86_64-3.7/dgl/_dataloading/dataloader.py -> build/bdist.linux-x86_64/egg/dgl/_dataloading
creating build/bdist.linux-x86_64/egg/dgl/_dataloading/pytorch
copying build/lib.linux-x86_64-3.7/dgl/_dataloading/pytorch/dataloader.py -> build/bdist.linux-x86_64/egg/dgl/_dataloading/pytorch
copying build/lib.linux-x86_64-3.7/dgl/_dataloading/pytorch/__init__.py -> build/bdist.linux-x86_64/egg/dgl/_dataloading/pytorch
copying build/lib.linux-x86_64-3.7/dgl/_dataloading/shadow.py -> build/bdist.linux-x86_64/egg/dgl/_dataloading
copying build/lib.linux-x86_64-3.7/dgl/_dataloading/negative_sampler.py -> build/bdist.linux-x86_64/egg/dgl/_dataloading
copying build/lib.linux-x86_64-3.7/dgl/_dataloading/__init__.py -> build/bdist.linux-x86_64/egg/dgl/_dataloading
copying build/lib.linux-x86_64-3.7/dgl/_dataloading/cluster_gcn.py -> build/bdist.linux-x86_64/egg/dgl/_dataloading
copying build/lib.linux-x86_64-3.7/dgl/_api_internal.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-3.7/dgl/heterograph.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/backend
copying build/lib.linux-x86_64-3.7/dgl/backend/backend.py -> build/bdist.linux-x86_64/egg/dgl/backend
creating build/bdist.linux-x86_64/egg/dgl/backend/pytorch
copying build/lib.linux-x86_64-3.7/dgl/backend/pytorch/sparse.py -> build/bdist.linux-x86_64/egg/dgl/backend/pytorch
copying build/lib.linux-x86_64-3.7/dgl/backend/pytorch/tensor.py -> build/bdist.linux-x86_64/egg/dgl/backend/pytorch
copying build/lib.linux-x86_64-3.7/dgl/backend/pytorch/__init__.py -> build/bdist.linux-x86_64/egg/dgl/backend/pytorch
creating build/bdist.linux-x86_64/egg/dgl/backend/mxnet
copying build/lib.linux-x86_64-3.7/dgl/backend/mxnet/sparse.py -> build/bdist.linux-x86_64/egg/dgl/backend/mxnet
copying build/lib.linux-x86_64-3.7/dgl/backend/mxnet/tensor.py -> build/bdist.linux-x86_64/egg/dgl/backend/mxnet
copying build/lib.linux-x86_64-3.7/dgl/backend/mxnet/__init__.py -> build/bdist.linux-x86_64/egg/dgl/backend/mxnet
copying build/lib.linux-x86_64-3.7/dgl/backend/mxnet/sparse_optim.py -> build/bdist.linux-x86_64/egg/dgl/backend/mxnet
copying build/lib.linux-x86_64-3.7/dgl/backend/__init__.py -> build/bdist.linux-x86_64/egg/dgl/backend
creating build/bdist.linux-x86_64/egg/dgl/backend/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/backend/tensorflow/sparse.py -> build/bdist.linux-x86_64/egg/dgl/backend/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/backend/tensorflow/tensor.py -> build/bdist.linux-x86_64/egg/dgl/backend/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/backend/tensorflow/__init__.py -> build/bdist.linux-x86_64/egg/dgl/backend/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/backend/tensorflow/sparse_optim.py -> build/bdist.linux-x86_64/egg/dgl/backend/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/backend/set_default_backend.py -> build/bdist.linux-x86_64/egg/dgl/backend
copying build/lib.linux-x86_64-3.7/dgl/network.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/ops
copying build/lib.linux-x86_64-3.7/dgl/ops/edge_softmax.py -> build/bdist.linux-x86_64/egg/dgl/ops
copying build/lib.linux-x86_64-3.7/dgl/ops/sddmm.py -> build/bdist.linux-x86_64/egg/dgl/ops
copying build/lib.linux-x86_64-3.7/dgl/ops/gather_mm.py -> build/bdist.linux-x86_64/egg/dgl/ops
copying build/lib.linux-x86_64-3.7/dgl/ops/segment.py -> build/bdist.linux-x86_64/egg/dgl/ops
copying build/lib.linux-x86_64-3.7/dgl/ops/__init__.py -> build/bdist.linux-x86_64/egg/dgl/ops
copying build/lib.linux-x86_64-3.7/dgl/ops/spmm.py -> build/bdist.linux-x86_64/egg/dgl/ops
copying build/lib.linux-x86_64-3.7/dgl/logging.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-3.7/dgl/graph_index.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/cuda
copying build/lib.linux-x86_64-3.7/dgl/cuda/nccl.py -> build/bdist.linux-x86_64/egg/dgl/cuda
copying build/lib.linux-x86_64-3.7/dgl/cuda/__init__.py -> build/bdist.linux-x86_64/egg/dgl/cuda
copying build/lib.linux-x86_64-3.7/dgl/udf.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/storages
copying build/lib.linux-x86_64-3.7/dgl/storages/pytorch_tensor.py -> build/bdist.linux-x86_64/egg/dgl/storages
copying build/lib.linux-x86_64-3.7/dgl/storages/base.py -> build/bdist.linux-x86_64/egg/dgl/storages
copying build/lib.linux-x86_64-3.7/dgl/storages/tensor.py -> build/bdist.linux-x86_64/egg/dgl/storages
copying build/lib.linux-x86_64-3.7/dgl/storages/__init__.py -> build/bdist.linux-x86_64/egg/dgl/storages
copying build/lib.linux-x86_64-3.7/dgl/storages/numpy.py -> build/bdist.linux-x86_64/egg/dgl/storages
copying build/lib.linux-x86_64-3.7/dgl/ndarray.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-3.7/dgl/init.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-3.7/dgl/view.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/sampling
copying build/lib.linux-x86_64-3.7/dgl/sampling/neighbor.py -> build/bdist.linux-x86_64/egg/dgl/sampling
copying build/lib.linux-x86_64-3.7/dgl/sampling/pinsage.py -> build/bdist.linux-x86_64/egg/dgl/sampling
copying build/lib.linux-x86_64-3.7/dgl/sampling/negative.py -> build/bdist.linux-x86_64/egg/dgl/sampling
copying build/lib.linux-x86_64-3.7/dgl/sampling/randomwalks.py -> build/bdist.linux-x86_64/egg/dgl/sampling
copying build/lib.linux-x86_64-3.7/dgl/sampling/utils.py -> build/bdist.linux-x86_64/egg/dgl/sampling
copying build/lib.linux-x86_64-3.7/dgl/sampling/node2vec_randomwalk.py -> build/bdist.linux-x86_64/egg/dgl/sampling
copying build/lib.linux-x86_64-3.7/dgl/sampling/__init__.py -> build/bdist.linux-x86_64/egg/dgl/sampling
copying build/lib.linux-x86_64-3.7/dgl/generators.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/qm7b.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/graph_serialize.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/dgl_dataset.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/csv_dataset_base.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/knowledge_graph.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/qm9_edge.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/gnn_benchmark.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/qm9.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/icews18.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/gdelt.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/utils.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/sbm.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/csv_dataset.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/karate.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/citation_graph.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/rdf.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/bitcoinotc.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/ppi.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/tensor_serialize.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/flickr.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/wikics.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/heterograph_serialize.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/tu.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/__init__.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/gindt.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/tree.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/synthetic.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/adapter.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/fakenews.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/fraud.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/minigc.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/yelp.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/reddit.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/convert.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-3.7/dgl/container.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-3.7/dgl/merge.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-3.7/dgl/readout.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-3.7/dgl/__init__.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/dist_graph.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/graph_partition_book.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/partition.py -> build/bdist.linux-x86_64/egg/dgl/distributed
creating build/bdist.linux-x86_64/egg/dgl/distributed/nn
creating build/bdist.linux-x86_64/egg/dgl/distributed/nn/pytorch
copying build/lib.linux-x86_64-3.7/dgl/distributed/nn/pytorch/sparse_emb.py -> build/bdist.linux-x86_64/egg/dgl/distributed/nn/pytorch
copying build/lib.linux-x86_64-3.7/dgl/distributed/nn/pytorch/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distributed/nn/pytorch
creating build/bdist.linux-x86_64/egg/dgl/distributed/nn/mxnet
copying build/lib.linux-x86_64-3.7/dgl/distributed/nn/mxnet/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distributed/nn/mxnet
copying build/lib.linux-x86_64-3.7/dgl/distributed/nn/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distributed/nn
creating build/bdist.linux-x86_64/egg/dgl/distributed/nn/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/distributed/nn/tensorflow/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distributed/nn/tensorflow
creating build/bdist.linux-x86_64/egg/dgl/distributed/optim
creating build/bdist.linux-x86_64/egg/dgl/distributed/optim/pytorch
copying build/lib.linux-x86_64-3.7/dgl/distributed/optim/pytorch/utils.py -> build/bdist.linux-x86_64/egg/dgl/distributed/optim/pytorch
copying build/lib.linux-x86_64-3.7/dgl/distributed/optim/pytorch/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distributed/optim/pytorch
copying build/lib.linux-x86_64-3.7/dgl/distributed/optim/pytorch/sparse_optim.py -> build/bdist.linux-x86_64/egg/dgl/distributed/optim/pytorch
creating build/bdist.linux-x86_64/egg/dgl/distributed/optim/mxnet
copying build/lib.linux-x86_64-3.7/dgl/distributed/optim/mxnet/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distributed/optim/mxnet
copying build/lib.linux-x86_64-3.7/dgl/distributed/optim/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distributed/optim
creating build/bdist.linux-x86_64/egg/dgl/distributed/optim/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/distributed/optim/tensorflow/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distributed/optim/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/distributed/standalone_kvstore.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/graph_services.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/role.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/constants.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/rpc_server.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/kvstore.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/dist_context.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/rpc.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/rpc_client.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/id_map.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/server_state.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/shared_mem_utils.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/dist_tensor.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/dist_dataloader.py -> build/bdist.linux-x86_64/egg/dgl/distributed
creating build/bdist.linux-x86_64/egg/dgl/_deprecate
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/nodeflow.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/udf.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/view.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/kernel.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/__init__.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate
creating build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/spmv.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/degree_bucketing.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/__init__.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/scheduler.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/adapter.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/runtime.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime
creating build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/ir
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/ir/program.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/ir
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/ir/executor.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/ir
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/ir/var.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/ir
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/ir/registry.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/ir
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/ir/__init__.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/ir
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/graph.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/frame.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate
copying build/lib.linux-x86_64-3.7/dgl/propagate.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-3.7/dgl/batch.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/transforms
copying build/lib.linux-x86_64-3.7/dgl/transforms/functional.py -> build/bdist.linux-x86_64/egg/dgl/transforms
copying build/lib.linux-x86_64-3.7/dgl/transforms/__init__.py -> build/bdist.linux-x86_64/egg/dgl/transforms
copying build/lib.linux-x86_64-3.7/dgl/transforms/module.py -> build/bdist.linux-x86_64/egg/dgl/transforms
creating build/bdist.linux-x86_64/egg/dgl/distgnn
creating build/bdist.linux-x86_64/egg/dgl/distgnn/partition
copying build/lib.linux-x86_64-3.7/dgl/distgnn/partition/libra_partition.py -> build/bdist.linux-x86_64/egg/dgl/distgnn/partition
copying build/lib.linux-x86_64-3.7/dgl/distgnn/partition/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distgnn/partition
creating build/bdist.linux-x86_64/egg/dgl/distgnn/tools
copying build/lib.linux-x86_64-3.7/dgl/distgnn/tools/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distgnn/tools
copying build/lib.linux-x86_64-3.7/dgl/distgnn/tools/tools.py -> build/bdist.linux-x86_64/egg/dgl/distgnn/tools
copying build/lib.linux-x86_64-3.7/dgl/distgnn/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distgnn
creating build/bdist.linux-x86_64/egg/dgl/_ffi
copying build/lib.linux-x86_64-3.7/dgl/_ffi/runtime_ctypes.py -> build/bdist.linux-x86_64/egg/dgl/_ffi
copying build/lib.linux-x86_64-3.7/dgl/_ffi/function.py -> build/bdist.linux-x86_64/egg/dgl/_ffi
copying build/lib.linux-x86_64-3.7/dgl/_ffi/libinfo.py -> build/bdist.linux-x86_64/egg/dgl/_ffi
copying build/lib.linux-x86_64-3.7/dgl/_ffi/streams.py -> build/bdist.linux-x86_64/egg/dgl/_ffi
copying build/lib.linux-x86_64-3.7/dgl/_ffi/base.py -> build/bdist.linux-x86_64/egg/dgl/_ffi
creating build/bdist.linux-x86_64/egg/dgl/_ffi/_ctypes
copying build/lib.linux-x86_64-3.7/dgl/_ffi/_ctypes/function.py -> build/bdist.linux-x86_64/egg/dgl/_ffi/_ctypes
copying build/lib.linux-x86_64-3.7/dgl/_ffi/_ctypes/types.py -> build/bdist.linux-x86_64/egg/dgl/_ffi/_ctypes
copying build/lib.linux-x86_64-3.7/dgl/_ffi/_ctypes/ndarray.py -> build/bdist.linux-x86_64/egg/dgl/_ffi/_ctypes
copying build/lib.linux-x86_64-3.7/dgl/_ffi/_ctypes/__init__.py -> build/bdist.linux-x86_64/egg/dgl/_ffi/_ctypes
copying build/lib.linux-x86_64-3.7/dgl/_ffi/_ctypes/object.py -> build/bdist.linux-x86_64/egg/dgl/_ffi/_ctypes
copying build/lib.linux-x86_64-3.7/dgl/_ffi/ndarray.py -> build/bdist.linux-x86_64/egg/dgl/_ffi
copying build/lib.linux-x86_64-3.7/dgl/_ffi/object_generic.py -> build/bdist.linux-x86_64/egg/dgl/_ffi
copying build/lib.linux-x86_64-3.7/dgl/_ffi/__init__.py -> build/bdist.linux-x86_64/egg/dgl/_ffi
copying build/lib.linux-x86_64-3.7/dgl/_ffi/object.py -> build/bdist.linux-x86_64/egg/dgl/_ffi
creating build/bdist.linux-x86_64/egg/dgl/_ffi/_cy3
copying build/lib.linux-x86_64-3.7/dgl/_ffi/_cy3/__init__.py -> build/bdist.linux-x86_64/egg/dgl/_ffi/_cy3
copying build/lib.linux-x86_64-3.7/dgl/_ffi/_cy3/core.cpython-37m-x86_64-linux-gnu.so -> build/bdist.linux-x86_64/egg/dgl/_ffi/_cy3
creating build/bdist.linux-x86_64/egg/dgl/_ffi/_cy2
copying build/lib.linux-x86_64-3.7/dgl/_ffi/_cy2/__init__.py -> build/bdist.linux-x86_64/egg/dgl/_ffi/_cy2
creating build/bdist.linux-x86_64/egg/dgl/_ffi/_cython
copying build/lib.linux-x86_64-3.7/dgl/_ffi/_cython/core.cpp -> build/bdist.linux-x86_64/egg/dgl/_ffi/_cython
copying build/lib.linux-x86_64-3.7/dgl/frame.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-3.7/dgl/random.py -> build/bdist.linux-x86_64/egg/dgl
byte-compiling build/bdist.linux-x86_64/egg/dgl/mock_sparse/sp_matrix.py to sp_matrix.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/mock_sparse/diag_matrix.py to diag_matrix.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/mock_sparse/reduction.py to reduction.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/mock_sparse/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/mock_sparse/elementwise_op_sp.py to elementwise_op_sp.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/geometry/edge_coarsening.py to edge_coarsening.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/geometry/fps.py to fps.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/geometry/capi.py to capi.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/geometry/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/function/base.py to base.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/function/message.py to message.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/function/reducer.py to reducer.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/function/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/partition.py to partition.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/glob.py to glob.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/link/transe.py to transe.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/link/edgepred.py to edgepred.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/link/transr.py to transr.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/link/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/hetero.py to hetero.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/sparse_emb.py to sparse_emb.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/linear.py to linear.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/utils.py to utils.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/agnnconv.py to agnnconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/relgraphconv.py to relgraphconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/appnpconv.py to appnpconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/dgnconv.py to dgnconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/grouprevres.py to grouprevres.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/twirlsconv.py to twirlsconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/egatconv.py to egatconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/graphconv.py to graphconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/densechebconv.py to densechebconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/gcn2conv.py to gcn2conv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/ginconv.py to ginconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/dotgatconv.py to dotgatconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/cfconv.py to cfconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/gmmconv.py to gmmconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/sageconv.py to sageconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/densesageconv.py to densesageconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/gineconv.py to gineconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/edgeconv.py to edgeconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/nnconv.py to nnconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/gatv2conv.py to gatv2conv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/hgtconv.py to hgtconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/pnaconv.py to pnaconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/gatedgraphconv.py to gatedgraphconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/atomicconv.py to atomicconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/tagconv.py to tagconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/densegraphconv.py to densegraphconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/egnnconv.py to egnnconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/chebconv.py to chebconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/sgconv.py to sgconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/gatconv.py to gatconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/explain/gnnexplainer.py to gnnexplainer.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/explain/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/factory.py to factory.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/softmax.py to softmax.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/glob.py to glob.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/hetero.py to hetero.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/utils.py to utils.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/agnnconv.py to agnnconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/relgraphconv.py to relgraphconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/appnpconv.py to appnpconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/graphconv.py to graphconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/densechebconv.py to densechebconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/ginconv.py to ginconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/gmmconv.py to gmmconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/sageconv.py to sageconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/densesageconv.py to densesageconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/edgeconv.py to edgeconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/nnconv.py to nnconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/gatedgraphconv.py to gatedgraphconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/tagconv.py to tagconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/densegraphconv.py to densegraphconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/chebconv.py to chebconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/sgconv.py to sgconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/gatconv.py to gatconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/softmax.py to softmax.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/glob.py to glob.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/hetero.py to hetero.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/utils.py to utils.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv/relgraphconv.py to relgraphconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv/appnpconv.py to appnpconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv/graphconv.py to graphconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv/densechebconv.py to densechebconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv/ginconv.py to ginconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv/sageconv.py to sageconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv/edgeconv.py to edgeconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv/chebconv.py to chebconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv/sgconv.py to sgconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv/gatconv.py to gatconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/softmax.py to softmax.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/functional/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/core.py to core.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/subgraph.py to subgraph.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/contrib/dis_kvstore.py to dis_kvstore.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/contrib/unified_tensor.py to unified_tensor.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/contrib/graph_store.py to graph_store.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/contrib/sampling/sampler.py to sampler.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/contrib/sampling/dis_sampler.py to dis_sampler.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/contrib/sampling/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/contrib/data/knowledge_graph.py to knowledge_graph.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/contrib/data/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/contrib/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/traversal.py to traversal.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/utils/checks.py to checks.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/utils/exception.py to exception.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/utils/pin_memory.py to pin_memory.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/utils/internal.py to internal.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/utils/filter.py to filter.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/utils/data.py to data.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/utils/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/utils/shared_mem.py to shared_mem.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/multiprocessing/pytorch.py to pytorch.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/multiprocessing/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/base.py to base.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/optim/pytorch/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/optim/pytorch/sparse_optim.py to sparse_optim.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/optim/mxnet/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/optim/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/optim/tensorflow/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/heterograph_index.py to heterograph_index.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/sparse.py to sparse.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/global_config.py to global_config.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/dataloading/base.py to base.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/dataloading/dataloader.py to dataloader.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/dataloading/graphsaint.py to graphsaint.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/dataloading/shadow.py to shadow.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/dataloading/negative_sampler.py to negative_sampler.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/dataloading/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/dataloading/cluster_gcn.py to cluster_gcn.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/dataloading/neighbor_sampler.py to neighbor_sampler.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/dataloading/dist_dataloader.py to dist_dataloader.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_dataloading/neighbor.py to neighbor.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_dataloading/dataloader.py to dataloader.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_dataloading/pytorch/dataloader.py to dataloader.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_dataloading/pytorch/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_dataloading/shadow.py to shadow.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_dataloading/negative_sampler.py to negative_sampler.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_dataloading/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_dataloading/cluster_gcn.py to cluster_gcn.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_api_internal.py to _api_internal.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/heterograph.py to heterograph.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/backend.py to backend.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/pytorch/sparse.py to sparse.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/pytorch/tensor.py to tensor.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/pytorch/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/mxnet/sparse.py to sparse.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/mxnet/tensor.py to tensor.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/mxnet/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/mxnet/sparse_optim.py to sparse_optim.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/tensorflow/sparse.py to sparse.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/tensorflow/tensor.py to tensor.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/tensorflow/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/tensorflow/sparse_optim.py to sparse_optim.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/set_default_backend.py to set_default_backend.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/network.py to network.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/ops/edge_softmax.py to edge_softmax.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/ops/sddmm.py to sddmm.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/ops/gather_mm.py to gather_mm.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/ops/segment.py to segment.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/ops/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/ops/spmm.py to spmm.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/logging.py to logging.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/graph_index.py to graph_index.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/cuda/nccl.py to nccl.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/cuda/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/udf.py to udf.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/storages/pytorch_tensor.py to pytorch_tensor.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/storages/base.py to base.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/storages/tensor.py to tensor.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/storages/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/storages/numpy.py to numpy.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/ndarray.py to ndarray.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/init.py to init.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/view.py to view.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/sampling/neighbor.py to neighbor.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/sampling/pinsage.py to pinsage.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/sampling/negative.py to negative.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/sampling/randomwalks.py to randomwalks.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/sampling/utils.py to utils.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/sampling/node2vec_randomwalk.py to node2vec_randomwalk.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/sampling/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/generators.py to generators.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/qm7b.py to qm7b.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/graph_serialize.py to graph_serialize.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/dgl_dataset.py to dgl_dataset.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/csv_dataset_base.py to csv_dataset_base.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/knowledge_graph.py to knowledge_graph.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/qm9_edge.py to qm9_edge.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/gnn_benchmark.py to gnn_benchmark.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/qm9.py to qm9.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/icews18.py to icews18.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/gdelt.py to gdelt.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/utils.py to utils.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/sbm.py to sbm.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/csv_dataset.py to csv_dataset.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/karate.py to karate.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/citation_graph.py to citation_graph.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/rdf.py to rdf.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/bitcoinotc.py to bitcoinotc.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/ppi.py to ppi.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/tensor_serialize.py to tensor_serialize.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/flickr.py to flickr.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/wikics.py to wikics.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/heterograph_serialize.py to heterograph_serialize.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/tu.py to tu.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/gindt.py to gindt.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/tree.py to tree.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/synthetic.py to synthetic.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/adapter.py to adapter.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/fakenews.py to fakenews.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/fraud.py to fraud.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/minigc.py to minigc.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/yelp.py to yelp.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/reddit.py to reddit.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/convert.py to convert.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/container.py to container.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/merge.py to merge.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/readout.py to readout.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/dist_graph.py to dist_graph.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/graph_partition_book.py to graph_partition_book.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/partition.py to partition.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/nn/pytorch/sparse_emb.py to sparse_emb.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/nn/pytorch/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/nn/mxnet/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/nn/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/nn/tensorflow/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/optim/pytorch/utils.py to utils.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/optim/pytorch/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/optim/pytorch/sparse_optim.py to sparse_optim.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/optim/mxnet/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/optim/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/optim/tensorflow/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/standalone_kvstore.py to standalone_kvstore.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/graph_services.py to graph_services.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/role.py to role.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/constants.py to constants.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/rpc_server.py to rpc_server.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/kvstore.py to kvstore.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/dist_context.py to dist_context.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/rpc.py to rpc.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/rpc_client.py to rpc_client.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/id_map.py to id_map.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/server_state.py to server_state.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/shared_mem_utils.py to shared_mem_utils.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/dist_tensor.py to dist_tensor.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/dist_dataloader.py to dist_dataloader.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/nodeflow.py to nodeflow.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/udf.py to udf.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/view.py to view.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/kernel.py to kernel.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/spmv.py to spmv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/degree_bucketing.py to degree_bucketing.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/scheduler.py to scheduler.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/adapter.py to adapter.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/runtime.py to runtime.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/ir/program.py to program.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/ir/executor.py to executor.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/ir/var.py to var.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/ir/registry.py to registry.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/ir/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/graph.py to graph.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/frame.py to frame.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/propagate.py to propagate.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/batch.py to batch.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/transforms/functional.py to functional.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/transforms/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/transforms/module.py to module.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distgnn/partition/libra_partition.py to libra_partition.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distgnn/partition/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distgnn/tools/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distgnn/tools/tools.py to tools.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distgnn/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/runtime_ctypes.py to runtime_ctypes.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/function.py to function.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/libinfo.py to libinfo.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/streams.py to streams.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/base.py to base.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/_ctypes/function.py to function.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/_ctypes/types.py to types.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/_ctypes/ndarray.py to ndarray.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/_ctypes/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/_ctypes/object.py to object.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/ndarray.py to ndarray.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/object_generic.py to object_generic.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/object.py to object.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/_cy3/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/_cy2/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/frame.py to frame.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/random.py to random.cpython-37.pyc
creating stub loader for dgl/_ffi/_cy3/core.cpython-37m-x86_64-linux-gnu.so
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/_cy3/core.py to core.cpython-37.pyc
installing package data to build/bdist.linux-x86_64/egg
running install_data
copying ../build/libdgl.so -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/tensoradapter
creating build/bdist.linux-x86_64/egg/dgl/tensoradapter/pytorch
copying ../build/tensoradapter/pytorch/libtensoradapter_pytorch_1.9.0.so -> build/bdist.linux-x86_64/egg/dgl/tensoradapter/pytorch
creating build/bdist.linux-x86_64/egg/EGG-INFO
copying dgl.egg-info/PKG-INFO -> build/bdist.linux-x86_64/egg/EGG-INFO
copying dgl.egg-info/SOURCES.txt -> build/bdist.linux-x86_64/egg/EGG-INFO
copying dgl.egg-info/dependency_links.txt -> build/bdist.linux-x86_64/egg/EGG-INFO
copying dgl.egg-info/not-zip-safe -> build/bdist.linux-x86_64/egg/EGG-INFO
copying dgl.egg-info/requires.txt -> build/bdist.linux-x86_64/egg/EGG-INFO
copying dgl.egg-info/top_level.txt -> build/bdist.linux-x86_64/egg/EGG-INFO
writing build/bdist.linux-x86_64/egg/EGG-INFO/native_libs.txt
creating dist
creating 'dist/dgl-0.9-py3.7-linux-x86_64.egg' and adding 'build/bdist.linux-x86_64/egg' to it
removing 'build/bdist.linux-x86_64/egg' (and everything under it)
Processing dgl-0.9-py3.7-linux-x86_64.egg
creating /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/dgl-0.9-py3.7-linux-x86_64.egg
Extracting dgl-0.9-py3.7-linux-x86_64.egg to /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages
Adding dgl 0.9 to easy-install.pth file

Installed /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/dgl-0.9-py3.7-linux-x86_64.egg
Processing dependencies for dgl==0.9
Searching for psutil>=5.8.0
Reading https://pypi.org/simple/psutil/
Downloading https://files.pythonhosted.org/packages/3d/73/d8c87b5612c58d1e6c6d91997c1590771d34e4ee27d9c11eb1e64ecbf365/psutil-5.9.2-cp37-cp37m-manylinux_2_12_x86_64.manylinux2010_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl#sha256=4fb54941aac044a61db9d8eb56fc5bee207db3bc58645d657249030e15ba3727
Best match: psutil 5.9.2
Processing psutil-5.9.2-cp37-cp37m-manylinux_2_12_x86_64.manylinux2010_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl
Installing psutil-5.9.2-cp37-cp37m-manylinux_2_12_x86_64.manylinux2010_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl to /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages
Adding psutil 5.9.2 to easy-install.pth file

Installed /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/psutil-5.9.2-py3.7-linux-x86_64.egg
Searching for tqdm==4.64.0
Best match: tqdm 4.64.0
Adding tqdm 4.64.0 to easy-install.pth file
Installing tqdm script to /opt/conda/envs/pytorch-ci/bin

Using /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages
Searching for requests==2.28.1
Best match: requests 2.28.1
Adding requests 2.28.1 to easy-install.pth file

Using /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages
Searching for networkx==2.6.3
Best match: networkx 2.6.3
Adding networkx 2.6.3 to easy-install.pth file

Using /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages
Searching for scipy==1.7.3
Best match: scipy 1.7.3
Adding scipy 1.7.3 to easy-install.pth file

Using /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages
Searching for numpy==1.21.6
Best match: numpy 1.21.6
Adding numpy 1.21.6 to easy-install.pth file
Installing f2py script to /opt/conda/envs/pytorch-ci/bin
Installing f2py3 script to /opt/conda/envs/pytorch-ci/bin
Installing f2py3.7 script to /opt/conda/envs/pytorch-ci/bin

Using /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages
Searching for certifi==2022.6.15
Best match: certifi 2022.6.15
Adding certifi 2022.6.15 to easy-install.pth file

Using /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages
Searching for urllib3==1.26.11
Best match: urllib3 1.26.11
Adding urllib3 1.26.11 to easy-install.pth file

Using /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages
Searching for idna==3.3
Best match: idna 3.3
Adding idna 3.3 to easy-install.pth file

Using /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages
Searching for charset-normalizer==2.1.0
Best match: charset-normalizer 2.1.0
Adding charset-normalizer 2.1.0 to easy-install.pth file
Installing normalizer script to /opt/conda/envs/pytorch-ci/bin

Using /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages
Finished processing dependencies for dgl==0.9
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/Cython/Compiler/Main.py:369: FutureWarning: Cython directive 'language_level' not set, using 2 for now (Py2). This will change in a later release! File: /root/jenkins/workspace/dgl_PR-4648@2/python/dgl/_ffi/_cython/core.pyx
  tree = Parsing.p_module(s, pxd, full_module_name)
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/setuptools/command/install.py:37: SetuptoolsDeprecationWarning: setup.py install is deprecated. Use build and pip and other standards-based tools.
  setuptools.SetuptoolsDeprecationWarning,
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/setuptools/command/easy_install.py:147: EasyInstallDeprecationWarning: easy_install command is deprecated. Use build and pip and other standards-based tools.
  EasyInstallDeprecationWarning,
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/pkg_resources/__init__.py:125: PkgResourcesDeprecationWarning:  is an invalid version and will not be supported in a future release
  PkgResourcesDeprecationWarning,
[1/1] Cythonizing dgl/_ffi/_cython/core.pyx
running build_ext
building 'dgl._ffi._cy3.core' extension
gcc -pthread -B /opt/conda/envs/pytorch-ci/compiler_compat -Wl,--sysroot=/ -Wsign-compare -DNDEBUG -g -fwrapv -O3 -Wall -Wstrict-prototypes -fPIC -I../include/ -I../third_party/dmlc-core/include -I../third_party/dlpack/include -I/opt/conda/envs/pytorch-ci/include/python3.7m -c dgl/_ffi/_cython/core.cpp -o build/temp.linux-x86_64-3.7/dgl/_ffi/_cython/core.o
cc1plus: warning: command line option ‘-Wstrict-prototypes’ is valid for C/ObjC but not for C++
g++ -pthread -B /opt/conda/envs/pytorch-ci/compiler_compat -Wl,--sysroot=/ -pthread -shared -B /opt/conda/envs/pytorch-ci/compiler_compat -L/opt/conda/envs/pytorch-ci/lib -Wl,-rpath=/opt/conda/envs/pytorch-ci/lib -Wl,--no-as-needed -Wl,--sysroot=/ build/temp.linux-x86_64-3.7/dgl/_ffi/_cython/core.o -Ldgl -L../build/Release -L../build -ldgl -o build/lib.linux-x86_64-3.7/dgl/_ffi/_cy3/core.cpython-37m-x86_64-linux-gnu.so
/opt/conda/envs/pytorch-ci/compiler_compat/ld: warning: /opt/conda/envs/pytorch-ci/lib/libstdc++.so: unsupported GNU_PROPERTY_TYPE (5) type: 0xc0010001
/opt/conda/envs/pytorch-ci/compiler_compat/ld: warning: /opt/conda/envs/pytorch-ci/lib/libstdc++.so: unsupported GNU_PROPERTY_TYPE (5) type: 0xc0010002
/opt/conda/envs/pytorch-ci/compiler_compat/ld: warning: /opt/conda/envs/pytorch-ci/lib/libgcc_s.so.1: unsupported GNU_PROPERTY_TYPE (5) type: 0xc0010001
/opt/conda/envs/pytorch-ci/compiler_compat/ld: warning: /opt/conda/envs/pytorch-ci/lib/libgcc_s.so.1: unsupported GNU_PROPERTY_TYPE (5) type: 0xc0010002
/opt/conda/envs/pytorch-ci/compiler_compat/ld: warning: /opt/conda/envs/pytorch-ci/lib/libgcc_s.so.1: unsupported GNU_PROPERTY_TYPE (5) type: 0xc0010001
/opt/conda/envs/pytorch-ci/compiler_compat/ld: warning: /opt/conda/envs/pytorch-ci/lib/libgcc_s.so.1: unsupported GNU_PROPERTY_TYPE (5) type: 0xc0010002
copying build/lib.linux-x86_64-3.7/dgl/_ffi/_cy3/core.cpython-37m-x86_64-linux-gnu.so -> dgl/_ffi/_cy3
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/Cython/Compiler/Main.py:369: FutureWarning: Cython directive 'language_level' not set, using 2 for now (Py2). This will change in a later release! File: /root/jenkins/workspace/dgl_PR-4648@2/python/dgl/_ffi/_cython/core.pyx
  tree = Parsing.p_module(s, pxd, full_module_name)
WARNING: Skipping dgl as it is not installed.
WARNING: Running pip as the 'root' user can result in broken permissions and conflicting behaviour with the system package manager. It is recommended to use a virtual environment instead: https://pip.pypa.io/warnings/venv
[1/1] Cythonizing dgl/_ffi/_cython/core.pyx
running install
running bdist_egg
running egg_info
creating dgl.egg-info
writing dgl.egg-info/PKG-INFO
writing dependency_links to dgl.egg-info/dependency_links.txt
writing requirements to dgl.egg-info/requires.txt
writing top-level names to dgl.egg-info/top_level.txt
writing manifest file 'dgl.egg-info/SOURCES.txt'
reading manifest file 'dgl.egg-info/SOURCES.txt'
writing manifest file 'dgl.egg-info/SOURCES.txt'
installing library code to build/bdist.linux-x86_64/egg
running install_lib
running build_py
creating build
creating build/lib.linux-x86_64-3.7
creating build/lib.linux-x86_64-3.7/dgl
copying dgl/partition.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/core.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/subgraph.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/traversal.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/base.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/heterograph_index.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/sparse.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/global_config.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/_api_internal.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/heterograph.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/network.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/logging.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/graph_index.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/udf.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/ndarray.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/init.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/view.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/generators.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/convert.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/container.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/merge.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/readout.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/__init__.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/propagate.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/batch.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/frame.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/random.py -> build/lib.linux-x86_64-3.7/dgl
creating build/lib.linux-x86_64-3.7/dgl/mock_sparse
copying dgl/mock_sparse/sp_matrix.py -> build/lib.linux-x86_64-3.7/dgl/mock_sparse
copying dgl/mock_sparse/diag_matrix.py -> build/lib.linux-x86_64-3.7/dgl/mock_sparse
copying dgl/mock_sparse/reduction.py -> build/lib.linux-x86_64-3.7/dgl/mock_sparse
copying dgl/mock_sparse/__init__.py -> build/lib.linux-x86_64-3.7/dgl/mock_sparse
copying dgl/mock_sparse/elementwise_op_sp.py -> build/lib.linux-x86_64-3.7/dgl/mock_sparse
creating build/lib.linux-x86_64-3.7/dgl/geometry
copying dgl/geometry/edge_coarsening.py -> build/lib.linux-x86_64-3.7/dgl/geometry
copying dgl/geometry/fps.py -> build/lib.linux-x86_64-3.7/dgl/geometry
copying dgl/geometry/capi.py -> build/lib.linux-x86_64-3.7/dgl/geometry
copying dgl/geometry/__init__.py -> build/lib.linux-x86_64-3.7/dgl/geometry
creating build/lib.linux-x86_64-3.7/dgl/function
copying dgl/function/base.py -> build/lib.linux-x86_64-3.7/dgl/function
copying dgl/function/message.py -> build/lib.linux-x86_64-3.7/dgl/function
copying dgl/function/reducer.py -> build/lib.linux-x86_64-3.7/dgl/function
copying dgl/function/__init__.py -> build/lib.linux-x86_64-3.7/dgl/function
creating build/lib.linux-x86_64-3.7/dgl/nn
copying dgl/nn/__init__.py -> build/lib.linux-x86_64-3.7/dgl/nn
creating build/lib.linux-x86_64-3.7/dgl/contrib
copying dgl/contrib/dis_kvstore.py -> build/lib.linux-x86_64-3.7/dgl/contrib
copying dgl/contrib/unified_tensor.py -> build/lib.linux-x86_64-3.7/dgl/contrib
copying dgl/contrib/graph_store.py -> build/lib.linux-x86_64-3.7/dgl/contrib
copying dgl/contrib/__init__.py -> build/lib.linux-x86_64-3.7/dgl/contrib
creating build/lib.linux-x86_64-3.7/dgl/utils
copying dgl/utils/checks.py -> build/lib.linux-x86_64-3.7/dgl/utils
copying dgl/utils/exception.py -> build/lib.linux-x86_64-3.7/dgl/utils
copying dgl/utils/pin_memory.py -> build/lib.linux-x86_64-3.7/dgl/utils
copying dgl/utils/internal.py -> build/lib.linux-x86_64-3.7/dgl/utils
copying dgl/utils/filter.py -> build/lib.linux-x86_64-3.7/dgl/utils
copying dgl/utils/data.py -> build/lib.linux-x86_64-3.7/dgl/utils
copying dgl/utils/__init__.py -> build/lib.linux-x86_64-3.7/dgl/utils
copying dgl/utils/shared_mem.py -> build/lib.linux-x86_64-3.7/dgl/utils
creating build/lib.linux-x86_64-3.7/dgl/multiprocessing
copying dgl/multiprocessing/pytorch.py -> build/lib.linux-x86_64-3.7/dgl/multiprocessing
copying dgl/multiprocessing/__init__.py -> build/lib.linux-x86_64-3.7/dgl/multiprocessing
creating build/lib.linux-x86_64-3.7/dgl/optim
copying dgl/optim/__init__.py -> build/lib.linux-x86_64-3.7/dgl/optim
creating build/lib.linux-x86_64-3.7/dgl/dataloading
copying dgl/dataloading/base.py -> build/lib.linux-x86_64-3.7/dgl/dataloading
copying dgl/dataloading/dataloader.py -> build/lib.linux-x86_64-3.7/dgl/dataloading
copying dgl/dataloading/graphsaint.py -> build/lib.linux-x86_64-3.7/dgl/dataloading
copying dgl/dataloading/shadow.py -> build/lib.linux-x86_64-3.7/dgl/dataloading
copying dgl/dataloading/negative_sampler.py -> build/lib.linux-x86_64-3.7/dgl/dataloading
copying dgl/dataloading/__init__.py -> build/lib.linux-x86_64-3.7/dgl/dataloading
copying dgl/dataloading/cluster_gcn.py -> build/lib.linux-x86_64-3.7/dgl/dataloading
copying dgl/dataloading/neighbor_sampler.py -> build/lib.linux-x86_64-3.7/dgl/dataloading
copying dgl/dataloading/dist_dataloader.py -> build/lib.linux-x86_64-3.7/dgl/dataloading
creating build/lib.linux-x86_64-3.7/dgl/_dataloading
copying dgl/_dataloading/neighbor.py -> build/lib.linux-x86_64-3.7/dgl/_dataloading
copying dgl/_dataloading/dataloader.py -> build/lib.linux-x86_64-3.7/dgl/_dataloading
copying dgl/_dataloading/shadow.py -> build/lib.linux-x86_64-3.7/dgl/_dataloading
copying dgl/_dataloading/negative_sampler.py -> build/lib.linux-x86_64-3.7/dgl/_dataloading
copying dgl/_dataloading/__init__.py -> build/lib.linux-x86_64-3.7/dgl/_dataloading
copying dgl/_dataloading/cluster_gcn.py -> build/lib.linux-x86_64-3.7/dgl/_dataloading
creating build/lib.linux-x86_64-3.7/dgl/backend
copying dgl/backend/backend.py -> build/lib.linux-x86_64-3.7/dgl/backend
copying dgl/backend/__init__.py -> build/lib.linux-x86_64-3.7/dgl/backend
copying dgl/backend/set_default_backend.py -> build/lib.linux-x86_64-3.7/dgl/backend
creating build/lib.linux-x86_64-3.7/dgl/ops
copying dgl/ops/edge_softmax.py -> build/lib.linux-x86_64-3.7/dgl/ops
copying dgl/ops/sddmm.py -> build/lib.linux-x86_64-3.7/dgl/ops
copying dgl/ops/gather_mm.py -> build/lib.linux-x86_64-3.7/dgl/ops
copying dgl/ops/segment.py -> build/lib.linux-x86_64-3.7/dgl/ops
copying dgl/ops/__init__.py -> build/lib.linux-x86_64-3.7/dgl/ops
copying dgl/ops/spmm.py -> build/lib.linux-x86_64-3.7/dgl/ops
creating build/lib.linux-x86_64-3.7/dgl/cuda
copying dgl/cuda/nccl.py -> build/lib.linux-x86_64-3.7/dgl/cuda
copying dgl/cuda/__init__.py -> build/lib.linux-x86_64-3.7/dgl/cuda
creating build/lib.linux-x86_64-3.7/dgl/storages
copying dgl/storages/pytorch_tensor.py -> build/lib.linux-x86_64-3.7/dgl/storages
copying dgl/storages/base.py -> build/lib.linux-x86_64-3.7/dgl/storages
copying dgl/storages/tensor.py -> build/lib.linux-x86_64-3.7/dgl/storages
copying dgl/storages/__init__.py -> build/lib.linux-x86_64-3.7/dgl/storages
copying dgl/storages/numpy.py -> build/lib.linux-x86_64-3.7/dgl/storages
creating build/lib.linux-x86_64-3.7/dgl/sampling
copying dgl/sampling/neighbor.py -> build/lib.linux-x86_64-3.7/dgl/sampling
copying dgl/sampling/pinsage.py -> build/lib.linux-x86_64-3.7/dgl/sampling
copying dgl/sampling/negative.py -> build/lib.linux-x86_64-3.7/dgl/sampling
copying dgl/sampling/randomwalks.py -> build/lib.linux-x86_64-3.7/dgl/sampling
copying dgl/sampling/utils.py -> build/lib.linux-x86_64-3.7/dgl/sampling
copying dgl/sampling/node2vec_randomwalk.py -> build/lib.linux-x86_64-3.7/dgl/sampling
copying dgl/sampling/__init__.py -> build/lib.linux-x86_64-3.7/dgl/sampling
creating build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/qm7b.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/graph_serialize.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/dgl_dataset.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/csv_dataset_base.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/knowledge_graph.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/qm9_edge.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/gnn_benchmark.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/qm9.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/icews18.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/gdelt.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/utils.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/sbm.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/csv_dataset.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/karate.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/citation_graph.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/rdf.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/bitcoinotc.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/ppi.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/tensor_serialize.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/flickr.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/wikics.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/heterograph_serialize.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/tu.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/__init__.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/gindt.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/tree.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/synthetic.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/adapter.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/fakenews.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/fraud.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/minigc.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/yelp.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/reddit.py -> build/lib.linux-x86_64-3.7/dgl/data
creating build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/dist_graph.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/graph_partition_book.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/partition.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/standalone_kvstore.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/graph_services.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/role.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/constants.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/rpc_server.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/kvstore.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/dist_context.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/rpc.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/rpc_client.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/id_map.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/server_state.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/__init__.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/shared_mem_utils.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/dist_tensor.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/dist_dataloader.py -> build/lib.linux-x86_64-3.7/dgl/distributed
creating build/lib.linux-x86_64-3.7/dgl/_deprecate
copying dgl/_deprecate/nodeflow.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate
copying dgl/_deprecate/udf.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate
copying dgl/_deprecate/view.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate
copying dgl/_deprecate/kernel.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate
copying dgl/_deprecate/__init__.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate
copying dgl/_deprecate/graph.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate
copying dgl/_deprecate/frame.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate
creating build/lib.linux-x86_64-3.7/dgl/transforms
copying dgl/transforms/functional.py -> build/lib.linux-x86_64-3.7/dgl/transforms
copying dgl/transforms/__init__.py -> build/lib.linux-x86_64-3.7/dgl/transforms
copying dgl/transforms/module.py -> build/lib.linux-x86_64-3.7/dgl/transforms
creating build/lib.linux-x86_64-3.7/dgl/distgnn
copying dgl/distgnn/__init__.py -> build/lib.linux-x86_64-3.7/dgl/distgnn
creating build/lib.linux-x86_64-3.7/dgl/_ffi
copying dgl/_ffi/runtime_ctypes.py -> build/lib.linux-x86_64-3.7/dgl/_ffi
copying dgl/_ffi/function.py -> build/lib.linux-x86_64-3.7/dgl/_ffi
copying dgl/_ffi/libinfo.py -> build/lib.linux-x86_64-3.7/dgl/_ffi
copying dgl/_ffi/streams.py -> build/lib.linux-x86_64-3.7/dgl/_ffi
copying dgl/_ffi/base.py -> build/lib.linux-x86_64-3.7/dgl/_ffi
copying dgl/_ffi/ndarray.py -> build/lib.linux-x86_64-3.7/dgl/_ffi
copying dgl/_ffi/object_generic.py -> build/lib.linux-x86_64-3.7/dgl/_ffi
copying dgl/_ffi/__init__.py -> build/lib.linux-x86_64-3.7/dgl/_ffi
copying dgl/_ffi/object.py -> build/lib.linux-x86_64-3.7/dgl/_ffi
creating build/lib.linux-x86_64-3.7/dgl/nn/pytorch
copying dgl/nn/pytorch/glob.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch
copying dgl/nn/pytorch/hetero.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch
copying dgl/nn/pytorch/sparse_emb.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch
copying dgl/nn/pytorch/linear.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch
copying dgl/nn/pytorch/utils.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch
copying dgl/nn/pytorch/__init__.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch
copying dgl/nn/pytorch/factory.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch
copying dgl/nn/pytorch/softmax.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch
creating build/lib.linux-x86_64-3.7/dgl/nn/mxnet
copying dgl/nn/mxnet/glob.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet
copying dgl/nn/mxnet/hetero.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet
copying dgl/nn/mxnet/utils.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet
copying dgl/nn/mxnet/__init__.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet
copying dgl/nn/mxnet/softmax.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet
creating build/lib.linux-x86_64-3.7/dgl/nn/tensorflow
copying dgl/nn/tensorflow/glob.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow
copying dgl/nn/tensorflow/hetero.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow
copying dgl/nn/tensorflow/utils.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow
copying dgl/nn/tensorflow/__init__.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow
copying dgl/nn/tensorflow/softmax.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow
creating build/lib.linux-x86_64-3.7/dgl/nn/functional
copying dgl/nn/functional/__init__.py -> build/lib.linux-x86_64-3.7/dgl/nn/functional
creating build/lib.linux-x86_64-3.7/dgl/nn/pytorch/link
copying dgl/nn/pytorch/link/transe.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/link
copying dgl/nn/pytorch/link/edgepred.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/link
copying dgl/nn/pytorch/link/transr.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/link
copying dgl/nn/pytorch/link/__init__.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/link
creating build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/agnnconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/relgraphconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/appnpconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/dgnconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/grouprevres.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/twirlsconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/egatconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/graphconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/densechebconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/gcn2conv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/ginconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/dotgatconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/cfconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/gmmconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/sageconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/densesageconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/gineconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/edgeconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/nnconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/__init__.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/gatv2conv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/hgtconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/pnaconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/gatedgraphconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/atomicconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/tagconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/densegraphconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/egnnconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/chebconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/sgconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/gatconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
creating build/lib.linux-x86_64-3.7/dgl/nn/pytorch/explain
copying dgl/nn/pytorch/explain/gnnexplainer.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/explain
copying dgl/nn/pytorch/explain/__init__.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/explain
creating build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/agnnconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/relgraphconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/appnpconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/graphconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/densechebconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/ginconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/gmmconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/sageconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/densesageconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/edgeconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/nnconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/__init__.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/gatedgraphconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/tagconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/densegraphconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/chebconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/sgconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/gatconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
creating build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv
copying dgl/nn/tensorflow/conv/relgraphconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv
copying dgl/nn/tensorflow/conv/appnpconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv
copying dgl/nn/tensorflow/conv/graphconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv
copying dgl/nn/tensorflow/conv/densechebconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv
copying dgl/nn/tensorflow/conv/ginconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv
copying dgl/nn/tensorflow/conv/sageconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv
copying dgl/nn/tensorflow/conv/edgeconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv
copying dgl/nn/tensorflow/conv/__init__.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv
copying dgl/nn/tensorflow/conv/chebconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv
copying dgl/nn/tensorflow/conv/sgconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv
copying dgl/nn/tensorflow/conv/gatconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv
creating build/lib.linux-x86_64-3.7/dgl/contrib/sampling
copying dgl/contrib/sampling/sampler.py -> build/lib.linux-x86_64-3.7/dgl/contrib/sampling
copying dgl/contrib/sampling/dis_sampler.py -> build/lib.linux-x86_64-3.7/dgl/contrib/sampling
copying dgl/contrib/sampling/__init__.py -> build/lib.linux-x86_64-3.7/dgl/contrib/sampling
creating build/lib.linux-x86_64-3.7/dgl/contrib/data
copying dgl/contrib/data/knowledge_graph.py -> build/lib.linux-x86_64-3.7/dgl/contrib/data
copying dgl/contrib/data/__init__.py -> build/lib.linux-x86_64-3.7/dgl/contrib/data
creating build/lib.linux-x86_64-3.7/dgl/optim/pytorch
copying dgl/optim/pytorch/__init__.py -> build/lib.linux-x86_64-3.7/dgl/optim/pytorch
copying dgl/optim/pytorch/sparse_optim.py -> build/lib.linux-x86_64-3.7/dgl/optim/pytorch
creating build/lib.linux-x86_64-3.7/dgl/optim/mxnet
copying dgl/optim/mxnet/__init__.py -> build/lib.linux-x86_64-3.7/dgl/optim/mxnet
creating build/lib.linux-x86_64-3.7/dgl/optim/tensorflow
copying dgl/optim/tensorflow/__init__.py -> build/lib.linux-x86_64-3.7/dgl/optim/tensorflow
creating build/lib.linux-x86_64-3.7/dgl/_dataloading/pytorch
copying dgl/_dataloading/pytorch/dataloader.py -> build/lib.linux-x86_64-3.7/dgl/_dataloading/pytorch
copying dgl/_dataloading/pytorch/__init__.py -> build/lib.linux-x86_64-3.7/dgl/_dataloading/pytorch
creating build/lib.linux-x86_64-3.7/dgl/backend/pytorch
copying dgl/backend/pytorch/sparse.py -> build/lib.linux-x86_64-3.7/dgl/backend/pytorch
copying dgl/backend/pytorch/tensor.py -> build/lib.linux-x86_64-3.7/dgl/backend/pytorch
copying dgl/backend/pytorch/__init__.py -> build/lib.linux-x86_64-3.7/dgl/backend/pytorch
creating build/lib.linux-x86_64-3.7/dgl/backend/mxnet
copying dgl/backend/mxnet/sparse.py -> build/lib.linux-x86_64-3.7/dgl/backend/mxnet
copying dgl/backend/mxnet/tensor.py -> build/lib.linux-x86_64-3.7/dgl/backend/mxnet
copying dgl/backend/mxnet/__init__.py -> build/lib.linux-x86_64-3.7/dgl/backend/mxnet
copying dgl/backend/mxnet/sparse_optim.py -> build/lib.linux-x86_64-3.7/dgl/backend/mxnet
creating build/lib.linux-x86_64-3.7/dgl/backend/tensorflow
copying dgl/backend/tensorflow/sparse.py -> build/lib.linux-x86_64-3.7/dgl/backend/tensorflow
copying dgl/backend/tensorflow/tensor.py -> build/lib.linux-x86_64-3.7/dgl/backend/tensorflow
copying dgl/backend/tensorflow/__init__.py -> build/lib.linux-x86_64-3.7/dgl/backend/tensorflow
copying dgl/backend/tensorflow/sparse_optim.py -> build/lib.linux-x86_64-3.7/dgl/backend/tensorflow
creating build/lib.linux-x86_64-3.7/dgl/distributed/nn
copying dgl/distributed/nn/__init__.py -> build/lib.linux-x86_64-3.7/dgl/distributed/nn
creating build/lib.linux-x86_64-3.7/dgl/distributed/optim
copying dgl/distributed/optim/__init__.py -> build/lib.linux-x86_64-3.7/dgl/distributed/optim
creating build/lib.linux-x86_64-3.7/dgl/distributed/nn/pytorch
copying dgl/distributed/nn/pytorch/sparse_emb.py -> build/lib.linux-x86_64-3.7/dgl/distributed/nn/pytorch
copying dgl/distributed/nn/pytorch/__init__.py -> build/lib.linux-x86_64-3.7/dgl/distributed/nn/pytorch
creating build/lib.linux-x86_64-3.7/dgl/distributed/nn/mxnet
copying dgl/distributed/nn/mxnet/__init__.py -> build/lib.linux-x86_64-3.7/dgl/distributed/nn/mxnet
creating build/lib.linux-x86_64-3.7/dgl/distributed/nn/tensorflow
copying dgl/distributed/nn/tensorflow/__init__.py -> build/lib.linux-x86_64-3.7/dgl/distributed/nn/tensorflow
creating build/lib.linux-x86_64-3.7/dgl/distributed/optim/pytorch
copying dgl/distributed/optim/pytorch/utils.py -> build/lib.linux-x86_64-3.7/dgl/distributed/optim/pytorch
copying dgl/distributed/optim/pytorch/__init__.py -> build/lib.linux-x86_64-3.7/dgl/distributed/optim/pytorch
copying dgl/distributed/optim/pytorch/sparse_optim.py -> build/lib.linux-x86_64-3.7/dgl/distributed/optim/pytorch
creating build/lib.linux-x86_64-3.7/dgl/distributed/optim/mxnet
copying dgl/distributed/optim/mxnet/__init__.py -> build/lib.linux-x86_64-3.7/dgl/distributed/optim/mxnet
creating build/lib.linux-x86_64-3.7/dgl/distributed/optim/tensorflow
copying dgl/distributed/optim/tensorflow/__init__.py -> build/lib.linux-x86_64-3.7/dgl/distributed/optim/tensorflow
creating build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime
copying dgl/_deprecate/runtime/spmv.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime
copying dgl/_deprecate/runtime/degree_bucketing.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime
copying dgl/_deprecate/runtime/__init__.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime
copying dgl/_deprecate/runtime/scheduler.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime
copying dgl/_deprecate/runtime/adapter.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime
copying dgl/_deprecate/runtime/runtime.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime
creating build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/ir
copying dgl/_deprecate/runtime/ir/program.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/ir
copying dgl/_deprecate/runtime/ir/executor.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/ir
copying dgl/_deprecate/runtime/ir/var.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/ir
copying dgl/_deprecate/runtime/ir/registry.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/ir
copying dgl/_deprecate/runtime/ir/__init__.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/ir
creating build/lib.linux-x86_64-3.7/dgl/distgnn/partition
copying dgl/distgnn/partition/libra_partition.py -> build/lib.linux-x86_64-3.7/dgl/distgnn/partition
copying dgl/distgnn/partition/__init__.py -> build/lib.linux-x86_64-3.7/dgl/distgnn/partition
creating build/lib.linux-x86_64-3.7/dgl/distgnn/tools
copying dgl/distgnn/tools/__init__.py -> build/lib.linux-x86_64-3.7/dgl/distgnn/tools
copying dgl/distgnn/tools/tools.py -> build/lib.linux-x86_64-3.7/dgl/distgnn/tools
creating build/lib.linux-x86_64-3.7/dgl/_ffi/_ctypes
copying dgl/_ffi/_ctypes/function.py -> build/lib.linux-x86_64-3.7/dgl/_ffi/_ctypes
copying dgl/_ffi/_ctypes/types.py -> build/lib.linux-x86_64-3.7/dgl/_ffi/_ctypes
copying dgl/_ffi/_ctypes/ndarray.py -> build/lib.linux-x86_64-3.7/dgl/_ffi/_ctypes
copying dgl/_ffi/_ctypes/__init__.py -> build/lib.linux-x86_64-3.7/dgl/_ffi/_ctypes
copying dgl/_ffi/_ctypes/object.py -> build/lib.linux-x86_64-3.7/dgl/_ffi/_ctypes
creating build/lib.linux-x86_64-3.7/dgl/_ffi/_cy3
copying dgl/_ffi/_cy3/__init__.py -> build/lib.linux-x86_64-3.7/dgl/_ffi/_cy3
creating build/lib.linux-x86_64-3.7/dgl/_ffi/_cy2
copying dgl/_ffi/_cy2/__init__.py -> build/lib.linux-x86_64-3.7/dgl/_ffi/_cy2
creating build/lib.linux-x86_64-3.7/dgl/_ffi/_cython
copying dgl/_ffi/_cython/core.cpp -> build/lib.linux-x86_64-3.7/dgl/_ffi/_cython
running build_ext
building 'dgl._ffi._cy3.core' extension
creating build/temp.linux-x86_64-3.7
creating build/temp.linux-x86_64-3.7/dgl
creating build/temp.linux-x86_64-3.7/dgl/_ffi
creating build/temp.linux-x86_64-3.7/dgl/_ffi/_cython
gcc -pthread -B /opt/conda/envs/mxnet-ci/compiler_compat -Wl,--sysroot=/ -Wsign-compare -DNDEBUG -g -fwrapv -O3 -Wall -Wstrict-prototypes -fPIC -I../include/ -I../third_party/dmlc-core/include -I../third_party/dlpack/include -I/opt/conda/envs/mxnet-ci/include/python3.7m -c dgl/_ffi/_cython/core.cpp -o build/temp.linux-x86_64-3.7/dgl/_ffi/_cython/core.o
cc1plus: warning: command line option ‘-Wstrict-prototypes’ is valid for C/ObjC but not for C++
g++ -pthread -B /opt/conda/envs/mxnet-ci/compiler_compat -Wl,--sysroot=/ -pthread -shared -B /opt/conda/envs/mxnet-ci/compiler_compat -L/opt/conda/envs/mxnet-ci/lib -Wl,-rpath=/opt/conda/envs/mxnet-ci/lib -Wl,--no-as-needed -Wl,--sysroot=/ build/temp.linux-x86_64-3.7/dgl/_ffi/_cython/core.o -Ldgl -L../build/Release -L../build -ldgl -o build/lib.linux-x86_64-3.7/dgl/_ffi/_cy3/core.cpython-37m-x86_64-linux-gnu.so
/opt/conda/envs/mxnet-ci/compiler_compat/ld: warning: /opt/conda/envs/mxnet-ci/lib/libstdc++.so: unsupported GNU_PROPERTY_TYPE (5) type: 0xc0010001
/opt/conda/envs/mxnet-ci/compiler_compat/ld: warning: /opt/conda/envs/mxnet-ci/lib/libstdc++.so: unsupported GNU_PROPERTY_TYPE (5) type: 0xc0010002
/opt/conda/envs/mxnet-ci/compiler_compat/ld: warning: /opt/conda/envs/mxnet-ci/lib/libgcc_s.so.1: unsupported GNU_PROPERTY_TYPE (5) type: 0xc0010001
/opt/conda/envs/mxnet-ci/compiler_compat/ld: warning: /opt/conda/envs/mxnet-ci/lib/libgcc_s.so.1: unsupported GNU_PROPERTY_TYPE (5) type: 0xc0010002
/opt/conda/envs/mxnet-ci/compiler_compat/ld: warning: /opt/conda/envs/mxnet-ci/lib/libgcc_s.so.1: unsupported GNU_PROPERTY_TYPE (5) type: 0xc0010001
/opt/conda/envs/mxnet-ci/compiler_compat/ld: warning: /opt/conda/envs/mxnet-ci/lib/libgcc_s.so.1: unsupported GNU_PROPERTY_TYPE (5) type: 0xc0010002
creating build/bdist.linux-x86_64
creating build/bdist.linux-x86_64/egg
creating build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/mock_sparse
copying build/lib.linux-x86_64-3.7/dgl/mock_sparse/sp_matrix.py -> build/bdist.linux-x86_64/egg/dgl/mock_sparse
copying build/lib.linux-x86_64-3.7/dgl/mock_sparse/diag_matrix.py -> build/bdist.linux-x86_64/egg/dgl/mock_sparse
copying build/lib.linux-x86_64-3.7/dgl/mock_sparse/reduction.py -> build/bdist.linux-x86_64/egg/dgl/mock_sparse
copying build/lib.linux-x86_64-3.7/dgl/mock_sparse/__init__.py -> build/bdist.linux-x86_64/egg/dgl/mock_sparse
copying build/lib.linux-x86_64-3.7/dgl/mock_sparse/elementwise_op_sp.py -> build/bdist.linux-x86_64/egg/dgl/mock_sparse
creating build/bdist.linux-x86_64/egg/dgl/geometry
copying build/lib.linux-x86_64-3.7/dgl/geometry/edge_coarsening.py -> build/bdist.linux-x86_64/egg/dgl/geometry
copying build/lib.linux-x86_64-3.7/dgl/geometry/fps.py -> build/bdist.linux-x86_64/egg/dgl/geometry
copying build/lib.linux-x86_64-3.7/dgl/geometry/capi.py -> build/bdist.linux-x86_64/egg/dgl/geometry
copying build/lib.linux-x86_64-3.7/dgl/geometry/__init__.py -> build/bdist.linux-x86_64/egg/dgl/geometry
creating build/bdist.linux-x86_64/egg/dgl/function
copying build/lib.linux-x86_64-3.7/dgl/function/base.py -> build/bdist.linux-x86_64/egg/dgl/function
copying build/lib.linux-x86_64-3.7/dgl/function/message.py -> build/bdist.linux-x86_64/egg/dgl/function
copying build/lib.linux-x86_64-3.7/dgl/function/reducer.py -> build/bdist.linux-x86_64/egg/dgl/function
copying build/lib.linux-x86_64-3.7/dgl/function/__init__.py -> build/bdist.linux-x86_64/egg/dgl/function
copying build/lib.linux-x86_64-3.7/dgl/partition.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/nn
creating build/bdist.linux-x86_64/egg/dgl/nn/pytorch
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/glob.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch
creating build/bdist.linux-x86_64/egg/dgl/nn/pytorch/link
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/link/transe.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/link
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/link/edgepred.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/link
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/link/transr.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/link
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/link/__init__.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/link
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/hetero.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/sparse_emb.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/linear.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/utils.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch
creating build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/agnnconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/relgraphconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/appnpconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/dgnconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/grouprevres.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/twirlsconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/egatconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/graphconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/densechebconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/gcn2conv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/ginconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/dotgatconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/cfconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/gmmconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/sageconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/densesageconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/gineconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/edgeconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/nnconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/__init__.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/gatv2conv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/hgtconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/pnaconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/gatedgraphconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/atomicconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/tagconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/densegraphconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/egnnconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/chebconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/sgconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/gatconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/__init__.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch
creating build/bdist.linux-x86_64/egg/dgl/nn/pytorch/explain
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/explain/gnnexplainer.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/explain
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/explain/__init__.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/explain
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/factory.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/softmax.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch
creating build/bdist.linux-x86_64/egg/dgl/nn/mxnet
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/glob.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/hetero.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/utils.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet
creating build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/agnnconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/relgraphconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/appnpconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/graphconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/densechebconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/ginconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/gmmconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/sageconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/densesageconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/edgeconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/nnconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/__init__.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/gatedgraphconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/tagconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/densegraphconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/chebconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/sgconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/gatconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/__init__.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/softmax.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet
copying build/lib.linux-x86_64-3.7/dgl/nn/__init__.py -> build/bdist.linux-x86_64/egg/dgl/nn
creating build/bdist.linux-x86_64/egg/dgl/nn/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/glob.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/hetero.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/utils.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow
creating build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv/relgraphconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv/appnpconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv/graphconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv/densechebconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv/ginconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv/sageconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv/edgeconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv/__init__.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv/chebconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv/sgconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv/gatconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/__init__.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/softmax.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow
creating build/bdist.linux-x86_64/egg/dgl/nn/functional
copying build/lib.linux-x86_64-3.7/dgl/nn/functional/__init__.py -> build/bdist.linux-x86_64/egg/dgl/nn/functional
copying build/lib.linux-x86_64-3.7/dgl/core.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-3.7/dgl/subgraph.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/contrib
copying build/lib.linux-x86_64-3.7/dgl/contrib/dis_kvstore.py -> build/bdist.linux-x86_64/egg/dgl/contrib
copying build/lib.linux-x86_64-3.7/dgl/contrib/unified_tensor.py -> build/bdist.linux-x86_64/egg/dgl/contrib
copying build/lib.linux-x86_64-3.7/dgl/contrib/graph_store.py -> build/bdist.linux-x86_64/egg/dgl/contrib
creating build/bdist.linux-x86_64/egg/dgl/contrib/sampling
copying build/lib.linux-x86_64-3.7/dgl/contrib/sampling/sampler.py -> build/bdist.linux-x86_64/egg/dgl/contrib/sampling
copying build/lib.linux-x86_64-3.7/dgl/contrib/sampling/dis_sampler.py -> build/bdist.linux-x86_64/egg/dgl/contrib/sampling
copying build/lib.linux-x86_64-3.7/dgl/contrib/sampling/__init__.py -> build/bdist.linux-x86_64/egg/dgl/contrib/sampling
creating build/bdist.linux-x86_64/egg/dgl/contrib/data
copying build/lib.linux-x86_64-3.7/dgl/contrib/data/knowledge_graph.py -> build/bdist.linux-x86_64/egg/dgl/contrib/data
copying build/lib.linux-x86_64-3.7/dgl/contrib/data/__init__.py -> build/bdist.linux-x86_64/egg/dgl/contrib/data
copying build/lib.linux-x86_64-3.7/dgl/contrib/__init__.py -> build/bdist.linux-x86_64/egg/dgl/contrib
copying build/lib.linux-x86_64-3.7/dgl/traversal.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/utils
copying build/lib.linux-x86_64-3.7/dgl/utils/checks.py -> build/bdist.linux-x86_64/egg/dgl/utils
copying build/lib.linux-x86_64-3.7/dgl/utils/exception.py -> build/bdist.linux-x86_64/egg/dgl/utils
copying build/lib.linux-x86_64-3.7/dgl/utils/pin_memory.py -> build/bdist.linux-x86_64/egg/dgl/utils
copying build/lib.linux-x86_64-3.7/dgl/utils/internal.py -> build/bdist.linux-x86_64/egg/dgl/utils
copying build/lib.linux-x86_64-3.7/dgl/utils/filter.py -> build/bdist.linux-x86_64/egg/dgl/utils
copying build/lib.linux-x86_64-3.7/dgl/utils/data.py -> build/bdist.linux-x86_64/egg/dgl/utils
copying build/lib.linux-x86_64-3.7/dgl/utils/__init__.py -> build/bdist.linux-x86_64/egg/dgl/utils
copying build/lib.linux-x86_64-3.7/dgl/utils/shared_mem.py -> build/bdist.linux-x86_64/egg/dgl/utils
creating build/bdist.linux-x86_64/egg/dgl/multiprocessing
copying build/lib.linux-x86_64-3.7/dgl/multiprocessing/pytorch.py -> build/bdist.linux-x86_64/egg/dgl/multiprocessing
copying build/lib.linux-x86_64-3.7/dgl/multiprocessing/__init__.py -> build/bdist.linux-x86_64/egg/dgl/multiprocessing
copying build/lib.linux-x86_64-3.7/dgl/base.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/optim
creating build/bdist.linux-x86_64/egg/dgl/optim/pytorch
copying build/lib.linux-x86_64-3.7/dgl/optim/pytorch/__init__.py -> build/bdist.linux-x86_64/egg/dgl/optim/pytorch
copying build/lib.linux-x86_64-3.7/dgl/optim/pytorch/sparse_optim.py -> build/bdist.linux-x86_64/egg/dgl/optim/pytorch
creating build/bdist.linux-x86_64/egg/dgl/optim/mxnet
copying build/lib.linux-x86_64-3.7/dgl/optim/mxnet/__init__.py -> build/bdist.linux-x86_64/egg/dgl/optim/mxnet
copying build/lib.linux-x86_64-3.7/dgl/optim/__init__.py -> build/bdist.linux-x86_64/egg/dgl/optim
creating build/bdist.linux-x86_64/egg/dgl/optim/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/optim/tensorflow/__init__.py -> build/bdist.linux-x86_64/egg/dgl/optim/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/heterograph_index.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-3.7/dgl/sparse.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-3.7/dgl/global_config.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/dataloading
copying build/lib.linux-x86_64-3.7/dgl/dataloading/base.py -> build/bdist.linux-x86_64/egg/dgl/dataloading
copying build/lib.linux-x86_64-3.7/dgl/dataloading/dataloader.py -> build/bdist.linux-x86_64/egg/dgl/dataloading
copying build/lib.linux-x86_64-3.7/dgl/dataloading/graphsaint.py -> build/bdist.linux-x86_64/egg/dgl/dataloading
copying build/lib.linux-x86_64-3.7/dgl/dataloading/shadow.py -> build/bdist.linux-x86_64/egg/dgl/dataloading
copying build/lib.linux-x86_64-3.7/dgl/dataloading/negative_sampler.py -> build/bdist.linux-x86_64/egg/dgl/dataloading
copying build/lib.linux-x86_64-3.7/dgl/dataloading/__init__.py -> build/bdist.linux-x86_64/egg/dgl/dataloading
copying build/lib.linux-x86_64-3.7/dgl/dataloading/cluster_gcn.py -> build/bdist.linux-x86_64/egg/dgl/dataloading
copying build/lib.linux-x86_64-3.7/dgl/dataloading/neighbor_sampler.py -> build/bdist.linux-x86_64/egg/dgl/dataloading
copying build/lib.linux-x86_64-3.7/dgl/dataloading/dist_dataloader.py -> build/bdist.linux-x86_64/egg/dgl/dataloading
creating build/bdist.linux-x86_64/egg/dgl/_dataloading
copying build/lib.linux-x86_64-3.7/dgl/_dataloading/neighbor.py -> build/bdist.linux-x86_64/egg/dgl/_dataloading
copying build/lib.linux-x86_64-3.7/dgl/_dataloading/dataloader.py -> build/bdist.linux-x86_64/egg/dgl/_dataloading
creating build/bdist.linux-x86_64/egg/dgl/_dataloading/pytorch
copying build/lib.linux-x86_64-3.7/dgl/_dataloading/pytorch/dataloader.py -> build/bdist.linux-x86_64/egg/dgl/_dataloading/pytorch
copying build/lib.linux-x86_64-3.7/dgl/_dataloading/pytorch/__init__.py -> build/bdist.linux-x86_64/egg/dgl/_dataloading/pytorch
copying build/lib.linux-x86_64-3.7/dgl/_dataloading/shadow.py -> build/bdist.linux-x86_64/egg/dgl/_dataloading
copying build/lib.linux-x86_64-3.7/dgl/_dataloading/negative_sampler.py -> build/bdist.linux-x86_64/egg/dgl/_dataloading
copying build/lib.linux-x86_64-3.7/dgl/_dataloading/__init__.py -> build/bdist.linux-x86_64/egg/dgl/_dataloading
copying build/lib.linux-x86_64-3.7/dgl/_dataloading/cluster_gcn.py -> build/bdist.linux-x86_64/egg/dgl/_dataloading
copying build/lib.linux-x86_64-3.7/dgl/_api_internal.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-3.7/dgl/heterograph.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/backend
copying build/lib.linux-x86_64-3.7/dgl/backend/backend.py -> build/bdist.linux-x86_64/egg/dgl/backend
creating build/bdist.linux-x86_64/egg/dgl/backend/pytorch
copying build/lib.linux-x86_64-3.7/dgl/backend/pytorch/sparse.py -> build/bdist.linux-x86_64/egg/dgl/backend/pytorch
copying build/lib.linux-x86_64-3.7/dgl/backend/pytorch/tensor.py -> build/bdist.linux-x86_64/egg/dgl/backend/pytorch
copying build/lib.linux-x86_64-3.7/dgl/backend/pytorch/__init__.py -> build/bdist.linux-x86_64/egg/dgl/backend/pytorch
creating build/bdist.linux-x86_64/egg/dgl/backend/mxnet
copying build/lib.linux-x86_64-3.7/dgl/backend/mxnet/sparse.py -> build/bdist.linux-x86_64/egg/dgl/backend/mxnet
copying build/lib.linux-x86_64-3.7/dgl/backend/mxnet/tensor.py -> build/bdist.linux-x86_64/egg/dgl/backend/mxnet
copying build/lib.linux-x86_64-3.7/dgl/backend/mxnet/__init__.py -> build/bdist.linux-x86_64/egg/dgl/backend/mxnet
copying build/lib.linux-x86_64-3.7/dgl/backend/mxnet/sparse_optim.py -> build/bdist.linux-x86_64/egg/dgl/backend/mxnet
copying build/lib.linux-x86_64-3.7/dgl/backend/__init__.py -> build/bdist.linux-x86_64/egg/dgl/backend
creating build/bdist.linux-x86_64/egg/dgl/backend/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/backend/tensorflow/sparse.py -> build/bdist.linux-x86_64/egg/dgl/backend/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/backend/tensorflow/tensor.py -> build/bdist.linux-x86_64/egg/dgl/backend/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/backend/tensorflow/__init__.py -> build/bdist.linux-x86_64/egg/dgl/backend/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/backend/tensorflow/sparse_optim.py -> build/bdist.linux-x86_64/egg/dgl/backend/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/backend/set_default_backend.py -> build/bdist.linux-x86_64/egg/dgl/backend
copying build/lib.linux-x86_64-3.7/dgl/network.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/ops
copying build/lib.linux-x86_64-3.7/dgl/ops/edge_softmax.py -> build/bdist.linux-x86_64/egg/dgl/ops
copying build/lib.linux-x86_64-3.7/dgl/ops/sddmm.py -> build/bdist.linux-x86_64/egg/dgl/ops
copying build/lib.linux-x86_64-3.7/dgl/ops/gather_mm.py -> build/bdist.linux-x86_64/egg/dgl/ops
copying build/lib.linux-x86_64-3.7/dgl/ops/segment.py -> build/bdist.linux-x86_64/egg/dgl/ops
copying build/lib.linux-x86_64-3.7/dgl/ops/__init__.py -> build/bdist.linux-x86_64/egg/dgl/ops
copying build/lib.linux-x86_64-3.7/dgl/ops/spmm.py -> build/bdist.linux-x86_64/egg/dgl/ops
copying build/lib.linux-x86_64-3.7/dgl/logging.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-3.7/dgl/graph_index.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/cuda
copying build/lib.linux-x86_64-3.7/dgl/cuda/nccl.py -> build/bdist.linux-x86_64/egg/dgl/cuda
copying build/lib.linux-x86_64-3.7/dgl/cuda/__init__.py -> build/bdist.linux-x86_64/egg/dgl/cuda
copying build/lib.linux-x86_64-3.7/dgl/udf.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/storages
copying build/lib.linux-x86_64-3.7/dgl/storages/pytorch_tensor.py -> build/bdist.linux-x86_64/egg/dgl/storages
copying build/lib.linux-x86_64-3.7/dgl/storages/base.py -> build/bdist.linux-x86_64/egg/dgl/storages
copying build/lib.linux-x86_64-3.7/dgl/storages/tensor.py -> build/bdist.linux-x86_64/egg/dgl/storages
copying build/lib.linux-x86_64-3.7/dgl/storages/__init__.py -> build/bdist.linux-x86_64/egg/dgl/storages
copying build/lib.linux-x86_64-3.7/dgl/storages/numpy.py -> build/bdist.linux-x86_64/egg/dgl/storages
copying build/lib.linux-x86_64-3.7/dgl/ndarray.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-3.7/dgl/init.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-3.7/dgl/view.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/sampling
copying build/lib.linux-x86_64-3.7/dgl/sampling/neighbor.py -> build/bdist.linux-x86_64/egg/dgl/sampling
copying build/lib.linux-x86_64-3.7/dgl/sampling/pinsage.py -> build/bdist.linux-x86_64/egg/dgl/sampling
copying build/lib.linux-x86_64-3.7/dgl/sampling/negative.py -> build/bdist.linux-x86_64/egg/dgl/sampling
copying build/lib.linux-x86_64-3.7/dgl/sampling/randomwalks.py -> build/bdist.linux-x86_64/egg/dgl/sampling
copying build/lib.linux-x86_64-3.7/dgl/sampling/utils.py -> build/bdist.linux-x86_64/egg/dgl/sampling
copying build/lib.linux-x86_64-3.7/dgl/sampling/node2vec_randomwalk.py -> build/bdist.linux-x86_64/egg/dgl/sampling
copying build/lib.linux-x86_64-3.7/dgl/sampling/__init__.py -> build/bdist.linux-x86_64/egg/dgl/sampling
copying build/lib.linux-x86_64-3.7/dgl/generators.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/qm7b.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/graph_serialize.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/dgl_dataset.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/csv_dataset_base.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/knowledge_graph.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/qm9_edge.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/gnn_benchmark.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/qm9.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/icews18.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/gdelt.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/utils.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/sbm.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/csv_dataset.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/karate.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/citation_graph.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/rdf.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/bitcoinotc.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/ppi.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/tensor_serialize.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/flickr.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/wikics.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/heterograph_serialize.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/tu.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/__init__.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/gindt.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/tree.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/synthetic.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/adapter.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/fakenews.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/fraud.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/minigc.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/yelp.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/reddit.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/convert.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-3.7/dgl/container.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-3.7/dgl/merge.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-3.7/dgl/readout.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-3.7/dgl/__init__.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/dist_graph.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/graph_partition_book.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/partition.py -> build/bdist.linux-x86_64/egg/dgl/distributed
creating build/bdist.linux-x86_64/egg/dgl/distributed/nn
creating build/bdist.linux-x86_64/egg/dgl/distributed/nn/pytorch
copying build/lib.linux-x86_64-3.7/dgl/distributed/nn/pytorch/sparse_emb.py -> build/bdist.linux-x86_64/egg/dgl/distributed/nn/pytorch
copying build/lib.linux-x86_64-3.7/dgl/distributed/nn/pytorch/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distributed/nn/pytorch
creating build/bdist.linux-x86_64/egg/dgl/distributed/nn/mxnet
copying build/lib.linux-x86_64-3.7/dgl/distributed/nn/mxnet/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distributed/nn/mxnet
copying build/lib.linux-x86_64-3.7/dgl/distributed/nn/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distributed/nn
creating build/bdist.linux-x86_64/egg/dgl/distributed/nn/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/distributed/nn/tensorflow/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distributed/nn/tensorflow
creating build/bdist.linux-x86_64/egg/dgl/distributed/optim
creating build/bdist.linux-x86_64/egg/dgl/distributed/optim/pytorch
copying build/lib.linux-x86_64-3.7/dgl/distributed/optim/pytorch/utils.py -> build/bdist.linux-x86_64/egg/dgl/distributed/optim/pytorch
copying build/lib.linux-x86_64-3.7/dgl/distributed/optim/pytorch/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distributed/optim/pytorch
copying build/lib.linux-x86_64-3.7/dgl/distributed/optim/pytorch/sparse_optim.py -> build/bdist.linux-x86_64/egg/dgl/distributed/optim/pytorch
creating build/bdist.linux-x86_64/egg/dgl/distributed/optim/mxnet
copying build/lib.linux-x86_64-3.7/dgl/distributed/optim/mxnet/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distributed/optim/mxnet
copying build/lib.linux-x86_64-3.7/dgl/distributed/optim/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distributed/optim
creating build/bdist.linux-x86_64/egg/dgl/distributed/optim/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/distributed/optim/tensorflow/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distributed/optim/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/distributed/standalone_kvstore.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/graph_services.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/role.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/constants.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/rpc_server.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/kvstore.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/dist_context.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/rpc.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/rpc_client.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/id_map.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/server_state.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/shared_mem_utils.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/dist_tensor.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/dist_dataloader.py -> build/bdist.linux-x86_64/egg/dgl/distributed
creating build/bdist.linux-x86_64/egg/dgl/_deprecate
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/nodeflow.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/udf.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/view.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/kernel.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/__init__.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate
creating build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/spmv.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/degree_bucketing.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/__init__.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/scheduler.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/adapter.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/runtime.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime
creating build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/ir
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/ir/program.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/ir
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/ir/executor.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/ir
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/ir/var.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/ir
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/ir/registry.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/ir
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/ir/__init__.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/ir
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/graph.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/frame.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate
copying build/lib.linux-x86_64-3.7/dgl/propagate.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-3.7/dgl/batch.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/transforms
copying build/lib.linux-x86_64-3.7/dgl/transforms/functional.py -> build/bdist.linux-x86_64/egg/dgl/transforms
copying build/lib.linux-x86_64-3.7/dgl/transforms/__init__.py -> build/bdist.linux-x86_64/egg/dgl/transforms
copying build/lib.linux-x86_64-3.7/dgl/transforms/module.py -> build/bdist.linux-x86_64/egg/dgl/transforms
creating build/bdist.linux-x86_64/egg/dgl/distgnn
creating build/bdist.linux-x86_64/egg/dgl/distgnn/partition
copying build/lib.linux-x86_64-3.7/dgl/distgnn/partition/libra_partition.py -> build/bdist.linux-x86_64/egg/dgl/distgnn/partition
copying build/lib.linux-x86_64-3.7/dgl/distgnn/partition/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distgnn/partition
creating build/bdist.linux-x86_64/egg/dgl/distgnn/tools
copying build/lib.linux-x86_64-3.7/dgl/distgnn/tools/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distgnn/tools
copying build/lib.linux-x86_64-3.7/dgl/distgnn/tools/tools.py -> build/bdist.linux-x86_64/egg/dgl/distgnn/tools
copying build/lib.linux-x86_64-3.7/dgl/distgnn/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distgnn
creating build/bdist.linux-x86_64/egg/dgl/_ffi
copying build/lib.linux-x86_64-3.7/dgl/_ffi/runtime_ctypes.py -> build/bdist.linux-x86_64/egg/dgl/_ffi
copying build/lib.linux-x86_64-3.7/dgl/_ffi/function.py -> build/bdist.linux-x86_64/egg/dgl/_ffi
copying build/lib.linux-x86_64-3.7/dgl/_ffi/libinfo.py -> build/bdist.linux-x86_64/egg/dgl/_ffi
copying build/lib.linux-x86_64-3.7/dgl/_ffi/streams.py -> build/bdist.linux-x86_64/egg/dgl/_ffi
copying build/lib.linux-x86_64-3.7/dgl/_ffi/base.py -> build/bdist.linux-x86_64/egg/dgl/_ffi
creating build/bdist.linux-x86_64/egg/dgl/_ffi/_ctypes
copying build/lib.linux-x86_64-3.7/dgl/_ffi/_ctypes/function.py -> build/bdist.linux-x86_64/egg/dgl/_ffi/_ctypes
copying build/lib.linux-x86_64-3.7/dgl/_ffi/_ctypes/types.py -> build/bdist.linux-x86_64/egg/dgl/_ffi/_ctypes
copying build/lib.linux-x86_64-3.7/dgl/_ffi/_ctypes/ndarray.py -> build/bdist.linux-x86_64/egg/dgl/_ffi/_ctypes
copying build/lib.linux-x86_64-3.7/dgl/_ffi/_ctypes/__init__.py -> build/bdist.linux-x86_64/egg/dgl/_ffi/_ctypes
copying build/lib.linux-x86_64-3.7/dgl/_ffi/_ctypes/object.py -> build/bdist.linux-x86_64/egg/dgl/_ffi/_ctypes
copying build/lib.linux-x86_64-3.7/dgl/_ffi/ndarray.py -> build/bdist.linux-x86_64/egg/dgl/_ffi
copying build/lib.linux-x86_64-3.7/dgl/_ffi/object_generic.py -> build/bdist.linux-x86_64/egg/dgl/_ffi
copying build/lib.linux-x86_64-3.7/dgl/_ffi/__init__.py -> build/bdist.linux-x86_64/egg/dgl/_ffi
copying build/lib.linux-x86_64-3.7/dgl/_ffi/object.py -> build/bdist.linux-x86_64/egg/dgl/_ffi
creating build/bdist.linux-x86_64/egg/dgl/_ffi/_cy3
copying build/lib.linux-x86_64-3.7/dgl/_ffi/_cy3/__init__.py -> build/bdist.linux-x86_64/egg/dgl/_ffi/_cy3
copying build/lib.linux-x86_64-3.7/dgl/_ffi/_cy3/core.cpython-37m-x86_64-linux-gnu.so -> build/bdist.linux-x86_64/egg/dgl/_ffi/_cy3
creating build/bdist.linux-x86_64/egg/dgl/_ffi/_cy2
copying build/lib.linux-x86_64-3.7/dgl/_ffi/_cy2/__init__.py -> build/bdist.linux-x86_64/egg/dgl/_ffi/_cy2
creating build/bdist.linux-x86_64/egg/dgl/_ffi/_cython
copying build/lib.linux-x86_64-3.7/dgl/_ffi/_cython/core.cpp -> build/bdist.linux-x86_64/egg/dgl/_ffi/_cython
copying build/lib.linux-x86_64-3.7/dgl/frame.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-3.7/dgl/random.py -> build/bdist.linux-x86_64/egg/dgl
byte-compiling build/bdist.linux-x86_64/egg/dgl/mock_sparse/sp_matrix.py to sp_matrix.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/mock_sparse/diag_matrix.py to diag_matrix.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/mock_sparse/reduction.py to reduction.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/mock_sparse/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/mock_sparse/elementwise_op_sp.py to elementwise_op_sp.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/geometry/edge_coarsening.py to edge_coarsening.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/geometry/fps.py to fps.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/geometry/capi.py to capi.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/geometry/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/function/base.py to base.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/function/message.py to message.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/function/reducer.py to reducer.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/function/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/partition.py to partition.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/glob.py to glob.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/link/transe.py to transe.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/link/edgepred.py to edgepred.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/link/transr.py to transr.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/link/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/hetero.py to hetero.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/sparse_emb.py to sparse_emb.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/linear.py to linear.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/utils.py to utils.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/agnnconv.py to agnnconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/relgraphconv.py to relgraphconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/appnpconv.py to appnpconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/dgnconv.py to dgnconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/grouprevres.py to grouprevres.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/twirlsconv.py to twirlsconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/egatconv.py to egatconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/graphconv.py to graphconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/densechebconv.py to densechebconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/gcn2conv.py to gcn2conv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/ginconv.py to ginconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/dotgatconv.py to dotgatconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/cfconv.py to cfconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/gmmconv.py to gmmconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/sageconv.py to sageconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/densesageconv.py to densesageconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/gineconv.py to gineconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/edgeconv.py to edgeconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/nnconv.py to nnconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/gatv2conv.py to gatv2conv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/hgtconv.py to hgtconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/pnaconv.py to pnaconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/gatedgraphconv.py to gatedgraphconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/atomicconv.py to atomicconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/tagconv.py to tagconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/densegraphconv.py to densegraphconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/egnnconv.py to egnnconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/chebconv.py to chebconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/sgconv.py to sgconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/gatconv.py to gatconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/explain/gnnexplainer.py to gnnexplainer.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/explain/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/factory.py to factory.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/softmax.py to softmax.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/glob.py to glob.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/hetero.py to hetero.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/utils.py to utils.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/agnnconv.py to agnnconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/relgraphconv.py to relgraphconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/appnpconv.py to appnpconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/graphconv.py to graphconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/densechebconv.py to densechebconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/ginconv.py to ginconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/gmmconv.py to gmmconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/sageconv.py to sageconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/densesageconv.py to densesageconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/edgeconv.py to edgeconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/nnconv.py to nnconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/gatedgraphconv.py to gatedgraphconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/tagconv.py to tagconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/densegraphconv.py to densegraphconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/chebconv.py to chebconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/sgconv.py to sgconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/gatconv.py to gatconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/softmax.py to softmax.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/glob.py to glob.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/hetero.py to hetero.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/utils.py to utils.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv/relgraphconv.py to relgraphconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv/appnpconv.py to appnpconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv/graphconv.py to graphconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv/densechebconv.py to densechebconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv/ginconv.py to ginconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv/sageconv.py to sageconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv/edgeconv.py to edgeconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv/chebconv.py to chebconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv/sgconv.py to sgconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv/gatconv.py to gatconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/softmax.py to softmax.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/functional/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/core.py to core.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/subgraph.py to subgraph.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/contrib/dis_kvstore.py to dis_kvstore.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/contrib/unified_tensor.py to unified_tensor.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/contrib/graph_store.py to graph_store.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/contrib/sampling/sampler.py to sampler.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/contrib/sampling/dis_sampler.py to dis_sampler.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/contrib/sampling/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/contrib/data/knowledge_graph.py to knowledge_graph.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/contrib/data/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/contrib/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/traversal.py to traversal.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/utils/checks.py to checks.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/utils/exception.py to exception.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/utils/pin_memory.py to pin_memory.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/utils/internal.py to internal.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/utils/filter.py to filter.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/utils/data.py to data.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/utils/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/utils/shared_mem.py to shared_mem.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/multiprocessing/pytorch.py to pytorch.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/multiprocessing/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/base.py to base.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/optim/pytorch/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/optim/pytorch/sparse_optim.py to sparse_optim.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/optim/mxnet/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/optim/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/optim/tensorflow/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/heterograph_index.py to heterograph_index.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/sparse.py to sparse.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/global_config.py to global_config.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/dataloading/base.py to base.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/dataloading/dataloader.py to dataloader.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/dataloading/graphsaint.py to graphsaint.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/dataloading/shadow.py to shadow.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/dataloading/negative_sampler.py to negative_sampler.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/dataloading/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/dataloading/cluster_gcn.py to cluster_gcn.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/dataloading/neighbor_sampler.py to neighbor_sampler.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/dataloading/dist_dataloader.py to dist_dataloader.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_dataloading/neighbor.py to neighbor.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_dataloading/dataloader.py to dataloader.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_dataloading/pytorch/dataloader.py to dataloader.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_dataloading/pytorch/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_dataloading/shadow.py to shadow.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_dataloading/negative_sampler.py to negative_sampler.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_dataloading/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_dataloading/cluster_gcn.py to cluster_gcn.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_api_internal.py to _api_internal.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/heterograph.py to heterograph.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/backend.py to backend.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/pytorch/sparse.py to sparse.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/pytorch/tensor.py to tensor.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/pytorch/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/mxnet/sparse.py to sparse.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/mxnet/tensor.py to tensor.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/mxnet/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/mxnet/sparse_optim.py to sparse_optim.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/tensorflow/sparse.py to sparse.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/tensorflow/tensor.py to tensor.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/tensorflow/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/tensorflow/sparse_optim.py to sparse_optim.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/set_default_backend.py to set_default_backend.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/network.py to network.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/ops/edge_softmax.py to edge_softmax.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/ops/sddmm.py to sddmm.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/ops/gather_mm.py to gather_mm.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/ops/segment.py to segment.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/ops/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/ops/spmm.py to spmm.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/logging.py to logging.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/graph_index.py to graph_index.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/cuda/nccl.py to nccl.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/cuda/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/udf.py to udf.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/storages/pytorch_tensor.py to pytorch_tensor.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/storages/base.py to base.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/storages/tensor.py to tensor.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/storages/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/storages/numpy.py to numpy.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/ndarray.py to ndarray.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/init.py to init.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/view.py to view.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/sampling/neighbor.py to neighbor.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/sampling/pinsage.py to pinsage.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/sampling/negative.py to negative.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/sampling/randomwalks.py to randomwalks.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/sampling/utils.py to utils.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/sampling/node2vec_randomwalk.py to node2vec_randomwalk.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/sampling/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/generators.py to generators.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/qm7b.py to qm7b.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/graph_serialize.py to graph_serialize.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/dgl_dataset.py to dgl_dataset.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/csv_dataset_base.py to csv_dataset_base.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/knowledge_graph.py to knowledge_graph.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/qm9_edge.py to qm9_edge.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/gnn_benchmark.py to gnn_benchmark.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/qm9.py to qm9.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/icews18.py to icews18.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/gdelt.py to gdelt.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/utils.py to utils.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/sbm.py to sbm.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/csv_dataset.py to csv_dataset.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/karate.py to karate.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/citation_graph.py to citation_graph.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/rdf.py to rdf.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/bitcoinotc.py to bitcoinotc.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/ppi.py to ppi.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/tensor_serialize.py to tensor_serialize.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/flickr.py to flickr.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/wikics.py to wikics.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/heterograph_serialize.py to heterograph_serialize.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/tu.py to tu.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/gindt.py to gindt.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/tree.py to tree.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/synthetic.py to synthetic.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/adapter.py to adapter.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/fakenews.py to fakenews.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/fraud.py to fraud.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/minigc.py to minigc.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/yelp.py to yelp.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/reddit.py to reddit.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/convert.py to convert.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/container.py to container.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/merge.py to merge.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/readout.py to readout.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/dist_graph.py to dist_graph.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/graph_partition_book.py to graph_partition_book.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/partition.py to partition.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/nn/pytorch/sparse_emb.py to sparse_emb.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/nn/pytorch/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/nn/mxnet/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/nn/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/nn/tensorflow/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/optim/pytorch/utils.py to utils.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/optim/pytorch/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/optim/pytorch/sparse_optim.py to sparse_optim.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/optim/mxnet/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/optim/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/optim/tensorflow/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/standalone_kvstore.py to standalone_kvstore.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/graph_services.py to graph_services.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/role.py to role.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/constants.py to constants.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/rpc_server.py to rpc_server.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/kvstore.py to kvstore.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/dist_context.py to dist_context.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/rpc.py to rpc.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/rpc_client.py to rpc_client.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/id_map.py to id_map.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/server_state.py to server_state.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/shared_mem_utils.py to shared_mem_utils.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/dist_tensor.py to dist_tensor.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/dist_dataloader.py to dist_dataloader.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/nodeflow.py to nodeflow.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/udf.py to udf.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/view.py to view.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/kernel.py to kernel.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/spmv.py to spmv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/degree_bucketing.py to degree_bucketing.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/scheduler.py to scheduler.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/adapter.py to adapter.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/runtime.py to runtime.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/ir/program.py to program.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/ir/executor.py to executor.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/ir/var.py to var.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/ir/registry.py to registry.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/ir/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/graph.py to graph.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/frame.py to frame.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/propagate.py to propagate.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/batch.py to batch.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/transforms/functional.py to functional.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/transforms/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/transforms/module.py to module.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distgnn/partition/libra_partition.py to libra_partition.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distgnn/partition/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distgnn/tools/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distgnn/tools/tools.py to tools.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distgnn/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/runtime_ctypes.py to runtime_ctypes.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/function.py to function.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/libinfo.py to libinfo.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/streams.py to streams.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/base.py to base.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/_ctypes/function.py to function.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/_ctypes/types.py to types.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/_ctypes/ndarray.py to ndarray.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/_ctypes/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/_ctypes/object.py to object.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/ndarray.py to ndarray.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/object_generic.py to object_generic.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/object.py to object.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/_cy3/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/_cy2/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/frame.py to frame.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/random.py to random.cpython-37.pyc
creating stub loader for dgl/_ffi/_cy3/core.cpython-37m-x86_64-linux-gnu.so
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/_cy3/core.py to core.cpython-37.pyc
installing package data to build/bdist.linux-x86_64/egg
running install_data
copying ../build/libdgl.so -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/tensoradapter
creating build/bdist.linux-x86_64/egg/dgl/tensoradapter/pytorch
copying ../build/tensoradapter/pytorch/libtensoradapter_pytorch_1.9.0.so -> build/bdist.linux-x86_64/egg/dgl/tensoradapter/pytorch
creating build/bdist.linux-x86_64/egg/EGG-INFO
copying dgl.egg-info/PKG-INFO -> build/bdist.linux-x86_64/egg/EGG-INFO
copying dgl.egg-info/SOURCES.txt -> build/bdist.linux-x86_64/egg/EGG-INFO
copying dgl.egg-info/dependency_links.txt -> build/bdist.linux-x86_64/egg/EGG-INFO
copying dgl.egg-info/not-zip-safe -> build/bdist.linux-x86_64/egg/EGG-INFO
copying dgl.egg-info/requires.txt -> build/bdist.linux-x86_64/egg/EGG-INFO
copying dgl.egg-info/top_level.txt -> build/bdist.linux-x86_64/egg/EGG-INFO
writing build/bdist.linux-x86_64/egg/EGG-INFO/native_libs.txt
creating dist
creating 'dist/dgl-0.9-py3.7-linux-x86_64.egg' and adding 'build/bdist.linux-x86_64/egg' to it
removing 'build/bdist.linux-x86_64/egg' (and everything under it)
Processing dgl-0.9-py3.7-linux-x86_64.egg
creating /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages/dgl-0.9-py3.7-linux-x86_64.egg
Extracting dgl-0.9-py3.7-linux-x86_64.egg to /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages
Adding dgl 0.9 to easy-install.pth file

Installed /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages/dgl-0.9-py3.7-linux-x86_64.egg
Processing dependencies for dgl==0.9
Searching for psutil>=5.8.0
Reading https://pypi.org/simple/psutil/
Downloading https://files.pythonhosted.org/packages/3d/73/d8c87b5612c58d1e6c6d91997c1590771d34e4ee27d9c11eb1e64ecbf365/psutil-5.9.2-cp37-cp37m-manylinux_2_12_x86_64.manylinux2010_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl#sha256=4fb54941aac044a61db9d8eb56fc5bee207db3bc58645d657249030e15ba3727
Best match: psutil 5.9.2
Processing psutil-5.9.2-cp37-cp37m-manylinux_2_12_x86_64.manylinux2010_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl
Installing psutil-5.9.2-cp37-cp37m-manylinux_2_12_x86_64.manylinux2010_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl to /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages
Adding psutil 5.9.2 to easy-install.pth file

Installed /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages/psutil-5.9.2-py3.7-linux-x86_64.egg
Searching for tqdm==4.64.0
Best match: tqdm 4.64.0
Adding tqdm 4.64.0 to easy-install.pth file
Installing tqdm script to /opt/conda/envs/mxnet-ci/bin

Using /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages
Searching for requests==2.28.1
Best match: requests 2.28.1
Adding requests 2.28.1 to easy-install.pth file

Using /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages
Searching for networkx==2.6.3
Best match: networkx 2.6.3
Adding networkx 2.6.3 to easy-install.pth file

Using /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages
Searching for scipy==1.7.3
Best match: scipy 1.7.3
Adding scipy 1.7.3 to easy-install.pth file

Using /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages
Searching for numpy==1.21.6
Best match: numpy 1.21.6
Adding numpy 1.21.6 to easy-install.pth file
Installing f2py script to /opt/conda/envs/mxnet-ci/bin
Installing f2py3 script to /opt/conda/envs/mxnet-ci/bin
Installing f2py3.7 script to /opt/conda/envs/mxnet-ci/bin

Using /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages
Searching for certifi==2022.6.15
Best match: certifi 2022.6.15
Adding certifi 2022.6.15 to easy-install.pth file

Using /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages
Searching for urllib3==1.26.11
Best match: urllib3 1.26.11
Adding urllib3 1.26.11 to easy-install.pth file

Using /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages
Searching for idna==3.3
Best match: idna 3.3
Adding idna 3.3 to easy-install.pth file

Using /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages
Searching for charset-normalizer==2.1.0
Best match: charset-normalizer 2.1.0
Adding charset-normalizer 2.1.0 to easy-install.pth file
Installing normalizer script to /opt/conda/envs/mxnet-ci/bin

Using /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages
Finished processing dependencies for dgl==0.9
/opt/conda/envs/mxnet-ci/lib/python3.7/site-packages/Cython/Compiler/Main.py:369: FutureWarning: Cython directive 'language_level' not set, using 2 for now (Py2). This will change in a later release! File: /root/jenkins/workspace/dgl_PR-4648@2/python/dgl/_ffi/_cython/core.pyx
  tree = Parsing.p_module(s, pxd, full_module_name)
/opt/conda/envs/mxnet-ci/lib/python3.7/site-packages/setuptools/command/install.py:37: SetuptoolsDeprecationWarning: setup.py install is deprecated. Use build and pip and other standards-based tools.
  setuptools.SetuptoolsDeprecationWarning,
/opt/conda/envs/mxnet-ci/lib/python3.7/site-packages/setuptools/command/easy_install.py:147: EasyInstallDeprecationWarning: easy_install command is deprecated. Use build and pip and other standards-based tools.
  EasyInstallDeprecationWarning,
/opt/conda/envs/mxnet-ci/lib/python3.7/site-packages/pkg_resources/__init__.py:125: PkgResourcesDeprecationWarning:  is an invalid version and will not be supported in a future release
  PkgResourcesDeprecationWarning,
[1/1] Cythonizing dgl/_ffi/_cython/core.pyx
running build_ext
building 'dgl._ffi._cy3.core' extension
gcc -pthread -B /opt/conda/envs/mxnet-ci/compiler_compat -Wl,--sysroot=/ -Wsign-compare -DNDEBUG -g -fwrapv -O3 -Wall -Wstrict-prototypes -fPIC -I../include/ -I../third_party/dmlc-core/include -I../third_party/dlpack/include -I/opt/conda/envs/mxnet-ci/include/python3.7m -c dgl/_ffi/_cython/core.cpp -o build/temp.linux-x86_64-3.7/dgl/_ffi/_cython/core.o
cc1plus: warning: command line option ‘-Wstrict-prototypes’ is valid for C/ObjC but not for C++
g++ -pthread -B /opt/conda/envs/mxnet-ci/compiler_compat -Wl,--sysroot=/ -pthread -shared -B /opt/conda/envs/mxnet-ci/compiler_compat -L/opt/conda/envs/mxnet-ci/lib -Wl,-rpath=/opt/conda/envs/mxnet-ci/lib -Wl,--no-as-needed -Wl,--sysroot=/ build/temp.linux-x86_64-3.7/dgl/_ffi/_cython/core.o -Ldgl -L../build/Release -L../build -ldgl -o build/lib.linux-x86_64-3.7/dgl/_ffi/_cy3/core.cpython-37m-x86_64-linux-gnu.so
/opt/conda/envs/mxnet-ci/compiler_compat/ld: warning: /opt/conda/envs/mxnet-ci/lib/libstdc++.so: unsupported GNU_PROPERTY_TYPE (5) type: 0xc0010001
/opt/conda/envs/mxnet-ci/compiler_compat/ld: warning: /opt/conda/envs/mxnet-ci/lib/libstdc++.so: unsupported GNU_PROPERTY_TYPE (5) type: 0xc0010002
/opt/conda/envs/mxnet-ci/compiler_compat/ld: warning: /opt/conda/envs/mxnet-ci/lib/libgcc_s.so.1: unsupported GNU_PROPERTY_TYPE (5) type: 0xc0010001
/opt/conda/envs/mxnet-ci/compiler_compat/ld: warning: /opt/conda/envs/mxnet-ci/lib/libgcc_s.so.1: unsupported GNU_PROPERTY_TYPE (5) type: 0xc0010002
/opt/conda/envs/mxnet-ci/compiler_compat/ld: warning: /opt/conda/envs/mxnet-ci/lib/libgcc_s.so.1: unsupported GNU_PROPERTY_TYPE (5) type: 0xc0010001
/opt/conda/envs/mxnet-ci/compiler_compat/ld: warning: /opt/conda/envs/mxnet-ci/lib/libgcc_s.so.1: unsupported GNU_PROPERTY_TYPE (5) type: 0xc0010002
copying build/lib.linux-x86_64-3.7/dgl/_ffi/_cy3/core.cpython-37m-x86_64-linux-gnu.so -> dgl/_ffi/_cy3
/opt/conda/envs/mxnet-ci/lib/python3.7/site-packages/Cython/Compiler/Main.py:369: FutureWarning: Cython directive 'language_level' not set, using 2 for now (Py2). This will change in a later release! File: /root/jenkins/workspace/dgl_PR-4648@2/python/dgl/_ffi/_cython/core.pyx
  tree = Parsing.p_module(s, pxd, full_module_name)
WARNING: Skipping dgl as it is not installed.
WARNING: Running pip as the 'root' user can result in broken permissions and conflicting behaviour with the system package manager. It is recommended to use a virtual environment instead: https://pip.pypa.io/warnings/venv
[1/1] Cythonizing dgl/_ffi/_cython/core.pyx
running install
running bdist_egg
running egg_info
creating dgl.egg-info
writing dgl.egg-info/PKG-INFO
writing dependency_links to dgl.egg-info/dependency_links.txt
writing requirements to dgl.egg-info/requires.txt
writing top-level names to dgl.egg-info/top_level.txt
writing manifest file 'dgl.egg-info/SOURCES.txt'
reading manifest file 'dgl.egg-info/SOURCES.txt'
writing manifest file 'dgl.egg-info/SOURCES.txt'
installing library code to build/bdist.linux-x86_64/egg
running install_lib
running build_py
creating build
creating build/lib.linux-x86_64-3.7
creating build/lib.linux-x86_64-3.7/dgl
copying dgl/partition.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/core.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/subgraph.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/traversal.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/base.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/heterograph_index.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/sparse.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/global_config.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/_api_internal.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/heterograph.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/network.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/logging.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/graph_index.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/udf.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/ndarray.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/init.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/view.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/generators.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/convert.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/container.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/merge.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/readout.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/__init__.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/propagate.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/batch.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/frame.py -> build/lib.linux-x86_64-3.7/dgl
copying dgl/random.py -> build/lib.linux-x86_64-3.7/dgl
creating build/lib.linux-x86_64-3.7/dgl/mock_sparse
copying dgl/mock_sparse/sp_matrix.py -> build/lib.linux-x86_64-3.7/dgl/mock_sparse
copying dgl/mock_sparse/diag_matrix.py -> build/lib.linux-x86_64-3.7/dgl/mock_sparse
copying dgl/mock_sparse/reduction.py -> build/lib.linux-x86_64-3.7/dgl/mock_sparse
copying dgl/mock_sparse/__init__.py -> build/lib.linux-x86_64-3.7/dgl/mock_sparse
copying dgl/mock_sparse/elementwise_op_sp.py -> build/lib.linux-x86_64-3.7/dgl/mock_sparse
creating build/lib.linux-x86_64-3.7/dgl/geometry
copying dgl/geometry/edge_coarsening.py -> build/lib.linux-x86_64-3.7/dgl/geometry
copying dgl/geometry/fps.py -> build/lib.linux-x86_64-3.7/dgl/geometry
copying dgl/geometry/capi.py -> build/lib.linux-x86_64-3.7/dgl/geometry
copying dgl/geometry/__init__.py -> build/lib.linux-x86_64-3.7/dgl/geometry
creating build/lib.linux-x86_64-3.7/dgl/function
copying dgl/function/base.py -> build/lib.linux-x86_64-3.7/dgl/function
copying dgl/function/message.py -> build/lib.linux-x86_64-3.7/dgl/function
copying dgl/function/reducer.py -> build/lib.linux-x86_64-3.7/dgl/function
copying dgl/function/__init__.py -> build/lib.linux-x86_64-3.7/dgl/function
creating build/lib.linux-x86_64-3.7/dgl/nn
copying dgl/nn/__init__.py -> build/lib.linux-x86_64-3.7/dgl/nn
creating build/lib.linux-x86_64-3.7/dgl/contrib
copying dgl/contrib/dis_kvstore.py -> build/lib.linux-x86_64-3.7/dgl/contrib
copying dgl/contrib/unified_tensor.py -> build/lib.linux-x86_64-3.7/dgl/contrib
copying dgl/contrib/graph_store.py -> build/lib.linux-x86_64-3.7/dgl/contrib
copying dgl/contrib/__init__.py -> build/lib.linux-x86_64-3.7/dgl/contrib
creating build/lib.linux-x86_64-3.7/dgl/utils
copying dgl/utils/checks.py -> build/lib.linux-x86_64-3.7/dgl/utils
copying dgl/utils/exception.py -> build/lib.linux-x86_64-3.7/dgl/utils
copying dgl/utils/pin_memory.py -> build/lib.linux-x86_64-3.7/dgl/utils
copying dgl/utils/internal.py -> build/lib.linux-x86_64-3.7/dgl/utils
copying dgl/utils/filter.py -> build/lib.linux-x86_64-3.7/dgl/utils
copying dgl/utils/data.py -> build/lib.linux-x86_64-3.7/dgl/utils
copying dgl/utils/__init__.py -> build/lib.linux-x86_64-3.7/dgl/utils
copying dgl/utils/shared_mem.py -> build/lib.linux-x86_64-3.7/dgl/utils
creating build/lib.linux-x86_64-3.7/dgl/multiprocessing
copying dgl/multiprocessing/pytorch.py -> build/lib.linux-x86_64-3.7/dgl/multiprocessing
copying dgl/multiprocessing/__init__.py -> build/lib.linux-x86_64-3.7/dgl/multiprocessing
creating build/lib.linux-x86_64-3.7/dgl/optim
copying dgl/optim/__init__.py -> build/lib.linux-x86_64-3.7/dgl/optim
creating build/lib.linux-x86_64-3.7/dgl/dataloading
copying dgl/dataloading/base.py -> build/lib.linux-x86_64-3.7/dgl/dataloading
copying dgl/dataloading/dataloader.py -> build/lib.linux-x86_64-3.7/dgl/dataloading
copying dgl/dataloading/graphsaint.py -> build/lib.linux-x86_64-3.7/dgl/dataloading
copying dgl/dataloading/shadow.py -> build/lib.linux-x86_64-3.7/dgl/dataloading
copying dgl/dataloading/negative_sampler.py -> build/lib.linux-x86_64-3.7/dgl/dataloading
copying dgl/dataloading/__init__.py -> build/lib.linux-x86_64-3.7/dgl/dataloading
copying dgl/dataloading/cluster_gcn.py -> build/lib.linux-x86_64-3.7/dgl/dataloading
copying dgl/dataloading/neighbor_sampler.py -> build/lib.linux-x86_64-3.7/dgl/dataloading
copying dgl/dataloading/dist_dataloader.py -> build/lib.linux-x86_64-3.7/dgl/dataloading
creating build/lib.linux-x86_64-3.7/dgl/_dataloading
copying dgl/_dataloading/neighbor.py -> build/lib.linux-x86_64-3.7/dgl/_dataloading
copying dgl/_dataloading/dataloader.py -> build/lib.linux-x86_64-3.7/dgl/_dataloading
copying dgl/_dataloading/shadow.py -> build/lib.linux-x86_64-3.7/dgl/_dataloading
copying dgl/_dataloading/negative_sampler.py -> build/lib.linux-x86_64-3.7/dgl/_dataloading
copying dgl/_dataloading/__init__.py -> build/lib.linux-x86_64-3.7/dgl/_dataloading
copying dgl/_dataloading/cluster_gcn.py -> build/lib.linux-x86_64-3.7/dgl/_dataloading
creating build/lib.linux-x86_64-3.7/dgl/backend
copying dgl/backend/backend.py -> build/lib.linux-x86_64-3.7/dgl/backend
copying dgl/backend/__init__.py -> build/lib.linux-x86_64-3.7/dgl/backend
copying dgl/backend/set_default_backend.py -> build/lib.linux-x86_64-3.7/dgl/backend
creating build/lib.linux-x86_64-3.7/dgl/ops
copying dgl/ops/edge_softmax.py -> build/lib.linux-x86_64-3.7/dgl/ops
copying dgl/ops/sddmm.py -> build/lib.linux-x86_64-3.7/dgl/ops
copying dgl/ops/gather_mm.py -> build/lib.linux-x86_64-3.7/dgl/ops
copying dgl/ops/segment.py -> build/lib.linux-x86_64-3.7/dgl/ops
copying dgl/ops/__init__.py -> build/lib.linux-x86_64-3.7/dgl/ops
copying dgl/ops/spmm.py -> build/lib.linux-x86_64-3.7/dgl/ops
creating build/lib.linux-x86_64-3.7/dgl/cuda
copying dgl/cuda/nccl.py -> build/lib.linux-x86_64-3.7/dgl/cuda
copying dgl/cuda/__init__.py -> build/lib.linux-x86_64-3.7/dgl/cuda
creating build/lib.linux-x86_64-3.7/dgl/storages
copying dgl/storages/pytorch_tensor.py -> build/lib.linux-x86_64-3.7/dgl/storages
copying dgl/storages/base.py -> build/lib.linux-x86_64-3.7/dgl/storages
copying dgl/storages/tensor.py -> build/lib.linux-x86_64-3.7/dgl/storages
copying dgl/storages/__init__.py -> build/lib.linux-x86_64-3.7/dgl/storages
copying dgl/storages/numpy.py -> build/lib.linux-x86_64-3.7/dgl/storages
creating build/lib.linux-x86_64-3.7/dgl/sampling
copying dgl/sampling/neighbor.py -> build/lib.linux-x86_64-3.7/dgl/sampling
copying dgl/sampling/pinsage.py -> build/lib.linux-x86_64-3.7/dgl/sampling
copying dgl/sampling/negative.py -> build/lib.linux-x86_64-3.7/dgl/sampling
copying dgl/sampling/randomwalks.py -> build/lib.linux-x86_64-3.7/dgl/sampling
copying dgl/sampling/utils.py -> build/lib.linux-x86_64-3.7/dgl/sampling
copying dgl/sampling/node2vec_randomwalk.py -> build/lib.linux-x86_64-3.7/dgl/sampling
copying dgl/sampling/__init__.py -> build/lib.linux-x86_64-3.7/dgl/sampling
creating build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/qm7b.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/graph_serialize.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/dgl_dataset.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/csv_dataset_base.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/knowledge_graph.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/qm9_edge.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/gnn_benchmark.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/qm9.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/icews18.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/gdelt.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/utils.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/sbm.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/csv_dataset.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/karate.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/citation_graph.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/rdf.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/bitcoinotc.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/ppi.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/tensor_serialize.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/flickr.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/wikics.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/heterograph_serialize.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/tu.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/__init__.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/gindt.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/tree.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/synthetic.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/adapter.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/fakenews.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/fraud.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/minigc.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/yelp.py -> build/lib.linux-x86_64-3.7/dgl/data
copying dgl/data/reddit.py -> build/lib.linux-x86_64-3.7/dgl/data
creating build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/dist_graph.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/graph_partition_book.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/partition.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/standalone_kvstore.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/graph_services.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/role.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/constants.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/rpc_server.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/kvstore.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/dist_context.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/rpc.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/rpc_client.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/id_map.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/server_state.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/__init__.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/shared_mem_utils.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/dist_tensor.py -> build/lib.linux-x86_64-3.7/dgl/distributed
copying dgl/distributed/dist_dataloader.py -> build/lib.linux-x86_64-3.7/dgl/distributed
creating build/lib.linux-x86_64-3.7/dgl/_deprecate
copying dgl/_deprecate/nodeflow.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate
copying dgl/_deprecate/udf.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate
copying dgl/_deprecate/view.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate
copying dgl/_deprecate/kernel.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate
copying dgl/_deprecate/__init__.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate
copying dgl/_deprecate/graph.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate
copying dgl/_deprecate/frame.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate
creating build/lib.linux-x86_64-3.7/dgl/transforms
copying dgl/transforms/functional.py -> build/lib.linux-x86_64-3.7/dgl/transforms
copying dgl/transforms/__init__.py -> build/lib.linux-x86_64-3.7/dgl/transforms
copying dgl/transforms/module.py -> build/lib.linux-x86_64-3.7/dgl/transforms
creating build/lib.linux-x86_64-3.7/dgl/distgnn
copying dgl/distgnn/__init__.py -> build/lib.linux-x86_64-3.7/dgl/distgnn
creating build/lib.linux-x86_64-3.7/dgl/_ffi
copying dgl/_ffi/runtime_ctypes.py -> build/lib.linux-x86_64-3.7/dgl/_ffi
copying dgl/_ffi/function.py -> build/lib.linux-x86_64-3.7/dgl/_ffi
copying dgl/_ffi/libinfo.py -> build/lib.linux-x86_64-3.7/dgl/_ffi
copying dgl/_ffi/streams.py -> build/lib.linux-x86_64-3.7/dgl/_ffi
copying dgl/_ffi/base.py -> build/lib.linux-x86_64-3.7/dgl/_ffi
copying dgl/_ffi/ndarray.py -> build/lib.linux-x86_64-3.7/dgl/_ffi
copying dgl/_ffi/object_generic.py -> build/lib.linux-x86_64-3.7/dgl/_ffi
copying dgl/_ffi/__init__.py -> build/lib.linux-x86_64-3.7/dgl/_ffi
copying dgl/_ffi/object.py -> build/lib.linux-x86_64-3.7/dgl/_ffi
creating build/lib.linux-x86_64-3.7/dgl/nn/pytorch
copying dgl/nn/pytorch/glob.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch
copying dgl/nn/pytorch/hetero.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch
copying dgl/nn/pytorch/sparse_emb.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch
copying dgl/nn/pytorch/linear.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch
copying dgl/nn/pytorch/utils.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch
copying dgl/nn/pytorch/__init__.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch
copying dgl/nn/pytorch/factory.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch
copying dgl/nn/pytorch/softmax.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch
creating build/lib.linux-x86_64-3.7/dgl/nn/mxnet
copying dgl/nn/mxnet/glob.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet
copying dgl/nn/mxnet/hetero.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet
copying dgl/nn/mxnet/utils.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet
copying dgl/nn/mxnet/__init__.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet
copying dgl/nn/mxnet/softmax.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet
creating build/lib.linux-x86_64-3.7/dgl/nn/tensorflow
copying dgl/nn/tensorflow/glob.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow
copying dgl/nn/tensorflow/hetero.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow
copying dgl/nn/tensorflow/utils.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow
copying dgl/nn/tensorflow/__init__.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow
copying dgl/nn/tensorflow/softmax.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow
creating build/lib.linux-x86_64-3.7/dgl/nn/functional
copying dgl/nn/functional/__init__.py -> build/lib.linux-x86_64-3.7/dgl/nn/functional
creating build/lib.linux-x86_64-3.7/dgl/nn/pytorch/link
copying dgl/nn/pytorch/link/transe.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/link
copying dgl/nn/pytorch/link/edgepred.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/link
copying dgl/nn/pytorch/link/transr.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/link
copying dgl/nn/pytorch/link/__init__.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/link
creating build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/agnnconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/relgraphconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/appnpconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/dgnconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/grouprevres.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/twirlsconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/egatconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/graphconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/densechebconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/gcn2conv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/ginconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/dotgatconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/cfconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/gmmconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/sageconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/densesageconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/gineconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/edgeconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/nnconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/__init__.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/gatv2conv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/hgtconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/pnaconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/gatedgraphconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/atomicconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/tagconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/densegraphconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/egnnconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/chebconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/sgconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/gatconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv
creating build/lib.linux-x86_64-3.7/dgl/nn/pytorch/explain
copying dgl/nn/pytorch/explain/gnnexplainer.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/explain
copying dgl/nn/pytorch/explain/__init__.py -> build/lib.linux-x86_64-3.7/dgl/nn/pytorch/explain
creating build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/agnnconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/relgraphconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/appnpconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/graphconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/densechebconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/ginconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/gmmconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/sageconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/densesageconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/edgeconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/nnconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/__init__.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/gatedgraphconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/tagconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/densegraphconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/chebconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/sgconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/gatconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv
creating build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv
copying dgl/nn/tensorflow/conv/relgraphconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv
copying dgl/nn/tensorflow/conv/appnpconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv
copying dgl/nn/tensorflow/conv/graphconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv
copying dgl/nn/tensorflow/conv/densechebconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv
copying dgl/nn/tensorflow/conv/ginconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv
copying dgl/nn/tensorflow/conv/sageconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv
copying dgl/nn/tensorflow/conv/edgeconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv
copying dgl/nn/tensorflow/conv/__init__.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv
copying dgl/nn/tensorflow/conv/chebconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv
copying dgl/nn/tensorflow/conv/sgconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv
copying dgl/nn/tensorflow/conv/gatconv.py -> build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv
creating build/lib.linux-x86_64-3.7/dgl/contrib/sampling
copying dgl/contrib/sampling/sampler.py -> build/lib.linux-x86_64-3.7/dgl/contrib/sampling
copying dgl/contrib/sampling/dis_sampler.py -> build/lib.linux-x86_64-3.7/dgl/contrib/sampling
copying dgl/contrib/sampling/__init__.py -> build/lib.linux-x86_64-3.7/dgl/contrib/sampling
creating build/lib.linux-x86_64-3.7/dgl/contrib/data
copying dgl/contrib/data/knowledge_graph.py -> build/lib.linux-x86_64-3.7/dgl/contrib/data
copying dgl/contrib/data/__init__.py -> build/lib.linux-x86_64-3.7/dgl/contrib/data
creating build/lib.linux-x86_64-3.7/dgl/optim/pytorch
copying dgl/optim/pytorch/__init__.py -> build/lib.linux-x86_64-3.7/dgl/optim/pytorch
copying dgl/optim/pytorch/sparse_optim.py -> build/lib.linux-x86_64-3.7/dgl/optim/pytorch
creating build/lib.linux-x86_64-3.7/dgl/optim/mxnet
copying dgl/optim/mxnet/__init__.py -> build/lib.linux-x86_64-3.7/dgl/optim/mxnet
creating build/lib.linux-x86_64-3.7/dgl/optim/tensorflow
copying dgl/optim/tensorflow/__init__.py -> build/lib.linux-x86_64-3.7/dgl/optim/tensorflow
creating build/lib.linux-x86_64-3.7/dgl/_dataloading/pytorch
copying dgl/_dataloading/pytorch/dataloader.py -> build/lib.linux-x86_64-3.7/dgl/_dataloading/pytorch
copying dgl/_dataloading/pytorch/__init__.py -> build/lib.linux-x86_64-3.7/dgl/_dataloading/pytorch
creating build/lib.linux-x86_64-3.7/dgl/backend/pytorch
copying dgl/backend/pytorch/sparse.py -> build/lib.linux-x86_64-3.7/dgl/backend/pytorch
copying dgl/backend/pytorch/tensor.py -> build/lib.linux-x86_64-3.7/dgl/backend/pytorch
copying dgl/backend/pytorch/__init__.py -> build/lib.linux-x86_64-3.7/dgl/backend/pytorch
creating build/lib.linux-x86_64-3.7/dgl/backend/mxnet
copying dgl/backend/mxnet/sparse.py -> build/lib.linux-x86_64-3.7/dgl/backend/mxnet
copying dgl/backend/mxnet/tensor.py -> build/lib.linux-x86_64-3.7/dgl/backend/mxnet
copying dgl/backend/mxnet/__init__.py -> build/lib.linux-x86_64-3.7/dgl/backend/mxnet
copying dgl/backend/mxnet/sparse_optim.py -> build/lib.linux-x86_64-3.7/dgl/backend/mxnet
creating build/lib.linux-x86_64-3.7/dgl/backend/tensorflow
copying dgl/backend/tensorflow/sparse.py -> build/lib.linux-x86_64-3.7/dgl/backend/tensorflow
copying dgl/backend/tensorflow/tensor.py -> build/lib.linux-x86_64-3.7/dgl/backend/tensorflow
copying dgl/backend/tensorflow/__init__.py -> build/lib.linux-x86_64-3.7/dgl/backend/tensorflow
copying dgl/backend/tensorflow/sparse_optim.py -> build/lib.linux-x86_64-3.7/dgl/backend/tensorflow
creating build/lib.linux-x86_64-3.7/dgl/distributed/nn
copying dgl/distributed/nn/__init__.py -> build/lib.linux-x86_64-3.7/dgl/distributed/nn
creating build/lib.linux-x86_64-3.7/dgl/distributed/optim
copying dgl/distributed/optim/__init__.py -> build/lib.linux-x86_64-3.7/dgl/distributed/optim
creating build/lib.linux-x86_64-3.7/dgl/distributed/nn/pytorch
copying dgl/distributed/nn/pytorch/sparse_emb.py -> build/lib.linux-x86_64-3.7/dgl/distributed/nn/pytorch
copying dgl/distributed/nn/pytorch/__init__.py -> build/lib.linux-x86_64-3.7/dgl/distributed/nn/pytorch
creating build/lib.linux-x86_64-3.7/dgl/distributed/nn/mxnet
copying dgl/distributed/nn/mxnet/__init__.py -> build/lib.linux-x86_64-3.7/dgl/distributed/nn/mxnet
creating build/lib.linux-x86_64-3.7/dgl/distributed/nn/tensorflow
copying dgl/distributed/nn/tensorflow/__init__.py -> build/lib.linux-x86_64-3.7/dgl/distributed/nn/tensorflow
creating build/lib.linux-x86_64-3.7/dgl/distributed/optim/pytorch
copying dgl/distributed/optim/pytorch/utils.py -> build/lib.linux-x86_64-3.7/dgl/distributed/optim/pytorch
copying dgl/distributed/optim/pytorch/__init__.py -> build/lib.linux-x86_64-3.7/dgl/distributed/optim/pytorch
copying dgl/distributed/optim/pytorch/sparse_optim.py -> build/lib.linux-x86_64-3.7/dgl/distributed/optim/pytorch
creating build/lib.linux-x86_64-3.7/dgl/distributed/optim/mxnet
copying dgl/distributed/optim/mxnet/__init__.py -> build/lib.linux-x86_64-3.7/dgl/distributed/optim/mxnet
creating build/lib.linux-x86_64-3.7/dgl/distributed/optim/tensorflow
copying dgl/distributed/optim/tensorflow/__init__.py -> build/lib.linux-x86_64-3.7/dgl/distributed/optim/tensorflow
creating build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime
copying dgl/_deprecate/runtime/spmv.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime
copying dgl/_deprecate/runtime/degree_bucketing.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime
copying dgl/_deprecate/runtime/__init__.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime
copying dgl/_deprecate/runtime/scheduler.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime
copying dgl/_deprecate/runtime/adapter.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime
copying dgl/_deprecate/runtime/runtime.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime
creating build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/ir
copying dgl/_deprecate/runtime/ir/program.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/ir
copying dgl/_deprecate/runtime/ir/executor.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/ir
copying dgl/_deprecate/runtime/ir/var.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/ir
copying dgl/_deprecate/runtime/ir/registry.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/ir
copying dgl/_deprecate/runtime/ir/__init__.py -> build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/ir
creating build/lib.linux-x86_64-3.7/dgl/distgnn/partition
copying dgl/distgnn/partition/libra_partition.py -> build/lib.linux-x86_64-3.7/dgl/distgnn/partition
copying dgl/distgnn/partition/__init__.py -> build/lib.linux-x86_64-3.7/dgl/distgnn/partition
creating build/lib.linux-x86_64-3.7/dgl/distgnn/tools
copying dgl/distgnn/tools/__init__.py -> build/lib.linux-x86_64-3.7/dgl/distgnn/tools
copying dgl/distgnn/tools/tools.py -> build/lib.linux-x86_64-3.7/dgl/distgnn/tools
creating build/lib.linux-x86_64-3.7/dgl/_ffi/_ctypes
copying dgl/_ffi/_ctypes/function.py -> build/lib.linux-x86_64-3.7/dgl/_ffi/_ctypes
copying dgl/_ffi/_ctypes/types.py -> build/lib.linux-x86_64-3.7/dgl/_ffi/_ctypes
copying dgl/_ffi/_ctypes/ndarray.py -> build/lib.linux-x86_64-3.7/dgl/_ffi/_ctypes
copying dgl/_ffi/_ctypes/__init__.py -> build/lib.linux-x86_64-3.7/dgl/_ffi/_ctypes
copying dgl/_ffi/_ctypes/object.py -> build/lib.linux-x86_64-3.7/dgl/_ffi/_ctypes
creating build/lib.linux-x86_64-3.7/dgl/_ffi/_cy3
copying dgl/_ffi/_cy3/__init__.py -> build/lib.linux-x86_64-3.7/dgl/_ffi/_cy3
creating build/lib.linux-x86_64-3.7/dgl/_ffi/_cy2
copying dgl/_ffi/_cy2/__init__.py -> build/lib.linux-x86_64-3.7/dgl/_ffi/_cy2
creating build/lib.linux-x86_64-3.7/dgl/_ffi/_cython
copying dgl/_ffi/_cython/core.cpp -> build/lib.linux-x86_64-3.7/dgl/_ffi/_cython
running build_ext
building 'dgl._ffi._cy3.core' extension
creating build/temp.linux-x86_64-3.7
creating build/temp.linux-x86_64-3.7/dgl
creating build/temp.linux-x86_64-3.7/dgl/_ffi
creating build/temp.linux-x86_64-3.7/dgl/_ffi/_cython
gcc -pthread -B /opt/conda/envs/tensorflow-ci/compiler_compat -Wl,--sysroot=/ -Wsign-compare -DNDEBUG -g -fwrapv -O3 -Wall -Wstrict-prototypes -fPIC -I../include/ -I../third_party/dmlc-core/include -I../third_party/dlpack/include -I/opt/conda/envs/tensorflow-ci/include/python3.7m -c dgl/_ffi/_cython/core.cpp -o build/temp.linux-x86_64-3.7/dgl/_ffi/_cython/core.o
cc1plus: warning: command line option ‘-Wstrict-prototypes’ is valid for C/ObjC but not for C++
g++ -pthread -B /opt/conda/envs/tensorflow-ci/compiler_compat -Wl,--sysroot=/ -pthread -shared -B /opt/conda/envs/tensorflow-ci/compiler_compat -L/opt/conda/envs/tensorflow-ci/lib -Wl,-rpath=/opt/conda/envs/tensorflow-ci/lib -Wl,--no-as-needed -Wl,--sysroot=/ build/temp.linux-x86_64-3.7/dgl/_ffi/_cython/core.o -Ldgl -L../build/Release -L../build -ldgl -o build/lib.linux-x86_64-3.7/dgl/_ffi/_cy3/core.cpython-37m-x86_64-linux-gnu.so
/opt/conda/envs/tensorflow-ci/compiler_compat/ld: warning: /opt/conda/envs/tensorflow-ci/lib/libstdc++.so: unsupported GNU_PROPERTY_TYPE (5) type: 0xc0010001
/opt/conda/envs/tensorflow-ci/compiler_compat/ld: warning: /opt/conda/envs/tensorflow-ci/lib/libstdc++.so: unsupported GNU_PROPERTY_TYPE (5) type: 0xc0010002
/opt/conda/envs/tensorflow-ci/compiler_compat/ld: warning: /opt/conda/envs/tensorflow-ci/lib/libgcc_s.so.1: unsupported GNU_PROPERTY_TYPE (5) type: 0xc0010001
/opt/conda/envs/tensorflow-ci/compiler_compat/ld: warning: /opt/conda/envs/tensorflow-ci/lib/libgcc_s.so.1: unsupported GNU_PROPERTY_TYPE (5) type: 0xc0010002
/opt/conda/envs/tensorflow-ci/compiler_compat/ld: warning: /opt/conda/envs/tensorflow-ci/lib/libgcc_s.so.1: unsupported GNU_PROPERTY_TYPE (5) type: 0xc0010001
/opt/conda/envs/tensorflow-ci/compiler_compat/ld: warning: /opt/conda/envs/tensorflow-ci/lib/libgcc_s.so.1: unsupported GNU_PROPERTY_TYPE (5) type: 0xc0010002
creating build/bdist.linux-x86_64
creating build/bdist.linux-x86_64/egg
creating build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/mock_sparse
copying build/lib.linux-x86_64-3.7/dgl/mock_sparse/sp_matrix.py -> build/bdist.linux-x86_64/egg/dgl/mock_sparse
copying build/lib.linux-x86_64-3.7/dgl/mock_sparse/diag_matrix.py -> build/bdist.linux-x86_64/egg/dgl/mock_sparse
copying build/lib.linux-x86_64-3.7/dgl/mock_sparse/reduction.py -> build/bdist.linux-x86_64/egg/dgl/mock_sparse
copying build/lib.linux-x86_64-3.7/dgl/mock_sparse/__init__.py -> build/bdist.linux-x86_64/egg/dgl/mock_sparse
copying build/lib.linux-x86_64-3.7/dgl/mock_sparse/elementwise_op_sp.py -> build/bdist.linux-x86_64/egg/dgl/mock_sparse
creating build/bdist.linux-x86_64/egg/dgl/geometry
copying build/lib.linux-x86_64-3.7/dgl/geometry/edge_coarsening.py -> build/bdist.linux-x86_64/egg/dgl/geometry
copying build/lib.linux-x86_64-3.7/dgl/geometry/fps.py -> build/bdist.linux-x86_64/egg/dgl/geometry
copying build/lib.linux-x86_64-3.7/dgl/geometry/capi.py -> build/bdist.linux-x86_64/egg/dgl/geometry
copying build/lib.linux-x86_64-3.7/dgl/geometry/__init__.py -> build/bdist.linux-x86_64/egg/dgl/geometry
creating build/bdist.linux-x86_64/egg/dgl/function
copying build/lib.linux-x86_64-3.7/dgl/function/base.py -> build/bdist.linux-x86_64/egg/dgl/function
copying build/lib.linux-x86_64-3.7/dgl/function/message.py -> build/bdist.linux-x86_64/egg/dgl/function
copying build/lib.linux-x86_64-3.7/dgl/function/reducer.py -> build/bdist.linux-x86_64/egg/dgl/function
copying build/lib.linux-x86_64-3.7/dgl/function/__init__.py -> build/bdist.linux-x86_64/egg/dgl/function
copying build/lib.linux-x86_64-3.7/dgl/partition.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/nn
creating build/bdist.linux-x86_64/egg/dgl/nn/pytorch
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/glob.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch
creating build/bdist.linux-x86_64/egg/dgl/nn/pytorch/link
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/link/transe.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/link
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/link/edgepred.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/link
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/link/transr.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/link
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/link/__init__.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/link
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/hetero.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/sparse_emb.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/linear.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/utils.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch
creating build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/agnnconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/relgraphconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/appnpconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/dgnconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/grouprevres.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/twirlsconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/egatconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/graphconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/densechebconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/gcn2conv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/ginconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/dotgatconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/cfconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/gmmconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/sageconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/densesageconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/gineconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/edgeconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/nnconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/__init__.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/gatv2conv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/hgtconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/pnaconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/gatedgraphconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/atomicconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/tagconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/densegraphconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/egnnconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/chebconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/sgconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/conv/gatconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/__init__.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch
creating build/bdist.linux-x86_64/egg/dgl/nn/pytorch/explain
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/explain/gnnexplainer.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/explain
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/explain/__init__.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/explain
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/factory.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch
copying build/lib.linux-x86_64-3.7/dgl/nn/pytorch/softmax.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch
creating build/bdist.linux-x86_64/egg/dgl/nn/mxnet
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/glob.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/hetero.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/utils.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet
creating build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/agnnconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/relgraphconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/appnpconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/graphconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/densechebconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/ginconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/gmmconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/sageconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/densesageconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/edgeconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/nnconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/__init__.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/gatedgraphconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/tagconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/densegraphconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/chebconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/sgconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/conv/gatconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/__init__.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet
copying build/lib.linux-x86_64-3.7/dgl/nn/mxnet/softmax.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet
copying build/lib.linux-x86_64-3.7/dgl/nn/__init__.py -> build/bdist.linux-x86_64/egg/dgl/nn
creating build/bdist.linux-x86_64/egg/dgl/nn/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/glob.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/hetero.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/utils.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow
creating build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv/relgraphconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv/appnpconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv/graphconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv/densechebconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv/ginconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv/sageconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv/edgeconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv/__init__.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv/chebconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv/sgconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/conv/gatconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/__init__.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/nn/tensorflow/softmax.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow
creating build/bdist.linux-x86_64/egg/dgl/nn/functional
copying build/lib.linux-x86_64-3.7/dgl/nn/functional/__init__.py -> build/bdist.linux-x86_64/egg/dgl/nn/functional
copying build/lib.linux-x86_64-3.7/dgl/core.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-3.7/dgl/subgraph.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/contrib
copying build/lib.linux-x86_64-3.7/dgl/contrib/dis_kvstore.py -> build/bdist.linux-x86_64/egg/dgl/contrib
copying build/lib.linux-x86_64-3.7/dgl/contrib/unified_tensor.py -> build/bdist.linux-x86_64/egg/dgl/contrib
copying build/lib.linux-x86_64-3.7/dgl/contrib/graph_store.py -> build/bdist.linux-x86_64/egg/dgl/contrib
creating build/bdist.linux-x86_64/egg/dgl/contrib/sampling
copying build/lib.linux-x86_64-3.7/dgl/contrib/sampling/sampler.py -> build/bdist.linux-x86_64/egg/dgl/contrib/sampling
copying build/lib.linux-x86_64-3.7/dgl/contrib/sampling/dis_sampler.py -> build/bdist.linux-x86_64/egg/dgl/contrib/sampling
copying build/lib.linux-x86_64-3.7/dgl/contrib/sampling/__init__.py -> build/bdist.linux-x86_64/egg/dgl/contrib/sampling
creating build/bdist.linux-x86_64/egg/dgl/contrib/data
copying build/lib.linux-x86_64-3.7/dgl/contrib/data/knowledge_graph.py -> build/bdist.linux-x86_64/egg/dgl/contrib/data
copying build/lib.linux-x86_64-3.7/dgl/contrib/data/__init__.py -> build/bdist.linux-x86_64/egg/dgl/contrib/data
copying build/lib.linux-x86_64-3.7/dgl/contrib/__init__.py -> build/bdist.linux-x86_64/egg/dgl/contrib
copying build/lib.linux-x86_64-3.7/dgl/traversal.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/utils
copying build/lib.linux-x86_64-3.7/dgl/utils/checks.py -> build/bdist.linux-x86_64/egg/dgl/utils
copying build/lib.linux-x86_64-3.7/dgl/utils/exception.py -> build/bdist.linux-x86_64/egg/dgl/utils
copying build/lib.linux-x86_64-3.7/dgl/utils/pin_memory.py -> build/bdist.linux-x86_64/egg/dgl/utils
copying build/lib.linux-x86_64-3.7/dgl/utils/internal.py -> build/bdist.linux-x86_64/egg/dgl/utils
copying build/lib.linux-x86_64-3.7/dgl/utils/filter.py -> build/bdist.linux-x86_64/egg/dgl/utils
copying build/lib.linux-x86_64-3.7/dgl/utils/data.py -> build/bdist.linux-x86_64/egg/dgl/utils
copying build/lib.linux-x86_64-3.7/dgl/utils/__init__.py -> build/bdist.linux-x86_64/egg/dgl/utils
copying build/lib.linux-x86_64-3.7/dgl/utils/shared_mem.py -> build/bdist.linux-x86_64/egg/dgl/utils
creating build/bdist.linux-x86_64/egg/dgl/multiprocessing
copying build/lib.linux-x86_64-3.7/dgl/multiprocessing/pytorch.py -> build/bdist.linux-x86_64/egg/dgl/multiprocessing
copying build/lib.linux-x86_64-3.7/dgl/multiprocessing/__init__.py -> build/bdist.linux-x86_64/egg/dgl/multiprocessing
copying build/lib.linux-x86_64-3.7/dgl/base.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/optim
creating build/bdist.linux-x86_64/egg/dgl/optim/pytorch
copying build/lib.linux-x86_64-3.7/dgl/optim/pytorch/__init__.py -> build/bdist.linux-x86_64/egg/dgl/optim/pytorch
copying build/lib.linux-x86_64-3.7/dgl/optim/pytorch/sparse_optim.py -> build/bdist.linux-x86_64/egg/dgl/optim/pytorch
creating build/bdist.linux-x86_64/egg/dgl/optim/mxnet
copying build/lib.linux-x86_64-3.7/dgl/optim/mxnet/__init__.py -> build/bdist.linux-x86_64/egg/dgl/optim/mxnet
copying build/lib.linux-x86_64-3.7/dgl/optim/__init__.py -> build/bdist.linux-x86_64/egg/dgl/optim
creating build/bdist.linux-x86_64/egg/dgl/optim/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/optim/tensorflow/__init__.py -> build/bdist.linux-x86_64/egg/dgl/optim/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/heterograph_index.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-3.7/dgl/sparse.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-3.7/dgl/global_config.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/dataloading
copying build/lib.linux-x86_64-3.7/dgl/dataloading/base.py -> build/bdist.linux-x86_64/egg/dgl/dataloading
copying build/lib.linux-x86_64-3.7/dgl/dataloading/dataloader.py -> build/bdist.linux-x86_64/egg/dgl/dataloading
copying build/lib.linux-x86_64-3.7/dgl/dataloading/graphsaint.py -> build/bdist.linux-x86_64/egg/dgl/dataloading
copying build/lib.linux-x86_64-3.7/dgl/dataloading/shadow.py -> build/bdist.linux-x86_64/egg/dgl/dataloading
copying build/lib.linux-x86_64-3.7/dgl/dataloading/negative_sampler.py -> build/bdist.linux-x86_64/egg/dgl/dataloading
copying build/lib.linux-x86_64-3.7/dgl/dataloading/__init__.py -> build/bdist.linux-x86_64/egg/dgl/dataloading
copying build/lib.linux-x86_64-3.7/dgl/dataloading/cluster_gcn.py -> build/bdist.linux-x86_64/egg/dgl/dataloading
copying build/lib.linux-x86_64-3.7/dgl/dataloading/neighbor_sampler.py -> build/bdist.linux-x86_64/egg/dgl/dataloading
copying build/lib.linux-x86_64-3.7/dgl/dataloading/dist_dataloader.py -> build/bdist.linux-x86_64/egg/dgl/dataloading
creating build/bdist.linux-x86_64/egg/dgl/_dataloading
copying build/lib.linux-x86_64-3.7/dgl/_dataloading/neighbor.py -> build/bdist.linux-x86_64/egg/dgl/_dataloading
copying build/lib.linux-x86_64-3.7/dgl/_dataloading/dataloader.py -> build/bdist.linux-x86_64/egg/dgl/_dataloading
creating build/bdist.linux-x86_64/egg/dgl/_dataloading/pytorch
copying build/lib.linux-x86_64-3.7/dgl/_dataloading/pytorch/dataloader.py -> build/bdist.linux-x86_64/egg/dgl/_dataloading/pytorch
copying build/lib.linux-x86_64-3.7/dgl/_dataloading/pytorch/__init__.py -> build/bdist.linux-x86_64/egg/dgl/_dataloading/pytorch
copying build/lib.linux-x86_64-3.7/dgl/_dataloading/shadow.py -> build/bdist.linux-x86_64/egg/dgl/_dataloading
copying build/lib.linux-x86_64-3.7/dgl/_dataloading/negative_sampler.py -> build/bdist.linux-x86_64/egg/dgl/_dataloading
copying build/lib.linux-x86_64-3.7/dgl/_dataloading/__init__.py -> build/bdist.linux-x86_64/egg/dgl/_dataloading
copying build/lib.linux-x86_64-3.7/dgl/_dataloading/cluster_gcn.py -> build/bdist.linux-x86_64/egg/dgl/_dataloading
copying build/lib.linux-x86_64-3.7/dgl/_api_internal.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-3.7/dgl/heterograph.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/backend
copying build/lib.linux-x86_64-3.7/dgl/backend/backend.py -> build/bdist.linux-x86_64/egg/dgl/backend
creating build/bdist.linux-x86_64/egg/dgl/backend/pytorch
copying build/lib.linux-x86_64-3.7/dgl/backend/pytorch/sparse.py -> build/bdist.linux-x86_64/egg/dgl/backend/pytorch
copying build/lib.linux-x86_64-3.7/dgl/backend/pytorch/tensor.py -> build/bdist.linux-x86_64/egg/dgl/backend/pytorch
copying build/lib.linux-x86_64-3.7/dgl/backend/pytorch/__init__.py -> build/bdist.linux-x86_64/egg/dgl/backend/pytorch
creating build/bdist.linux-x86_64/egg/dgl/backend/mxnet
copying build/lib.linux-x86_64-3.7/dgl/backend/mxnet/sparse.py -> build/bdist.linux-x86_64/egg/dgl/backend/mxnet
copying build/lib.linux-x86_64-3.7/dgl/backend/mxnet/tensor.py -> build/bdist.linux-x86_64/egg/dgl/backend/mxnet
copying build/lib.linux-x86_64-3.7/dgl/backend/mxnet/__init__.py -> build/bdist.linux-x86_64/egg/dgl/backend/mxnet
copying build/lib.linux-x86_64-3.7/dgl/backend/mxnet/sparse_optim.py -> build/bdist.linux-x86_64/egg/dgl/backend/mxnet
copying build/lib.linux-x86_64-3.7/dgl/backend/__init__.py -> build/bdist.linux-x86_64/egg/dgl/backend
creating build/bdist.linux-x86_64/egg/dgl/backend/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/backend/tensorflow/sparse.py -> build/bdist.linux-x86_64/egg/dgl/backend/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/backend/tensorflow/tensor.py -> build/bdist.linux-x86_64/egg/dgl/backend/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/backend/tensorflow/__init__.py -> build/bdist.linux-x86_64/egg/dgl/backend/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/backend/tensorflow/sparse_optim.py -> build/bdist.linux-x86_64/egg/dgl/backend/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/backend/set_default_backend.py -> build/bdist.linux-x86_64/egg/dgl/backend
copying build/lib.linux-x86_64-3.7/dgl/network.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/ops
copying build/lib.linux-x86_64-3.7/dgl/ops/edge_softmax.py -> build/bdist.linux-x86_64/egg/dgl/ops
copying build/lib.linux-x86_64-3.7/dgl/ops/sddmm.py -> build/bdist.linux-x86_64/egg/dgl/ops
copying build/lib.linux-x86_64-3.7/dgl/ops/gather_mm.py -> build/bdist.linux-x86_64/egg/dgl/ops
copying build/lib.linux-x86_64-3.7/dgl/ops/segment.py -> build/bdist.linux-x86_64/egg/dgl/ops
copying build/lib.linux-x86_64-3.7/dgl/ops/__init__.py -> build/bdist.linux-x86_64/egg/dgl/ops
copying build/lib.linux-x86_64-3.7/dgl/ops/spmm.py -> build/bdist.linux-x86_64/egg/dgl/ops
copying build/lib.linux-x86_64-3.7/dgl/logging.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-3.7/dgl/graph_index.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/cuda
copying build/lib.linux-x86_64-3.7/dgl/cuda/nccl.py -> build/bdist.linux-x86_64/egg/dgl/cuda
copying build/lib.linux-x86_64-3.7/dgl/cuda/__init__.py -> build/bdist.linux-x86_64/egg/dgl/cuda
copying build/lib.linux-x86_64-3.7/dgl/udf.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/storages
copying build/lib.linux-x86_64-3.7/dgl/storages/pytorch_tensor.py -> build/bdist.linux-x86_64/egg/dgl/storages
copying build/lib.linux-x86_64-3.7/dgl/storages/base.py -> build/bdist.linux-x86_64/egg/dgl/storages
copying build/lib.linux-x86_64-3.7/dgl/storages/tensor.py -> build/bdist.linux-x86_64/egg/dgl/storages
copying build/lib.linux-x86_64-3.7/dgl/storages/__init__.py -> build/bdist.linux-x86_64/egg/dgl/storages
copying build/lib.linux-x86_64-3.7/dgl/storages/numpy.py -> build/bdist.linux-x86_64/egg/dgl/storages
copying build/lib.linux-x86_64-3.7/dgl/ndarray.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-3.7/dgl/init.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-3.7/dgl/view.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/sampling
copying build/lib.linux-x86_64-3.7/dgl/sampling/neighbor.py -> build/bdist.linux-x86_64/egg/dgl/sampling
copying build/lib.linux-x86_64-3.7/dgl/sampling/pinsage.py -> build/bdist.linux-x86_64/egg/dgl/sampling
copying build/lib.linux-x86_64-3.7/dgl/sampling/negative.py -> build/bdist.linux-x86_64/egg/dgl/sampling
copying build/lib.linux-x86_64-3.7/dgl/sampling/randomwalks.py -> build/bdist.linux-x86_64/egg/dgl/sampling
copying build/lib.linux-x86_64-3.7/dgl/sampling/utils.py -> build/bdist.linux-x86_64/egg/dgl/sampling
copying build/lib.linux-x86_64-3.7/dgl/sampling/node2vec_randomwalk.py -> build/bdist.linux-x86_64/egg/dgl/sampling
copying build/lib.linux-x86_64-3.7/dgl/sampling/__init__.py -> build/bdist.linux-x86_64/egg/dgl/sampling
copying build/lib.linux-x86_64-3.7/dgl/generators.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/qm7b.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/graph_serialize.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/dgl_dataset.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/csv_dataset_base.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/knowledge_graph.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/qm9_edge.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/gnn_benchmark.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/qm9.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/icews18.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/gdelt.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/utils.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/sbm.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/csv_dataset.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/karate.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/citation_graph.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/rdf.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/bitcoinotc.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/ppi.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/tensor_serialize.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/flickr.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/wikics.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/heterograph_serialize.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/tu.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/__init__.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/gindt.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/tree.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/synthetic.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/adapter.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/fakenews.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/fraud.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/minigc.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/yelp.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/data/reddit.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-3.7/dgl/convert.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-3.7/dgl/container.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-3.7/dgl/merge.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-3.7/dgl/readout.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-3.7/dgl/__init__.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/dist_graph.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/graph_partition_book.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/partition.py -> build/bdist.linux-x86_64/egg/dgl/distributed
creating build/bdist.linux-x86_64/egg/dgl/distributed/nn
creating build/bdist.linux-x86_64/egg/dgl/distributed/nn/pytorch
copying build/lib.linux-x86_64-3.7/dgl/distributed/nn/pytorch/sparse_emb.py -> build/bdist.linux-x86_64/egg/dgl/distributed/nn/pytorch
copying build/lib.linux-x86_64-3.7/dgl/distributed/nn/pytorch/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distributed/nn/pytorch
creating build/bdist.linux-x86_64/egg/dgl/distributed/nn/mxnet
copying build/lib.linux-x86_64-3.7/dgl/distributed/nn/mxnet/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distributed/nn/mxnet
copying build/lib.linux-x86_64-3.7/dgl/distributed/nn/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distributed/nn
creating build/bdist.linux-x86_64/egg/dgl/distributed/nn/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/distributed/nn/tensorflow/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distributed/nn/tensorflow
creating build/bdist.linux-x86_64/egg/dgl/distributed/optim
creating build/bdist.linux-x86_64/egg/dgl/distributed/optim/pytorch
copying build/lib.linux-x86_64-3.7/dgl/distributed/optim/pytorch/utils.py -> build/bdist.linux-x86_64/egg/dgl/distributed/optim/pytorch
copying build/lib.linux-x86_64-3.7/dgl/distributed/optim/pytorch/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distributed/optim/pytorch
copying build/lib.linux-x86_64-3.7/dgl/distributed/optim/pytorch/sparse_optim.py -> build/bdist.linux-x86_64/egg/dgl/distributed/optim/pytorch
creating build/bdist.linux-x86_64/egg/dgl/distributed/optim/mxnet
copying build/lib.linux-x86_64-3.7/dgl/distributed/optim/mxnet/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distributed/optim/mxnet
copying build/lib.linux-x86_64-3.7/dgl/distributed/optim/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distributed/optim
creating build/bdist.linux-x86_64/egg/dgl/distributed/optim/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/distributed/optim/tensorflow/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distributed/optim/tensorflow
copying build/lib.linux-x86_64-3.7/dgl/distributed/standalone_kvstore.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/graph_services.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/role.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/constants.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/rpc_server.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/kvstore.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/dist_context.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/rpc.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/rpc_client.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/id_map.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/server_state.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/shared_mem_utils.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/dist_tensor.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-3.7/dgl/distributed/dist_dataloader.py -> build/bdist.linux-x86_64/egg/dgl/distributed
creating build/bdist.linux-x86_64/egg/dgl/_deprecate
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/nodeflow.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/udf.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/view.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/kernel.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/__init__.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate
creating build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/spmv.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/degree_bucketing.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/__init__.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/scheduler.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/adapter.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/runtime.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime
creating build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/ir
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/ir/program.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/ir
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/ir/executor.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/ir
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/ir/var.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/ir
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/ir/registry.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/ir
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/runtime/ir/__init__.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/ir
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/graph.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate
copying build/lib.linux-x86_64-3.7/dgl/_deprecate/frame.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate
copying build/lib.linux-x86_64-3.7/dgl/propagate.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-3.7/dgl/batch.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/transforms
copying build/lib.linux-x86_64-3.7/dgl/transforms/functional.py -> build/bdist.linux-x86_64/egg/dgl/transforms
copying build/lib.linux-x86_64-3.7/dgl/transforms/__init__.py -> build/bdist.linux-x86_64/egg/dgl/transforms
copying build/lib.linux-x86_64-3.7/dgl/transforms/module.py -> build/bdist.linux-x86_64/egg/dgl/transforms
creating build/bdist.linux-x86_64/egg/dgl/distgnn
creating build/bdist.linux-x86_64/egg/dgl/distgnn/partition
copying build/lib.linux-x86_64-3.7/dgl/distgnn/partition/libra_partition.py -> build/bdist.linux-x86_64/egg/dgl/distgnn/partition
copying build/lib.linux-x86_64-3.7/dgl/distgnn/partition/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distgnn/partition
creating build/bdist.linux-x86_64/egg/dgl/distgnn/tools
copying build/lib.linux-x86_64-3.7/dgl/distgnn/tools/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distgnn/tools
copying build/lib.linux-x86_64-3.7/dgl/distgnn/tools/tools.py -> build/bdist.linux-x86_64/egg/dgl/distgnn/tools
copying build/lib.linux-x86_64-3.7/dgl/distgnn/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distgnn
creating build/bdist.linux-x86_64/egg/dgl/_ffi
copying build/lib.linux-x86_64-3.7/dgl/_ffi/runtime_ctypes.py -> build/bdist.linux-x86_64/egg/dgl/_ffi
copying build/lib.linux-x86_64-3.7/dgl/_ffi/function.py -> build/bdist.linux-x86_64/egg/dgl/_ffi
copying build/lib.linux-x86_64-3.7/dgl/_ffi/libinfo.py -> build/bdist.linux-x86_64/egg/dgl/_ffi
copying build/lib.linux-x86_64-3.7/dgl/_ffi/streams.py -> build/bdist.linux-x86_64/egg/dgl/_ffi
copying build/lib.linux-x86_64-3.7/dgl/_ffi/base.py -> build/bdist.linux-x86_64/egg/dgl/_ffi
creating build/bdist.linux-x86_64/egg/dgl/_ffi/_ctypes
copying build/lib.linux-x86_64-3.7/dgl/_ffi/_ctypes/function.py -> build/bdist.linux-x86_64/egg/dgl/_ffi/_ctypes
copying build/lib.linux-x86_64-3.7/dgl/_ffi/_ctypes/types.py -> build/bdist.linux-x86_64/egg/dgl/_ffi/_ctypes
copying build/lib.linux-x86_64-3.7/dgl/_ffi/_ctypes/ndarray.py -> build/bdist.linux-x86_64/egg/dgl/_ffi/_ctypes
copying build/lib.linux-x86_64-3.7/dgl/_ffi/_ctypes/__init__.py -> build/bdist.linux-x86_64/egg/dgl/_ffi/_ctypes
copying build/lib.linux-x86_64-3.7/dgl/_ffi/_ctypes/object.py -> build/bdist.linux-x86_64/egg/dgl/_ffi/_ctypes
copying build/lib.linux-x86_64-3.7/dgl/_ffi/ndarray.py -> build/bdist.linux-x86_64/egg/dgl/_ffi
copying build/lib.linux-x86_64-3.7/dgl/_ffi/object_generic.py -> build/bdist.linux-x86_64/egg/dgl/_ffi
copying build/lib.linux-x86_64-3.7/dgl/_ffi/__init__.py -> build/bdist.linux-x86_64/egg/dgl/_ffi
copying build/lib.linux-x86_64-3.7/dgl/_ffi/object.py -> build/bdist.linux-x86_64/egg/dgl/_ffi
creating build/bdist.linux-x86_64/egg/dgl/_ffi/_cy3
copying build/lib.linux-x86_64-3.7/dgl/_ffi/_cy3/__init__.py -> build/bdist.linux-x86_64/egg/dgl/_ffi/_cy3
copying build/lib.linux-x86_64-3.7/dgl/_ffi/_cy3/core.cpython-37m-x86_64-linux-gnu.so -> build/bdist.linux-x86_64/egg/dgl/_ffi/_cy3
creating build/bdist.linux-x86_64/egg/dgl/_ffi/_cy2
copying build/lib.linux-x86_64-3.7/dgl/_ffi/_cy2/__init__.py -> build/bdist.linux-x86_64/egg/dgl/_ffi/_cy2
creating build/bdist.linux-x86_64/egg/dgl/_ffi/_cython
copying build/lib.linux-x86_64-3.7/dgl/_ffi/_cython/core.cpp -> build/bdist.linux-x86_64/egg/dgl/_ffi/_cython
copying build/lib.linux-x86_64-3.7/dgl/frame.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-3.7/dgl/random.py -> build/bdist.linux-x86_64/egg/dgl
byte-compiling build/bdist.linux-x86_64/egg/dgl/mock_sparse/sp_matrix.py to sp_matrix.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/mock_sparse/diag_matrix.py to diag_matrix.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/mock_sparse/reduction.py to reduction.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/mock_sparse/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/mock_sparse/elementwise_op_sp.py to elementwise_op_sp.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/geometry/edge_coarsening.py to edge_coarsening.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/geometry/fps.py to fps.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/geometry/capi.py to capi.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/geometry/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/function/base.py to base.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/function/message.py to message.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/function/reducer.py to reducer.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/function/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/partition.py to partition.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/glob.py to glob.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/link/transe.py to transe.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/link/edgepred.py to edgepred.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/link/transr.py to transr.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/link/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/hetero.py to hetero.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/sparse_emb.py to sparse_emb.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/linear.py to linear.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/utils.py to utils.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/agnnconv.py to agnnconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/relgraphconv.py to relgraphconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/appnpconv.py to appnpconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/dgnconv.py to dgnconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/grouprevres.py to grouprevres.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/twirlsconv.py to twirlsconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/egatconv.py to egatconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/graphconv.py to graphconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/densechebconv.py to densechebconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/gcn2conv.py to gcn2conv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/ginconv.py to ginconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/dotgatconv.py to dotgatconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/cfconv.py to cfconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/gmmconv.py to gmmconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/sageconv.py to sageconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/densesageconv.py to densesageconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/gineconv.py to gineconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/edgeconv.py to edgeconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/nnconv.py to nnconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/gatv2conv.py to gatv2conv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/hgtconv.py to hgtconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/pnaconv.py to pnaconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/gatedgraphconv.py to gatedgraphconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/atomicconv.py to atomicconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/tagconv.py to tagconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/densegraphconv.py to densegraphconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/egnnconv.py to egnnconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/chebconv.py to chebconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/sgconv.py to sgconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/gatconv.py to gatconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/explain/gnnexplainer.py to gnnexplainer.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/explain/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/factory.py to factory.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/softmax.py to softmax.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/glob.py to glob.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/hetero.py to hetero.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/utils.py to utils.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/agnnconv.py to agnnconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/relgraphconv.py to relgraphconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/appnpconv.py to appnpconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/graphconv.py to graphconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/densechebconv.py to densechebconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/ginconv.py to ginconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/gmmconv.py to gmmconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/sageconv.py to sageconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/densesageconv.py to densesageconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/edgeconv.py to edgeconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/nnconv.py to nnconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/gatedgraphconv.py to gatedgraphconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/tagconv.py to tagconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/densegraphconv.py to densegraphconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/chebconv.py to chebconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/sgconv.py to sgconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/gatconv.py to gatconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/softmax.py to softmax.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/glob.py to glob.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/hetero.py to hetero.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/utils.py to utils.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv/relgraphconv.py to relgraphconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv/appnpconv.py to appnpconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv/graphconv.py to graphconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv/densechebconv.py to densechebconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv/ginconv.py to ginconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv/sageconv.py to sageconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv/edgeconv.py to edgeconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv/chebconv.py to chebconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv/sgconv.py to sgconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv/gatconv.py to gatconv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/softmax.py to softmax.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/functional/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/core.py to core.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/subgraph.py to subgraph.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/contrib/dis_kvstore.py to dis_kvstore.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/contrib/unified_tensor.py to unified_tensor.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/contrib/graph_store.py to graph_store.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/contrib/sampling/sampler.py to sampler.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/contrib/sampling/dis_sampler.py to dis_sampler.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/contrib/sampling/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/contrib/data/knowledge_graph.py to knowledge_graph.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/contrib/data/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/contrib/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/traversal.py to traversal.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/utils/checks.py to checks.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/utils/exception.py to exception.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/utils/pin_memory.py to pin_memory.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/utils/internal.py to internal.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/utils/filter.py to filter.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/utils/data.py to data.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/utils/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/utils/shared_mem.py to shared_mem.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/multiprocessing/pytorch.py to pytorch.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/multiprocessing/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/base.py to base.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/optim/pytorch/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/optim/pytorch/sparse_optim.py to sparse_optim.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/optim/mxnet/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/optim/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/optim/tensorflow/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/heterograph_index.py to heterograph_index.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/sparse.py to sparse.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/global_config.py to global_config.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/dataloading/base.py to base.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/dataloading/dataloader.py to dataloader.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/dataloading/graphsaint.py to graphsaint.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/dataloading/shadow.py to shadow.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/dataloading/negative_sampler.py to negative_sampler.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/dataloading/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/dataloading/cluster_gcn.py to cluster_gcn.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/dataloading/neighbor_sampler.py to neighbor_sampler.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/dataloading/dist_dataloader.py to dist_dataloader.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_dataloading/neighbor.py to neighbor.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_dataloading/dataloader.py to dataloader.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_dataloading/pytorch/dataloader.py to dataloader.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_dataloading/pytorch/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_dataloading/shadow.py to shadow.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_dataloading/negative_sampler.py to negative_sampler.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_dataloading/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_dataloading/cluster_gcn.py to cluster_gcn.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_api_internal.py to _api_internal.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/heterograph.py to heterograph.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/backend.py to backend.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/pytorch/sparse.py to sparse.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/pytorch/tensor.py to tensor.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/pytorch/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/mxnet/sparse.py to sparse.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/mxnet/tensor.py to tensor.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/mxnet/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/mxnet/sparse_optim.py to sparse_optim.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/tensorflow/sparse.py to sparse.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/tensorflow/tensor.py to tensor.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/tensorflow/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/tensorflow/sparse_optim.py to sparse_optim.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/set_default_backend.py to set_default_backend.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/network.py to network.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/ops/edge_softmax.py to edge_softmax.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/ops/sddmm.py to sddmm.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/ops/gather_mm.py to gather_mm.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/ops/segment.py to segment.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/ops/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/ops/spmm.py to spmm.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/logging.py to logging.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/graph_index.py to graph_index.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/cuda/nccl.py to nccl.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/cuda/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/udf.py to udf.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/storages/pytorch_tensor.py to pytorch_tensor.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/storages/base.py to base.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/storages/tensor.py to tensor.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/storages/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/storages/numpy.py to numpy.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/ndarray.py to ndarray.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/init.py to init.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/view.py to view.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/sampling/neighbor.py to neighbor.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/sampling/pinsage.py to pinsage.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/sampling/negative.py to negative.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/sampling/randomwalks.py to randomwalks.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/sampling/utils.py to utils.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/sampling/node2vec_randomwalk.py to node2vec_randomwalk.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/sampling/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/generators.py to generators.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/qm7b.py to qm7b.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/graph_serialize.py to graph_serialize.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/dgl_dataset.py to dgl_dataset.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/csv_dataset_base.py to csv_dataset_base.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/knowledge_graph.py to knowledge_graph.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/qm9_edge.py to qm9_edge.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/gnn_benchmark.py to gnn_benchmark.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/qm9.py to qm9.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/icews18.py to icews18.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/gdelt.py to gdelt.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/utils.py to utils.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/sbm.py to sbm.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/csv_dataset.py to csv_dataset.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/karate.py to karate.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/citation_graph.py to citation_graph.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/rdf.py to rdf.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/bitcoinotc.py to bitcoinotc.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/ppi.py to ppi.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/tensor_serialize.py to tensor_serialize.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/flickr.py to flickr.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/wikics.py to wikics.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/heterograph_serialize.py to heterograph_serialize.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/tu.py to tu.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/gindt.py to gindt.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/tree.py to tree.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/synthetic.py to synthetic.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/adapter.py to adapter.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/fakenews.py to fakenews.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/fraud.py to fraud.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/minigc.py to minigc.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/yelp.py to yelp.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/reddit.py to reddit.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/convert.py to convert.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/container.py to container.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/merge.py to merge.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/readout.py to readout.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/dist_graph.py to dist_graph.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/graph_partition_book.py to graph_partition_book.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/partition.py to partition.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/nn/pytorch/sparse_emb.py to sparse_emb.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/nn/pytorch/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/nn/mxnet/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/nn/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/nn/tensorflow/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/optim/pytorch/utils.py to utils.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/optim/pytorch/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/optim/pytorch/sparse_optim.py to sparse_optim.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/optim/mxnet/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/optim/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/optim/tensorflow/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/standalone_kvstore.py to standalone_kvstore.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/graph_services.py to graph_services.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/role.py to role.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/constants.py to constants.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/rpc_server.py to rpc_server.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/kvstore.py to kvstore.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/dist_context.py to dist_context.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/rpc.py to rpc.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/rpc_client.py to rpc_client.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/id_map.py to id_map.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/server_state.py to server_state.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/shared_mem_utils.py to shared_mem_utils.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/dist_tensor.py to dist_tensor.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/dist_dataloader.py to dist_dataloader.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/nodeflow.py to nodeflow.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/udf.py to udf.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/view.py to view.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/kernel.py to kernel.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/spmv.py to spmv.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/degree_bucketing.py to degree_bucketing.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/scheduler.py to scheduler.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/adapter.py to adapter.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/runtime.py to runtime.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/ir/program.py to program.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/ir/executor.py to executor.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/ir/var.py to var.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/ir/registry.py to registry.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/ir/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/graph.py to graph.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/frame.py to frame.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/propagate.py to propagate.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/batch.py to batch.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/transforms/functional.py to functional.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/transforms/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/transforms/module.py to module.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distgnn/partition/libra_partition.py to libra_partition.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distgnn/partition/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distgnn/tools/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distgnn/tools/tools.py to tools.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distgnn/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/runtime_ctypes.py to runtime_ctypes.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/function.py to function.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/libinfo.py to libinfo.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/streams.py to streams.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/base.py to base.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/_ctypes/function.py to function.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/_ctypes/types.py to types.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/_ctypes/ndarray.py to ndarray.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/_ctypes/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/_ctypes/object.py to object.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/ndarray.py to ndarray.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/object_generic.py to object_generic.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/object.py to object.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/_cy3/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/_cy2/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/frame.py to frame.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/random.py to random.cpython-37.pyc
creating stub loader for dgl/_ffi/_cy3/core.cpython-37m-x86_64-linux-gnu.so
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/_cy3/core.py to core.cpython-37.pyc
installing package data to build/bdist.linux-x86_64/egg
running install_data
copying ../build/libdgl.so -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/tensoradapter
creating build/bdist.linux-x86_64/egg/dgl/tensoradapter/pytorch
copying ../build/tensoradapter/pytorch/libtensoradapter_pytorch_1.9.0.so -> build/bdist.linux-x86_64/egg/dgl/tensoradapter/pytorch
creating build/bdist.linux-x86_64/egg/EGG-INFO
copying dgl.egg-info/PKG-INFO -> build/bdist.linux-x86_64/egg/EGG-INFO
copying dgl.egg-info/SOURCES.txt -> build/bdist.linux-x86_64/egg/EGG-INFO
copying dgl.egg-info/dependency_links.txt -> build/bdist.linux-x86_64/egg/EGG-INFO
copying dgl.egg-info/not-zip-safe -> build/bdist.linux-x86_64/egg/EGG-INFO
copying dgl.egg-info/requires.txt -> build/bdist.linux-x86_64/egg/EGG-INFO
copying dgl.egg-info/top_level.txt -> build/bdist.linux-x86_64/egg/EGG-INFO
writing build/bdist.linux-x86_64/egg/EGG-INFO/native_libs.txt
creating dist
creating 'dist/dgl-0.9-py3.7-linux-x86_64.egg' and adding 'build/bdist.linux-x86_64/egg' to it
removing 'build/bdist.linux-x86_64/egg' (and everything under it)
Processing dgl-0.9-py3.7-linux-x86_64.egg
creating /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/dgl-0.9-py3.7-linux-x86_64.egg
Extracting dgl-0.9-py3.7-linux-x86_64.egg to /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages
Adding dgl 0.9 to easy-install.pth file

Installed /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/dgl-0.9-py3.7-linux-x86_64.egg
Processing dependencies for dgl==0.9
Searching for psutil>=5.8.0
Reading https://pypi.org/simple/psutil/
Downloading https://files.pythonhosted.org/packages/3d/73/d8c87b5612c58d1e6c6d91997c1590771d34e4ee27d9c11eb1e64ecbf365/psutil-5.9.2-cp37-cp37m-manylinux_2_12_x86_64.manylinux2010_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl#sha256=4fb54941aac044a61db9d8eb56fc5bee207db3bc58645d657249030e15ba3727
Best match: psutil 5.9.2
Processing psutil-5.9.2-cp37-cp37m-manylinux_2_12_x86_64.manylinux2010_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl
Installing psutil-5.9.2-cp37-cp37m-manylinux_2_12_x86_64.manylinux2010_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl to /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages
Adding psutil 5.9.2 to easy-install.pth file

Installed /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/psutil-5.9.2-py3.7-linux-x86_64.egg
Searching for tqdm==4.64.0
Best match: tqdm 4.64.0
Adding tqdm 4.64.0 to easy-install.pth file
Installing tqdm script to /opt/conda/envs/tensorflow-ci/bin

Using /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages
Searching for requests==2.28.1
Best match: requests 2.28.1
Adding requests 2.28.1 to easy-install.pth file

Using /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages
Searching for networkx==2.6.3
Best match: networkx 2.6.3
Adding networkx 2.6.3 to easy-install.pth file

Using /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages
Searching for scipy==1.4.1
Best match: scipy 1.4.1
Adding scipy 1.4.1 to easy-install.pth file

Using /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages
Searching for numpy==1.18.5
Best match: numpy 1.18.5
Adding numpy 1.18.5 to easy-install.pth file
Installing f2py script to /opt/conda/envs/tensorflow-ci/bin
Installing f2py3 script to /opt/conda/envs/tensorflow-ci/bin
Installing f2py3.7 script to /opt/conda/envs/tensorflow-ci/bin

Using /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages
Searching for certifi==2022.6.15
Best match: certifi 2022.6.15
Adding certifi 2022.6.15 to easy-install.pth file

Using /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages
Searching for urllib3==1.26.11
Best match: urllib3 1.26.11
Adding urllib3 1.26.11 to easy-install.pth file

Using /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages
Searching for idna==3.3
Best match: idna 3.3
Adding idna 3.3 to easy-install.pth file

Using /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages
Searching for charset-normalizer==2.1.0
Best match: charset-normalizer 2.1.0
Adding charset-normalizer 2.1.0 to easy-install.pth file
Installing normalizer script to /opt/conda/envs/tensorflow-ci/bin

Using /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages
Finished processing dependencies for dgl==0.9
/opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/Cython/Compiler/Main.py:369: FutureWarning: Cython directive 'language_level' not set, using 2 for now (Py2). This will change in a later release! File: /root/jenkins/workspace/dgl_PR-4648@2/python/dgl/_ffi/_cython/core.pyx
  tree = Parsing.p_module(s, pxd, full_module_name)
/opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/setuptools/command/install.py:37: SetuptoolsDeprecationWarning: setup.py install is deprecated. Use build and pip and other standards-based tools.
  setuptools.SetuptoolsDeprecationWarning,
/opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/setuptools/command/easy_install.py:147: EasyInstallDeprecationWarning: easy_install command is deprecated. Use build and pip and other standards-based tools.
  EasyInstallDeprecationWarning,
/opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/pkg_resources/__init__.py:125: PkgResourcesDeprecationWarning:  is an invalid version and will not be supported in a future release
  PkgResourcesDeprecationWarning,
[1/1] Cythonizing dgl/_ffi/_cython/core.pyx
running build_ext
building 'dgl._ffi._cy3.core' extension
gcc -pthread -B /opt/conda/envs/tensorflow-ci/compiler_compat -Wl,--sysroot=/ -Wsign-compare -DNDEBUG -g -fwrapv -O3 -Wall -Wstrict-prototypes -fPIC -I../include/ -I../third_party/dmlc-core/include -I../third_party/dlpack/include -I/opt/conda/envs/tensorflow-ci/include/python3.7m -c dgl/_ffi/_cython/core.cpp -o build/temp.linux-x86_64-3.7/dgl/_ffi/_cython/core.o
cc1plus: warning: command line option ‘-Wstrict-prototypes’ is valid for C/ObjC but not for C++
g++ -pthread -B /opt/conda/envs/tensorflow-ci/compiler_compat -Wl,--sysroot=/ -pthread -shared -B /opt/conda/envs/tensorflow-ci/compiler_compat -L/opt/conda/envs/tensorflow-ci/lib -Wl,-rpath=/opt/conda/envs/tensorflow-ci/lib -Wl,--no-as-needed -Wl,--sysroot=/ build/temp.linux-x86_64-3.7/dgl/_ffi/_cython/core.o -Ldgl -L../build/Release -L../build -ldgl -o build/lib.linux-x86_64-3.7/dgl/_ffi/_cy3/core.cpython-37m-x86_64-linux-gnu.so
/opt/conda/envs/tensorflow-ci/compiler_compat/ld: warning: /opt/conda/envs/tensorflow-ci/lib/libstdc++.so: unsupported GNU_PROPERTY_TYPE (5) type: 0xc0010001
/opt/conda/envs/tensorflow-ci/compiler_compat/ld: warning: /opt/conda/envs/tensorflow-ci/lib/libstdc++.so: unsupported GNU_PROPERTY_TYPE (5) type: 0xc0010002
/opt/conda/envs/tensorflow-ci/compiler_compat/ld: warning: /opt/conda/envs/tensorflow-ci/lib/libgcc_s.so.1: unsupported GNU_PROPERTY_TYPE (5) type: 0xc0010001
/opt/conda/envs/tensorflow-ci/compiler_compat/ld: warning: /opt/conda/envs/tensorflow-ci/lib/libgcc_s.so.1: unsupported GNU_PROPERTY_TYPE (5) type: 0xc0010002
/opt/conda/envs/tensorflow-ci/compiler_compat/ld: warning: /opt/conda/envs/tensorflow-ci/lib/libgcc_s.so.1: unsupported GNU_PROPERTY_TYPE (5) type: 0xc0010001
/opt/conda/envs/tensorflow-ci/compiler_compat/ld: warning: /opt/conda/envs/tensorflow-ci/lib/libgcc_s.so.1: unsupported GNU_PROPERTY_TYPE (5) type: 0xc0010002
copying build/lib.linux-x86_64-3.7/dgl/_ffi/_cy3/core.cpython-37m-x86_64-linux-gnu.so -> dgl/_ffi/_cy3
/opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/Cython/Compiler/Main.py:369: FutureWarning: Cython directive 'language_level' not set, using 2 for now (Py2). This will change in a later release! File: /root/jenkins/workspace/dgl_PR-4648@2/python/dgl/_ffi/_cython/core.pyx
  tree = Parsing.p_module(s, pxd, full_module_name)
~/jenkins/workspace/dgl_PR-4648@2
[Pipeline] sh
+ ls -lh /usr/lib/x86_64-linux-gnu/
total 1.8G
-rw-r--r--  1 root root   496 May  3 10:19 Mcrt1.o
-rw-r--r--  1 root root  1.8K May  3 10:19 Scrt1.o
drwxr-xr-x  1 root root  4.0K May 28 19:15 audit
drwxr-xr-x  2 root root  4.0K Aug 16 08:53 avahi
drwxr-xr-x  2 root root  4.0K Apr 28 00:03 coreutils
-rw-r--r--  1 root root  1.9K May  3 10:19 crt1.o
-rw-r--r--  1 root root  1.2K May  3 10:19 crti.o
-rw-r--r--  1 root root   648 May  3 10:19 crtn.o
drwxr-xr-x  2 root root  4.0K Aug 16 08:53 dri
drwxr-xr-x  2 root root  4.0K May 23 21:53 engines-1.1
drwxr-xr-x  1 root root   12K May 28 19:15 gconv
-rw-r--r--  1 root root  2.5K May  3 10:19 gcrt1.o
drwxr-xr-x  2 root root  4.0K Aug 16 08:53 gdcm-2.8
drwxr-xr-x  3 root root  4.0K Aug 16 08:53 gdk-pixbuf-2.0
drwxr-xr-x  3 root root  4.0K Aug 16 08:53 gio
drwxr-xr-x  2 root root  4.0K Aug 16 08:53 girepository-1.0
drwxr-xr-x  3 root root  4.0K Aug 16 08:53 glib-2.0
drwxr-xr-x  2 root root  4.0K Aug 16 08:54 graphviz
-rw-r--r--  1 root root  2.3K May  3 10:19 grcrt1.o
drwxr-xr-x  3 root root  4.0K Aug 16 08:53 gtk-2.0
drwxr-xr-x  3 root root  4.0K Aug 16 08:53 gtk-3.0
drwxr-xr-x  2 root root  4.0K Aug 16 08:53 hwloc
drwxr-xr-x  3 root root  4.0K Aug 16 08:53 icu
drwxr-xr-x  3 root root  4.0K Aug 16 08:53 krb5
drwxr-xr-x  2 root root  4.0K May 28 19:15 ldscripts
-rw-r--r--  1 root root  2.0K May  3 10:19 libBrokenLocale.a
lrwxrwxrwx  1 root root    42 May  3 10:19 libBrokenLocale.so -> /lib/x86_64-linux-gnu/libBrokenLocale.so.1
lrwxrwxrwx  1 root root    16 May 23  2016 libCharLS.so.1 -> libCharLS.so.1.0
-rw-r--r--  1 root root  219K May 23  2016 libCharLS.so.1.0
lrwxrwxrwx  1 root root    14 May 10  2019 libGL.so.1 -> libGL.so.1.0.0
-rw-r--r--  1 root root  555K May 10  2019 libGL.so.1.0.0
lrwxrwxrwx  1 root root    15 May 10  2019 libGLX.so.0 -> libGLX.so.0.0.0
-rw-r--r--  1 root root   67K May 10  2019 libGLX.so.0.0.0
lrwxrwxrwx  1 root root    16 Jun 12  2020 libGLX_indirect.so.0 -> libGLX_mesa.so.0
lrwxrwxrwx  1 root root    20 Jun 12  2020 libGLX_mesa.so.0 -> libGLX_mesa.so.0.0.0
-rw-r--r--  1 root root  477K Jun 12  2020 libGLX_mesa.so.0.0.0
lrwxrwxrwx  1 root root    22 May 10  2019 libGLdispatch.so.0 -> libGLdispatch.so.0.0.0
-rw-r--r--  1 root root  599K May 10  2019 libGLdispatch.so.0.0.0
lrwxrwxrwx  1 root root    17 Apr 16  2016 libHalf.so -> libHalf.so.12.0.0
lrwxrwxrwx  1 root root    17 Apr 16  2016 libHalf.so.12 -> libHalf.so.12.0.0
-rw-r--r--  1 root root  267K Apr 16  2016 libHalf.so.12.0.0
lrwxrwxrwx  1 root root    15 Feb 28  2017 libICE.so.6 -> libICE.so.6.3.0
-rw-r--r--  1 root root   96K Feb 28  2017 libICE.so.6.3.0
lrwxrwxrwx  1 root root    20 Apr 16  2016 libIex-2_2.so.12 -> libIex-2_2.so.12.0.0
-rw-r--r--  1 root root  119K Apr 16  2016 libIex-2_2.so.12.0.0
lrwxrwxrwx  1 root root    20 Apr 16  2016 libIex.so -> libIex-2_2.so.12.0.0
lrwxrwxrwx  1 root root    24 Apr 16  2016 libIexMath-2_2.so.12 -> libIexMath-2_2.so.12.0.0
-rw-r--r--  1 root root   15K Apr 16  2016 libIexMath-2_2.so.12.0.0
lrwxrwxrwx  1 root root    24 Apr 16  2016 libIexMath.so -> libIexMath-2_2.so.12.0.0
lrwxrwxrwx  1 root root    23 Nov 16  2021 libIlmImf-2_2.so.22 -> libIlmImf-2_2.so.22.0.0
-rw-r--r--  1 root root  2.8M Nov 16  2021 libIlmImf-2_2.so.22.0.0
-rw-r--r--  1 root root  4.4M Nov 16  2021 libIlmImf.a
lrwxrwxrwx  1 root root    23 Nov 16  2021 libIlmImf.so -> libIlmImf-2_2.so.22.0.0
lrwxrwxrwx  1 root root    27 Nov 16  2021 libIlmImfUtil-2_2.so.22 -> libIlmImfUtil-2_2.so.22.0.0
-rw-r--r--  1 root root  140K Nov 16  2021 libIlmImfUtil-2_2.so.22.0.0
-rw-r--r--  1 root root  319K Nov 16  2021 libIlmImfUtil.a
lrwxrwxrwx  1 root root    27 Nov 16  2021 libIlmImfUtil.so -> libIlmImfUtil-2_2.so.22.0.0
lrwxrwxrwx  1 root root    26 Apr 16  2016 libIlmThread-2_2.so.12 -> libIlmThread-2_2.so.12.0.0
-rw-r--r--  1 root root   27K Apr 16  2016 libIlmThread-2_2.so.12.0.0
lrwxrwxrwx  1 root root    26 Apr 16  2016 libIlmThread.so -> libIlmThread-2_2.so.12.0.0
lrwxrwxrwx  1 root root    22 Apr 16  2016 libImath-2_2.so.12 -> libImath-2_2.so.12.0.0
-rw-r--r--  1 root root   71K Apr 16  2016 libImath-2_2.so.12.0.0
lrwxrwxrwx  1 root root    22 Apr 16  2016 libImath.so -> libImath-2_2.so.12.0.0
lrwxrwxrwx  1 root root    15 Jul 14  2020 libLLVM-10.so -> libLLVM-10.so.1
-rw-r--r--  1 root root   71M Jul 14  2020 libLLVM-10.so.1
lrwxrwxrwx  1 root root    18 Apr  5  2017 libOpenCL.so.1 -> libOpenCL.so.1.0.0
-rw-r--r--  1 root root   43K Apr  5  2017 libOpenCL.so.1.0.0
lrwxrwxrwx  1 root root    14 Jul 13  2014 libSM.so.6 -> libSM.so.6.0.1
-rw-r--r--  1 root root   31K Jul 13  2014 libSM.so.6.0.1
lrwxrwxrwx  1 root root    19 May 19  2021 libX11-xcb.so.1 -> libX11-xcb.so.1.0.0
-rw-r--r--  1 root root  5.7K May 19  2021 libX11-xcb.so.1.0.0
lrwxrwxrwx  1 root root    15 May 19  2021 libX11.so.6 -> libX11.so.6.3.0
-rw-r--r--  1 root root  1.3M May 19  2021 libX11.so.6.3.0
lrwxrwxrwx  1 root root    15 Apr 21  2020 libXau.so.6 -> libXau.so.6.0.0
-rw-r--r--  1 root root   15K Apr 21  2020 libXau.so.6.0.0
lrwxrwxrwx  1 root root    12 Aug  3  2015 libXaw.so.7 -> libXaw7.so.7
lrwxrwxrwx  1 root root    16 Aug  3  2015 libXaw7.so.7 -> libXaw7.so.7.0.0
-rw-r--r--  1 root root  461K Aug  3  2015 libXaw7.so.7.0.0
lrwxrwxrwx  1 root root    22 Jan 20  2017 libXcomposite.so.1 -> libXcomposite.so.1.0.0
-rw-r--r--  1 root root   11K Jan 20  2017 libXcomposite.so.1.0.0
lrwxrwxrwx  1 root root    19 Dec 19  2017 libXcursor.so.1 -> libXcursor.so.1.0.2
-rw-r--r--  1 root root   39K Dec 19  2017 libXcursor.so.1.0.2
lrwxrwxrwx  1 root root    19 Aug 20  2017 libXdamage.so.1 -> libXdamage.so.1.1.0
-rw-r--r--  1 root root   11K Aug 20  2017 libXdamage.so.1.1.0
lrwxrwxrwx  1 root root    17 Mar  2  2017 libXdmcp.so.6 -> libXdmcp.so.6.0.0
-rw-r--r--  1 root root   23K Mar  2  2017 libXdmcp.so.6.0.0
lrwxrwxrwx  1 root root    16 Oct 25  2014 libXext.so.6 -> libXext.so.6.4.0
-rw-r--r--  1 root root   72K Oct 25  2014 libXext.so.6.4.0
lrwxrwxrwx  1 root root    18 Dec  5  2016 libXfixes.so.3 -> libXfixes.so.3.1.0
-rw-r--r--  1 root root   23K Dec  5  2016 libXfixes.so.3.1.0
lrwxrwxrwx  1 root root    15 Jul 13  2014 libXft.so.2 -> libXft.so.2.3.2
-rw-r--r--  1 root root   84K Jul 13  2014 libXft.so.2.3.2
lrwxrwxrwx  1 root root    14 Jan 23  2017 libXi.so.6 -> libXi.so.6.1.0
-rw-r--r--  1 root root   63K Jan 23  2017 libXi.so.6.1.0
lrwxrwxrwx  1 root root    20 Jun 30  2013 libXinerama.so.1 -> libXinerama.so.1.0.0
-rw-r--r--  1 root root   11K Jun 30  2013 libXinerama.so.1.0.0
lrwxrwxrwx  1 root root    15 Dec 10  2015 libXmu.so.6 -> libXmu.so.6.2.0
-rw-r--r--  1 root root  101K Dec 10  2015 libXmu.so.6.2.0
lrwxrwxrwx  1 root root    16 Dec 22  2016 libXpm.so.4 -> libXpm.so.4.11.0
-rw-r--r--  1 root root   71K Dec 22  2016 libXpm.so.4.11.0
lrwxrwxrwx  1 root root    18 Dec  6  2016 libXrandr.so.2 -> libXrandr.so.2.2.0
-rw-r--r--  1 root root   43K Dec  6  2016 libXrandr.so.2.2.0
lrwxrwxrwx  1 root root    19 Dec  5  2016 libXrender.so.1 -> libXrender.so.1.3.0
-rw-r--r--  1 root root   39K Dec  5  2016 libXrender.so.1.3.0
lrwxrwxrwx  1 root root    15 May  9  2012 libXss.so.1 -> libXss.so.1.0.0
-rw-r--r--  1 root root   15K May  9  2012 libXss.so.1.0.0
lrwxrwxrwx  1 root root    14 May 24  2016 libXt.so.6 -> libXt.so.6.0.0
-rw-r--r--  1 root root  416K May 24  2016 libXt.so.6.0.0
lrwxrwxrwx  1 root root    19 May  6  2015 libXxf86vm.so.1 -> libXxf86vm.so.1.0.0
-rw-r--r--  1 root root   23K May  6  2015 libXxf86vm.so.1.0.0
lrwxrwxrwx  1 root root    15 Jul 28  2017 libaec.so.0 -> libaec.so.0.0.3
-rw-r--r--  1 root root   30K Jul 28  2017 libaec.so.0.0.3
-rw-r--r--  1 root root   23K May  3 10:19 libanl.a
lrwxrwxrwx  1 root root    33 May  3 10:19 libanl.so -> /lib/x86_64-linux-gnu/libanl.so.1
lrwxrwxrwx  1 root root    20 Jun 15  2021 libapt-inst.so.2.0 -> libapt-inst.so.2.0.0
-rw-r--r--  1 root root   51K Jun 15  2021 libapt-inst.so.2.0.0
lrwxrwxrwx  1 root root    19 Jun 15  2021 libapt-pkg.so.5.0 -> libapt-pkg.so.5.0.2
-rw-r--r--  1 root root  1.8M Jun 15  2021 libapt-pkg.so.5.0.2
lrwxrwxrwx  1 root root    23 Jun 15  2021 libapt-private.so.0.0 -> libapt-private.so.0.0.0
-rw-r--r--  1 root root  415K Jun 15  2021 libapt-private.so.0.0.0
lrwxrwxrwx  1 root root    18 Jan 24  2018 libarpack.so.2 -> libarpack.so.2.0.0
-rw-r--r--  1 root root  295K Jan 24  2018 libarpack.so.2.0.0
lrwxrwxrwx  1 root root    16 Dec  4  2019 libasan.so.4 -> libasan.so.4.0.0
-rw-r--r--  1 root root  1.4M Dec  4  2019 libasan.so.4.0.0
lrwxrwxrwx  1 root root    16 Dec 15  2017 libasn1.so.8 -> libasn1.so.8.0.0
-rw-r--r--  1 root root  647K Dec 15  2017 libasn1.so.8.0.0
lrwxrwxrwx  1 root root    18 Feb  6  2018 libassuan.so.0 -> libassuan.so.0.8.1
-rw-r--r--  1 root root   75K Feb  6  2018 libassuan.so.0.8.1
lrwxrwxrwx  1 root root    23 Mar 13  2018 libatk-1.0.so.0 -> libatk-1.0.so.0.22810.1
-rw-r--r--  1 root root  150K Mar 13  2018 libatk-1.0.so.0.22810.1
lrwxrwxrwx  1 root root    26 Mar 13  2018 libatk-bridge-2.0.so.0 -> libatk-bridge-2.0.so.0.0.0
-rw-r--r--  1 root root  194K Mar 13  2018 libatk-bridge-2.0.so.0.0.0
lrwxrwxrwx  1 root root    18 Mar 10  2020 libatomic.so.1 -> libatomic.so.1.2.0
-rw-r--r--  1 root root   27K Mar 10  2020 libatomic.so.1.2.0
lrwxrwxrwx  1 root root    17 Mar 13  2018 libatspi.so.0 -> libatspi.so.0.0.1
-rw-r--r--  1 root root  191K Mar 13  2018 libatspi.so.0.0.1
lrwxrwxrwx  1 root root    24 Jul  6  2021 libavahi-client.so.3 -> libavahi-client.so.3.2.9
-rw-r--r--  1 root root   67K Jul  6  2021 libavahi-client.so.3.2.9
lrwxrwxrwx  1 root root    24 Jul  6  2021 libavahi-common.so.3 -> libavahi-common.so.3.5.3
-rw-r--r--  1 root root   47K Jul  6  2021 libavahi-common.so.3.5.3
-rw-r--r--  1 root root   19M May 18 20:01 libavcodec.a
lrwxrwxrwx  1 root root    24 May 18 20:01 libavcodec.so -> libavcodec.so.57.107.100
lrwxrwxrwx  1 root root    24 May 18 20:01 libavcodec.so.57 -> libavcodec.so.57.107.100
-rw-r--r--  1 root root   14M May 18 20:01 libavcodec.so.57.107.100
-rw-r--r--  1 root root  4.7M May 18 20:01 libavformat.a
lrwxrwxrwx  1 root root    24 May 18 20:01 libavformat.so -> libavformat.so.57.83.100
lrwxrwxrwx  1 root root    24 May 18 20:01 libavformat.so.57 -> libavformat.so.57.83.100
-rw-r--r--  1 root root  2.4M May 18 20:01 libavformat.so.57.83.100
-rw-r--r--  1 root root  214K May 18 20:01 libavresample.a
lrwxrwxrwx  1 root root    22 May 18 20:01 libavresample.so -> libavresample.so.3.7.0
lrwxrwxrwx  1 root root    22 May 18 20:01 libavresample.so.3 -> libavresample.so.3.7.0
-rw-r--r--  1 root root  130K May 18 20:01 libavresample.so.3.7.0
-rw-r--r--  1 root root  755K May 18 20:01 libavutil.a
lrwxrwxrwx  1 root root    22 May 18 20:01 libavutil.so -> libavutil.so.55.78.100
lrwxrwxrwx  1 root root    22 May 18 20:01 libavutil.so.55 -> libavutil.so.55.78.100
-rw-r--r--  1 root root  487K May 18 20:01 libavutil.so.55.78.100
-rw-r--r--  1 root root  1.3M Oct 20  2021 libbfd-2.30-system.so
lrwxrwxrwx  1 root root    44 Aug 16 08:54 libblas.a -> /etc/alternatives/libblas.a-x86_64-linux-gnu
lrwxrwxrwx  1 root root    45 Aug 16 08:54 libblas.so -> /etc/alternatives/libblas.so-x86_64-linux-gnu
lrwxrwxrwx  1 root root    47 Aug 16 08:54 libblas.so.3 -> /etc/alternatives/libblas.so.3-x86_64-linux-gnu
lrwxrwxrwx  1 root root    18 Apr  3  2018 libbluray.so.2 -> libbluray.so.2.0.2
-rw-r--r--  1 root root  319K Apr  3  2018 libbluray.so.2.0.2
-rw-r--r--  1 root root  5.3M May  3 10:19 libc.a
-rw-r--r--  1 root root   298 May  3 10:19 libc.so
-rw-r--r--  1 root root   20K May  3 10:19 libc_nonshared.a
lrwxrwxrwx  1 root root    29 Jan 22  2019 libcairo-gobject.so.2 -> libcairo-gobject.so.2.11510.0
-rw-r--r--  1 root root   34K Jan 22  2019 libcairo-gobject.so.2.11510.0
lrwxrwxrwx  1 root root    21 Jan 22  2019 libcairo.so.2 -> libcairo.so.2.11510.0
-rw-r--r--  1 root root  1.2M Jan 22  2019 libcairo.so.2.11510.0
lrwxrwxrwx  1 root root    15 Mar 10  2020 libcc1.so.0 -> libcc1.so.0.0.0
-rw-r--r--  1 root root  116K Mar 10  2020 libcc1.so.0.0.0
lrwxrwxrwx  1 root root    15 Mar 24  2018 libcdt.so -> libcdt.so.5.0.0
lrwxrwxrwx  1 root root    15 Mar 24  2018 libcdt.so.5 -> libcdt.so.5.0.0
-rw-r--r--  1 root root   27K Mar 24  2018 libcdt.so.5.0.0
lrwxrwxrwx  1 root root    18 Mar 24  2018 libcgraph.so -> libcgraph.so.6.0.0
lrwxrwxrwx  1 root root    18 Mar 24  2018 libcgraph.so.6 -> libcgraph.so.6.0.0
-rw-r--r--  1 root root   88K Mar 24  2018 libcgraph.so.6.0.0
lrwxrwxrwx  1 root root    23 Feb  7  2018 libchromaprint.so.1 -> libchromaprint.so.1.4.3
-rw-r--r--  1 root root   75K Feb  7  2018 libchromaprint.so.1.4.3
lrwxrwxrwx  1 root root    34 May  3 10:19 libcidn.so -> /lib/x86_64-linux-gnu/libcidn.so.1
lrwxrwxrwx  1 root root    19 Dec  4  2019 libcilkrts.so.5 -> libcilkrts.so.5.0.0
-rw-r--r--  1 root root  117K Dec  4  2019 libcilkrts.so.5.0.0
lrwxrwxrwx  1 root root    18 Jul 23  2017 libcolord.so.2 -> libcolord.so.2.0.5
-rw-r--r--  1 root root  126K Jul 23  2017 libcolord.so.2.0.5
lrwxrwxrwx  1 root root    25 Jul 23  2017 libcolordprivate.so.2 -> libcolordprivate.so.2.0.5
-rw-r--r--  1 root root  188K Jul 23  2017 libcolordprivate.so.2.0.5
lrwxrwxrwx  1 root root    21 Dec 16  2017 libcroco-0.6.so.3 -> libcroco-0.6.so.3.0.1
-rw-r--r--  1 root root  235K Dec 16  2017 libcroco-0.6.so.3.0.1
-rw-r--r--  1 root root   61K May  3 10:19 libcrypt.a
lrwxrwxrwx  1 root root    35 May  3 10:19 libcrypt.so -> /lib/x86_64-linux-gnu/libcrypt.so.1
-rw-r--r--  1 root root  2.8M May  3 17:51 libcrypto.so.1.1
lrwxrwxrwx  1 root root    19 Mar 13  2017 libcrystalhd.so.3 -> libcrystalhd.so.3.6
-rw-r--r--  1 root root  109K Mar 13  2017 libcrystalhd.so.3.6
lrwxrwxrwx  1 root root    15 Aug  9  2019 libcublas.so -> libcublas.so.10
lrwxrwxrwx  1 root root    23 Aug  9  2019 libcublas.so.10 -> libcublas.so.10.2.1.243
-rw-r--r--  1 root root   60M Aug  9  2019 libcublas.so.10.2.1.243
lrwxrwxrwx  1 root root    17 Aug  9  2019 libcublasLt.so -> libcublasLt.so.10
lrwxrwxrwx  1 root root    25 Aug  9  2019 libcublasLt.so.10 -> libcublasLt.so.10.2.1.243
-rw-r--r--  1 root root   29M Aug  9  2019 libcublasLt.so.10.2.1.243
-rw-r--r--  1 root root   32M Aug  9  2019 libcublasLt_static.a
-rw-r--r--  1 root root   72M Aug  9  2019 libcublas_static.a
lrwxrwxrwx  1 root root    29 May 28 20:19 libcudnn.so -> /etc/alternatives/libcudnn_so
lrwxrwxrwx  1 root root    17 Oct 27  2019 libcudnn.so.7 -> libcudnn.so.7.6.5
-rw-r--r--  1 root root  409M Oct 27  2019 libcudnn.so.7.6.5
lrwxrwxrwx  1 root root    32 May 28 20:19 libcudnn_static.a -> /etc/alternatives/libcudnn_stlib
-rw-r--r--  1 root root  386M Oct 27  2019 libcudnn_static_v7.a
-rw-r--r--  1 root root  558K May 27 15:03 libcups.so.2
lrwxrwxrwx  1 root root    19 Jun 22 16:00 libcurl-gnutls.so.3 -> libcurl-gnutls.so.4
lrwxrwxrwx  1 root root    23 Jun 22 16:00 libcurl-gnutls.so.4 -> libcurl-gnutls.so.4.5.0
-rw-r--r--  1 root root  515K Jun 22 16:00 libcurl-gnutls.so.4.5.0
lrwxrwxrwx  1 root root    16 Jun 22 16:00 libcurl.so.4 -> libcurl.so.4.5.0
-rw-r--r--  1 root root  519K Jun 22 16:00 libcurl.so.4.5.0
lrwxrwxrwx  1 root root    16 Feb 20  2018 libdap.so.25 -> libdap.so.25.0.1
-rw-r--r--  1 root root  1.6M Feb 20  2018 libdap.so.25.0.1
lrwxrwxrwx  1 root root    21 Feb 20  2018 libdapclient.so.6 -> libdapclient.so.6.1.7
-rw-r--r--  1 root root  259K Feb 20  2018 libdapclient.so.6.1.7
lrwxrwxrwx  1 root root    18 Feb 14  2018 libdatrie.so.1 -> libdatrie.so.1.3.3
-rw-r--r--  1 root root   26K Feb 14  2018 libdatrie.so.1.3.3
-rw-r--r--  1 root root  1.7M Jun  3  2019 libdb-5.3.so
-rw-r--r--  1 root root  365K Jan 15  2017 libdc1394.a
lrwxrwxrwx  1 root root    19 Jan 15  2017 libdc1394.so -> libdc1394.so.22.2.1
lrwxrwxrwx  1 root root    19 Jan 15  2017 libdc1394.so.22 -> libdc1394.so.22.2.1
-rw-r--r--  1 root root  216K Jan 15  2017 libdc1394.so.22.2.1
lrwxrwxrwx  1 root root    17 Mar 29  2018 libdconf.so.1 -> libdconf.so.1.0.0
-rw-r--r--  1 root root   56K Mar 29  2018 libdconf.so.1.0.0
lrwxrwxrwx  1 root root    25 Jun  8  2016 libdebconfclient.so.0 -> libdebconfclient.so.0.0.0
-rw-r--r--  1 root root   11K Jun  8  2016 libdebconfclient.so.0.0.0
-rw-r--r--  1 root root   12K May  3 10:19 libdl.a
lrwxrwxrwx  1 root root    32 May  3 10:19 libdl.so -> /lib/x86_64-linux-gnu/libdl.so.2
lrwxrwxrwx  1 root root    15 May 11  2020 libdrm.so.2 -> libdrm.so.2.4.0
-rw-r--r--  1 root root   67K May 11  2020 libdrm.so.2.4.0
lrwxrwxrwx  1 root root    22 May 11  2020 libdrm_amdgpu.so.1 -> libdrm_amdgpu.so.1.0.0
-rw-r--r--  1 root root   39K May 11  2020 libdrm_amdgpu.so.1.0.0
lrwxrwxrwx  1 root root    21 May 11  2020 libdrm_intel.so.1 -> libdrm_intel.so.1.0.0
-rw-r--r--  1 root root  140K May 11  2020 libdrm_intel.so.1.0.0
lrwxrwxrwx  1 root root    23 May 11  2020 libdrm_nouveau.so.2 -> libdrm_nouveau.so.2.0.0
-rw-r--r--  1 root root   31K May 11  2020 libdrm_nouveau.so.2.0.0
lrwxrwxrwx  1 root root    22 May 11  2020 libdrm_radeon.so.1 -> libdrm_radeon.so.1.0.1
-rw-r--r--  1 root root   47K May 11  2020 libdrm_radeon.so.1.0.1
lrwxrwxrwx  1 root root    17 Jun 19  2017 libedit.so.2 -> libedit.so.2.0.56
-rw-r--r--  1 root root  204K Jun 19  2017 libedit.so.2.0.56
-rw-r--r--  1 root root  103K Jun  7  2019 libelf-0.170.so
lrwxrwxrwx  1 root root    15 Jun  7  2019 libelf.so.1 -> libelf-0.170.so
lrwxrwxrwx  1 root root    17 Nov 13  2017 libepoxy.so.0 -> libepoxy.so.0.0.0
-rw-r--r--  1 root root  1.1M Nov 13  2017 libepoxy.so.0.0.0
lrwxrwxrwx  1 root root    19 Apr 27  2016 libepsilon.so.1 -> libepsilon.so.1.0.0
-rw-r--r--  1 root root   97K Apr 27  2016 libepsilon.so.1.0.0
-rw-r--r--  1 root root  400K Nov  6  2020 libexif.a
lrwxrwxrwx  1 root root    17 Nov  6  2020 libexif.so -> libexif.so.12.3.3
lrwxrwxrwx  1 root root    17 Nov  6  2020 libexif.so.12 -> libexif.so.12.3.3
-rw-r--r--  1 root root  274K Nov  6  2020 libexif.so.12.3.3
lrwxrwxrwx  1 root root    18 Mar  8  2022 libexpatw.so.1 -> libexpatw.so.1.6.7
-rw-r--r--  1 root root  203K Mar  8  2022 libexpatw.so.1.6.7
lrwxrwxrwx  1 root root    18 Jan  4  2018 libfabric.so.1 -> libfabric.so.1.9.3
-rw-r--r--  1 root root  676K Jan  4  2018 libfabric.so.1.9.3
lrwxrwxrwx  1 root root    15 Jan  7  2018 libffi.so.6 -> libffi.so.6.0.4
-rw-r--r--  1 root root   31K Jan  7  2018 libffi.so.6.0.4
lrwxrwxrwx  1 root root    23 Apr  5  2018 libfontconfig.so.1 -> libfontconfig.so.1.10.1
-rw-r--r--  1 root root  276K Apr  5  2018 libfontconfig.so.1.10.1
lrwxrwxrwx  1 root root    14 May 23  2018 libform.so.5 -> libform.so.5.9
-rw-r--r--  1 root root   60K May 23  2018 libform.so.5.9
lrwxrwxrwx  1 root root    15 May 23  2018 libformw.so.5 -> libformw.so.5.9
-rw-r--r--  1 root root   68K May 23  2018 libformw.so.5.9
lrwxrwxrwx  1 root root    18 Jul  6 11:25 libfreebl3.chk -> nss/libfreebl3.chk
lrwxrwxrwx  1 root root    17 Jul  6 11:25 libfreebl3.so -> nss/libfreebl3.so
lrwxrwxrwx  1 root root    22 Jul  6 11:25 libfreeblpriv3.chk -> nss/libfreeblpriv3.chk
lrwxrwxrwx  1 root root    21 Jul  6 11:25 libfreeblpriv3.so -> nss/libfreeblpriv3.so
lrwxrwxrwx  1 root root    21 Jul 19 16:39 libfreetype.so.6 -> libfreetype.so.6.15.0
-rw-r--r--  1 root root  719K Jul 19 16:39 libfreetype.so.6.15.0
lrwxrwxrwx  1 root root    18 Feb 22  2018 libfreexl.so.1 -> libfreexl.so.1.1.0
-rw-r--r--  1 root root   34K Feb 22  2018 libfreexl.so.1.1.0
lrwxrwxrwx  1 root root    16 Aug 19  2016 libfyba.so.0 -> libfyba.so.0.0.0
-rw-r--r--  1 root root  215K Aug 19  2016 libfyba.so.0.0.0
lrwxrwxrwx  1 root root    16 Aug 19  2016 libfygm.so.0 -> libfygm.so.0.0.0
-rw-r--r--  1 root root   27K Aug 19  2016 libfygm.so.0.0.0
lrwxrwxrwx  1 root root    16 Aug 19  2016 libfyut.so.0 -> libfyut.so.0.0.0
-rw-r--r--  1 root root   35K Aug 19  2016 libfyut.so.0.0.0
-rw-r--r--  1 root root  1.2K May  3 10:19 libg.a
lrwxrwxrwx  1 root root    14 Aug 30  2021 libgd.so.3 -> libgd.so.3.0.5
-rw-r--r--  1 root root  397K Aug 30  2021 libgd.so.3.0.5
lrwxrwxrwx  1 root root    16 Mar 14  2018 libgdbm.so.5 -> libgdbm.so.5.0.0
-rw-r--r--  1 root root   51K Mar 14  2018 libgdbm.so.5.0.0
lrwxrwxrwx  1 root root    23 Mar 14  2018 libgdbm_compat.so.4 -> libgdbm_compat.so.4.0.0
-rw-r--r--  1 root root   14K Mar 14  2018 libgdbm_compat.so.4.0.0
lrwxrwxrwx  1 root root    20 Feb  5  2018 libgdcmCommon.so -> libgdcmCommon.so.2.8
lrwxrwxrwx  1 root root    22 Feb  5  2018 libgdcmCommon.so.2.8 -> libgdcmCommon.so.2.8.4
-rw-r--r--  1 root root  155K Feb  5  2018 libgdcmCommon.so.2.8.4
lrwxrwxrwx  1 root root    18 Feb  5  2018 libgdcmDICT.so -> libgdcmDICT.so.2.8
lrwxrwxrwx  1 root root    20 Feb  5  2018 libgdcmDICT.so.2.8 -> libgdcmDICT.so.2.8.4
-rw-r--r--  1 root root  2.4M Feb  5  2018 libgdcmDICT.so.2.8.4
lrwxrwxrwx  1 root root    18 Feb  5  2018 libgdcmDSED.so -> libgdcmDSED.so.2.8
lrwxrwxrwx  1 root root    20 Feb  5  2018 libgdcmDSED.so.2.8 -> libgdcmDSED.so.2.8.4
-rw-r--r--  1 root root  1.1M Feb  5  2018 libgdcmDSED.so.2.8.4
lrwxrwxrwx  1 root root    17 Feb  5  2018 libgdcmIOD.so -> libgdcmIOD.so.2.8
lrwxrwxrwx  1 root root    19 Feb  5  2018 libgdcmIOD.so.2.8 -> libgdcmIOD.so.2.8.4
-rw-r--r--  1 root root   87K Feb  5  2018 libgdcmIOD.so.2.8.4
lrwxrwxrwx  1 root root    18 Feb  5  2018 libgdcmMEXD.so -> libgdcmMEXD.so.2.8
lrwxrwxrwx  1 root root    20 Feb  5  2018 libgdcmMEXD.so.2.8 -> libgdcmMEXD.so.2.8.4
-rw-r--r--  1 root root  1.1M Feb  5  2018 libgdcmMEXD.so.2.8.4
lrwxrwxrwx  1 root root    18 Feb  5  2018 libgdcmMSFF.so -> libgdcmMSFF.so.2.8
lrwxrwxrwx  1 root root    20 Feb  5  2018 libgdcmMSFF.so.2.8 -> libgdcmMSFF.so.2.8.4
-rw-r--r--  1 root root  2.6M Feb  5  2018 libgdcmMSFF.so.2.8.4
lrwxrwxrwx  1 root root    20 Feb  5  2018 libgdcmjpeg12.so -> libgdcmjpeg12.so.2.8
lrwxrwxrwx  1 root root    22 Feb  5  2018 libgdcmjpeg12.so.2.8 -> libgdcmjpeg12.so.2.8.4
-rw-r--r--  1 root root  158K Feb  5  2018 libgdcmjpeg12.so.2.8.4
lrwxrwxrwx  1 root root    20 Feb  5  2018 libgdcmjpeg16.so -> libgdcmjpeg16.so.2.8
lrwxrwxrwx  1 root root    22 Feb  5  2018 libgdcmjpeg16.so.2.8 -> libgdcmjpeg16.so.2.8.4
-rw-r--r--  1 root root  158K Feb  5  2018 libgdcmjpeg16.so.2.8.4
lrwxrwxrwx  1 root root    19 Feb  5  2018 libgdcmjpeg8.so -> libgdcmjpeg8.so.2.8
lrwxrwxrwx  1 root root    21 Feb  5  2018 libgdcmjpeg8.so.2.8 -> libgdcmjpeg8.so.2.8.4
-rw-r--r--  1 root root  158K Feb  5  2018 libgdcmjpeg8.so.2.8.4
lrwxrwxrwx  1 root root    21 Jun  5  2019 libgdk-3.so.0 -> libgdk-3.so.0.2200.30
-rw-r--r--  1 root root  981K Jun  5  2019 libgdk-3.so.0.2200.30
lrwxrwxrwx  1 root root    27 Mar 24  2018 libgdk-x11-2.0.so.0 -> libgdk-x11-2.0.so.0.2400.32
-rw-r--r--  1 root root  724K Mar 24  2018 libgdk-x11-2.0.so.0.2400.32
lrwxrwxrwx  1 root root    29 Mar 16  2018 libgdk_pixbuf-2.0.so.0 -> libgdk_pixbuf-2.0.so.0.3611.0
-rw-r--r--  1 root root  144K Mar 16  2018 libgdk_pixbuf-2.0.so.0.3611.0
lrwxrwxrwx  1 root root    34 Mar 16  2018 libgdk_pixbuf_xlib-2.0.so.0 -> libgdk_pixbuf_xlib-2.0.so.0.3611.0
-rw-r--r--  1 root root   67K Mar 16  2018 libgdk_pixbuf_xlib-2.0.so.0.3611.0
-rw-r--r--  1 root root  1.6M Mar  1  2018 libgeos-3.6.2.so
lrwxrwxrwx  1 root root    19 Mar  1  2018 libgeos_c.so.1 -> libgeos_c.so.1.10.2
-rw-r--r--  1 root root  187K Mar  1  2018 libgeos_c.so.1.10.2
lrwxrwxrwx  1 root root    19 Nov 17  2016 libgeotiff.so.2 -> libgeotiff.so.2.1.2
-rw-r--r--  1 root root  215K Nov 17  2016 libgeotiff.so.2.1.2
lrwxrwxrwx  1 root root    20 Dec  4  2019 libgfortran.so.4 -> libgfortran.so.4.0.0
-rw-r--r--  1 root root  1.9M Dec  4  2019 libgfortran.so.4.0.0
lrwxrwxrwx  1 root root    15 Aug 19  2019 libgif.so.7 -> libgif.so.7.0.0
-rw-r--r--  1 root root   34K Aug 19  2019 libgif.so.7.0.0
-rw-r--r--  1 root root  4.0M Nov 29  2021 libgio-2.0.a
lrwxrwxrwx  1 root root    22 Nov 29  2021 libgio-2.0.so -> libgio-2.0.so.0.5600.4
lrwxrwxrwx  1 root root    22 Nov 29  2021 libgio-2.0.so.0 -> libgio-2.0.so.0.5600.4
-rw-r--r--  1 root root  1.7M Nov 29  2021 libgio-2.0.so.0.5600.4
lrwxrwxrwx  1 root root    28 Apr 15  2018 libgirepository-1.0.so.1 -> libgirepository-1.0.so.1.0.0
-rw-r--r--  1 root root  208K Apr 15  2018 libgirepository-1.0.so.1.0.0
lrwxrwxrwx  1 root root    17 Sep 27  2017 libgl2ps.so.1.4 -> libgl2ps.so.1.4.0
-rw-r--r--  1 root root   79K Sep 27  2017 libgl2ps.so.1.4.0
lrwxrwxrwx  1 root root    17 Jun 12  2020 libglapi.so.0 -> libglapi.so.0.0.0
-rw-r--r--  1 root root  215K Jun 12  2020 libglapi.so.0.0.0
-rw-r--r--  1 root root  2.0M Nov 29  2021 libglib-2.0.a
lrwxrwxrwx  1 root root    23 Nov 29  2021 libglib-2.0.so -> libglib-2.0.so.0.5600.4
lrwxrwxrwx  1 root root    23 Nov 29  2021 libglib-2.0.so.0 -> libglib-2.0.so.0.5600.4
-rw-r--r--  1 root root  1.1M Nov 29  2021 libglib-2.0.so.0.5600.4
lrwxrwxrwx  1 root root    15 Dec  7  2017 libgme.so.0 -> libgme.so.0.6.2
-rw-r--r--  1 root root  303K Dec  7  2017 libgme.so.0.6.2
-rw-r--r--  1 root root   16K Nov 29  2021 libgmodule-2.0.a
lrwxrwxrwx  1 root root    26 Nov 29  2021 libgmodule-2.0.so -> libgmodule-2.0.so.0.5600.4
lrwxrwxrwx  1 root root    26 Nov 29  2021 libgmodule-2.0.so.0 -> libgmodule-2.0.so.0.5600.4
-rw-r--r--  1 root root   14K Nov 29  2021 libgmodule-2.0.so.0.5600.4
lrwxrwxrwx  1 root root    16 Jan 24  2018 libgmp.so.10 -> libgmp.so.10.3.2
-rw-r--r--  1 root root  515K Jan 24  2018 libgmp.so.10.3.2
lrwxrwxrwx  1 root root    21 Aug 25  2021 libgnutls.so.30 -> libgnutls.so.30.14.10
-rw-r--r--  1 root root  1.4M Aug 25  2021 libgnutls.so.30.14.10
-rw-r--r--  1 root root  667K Nov 29  2021 libgobject-2.0.a
lrwxrwxrwx  1 root root    26 Nov 29  2021 libgobject-2.0.so -> libgobject-2.0.so.0.5600.4
lrwxrwxrwx  1 root root    26 Nov 29  2021 libgobject-2.0.so.0 -> libgobject-2.0.so.0.5600.4
-rw-r--r--  1 root root  335K Nov 29  2021 libgobject-2.0.so.0.5600.4
lrwxrwxrwx  1 root root    16 Mar 10  2020 libgomp.so.1 -> libgomp.so.1.0.0
-rw-r--r--  1 root root  188K Mar 10  2020 libgomp.so.1.0.0
drwxr-xr-x  3 root root  4.0K Aug 16 08:53 libgphoto2
drwxr-xr-x  3 root root  4.0K Aug 16 08:53 libgphoto2-dev
-rw-r--r--  1 root root  262K Dec 23  2017 libgphoto2.a
lrwxrwxrwx  1 root root    19 Dec 23  2017 libgphoto2.so -> libgphoto2.so.6.0.0
lrwxrwxrwx  1 root root    19 Dec 23  2017 libgphoto2.so.6 -> libgphoto2.so.6.0.0
-rw-r--r--  1 root root  143K Dec 23  2017 libgphoto2.so.6.0.0
drwxr-xr-x  3 root root  4.0K Aug 16 08:53 libgphoto2_port
-rw-r--r--  1 root root   79K Dec 23  2017 libgphoto2_port.a
lrwxrwxrwx  1 root root    25 Dec 23  2017 libgphoto2_port.so -> libgphoto2_port.so.12.0.0
lrwxrwxrwx  1 root root    25 Dec 23  2017 libgphoto2_port.so.12 -> libgphoto2_port.so.12.0.0
-rw-r--r--  1 root root   43K Dec 23  2017 libgphoto2_port.so.12.0.0
lrwxrwxrwx  1 root root    17 Mar 11  2018 libgraphite2.so -> libgraphite2.so.3
lrwxrwxrwx  1 root root    17 Mar 11  2018 libgraphite2.so.2.0.0 -> libgraphite2.so.3
lrwxrwxrwx  1 root root    21 Mar 11  2018 libgraphite2.so.3 -> libgraphite2.so.3.0.1
-rw-r--r--  1 root root  179K Mar 11  2018 libgraphite2.so.3.0.1
lrwxrwxrwx  1 root root    16 Apr  3  2018 libgsm.so.1 -> libgsm.so.1.0.12
-rw-r--r--  1 root root   51K Apr  3  2018 libgsm.so.1.0.12
lrwxrwxrwx  1 root root    18 Dec 15  2017 libgssapi.so.3 -> libgssapi.so.3.0.0
-rw-r--r--  1 root root  260K Dec 15  2017 libgssapi.so.3.0.0
lrwxrwxrwx  1 root root    21 Nov 11  2020 libgssapi_krb5.so.2 -> libgssapi_krb5.so.2.2
-rw-r--r--  1 root root  299K Nov 11  2020 libgssapi_krb5.so.2.2
-rw-r--r--  1 root root  2.7K Nov 29  2021 libgthread-2.0.a
lrwxrwxrwx  1 root root    26 Nov 29  2021 libgthread-2.0.so -> libgthread-2.0.so.0.5600.4
lrwxrwxrwx  1 root root    26 Nov 29  2021 libgthread-2.0.so.0 -> libgthread-2.0.so.0.5600.4
-rw-r--r--  1 root root  5.9K Nov 29  2021 libgthread-2.0.so.0.5600.4
drwxr-xr-x  2 root root  4.0K Aug 16 08:53 libgtk-3-0
lrwxrwxrwx  1 root root    21 Jun  5  2019 libgtk-3.so.0 -> libgtk-3.so.0.2200.30
-rw-r--r--  1 root root  7.1M Jun  5  2019 libgtk-3.so.0.2200.30
lrwxrwxrwx  1 root root    27 Mar 24  2018 libgtk-x11-2.0.so.0 -> libgtk-x11-2.0.so.0.2400.32
-rw-r--r--  1 root root  4.3M Mar 24  2018 libgtk-x11-2.0.so.0.2400.32
drwxr-xr-x  2 root root  4.0K Aug 16 08:53 libgtk2.0-0
lrwxrwxrwx  1 root root    19 Feb  3  2017 libgts-0.7.so.5 -> libgts-0.7.so.5.0.1
-rw-r--r--  1 root root  371K Feb  3  2017 libgts-0.7.so.5.0.1
lrwxrwxrwx  1 root root    15 Mar 24  2018 libgvc.so -> libgvc.so.6.0.0
lrwxrwxrwx  1 root root    15 Mar 24  2018 libgvc.so.6 -> libgvc.so.6.0.0
-rw-r--r--  1 root root  613K Mar 24  2018 libgvc.so.6.0.0
lrwxrwxrwx  1 root root    16 Mar 24  2018 libgvpr.so -> libgvpr.so.2.0.0
lrwxrwxrwx  1 root root    16 Mar 24  2018 libgvpr.so.2 -> libgvpr.so.2.0.0
-rw-r--r--  1 root root  480K Mar 24  2018 libgvpr.so.2.0.0
-rw-r--r--  1 root root   67K Apr 13  2018 libharfbuzz-gobject.a
lrwxrwxrwx  1 root root    32 Apr 13  2018 libharfbuzz-gobject.so -> libharfbuzz-gobject.so.0.10702.0
lrwxrwxrwx  1 root root    32 Apr 13  2018 libharfbuzz-gobject.so.0 -> libharfbuzz-gobject.so.0.10702.0
-rw-r--r--  1 root root   55K Apr 13  2018 libharfbuzz-gobject.so.0.10702.0
-rw-r--r--  1 root root  7.0K Apr 13  2018 libharfbuzz-icu.a
lrwxrwxrwx  1 root root    28 Apr 13  2018 libharfbuzz-icu.so -> libharfbuzz-icu.so.0.10702.0
lrwxrwxrwx  1 root root    28 Apr 13  2018 libharfbuzz-icu.so.0 -> libharfbuzz-icu.so.0.10702.0
-rw-r--r--  1 root root   11K Apr 13  2018 libharfbuzz-icu.so.0.10702.0
-rw-r--r--  1 root root  965K Apr 13  2018 libharfbuzz.a
lrwxrwxrwx  1 root root    24 Apr 13  2018 libharfbuzz.so -> libharfbuzz.so.0.10702.0
lrwxrwxrwx  1 root root    24 Apr 13  2018 libharfbuzz.so.0 -> libharfbuzz.so.0.10702.0
-rw-r--r--  1 root root  631K Apr 13  2018 libharfbuzz.so.0.10702.0
lrwxrwxrwx  1 root root    19 Dec 15  2017 libhcrypto.so.4 -> libhcrypto.so.4.1.0
-rw-r--r--  1 root root  213K Dec 15  2017 libhcrypto.so.4.1.0
lrwxrwxrwx  1 root root    26 Aug 13  2017 libhdf5_openmpi.so.100 -> libhdf5_openmpi.so.100.0.1
-rw-r--r--  1 root root  3.5M Aug 13  2017 libhdf5_openmpi.so.100.0.1
lrwxrwxrwx  1 root root    34 Aug 13  2017 libhdf5_openmpi_fortran.so.100 -> libhdf5_openmpi_fortran.so.100.0.1
-rw-r--r--  1 root root  249K Aug 13  2017 libhdf5_openmpi_fortran.so.100.0.1
lrwxrwxrwx  1 root root    29 Aug 13  2017 libhdf5_openmpi_hl.so.100 -> libhdf5_openmpi_hl.so.100.0.0
-rw-r--r--  1 root root  136K Aug 13  2017 libhdf5_openmpi_hl.so.100.0.0
lrwxrwxrwx  1 root root    36 Aug 13  2017 libhdf5_openmpihl_fortran.so.100 -> libhdf5_openmpihl_fortran.so.100.0.0
-rw-r--r--  1 root root  119K Aug 13  2017 libhdf5_openmpihl_fortran.so.100.0.0
lrwxrwxrwx  1 root root    25 Aug 13  2017 libhdf5_serial.so.100 -> libhdf5_serial.so.100.0.1
-rw-r--r--  1 root root  3.4M Aug 13  2017 libhdf5_serial.so.100.0.1
lrwxrwxrwx  1 root root    33 Aug 13  2017 libhdf5_serial_fortran.so.100 -> libhdf5_serial_fortran.so.100.0.1
-rw-r--r--  1 root root  245K Aug 13  2017 libhdf5_serial_fortran.so.100.0.1
lrwxrwxrwx  1 root root    28 Aug 13  2017 libhdf5_serial_hl.so.100 -> libhdf5_serial_hl.so.100.0.0
-rw-r--r--  1 root root  136K Aug 13  2017 libhdf5_serial_hl.so.100.0.0
lrwxrwxrwx  1 root root    35 Aug 13  2017 libhdf5_serialhl_fortran.so.100 -> libhdf5_serialhl_fortran.so.100.0.0
-rw-r--r--  1 root root  119K Aug 13  2017 libhdf5_serialhl_fortran.so.100.0.0
lrwxrwxrwx  1 root root    20 Dec 15  2017 libheimbase.so.1 -> libheimbase.so.1.0.0
-rw-r--r--  1 root root   59K Dec 15  2017 libheimbase.so.1.0.0
lrwxrwxrwx  1 root root    20 Dec 15  2017 libheimntlm.so.0 -> libheimntlm.so.0.1.0
-rw-r--r--  1 root root   35K Dec 15  2017 libheimntlm.so.0.1.0
lrwxrwxrwx  1 root root    17 Jun 14  2021 libhogweed.so.4 -> libhogweed.so.4.5
-rw-r--r--  1 root root  215K Jun 14  2021 libhogweed.so.4.5
lrwxrwxrwx  1 root root    13 Jan 19  2018 libhwloc.so.0 -> libhwloc.so.5
lrwxrwxrwx  1 root root    13 Jan 19  2018 libhwloc.so.1 -> libhwloc.so.5
lrwxrwxrwx  1 root root    13 Jan 19  2018 libhwloc.so.2 -> libhwloc.so.5
lrwxrwxrwx  1 root root    13 Jan 19  2018 libhwloc.so.3 -> libhwloc.so.5
lrwxrwxrwx  1 root root    13 Jan 19  2018 libhwloc.so.4 -> libhwloc.so.5
lrwxrwxrwx  1 root root    17 Jan 19  2018 libhwloc.so.5 -> libhwloc.so.5.7.6
-rw-r--r--  1 root root  244K Jan 19  2018 libhwloc.so.5.7.6
lrwxrwxrwx  1 root root    17 Dec 15  2017 libhx509.so.5 -> libhx509.so.5.0.0
-rw-r--r--  1 root root  294K Dec 15  2017 libhx509.so.5.0.0
lrwxrwxrwx  1 root root    22 Aug  5  2019 libibverbs.so.1 -> libibverbs.so.1.1.17.1
-rw-r--r--  1 root root   87K Aug  5  2019 libibverbs.so.1.1.17.1
-rw-r--r--  1 root root   55K Dec  9  2017 libicu-le-hb.a
lrwxrwxrwx  1 root root    21 Dec  9  2017 libicu-le-hb.so -> libicu-le-hb.so.0.0.0
lrwxrwxrwx  1 root root    21 Dec  9  2017 libicu-le-hb.so.0 -> libicu-le-hb.so.0.0.0
-rw-r--r--  1 root root   35K Dec  9  2017 libicu-le-hb.so.0.0.0
-rw-r--r--  1 root root   26M Oct 19  2021 libicudata.a
lrwxrwxrwx  1 root root    18 Oct 19  2021 libicudata.so -> libicudata.so.60.2
lrwxrwxrwx  1 root root    18 Oct 19  2021 libicudata.so.60 -> libicudata.so.60.2
-rw-r--r--  1 root root   26M Oct 19  2021 libicudata.so.60.2
-rw-r--r--  1 root root  5.9M Oct 19  2021 libicui18n.a
lrwxrwxrwx  1 root root    18 Oct 19  2021 libicui18n.so -> libicui18n.so.60.2
lrwxrwxrwx  1 root root    18 Oct 19  2021 libicui18n.so.60 -> libicui18n.so.60.2
-rw-r--r--  1 root root  2.7M Oct 19  2021 libicui18n.so.60.2
-rw-r--r--  1 root root   85K Oct 19  2021 libicuio.a
lrwxrwxrwx  1 root root    16 Oct 19  2021 libicuio.so -> libicuio.so.60.2
lrwxrwxrwx  1 root root    16 Oct 19  2021 libicuio.so.60 -> libicuio.so.60.2
-rw-r--r--  1 root root   55K Oct 19  2021 libicuio.so.60.2
-rw-r--r--  1 root root   75K Oct 19  2021 libiculx.a
lrwxrwxrwx  1 root root    16 Oct 19  2021 libiculx.so -> libiculx.so.60.2
lrwxrwxrwx  1 root root    16 Oct 19  2021 libiculx.so.60 -> libiculx.so.60.2
-rw-r--r--  1 root root   47K Oct 19  2021 libiculx.so.60.2
-rw-r--r--  1 root root  109K Oct 19  2021 libicutest.a
lrwxrwxrwx  1 root root    18 Oct 19  2021 libicutest.so -> libicutest.so.60.2
lrwxrwxrwx  1 root root    18 Oct 19  2021 libicutest.so.60 -> libicutest.so.60.2
-rw-r--r--  1 root root   64K Oct 19  2021 libicutest.so.60.2
-rw-r--r--  1 root root  359K Oct 19  2021 libicutu.a
lrwxrwxrwx  1 root root    16 Oct 19  2021 libicutu.so -> libicutu.so.60.2
lrwxrwxrwx  1 root root    16 Oct 19  2021 libicutu.so.60 -> libicutu.so.60.2
-rw-r--r--  1 root root  199K Oct 19  2021 libicutu.so.60.2
-rw-r--r--  1 root root  3.1M Oct 19  2021 libicuuc.a
lrwxrwxrwx  1 root root    16 Oct 19  2021 libicuuc.so -> libicuuc.so.60.2
lrwxrwxrwx  1 root root    16 Oct 19  2021 libicuuc.so.60 -> libicuuc.so.60.2
-rw-r--r--  1 root root  1.8M Oct 19  2021 libicuuc.so.60.2
lrwxrwxrwx  1 root root    16 Oct 24  2019 libidn2.so.0 -> libidn2.so.0.3.3
-rw-r--r--  1 root root  114K Oct 24  2019 libidn2.so.0.3.3
lrwxrwxrwx  1 root root    20 Jan 14  2018 libinfinipath.so.4 -> libinfinipath.so.4.0
-rw-r--r--  1 root root   59K Jan 14  2018 libinfinipath.so.4.0
lrwxrwxrwx  1 root root    16 Mar  9  2018 libisl.so.19 -> libisl.so.19.0.0
-rw-r--r--  1 root root  1.6M Mar  9  2018 libisl.so.19.0.0
lrwxrwxrwx  1 root root    15 Mar 10  2020 libitm.so.1 -> libitm.so.1.0.0
-rw-r--r--  1 root root  111K Mar 10  2020 libitm.so.1.0.0
-rw-r--r--  1 root root   51K Apr  3  2018 libjbig.a
lrwxrwxrwx  1 root root    12 Apr  3  2018 libjbig.so -> libjbig.so.0
-rw-r--r--  1 root root   57K Apr  3  2018 libjbig.so.0
-rw-r--r--  1 root root  579K Jun  4  2020 libjpeg.a
lrwxrwxrwx  1 root root    16 Jun  4  2020 libjpeg.so -> libjpeg.so.8.1.2
lrwxrwxrwx  1 root root    16 Jun  4  2020 libjpeg.so.8 -> libjpeg.so.8.1.2
-rw-r--r--  1 root root  415K Jun  4  2020 libjpeg.so.8.1.2
lrwxrwxrwx  1 root root    27 Apr  6  2020 libjson-glib-1.0.so.0 -> libjson-glib-1.0.so.0.400.2
-rw-r--r--  1 root root  154K Apr  6  2020 libjson-glib-1.0.so.0.400.2
lrwxrwxrwx  1 root root    19 Aug 23  2016 libjsoncpp.so.1 -> libjsoncpp.so.1.7.4
-rw-r--r--  1 root root  199K Aug 23  2016 libjsoncpp.so.1.7.4
lrwxrwxrwx  1 root root    18 Nov 11  2020 libk5crypto.so.3 -> libk5crypto.so.3.1
-rw-r--r--  1 root root  195K Nov 11  2020 libk5crypto.so.3.1
lrwxrwxrwx  1 root root    19 Sep 25  2017 libkmlbase.so.1 -> libkmlbase.so.1.3.0
-rw-r--r--  1 root root  107K Sep 25  2017 libkmlbase.so.1.3.0
lrwxrwxrwx  1 root root    18 Sep 25  2017 libkmldom.so.1 -> libkmldom.so.1.3.0
-rw-r--r--  1 root root  731K Sep 25  2017 libkmldom.so.1.3.0
lrwxrwxrwx  1 root root    21 Sep 25  2017 libkmlengine.so.1 -> libkmlengine.so.1.3.0
-rw-r--r--  1 root root  223K Sep 25  2017 libkmlengine.so.1.3.0
lrwxrwxrwx  1 root root    17 Dec 15  2017 libkrb5.so.26 -> libkrb5.so.26.0.0
-rw-r--r--  1 root root  561K Dec 15  2017 libkrb5.so.26.0.0
lrwxrwxrwx  1 root root    14 Nov 11  2020 libkrb5.so.3 -> libkrb5.so.3.3
-rw-r--r--  1 root root  857K Nov 11  2020 libkrb5.so.3.3
lrwxrwxrwx  1 root root    21 Nov 11  2020 libkrb5support.so.0 -> libkrb5support.so.0.1
-rw-r--r--  1 root root   43K Nov 11  2020 libkrb5support.so.0.1
lrwxrwxrwx  1 root root    17 Sep  3  2016 libksba.so.8 -> libksba.so.8.11.6
-rw-r--r--  1 root root  223K Sep  3  2016 libksba.so.8.11.6
lrwxrwxrwx  1 root root    21 Mar 24  2018 liblab_gamut.so -> liblab_gamut.so.1.0.0
lrwxrwxrwx  1 root root    21 Mar 24  2018 liblab_gamut.so.1 -> liblab_gamut.so.1.0.0
-rw-r--r--  1 root root  2.4M Mar 24  2018 liblab_gamut.so.1.0.0
lrwxrwxrwx  1 root root    46 Aug 16 08:54 liblapack.a -> /etc/alternatives/liblapack.a-x86_64-linux-gnu
lrwxrwxrwx  1 root root    47 Aug 16 08:54 liblapack.so -> /etc/alternatives/liblapack.so-x86_64-linux-gnu
lrwxrwxrwx  1 root root    49 Aug 16 08:54 liblapack.so.3 -> /etc/alternatives/liblapack.so.3-x86_64-linux-gnu
lrwxrwxrwx  1 root root    21 May 12 13:52 liblber-2.4.so.2 -> liblber-2.4.so.2.10.8
-rw-r--r--  1 root root   55K May 12 13:52 liblber-2.4.so.2.10.8
lrwxrwxrwx  1 root root    17 Sep 20  2018 liblcms2.so.2 -> liblcms2.so.2.0.8
-rw-r--r--  1 root root  349K Sep 20  2018 liblcms2.so.2.0.8
lrwxrwxrwx  1 root root    18 May 12 13:52 libldap-2.4.so.2 -> libldap_r-2.4.so.2
lrwxrwxrwx  1 root root    23 May 12 13:52 libldap_r-2.4.so.2 -> libldap_r-2.4.so.2.10.8
-rw-r--r--  1 root root  320K May 12 13:52 libldap_r-2.4.so.2.10.8
lrwxrwxrwx  1 root root    16 Mar  2  2018 liblept.so.5 -> liblept.so.5.0.2
-rw-r--r--  1 root root  2.4M Mar  2  2018 liblept.so.5.0.2
lrwxrwxrwx  1 root root    16 Mar 10  2020 liblsan.so.0 -> liblsan.so.0.0.0
-rw-r--r--  1 root root  338K Mar 10  2020 liblsan.so.0.0.0
lrwxrwxrwx  1 root root    16 Aug 20  2016 libltdl.so.7 -> libltdl.so.7.3.1
-rw-r--r--  1 root root   39K Aug 20  2016 libltdl.so.7.3.1
lrwxrwxrwx  1 root root    15 May 20  2021 liblz4.so.1 -> liblz4.so.1.7.1
-rw-r--r--  1 root root  111K May 20  2021 liblz4.so.1.7.1
-rw-r--r--  1 root root  261K Apr  8 12:56 liblzma.a
lrwxrwxrwx  1 root root    38 Apr  8 12:56 liblzma.so -> /lib/x86_64-linux-gnu/liblzma.so.5.2.2
-rw-r--r--  1 root root  3.5M May  3 10:19 libm-2.27.a
-rw-r--r--  1 root root   132 May  3 10:19 libm.a
-rw-r--r--  1 root root   186 May  3 10:19 libm.so
lrwxrwxrwx  1 root root    34 Feb  5  2018 libmca_common_libfabric.so.20 -> libmca_common_libfabric.so.20.10.0
lrwxrwxrwx  1 root root    46 Feb  5  2018 libmca_common_libfabric.so.20.10.0 -> openmpi/lib/libmca_common_libfabric.so.20.10.0
lrwxrwxrwx  1 root root    27 Feb  5  2018 libmca_common_sm.so.20 -> libmca_common_sm.so.20.10.1
lrwxrwxrwx  1 root root    39 Feb  5  2018 libmca_common_sm.so.20.10.1 -> openmpi/lib/libmca_common_sm.so.20.10.1
lrwxrwxrwx  1 root root    30 Feb  5  2018 libmca_common_verbs.so.20 -> libmca_common_verbs.so.20.10.0
lrwxrwxrwx  1 root root    42 Feb  5  2018 libmca_common_verbs.so.20.10.0 -> openmpi/lib/libmca_common_verbs.so.20.10.0
-rw-r--r--  1 root root  1.5K May  3 10:19 libmcheck.a
lrwxrwxrwx  1 root root    14 May 23  2018 libmenu.so.5 -> libmenu.so.5.9
-rw-r--r--  1 root root   35K May 23  2018 libmenu.so.5.9
lrwxrwxrwx  1 root root    15 May 23  2018 libmenuw.so.5 -> libmenuw.so.5.9
-rw-r--r--  1 root root   35K May 23  2018 libmenuw.so.5.9
lrwxrwxrwx  1 root root    19 Apr  3  2018 libminizip.so.1 -> libminizip.so.1.0.0
-rw-r--r--  1 root root   43K Apr  3  2018 libminizip.so.1.0.0
lrwxrwxrwx  1 root root    19 Oct 21  2017 libmp3lame.so.0 -> libmp3lame.so.0.0.0
-rw-r--r--  1 root root  294K Oct 21  2017 libmp3lame.so.0.0.0
lrwxrwxrwx  1 root root    15 Jan 23  2018 libmpc.so.3 -> libmpc.so.3.1.0
-rw-r--r--  1 root root   96K Jan 23  2018 libmpc.so.3.1.0
lrwxrwxrwx  1 root root    17 Apr 23  2018 libmpdec.so.2 -> libmpdec.so.2.4.2
-rw-r--r--  1 root root  223K Apr 23  2018 libmpdec.so.2.4.2
lrwxrwxrwx  1 root root    16 Feb  7  2018 libmpfr.so.6 -> libmpfr.so.6.0.1
-rw-r--r--  1 root root  512K Feb  7  2018 libmpfr.so.6.0.1
lrwxrwxrwx  1 root root    19 Mar 10  2018 libmpg123.so.0 -> libmpg123.so.0.44.8
-rw-r--r--  1 root root  316K Mar 10  2018 libmpg123.so.0.44.8
lrwxrwxrwx  1 root root    17 Feb  5  2018 libmpi.so.20 -> libmpi.so.20.10.1
lrwxrwxrwx  1 root root    29 Feb  5  2018 libmpi.so.20.10.1 -> openmpi/lib/libmpi.so.20.10.1
lrwxrwxrwx  1 root root    21 Feb  5  2018 libmpi_cxx.so.20 -> libmpi_cxx.so.20.10.0
lrwxrwxrwx  1 root root    33 Feb  5  2018 libmpi_cxx.so.20.10.0 -> openmpi/lib/libmpi_cxx.so.20.10.0
lrwxrwxrwx  1 root root    22 Feb  5  2018 libmpi_java.so.20 -> libmpi_java.so.20.10.0
lrwxrwxrwx  1 root root    34 Feb  5  2018 libmpi_java.so.20.10.0 -> openmpi/lib/libmpi_java.so.20.10.0
lrwxrwxrwx  1 root root    23 Feb  5  2018 libmpi_mpifh.so.20 -> libmpi_mpifh.so.20.11.0
lrwxrwxrwx  1 root root    35 Feb  5  2018 libmpi_mpifh.so.20.11.0 -> openmpi/lib/libmpi_mpifh.so.20.11.0
lrwxrwxrwx  1 root root    35 Feb  5  2018 libmpi_usempi_ignore_tkr.so.20 -> libmpi_usempi_ignore_tkr.so.20.10.0
lrwxrwxrwx  1 root root    47 Feb  5  2018 libmpi_usempi_ignore_tkr.so.20.10.0 -> openmpi/lib/libmpi_usempi_ignore_tkr.so.20.10.0
lrwxrwxrwx  1 root root    27 Feb  5  2018 libmpi_usempif08.so.20 -> libmpi_usempif08.so.20.10.0
lrwxrwxrwx  1 root root    39 Feb  5  2018 libmpi_usempif08.so.20.10.0 -> openmpi/lib/libmpi_usempif08.so.20.10.0
lrwxrwxrwx  1 root root    15 Mar 10  2020 libmpx.so.2 -> libmpx.so.2.0.1
-rw-r--r--  1 root root   19K Mar 10  2020 libmpx.so.2.0.1
lrwxrwxrwx  1 root root    23 Mar 10  2020 libmpxwrappers.so.2 -> libmpxwrappers.so.2.0.1
-rw-r--r--  1 root root   15K Mar 10  2020 libmpxwrappers.so.2.0.1
-rw-r--r--  1 root root  337K May  3 10:19 libmvec.a
lrwxrwxrwx  1 root root    34 May  3 10:19 libmvec.so -> /lib/x86_64-linux-gnu/libmvec.so.1
-rw-r--r--  1 root root  5.9K May  3 10:19 libmvec_nonshared.a
lrwxrwxrwx  1 root root    25 Jul 26 11:28 libmysqlclient.so.20 -> libmysqlclient.so.20.3.26
-rw-r--r--  1 root root  3.8M Jul 26 11:28 libmysqlclient.so.20.3.26
lrwxrwxrwx  1 root root    12 Oct 26  2020 libnccl.so -> libnccl.so.2
lrwxrwxrwx  1 root root    16 Oct 26  2020 libnccl.so.2 -> libnccl.so.2.8.3
-rw-r--r--  1 root root   78M Oct 26  2020 libnccl.so.2.8.3
-rw-r--r--  1 root root   95M Oct 26  2020 libnccl_static.a
-rw-r--r--  1 root root  1.1M Feb  9  2018 libnetcdf.so.13
lrwxrwxrwx  1 root root    22 Nov  9  2017 libnetcdf_c++.so.4 -> libnetcdf_c++.so.4.2.0
-rw-r--r--  1 root root  119K Nov  9  2017 libnetcdf_c++.so.4.2.0
lrwxrwxrwx  1 root root    16 Jun 14  2021 libnettle.so.6 -> libnettle.so.6.5
-rw-r--r--  1 root root  215K Jun 14  2021 libnettle.so.6.5
lrwxrwxrwx  1 root root    21 Apr 10  2018 libnghttp2.so.14 -> libnghttp2.so.14.15.2
-rw-r--r--  1 root root  150K Apr 10  2018 libnghttp2.so.14.15.2
lrwxrwxrwx  1 root root    25 Jun  2  2017 libnl-route-3.so.200 -> libnl-route-3.so.200.24.0
-rw-r--r--  1 root root  458K Jun  2  2017 libnl-route-3.so.200.24.0
lrwxrwxrwx  1 root root    16 Nov 12  2017 libnpth.so.0 -> libnpth.so.0.1.1
-rw-r--r--  1 root root   15K Nov 12  2017 libnpth.so.0.1.1
-rw-r--r--  1 root root  201K May  3 10:19 libnsl.a
lrwxrwxrwx  1 root root    33 May  3 10:19 libnsl.so -> /lib/x86_64-linux-gnu/libnsl.so.1
-rw-r--r--  1 root root  234K Feb 22  2018 libnspr4.so
-rw-r--r--  1 root root  1.3M Jul  6 11:25 libnss3.so
lrwxrwxrwx  1 root root    40 May  3 10:19 libnss_compat.so -> /lib/x86_64-linux-gnu/libnss_compat.so.2
lrwxrwxrwx  1 root root    37 May  3 10:19 libnss_dns.so -> /lib/x86_64-linux-gnu/libnss_dns.so.2
lrwxrwxrwx  1 root root    39 May  3 10:19 libnss_files.so -> /lib/x86_64-linux-gnu/libnss_files.so.2
lrwxrwxrwx  1 root root    40 May  3 10:19 libnss_hesiod.so -> /lib/x86_64-linux-gnu/libnss_hesiod.so.2
lrwxrwxrwx  1 root root    37 May  3 10:19 libnss_nis.so -> /lib/x86_64-linux-gnu/libnss_nis.so.2
lrwxrwxrwx  1 root root    41 May  3 10:19 libnss_nisplus.so -> /lib/x86_64-linux-gnu/libnss_nisplus.so.2
-rw-r--r--  1 root root  187K Jul  6 11:25 libnssutil3.so
lrwxrwxrwx  1 root root    16 Jun 20  2018 libnuma.so.1 -> libnuma.so.1.0.0
-rw-r--r--  1 root root   43K Jun 20  2018 libnuma.so.1.0.0
lrwxrwxrwx  1 root root    15 Aug  9  2019 libnvblas.so -> libnvblas.so.10
lrwxrwxrwx  1 root root    23 Aug  9  2019 libnvblas.so.10 -> libnvblas.so.10.2.1.243
-rw-r--r--  1 root root  526K Aug  9  2019 libnvblas.so.10.2.1.243
lrwxrwxrwx  1 root root    12 Mar 26  2018 libodbc.so.1 -> libodbc.so.2
lrwxrwxrwx  1 root root    16 Mar 26  2018 libodbc.so.2 -> libodbc.so.2.0.0
-rw-r--r--  1 root root  422K Mar 26  2018 libodbc.so.2.0.0
lrwxrwxrwx  1 root root    14 Mar 26  2018 libodbccr.so.1 -> libodbccr.so.2
lrwxrwxrwx  1 root root    18 Mar 26  2018 libodbccr.so.2 -> libodbccr.so.2.0.0
-rw-r--r--  1 root root   43K Mar 26  2018 libodbccr.so.2.0.0
lrwxrwxrwx  1 root root    16 Mar 26  2018 libodbcinst.so.1 -> libodbcinst.so.2
lrwxrwxrwx  1 root root    20 Mar 26  2018 libodbcinst.so.2 -> libodbcinst.so.2.0.0
-rw-r--r--  1 root root   71K Mar 26  2018 libodbcinst.so.2.0.0
lrwxrwxrwx  1 root root    15 Jun  2  2014 libogg.so.0 -> libogg.so.0.8.2
-rw-r--r--  1 root root   34K Jun  2  2014 libogg.so.0.8.2
lrwxrwxrwx  1 root root    23 Feb  5  2018 libompitrace.so.20 -> libompitrace.so.20.10.0
lrwxrwxrwx  1 root root    35 Feb  5  2018 libompitrace.so.20.10.0 -> openmpi/lib/libompitrace.so.20.10.0
-rw-r--r--  1 root root  1.7M Oct 20  2021 libopcodes-2.30-system.so
lrwxrwxrwx  1 root root    22 Feb  5  2018 libopen-pal.so.20 -> libopen-pal.so.20.10.1
lrwxrwxrwx  1 root root    34 Feb  5  2018 libopen-pal.so.20.10.1 -> openmpi/lib/libopen-pal.so.20.10.1
lrwxrwxrwx  1 root root    22 Feb  5  2018 libopen-rte.so.20 -> libopen-rte.so.20.10.1
lrwxrwxrwx  1 root root    34 Feb  5  2018 libopen-rte.so.20.10.1 -> openmpi/lib/libopen-rte.so.20.10.1
lrwxrwxrwx  1 root root    22 Sep 19  2017 libopenblas.a -> libopenblasp-r0.2.20.a
lrwxrwxrwx  1 root root    23 Sep 19  2017 libopenblas.so -> libopenblasp-r0.2.20.so
lrwxrwxrwx  1 root root    23 Sep 19  2017 libopenblas.so.0 -> libopenblasp-r0.2.20.so
-rw-r--r--  1 root root   51M Sep 19  2017 libopenblasp-r0.2.20.a
-rw-r--r--  1 root root   33M Sep 19  2017 libopenblasp-r0.2.20.so
-rw-r--r--  1 root root  483K Sep 20  2018 libopencv_aruco.a
lrwxrwxrwx  1 root root    22 Sep 20  2018 libopencv_aruco.so -> libopencv_aruco.so.3.2
lrwxrwxrwx  1 root root    24 Sep 20  2018 libopencv_aruco.so.3.2 -> libopencv_aruco.so.3.2.0
-rw-r--r--  1 root root  285K Sep 20  2018 libopencv_aruco.so.3.2.0
-rw-r--r--  1 root root  116K Sep 20  2018 libopencv_bgsegm.a
lrwxrwxrwx  1 root root    23 Sep 20  2018 libopencv_bgsegm.so -> libopencv_bgsegm.so.3.2
lrwxrwxrwx  1 root root    25 Sep 20  2018 libopencv_bgsegm.so.3.2 -> libopencv_bgsegm.so.3.2.0
-rw-r--r--  1 root root   51K Sep 20  2018 libopencv_bgsegm.so.3.2.0
-rw-r--r--  1 root root  449K Sep 20  2018 libopencv_bioinspired.a
lrwxrwxrwx  1 root root    28 Sep 20  2018 libopencv_bioinspired.so -> libopencv_bioinspired.so.3.2
lrwxrwxrwx  1 root root    30 Sep 20  2018 libopencv_bioinspired.so.3.2 -> libopencv_bioinspired.so.3.2.0
-rw-r--r--  1 root root  171K Sep 20  2018 libopencv_bioinspired.so.3.2.0
-rw-r--r--  1 root root  2.3M Sep 20  2018 libopencv_calib3d.a
lrwxrwxrwx  1 root root    24 Sep 20  2018 libopencv_calib3d.so -> libopencv_calib3d.so.3.2
lrwxrwxrwx  1 root root    26 Sep 20  2018 libopencv_calib3d.so.3.2 -> libopencv_calib3d.so.3.2.0
-rw-r--r--  1 root root  1.4M Sep 20  2018 libopencv_calib3d.so.3.2.0
-rw-r--r--  1 root root  633K Sep 20  2018 libopencv_ccalib.a
lrwxrwxrwx  1 root root    23 Sep 20  2018 libopencv_ccalib.so -> libopencv_ccalib.so.3.2
lrwxrwxrwx  1 root root    25 Sep 20  2018 libopencv_ccalib.so.3.2 -> libopencv_ccalib.so.3.2.0
-rw-r--r--  1 root root  363K Sep 20  2018 libopencv_ccalib.so.3.2.0
-rw-r--r--  1 root root  4.6M Sep 20  2018 libopencv_core.a
lrwxrwxrwx  1 root root    21 Sep 20  2018 libopencv_core.so -> libopencv_core.so.3.2
lrwxrwxrwx  1 root root    23 Sep 20  2018 libopencv_core.so.3.2 -> libopencv_core.so.3.2.0
-rw-r--r--  1 root root  2.3M Sep 20  2018 libopencv_core.so.3.2.0
-rw-r--r--  1 root root  1.4M Sep 20  2018 libopencv_datasets.a
lrwxrwxrwx  1 root root    25 Sep 20  2018 libopencv_datasets.so -> libopencv_datasets.so.3.2
lrwxrwxrwx  1 root root    27 Sep 20  2018 libopencv_datasets.so.3.2 -> libopencv_datasets.so.3.2.0
-rw-r--r--  1 root root  399K Sep 20  2018 libopencv_datasets.so.3.2.0
-rw-r--r--  1 root root  206K Sep 20  2018 libopencv_dpm.a
lrwxrwxrwx  1 root root    20 Sep 20  2018 libopencv_dpm.so -> libopencv_dpm.so.3.2
lrwxrwxrwx  1 root root    22 Sep 20  2018 libopencv_dpm.so.3.2 -> libopencv_dpm.so.3.2.0
-rw-r--r--  1 root root   99K Sep 20  2018 libopencv_dpm.so.3.2.0
-rw-r--r--  1 root root  331K Sep 20  2018 libopencv_face.a
lrwxrwxrwx  1 root root    21 Sep 20  2018 libopencv_face.so -> libopencv_face.so.3.2
lrwxrwxrwx  1 root root    23 Sep 20  2018 libopencv_face.so.3.2 -> libopencv_face.so.3.2.0
-rw-r--r--  1 root root  143K Sep 20  2018 libopencv_face.so.3.2.0
-rw-r--r--  1 root root  1.4M Sep 20  2018 libopencv_features2d.a
lrwxrwxrwx  1 root root    27 Sep 20  2018 libopencv_features2d.so -> libopencv_features2d.so.3.2
lrwxrwxrwx  1 root root    29 Sep 20  2018 libopencv_features2d.so.3.2 -> libopencv_features2d.so.3.2.0
-rw-r--r--  1 root root  723K Sep 20  2018 libopencv_features2d.so.3.2.0
-rw-r--r--  1 root root 1007K Sep 20  2018 libopencv_flann.a
lrwxrwxrwx  1 root root    22 Sep 20  2018 libopencv_flann.so -> libopencv_flann.so.3.2
lrwxrwxrwx  1 root root    24 Sep 20  2018 libopencv_flann.so.3.2 -> libopencv_flann.so.3.2.0
-rw-r--r--  1 root root  319K Sep 20  2018 libopencv_flann.so.3.2.0
-rw-r--r--  1 root root   43K Sep 20  2018 libopencv_freetype.a
lrwxrwxrwx  1 root root    25 Sep 20  2018 libopencv_freetype.so -> libopencv_freetype.so.3.2
lrwxrwxrwx  1 root root    27 Sep 20  2018 libopencv_freetype.so.3.2 -> libopencv_freetype.so.3.2.0
-rw-r--r--  1 root root   27K Sep 20  2018 libopencv_freetype.so.3.2.0
-rw-r--r--  1 root root   73K Sep 20  2018 libopencv_fuzzy.a
lrwxrwxrwx  1 root root    22 Sep 20  2018 libopencv_fuzzy.so -> libopencv_fuzzy.so.3.2
lrwxrwxrwx  1 root root    24 Sep 20  2018 libopencv_fuzzy.so.3.2 -> libopencv_fuzzy.so.3.2.0
-rw-r--r--  1 root root   54K Sep 20  2018 libopencv_fuzzy.so.3.2.0
-rw-r--r--  1 root root   75K Sep 20  2018 libopencv_hdf.a
lrwxrwxrwx  1 root root    20 Sep 20  2018 libopencv_hdf.so -> libopencv_hdf.so.3.2
lrwxrwxrwx  1 root root    22 Sep 20  2018 libopencv_hdf.so.3.2 -> libopencv_hdf.so.3.2.0
-rw-r--r--  1 root root   39K Sep 20  2018 libopencv_hdf.so.3.2.0
-rw-r--r--  1 root root   94K Sep 20  2018 libopencv_highgui.a
lrwxrwxrwx  1 root root    24 Sep 20  2018 libopencv_highgui.so -> libopencv_highgui.so.3.2
lrwxrwxrwx  1 root root    26 Sep 20  2018 libopencv_highgui.so.3.2 -> libopencv_highgui.so.3.2.0
-rw-r--r--  1 root root   47K Sep 20  2018 libopencv_highgui.so.3.2.0
-rw-r--r--  1 root root  756K Sep 20  2018 libopencv_imgcodecs.a
lrwxrwxrwx  1 root root    26 Sep 20  2018 libopencv_imgcodecs.so -> libopencv_imgcodecs.so.3.2
lrwxrwxrwx  1 root root    28 Sep 20  2018 libopencv_imgcodecs.so.3.2 -> libopencv_imgcodecs.so.3.2.0
-rw-r--r--  1 root root  227K Sep 20  2018 libopencv_imgcodecs.so.3.2.0
-rw-r--r--  1 root root  5.5M Sep 20  2018 libopencv_imgproc.a
lrwxrwxrwx  1 root root    24 Sep 20  2018 libopencv_imgproc.so -> libopencv_imgproc.so.3.2
lrwxrwxrwx  1 root root    26 Sep 20  2018 libopencv_imgproc.so.3.2 -> libopencv_imgproc.so.3.2.0
-rw-r--r--  1 root root  2.8M Sep 20  2018 libopencv_imgproc.so.3.2.0
-rw-r--r--  1 root root  312K Sep 20  2018 libopencv_line_descriptor.a
lrwxrwxrwx  1 root root    32 Sep 20  2018 libopencv_line_descriptor.so -> libopencv_line_descriptor.so.3.2
lrwxrwxrwx  1 root root    34 Sep 20  2018 libopencv_line_descriptor.so.3.2 -> libopencv_line_descriptor.so.3.2.0
-rw-r--r--  1 root root  159K Sep 20  2018 libopencv_line_descriptor.so.3.2.0
-rw-r--r--  1 root root  1.5M Sep 20  2018 libopencv_ml.a
lrwxrwxrwx  1 root root    19 Sep 20  2018 libopencv_ml.so -> libopencv_ml.so.3.2
lrwxrwxrwx  1 root root    21 Sep 20  2018 libopencv_ml.so.3.2 -> libopencv_ml.so.3.2.0
-rw-r--r--  1 root root  679K Sep 20  2018 libopencv_ml.so.3.2.0
-rw-r--r--  1 root root  689K Sep 20  2018 libopencv_objdetect.a
lrwxrwxrwx  1 root root    26 Sep 20  2018 libopencv_objdetect.so -> libopencv_objdetect.so.3.2
lrwxrwxrwx  1 root root    28 Sep 20  2018 libopencv_objdetect.so.3.2 -> libopencv_objdetect.so.3.2.0
-rw-r--r--  1 root root  343K Sep 20  2018 libopencv_objdetect.so.3.2.0
-rw-r--r--  1 root root  648K Sep 20  2018 libopencv_optflow.a
lrwxrwxrwx  1 root root    24 Sep 20  2018 libopencv_optflow.so -> libopencv_optflow.so.3.2
lrwxrwxrwx  1 root root    26 Sep 20  2018 libopencv_optflow.so.3.2 -> libopencv_optflow.so.3.2.0
-rw-r--r--  1 root root  327K Sep 20  2018 libopencv_optflow.so.3.2.0
-rw-r--r--  1 root root   62K Sep 20  2018 libopencv_phase_unwrapping.a
lrwxrwxrwx  1 root root    33 Sep 20  2018 libopencv_phase_unwrapping.so -> libopencv_phase_unwrapping.so.3.2
lrwxrwxrwx  1 root root    35 Sep 20  2018 libopencv_phase_unwrapping.so.3.2 -> libopencv_phase_unwrapping.so.3.2.0
-rw-r--r--  1 root root   35K Sep 20  2018 libopencv_phase_unwrapping.so.3.2.0
-rw-r--r--  1 root root  1.5M Sep 20  2018 libopencv_photo.a
lrwxrwxrwx  1 root root    22 Sep 20  2018 libopencv_photo.so -> libopencv_photo.so.3.2
lrwxrwxrwx  1 root root    24 Sep 20  2018 libopencv_photo.so.3.2 -> libopencv_photo.so.3.2.0
-rw-r--r--  1 root root  771K Sep 20  2018 libopencv_photo.so.3.2.0
-rw-r--r--  1 root root   66K Sep 20  2018 libopencv_plot.a
lrwxrwxrwx  1 root root    21 Sep 20  2018 libopencv_plot.so -> libopencv_plot.so.3.2
lrwxrwxrwx  1 root root    23 Sep 20  2018 libopencv_plot.so.3.2 -> libopencv_plot.so.3.2.0
-rw-r--r--  1 root root   43K Sep 20  2018 libopencv_plot.so.3.2.0
-rw-r--r--  1 root root  268K Sep 20  2018 libopencv_reg.a
lrwxrwxrwx  1 root root    20 Sep 20  2018 libopencv_reg.so -> libopencv_reg.so.3.2
lrwxrwxrwx  1 root root    22 Sep 20  2018 libopencv_reg.so.3.2 -> libopencv_reg.so.3.2.0
-rw-r--r--  1 root root  127K Sep 20  2018 libopencv_reg.so.3.2.0
-rw-r--r--  1 root root  982K Sep 20  2018 libopencv_rgbd.a
lrwxrwxrwx  1 root root    21 Sep 20  2018 libopencv_rgbd.so -> libopencv_rgbd.so.3.2
lrwxrwxrwx  1 root root    23 Sep 20  2018 libopencv_rgbd.so.3.2 -> libopencv_rgbd.so.3.2.0
-rw-r--r--  1 root root  535K Sep 20  2018 libopencv_rgbd.so.3.2.0
-rw-r--r--  1 root root  353K Sep 20  2018 libopencv_saliency.a
lrwxrwxrwx  1 root root    25 Sep 20  2018 libopencv_saliency.so -> libopencv_saliency.so.3.2
lrwxrwxrwx  1 root root    27 Sep 20  2018 libopencv_saliency.so.3.2 -> libopencv_saliency.so.3.2.0
-rw-r--r--  1 root root  179K Sep 20  2018 libopencv_saliency.so.3.2.0
-rw-r--r--  1 root root  428K Sep 20  2018 libopencv_shape.a
lrwxrwxrwx  1 root root    22 Sep 20  2018 libopencv_shape.so -> libopencv_shape.so.3.2
lrwxrwxrwx  1 root root    24 Sep 20  2018 libopencv_shape.so.3.2 -> libopencv_shape.so.3.2.0
-rw-r--r--  1 root root  187K Sep 20  2018 libopencv_shape.so.3.2.0
-rw-r--r--  1 root root  327K Sep 20  2018 libopencv_stereo.a
lrwxrwxrwx  1 root root    23 Sep 20  2018 libopencv_stereo.so -> libopencv_stereo.so.3.2
lrwxrwxrwx  1 root root    25 Sep 20  2018 libopencv_stereo.so.3.2 -> libopencv_stereo.so.3.2.0
-rw-r--r--  1 root root  123K Sep 20  2018 libopencv_stereo.so.3.2.0
-rw-r--r--  1 root root  1.1M Sep 20  2018 libopencv_stitching.a
lrwxrwxrwx  1 root root    26 Sep 20  2018 libopencv_stitching.so -> libopencv_stitching.so.3.2
lrwxrwxrwx  1 root root    28 Sep 20  2018 libopencv_stitching.so.3.2 -> libopencv_stitching.so.3.2.0
-rw-r--r--  1 root root  519K Sep 20  2018 libopencv_stitching.so.3.2.0
-rw-r--r--  1 root root  156K Sep 20  2018 libopencv_structured_light.a
lrwxrwxrwx  1 root root    33 Sep 20  2018 libopencv_structured_light.so -> libopencv_structured_light.so.3.2
lrwxrwxrwx  1 root root    35 Sep 20  2018 libopencv_structured_light.so.3.2 -> libopencv_structured_light.so.3.2.0
-rw-r--r--  1 root root   91K Sep 20  2018 libopencv_structured_light.so.3.2.0
-rw-r--r--  1 root root  295K Sep 20  2018 libopencv_superres.a
lrwxrwxrwx  1 root root    25 Sep 20  2018 libopencv_superres.so -> libopencv_superres.so.3.2
lrwxrwxrwx  1 root root    27 Sep 20  2018 libopencv_superres.so.3.2 -> libopencv_superres.so.3.2.0
-rw-r--r--  1 root root  151K Sep 20  2018 libopencv_superres.so.3.2.0
-rw-r--r--  1 root root  1.1M Sep 20  2018 libopencv_surface_matching.a
lrwxrwxrwx  1 root root    33 Sep 20  2018 libopencv_surface_matching.so -> libopencv_surface_matching.so.3.2
lrwxrwxrwx  1 root root    35 Sep 20  2018 libopencv_surface_matching.so.3.2 -> libopencv_surface_matching.so.3.2.0
-rw-r--r--  1 root root  259K Sep 20  2018 libopencv_surface_matching.so.3.2.0
-rw-r--r--  1 root root  761K Sep 20  2018 libopencv_text.a
lrwxrwxrwx  1 root root    21 Sep 20  2018 libopencv_text.so -> libopencv_text.so.3.2
lrwxrwxrwx  1 root root    23 Sep 20  2018 libopencv_text.so.3.2 -> libopencv_text.so.3.2.0
-rw-r--r--  1 root root  367K Sep 20  2018 libopencv_text.so.3.2.0
-rw-r--r--  1 root root  1.6M Sep 20  2018 libopencv_ts.a
-rw-r--r--  1 root root  700K Sep 20  2018 libopencv_video.a
lrwxrwxrwx  1 root root    22 Sep 20  2018 libopencv_video.so -> libopencv_video.so.3.2
lrwxrwxrwx  1 root root    24 Sep 20  2018 libopencv_video.so.3.2 -> libopencv_video.so.3.2.0
-rw-r--r--  1 root root  395K Sep 20  2018 libopencv_video.so.3.2.0
-rw-r--r--  1 root root  488K Sep 20  2018 libopencv_videoio.a
lrwxrwxrwx  1 root root    24 Sep 20  2018 libopencv_videoio.so -> libopencv_videoio.so.3.2
lrwxrwxrwx  1 root root    26 Sep 20  2018 libopencv_videoio.so.3.2 -> libopencv_videoio.so.3.2.0
-rw-r--r--  1 root root  211K Sep 20  2018 libopencv_videoio.so.3.2.0
-rw-r--r--  1 root root  688K Sep 20  2018 libopencv_videostab.a
lrwxrwxrwx  1 root root    26 Sep 20  2018 libopencv_videostab.so -> libopencv_videostab.so.3.2
lrwxrwxrwx  1 root root    28 Sep 20  2018 libopencv_videostab.so.3.2 -> libopencv_videostab.so.3.2.0
-rw-r--r--  1 root root  347K Sep 20  2018 libopencv_videostab.so.3.2.0
-rw-r--r--  1 root root 1019K Sep 20  2018 libopencv_viz.a
lrwxrwxrwx  1 root root    20 Sep 20  2018 libopencv_viz.so -> libopencv_viz.so.3.2
lrwxrwxrwx  1 root root    22 Sep 20  2018 libopencv_viz.so.3.2 -> libopencv_viz.so.3.2.0
-rw-r--r--  1 root root  363K Sep 20  2018 libopencv_viz.so.3.2.0
-rw-r--r--  1 root root  2.3M Sep 20  2018 libopencv_ximgproc.a
lrwxrwxrwx  1 root root    25 Sep 20  2018 libopencv_ximgproc.so -> libopencv_ximgproc.so.3.2
lrwxrwxrwx  1 root root    27 Sep 20  2018 libopencv_ximgproc.so.3.2 -> libopencv_ximgproc.so.3.2.0
-rw-r--r--  1 root root  983K Sep 20  2018 libopencv_ximgproc.so.3.2.0
-rw-r--r--  1 root root  186K Sep 20  2018 libopencv_xobjdetect.a
lrwxrwxrwx  1 root root    27 Sep 20  2018 libopencv_xobjdetect.so -> libopencv_xobjdetect.so.3.2
lrwxrwxrwx  1 root root    29 Sep 20  2018 libopencv_xobjdetect.so.3.2 -> libopencv_xobjdetect.so.3.2.0
-rw-r--r--  1 root root   87K Sep 20  2018 libopencv_xobjdetect.so.3.2.0
-rw-r--r--  1 root root  441K Sep 20  2018 libopencv_xphoto.a
lrwxrwxrwx  1 root root    23 Sep 20  2018 libopencv_xphoto.so -> libopencv_xphoto.so.3.2
lrwxrwxrwx  1 root root    25 Sep 20  2018 libopencv_xphoto.so.3.2 -> libopencv_xphoto.so.3.2.0
-rw-r--r--  1 root root  220K Sep 20  2018 libopencv_xphoto.so.3.2.0
-rw-r--r--  1 root root  343K Aug 20  2019 libopenjp2.so.2.3.0
lrwxrwxrwx  1 root root    19 Aug 20  2019 libopenjp2.so.7 -> libopenjp2.so.2.3.0
lrwxrwxrwx  1 root root    19 Feb  4  2018 libopenmpt.so.0 -> libopenmpt.so.0.1.1
-rw-r--r--  1 root root  1.5M Feb  4  2018 libopenmpt.so.0.1.1
lrwxrwxrwx  1 root root    16 Feb 17  2016 libopus.so.0 -> libopus.so.0.5.2
-rw-r--r--  1 root root  295K Feb 17  2016 libopus.so.0.5.2
lrwxrwxrwx  1 root root    32 Feb  5  2018 liboshmem.so.20 -> openmpi/lib/liboshmem.so.20.10.1
lrwxrwxrwx  1 root root    19 Jan  4  2021 libp11-kit.so.0 -> libp11-kit.so.0.3.0
-rw-r--r--  1 root root  1.2M Jan  4  2021 libp11-kit.so.0.3.0
lrwxrwxrwx  1 root root    15 May 23  2018 libpanel.so.5 -> libpanel.so.5.9
-rw-r--r--  1 root root   14K May 23  2018 libpanel.so.5.9
lrwxrwxrwx  1 root root    16 May 23  2018 libpanelw.so.5 -> libpanelw.so.5.9
-rw-r--r--  1 root root   14K May 23  2018 libpanelw.so.5.9
lrwxrwxrwx  1 root root    25 Aug 21  2018 libpango-1.0.so.0 -> libpango-1.0.so.0.4000.14
-rw-r--r--  1 root root  306K Aug 21  2018 libpango-1.0.so.0.4000.14
lrwxrwxrwx  1 root root    30 Aug 21  2018 libpangocairo-1.0.so.0 -> libpangocairo-1.0.so.0.4000.14
-rw-r--r--  1 root root   50K Aug 21  2018 libpangocairo-1.0.so.0.4000.14
lrwxrwxrwx  1 root root    28 Aug 21  2018 libpangoft2-1.0.so.0 -> libpangoft2-1.0.so.0.4000.14
-rw-r--r--  1 root root   86K Aug 21  2018 libpangoft2-1.0.so.0.4000.14
lrwxrwxrwx  1 root root    20 Mar 24  2018 libpathplan.so -> libpathplan.so.4.0.0
lrwxrwxrwx  1 root root    20 Mar 24  2018 libpathplan.so.4 -> libpathplan.so.4.0.0
-rw-r--r--  1 root root   31K Mar 24  2018 libpathplan.so.4.0.0
lrwxrwxrwx  1 root root    22 Mar 18  2018 libpciaccess.so.0 -> libpciaccess.so.0.11.1
-rw-r--r--  1 root root   35K Mar 18  2018 libpciaccess.so.0.11.1
-rw-r--r--  1 root root  615K May 17 07:42 libpcre.a
lrwxrwxrwx  1 root root    34 May 17 07:42 libpcre.so -> /lib/x86_64-linux-gnu/libpcre.so.3
-rw-r--r--  1 root root  576K May 17 07:42 libpcre16.a
lrwxrwxrwx  1 root root    19 May 17 07:42 libpcre16.so -> libpcre16.so.3.13.3
lrwxrwxrwx  1 root root    19 May 17 07:42 libpcre16.so.3 -> libpcre16.so.3.13.3
-rw-r--r--  1 root root  410K May 17 07:42 libpcre16.so.3.13.3
-rw-r--r--  1 root root  558K May 17 07:42 libpcre32.a
lrwxrwxrwx  1 root root    19 May 17 07:42 libpcre32.so -> libpcre32.so.3.13.3
lrwxrwxrwx  1 root root    19 May 17 07:42 libpcre32.so.3 -> libpcre32.so.3.13.3
-rw-r--r--  1 root root  390K May 17 07:42 libpcre32.so.3.13.3
-rw-r--r--  1 root root   51K May 17 07:42 libpcrecpp.a
lrwxrwxrwx  1 root root    19 May 17 07:42 libpcrecpp.so -> libpcrecpp.so.0.0.1
lrwxrwxrwx  1 root root    19 May 17 07:42 libpcrecpp.so.0 -> libpcrecpp.so.0.0.1
-rw-r--r--  1 root root   34K May 17 07:42 libpcrecpp.so.0.0.1
-rw-r--r--  1 root root  6.6K May 17 07:42 libpcreposix.a
lrwxrwxrwx  1 root root    22 May 17 07:42 libpcreposix.so -> libpcreposix.so.3.13.3
lrwxrwxrwx  1 root root    22 May 17 07:42 libpcreposix.so.3 -> libpcreposix.so.3.13.3
-rw-r--r--  1 root root  9.9K May 17 07:42 libpcreposix.so.3.13.3
lrwxrwxrwx  1 root root    17 Oct 19  2020 libperl.so.5.26 -> libperl.so.5.26.1
-rw-r--r--  1 root root  2.0M Oct 19  2020 libperl.so.5.26.1
lrwxrwxrwx  1 root root    21 Dec 17  2017 libpixman-1.so.0 -> libpixman-1.so.0.34.0
-rw-r--r--  1 root root  659K Dec 17  2017 libpixman-1.so.0.34.0
-rw-r--r--  1 root root   19K Feb 22  2018 libplc4.so
-rw-r--r--  1 root root   15K Feb 22  2018 libplds4.so
lrwxrwxrwx  1 root root    10 Apr 30  2019 libpng.a -> libpng16.a
lrwxrwxrwx  1 root root    11 Apr 30  2019 libpng.so -> libpng16.so
-rw-r--r--  1 root root  328K Apr 30  2019 libpng16.a
lrwxrwxrwx  1 root root    19 Apr 30  2019 libpng16.so -> libpng16.so.16.34.0
lrwxrwxrwx  1 root root    19 Apr 30  2019 libpng16.so.16 -> libpng16.so.16.34.0
-rw-r--r--  1 root root  198K Apr 30  2019 libpng16.so.16.34.0
lrwxrwxrwx  1 root root    20 Nov 26  2020 libpoppler.so.73 -> libpoppler.so.73.0.0
-rw-r--r--  1 root root  2.6M Nov 26  2020 libpoppler.so.73.0.0
lrwxrwxrwx  1 root root    13 May 18 01:58 libpq.so.5 -> libpq.so.5.10
-rw-r--r--  1 root root  290K May 18 01:58 libpq.so.5.10
lrwxrwxrwx  1 root root    17 Jun 18  2017 libproj.so.12 -> libproj.so.12.0.0
-rw-r--r--  1 root root  420K Jun 18  2017 libproj.so.12.0.0
lrwxrwxrwx  1 root root    17 Dec  8  2020 libproxy.so.1 -> libproxy.so.1.0.0
-rw-r--r--  1 root root  123K Dec  8  2020 libproxy.so.1.0.0
lrwxrwxrwx  1 root root    15 Mar  4  2018 libpsl.so.5 -> libpsl.so.5.2.0
-rw-r--r--  1 root root   54K Mar  4  2018 libpsl.so.5.2.0
lrwxrwxrwx  1 root root    40 Aug 16 08:53 libpsm_infinipath.so.1 -> /etc/alternatives/libpsm_infinipath.so.1
-rw-r--r--  1 root root  6.0M May  3 10:19 libpthread.a
-rw-r--r--  1 root root   252 May  3 10:19 libpthread.so
-rw-r--r--  1 root root   29K May  3 10:19 libpthread_nonshared.a
lrwxrwxrwx  1 root root    19 Jul  1 15:56 libpython2.7.so.1 -> libpython2.7.so.1.0
-rw-r--r--  1 root root  3.4M Jul  1 15:56 libpython2.7.so.1.0
lrwxrwxrwx  1 root root    17 Dec  9  2017 libqhull.so.7 -> libqhull.so.7.2.0
-rw-r--r--  1 root root  358K Dec  9  2017 libqhull.so.7.2.0
lrwxrwxrwx  1 root root    20 Mar 10  2020 libquadmath.so.0 -> libquadmath.so.0.0.0
-rw-r--r--  1 root root  256K Mar 10  2020 libquadmath.so.0.0.0
-rw-r--r--  1 root root   91K Apr 26  2016 libraw1394.a
-rw-r--r--  1 root root   969 Apr 26  2016 libraw1394.la
lrwxrwxrwx  1 root root    20 Apr 26  2016 libraw1394.so -> libraw1394.so.11.1.0
lrwxrwxrwx  1 root root    20 Apr 26  2016 libraw1394.so.11 -> libraw1394.so.11.1.0
-rw-r--r--  1 root root   59K Apr 26  2016 libraw1394.so.11.1.0
lrwxrwxrwx  1 root root    21 Aug  5  2019 librdmacm.so.1 -> librdmacm.so.1.1.17.1
-rw-r--r--  1 root root   87K Aug  5  2019 librdmacm.so.1.1.17.1
-rw-r--r--  1 root root  136K May  3 10:19 libresolv.a
lrwxrwxrwx  1 root root    36 May  3 10:19 libresolv.so -> /lib/x86_64-linux-gnu/libresolv.so.2
lrwxrwxrwx  1 root root    20 Oct 24  2016 librest-0.7.so.0 -> librest-0.7.so.0.0.0
-rw-r--r--  1 root root   88K Oct 24  2016 librest-0.7.so.0.0.0
lrwxrwxrwx  1 root root    18 Dec 15  2017 libroken.so.18 -> libroken.so.18.1.0
-rw-r--r--  1 root root   87K Dec 15  2017 libroken.so.18.1.0
-rw-r--r--  1 root root   52K May  3 10:19 librpcsvc.a
lrwxrwxrwx  1 root root    20 Jul 28  2020 librsvg-2.so.2 -> librsvg-2.so.2.40.20
-rw-r--r--  1 root root  223K Jul 28  2020 librsvg-2.so.2.40.20
-rw-r--r--  1 root root   76K May  3 10:19 librt.a
lrwxrwxrwx  1 root root    32 May  3 10:19 librt.so -> /lib/x86_64-linux-gnu/librt.so.1
-rw-r--r--  1 root root  111K Apr 27  2016 librtmp.so.1
lrwxrwxrwx  1 root root    18 Feb 15  2022 libsasl2.so.2 -> libsasl2.so.2.0.25
-rw-r--r--  1 root root  107K Feb 15  2022 libsasl2.so.2.0.25
-rw-r--r--  1 root root  251K Mar  1  2018 libsemanage.so.1
lrwxrwxrwx  1 root root    19 Mar 31 20:53 libsensors.so.4 -> libsensors.so.4.4.0
-rw-r--r--  1 root root   59K Mar 31 20:53 libsensors.so.4.4.0
lrwxrwxrwx  1 root root    17 Aug  6  2017 libshine.so.3 -> libshine.so.3.0.1
-rw-r--r--  1 root root   43K Aug  6  2017 libshine.so.3.0.1
-rw-r--r--  1 root root  181K Jul  6 11:25 libsmime3.so
lrwxrwxrwx  1 root root    18 Sep  5  2017 libsnappy.so.1 -> libsnappy.so.1.1.7
-rw-r--r--  1 root root   31K Sep  5  2017 libsnappy.so.1.1.7
lrwxrwxrwx  1 root root    20 Oct  5  2017 libsocket++.so.1 -> libsocket++.so.1.0.2
-rw-r--r--  1 root root  132K Oct  5  2017 libsocket++.so.1.0.2
lrwxrwxrwx  1 root root    20 Oct  9  2019 libsoup-2.4.so.1 -> libsoup-2.4.so.1.8.0
-rw-r--r--  1 root root  971K Oct  9  2019 libsoup-2.4.so.1.8.0
lrwxrwxrwx  1 root root    26 Oct  9  2019 libsoup-gnome-2.4.so.1 -> libsoup-gnome-2.4.so.1.8.0
-rw-r--r--  1 root root  9.9K Oct  9  2019 libsoup-gnome-2.4.so.1.8.0
lrwxrwxrwx  1 root root    16 Nov  4  2017 libsoxr.so.0 -> libsoxr.so.0.1.1
-rw-r--r--  1 root root  183K Nov  4  2017 libsoxr.so.0.1.1
lrwxrwxrwx  1 root root    22 Nov 17  2016 libspatialite.so.7 -> libspatialite.so.7.1.0
-rw-r--r--  1 root root  5.6M Nov 17  2016 libspatialite.so.7.1.0
lrwxrwxrwx  1 root root    17 Feb  8  2022 libspeex.so.1 -> libspeex.so.1.5.0
-rw-r--r--  1 root root  102K Feb  8  2022 libspeex.so.1.5.0
lrwxrwxrwx  1 root root    19 Apr 28 14:04 libsqlite3.so.0 -> libsqlite3.so.0.8.6
-rw-r--r--  1 root root  1.1M Apr 28 14:04 libsqlite3.so.0.8.6
lrwxrwxrwx  1 root root    22 Jul 31  2020 libssh-gcrypt.so.4 -> libssh-gcrypt.so.4.5.0
-rw-r--r--  1 root root  465K Jul 31  2020 libssh-gcrypt.so.4.5.0
lrwxrwxrwx  1 root root    30 Jul 31  2020 libssh-gcrypt_threads.so.4 -> libssh-gcrypt_threads.so.4.5.0
-rw-r--r--  1 root root  6.0K Jul 31  2020 libssh-gcrypt_threads.so.4.5.0
-rw-r--r--  1 root root  564K May  3 17:51 libssl.so.1.1
-rw-r--r--  1 root root  321K Jul  6 11:25 libssl3.so
lrwxrwxrwx  1 root root    19 Mar 10  2020 libstdc++.so.6 -> libstdc++.so.6.0.25
-rw-r--r--  1 root root  1.6M Mar 10  2020 libstdc++.so.6.0.25
lrwxrwxrwx  1 root root    19 Nov  6  2017 libsuperlu.so.5 -> libsuperlu.so.5.2.1
-rw-r--r--  1 root root  447K Nov  6  2017 libsuperlu.so.5.2.1
-rw-r--r--  1 root root  220K May 18 20:01 libswresample.a
lrwxrwxrwx  1 root root    24 May 18 20:01 libswresample.so -> libswresample.so.2.9.100
lrwxrwxrwx  1 root root    24 May 18 20:01 libswresample.so.2 -> libswresample.so.2.9.100
-rw-r--r--  1 root root  122K May 18 20:01 libswresample.so.2.9.100
-rw-r--r--  1 root root  780K May 18 20:01 libswscale.a
lrwxrwxrwx  1 root root    21 May 18 20:01 libswscale.so -> libswscale.so.4.8.100
lrwxrwxrwx  1 root root    21 May 18 20:01 libswscale.so.4 -> libswscale.so.4.8.100
-rw-r--r--  1 root root  538K May 18 20:01 libswscale.so.4.8.100
lrwxrwxrwx  1 root root    14 Jul 28  2017 libsz.so.2 -> libsz.so.2.0.1
-rw-r--r--  1 root root  9.9K Jul 28  2017 libsz.so.2.0.1
lrwxrwxrwx  1 root root    17 Jan 21  2018 libtasn1.so.6 -> libtasn1.so.6.5.5
-rw-r--r--  1 root root   74K Jan 21  2018 libtasn1.so.6.5.5
lrwxrwxrwx  1 root root    11 Oct  4  2017 libtbb.so -> libtbb.so.2
-rw-r--r--  1 root root  230K Oct  4  2017 libtbb.so.2
lrwxrwxrwx  1 root root    17 Oct  4  2017 libtbbmalloc.so -> libtbbmalloc.so.2
-rw-r--r--  1 root root  110K Oct  4  2017 libtbbmalloc.so.2
lrwxrwxrwx  1 root root    23 Oct  4  2017 libtbbmalloc_proxy.so -> libtbbmalloc_proxy.so.2
-rw-r--r--  1 root root   11K Oct  4  2017 libtbbmalloc_proxy.so.2
-rw-r--r--  1 root root  1.7M Feb 22  2018 libtcl8.6.so
lrwxrwxrwx  1 root root    12 Feb 22  2018 libtcl8.6.so.0 -> libtcl8.6.so
lrwxrwxrwx  1 root root    21 Apr  7  2018 libtesseract.so.4 -> libtesseract.so.4.0.0
-rw-r--r--  1 root root  3.1M Apr  7  2018 libtesseract.so.4.0.0
lrwxrwxrwx  1 root root    16 Feb 16  2018 libthai.so.0 -> libthai.so.0.3.0
-rw-r--r--  1 root root   37K Feb 16  2018 libthai.so.0.3.0
lrwxrwxrwx  1 root root    19 May 27  2016 libtheora.so.0 -> libtheora.so.0.3.10
-rw-r--r--  1 root root  327K May 27  2016 libtheora.so.0.3.10
lrwxrwxrwx  1 root root    21 May 27  2016 libtheoradec.so.1 -> libtheoradec.so.1.1.4
-rw-r--r--  1 root root  119K May 27  2016 libtheoradec.so.1.1.4
lrwxrwxrwx  1 root root    21 May 27  2016 libtheoraenc.so.1 -> libtheoraenc.so.1.1.2
-rw-r--r--  1 root root  251K May 27  2016 libtheoraenc.so.1.1.2
lrwxrwxrwx  1 root root    39 May  3 10:19 libthread_db.so -> /lib/x86_64-linux-gnu/libthread_db.so.1
lrwxrwxrwx  1 root root    13 May 23  2018 libtic.so.5 -> libtic.so.5.9
-rw-r--r--  1 root root   62K May 23  2018 libtic.so.5.9
-rw-r--r--  1 root root  752K May 11 15:09 libtiff.a
lrwxrwxrwx  1 root root    16 May 11 15:09 libtiff.so -> libtiff.so.5.3.0
lrwxrwxrwx  1 root root    16 May 11 15:09 libtiff.so.5 -> libtiff.so.5.3.0
-rw-r--r--  1 root root  479K May 11 15:09 libtiff.so.5.3.0
-rw-r--r--  1 root root  8.5K May 11 15:09 libtiffxx.a
lrwxrwxrwx  1 root root    18 May 11 15:09 libtiffxx.so -> libtiffxx.so.5.3.0
lrwxrwxrwx  1 root root    18 May 11 15:09 libtiffxx.so.5 -> libtiffxx.so.5.3.0
-rw-r--r--  1 root root   10K May 11 15:09 libtiffxx.so.5.3.0
-rw-r--r--  1 root root  1.4M Mar 24  2018 libtk8.6.so
lrwxrwxrwx  1 root root    11 Mar 24  2018 libtk8.6.so.0 -> libtk8.6.so
lrwxrwxrwx  1 root root    16 Mar 10  2020 libtsan.so.0 -> libtsan.so.0.0.0
-rw-r--r--  1 root root  942K Mar 10  2020 libtsan.so.0.0.0
-rw-r--r--  1 root root  3.6K Mar 10  2020 libtsan_preinit.o
lrwxrwxrwx  1 root root    19 Aug 29  2017 libtwolame.so.0 -> libtwolame.so.0.0.0
-rw-r--r--  1 root root  122K Aug 29  2017 libtwolame.so.0.0.0
lrwxrwxrwx  1 root root    17 Dec  4  2019 libubsan.so.0 -> libubsan.so.0.0.0
-rw-r--r--  1 root root  322K Dec  4  2019 libubsan.so.0.0.0
lrwxrwxrwx  1 root root    21 Mar 21  2019 libunistring.so.2 -> libunistring.so.2.1.0
-rw-r--r--  1 root root  1.5M Mar 21  2019 libunistring.so.2.1.0
lrwxrwxrwx  1 root root    22 Jul 12 13:52 liburiparser.so.1 -> liburiparser.so.1.0.20
-rw-r--r--  1 root root  110K Jul 12 13:52 liburiparser.so.1.0.20
-rw-r--r--  1 root root   15K May  3 10:19 libutil.a
lrwxrwxrwx  1 root root    34 May  3 10:19 libutil.so -> /lib/x86_64-linux-gnu/libutil.so.1
lrwxrwxrwx  1 root root    20 Apr  3  2018 libva-drm.so.2 -> libva-drm.so.2.100.0
-rw-r--r--  1 root root   11K Apr  3  2018 libva-drm.so.2.100.0
lrwxrwxrwx  1 root root    20 Apr  3  2018 libva-x11.so.2 -> libva-x11.so.2.100.0
-rw-r--r--  1 root root   23K Apr  3  2018 libva-x11.so.2.100.0
lrwxrwxrwx  1 root root    16 Apr  3  2018 libva.so.2 -> libva.so.2.100.0
-rw-r--r--  1 root root  131K Apr  3  2018 libva.so.2.100.0
lrwxrwxrwx  1 root root    17 Feb 19  2016 libvdpau.so.1 -> libvdpau.so.1.0.0
-rw-r--r--  1 root root   15K Feb 19  2016 libvdpau.so.1.0.0
lrwxrwxrwx  1 root root    18 Mar 16  2018 libvorbis.so.0 -> libvorbis.so.0.4.8
-rw-r--r--  1 root root  171K Mar 16  2018 libvorbis.so.0.4.8
lrwxrwxrwx  1 root root    22 Mar 16  2018 libvorbisenc.so.2 -> libvorbisenc.so.2.0.11
-rw-r--r--  1 root root  675K Mar 16  2018 libvorbisenc.so.2.0.11
lrwxrwxrwx  1 root root    22 Mar 16  2018 libvorbisfile.so.3 -> libvorbisfile.so.3.3.7
-rw-r--r--  1 root root   31K Mar 16  2018 libvorbisfile.so.3.3.7
lrwxrwxrwx  1 root root    15 Nov 19  2019 libvpx.so.5 -> libvpx.so.5.0.0
lrwxrwxrwx  1 root root    15 Nov 19  2019 libvpx.so.5.0 -> libvpx.so.5.0.0
-rw-r--r--  1 root root  2.3M Nov 19  2019 libvpx.so.5.0.0
lrwxrwxrwx  1 root root    29 Dec 17  2017 libvtkChartsCore-6.3.so.6.3 -> libvtkChartsCore-6.3.so.6.3.0
-rw-r--r--  1 root root  1.4M Dec 17  2017 libvtkChartsCore-6.3.so.6.3.0
lrwxrwxrwx  1 root root    38 Dec 17  2017 libvtkChartsCorePython27D-6.3.so.6.3 -> libvtkChartsCorePython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  939K Dec 17  2017 libvtkChartsCorePython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    32 Dec 17  2017 libvtkChartsCoreTCL-6.3.so.6.3 -> libvtkChartsCoreTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  573K Dec 17  2017 libvtkChartsCoreTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    30 Dec 17  2017 libvtkCommonColor-6.3.so.6.3 -> libvtkCommonColor-6.3.so.6.3.0
-rw-r--r--  1 root root  106K Dec 17  2017 libvtkCommonColor-6.3.so.6.3.0
lrwxrwxrwx  1 root root    39 Dec 17  2017 libvtkCommonColorPython27D-6.3.so.6.3 -> libvtkCommonColorPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   70K Dec 17  2017 libvtkCommonColorPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    33 Dec 17  2017 libvtkCommonColorTCL-6.3.so.6.3 -> libvtkCommonColorTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   27K Dec 17  2017 libvtkCommonColorTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    46 Dec 17  2017 libvtkCommonComputationalGeometry-6.3.so.6.3 -> libvtkCommonComputationalGeometry-6.3.so.6.3.0
-rw-r--r--  1 root root  240K Dec 17  2017 libvtkCommonComputationalGeometry-6.3.so.6.3.0
lrwxrwxrwx  1 root root    55 Dec 17  2017 libvtkCommonComputationalGeometryPython27D-6.3.so.6.3 -> libvtkCommonComputationalGeometryPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  423K Dec 17  2017 libvtkCommonComputationalGeometryPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    49 Dec 17  2017 libvtkCommonComputationalGeometryTCL-6.3.so.6.3 -> libvtkCommonComputationalGeometryTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  259K Dec 17  2017 libvtkCommonComputationalGeometryTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    29 Dec 17  2017 libvtkCommonCore-6.3.so.6.3 -> libvtkCommonCore-6.3.so.6.3.0
-rw-r--r--  1 root root  3.2M Dec 17  2017 libvtkCommonCore-6.3.so.6.3.0
lrwxrwxrwx  1 root root    38 Dec 17  2017 libvtkCommonCorePython27D-6.3.so.6.3 -> libvtkCommonCorePython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  3.0M Dec 17  2017 libvtkCommonCorePython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    32 Dec 17  2017 libvtkCommonCoreTCL-6.3.so.6.3 -> libvtkCommonCoreTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  900K Dec 17  2017 libvtkCommonCoreTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    34 Dec 17  2017 libvtkCommonDataModel-6.3.so.6.3 -> libvtkCommonDataModel-6.3.so.6.3.0
-rw-r--r--  1 root root  3.4M Dec 17  2017 libvtkCommonDataModel-6.3.so.6.3.0
lrwxrwxrwx  1 root root    43 Dec 17  2017 libvtkCommonDataModelPython27D-6.3.so.6.3 -> libvtkCommonDataModelPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  4.2M Dec 17  2017 libvtkCommonDataModelPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    37 Dec 17  2017 libvtkCommonDataModelTCL-6.3.so.6.3 -> libvtkCommonDataModelTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  2.2M Dec 17  2017 libvtkCommonDataModelTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    39 Dec 17  2017 libvtkCommonExecutionModel-6.3.so.6.3 -> libvtkCommonExecutionModel-6.3.so.6.3.0
-rw-r--r--  1 root root  669K Dec 17  2017 libvtkCommonExecutionModel-6.3.so.6.3.0
lrwxrwxrwx  1 root root    48 Dec 17  2017 libvtkCommonExecutionModelPython27D-6.3.so.6.3 -> libvtkCommonExecutionModelPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  621K Dec 17  2017 libvtkCommonExecutionModelPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    42 Dec 17  2017 libvtkCommonExecutionModelTCL-6.3.so.6.3 -> libvtkCommonExecutionModelTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  474K Dec 17  2017 libvtkCommonExecutionModelTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    29 Dec 17  2017 libvtkCommonMath-6.3.so.6.3 -> libvtkCommonMath-6.3.so.6.3.0
-rw-r--r--  1 root root  128K Dec 17  2017 libvtkCommonMath-6.3.so.6.3.0
lrwxrwxrwx  1 root root    38 Dec 17  2017 libvtkCommonMathPython27D-6.3.so.6.3 -> libvtkCommonMathPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  347K Dec 17  2017 libvtkCommonMathPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    32 Dec 17  2017 libvtkCommonMathTCL-6.3.so.6.3 -> libvtkCommonMathTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  103K Dec 17  2017 libvtkCommonMathTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    29 Dec 17  2017 libvtkCommonMisc-6.3.so.6.3 -> libvtkCommonMisc-6.3.so.6.3.0
-rw-r--r--  1 root root   88K Dec 17  2017 libvtkCommonMisc-6.3.so.6.3.0
lrwxrwxrwx  1 root root    38 Dec 17  2017 libvtkCommonMiscPython27D-6.3.so.6.3 -> libvtkCommonMiscPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   88K Dec 17  2017 libvtkCommonMiscPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    32 Dec 17  2017 libvtkCommonMiscTCL-6.3.so.6.3 -> libvtkCommonMiscTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   47K Dec 17  2017 libvtkCommonMiscTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    31 Dec 17  2017 libvtkCommonSystem-6.3.so.6.3 -> libvtkCommonSystem-6.3.so.6.3.0
-rw-r--r--  1 root root   76K Dec 17  2017 libvtkCommonSystem-6.3.so.6.3.0
lrwxrwxrwx  1 root root    40 Dec 17  2017 libvtkCommonSystemPython27D-6.3.so.6.3 -> libvtkCommonSystemPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   71K Dec 17  2017 libvtkCommonSystemPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    34 Dec 17  2017 libvtkCommonSystemTCL-6.3.so.6.3 -> libvtkCommonSystemTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   67K Dec 17  2017 libvtkCommonSystemTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    35 Dec 17  2017 libvtkCommonTransforms-6.3.so.6.3 -> libvtkCommonTransforms-6.3.so.6.3.0
-rw-r--r--  1 root root  184K Dec 17  2017 libvtkCommonTransforms-6.3.so.6.3.0
lrwxrwxrwx  1 root root    44 Dec 17  2017 libvtkCommonTransformsPython27D-6.3.so.6.3 -> libvtkCommonTransformsPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  263K Dec 17  2017 libvtkCommonTransformsPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    38 Dec 17  2017 libvtkCommonTransformsTCL-6.3.so.6.3 -> libvtkCommonTransformsTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  200K Dec 17  2017 libvtkCommonTransformsTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    30 Dec 17  2017 libvtkDICOMParser-6.3.so.6.3 -> libvtkDICOMParser-6.3.so.6.3.0
-rw-r--r--  1 root root   96K Dec 17  2017 libvtkDICOMParser-6.3.so.6.3.0
lrwxrwxrwx  1 root root    35 Dec 17  2017 libvtkDomainsChemistry-6.3.so.6.3 -> libvtkDomainsChemistry-6.3.so.6.3.0
-rw-r--r--  1 root root  293K Dec 17  2017 libvtkDomainsChemistry-6.3.so.6.3.0
lrwxrwxrwx  1 root root    44 Dec 17  2017 libvtkDomainsChemistryPython27D-6.3.so.6.3 -> libvtkDomainsChemistryPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  181K Dec 17  2017 libvtkDomainsChemistryPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    38 Dec 17  2017 libvtkDomainsChemistryTCL-6.3.so.6.3 -> libvtkDomainsChemistryTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  127K Dec 17  2017 libvtkDomainsChemistryTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    29 Dec 17  2017 libvtkFiltersAMR-6.3.so.6.3 -> libvtkFiltersAMR-6.3.so.6.3.0
-rw-r--r--  1 root root  273K Dec 17  2017 libvtkFiltersAMR-6.3.so.6.3.0
lrwxrwxrwx  1 root root    38 Dec 17  2017 libvtkFiltersAMRPython27D-6.3.so.6.3 -> libvtkFiltersAMRPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  183K Dec 17  2017 libvtkFiltersAMRPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    32 Dec 17  2017 libvtkFiltersAMRTCL-6.3.so.6.3 -> libvtkFiltersAMRTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  139K Dec 17  2017 libvtkFiltersAMRTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    30 Dec 17  2017 libvtkFiltersCore-6.3.so.6.3 -> libvtkFiltersCore-6.3.so.6.3.0
-rw-r--r--  1 root root  4.9M Dec 17  2017 libvtkFiltersCore-6.3.so.6.3.0
lrwxrwxrwx  1 root root    39 Dec 17  2017 libvtkFiltersCorePython27D-6.3.so.6.3 -> libvtkFiltersCorePython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  2.6M Dec 17  2017 libvtkFiltersCorePython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    33 Dec 17  2017 libvtkFiltersCoreTCL-6.3.so.6.3 -> libvtkFiltersCoreTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  1.5M Dec 17  2017 libvtkFiltersCoreTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    36 Dec 17  2017 libvtkFiltersExtraction-6.3.so.6.3 -> libvtkFiltersExtraction-6.3.so.6.3.0
-rw-r--r--  1 root root  1.1M Dec 17  2017 libvtkFiltersExtraction-6.3.so.6.3.0
lrwxrwxrwx  1 root root    45 Dec 17  2017 libvtkFiltersExtractionPython27D-6.3.so.6.3 -> libvtkFiltersExtractionPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  497K Dec 17  2017 libvtkFiltersExtractionPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    39 Dec 17  2017 libvtkFiltersExtractionTCL-6.3.so.6.3 -> libvtkFiltersExtractionTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  360K Dec 17  2017 libvtkFiltersExtractionTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    35 Dec 17  2017 libvtkFiltersFlowPaths-6.3.so.6.3 -> libvtkFiltersFlowPaths-6.3.so.6.3.0
-rw-r--r--  1 root root  473K Dec 17  2017 libvtkFiltersFlowPaths-6.3.so.6.3.0
lrwxrwxrwx  1 root root    44 Dec 17  2017 libvtkFiltersFlowPathsPython27D-6.3.so.6.3 -> libvtkFiltersFlowPathsPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  495K Dec 17  2017 libvtkFiltersFlowPathsPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    38 Dec 17  2017 libvtkFiltersFlowPathsTCL-6.3.so.6.3 -> libvtkFiltersFlowPathsTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  264K Dec 17  2017 libvtkFiltersFlowPathsTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    33 Dec 17  2017 libvtkFiltersGeneral-6.3.so.6.3 -> libvtkFiltersGeneral-6.3.so.6.3.0
-rw-r--r--  1 root root  3.2M Dec 17  2017 libvtkFiltersGeneral-6.3.so.6.3.0
lrwxrwxrwx  1 root root    42 Dec 17  2017 libvtkFiltersGeneralPython27D-6.3.so.6.3 -> libvtkFiltersGeneralPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  2.2M Dec 17  2017 libvtkFiltersGeneralPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    36 Dec 17  2017 libvtkFiltersGeneralTCL-6.3.so.6.3 -> libvtkFiltersGeneralTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  1.5M Dec 17  2017 libvtkFiltersGeneralTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    33 Dec 17  2017 libvtkFiltersGeneric-6.3.so.6.3 -> libvtkFiltersGeneric-6.3.so.6.3.0
-rw-r--r--  1 root root  296K Dec 17  2017 libvtkFiltersGeneric-6.3.so.6.3.0
lrwxrwxrwx  1 root root    42 Dec 17  2017 libvtkFiltersGenericPython27D-6.3.so.6.3 -> libvtkFiltersGenericPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  391K Dec 17  2017 libvtkFiltersGenericPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    36 Dec 17  2017 libvtkFiltersGenericTCL-6.3.so.6.3 -> libvtkFiltersGenericTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  220K Dec 17  2017 libvtkFiltersGenericTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    34 Dec 17  2017 libvtkFiltersGeometry-6.3.so.6.3 -> libvtkFiltersGeometry-6.3.so.6.3.0
-rw-r--r--  1 root root  546K Dec 17  2017 libvtkFiltersGeometry-6.3.so.6.3.0
lrwxrwxrwx  1 root root    43 Dec 17  2017 libvtkFiltersGeometryPython27D-6.3.so.6.3 -> libvtkFiltersGeometryPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  423K Dec 17  2017 libvtkFiltersGeometryPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    37 Dec 17  2017 libvtkFiltersGeometryTCL-6.3.so.6.3 -> libvtkFiltersGeometryTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  280K Dec 17  2017 libvtkFiltersGeometryTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    32 Dec 17  2017 libvtkFiltersHybrid-6.3.so.6.3 -> libvtkFiltersHybrid-6.3.so.6.3.0
-rw-r--r--  1 root root  867K Dec 17  2017 libvtkFiltersHybrid-6.3.so.6.3.0
lrwxrwxrwx  1 root root    41 Dec 17  2017 libvtkFiltersHybridPython27D-6.3.so.6.3 -> libvtkFiltersHybridPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  701K Dec 17  2017 libvtkFiltersHybridPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    35 Dec 17  2017 libvtkFiltersHybridTCL-6.3.so.6.3 -> libvtkFiltersHybridTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  432K Dec 17  2017 libvtkFiltersHybridTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    35 Dec 17  2017 libvtkFiltersHyperTree-6.3.so.6.3 -> libvtkFiltersHyperTree-6.3.so.6.3.0
-rw-r--r--  1 root root  297K Dec 17  2017 libvtkFiltersHyperTree-6.3.so.6.3.0
lrwxrwxrwx  1 root root    44 Dec 17  2017 libvtkFiltersHyperTreePython27D-6.3.so.6.3 -> libvtkFiltersHyperTreePython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  218K Dec 17  2017 libvtkFiltersHyperTreePython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    38 Dec 17  2017 libvtkFiltersHyperTreeTCL-6.3.so.6.3 -> libvtkFiltersHyperTreeTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  163K Dec 17  2017 libvtkFiltersHyperTreeTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    33 Dec 17  2017 libvtkFiltersImaging-6.3.so.6.3 -> libvtkFiltersImaging-6.3.so.6.3.0
-rw-r--r--  1 root root  136K Dec 17  2017 libvtkFiltersImaging-6.3.so.6.3.0
lrwxrwxrwx  1 root root    42 Dec 17  2017 libvtkFiltersImagingPython27D-6.3.so.6.3 -> libvtkFiltersImagingPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  111K Dec 17  2017 libvtkFiltersImagingPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    36 Dec 17  2017 libvtkFiltersImagingTCL-6.3.so.6.3 -> libvtkFiltersImagingTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   71K Dec 17  2017 libvtkFiltersImagingTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    34 Dec 17  2017 libvtkFiltersModeling-6.3.so.6.3 -> libvtkFiltersModeling-6.3.so.6.3.0
-rw-r--r--  1 root root  481K Dec 17  2017 libvtkFiltersModeling-6.3.so.6.3.0
lrwxrwxrwx  1 root root    43 Dec 17  2017 libvtkFiltersModelingPython27D-6.3.so.6.3 -> libvtkFiltersModelingPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  607K Dec 17  2017 libvtkFiltersModelingPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    37 Dec 17  2017 libvtkFiltersModelingTCL-6.3.so.6.3 -> libvtkFiltersModelingTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  388K Dec 17  2017 libvtkFiltersModelingTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    34 Dec 17  2017 libvtkFiltersParallel-6.3.so.6.3 -> libvtkFiltersParallel-6.3.so.6.3.0
-rw-r--r--  1 root root  860K Dec 17  2017 libvtkFiltersParallel-6.3.so.6.3.0
lrwxrwxrwx  1 root root    43 Dec 17  2017 libvtkFiltersParallelFlowPaths-6.3.so.6.3 -> libvtkFiltersParallelFlowPaths-6.3.so.6.3.0
-rw-r--r--  1 root root  232K Dec 17  2017 libvtkFiltersParallelFlowPaths-6.3.so.6.3.0
lrwxrwxrwx  1 root root    52 Dec 17  2017 libvtkFiltersParallelFlowPathsPython27D-6.3.so.6.3 -> libvtkFiltersParallelFlowPathsPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   40K Dec 17  2017 libvtkFiltersParallelFlowPathsPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    46 Dec 17  2017 libvtkFiltersParallelFlowPathsTCL-6.3.so.6.3 -> libvtkFiltersParallelFlowPathsTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   43K Dec 17  2017 libvtkFiltersParallelFlowPathsTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    42 Dec 17  2017 libvtkFiltersParallelGeometry-6.3.so.6.3 -> libvtkFiltersParallelGeometry-6.3.so.6.3.0
-rw-r--r--  1 root root  265K Dec 17  2017 libvtkFiltersParallelGeometry-6.3.so.6.3.0
lrwxrwxrwx  1 root root    51 Dec 17  2017 libvtkFiltersParallelGeometryPython27D-6.3.so.6.3 -> libvtkFiltersParallelGeometryPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   65K Dec 17  2017 libvtkFiltersParallelGeometryPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    45 Dec 17  2017 libvtkFiltersParallelGeometryTCL-6.3.so.6.3 -> libvtkFiltersParallelGeometryTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   55K Dec 17  2017 libvtkFiltersParallelGeometryTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    41 Dec 17  2017 libvtkFiltersParallelImaging-6.3.so.6.3 -> libvtkFiltersParallelImaging-6.3.so.6.3.0
-rw-r--r--  1 root root  140K Dec 17  2017 libvtkFiltersParallelImaging-6.3.so.6.3.0
lrwxrwxrwx  1 root root    50 Dec 17  2017 libvtkFiltersParallelImagingPython27D-6.3.so.6.3 -> libvtkFiltersParallelImagingPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   45K Dec 17  2017 libvtkFiltersParallelImagingPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    44 Dec 17  2017 libvtkFiltersParallelImagingTCL-6.3.so.6.3 -> libvtkFiltersParallelImagingTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   47K Dec 17  2017 libvtkFiltersParallelImagingTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    37 Dec 17  2017 libvtkFiltersParallelMPI-6.3.so.6.3 -> libvtkFiltersParallelMPI-6.3.so.6.3.0
-rw-r--r--  1 root root  237K Dec 17  2017 libvtkFiltersParallelMPI-6.3.so.6.3.0
lrwxrwxrwx  1 root root    46 Dec 17  2017 libvtkFiltersParallelMPIPython27D-6.3.so.6.3 -> libvtkFiltersParallelMPIPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   94K Dec 17  2017 libvtkFiltersParallelMPIPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    40 Dec 17  2017 libvtkFiltersParallelMPITCL-6.3.so.6.3 -> libvtkFiltersParallelMPITCL-6.3.so.6.3.0
-rw-r--r--  1 root root   67K Dec 17  2017 libvtkFiltersParallelMPITCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    43 Dec 17  2017 libvtkFiltersParallelPython27D-6.3.so.6.3 -> libvtkFiltersParallelPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  484K Dec 17  2017 libvtkFiltersParallelPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    44 Dec 17  2017 libvtkFiltersParallelStatistics-6.3.so.6.3 -> libvtkFiltersParallelStatistics-6.3.so.6.3.0
-rw-r--r--  1 root root  220K Dec 17  2017 libvtkFiltersParallelStatistics-6.3.so.6.3.0
lrwxrwxrwx  1 root root    53 Dec 17  2017 libvtkFiltersParallelStatisticsPython27D-6.3.so.6.3 -> libvtkFiltersParallelStatisticsPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   74K Dec 17  2017 libvtkFiltersParallelStatisticsPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    47 Dec 17  2017 libvtkFiltersParallelStatisticsTCL-6.3.so.6.3 -> libvtkFiltersParallelStatisticsTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   71K Dec 17  2017 libvtkFiltersParallelStatisticsTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    37 Dec 17  2017 libvtkFiltersParallelTCL-6.3.so.6.3 -> libvtkFiltersParallelTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  380K Dec 17  2017 libvtkFiltersParallelTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    38 Dec 17  2017 libvtkFiltersProgrammable-6.3.so.6.3 -> libvtkFiltersProgrammable-6.3.so.6.3.0
-rw-r--r--  1 root root   63K Dec 17  2017 libvtkFiltersProgrammable-6.3.so.6.3.0
lrwxrwxrwx  1 root root    47 Dec 17  2017 libvtkFiltersProgrammablePython27D-6.3.so.6.3 -> libvtkFiltersProgrammablePython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   61K Dec 17  2017 libvtkFiltersProgrammablePython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    41 Dec 17  2017 libvtkFiltersProgrammableTCL-6.3.so.6.3 -> libvtkFiltersProgrammableTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   47K Dec 17  2017 libvtkFiltersProgrammableTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    32 Dec 17  2017 libvtkFiltersPython-6.3.so.6.3 -> libvtkFiltersPython-6.3.so.6.3.0
-rw-r--r--  1 root root   35K Dec 17  2017 libvtkFiltersPython-6.3.so.6.3.0
lrwxrwxrwx  1 root root    41 Dec 17  2017 libvtkFiltersPythonPython27D-6.3.so.6.3 -> libvtkFiltersPythonPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   15K Dec 17  2017 libvtkFiltersPythonPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    35 Dec 17  2017 libvtkFiltersReebGraph-6.3.so.6.3 -> libvtkFiltersReebGraph-6.3.so.6.3.0
-rw-r--r--  1 root root  108K Dec 17  2017 libvtkFiltersReebGraph-6.3.so.6.3.0
lrwxrwxrwx  1 root root    44 Dec 17  2017 libvtkFiltersReebGraphPython27D-6.3.so.6.3 -> libvtkFiltersReebGraphPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   73K Dec 17  2017 libvtkFiltersReebGraphPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    38 Dec 17  2017 libvtkFiltersReebGraphTCL-6.3.so.6.3 -> libvtkFiltersReebGraphTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   63K Dec 17  2017 libvtkFiltersReebGraphTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    29 Dec 17  2017 libvtkFiltersSMP-6.3.so.6.3 -> libvtkFiltersSMP-6.3.so.6.3.0
-rw-r--r--  1 root root  407K Dec 17  2017 libvtkFiltersSMP-6.3.so.6.3.0
lrwxrwxrwx  1 root root    38 Dec 17  2017 libvtkFiltersSMPPython27D-6.3.so.6.3 -> libvtkFiltersSMPPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  110K Dec 17  2017 libvtkFiltersSMPPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    32 Dec 17  2017 libvtkFiltersSMPTCL-6.3.so.6.3 -> libvtkFiltersSMPTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   87K Dec 17  2017 libvtkFiltersSMPTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    35 Dec 17  2017 libvtkFiltersSelection-6.3.so.6.3 -> libvtkFiltersSelection-6.3.so.6.3.0
-rw-r--r--  1 root root  100K Dec 17  2017 libvtkFiltersSelection-6.3.so.6.3.0
lrwxrwxrwx  1 root root    44 Dec 17  2017 libvtkFiltersSelectionPython27D-6.3.so.6.3 -> libvtkFiltersSelectionPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  109K Dec 17  2017 libvtkFiltersSelectionPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    38 Dec 17  2017 libvtkFiltersSelectionTCL-6.3.so.6.3 -> libvtkFiltersSelectionTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   83K Dec 17  2017 libvtkFiltersSelectionTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    33 Dec 17  2017 libvtkFiltersSources-6.3.so.6.3 -> libvtkFiltersSources-6.3.so.6.3.0
-rw-r--r--  1 root root  773K Dec 17  2017 libvtkFiltersSources-6.3.so.6.3.0
lrwxrwxrwx  1 root root    42 Dec 17  2017 libvtkFiltersSourcesPython27D-6.3.so.6.3 -> libvtkFiltersSourcesPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  1.2M Dec 17  2017 libvtkFiltersSourcesPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    36 Dec 17  2017 libvtkFiltersSourcesTCL-6.3.so.6.3 -> libvtkFiltersSourcesTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  760K Dec 17  2017 libvtkFiltersSourcesTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    36 Dec 17  2017 libvtkFiltersStatistics-6.3.so.6.3 -> libvtkFiltersStatistics-6.3.so.6.3.0
-rw-r--r--  1 root root  549K Dec 17  2017 libvtkFiltersStatistics-6.3.so.6.3.0
lrwxrwxrwx  1 root root    45 Dec 17  2017 libvtkFiltersStatisticsPython27D-6.3.so.6.3 -> libvtkFiltersStatisticsPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  341K Dec 17  2017 libvtkFiltersStatisticsPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    39 Dec 17  2017 libvtkFiltersStatisticsTCL-6.3.so.6.3 -> libvtkFiltersStatisticsTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  224K Dec 17  2017 libvtkFiltersStatisticsTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    33 Dec 17  2017 libvtkFiltersTexture-6.3.so.6.3 -> libvtkFiltersTexture-6.3.so.6.3.0
-rw-r--r--  1 root root  151K Dec 17  2017 libvtkFiltersTexture-6.3.so.6.3.0
lrwxrwxrwx  1 root root    42 Dec 17  2017 libvtkFiltersTexturePython27D-6.3.so.6.3 -> libvtkFiltersTexturePython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  219K Dec 17  2017 libvtkFiltersTexturePython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    36 Dec 17  2017 libvtkFiltersTextureTCL-6.3.so.6.3 -> libvtkFiltersTextureTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  167K Dec 17  2017 libvtkFiltersTextureTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    33 Dec 17  2017 libvtkFiltersVerdict-6.3.so.6.3 -> libvtkFiltersVerdict-6.3.so.6.3.0
-rw-r--r--  1 root root  124K Dec 17  2017 libvtkFiltersVerdict-6.3.so.6.3.0
lrwxrwxrwx  1 root root    42 Dec 17  2017 libvtkFiltersVerdictPython27D-6.3.so.6.3 -> libvtkFiltersVerdictPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  267K Dec 17  2017 libvtkFiltersVerdictPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    36 Dec 17  2017 libvtkFiltersVerdictTCL-6.3.so.6.3 -> libvtkFiltersVerdictTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  156K Dec 17  2017 libvtkFiltersVerdictTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    29 Dec 17  2017 libvtkGeovisCore-6.3.so.6.3 -> libvtkGeovisCore-6.3.so.6.3.0
-rw-r--r--  1 root root  707K Dec 17  2017 libvtkGeovisCore-6.3.so.6.3.0
lrwxrwxrwx  1 root root    38 Dec 17  2017 libvtkGeovisCorePython27D-6.3.so.6.3 -> libvtkGeovisCorePython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  555K Dec 17  2017 libvtkGeovisCorePython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    32 Dec 17  2017 libvtkGeovisCoreTCL-6.3.so.6.3 -> libvtkGeovisCoreTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  392K Dec 17  2017 libvtkGeovisCoreTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    24 Dec 17  2017 libvtkIOAMR-6.3.so.6.3 -> libvtkIOAMR-6.3.so.6.3.0
-rw-r--r--  1 root root  282K Dec 17  2017 libvtkIOAMR-6.3.so.6.3.0
lrwxrwxrwx  1 root root    33 Dec 17  2017 libvtkIOAMRPython27D-6.3.so.6.3 -> libvtkIOAMRPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  107K Dec 17  2017 libvtkIOAMRPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    27 Dec 17  2017 libvtkIOAMRTCL-6.3.so.6.3 -> libvtkIOAMRTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   91K Dec 17  2017 libvtkIOAMRTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    25 Dec 17  2017 libvtkIOCore-6.3.so.6.3 -> libvtkIOCore-6.3.so.6.3.0
-rw-r--r--  1 root root  458K Dec 17  2017 libvtkIOCore-6.3.so.6.3.0
lrwxrwxrwx  1 root root    34 Dec 17  2017 libvtkIOCorePython27D-6.3.so.6.3 -> libvtkIOCorePython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  322K Dec 17  2017 libvtkIOCorePython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    28 Dec 17  2017 libvtkIOCoreTCL-6.3.so.6.3 -> libvtkIOCoreTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  240K Dec 17  2017 libvtkIOCoreTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    28 Dec 17  2017 libvtkIOEnSight-6.3.so.6.3 -> libvtkIOEnSight-6.3.so.6.3.0
-rw-r--r--  1 root root  549K Dec 17  2017 libvtkIOEnSight-6.3.so.6.3.0
lrwxrwxrwx  1 root root    37 Dec 17  2017 libvtkIOEnSightPython27D-6.3.so.6.3 -> libvtkIOEnSightPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  141K Dec 17  2017 libvtkIOEnSightPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    31 Dec 17  2017 libvtkIOEnSightTCL-6.3.so.6.3 -> libvtkIOEnSightTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   95K Dec 17  2017 libvtkIOEnSightTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    27 Dec 17  2017 libvtkIOExodus-6.3.so.6.3 -> libvtkIOExodus-6.3.so.6.3.0
-rw-r--r--  1 root root  660K Dec 17  2017 libvtkIOExodus-6.3.so.6.3.0
lrwxrwxrwx  1 root root    36 Dec 17  2017 libvtkIOExodusPython27D-6.3.so.6.3 -> libvtkIOExodusPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  385K Dec 17  2017 libvtkIOExodusPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    30 Dec 17  2017 libvtkIOExodusTCL-6.3.so.6.3 -> libvtkIOExodusTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  220K Dec 17  2017 libvtkIOExodusTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    27 Dec 17  2017 libvtkIOExport-6.3.so.6.3 -> libvtkIOExport-6.3.so.6.3.0
-rw-r--r--  1 root root  513K Dec 17  2017 libvtkIOExport-6.3.so.6.3.0
lrwxrwxrwx  1 root root    36 Dec 17  2017 libvtkIOExportPython27D-6.3.so.6.3 -> libvtkIOExportPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  288K Dec 17  2017 libvtkIOExportPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    30 Dec 17  2017 libvtkIOExportTCL-6.3.so.6.3 -> libvtkIOExportTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  183K Dec 17  2017 libvtkIOExportTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    27 Dec 17  2017 libvtkIOFFMPEG-6.3.so.6.3 -> libvtkIOFFMPEG-6.3.so.6.3.0
-rw-r--r--  1 root root   55K Dec 17  2017 libvtkIOFFMPEG-6.3.so.6.3.0
lrwxrwxrwx  1 root root    36 Dec 17  2017 libvtkIOFFMPEGPython27D-6.3.so.6.3 -> libvtkIOFFMPEGPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   44K Dec 17  2017 libvtkIOFFMPEGPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    30 Dec 17  2017 libvtkIOFFMPEGTCL-6.3.so.6.3 -> libvtkIOFFMPEGTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   31K Dec 17  2017 libvtkIOFFMPEGTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    25 Dec 17  2017 libvtkIOGDAL-6.3.so.6.3 -> libvtkIOGDAL-6.3.so.6.3.0
-rw-r--r--  1 root root  152K Dec 17  2017 libvtkIOGDAL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    34 Dec 17  2017 libvtkIOGDALPython27D-6.3.so.6.3 -> libvtkIOGDALPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   60K Dec 17  2017 libvtkIOGDALPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    28 Dec 17  2017 libvtkIOGDALTCL-6.3.so.6.3 -> libvtkIOGDALTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   47K Dec 17  2017 libvtkIOGDALTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    28 Dec 17  2017 libvtkIOGeoJSON-6.3.so.6.3 -> libvtkIOGeoJSON-6.3.so.6.3.0
-rw-r--r--  1 root root   63K Dec 17  2017 libvtkIOGeoJSON-6.3.so.6.3.0
lrwxrwxrwx  1 root root    37 Dec 17  2017 libvtkIOGeoJSONPython27D-6.3.so.6.3 -> libvtkIOGeoJSONPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   40K Dec 17  2017 libvtkIOGeoJSONPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    31 Dec 17  2017 libvtkIOGeoJSONTCL-6.3.so.6.3 -> libvtkIOGeoJSONTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   35K Dec 17  2017 libvtkIOGeoJSONTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    29 Dec 17  2017 libvtkIOGeometry-6.3.so.6.3 -> libvtkIOGeometry-6.3.so.6.3.0
-rw-r--r--  1 root root  1.1M Dec 17  2017 libvtkIOGeometry-6.3.so.6.3.0
lrwxrwxrwx  1 root root    38 Dec 17  2017 libvtkIOGeometryPython27D-6.3.so.6.3 -> libvtkIOGeometryPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  621K Dec 17  2017 libvtkIOGeometryPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    32 Dec 17  2017 libvtkIOGeometryTCL-6.3.so.6.3 -> libvtkIOGeometryTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  400K Dec 17  2017 libvtkIOGeometryTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    26 Dec 17  2017 libvtkIOImage-6.3.so.6.3 -> libvtkIOImage-6.3.so.6.3.0
-rw-r--r--  1 root root  1.6M Dec 17  2017 libvtkIOImage-6.3.so.6.3.0
lrwxrwxrwx  1 root root    35 Dec 17  2017 libvtkIOImagePython27D-6.3.so.6.3 -> libvtkIOImagePython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  900K Dec 17  2017 libvtkIOImagePython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    29 Dec 17  2017 libvtkIOImageTCL-6.3.so.6.3 -> libvtkIOImageTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  540K Dec 17  2017 libvtkIOImageTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    27 Dec 17  2017 libvtkIOImport-6.3.so.6.3 -> libvtkIOImport-6.3.so.6.3.0
-rw-r--r--  1 root root  276K Dec 17  2017 libvtkIOImport-6.3.so.6.3.0
lrwxrwxrwx  1 root root    36 Dec 17  2017 libvtkIOImportPython27D-6.3.so.6.3 -> libvtkIOImportPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   57K Dec 17  2017 libvtkIOImportPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    30 Dec 17  2017 libvtkIOImportTCL-6.3.so.6.3 -> libvtkIOImportTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   47K Dec 17  2017 libvtkIOImportTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    28 Dec 17  2017 libvtkIOInfovis-6.3.so.6.3 -> libvtkIOInfovis-6.3.so.6.3.0
-rw-r--r--  1 root root  562K Dec 17  2017 libvtkIOInfovis-6.3.so.6.3.0
lrwxrwxrwx  1 root root    37 Dec 17  2017 libvtkIOInfovisPython27D-6.3.so.6.3 -> libvtkIOInfovisPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  320K Dec 17  2017 libvtkIOInfovisPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    31 Dec 17  2017 libvtkIOInfovisTCL-6.3.so.6.3 -> libvtkIOInfovisTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  212K Dec 17  2017 libvtkIOInfovisTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    27 Dec 17  2017 libvtkIOLSDyna-6.3.so.6.3 -> libvtkIOLSDyna-6.3.so.6.3.0
-rw-r--r--  1 root root  313K Dec 17  2017 libvtkIOLSDyna-6.3.so.6.3.0
lrwxrwxrwx  1 root root    36 Dec 17  2017 libvtkIOLSDynaPython27D-6.3.so.6.3 -> libvtkIOLSDynaPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  130K Dec 17  2017 libvtkIOLSDynaPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    30 Dec 17  2017 libvtkIOLSDynaTCL-6.3.so.6.3 -> libvtkIOLSDynaTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   71K Dec 17  2017 libvtkIOLSDynaTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    27 Dec 17  2017 libvtkIOLegacy-6.3.so.6.3 -> libvtkIOLegacy-6.3.so.6.3.0
-rw-r--r--  1 root root  655K Dec 17  2017 libvtkIOLegacy-6.3.so.6.3.0
lrwxrwxrwx  1 root root    36 Dec 17  2017 libvtkIOLegacyPython27D-6.3.so.6.3 -> libvtkIOLegacyPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  354K Dec 17  2017 libvtkIOLegacyPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    30 Dec 17  2017 libvtkIOLegacyTCL-6.3.so.6.3 -> libvtkIOLegacyTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  240K Dec 17  2017 libvtkIOLegacyTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    25 Dec 17  2017 libvtkIOMINC-6.3.so.6.3 -> libvtkIOMINC-6.3.so.6.3.0
-rw-r--r--  1 root root  450K Dec 17  2017 libvtkIOMINC-6.3.so.6.3.0
lrwxrwxrwx  1 root root    34 Dec 17  2017 libvtkIOMINCPython27D-6.3.so.6.3 -> libvtkIOMINCPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  214K Dec 17  2017 libvtkIOMINCPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    28 Dec 17  2017 libvtkIOMINCTCL-6.3.so.6.3 -> libvtkIOMINCTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  151K Dec 17  2017 libvtkIOMINCTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    29 Dec 17  2017 libvtkIOMPIImage-6.3.so.6.3 -> libvtkIOMPIImage-6.3.so.6.3.0
-rw-r--r--  1 root root  161K Dec 17  2017 libvtkIOMPIImage-6.3.so.6.3.0
lrwxrwxrwx  1 root root    38 Dec 17  2017 libvtkIOMPIImagePython27D-6.3.so.6.3 -> libvtkIOMPIImagePython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   28K Dec 17  2017 libvtkIOMPIImagePython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    32 Dec 17  2017 libvtkIOMPIImageTCL-6.3.so.6.3 -> libvtkIOMPIImageTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   23K Dec 17  2017 libvtkIOMPIImageTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    32 Dec 17  2017 libvtkIOMPIParallel-6.3.so.6.3 -> libvtkIOMPIParallel-6.3.so.6.3.0
-rw-r--r--  1 root root  217K Dec 17  2017 libvtkIOMPIParallel-6.3.so.6.3.0
lrwxrwxrwx  1 root root    41 Dec 17  2017 libvtkIOMPIParallelPython27D-6.3.so.6.3 -> libvtkIOMPIParallelPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   27K Dec 17  2017 libvtkIOMPIParallelPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    35 Dec 17  2017 libvtkIOMPIParallelTCL-6.3.so.6.3 -> libvtkIOMPIParallelTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   27K Dec 17  2017 libvtkIOMPIParallelTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    26 Dec 17  2017 libvtkIOMovie-6.3.so.6.3 -> libvtkIOMovie-6.3.so.6.3.0
-rw-r--r--  1 root root   59K Dec 17  2017 libvtkIOMovie-6.3.so.6.3.0
lrwxrwxrwx  1 root root    35 Dec 17  2017 libvtkIOMoviePython27D-6.3.so.6.3 -> libvtkIOMoviePython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   49K Dec 17  2017 libvtkIOMoviePython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    29 Dec 17  2017 libvtkIOMovieTCL-6.3.so.6.3 -> libvtkIOMovieTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   39K Dec 17  2017 libvtkIOMovieTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    26 Dec 17  2017 libvtkIOMySQL-6.3.so.6.3 -> libvtkIOMySQL-6.3.so.6.3.0
-rw-r--r--  1 root root  144K Dec 17  2017 libvtkIOMySQL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    35 Dec 17  2017 libvtkIOMySQLPython27D-6.3.so.6.3 -> libvtkIOMySQLPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   86K Dec 17  2017 libvtkIOMySQLPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    29 Dec 17  2017 libvtkIOMySQLTCL-6.3.so.6.3 -> libvtkIOMySQLTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   63K Dec 17  2017 libvtkIOMySQLTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    27 Dec 17  2017 libvtkIONetCDF-6.3.so.6.3 -> libvtkIONetCDF-6.3.so.6.3.0
-rw-r--r--  1 root root  410K Dec 17  2017 libvtkIONetCDF-6.3.so.6.3.0
lrwxrwxrwx  1 root root    36 Dec 17  2017 libvtkIONetCDFPython27D-6.3.so.6.3 -> libvtkIONetCDFPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  201K Dec 17  2017 libvtkIONetCDFPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    30 Dec 17  2017 libvtkIONetCDFTCL-6.3.so.6.3 -> libvtkIONetCDFTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  139K Dec 17  2017 libvtkIONetCDFTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    25 Dec 17  2017 libvtkIOODBC-6.3.so.6.3 -> libvtkIOODBC-6.3.so.6.3.0
-rw-r--r--  1 root root  156K Dec 17  2017 libvtkIOODBC-6.3.so.6.3.0
lrwxrwxrwx  1 root root    34 Dec 17  2017 libvtkIOODBCPython27D-6.3.so.6.3 -> libvtkIOODBCPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   65K Dec 17  2017 libvtkIOODBCPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    28 Dec 17  2017 libvtkIOODBCTCL-6.3.so.6.3 -> libvtkIOODBCTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   47K Dec 17  2017 libvtkIOODBCTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    24 Dec 17  2017 libvtkIOPLY-6.3.so.6.3 -> libvtkIOPLY-6.3.so.6.3.0
-rw-r--r--  1 root root   97K Dec 17  2017 libvtkIOPLY-6.3.so.6.3.0
lrwxrwxrwx  1 root root    33 Dec 17  2017 libvtkIOPLYPython27D-6.3.so.6.3 -> libvtkIOPLYPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   73K Dec 17  2017 libvtkIOPLYPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    27 Dec 17  2017 libvtkIOPLYTCL-6.3.so.6.3 -> libvtkIOPLYTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   43K Dec 17  2017 libvtkIOPLYTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    29 Dec 17  2017 libvtkIOParallel-6.3.so.6.3 -> libvtkIOParallel-6.3.so.6.3.0
-rw-r--r--  1 root root  526K Dec 17  2017 libvtkIOParallel-6.3.so.6.3.0
lrwxrwxrwx  1 root root    35 Dec 17  2017 libvtkIOParallelExodus-6.3.so.6.3 -> libvtkIOParallelExodus-6.3.so.6.3.0
-rw-r--r--  1 root root  148K Dec 17  2017 libvtkIOParallelExodus-6.3.so.6.3.0
lrwxrwxrwx  1 root root    44 Dec 17  2017 libvtkIOParallelExodusPython27D-6.3.so.6.3 -> libvtkIOParallelExodusPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   48K Dec 17  2017 libvtkIOParallelExodusPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    38 Dec 17  2017 libvtkIOParallelExodusTCL-6.3.so.6.3 -> libvtkIOParallelExodusTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   35K Dec 17  2017 libvtkIOParallelExodusTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    35 Dec 17  2017 libvtkIOParallelLSDyna-6.3.so.6.3 -> libvtkIOParallelLSDyna-6.3.so.6.3.0
-rw-r--r--  1 root root   55K Dec 17  2017 libvtkIOParallelLSDyna-6.3.so.6.3.0
lrwxrwxrwx  1 root root    44 Dec 17  2017 libvtkIOParallelLSDynaPython27D-6.3.so.6.3 -> libvtkIOParallelLSDynaPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   19K Dec 17  2017 libvtkIOParallelLSDynaPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    38 Dec 17  2017 libvtkIOParallelLSDynaTCL-6.3.so.6.3 -> libvtkIOParallelLSDynaTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   19K Dec 17  2017 libvtkIOParallelLSDynaTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    35 Dec 17  2017 libvtkIOParallelNetCDF-6.3.so.6.3 -> libvtkIOParallelNetCDF-6.3.so.6.3.0
-rw-r--r--  1 root root  113K Dec 17  2017 libvtkIOParallelNetCDF-6.3.so.6.3.0
lrwxrwxrwx  1 root root    44 Dec 17  2017 libvtkIOParallelNetCDFPython27D-6.3.so.6.3 -> libvtkIOParallelNetCDFPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   32K Dec 17  2017 libvtkIOParallelNetCDFPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    38 Dec 17  2017 libvtkIOParallelNetCDFTCL-6.3.so.6.3 -> libvtkIOParallelNetCDFTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   27K Dec 17  2017 libvtkIOParallelNetCDFTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    38 Dec 17  2017 libvtkIOParallelPython27D-6.3.so.6.3 -> libvtkIOParallelPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  221K Dec 17  2017 libvtkIOParallelPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    32 Dec 17  2017 libvtkIOParallelTCL-6.3.so.6.3 -> libvtkIOParallelTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  139K Dec 17  2017 libvtkIOParallelTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    32 Dec 17  2017 libvtkIOParallelXML-6.3.so.6.3 -> libvtkIOParallelXML-6.3.so.6.3.0
-rw-r--r--  1 root root  204K Dec 17  2017 libvtkIOParallelXML-6.3.so.6.3.0
lrwxrwxrwx  1 root root    41 Dec 17  2017 libvtkIOParallelXMLPython27D-6.3.so.6.3 -> libvtkIOParallelXMLPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   95K Dec 17  2017 libvtkIOParallelXMLPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    35 Dec 17  2017 libvtkIOParallelXMLTCL-6.3.so.6.3 -> libvtkIOParallelXMLTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   87K Dec 17  2017 libvtkIOParallelXMLTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    31 Dec 17  2017 libvtkIOPostgreSQL-6.3.so.6.3 -> libvtkIOPostgreSQL-6.3.so.6.3.0
-rw-r--r--  1 root root  160K Dec 17  2017 libvtkIOPostgreSQL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    40 Dec 17  2017 libvtkIOPostgreSQLPython27D-6.3.so.6.3 -> libvtkIOPostgreSQLPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   73K Dec 17  2017 libvtkIOPostgreSQLPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    34 Dec 17  2017 libvtkIOPostgreSQLTCL-6.3.so.6.3 -> libvtkIOPostgreSQLTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   59K Dec 17  2017 libvtkIOPostgreSQLTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    24 Dec 17  2017 libvtkIOSQL-6.3.so.6.3 -> libvtkIOSQL-6.3.so.6.3.0
-rw-r--r--  1 root root  277K Dec 17  2017 libvtkIOSQL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    33 Dec 17  2017 libvtkIOSQLPython27D-6.3.so.6.3 -> libvtkIOSQLPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  202K Dec 17  2017 libvtkIOSQLPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    27 Dec 17  2017 libvtkIOSQLTCL-6.3.so.6.3 -> libvtkIOSQLTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  148K Dec 17  2017 libvtkIOSQLTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    25 Dec 17  2017 libvtkIOVPIC-6.3.so.6.3 -> libvtkIOVPIC-6.3.so.6.3.0
-rw-r--r--  1 root root   63K Dec 17  2017 libvtkIOVPIC-6.3.so.6.3.0
lrwxrwxrwx  1 root root    34 Dec 17  2017 libvtkIOVPICPython27D-6.3.so.6.3 -> libvtkIOVPICPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   44K Dec 17  2017 libvtkIOVPICPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    28 Dec 17  2017 libvtkIOVPICTCL-6.3.so.6.3 -> libvtkIOVPICTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   39K Dec 17  2017 libvtkIOVPICTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    26 Dec 17  2017 libvtkIOVideo-6.3.so.6.3 -> libvtkIOVideo-6.3.so.6.3.0
-rw-r--r--  1 root root   71K Dec 17  2017 libvtkIOVideo-6.3.so.6.3.0
lrwxrwxrwx  1 root root    35 Dec 17  2017 libvtkIOVideoPython27D-6.3.so.6.3 -> libvtkIOVideoPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   85K Dec 17  2017 libvtkIOVideoPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    29 Dec 17  2017 libvtkIOVideoTCL-6.3.so.6.3 -> libvtkIOVideoTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   43K Dec 17  2017 libvtkIOVideoTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    24 Dec 17  2017 libvtkIOXML-6.3.so.6.3 -> libvtkIOXML-6.3.so.6.3.0
-rw-r--r--  1 root root  945K Dec 17  2017 libvtkIOXML-6.3.so.6.3.0
lrwxrwxrwx  1 root root    30 Dec 17  2017 libvtkIOXMLParser-6.3.so.6.3 -> libvtkIOXMLParser-6.3.so.6.3.0
-rw-r--r--  1 root root  112K Dec 17  2017 libvtkIOXMLParser-6.3.so.6.3.0
lrwxrwxrwx  1 root root    39 Dec 17  2017 libvtkIOXMLParserPython27D-6.3.so.6.3 -> libvtkIOXMLParserPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   69K Dec 17  2017 libvtkIOXMLParserPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    33 Dec 17  2017 libvtkIOXMLParserTCL-6.3.so.6.3 -> libvtkIOXMLParserTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   51K Dec 17  2017 libvtkIOXMLParserTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    33 Dec 17  2017 libvtkIOXMLPython27D-6.3.so.6.3 -> libvtkIOXMLPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  375K Dec 17  2017 libvtkIOXMLPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    27 Dec 17  2017 libvtkIOXMLTCL-6.3.so.6.3 -> libvtkIOXMLTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  304K Dec 17  2017 libvtkIOXMLTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    26 Dec 17  2017 libvtkIOXdmf2-6.3.so.6.3 -> libvtkIOXdmf2-6.3.so.6.3.0
-rw-r--r--  1 root root  335K Dec 17  2017 libvtkIOXdmf2-6.3.so.6.3.0
lrwxrwxrwx  1 root root    35 Dec 17  2017 libvtkIOXdmf2Python27D-6.3.so.6.3 -> libvtkIOXdmf2Python27D-6.3.so.6.3.0
-rw-r--r--  1 root root   93K Dec 17  2017 libvtkIOXdmf2Python27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    30 Dec 17  2017 libvtkIOXdmfIITCL-6.3.so.6.3 -> libvtkIOXdmfIITCL-6.3.so.6.3.0
-rw-r--r--  1 root root   67K Dec 17  2017 libvtkIOXdmfIITCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    31 Dec 17  2017 libvtkImagingColor-6.3.so.6.3 -> libvtkImagingColor-6.3.so.6.3.0
-rw-r--r--  1 root root  280K Dec 17  2017 libvtkImagingColor-6.3.so.6.3.0
lrwxrwxrwx  1 root root    40 Dec 17  2017 libvtkImagingColorPython27D-6.3.so.6.3 -> libvtkImagingColorPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   85K Dec 17  2017 libvtkImagingColorPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    34 Dec 17  2017 libvtkImagingColorTCL-6.3.so.6.3 -> libvtkImagingColorTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   75K Dec 17  2017 libvtkImagingColorTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    30 Dec 17  2017 libvtkImagingCore-6.3.so.6.3 -> libvtkImagingCore-6.3.so.6.3.0
-rw-r--r--  1 root root  1.9M Dec 17  2017 libvtkImagingCore-6.3.so.6.3.0
lrwxrwxrwx  1 root root    39 Dec 17  2017 libvtkImagingCorePython27D-6.3.so.6.3 -> libvtkImagingCorePython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  934K Dec 17  2017 libvtkImagingCorePython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    33 Dec 17  2017 libvtkImagingCoreTCL-6.3.so.6.3 -> libvtkImagingCoreTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  625K Dec 17  2017 libvtkImagingCoreTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    33 Dec 17  2017 libvtkImagingFourier-6.3.so.6.3 -> libvtkImagingFourier-6.3.so.6.3.0
-rw-r--r--  1 root root  187K Dec 17  2017 libvtkImagingFourier-6.3.so.6.3.0
lrwxrwxrwx  1 root root    42 Dec 17  2017 libvtkImagingFourierPython27D-6.3.so.6.3 -> libvtkImagingFourierPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  103K Dec 17  2017 libvtkImagingFourierPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    36 Dec 17  2017 libvtkImagingFourierTCL-6.3.so.6.3 -> libvtkImagingFourierTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   87K Dec 17  2017 libvtkImagingFourierTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    33 Dec 17  2017 libvtkImagingGeneral-6.3.so.6.3 -> libvtkImagingGeneral-6.3.so.6.3.0
-rw-r--r--  1 root root  1.1M Dec 17  2017 libvtkImagingGeneral-6.3.so.6.3.0
lrwxrwxrwx  1 root root    42 Dec 17  2017 libvtkImagingGeneralPython27D-6.3.so.6.3 -> libvtkImagingGeneralPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  378K Dec 17  2017 libvtkImagingGeneralPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    36 Dec 17  2017 libvtkImagingGeneralTCL-6.3.so.6.3 -> libvtkImagingGeneralTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  348K Dec 17  2017 libvtkImagingGeneralTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    32 Dec 17  2017 libvtkImagingHybrid-6.3.so.6.3 -> libvtkImagingHybrid-6.3.so.6.3.0
-rw-r--r--  1 root root  505K Dec 17  2017 libvtkImagingHybrid-6.3.so.6.3.0
lrwxrwxrwx  1 root root    41 Dec 17  2017 libvtkImagingHybridPython27D-6.3.so.6.3 -> libvtkImagingHybridPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  539K Dec 17  2017 libvtkImagingHybridPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    35 Dec 17  2017 libvtkImagingHybridTCL-6.3.so.6.3 -> libvtkImagingHybridTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  315K Dec 17  2017 libvtkImagingHybridTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    30 Dec 17  2017 libvtkImagingMath-6.3.so.6.3 -> libvtkImagingMath-6.3.so.6.3.0
-rw-r--r--  1 root root  296K Dec 17  2017 libvtkImagingMath-6.3.so.6.3.0
lrwxrwxrwx  1 root root    39 Dec 17  2017 libvtkImagingMathPython27D-6.3.so.6.3 -> libvtkImagingMathPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  143K Dec 17  2017 libvtkImagingMathPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    33 Dec 17  2017 libvtkImagingMathTCL-6.3.so.6.3 -> libvtkImagingMathTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   99K Dec 17  2017 libvtkImagingMathTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    39 Dec 17  2017 libvtkImagingMorphological-6.3.so.6.3 -> libvtkImagingMorphological-6.3.so.6.3.0
-rw-r--r--  1 root root  536K Dec 17  2017 libvtkImagingMorphological-6.3.so.6.3.0
lrwxrwxrwx  1 root root    48 Dec 17  2017 libvtkImagingMorphologicalPython27D-6.3.so.6.3 -> libvtkImagingMorphologicalPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  205K Dec 17  2017 libvtkImagingMorphologicalPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    42 Dec 17  2017 libvtkImagingMorphologicalTCL-6.3.so.6.3 -> libvtkImagingMorphologicalTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  143K Dec 17  2017 libvtkImagingMorphologicalTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    33 Dec 17  2017 libvtkImagingSources-6.3.so.6.3 -> libvtkImagingSources-6.3.so.6.3.0
-rw-r--r--  1 root root  331K Dec 17  2017 libvtkImagingSources-6.3.so.6.3.0
lrwxrwxrwx  1 root root    42 Dec 17  2017 libvtkImagingSourcesPython27D-6.3.so.6.3 -> libvtkImagingSourcesPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  253K Dec 17  2017 libvtkImagingSourcesPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    36 Dec 17  2017 libvtkImagingSourcesTCL-6.3.so.6.3 -> libvtkImagingSourcesTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  183K Dec 17  2017 libvtkImagingSourcesTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    36 Dec 17  2017 libvtkImagingStatistics-6.3.so.6.3 -> libvtkImagingStatistics-6.3.so.6.3.0
-rw-r--r--  1 root root  156K Dec 17  2017 libvtkImagingStatistics-6.3.so.6.3.0
lrwxrwxrwx  1 root root    45 Dec 17  2017 libvtkImagingStatisticsPython27D-6.3.so.6.3 -> libvtkImagingStatisticsPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  138K Dec 17  2017 libvtkImagingStatisticsPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    39 Dec 17  2017 libvtkImagingStatisticsTCL-6.3.so.6.3 -> libvtkImagingStatisticsTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   99K Dec 17  2017 libvtkImagingStatisticsTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    33 Dec 17  2017 libvtkImagingStencil-6.3.so.6.3 -> libvtkImagingStencil-6.3.so.6.3.0
-rw-r--r--  1 root root  212K Dec 17  2017 libvtkImagingStencil-6.3.so.6.3.0
lrwxrwxrwx  1 root root    42 Dec 17  2017 libvtkImagingStencilPython27D-6.3.so.6.3 -> libvtkImagingStencilPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  147K Dec 17  2017 libvtkImagingStencilPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    36 Dec 17  2017 libvtkImagingStencilTCL-6.3.so.6.3 -> libvtkImagingStencilTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  107K Dec 17  2017 libvtkImagingStencilTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    46 Dec 17  2017 libvtkInfovisBoostGraphAlgorithms-6.3.so.6.3 -> libvtkInfovisBoostGraphAlgorithms-6.3.so.6.3.0
-rw-r--r--  1 root root  477K Dec 17  2017 libvtkInfovisBoostGraphAlgorithms-6.3.so.6.3.0
lrwxrwxrwx  1 root root    55 Dec 17  2017 libvtkInfovisBoostGraphAlgorithmsPython27D-6.3.so.6.3 -> libvtkInfovisBoostGraphAlgorithmsPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  205K Dec 17  2017 libvtkInfovisBoostGraphAlgorithmsPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    49 Dec 17  2017 libvtkInfovisBoostGraphAlgorithmsTCL-6.3.so.6.3 -> libvtkInfovisBoostGraphAlgorithmsTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  163K Dec 17  2017 libvtkInfovisBoostGraphAlgorithmsTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    30 Dec 17  2017 libvtkInfovisCore-6.3.so.6.3 -> libvtkInfovisCore-6.3.so.6.3.0
-rw-r--r--  1 root root 1008K Dec 17  2017 libvtkInfovisCore-6.3.so.6.3.0
lrwxrwxrwx  1 root root    39 Dec 17  2017 libvtkInfovisCorePython27D-6.3.so.6.3 -> libvtkInfovisCorePython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  752K Dec 17  2017 libvtkInfovisCorePython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    33 Dec 17  2017 libvtkInfovisCoreTCL-6.3.so.6.3 -> libvtkInfovisCoreTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  536K Dec 17  2017 libvtkInfovisCoreTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    32 Dec 17  2017 libvtkInfovisLayout-6.3.so.6.3 -> libvtkInfovisLayout-6.3.so.6.3.0
-rw-r--r--  1 root root  670K Dec 17  2017 libvtkInfovisLayout-6.3.so.6.3.0
lrwxrwxrwx  1 root root    41 Dec 17  2017 libvtkInfovisLayoutPython27D-6.3.so.6.3 -> libvtkInfovisLayoutPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  897K Dec 17  2017 libvtkInfovisLayoutPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    35 Dec 17  2017 libvtkInfovisLayoutTCL-6.3.so.6.3 -> libvtkInfovisLayoutTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  600K Dec 17  2017 libvtkInfovisLayoutTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    35 Dec 17  2017 libvtkInteractionImage-6.3.so.6.3 -> libvtkInteractionImage-6.3.so.6.3.0
-rw-r--r--  1 root root  152K Dec 17  2017 libvtkInteractionImage-6.3.so.6.3.0
lrwxrwxrwx  1 root root    44 Dec 17  2017 libvtkInteractionImagePython27D-6.3.so.6.3 -> libvtkInteractionImagePython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  144K Dec 17  2017 libvtkInteractionImagePython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    38 Dec 17  2017 libvtkInteractionImageTCL-6.3.so.6.3 -> libvtkInteractionImageTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   91K Dec 17  2017 libvtkInteractionImageTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    35 Dec 17  2017 libvtkInteractionStyle-6.3.so.6.3 -> libvtkInteractionStyle-6.3.so.6.3.0
-rw-r--r--  1 root root  537K Dec 17  2017 libvtkInteractionStyle-6.3.so.6.3.0
lrwxrwxrwx  1 root root    44 Dec 17  2017 libvtkInteractionStylePython27D-6.3.so.6.3 -> libvtkInteractionStylePython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  349K Dec 17  2017 libvtkInteractionStylePython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    38 Dec 17  2017 libvtkInteractionStyleTCL-6.3.so.6.3 -> libvtkInteractionStyleTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  219K Dec 17  2017 libvtkInteractionStyleTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    37 Dec 17  2017 libvtkInteractionWidgets-6.3.so.6.3 -> libvtkInteractionWidgets-6.3.so.6.3.0
-rw-r--r--  1 root root  3.2M Dec 17  2017 libvtkInteractionWidgets-6.3.so.6.3.0
lrwxrwxrwx  1 root root    46 Dec 17  2017 libvtkInteractionWidgetsPython27D-6.3.so.6.3 -> libvtkInteractionWidgetsPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  3.6M Dec 17  2017 libvtkInteractionWidgetsPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    40 Dec 17  2017 libvtkInteractionWidgetsTCL-6.3.so.6.3 -> libvtkInteractionWidgetsTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  2.1M Dec 17  2017 libvtkInteractionWidgetsTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    31 Dec 17  2017 libvtkLocalExample-6.3.so.6.3 -> libvtkLocalExample-6.3.so.6.3.0
-rw-r--r--  1 root root   15K Dec 17  2017 libvtkLocalExample-6.3.so.6.3.0
lrwxrwxrwx  1 root root    34 Dec 17  2017 libvtkLocalExampleTCL-6.3.so.6.3 -> libvtkLocalExampleTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   15K Dec 17  2017 libvtkLocalExampleTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    31 Dec 17  2017 libvtkParallelCore-6.3.so.6.3 -> libvtkParallelCore-6.3.so.6.3.0
-rw-r--r--  1 root root  305K Dec 17  2017 libvtkParallelCore-6.3.so.6.3.0
lrwxrwxrwx  1 root root    40 Dec 17  2017 libvtkParallelCorePython27D-6.3.so.6.3 -> libvtkParallelCorePython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  376K Dec 17  2017 libvtkParallelCorePython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    34 Dec 17  2017 libvtkParallelCoreTCL-6.3.so.6.3 -> libvtkParallelCoreTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  152K Dec 17  2017 libvtkParallelCoreTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    30 Dec 17  2017 libvtkParallelMPI-6.3.so.6.3 -> libvtkParallelMPI-6.3.so.6.3.0
-rw-r--r--  1 root root  149K Dec 17  2017 libvtkParallelMPI-6.3.so.6.3.0
lrwxrwxrwx  1 root root    33 Dec 17  2017 libvtkParallelMPI4Py-6.3.so.6.3 -> libvtkParallelMPI4Py-6.3.so.6.3.0
-rw-r--r--  1 root root   68K Dec 17  2017 libvtkParallelMPI4Py-6.3.so.6.3.0
lrwxrwxrwx  1 root root    42 Dec 17  2017 libvtkParallelMPI4PyPython27D-6.3.so.6.3 -> libvtkParallelMPI4PyPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   15K Dec 17  2017 libvtkParallelMPI4PyPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    39 Dec 17  2017 libvtkParallelMPIPython27D-6.3.so.6.3 -> libvtkParallelMPIPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   73K Dec 17  2017 libvtkParallelMPIPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    33 Dec 17  2017 libvtkParallelMPITCL-6.3.so.6.3 -> libvtkParallelMPITCL-6.3.so.6.3.0
-rw-r--r--  1 root root   35K Dec 17  2017 libvtkParallelMPITCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    36 Dec 17  2017 libvtkPythonInterpreter-6.3.so.6.3 -> libvtkPythonInterpreter-6.3.so.6.3.0
-rw-r--r--  1 root root   40K Dec 17  2017 libvtkPythonInterpreter-6.3.so.6.3.0
lrwxrwxrwx  1 root root    45 Dec 17  2017 libvtkPythonInterpreterPython27D-6.3.so.6.3 -> libvtkPythonInterpreterPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   28K Dec 17  2017 libvtkPythonInterpreterPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    39 Dec 17  2017 libvtkPythonInterpreterTCL-6.3.so.6.3 -> libvtkPythonInterpreterTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   27K Dec 17  2017 libvtkPythonInterpreterTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    38 Dec 17  2017 libvtkRenderingAnnotation-6.3.so.6.3 -> libvtkRenderingAnnotation-6.3.so.6.3.0
-rw-r--r--  1 root root  1.3M Dec 17  2017 libvtkRenderingAnnotation-6.3.so.6.3.0
lrwxrwxrwx  1 root root    47 Dec 17  2017 libvtkRenderingAnnotationPython27D-6.3.so.6.3 -> libvtkRenderingAnnotationPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  1.9M Dec 17  2017 libvtkRenderingAnnotationPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    41 Dec 17  2017 libvtkRenderingAnnotationTCL-6.3.so.6.3 -> libvtkRenderingAnnotationTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  945K Dec 17  2017 libvtkRenderingAnnotationTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    37 Dec 17  2017 libvtkRenderingContext2D-6.3.so.6.3 -> libvtkRenderingContext2D-6.3.so.6.3.0
-rw-r--r--  1 root root  252K Dec 17  2017 libvtkRenderingContext2D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    46 Dec 17  2017 libvtkRenderingContext2DPython27D-6.3.so.6.3 -> libvtkRenderingContext2DPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  320K Dec 17  2017 libvtkRenderingContext2DPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    41 Dec 17  2017 libvtkRenderingContextIIDTCL-6.3.so.6.3 -> libvtkRenderingContextIIDTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  212K Dec 17  2017 libvtkRenderingContextIIDTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    41 Dec 17  2017 libvtkRenderingContextOpenGL-6.3.so.6.3 -> libvtkRenderingContextOpenGL-6.3.so.6.3.0
-rw-r--r--  1 root root  196K Dec 17  2017 libvtkRenderingContextOpenGL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    50 Dec 17  2017 libvtkRenderingContextOpenGLPython27D-6.3.so.6.3 -> libvtkRenderingContextOpenGLPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   15K Dec 17  2017 libvtkRenderingContextOpenGLPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    44 Dec 17  2017 libvtkRenderingContextOpenGLTCL-6.3.so.6.3 -> libvtkRenderingContextOpenGLTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   15K Dec 17  2017 libvtkRenderingContextOpenGLTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    32 Dec 17  2017 libvtkRenderingCore-6.3.so.6.3 -> libvtkRenderingCore-6.3.so.6.3.0
-rw-r--r--  1 root root  2.5M Dec 17  2017 libvtkRenderingCore-6.3.so.6.3.0
lrwxrwxrwx  1 root root    41 Dec 17  2017 libvtkRenderingCorePython27D-6.3.so.6.3 -> libvtkRenderingCorePython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  3.6M Dec 17  2017 libvtkRenderingCorePython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    35 Dec 17  2017 libvtkRenderingCoreTCL-6.3.so.6.3 -> libvtkRenderingCoreTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  2.2M Dec 17  2017 libvtkRenderingCoreTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    36 Dec 17  2017 libvtkRenderingExternal-6.3.so.6.3 -> libvtkRenderingExternal-6.3.so.6.3.0
-rw-r--r--  1 root root  236K Dec 17  2017 libvtkRenderingExternal-6.3.so.6.3.0
lrwxrwxrwx  1 root root    45 Dec 17  2017 libvtkRenderingExternalPython27D-6.3.so.6.3 -> libvtkRenderingExternalPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   86K Dec 17  2017 libvtkRenderingExternalPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    39 Dec 17  2017 libvtkRenderingExternalTCL-6.3.so.6.3 -> libvtkRenderingExternalTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   71K Dec 17  2017 libvtkRenderingExternalTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    36 Dec 17  2017 libvtkRenderingFreeType-6.3.so.6.3 -> libvtkRenderingFreeType-6.3.so.6.3.0
-rw-r--r--  1 root root  781K Dec 17  2017 libvtkRenderingFreeType-6.3.so.6.3.0
lrwxrwxrwx  1 root root    46 Dec 17  2017 libvtkRenderingFreeTypeFontConfig-6.3.so.6.3 -> libvtkRenderingFreeTypeFontConfig-6.3.so.6.3.0
-rw-r--r--  1 root root   39K Dec 17  2017 libvtkRenderingFreeTypeFontConfig-6.3.so.6.3.0
lrwxrwxrwx  1 root root    45 Dec 17  2017 libvtkRenderingFreeTypePython27D-6.3.so.6.3 -> libvtkRenderingFreeTypePython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   61K Dec 17  2017 libvtkRenderingFreeTypePython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    39 Dec 17  2017 libvtkRenderingFreeTypeTCL-6.3.so.6.3 -> libvtkRenderingFreeTypeTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   47K Dec 17  2017 libvtkRenderingFreeTypeTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    33 Dec 17  2017 libvtkRenderingGL2PS-6.3.so.6.3 -> libvtkRenderingGL2PS-6.3.so.6.3.0
-rw-r--r--  1 root root  112K Dec 17  2017 libvtkRenderingGL2PS-6.3.so.6.3.0
lrwxrwxrwx  1 root root    42 Dec 17  2017 libvtkRenderingGL2PSPython27D-6.3.so.6.3 -> libvtkRenderingGL2PSPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   23K Dec 17  2017 libvtkRenderingGL2PSPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    37 Dec 17  2017 libvtkRenderingGLtoPSTCL-6.3.so.6.3 -> libvtkRenderingGLtoPSTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   19K Dec 17  2017 libvtkRenderingGLtoPSTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    33 Dec 17  2017 libvtkRenderingImage-6.3.so.6.3 -> libvtkRenderingImage-6.3.so.6.3.0
-rw-r--r--  1 root root  136K Dec 17  2017 libvtkRenderingImage-6.3.so.6.3.0
lrwxrwxrwx  1 root root    42 Dec 17  2017 libvtkRenderingImagePython27D-6.3.so.6.3 -> libvtkRenderingImagePython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  110K Dec 17  2017 libvtkRenderingImagePython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    36 Dec 17  2017 libvtkRenderingImageTCL-6.3.so.6.3 -> libvtkRenderingImageTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   63K Dec 17  2017 libvtkRenderingImageTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    31 Dec 17  2017 libvtkRenderingLIC-6.3.so.6.3 -> libvtkRenderingLIC-6.3.so.6.3.0
-rw-r--r--  1 root root  647K Dec 17  2017 libvtkRenderingLIC-6.3.so.6.3.0
lrwxrwxrwx  1 root root    40 Dec 17  2017 libvtkRenderingLICPython27D-6.3.so.6.3 -> libvtkRenderingLICPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  164K Dec 17  2017 libvtkRenderingLICPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    34 Dec 17  2017 libvtkRenderingLICTCL-6.3.so.6.3 -> libvtkRenderingLICTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   99K Dec 17  2017 libvtkRenderingLICTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    31 Dec 17  2017 libvtkRenderingLOD-6.3.so.6.3 -> libvtkRenderingLOD-6.3.so.6.3.0
-rw-r--r--  1 root root   99K Dec 17  2017 libvtkRenderingLOD-6.3.so.6.3.0
lrwxrwxrwx  1 root root    40 Dec 17  2017 libvtkRenderingLODPython27D-6.3.so.6.3 -> libvtkRenderingLODPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  106K Dec 17  2017 libvtkRenderingLODPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    34 Dec 17  2017 libvtkRenderingLODTCL-6.3.so.6.3 -> libvtkRenderingLODTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   59K Dec 17  2017 libvtkRenderingLODTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    33 Dec 17  2017 libvtkRenderingLabel-6.3.so.6.3 -> libvtkRenderingLabel-6.3.so.6.3.0
-rw-r--r--  1 root root  689K Dec 17  2017 libvtkRenderingLabel-6.3.so.6.3.0
lrwxrwxrwx  1 root root    42 Dec 17  2017 libvtkRenderingLabelPython27D-6.3.so.6.3 -> libvtkRenderingLabelPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  434K Dec 17  2017 libvtkRenderingLabelPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    36 Dec 17  2017 libvtkRenderingLabelTCL-6.3.so.6.3 -> libvtkRenderingLabelTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  239K Dec 17  2017 libvtkRenderingLabelTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    38 Dec 17  2017 libvtkRenderingMatplotlib-6.3.so.6.3 -> libvtkRenderingMatplotlib-6.3.so.6.3.0
-rw-r--r--  1 root root   67K Dec 17  2017 libvtkRenderingMatplotlib-6.3.so.6.3.0
lrwxrwxrwx  1 root root    47 Dec 17  2017 libvtkRenderingMatplotlibPython27D-6.3.so.6.3 -> libvtkRenderingMatplotlibPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   23K Dec 17  2017 libvtkRenderingMatplotlibPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    41 Dec 17  2017 libvtkRenderingMatplotlibTCL-6.3.so.6.3 -> libvtkRenderingMatplotlibTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   23K Dec 17  2017 libvtkRenderingMatplotlibTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    34 Dec 17  2017 libvtkRenderingOpenGL-6.3.so.6.3 -> libvtkRenderingOpenGL-6.3.so.6.3.0
-rw-r--r--  1 root root  2.7M Dec 17  2017 libvtkRenderingOpenGL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    43 Dec 17  2017 libvtkRenderingOpenGLPython27D-6.3.so.6.3 -> libvtkRenderingOpenGLPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  983K Dec 17  2017 libvtkRenderingOpenGLPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    37 Dec 17  2017 libvtkRenderingOpenGLTCL-6.3.so.6.3 -> libvtkRenderingOpenGLTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  711K Dec 17  2017 libvtkRenderingOpenGLTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    36 Dec 17  2017 libvtkRenderingParallel-6.3.so.6.3 -> libvtkRenderingParallel-6.3.so.6.3.0
-rw-r--r--  1 root root  341K Dec 17  2017 libvtkRenderingParallel-6.3.so.6.3.0
lrwxrwxrwx  1 root root    39 Dec 17  2017 libvtkRenderingParallelLIC-6.3.so.6.3 -> libvtkRenderingParallelLIC-6.3.so.6.3.0
-rw-r--r--  1 root root  621K Dec 17  2017 libvtkRenderingParallelLIC-6.3.so.6.3.0
lrwxrwxrwx  1 root root    48 Dec 17  2017 libvtkRenderingParallelLICPython27D-6.3.so.6.3 -> libvtkRenderingParallelLICPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   48K Dec 17  2017 libvtkRenderingParallelLICPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    42 Dec 17  2017 libvtkRenderingParallelLICTCL-6.3.so.6.3 -> libvtkRenderingParallelLICTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   39K Dec 17  2017 libvtkRenderingParallelLICTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    45 Dec 17  2017 libvtkRenderingParallelPython27D-6.3.so.6.3 -> libvtkRenderingParallelPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  284K Dec 17  2017 libvtkRenderingParallelPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    39 Dec 17  2017 libvtkRenderingParallelTCL-6.3.so.6.3 -> libvtkRenderingParallelTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  179K Dec 17  2017 libvtkRenderingParallelTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    33 Dec 17  2017 libvtkRenderingTkTCL-6.3.so.6.3 -> libvtkRenderingTkTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  104K Dec 17  2017 libvtkRenderingTkTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    34 Dec 17  2017 libvtkRenderingVolume-6.3.so.6.3 -> libvtkRenderingVolume-6.3.so.6.3.0
-rw-r--r--  1 root root  4.2M Dec 17  2017 libvtkRenderingVolume-6.3.so.6.3.0
lrwxrwxrwx  1 root root    37 Dec 17  2017 libvtkRenderingVolumeAMR-6.3.so.6.3 -> libvtkRenderingVolumeAMR-6.3.so.6.3.0
-rw-r--r--  1 root root   87K Dec 17  2017 libvtkRenderingVolumeAMR-6.3.so.6.3.0
lrwxrwxrwx  1 root root    46 Dec 17  2017 libvtkRenderingVolumeAMRPython27D-6.3.so.6.3 -> libvtkRenderingVolumeAMRPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   69K Dec 17  2017 libvtkRenderingVolumeAMRPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    40 Dec 17  2017 libvtkRenderingVolumeAMRTCL-6.3.so.6.3 -> libvtkRenderingVolumeAMRTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   39K Dec 17  2017 libvtkRenderingVolumeAMRTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    40 Dec 17  2017 libvtkRenderingVolumeOpenGL-6.3.so.6.3 -> libvtkRenderingVolumeOpenGL-6.3.so.6.3.0
-rw-r--r--  1 root root  670K Dec 17  2017 libvtkRenderingVolumeOpenGL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    49 Dec 17  2017 libvtkRenderingVolumeOpenGLPython27D-6.3.so.6.3 -> libvtkRenderingVolumeOpenGLPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  127K Dec 17  2017 libvtkRenderingVolumeOpenGLPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    43 Dec 17  2017 libvtkRenderingVolumeOpenGLTCL-6.3.so.6.3 -> libvtkRenderingVolumeOpenGLTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   91K Dec 17  2017 libvtkRenderingVolumeOpenGLTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    43 Dec 17  2017 libvtkRenderingVolumePython27D-6.3.so.6.3 -> libvtkRenderingVolumePython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  1.1M Dec 17  2017 libvtkRenderingVolumePython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    37 Dec 17  2017 libvtkRenderingVolumeTCL-6.3.so.6.3 -> libvtkRenderingVolumeTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  620K Dec 17  2017 libvtkRenderingVolumeTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    39 Dec 17  2017 libvtkTestingGenericBridge-6.3.so.6.3 -> libvtkTestingGenericBridge-6.3.so.6.3.0
-rw-r--r--  1 root root  132K Dec 17  2017 libvtkTestingGenericBridge-6.3.so.6.3.0
lrwxrwxrwx  1 root root    31 Dec 17  2017 libvtkTestingIOSQL-6.3.so.6.3 -> libvtkTestingIOSQL-6.3.so.6.3.0
-rw-r--r--  1 root root   15K Dec 17  2017 libvtkTestingIOSQL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    35 Dec 17  2017 libvtkTestingRendering-6.3.so.6.3 -> libvtkTestingRendering-6.3.so.6.3.0
-rw-r--r--  1 root root  136K Dec 17  2017 libvtkTestingRendering-6.3.so.6.3.0
lrwxrwxrwx  1 root root    44 Dec 17  2017 libvtkTestingRenderingPython27D-6.3.so.6.3 -> libvtkTestingRenderingPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   61K Dec 17  2017 libvtkTestingRenderingPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    38 Dec 17  2017 libvtkTestingRenderingTCL-6.3.so.6.3 -> libvtkTestingRenderingTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   43K Dec 17  2017 libvtkTestingRenderingTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    23 Dec 17  2017 libvtkVPIC-6.3.so.6.3 -> libvtkVPIC-6.3.so.6.3.0
-rw-r--r--  1 root root   95K Dec 17  2017 libvtkVPIC-6.3.so.6.3.0
lrwxrwxrwx  1 root root    33 Dec 17  2017 libvtkViewsContext2D-6.3.so.6.3 -> libvtkViewsContext2D-6.3.so.6.3.0
-rw-r--r--  1 root root   75K Dec 17  2017 libvtkViewsContext2D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    42 Dec 17  2017 libvtkViewsContext2DPython27D-6.3.so.6.3 -> libvtkViewsContext2DPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   32K Dec 17  2017 libvtkViewsContext2DPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    37 Dec 17  2017 libvtkViewsContextIIDTCL-6.3.so.6.3 -> libvtkViewsContextIIDTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   27K Dec 17  2017 libvtkViewsContextIIDTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    28 Dec 17  2017 libvtkViewsCore-6.3.so.6.3 -> libvtkViewsCore-6.3.so.6.3.0
-rw-r--r--  1 root root  172K Dec 17  2017 libvtkViewsCore-6.3.so.6.3.0
lrwxrwxrwx  1 root root    37 Dec 17  2017 libvtkViewsCorePython27D-6.3.so.6.3 -> libvtkViewsCorePython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  181K Dec 17  2017 libvtkViewsCorePython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    31 Dec 17  2017 libvtkViewsCoreTCL-6.3.so.6.3 -> libvtkViewsCoreTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  119K Dec 17  2017 libvtkViewsCoreTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    30 Dec 17  2017 libvtkViewsGeovis-6.3.so.6.3 -> libvtkViewsGeovis-6.3.so.6.3.0
-rw-r--r--  1 root root   71K Dec 17  2017 libvtkViewsGeovis-6.3.so.6.3.0
lrwxrwxrwx  1 root root    39 Dec 17  2017 libvtkViewsGeovisPython27D-6.3.so.6.3 -> libvtkViewsGeovisPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root   32K Dec 17  2017 libvtkViewsGeovisPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    33 Dec 17  2017 libvtkViewsGeovisTCL-6.3.so.6.3 -> libvtkViewsGeovisTCL-6.3.so.6.3.0
-rw-r--r--  1 root root   27K Dec 17  2017 libvtkViewsGeovisTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    31 Dec 17  2017 libvtkViewsInfovis-6.3.so.6.3 -> libvtkViewsInfovis-6.3.so.6.3.0
-rw-r--r--  1 root root  1.1M Dec 17  2017 libvtkViewsInfovis-6.3.so.6.3.0
lrwxrwxrwx  1 root root    40 Dec 17  2017 libvtkViewsInfovisPython27D-6.3.so.6.3 -> libvtkViewsInfovisPython27D-6.3.so.6.3.0
-rw-r--r--  1 root root  927K Dec 17  2017 libvtkViewsInfovisPython27D-6.3.so.6.3.0
lrwxrwxrwx  1 root root    34 Dec 17  2017 libvtkViewsInfovisTCL-6.3.so.6.3 -> libvtkViewsInfovisTCL-6.3.so.6.3.0
-rw-r--r--  1 root root  565K Dec 17  2017 libvtkViewsInfovisTCL-6.3.so.6.3.0
lrwxrwxrwx  1 root root    31 Dec 17  2017 libvtkWrappingJava-6.3.so.6.3 -> libvtkWrappingJava-6.3.so.6.3.0
-rw-r--r--  1 root root   19K Dec 17  2017 libvtkWrappingJava-6.3.so.6.3.0
lrwxrwxrwx  1 root root    39 Dec 17  2017 libvtkWrappingPython27Core-6.3.so.6.3 -> libvtkWrappingPython27Core-6.3.so.6.3.0
-rw-r--r--  1 root root  180K Dec 17  2017 libvtkWrappingPython27Core-6.3.so.6.3.0
lrwxrwxrwx  1 root root    25 Dec 17  2017 libvtkalglib-6.3.so.6.3 -> libvtkalglib-6.3.so.6.3.0
-rw-r--r--  1 root root  139K Dec 17  2017 libvtkalglib-6.3.so.6.3.0
lrwxrwxrwx  1 root root    25 Dec 17  2017 libvtkexoIIc-6.3.so.6.3 -> libvtkexoIIc-6.3.so.6.3.0
-rw-r--r--  1 root root  288K Dec 17  2017 libvtkexoIIc-6.3.so.6.3.0
lrwxrwxrwx  1 root root    23 Dec 17  2017 libvtkftgl-6.3.so.6.3 -> libvtkftgl-6.3.so.6.3.0
-rw-r--r--  1 root root   43K Dec 17  2017 libvtkftgl-6.3.so.6.3.0
lrwxrwxrwx  1 root root    25 Dec 17  2017 libvtkmetaio-6.3.so.6.3 -> libvtkmetaio-6.3.so.6.3.0
-rw-r--r--  1 root root  591K Dec 17  2017 libvtkmetaio-6.3.so.6.3.0
lrwxrwxrwx  1 root root    22 Dec 17  2017 libvtksys-6.3.so.6.3 -> libvtksys-6.3.so.6.3.0
-rw-r--r--  1 root root  278K Dec 17  2017 libvtksys-6.3.so.6.3.0
lrwxrwxrwx  1 root root    26 Dec 17  2017 libvtkverdict-6.3.so.6.3 -> libvtkverdict-6.3.so.6.3.0
-rw-r--r--  1 root root  195K Dec 17  2017 libvtkverdict-6.3.so.6.3.0
lrwxrwxrwx  1 root root    24 Dec 17  2017 libvtkxdmf2-6.3.so.6.3 -> libvtkxdmf2-6.3.so.6.3.0
-rw-r--r--  1 root root  472K Dec 17  2017 libvtkxdmf2-6.3.so.6.3.0
lrwxrwxrwx  1 root root    19 Jan  5  2021 libwavpack.so.1 -> libwavpack.so.1.2.0
-rw-r--r--  1 root root  167K Jan  5  2021 libwavpack.so.1.2.0
lrwxrwxrwx  1 root root    26 Feb  1  2020 libwayland-client.so.0 -> libwayland-client.so.0.3.0
-rw-r--r--  1 root root   60K Feb  1  2020 libwayland-client.so.0.3.0
lrwxrwxrwx  1 root root    26 Feb  1  2020 libwayland-cursor.so.0 -> libwayland-cursor.so.0.0.0
-rw-r--r--  1 root root   31K Feb  1  2020 libwayland-cursor.so.0.0.0
lrwxrwxrwx  1 root root    23 Feb  1  2020 libwayland-egl.so.1 -> libwayland-egl.so.1.0.0
-rw-r--r--  1 root root  5.9K Feb  1  2020 libwayland-egl.so.1.0.0
lrwxrwxrwx  1 root root    16 May 20  2021 libwebp.so.6 -> libwebp.so.6.0.2
-rw-r--r--  1 root root  411K May 20  2021 libwebp.so.6.0.2
lrwxrwxrwx  1 root root    19 May 20  2021 libwebpmux.so.3 -> libwebpmux.so.3.0.1
-rw-r--r--  1 root root   39K May 20  2021 libwebpmux.so.3.0.1
lrwxrwxrwx  1 root root    16 Dec 15  2017 libwind.so.0 -> libwind.so.0.0.0
-rw-r--r--  1 root root  162K Dec 15  2017 libwind.so.0.0.0
-rw-r--r--  1 root root  1.2M Jan 19  2018 libx264.so.152
-rw-r--r--  1 root root   11M Dec 31  2017 libx265.so.146
lrwxrwxrwx  1 root root    20 Jun 21  2018 libxcb-dri2.so.0 -> libxcb-dri2.so.0.0.0
-rw-r--r--  1 root root   19K Jun 21  2018 libxcb-dri2.so.0.0.0
lrwxrwxrwx  1 root root    20 Jun 21  2018 libxcb-dri3.so.0 -> libxcb-dri3.so.0.0.0
-rw-r--r--  1 root root   15K Jun 21  2018 libxcb-dri3.so.0.0.0
lrwxrwxrwx  1 root root    19 Jun 21  2018 libxcb-glx.so.0 -> libxcb-glx.so.0.0.0
-rw-r--r--  1 root root  107K Jun 21  2018 libxcb-glx.so.0.0.0
lrwxrwxrwx  1 root root    23 Jun 21  2018 libxcb-present.so.0 -> libxcb-present.so.0.0.0
-rw-r--r--  1 root root   11K Jun 21  2018 libxcb-present.so.0.0.0
lrwxrwxrwx  1 root root    22 Jun 21  2018 libxcb-render.so.0 -> libxcb-render.so.0.0.0
-rw-r--r--  1 root root   51K Jun 21  2018 libxcb-render.so.0.0.0
lrwxrwxrwx  1 root root    19 Jun 21  2018 libxcb-shm.so.0 -> libxcb-shm.so.0.0.0
-rw-r--r--  1 root root   11K Jun 21  2018 libxcb-shm.so.0.0.0
lrwxrwxrwx  1 root root    20 Jun 21  2018 libxcb-sync.so.1 -> libxcb-sync.so.1.0.0
-rw-r--r--  1 root root   27K Jun 21  2018 libxcb-sync.so.1.0.0
lrwxrwxrwx  1 root root    15 Jun 21  2018 libxcb.so.1 -> libxcb.so.1.1.0
-rw-r--r--  1 root root  159K Jun 21  2018 libxcb.so.1.1.0
lrwxrwxrwx  1 root root    16 Mar 24  2018 libxdot.so -> libxdot.so.4.0.0
lrwxrwxrwx  1 root root    16 Mar 24  2018 libxdot.so.4 -> libxdot.so.4.0.0
-rw-r--r--  1 root root   23K Mar 24  2018 libxdot.so.4.0.0
-rw-r--r--  1 root root  3.5M Nov 10  2017 libxerces-c-3.2.so
lrwxrwxrwx  1 root root    21 Mar  5  2019 libxkbcommon.so.0 -> libxkbcommon.so.0.0.0
-rw-r--r--  1 root root  251K Mar  5  2019 libxkbcommon.so.0.0.0
-rw-r--r--  1 root root  2.9M Aug  1 14:25 libxml2.a
lrwxrwxrwx  1 root root    16 Aug  1 14:25 libxml2.so -> libxml2.so.2.9.4
lrwxrwxrwx  1 root root    16 Aug  1 14:25 libxml2.so.2 -> libxml2.so.2.9.4
-rw-r--r--  1 root root  1.8M Aug  1 14:25 libxml2.so.2.9.4
lrwxrwxrwx  1 root root    21 Mar 18  2018 libxshmfence.so.1 -> libxshmfence.so.1.0.0
-rw-r--r--  1 root root  6.1K Mar 18  2018 libxshmfence.so.1.0.0
lrwxrwxrwx  1 root root    18 Jan 28  2018 libxvidcore.so.4 -> libxvidcore.so.4.3
-rw-r--r--  1 root root  672K Jan 28  2018 libxvidcore.so.4.3
-rw-r--r--  1 root root  162K Mar 26  2022 libz.a
lrwxrwxrwx  1 root root    36 Mar 26  2022 libz.so -> /lib/x86_64-linux-gnu/libz.so.1.2.11
lrwxrwxrwx  1 root root    16 Mar  3  2021 libzstd.so.1 -> libzstd.so.1.3.3
-rw-r--r--  1 root root  490K Mar  3  2021 libzstd.so.1.3.3
lrwxrwxrwx  1 root root    23 Nov 17  2016 libzvbi-chains.so.0 -> libzvbi-chains.so.0.0.0
-rw-r--r--  1 root root   58K Nov 17  2016 libzvbi-chains.so.0.0.0
lrwxrwxrwx  1 root root    17 Nov 17  2016 libzvbi.so.0 -> libzvbi.so.0.13.2
-rw-r--r--  1 root root  556K Nov 17  2016 libzvbi.so.0.13.2
drwxr-xr-x  2 root root  4.0K Aug 16 08:53 nss
drwxr-xr-x  2 root root  4.0K Aug 16 08:53 odbc
drwxr-xr-x  2 root root  4.0K Aug 16 08:53 openblas
drwxr-xr-x  3 root root  4.0K Aug 16 08:53 openmpi
drwxr-xr-x  1 root root  4.0K May 28 19:15 perl
drwxr-xr-x 16 root root  4.0K Apr 28 00:03 perl-base
drwxr-xr-x  2 root root  4.0K Aug 16 08:54 pkgconfig
-rw-r--r--  1 root root  1.8K May  3 10:19 rcrt1.o
drwxr-xr-x  2 root root  4.0K Aug 16 08:53 rsocket
drwxr-xr-x  2 root root  4.0K May 23 21:53 sasl2
drwxr-xr-x  2 root root  4.0K May 28 19:16 stubs
drwxr-xr-x  2 root root  4.0K Aug 16 08:53 vdpau
drwxr-xr-x  2 root root  4.0K Aug 16 08:53 x264-10bit
-rw-r--r--  1 root root   204 Aug  1 14:25 xml2Conf.sh
[Pipeline] echo
Packing build/libdgl.so, build/runUnitTests, python/dgl/_ffi/_cy3/core.cpython-*-x86_64-linux-gnu.so, build/tensoradapter/pytorch/*.so into dgl-gpu-linux
[Pipeline] stash
Stashed 4 file(s)
Post stage
[Pipeline] cleanWs
[WS-CLEANUP] Deleting project workspace...
[WS-CLEANUP] Deferred wipeout is disabled by the job configuration...
[WS-CLEANUP] done
[Pipeline] }
$ docker stop --time=1 681e11859cd167f00c1889ac9c7c8c96d552be64f67c250e1cf0b5bc1d83bafe
$ docker rm -f 681e11859cd167f00c1889ac9c7c8c96d552be64f67c250e1cf0b5bc1d83bafe
[Pipeline] // withDockerContainer
[Pipeline] }
[Pipeline] // withEnv
[Pipeline] }
[Pipeline] // node
[Pipeline] }
[Pipeline] // stage
[Pipeline] }
[ 61%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/array_cumsum.cc.o
[ 61%] Building CXX object CMakeFiles/dgl.dir/src/array/array.cc.o
[ 61%] Building CXX object CMakeFiles/dgl.dir/src/array/array_arith.cc.o
[ 61%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/array_index_select.cc.o
[ 61%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/array_pack.cc.o
[ 61%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/array_nonzero.cc.o
[ 61%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/array_op_impl.cc.o
[ 62%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/csr_mm.cc.o
[ 62%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/coo_sort.cc.o
[ 62%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/csr_remove.cc.o
[ 62%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/csr_sort.cc.o
[ 63%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/array_repeat.cc.o
[ 63%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/array_scatter.cc.o
[ 63%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/coo_coalesce.cc.o
[ 63%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/array_sort.cc.o
[ 65%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/coo_linegraph.cc.o
[ 65%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/sddmm.cc.o
[ 65%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/rowwise_topk.cc.o
[ 65%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/coo_remove.cc.o
[ 65%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/segment_reduce.cc.o
[ 65%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/csr_to_simple.cc.o
[ 65%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/gather_mm.cc.o
[ 65%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/csr_get_data.cc.o
[ 65%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/csr_sum.cc.o
[ 66%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/csr_union.cc.o
[ 66%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/negative_sampling.cc.o
[ 66%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/disjoint_union.cc.o
[ 67%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/rowwise_sampling.cc.o
[ 67%] Building CXX object CMakeFiles/dgl.dir/src/array/filter.cc.o
[ 68%] Building CXX object CMakeFiles/dgl.dir/src/array/kernel.cc.o
[ 69%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/spmat_op_impl_csr.cc.o
[ 69%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/spmat_op_impl_coo.cc.o
[ 68%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/spmm.cc.o
[ 69%] Building CXX object CMakeFiles/dgl.dir/src/array/union_partition.cc.o
[ 69%] Building CXX object CMakeFiles/dgl.dir/src/array/cpu/traversal.cc.o
[ 69%] Building CXX object CMakeFiles/dgl.dir/src/array/uvm_array.cc.o
[ 69%] Building CXX object CMakeFiles/dgl.dir/src/geometry/geometry.cc.o
[ 71%] Building CXX object CMakeFiles/dgl.dir/src/bcast.cc.o
[ 71%] Building CXX object CMakeFiles/dgl.dir/src/array/libra_partition.cc.o
[ 71%] Building CXX object CMakeFiles/dgl.dir/src/c_api_common.cc.o
[ 71%] Building CXX object CMakeFiles/dgl.dir/src/geometry/cpu/geometry_op_impl.cc.o
[ 72%] Building CXX object CMakeFiles/dgl.dir/src/partition/ndarray_partition.cc.o
[ 72%] Building CXX object CMakeFiles/dgl.dir/src/random/cpu/choice.cc.o
[ 73%] Building CXX object CMakeFiles/dgl.dir/src/runtime/cpu_device_api.cc.o
[ 73%] Building CXX object CMakeFiles/dgl.dir/src/runtime/c_runtime_api.cc.o
[ 73%] Building CXX object CMakeFiles/dgl.dir/src/runtime/config.cc.o
[ 73%] Building CXX object CMakeFiles/dgl.dir/src/runtime/dlpack_convert.cc.o
[ 74%] Building CXX object CMakeFiles/dgl.dir/src/runtime/file_util.cc.o
[ 74%] Building CXX object CMakeFiles/dgl.dir/src/runtime/module.cc.o
[ 74%] Building CXX object CMakeFiles/dgl.dir/src/runtime/module_util.cc.o
[ 74%] Building CXX object CMakeFiles/dgl.dir/src/random/random.cc.o
[ 74%] Building CXX object CMakeFiles/dgl.dir/src/runtime/semaphore_wrapper.cc.o
[ 74%] Building CXX object CMakeFiles/dgl.dir/src/runtime/c_object_api.cc.o
[ 74%] Building CXX object CMakeFiles/dgl.dir/src/runtime/ndarray.cc.o
[ 75%] Building CXX object CMakeFiles/dgl.dir/src/runtime/object.cc.o
[ 75%] Building CXX object CMakeFiles/dgl.dir/src/runtime/dso_module.cc.o
[ 75%] Building CXX object CMakeFiles/dgl.dir/src/runtime/system_lib_module.cc.o
[ 75%] Building CXX object CMakeFiles/dgl.dir/src/runtime/resource_manager.cc.o
[ 77%] Building CXX object CMakeFiles/dgl.dir/src/runtime/shared_mem.cc.o
[ 77%] Building CXX object CMakeFiles/dgl.dir/src/runtime/thread_pool.cc.o
[ 77%] Building CXX object CMakeFiles/dgl.dir/src/runtime/tensordispatch.cc.o
[ 77%] Building CXX object CMakeFiles/dgl.dir/src/runtime/threading_backend.cc.o
[ 78%] Building CXX object CMakeFiles/dgl.dir/src/runtime/utils.cc.o
[ 78%] Building CXX object CMakeFiles/dgl.dir/src/runtime/workspace_pool.cc.o
[ 78%] Building CXX object CMakeFiles/dgl.dir/src/runtime/registry.cc.o
[ 78%] Building CXX object CMakeFiles/dgl.dir/src/graph/sampling/randomwalks/get_node_types_cpu.cc.o
[ 78%] Building CXX object CMakeFiles/dgl.dir/src/graph/sampling/randomwalks/node2vec_cpu.cc.o
[ 78%] Building CXX object CMakeFiles/dgl.dir/src/graph/serialize/heterograph_serialize.cc.o
[ 78%] Building CXX object CMakeFiles/dgl.dir/src/graph/sampling/randomwalks/randomwalk_cpu.cc.o
[ 79%] Building CXX object CMakeFiles/dgl.dir/src/graph/sampling/randomwalks/randomwalk_with_restart_cpu.cc.o
[ 79%] Building CXX object CMakeFiles/dgl.dir/src/graph/serialize/dglgraph_serialize.cc.o
[ 79%] Building CXX object CMakeFiles/dgl.dir/src/graph/sampling/randomwalks/randomwalks.cc.o
[ 80%] Building CXX object CMakeFiles/dgl.dir/src/graph/sampler.cc.o
[ 81%] Building CXX object CMakeFiles/dgl.dir/src/graph/subgraph.cc.o
[ 81%] Building CXX object CMakeFiles/dgl.dir/src/graph/serialize/zerocopy_serializer.cc.o
[ 83%] Building CXX object CMakeFiles/dgl.dir/src/graph/sampling/randomwalks/node2vec.cc.o
[ 83%] Building CXX object CMakeFiles/dgl.dir/src/graph/sampling/neighbor/neighbor.cc.o
[ 83%] Building CXX object CMakeFiles/dgl.dir/src/graph/transform/knn.cc.o
[ 83%] Building CXX object CMakeFiles/dgl.dir/src/graph/transform/metis_partition_hetero.cc.o
[ 83%] Building CXX object CMakeFiles/dgl.dir/src/graph/serialize/graph_serialize.cc.o
[ 83%] Building CXX object CMakeFiles/dgl.dir/src/graph/serialize/tensor_serialize.cc.o
[ 83%] Building CXX object CMakeFiles/dgl.dir/src/graph/shared_mem_manager.cc.o
[ 83%] Building CXX object CMakeFiles/dgl.dir/src/graph/transform/line_graph.cc.o
[ 84%] Building CXX object CMakeFiles/dgl.dir/src/graph/transform/cpu/knn.cc.o
[ 84%] Building CXX object CMakeFiles/dgl.dir/src/graph/transform/to_bipartite.cc.o
[ 84%] Building CXX object CMakeFiles/dgl.dir/src/api/api_container.cc.o
[ 84%] Building CXX object CMakeFiles/dgl.dir/src/graph/transform/to_simple.cc.o
[ 84%] Building CXX object CMakeFiles/dgl.dir/src/scheduler/scheduler_apis.cc.o
[ 84%] Building CXX object CMakeFiles/dgl.dir/src/scheduler/scheduler.cc.o
[ 84%] Building CXX object CMakeFiles/dgl.dir/src/graph/transform/remove_edges.cc.o
[ 84%] Building CXX object CMakeFiles/dgl.dir/src/graph/unit_graph.cc.o
[ 85%] Building CXX object CMakeFiles/dgl.dir/src/graph/metis_partition.cc.o
[ 85%] Building CXX object CMakeFiles/dgl.dir/src/graph/transform/union_partition.cc.o
[ 85%] Building CXX object CMakeFiles/dgl.dir/src/rpc/network/msg_queue.cc.o
[ 85%] Building CXX object CMakeFiles/dgl.dir/src/graph/nodeflow.cc.o
[ 85%] Building CXX object CMakeFiles/dgl.dir/src/graph/gk_ops.cc.o
[ 87%] Building CXX object CMakeFiles/dgl.dir/src/graph/transform/partition_hetero.cc.o
[ 86%] Building CXX object CMakeFiles/dgl.dir/src/graph/creators.cc.o
[ 87%] Building CXX object CMakeFiles/dgl.dir/src/rpc/network/socket_pool.cc.o
[ 87%] Building CXX object CMakeFiles/dgl.dir/src/api/api_test.cc.o
[ 87%] Building CXX object CMakeFiles/dgl.dir/src/graph/heterograph_capi.cc.o
[ 87%] Building CXX object CMakeFiles/dgl.dir/src/graph/graph_traversal.cc.o
[ 87%] Building CXX object CMakeFiles/dgl.dir/src/graph/heterograph.cc.o
[ 87%] Building CXX object CMakeFiles/dgl.dir/src/graph/immutable_graph.cc.o
[ 89%] Building CXX object CMakeFiles/dgl.dir/src/graph/graph_op.cc.o
[ 89%] Building CXX object CMakeFiles/dgl.dir/src/graph/graph_apis.cc.o
[ 89%] Building CXX object CMakeFiles/dgl.dir/src/graph/sampling/negative/global_uniform.cc.o
[ 89%] Building CXX object CMakeFiles/dgl.dir/src/graph/pickle.cc.o
[ 89%] Building CXX object CMakeFiles/dgl.dir/src/graph/network.cc.o
[ 89%] Building CXX object CMakeFiles/dgl.dir/src/graph/graph.cc.o
[ 89%] Building CXX object CMakeFiles/dgl.dir/src/graph/transform/compact.cc.o
[ 89%] Building CXX object CMakeFiles/dgl.dir/src/rpc/network/socket_communicator.cc.o
[ 90%] Building CXX object CMakeFiles/dgl.dir/src/graph/traversal.cc.o
[ 91%] Building CXX object CMakeFiles/dgl.dir/src/rpc/network/common.cc.o
[ 91%] Building CXX object CMakeFiles/dgl.dir/src/rpc/rpc.cc.o
[ 92%] Building CXX object CMakeFiles/dgl.dir/src/rpc/network/tcp_socket.cc.o
[ 92%] Building CXX object CMakeFiles/dgl.dir/src/array/cuda/csr_transpose.cc.o
[ 92%] Building CXX object CMakeFiles/dgl.dir/src/rpc/tensorpipe/tp_communicator.cc.o
[ 92%] Building CXX object CMakeFiles/dgl.dir/src/runtime/cuda/cuda_device_api.cc.o
In file included from /root/jenkins/workspace/dgl_PR-4648@3/src/geometry/../c_api_common.h:10:0,
                 from /root/jenkins/workspace/dgl_PR-4648@3/src/geometry/geometry.cc:9:
/root/jenkins/workspace/dgl_PR-4648@3/include/dgl/runtime/packed_func.h:502:21: warning: inline function 'TObjectRef dgl::runtime::DGLArgValue::AsObjectRef() const [with TObjectRef = dgl::HeteroGraphRef]' used but never defined
   inline TObjectRef AsObjectRef() const;
                     ^~~~~~~~~~~
/root/jenkins/workspace/dgl_PR-4648@3/src/partition/ndarray_partition.cc: In member function 'virtual int64_t dgl::partition::RangePartition::PartSize(int) const':
/root/jenkins/workspace/dgl_PR-4648@3/src/partition/ndarray_partition.cc:202:3: warning: control reaches end of non-void function [-Wreturn-type]
   }
   ^
In file included from /root/jenkins/workspace/dgl_PR-4648@3/third_party/dmlc-core/include/dmlc/logging.h:132:0,
                 from /root/jenkins/workspace/dgl_PR-4648@3/include/dgl/./runtime/object.h:9,
                 from /root/jenkins/workspace/dgl_PR-4648@3/include/dgl/graph_interface.h:15,
                 from /root/jenkins/workspace/dgl_PR-4648@3/include/dgl/sampler.h:13,
                 from /root/jenkins/workspace/dgl_PR-4648@3/src/graph/sampler.cc:6:
/root/jenkins/workspace/dgl_PR-4648@3/src/graph/sampler.cc: In member function 'dgl::NegSubgraph dgl::{anonymous}::EdgeSamplerObject::genNegEdgeSubgraph(const dgl::Subgraph&, const string&, int64_t, bool, bool)':
/root/jenkins/workspace/dgl_PR-4648@3/src/graph/sampler.cc:1189:48: warning: comparison between signed and unsigned integer expressions [-Wsign-compare]
       assert(prev_neg_offset + neg_sample_size == neg_vids.size());
              ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^~~~~~~~~~~~~~~~~~
/root/jenkins/workspace/dgl_PR-4648@3/src/graph/sampler.cc:1193:48: warning: comparison between signed and unsigned integer expressions [-Wsign-compare]
       assert(prev_neg_offset + neg_sample_size == neg_vids.size());
              ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^~~~~~~~~~~~~~~~~~
/root/jenkins/workspace/dgl_PR-4648@3/src/array/libra_partition.cc: In function 'dgl::runtime::List<dgl::runtime::Value> dgl::aten::Libra2dglBuildDict(dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray, dgl::runtime::NDArray, int32_t, int32_t, int64_t, const string&)':
/root/jenkins/workspace/dgl_PR-4648@3/src/array/libra_partition.cc:396:11: warning: ignoring return value of 'int fscanf(FILE*, const char*, ...)', declared with attribute warn_unused_result [-Wunused-result]
     fscanf(fp, "%ld,%ld,%f\n", &u, &v, &w);   // reading an edge - the src and dst global node IDs
     ~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
/root/jenkins/workspace/dgl_PR-4648@3/src/array/libra_partition.cc: In function 'int32_t dgl::aten::Ver2partition(IdType, int64_t*, int32_t) [with IdType = long int]':
/root/jenkins/workspace/dgl_PR-4648@3/src/array/libra_partition.cc:43:1: warning: control reaches end of non-void function [-Wreturn-type]
 }
 ^
[ 93%] Linking CXX shared library libdgl.so
[ 93%] Built target dgl
[ 95%] Building CXX object CMakeFiles/rpc_server.dir/tests/dist/cpp/rpc_server.cc.o
[ 95%] Building CXX object CMakeFiles/rpc_client.dir/tests/dist/cpp/rpc_client.cc.o
[ 95%] Building CXX object CMakeFiles/runUnitTests.dir/tests/cpp/graph_index_test.cc.o
[ 96%] Building CXX object CMakeFiles/runUnitTests.dir/tests/cpp/socket_communicator_test.cc.o
[ 96%] Building CXX object CMakeFiles/runUnitTests.dir/tests/cpp/string_test.cc.o
[ 96%] Building CXX object CMakeFiles/runUnitTests.dir/tests/cpp/test_aten.cc.o
[ 96%] Building CXX object CMakeFiles/runUnitTests.dir/tests/cpp/test_csrmm.cc.o
[ 96%] Building CXX object CMakeFiles/runUnitTests.dir/tests/cpp/message_queue_test.cc.o
[ 97%] Building CXX object CMakeFiles/runUnitTests.dir/tests/cpp/test_partition.cc.o
[ 97%] Building CXX object CMakeFiles/runUnitTests.dir/tests/cpp/test_rowwise.cc.o
[ 97%] Building CXX object CMakeFiles/runUnitTests.dir/tests/cpp/test_smart_ptr_serialize.cc.o
[ 97%] Building CXX object CMakeFiles/runUnitTests.dir/tests/cpp/test_serialize.cc.o
[ 98%] Building CXX object CMakeFiles/runUnitTests.dir/tests/cpp/test_spmat_coo.cc.o
[ 98%] Building CXX object CMakeFiles/runUnitTests.dir/tests/cpp/test_spmm.cc.o
[100%] Building CXX object CMakeFiles/runUnitTests.dir/tests/cpp/test_zerocopy_serialize.cc.o
[100%] Building CXX object CMakeFiles/runUnitTests.dir/tests/cpp/test_sampler.cc.o
[100%] Building CXX object CMakeFiles/runUnitTests.dir/tests/cpp/test_unit_graph.cc.o
[100%] Building CXX object CMakeFiles/runUnitTests.dir/tests/cpp/test_spmat_csr.cc.o
/root/jenkins/workspace/dgl_PR-4648@3/tests/dist/cpp/rpc_client.cc: In member function 'void RPCClient::StartClient()':
/root/jenkins/workspace/dgl_PR-4648@3/tests/dist/cpp/rpc_client.cc:50:15: warning: unused variable 'num_machines' [-Wunused-variable]
     const int num_machines = ips_.size();
               ^~~~~~~~~~~~
/root/jenkins/workspace/dgl_PR-4648@3/tests/cpp/test_partition.cc: In instantiation of 'void _TestRemainder_MapToX() [with DGLDeviceType XPU = (DGLDeviceType)2; IdType = int]':
/root/jenkins/workspace/dgl_PR-4648@3/tests/cpp/test_partition.cc:86:44:   required from here
/root/jenkins/workspace/dgl_PR-4648@3/tests/cpp/test_partition.cc:66:26: warning: comparison between signed and unsigned integer expressions [-Wsign-compare]
     for (size_t i = 0; i < global->shape[0]; ++i) {
/root/jenkins/workspace/dgl_PR-4648@3/tests/cpp/test_partition.cc:74:26: warning: comparison between signed and unsigned integer expressions [-Wsign-compare]
     for (size_t i = 0; i < act_local->shape[0]; ++i) {
/root/jenkins/workspace/dgl_PR-4648@3/tests/cpp/test_partition.cc: In instantiation of 'void _TestRemainder_MapToX() [with DGLDeviceType XPU = (DGLDeviceType)2; IdType = long int]':
/root/jenkins/workspace/dgl_PR-4648@3/tests/cpp/test_partition.cc:87:44:   required from here
/root/jenkins/workspace/dgl_PR-4648@3/tests/cpp/test_partition.cc:66:26: warning: comparison between signed and unsigned integer expressions [-Wsign-compare]
     for (size_t i = 0; i < global->shape[0]; ++i) {
/root/jenkins/workspace/dgl_PR-4648@3/tests/cpp/test_partition.cc:74:26: warning: comparison between signed and unsigned integer expressions [-Wsign-compare]
     for (size_t i = 0; i < act_local->shape[0]; ++i) {
/root/jenkins/workspace/dgl_PR-4648@3/tests/cpp/test_partition.cc: In instantiation of 'void _TestRange_MapToX() [with DGLDeviceType XPU = (DGLDeviceType)2; IdType = int]':
/root/jenkins/workspace/dgl_PR-4648@3/tests/cpp/test_partition.cc:195:40:   required from here
/root/jenkins/workspace/dgl_PR-4648@3/tests/cpp/test_partition.cc:176:26: warning: comparison between signed and unsigned integer expressions [-Wsign-compare]
     for (size_t i = 0; i < global->shape[0]; ++i) {
/root/jenkins/workspace/dgl_PR-4648@3/tests/cpp/test_partition.cc:184:26: warning: comparison between signed and unsigned integer expressions [-Wsign-compare]
     for (size_t i = 0; i < act_local->shape[0]; ++i) {
/root/jenkins/workspace/dgl_PR-4648@3/tests/cpp/test_partition.cc: In instantiation of 'void _TestRange_MapToX() [with DGLDeviceType XPU = (DGLDeviceType)2; IdType = long int]':
/root/jenkins/workspace/dgl_PR-4648@3/tests/cpp/test_partition.cc:196:40:   required from here
/root/jenkins/workspace/dgl_PR-4648@3/tests/cpp/test_partition.cc:176:26: warning: comparison between signed and unsigned integer expressions [-Wsign-compare]
     for (size_t i = 0; i < global->shape[0]; ++i) {
/root/jenkins/workspace/dgl_PR-4648@3/tests/cpp/test_partition.cc:184:26: warning: comparison between signed and unsigned integer expressions [-Wsign-compare]
     for (size_t i = 0; i < act_local->shape[0]; ++i) {
[100%] Linking CXX executable rpc_client
[100%] Linking CXX executable rpc_server
[100%] Built target rpc_client
[100%] Built target rpc_server
[100%] Linking CXX executable runUnitTests
[100%] Built target runUnitTests
~/jenkins/workspace/dgl_PR-4648@3
~/jenkins/workspace/dgl_PR-4648@3/python ~/jenkins/workspace/dgl_PR-4648@3
WARNING: Skipping dgl as it is not installed.
WARNING: Running pip as the 'root' user can result in broken permissions and conflicting behaviour with the system package manager. It is recommended to use a virtual environment instead: https://pip.pypa.io/warnings/venv
WARNING: Cython is not installed, will compile without cython module
running install
/opt/conda/lib/python3.9/site-packages/setuptools/command/install.py:34: SetuptoolsDeprecationWarning: setup.py install is deprecated. Use build and pip and other standards-based tools.
  warnings.warn(
/opt/conda/lib/python3.9/site-packages/setuptools/command/easy_install.py:144: EasyInstallDeprecationWarning: easy_install command is deprecated. Use build and pip and other standards-based tools.
  warnings.warn(
running bdist_egg
running egg_info
creating dgl.egg-info
writing dgl.egg-info/PKG-INFO
writing dependency_links to dgl.egg-info/dependency_links.txt
writing requirements to dgl.egg-info/requires.txt
writing top-level names to dgl.egg-info/top_level.txt
writing manifest file 'dgl.egg-info/SOURCES.txt'
reading manifest file 'dgl.egg-info/SOURCES.txt'
writing manifest file 'dgl.egg-info/SOURCES.txt'
installing library code to build/bdist.linux-x86_64/egg
running install_lib
running build_py
creating build
creating build/lib.linux-x86_64-cpython-39
creating build/lib.linux-x86_64-cpython-39/dgl
copying dgl/partition.py -> build/lib.linux-x86_64-cpython-39/dgl
copying dgl/core.py -> build/lib.linux-x86_64-cpython-39/dgl
copying dgl/subgraph.py -> build/lib.linux-x86_64-cpython-39/dgl
copying dgl/traversal.py -> build/lib.linux-x86_64-cpython-39/dgl
copying dgl/base.py -> build/lib.linux-x86_64-cpython-39/dgl
copying dgl/heterograph_index.py -> build/lib.linux-x86_64-cpython-39/dgl
copying dgl/sparse.py -> build/lib.linux-x86_64-cpython-39/dgl
copying dgl/global_config.py -> build/lib.linux-x86_64-cpython-39/dgl
copying dgl/_api_internal.py -> build/lib.linux-x86_64-cpython-39/dgl
copying dgl/heterograph.py -> build/lib.linux-x86_64-cpython-39/dgl
copying dgl/network.py -> build/lib.linux-x86_64-cpython-39/dgl
copying dgl/logging.py -> build/lib.linux-x86_64-cpython-39/dgl
copying dgl/graph_index.py -> build/lib.linux-x86_64-cpython-39/dgl
copying dgl/udf.py -> build/lib.linux-x86_64-cpython-39/dgl
copying dgl/ndarray.py -> build/lib.linux-x86_64-cpython-39/dgl
copying dgl/init.py -> build/lib.linux-x86_64-cpython-39/dgl
copying dgl/view.py -> build/lib.linux-x86_64-cpython-39/dgl
copying dgl/generators.py -> build/lib.linux-x86_64-cpython-39/dgl
copying dgl/convert.py -> build/lib.linux-x86_64-cpython-39/dgl
copying dgl/container.py -> build/lib.linux-x86_64-cpython-39/dgl
copying dgl/merge.py -> build/lib.linux-x86_64-cpython-39/dgl
copying dgl/readout.py -> build/lib.linux-x86_64-cpython-39/dgl
copying dgl/__init__.py -> build/lib.linux-x86_64-cpython-39/dgl
copying dgl/propagate.py -> build/lib.linux-x86_64-cpython-39/dgl
copying dgl/batch.py -> build/lib.linux-x86_64-cpython-39/dgl
copying dgl/frame.py -> build/lib.linux-x86_64-cpython-39/dgl
copying dgl/random.py -> build/lib.linux-x86_64-cpython-39/dgl
creating build/lib.linux-x86_64-cpython-39/dgl/mock_sparse
copying dgl/mock_sparse/sp_matrix.py -> build/lib.linux-x86_64-cpython-39/dgl/mock_sparse
copying dgl/mock_sparse/diag_matrix.py -> build/lib.linux-x86_64-cpython-39/dgl/mock_sparse
copying dgl/mock_sparse/reduction.py -> build/lib.linux-x86_64-cpython-39/dgl/mock_sparse
copying dgl/mock_sparse/__init__.py -> build/lib.linux-x86_64-cpython-39/dgl/mock_sparse
copying dgl/mock_sparse/elementwise_op_sp.py -> build/lib.linux-x86_64-cpython-39/dgl/mock_sparse
creating build/lib.linux-x86_64-cpython-39/dgl/geometry
copying dgl/geometry/edge_coarsening.py -> build/lib.linux-x86_64-cpython-39/dgl/geometry
copying dgl/geometry/fps.py -> build/lib.linux-x86_64-cpython-39/dgl/geometry
copying dgl/geometry/capi.py -> build/lib.linux-x86_64-cpython-39/dgl/geometry
copying dgl/geometry/__init__.py -> build/lib.linux-x86_64-cpython-39/dgl/geometry
creating build/lib.linux-x86_64-cpython-39/dgl/function
copying dgl/function/base.py -> build/lib.linux-x86_64-cpython-39/dgl/function
copying dgl/function/message.py -> build/lib.linux-x86_64-cpython-39/dgl/function
copying dgl/function/reducer.py -> build/lib.linux-x86_64-cpython-39/dgl/function
copying dgl/function/__init__.py -> build/lib.linux-x86_64-cpython-39/dgl/function
creating build/lib.linux-x86_64-cpython-39/dgl/nn
copying dgl/nn/__init__.py -> build/lib.linux-x86_64-cpython-39/dgl/nn
creating build/lib.linux-x86_64-cpython-39/dgl/contrib
copying dgl/contrib/dis_kvstore.py -> build/lib.linux-x86_64-cpython-39/dgl/contrib
copying dgl/contrib/unified_tensor.py -> build/lib.linux-x86_64-cpython-39/dgl/contrib
copying dgl/contrib/graph_store.py -> build/lib.linux-x86_64-cpython-39/dgl/contrib
copying dgl/contrib/__init__.py -> build/lib.linux-x86_64-cpython-39/dgl/contrib
creating build/lib.linux-x86_64-cpython-39/dgl/utils
copying dgl/utils/checks.py -> build/lib.linux-x86_64-cpython-39/dgl/utils
copying dgl/utils/exception.py -> build/lib.linux-x86_64-cpython-39/dgl/utils
copying dgl/utils/pin_memory.py -> build/lib.linux-x86_64-cpython-39/dgl/utils
copying dgl/utils/internal.py -> build/lib.linux-x86_64-cpython-39/dgl/utils
copying dgl/utils/filter.py -> build/lib.linux-x86_64-cpython-39/dgl/utils
copying dgl/utils/data.py -> build/lib.linux-x86_64-cpython-39/dgl/utils
copying dgl/utils/__init__.py -> build/lib.linux-x86_64-cpython-39/dgl/utils
copying dgl/utils/shared_mem.py -> build/lib.linux-x86_64-cpython-39/dgl/utils
creating build/lib.linux-x86_64-cpython-39/dgl/multiprocessing
copying dgl/multiprocessing/pytorch.py -> build/lib.linux-x86_64-cpython-39/dgl/multiprocessing
copying dgl/multiprocessing/__init__.py -> build/lib.linux-x86_64-cpython-39/dgl/multiprocessing
creating build/lib.linux-x86_64-cpython-39/dgl/optim
copying dgl/optim/__init__.py -> build/lib.linux-x86_64-cpython-39/dgl/optim
creating build/lib.linux-x86_64-cpython-39/dgl/dataloading
copying dgl/dataloading/base.py -> build/lib.linux-x86_64-cpython-39/dgl/dataloading
copying dgl/dataloading/dataloader.py -> build/lib.linux-x86_64-cpython-39/dgl/dataloading
copying dgl/dataloading/graphsaint.py -> build/lib.linux-x86_64-cpython-39/dgl/dataloading
copying dgl/dataloading/shadow.py -> build/lib.linux-x86_64-cpython-39/dgl/dataloading
copying dgl/dataloading/negative_sampler.py -> build/lib.linux-x86_64-cpython-39/dgl/dataloading
copying dgl/dataloading/__init__.py -> build/lib.linux-x86_64-cpython-39/dgl/dataloading
copying dgl/dataloading/cluster_gcn.py -> build/lib.linux-x86_64-cpython-39/dgl/dataloading
copying dgl/dataloading/neighbor_sampler.py -> build/lib.linux-x86_64-cpython-39/dgl/dataloading
copying dgl/dataloading/dist_dataloader.py -> build/lib.linux-x86_64-cpython-39/dgl/dataloading
creating build/lib.linux-x86_64-cpython-39/dgl/_dataloading
copying dgl/_dataloading/neighbor.py -> build/lib.linux-x86_64-cpython-39/dgl/_dataloading
copying dgl/_dataloading/dataloader.py -> build/lib.linux-x86_64-cpython-39/dgl/_dataloading
copying dgl/_dataloading/shadow.py -> build/lib.linux-x86_64-cpython-39/dgl/_dataloading
copying dgl/_dataloading/negative_sampler.py -> build/lib.linux-x86_64-cpython-39/dgl/_dataloading
copying dgl/_dataloading/__init__.py -> build/lib.linux-x86_64-cpython-39/dgl/_dataloading
copying dgl/_dataloading/cluster_gcn.py -> build/lib.linux-x86_64-cpython-39/dgl/_dataloading
creating build/lib.linux-x86_64-cpython-39/dgl/backend
copying dgl/backend/backend.py -> build/lib.linux-x86_64-cpython-39/dgl/backend
copying dgl/backend/__init__.py -> build/lib.linux-x86_64-cpython-39/dgl/backend
copying dgl/backend/set_default_backend.py -> build/lib.linux-x86_64-cpython-39/dgl/backend
creating build/lib.linux-x86_64-cpython-39/dgl/ops
copying dgl/ops/edge_softmax.py -> build/lib.linux-x86_64-cpython-39/dgl/ops
copying dgl/ops/sddmm.py -> build/lib.linux-x86_64-cpython-39/dgl/ops
copying dgl/ops/gather_mm.py -> build/lib.linux-x86_64-cpython-39/dgl/ops
copying dgl/ops/segment.py -> build/lib.linux-x86_64-cpython-39/dgl/ops
copying dgl/ops/__init__.py -> build/lib.linux-x86_64-cpython-39/dgl/ops
copying dgl/ops/spmm.py -> build/lib.linux-x86_64-cpython-39/dgl/ops
creating build/lib.linux-x86_64-cpython-39/dgl/cuda
copying dgl/cuda/nccl.py -> build/lib.linux-x86_64-cpython-39/dgl/cuda
copying dgl/cuda/__init__.py -> build/lib.linux-x86_64-cpython-39/dgl/cuda
creating build/lib.linux-x86_64-cpython-39/dgl/storages
copying dgl/storages/pytorch_tensor.py -> build/lib.linux-x86_64-cpython-39/dgl/storages
copying dgl/storages/base.py -> build/lib.linux-x86_64-cpython-39/dgl/storages
copying dgl/storages/tensor.py -> build/lib.linux-x86_64-cpython-39/dgl/storages
copying dgl/storages/__init__.py -> build/lib.linux-x86_64-cpython-39/dgl/storages
copying dgl/storages/numpy.py -> build/lib.linux-x86_64-cpython-39/dgl/storages
creating build/lib.linux-x86_64-cpython-39/dgl/sampling
copying dgl/sampling/neighbor.py -> build/lib.linux-x86_64-cpython-39/dgl/sampling
copying dgl/sampling/pinsage.py -> build/lib.linux-x86_64-cpython-39/dgl/sampling
copying dgl/sampling/negative.py -> build/lib.linux-x86_64-cpython-39/dgl/sampling
copying dgl/sampling/randomwalks.py -> build/lib.linux-x86_64-cpython-39/dgl/sampling
copying dgl/sampling/utils.py -> build/lib.linux-x86_64-cpython-39/dgl/sampling
copying dgl/sampling/node2vec_randomwalk.py -> build/lib.linux-x86_64-cpython-39/dgl/sampling
copying dgl/sampling/__init__.py -> build/lib.linux-x86_64-cpython-39/dgl/sampling
creating build/lib.linux-x86_64-cpython-39/dgl/data
copying dgl/data/qm7b.py -> build/lib.linux-x86_64-cpython-39/dgl/data
copying dgl/data/graph_serialize.py -> build/lib.linux-x86_64-cpython-39/dgl/data
copying dgl/data/dgl_dataset.py -> build/lib.linux-x86_64-cpython-39/dgl/data
copying dgl/data/csv_dataset_base.py -> build/lib.linux-x86_64-cpython-39/dgl/data
copying dgl/data/knowledge_graph.py -> build/lib.linux-x86_64-cpython-39/dgl/data
copying dgl/data/qm9_edge.py -> build/lib.linux-x86_64-cpython-39/dgl/data
copying dgl/data/gnn_benchmark.py -> build/lib.linux-x86_64-cpython-39/dgl/data
copying dgl/data/qm9.py -> build/lib.linux-x86_64-cpython-39/dgl/data
copying dgl/data/icews18.py -> build/lib.linux-x86_64-cpython-39/dgl/data
copying dgl/data/gdelt.py -> build/lib.linux-x86_64-cpython-39/dgl/data
copying dgl/data/utils.py -> build/lib.linux-x86_64-cpython-39/dgl/data
copying dgl/data/sbm.py -> build/lib.linux-x86_64-cpython-39/dgl/data
copying dgl/data/csv_dataset.py -> build/lib.linux-x86_64-cpython-39/dgl/data
copying dgl/data/karate.py -> build/lib.linux-x86_64-cpython-39/dgl/data
copying dgl/data/citation_graph.py -> build/lib.linux-x86_64-cpython-39/dgl/data
copying dgl/data/rdf.py -> build/lib.linux-x86_64-cpython-39/dgl/data
copying dgl/data/bitcoinotc.py -> build/lib.linux-x86_64-cpython-39/dgl/data
copying dgl/data/ppi.py -> build/lib.linux-x86_64-cpython-39/dgl/data
copying dgl/data/tensor_serialize.py -> build/lib.linux-x86_64-cpython-39/dgl/data
copying dgl/data/flickr.py -> build/lib.linux-x86_64-cpython-39/dgl/data
copying dgl/data/wikics.py -> build/lib.linux-x86_64-cpython-39/dgl/data
copying dgl/data/heterograph_serialize.py -> build/lib.linux-x86_64-cpython-39/dgl/data
copying dgl/data/tu.py -> build/lib.linux-x86_64-cpython-39/dgl/data
copying dgl/data/__init__.py -> build/lib.linux-x86_64-cpython-39/dgl/data
copying dgl/data/gindt.py -> build/lib.linux-x86_64-cpython-39/dgl/data
copying dgl/data/tree.py -> build/lib.linux-x86_64-cpython-39/dgl/data
copying dgl/data/synthetic.py -> build/lib.linux-x86_64-cpython-39/dgl/data
copying dgl/data/adapter.py -> build/lib.linux-x86_64-cpython-39/dgl/data
copying dgl/data/fakenews.py -> build/lib.linux-x86_64-cpython-39/dgl/data
copying dgl/data/fraud.py -> build/lib.linux-x86_64-cpython-39/dgl/data
copying dgl/data/minigc.py -> build/lib.linux-x86_64-cpython-39/dgl/data
copying dgl/data/yelp.py -> build/lib.linux-x86_64-cpython-39/dgl/data
copying dgl/data/reddit.py -> build/lib.linux-x86_64-cpython-39/dgl/data
creating build/lib.linux-x86_64-cpython-39/dgl/distributed
copying dgl/distributed/dist_graph.py -> build/lib.linux-x86_64-cpython-39/dgl/distributed
copying dgl/distributed/graph_partition_book.py -> build/lib.linux-x86_64-cpython-39/dgl/distributed
copying dgl/distributed/partition.py -> build/lib.linux-x86_64-cpython-39/dgl/distributed
copying dgl/distributed/standalone_kvstore.py -> build/lib.linux-x86_64-cpython-39/dgl/distributed
copying dgl/distributed/graph_services.py -> build/lib.linux-x86_64-cpython-39/dgl/distributed
copying dgl/distributed/role.py -> build/lib.linux-x86_64-cpython-39/dgl/distributed
copying dgl/distributed/constants.py -> build/lib.linux-x86_64-cpython-39/dgl/distributed
copying dgl/distributed/rpc_server.py -> build/lib.linux-x86_64-cpython-39/dgl/distributed
copying dgl/distributed/kvstore.py -> build/lib.linux-x86_64-cpython-39/dgl/distributed
copying dgl/distributed/dist_context.py -> build/lib.linux-x86_64-cpython-39/dgl/distributed
copying dgl/distributed/rpc.py -> build/lib.linux-x86_64-cpython-39/dgl/distributed
copying dgl/distributed/rpc_client.py -> build/lib.linux-x86_64-cpython-39/dgl/distributed
copying dgl/distributed/id_map.py -> build/lib.linux-x86_64-cpython-39/dgl/distributed
copying dgl/distributed/server_state.py -> build/lib.linux-x86_64-cpython-39/dgl/distributed
copying dgl/distributed/__init__.py -> build/lib.linux-x86_64-cpython-39/dgl/distributed
copying dgl/distributed/shared_mem_utils.py -> build/lib.linux-x86_64-cpython-39/dgl/distributed
copying dgl/distributed/dist_tensor.py -> build/lib.linux-x86_64-cpython-39/dgl/distributed
copying dgl/distributed/dist_dataloader.py -> build/lib.linux-x86_64-cpython-39/dgl/distributed
creating build/lib.linux-x86_64-cpython-39/dgl/_deprecate
copying dgl/_deprecate/nodeflow.py -> build/lib.linux-x86_64-cpython-39/dgl/_deprecate
copying dgl/_deprecate/udf.py -> build/lib.linux-x86_64-cpython-39/dgl/_deprecate
copying dgl/_deprecate/view.py -> build/lib.linux-x86_64-cpython-39/dgl/_deprecate
copying dgl/_deprecate/kernel.py -> build/lib.linux-x86_64-cpython-39/dgl/_deprecate
copying dgl/_deprecate/__init__.py -> build/lib.linux-x86_64-cpython-39/dgl/_deprecate
copying dgl/_deprecate/graph.py -> build/lib.linux-x86_64-cpython-39/dgl/_deprecate
copying dgl/_deprecate/frame.py -> build/lib.linux-x86_64-cpython-39/dgl/_deprecate
creating build/lib.linux-x86_64-cpython-39/dgl/transforms
copying dgl/transforms/functional.py -> build/lib.linux-x86_64-cpython-39/dgl/transforms
copying dgl/transforms/__init__.py -> build/lib.linux-x86_64-cpython-39/dgl/transforms
copying dgl/transforms/module.py -> build/lib.linux-x86_64-cpython-39/dgl/transforms
creating build/lib.linux-x86_64-cpython-39/dgl/distgnn
copying dgl/distgnn/__init__.py -> build/lib.linux-x86_64-cpython-39/dgl/distgnn
creating build/lib.linux-x86_64-cpython-39/dgl/_ffi
copying dgl/_ffi/runtime_ctypes.py -> build/lib.linux-x86_64-cpython-39/dgl/_ffi
copying dgl/_ffi/function.py -> build/lib.linux-x86_64-cpython-39/dgl/_ffi
copying dgl/_ffi/libinfo.py -> build/lib.linux-x86_64-cpython-39/dgl/_ffi
copying dgl/_ffi/streams.py -> build/lib.linux-x86_64-cpython-39/dgl/_ffi
copying dgl/_ffi/base.py -> build/lib.linux-x86_64-cpython-39/dgl/_ffi
copying dgl/_ffi/ndarray.py -> build/lib.linux-x86_64-cpython-39/dgl/_ffi
copying dgl/_ffi/object_generic.py -> build/lib.linux-x86_64-cpython-39/dgl/_ffi
copying dgl/_ffi/__init__.py -> build/lib.linux-x86_64-cpython-39/dgl/_ffi
copying dgl/_ffi/object.py -> build/lib.linux-x86_64-cpython-39/dgl/_ffi
creating build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch
copying dgl/nn/pytorch/glob.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch
copying dgl/nn/pytorch/hetero.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch
copying dgl/nn/pytorch/sparse_emb.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch
copying dgl/nn/pytorch/linear.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch
copying dgl/nn/pytorch/utils.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch
copying dgl/nn/pytorch/__init__.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch
copying dgl/nn/pytorch/factory.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch
copying dgl/nn/pytorch/softmax.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch
creating build/lib.linux-x86_64-cpython-39/dgl/nn/mxnet
copying dgl/nn/mxnet/glob.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/mxnet
copying dgl/nn/mxnet/hetero.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/mxnet
copying dgl/nn/mxnet/utils.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/mxnet
copying dgl/nn/mxnet/__init__.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/mxnet
copying dgl/nn/mxnet/softmax.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/mxnet
creating build/lib.linux-x86_64-cpython-39/dgl/nn/tensorflow
copying dgl/nn/tensorflow/glob.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/tensorflow
copying dgl/nn/tensorflow/hetero.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/tensorflow
copying dgl/nn/tensorflow/utils.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/tensorflow
copying dgl/nn/tensorflow/__init__.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/tensorflow
copying dgl/nn/tensorflow/softmax.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/tensorflow
creating build/lib.linux-x86_64-cpython-39/dgl/nn/functional
copying dgl/nn/functional/__init__.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/functional
creating build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/link
copying dgl/nn/pytorch/link/transe.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/link
copying dgl/nn/pytorch/link/edgepred.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/link
copying dgl/nn/pytorch/link/transr.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/link
copying dgl/nn/pytorch/link/__init__.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/link
creating build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/agnnconv.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/relgraphconv.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/appnpconv.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/dgnconv.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/grouprevres.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/twirlsconv.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/egatconv.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/graphconv.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/densechebconv.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/gcn2conv.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/ginconv.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/dotgatconv.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/cfconv.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/gmmconv.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/sageconv.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/densesageconv.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/gineconv.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/edgeconv.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/nnconv.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/__init__.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/gatv2conv.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/hgtconv.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/pnaconv.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/gatedgraphconv.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/atomicconv.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/tagconv.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/densegraphconv.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/egnnconv.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/chebconv.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/sgconv.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/conv
copying dgl/nn/pytorch/conv/gatconv.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/conv
creating build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/explain
copying dgl/nn/pytorch/explain/gnnexplainer.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/explain
copying dgl/nn/pytorch/explain/__init__.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/explain
creating build/lib.linux-x86_64-cpython-39/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/agnnconv.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/relgraphconv.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/appnpconv.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/graphconv.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/densechebconv.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/ginconv.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/gmmconv.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/sageconv.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/densesageconv.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/edgeconv.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/nnconv.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/__init__.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/gatedgraphconv.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/tagconv.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/densegraphconv.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/chebconv.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/sgconv.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/mxnet/conv
copying dgl/nn/mxnet/conv/gatconv.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/mxnet/conv
creating build/lib.linux-x86_64-cpython-39/dgl/nn/tensorflow/conv
copying dgl/nn/tensorflow/conv/relgraphconv.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/tensorflow/conv
copying dgl/nn/tensorflow/conv/appnpconv.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/tensorflow/conv
copying dgl/nn/tensorflow/conv/graphconv.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/tensorflow/conv
copying dgl/nn/tensorflow/conv/densechebconv.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/tensorflow/conv
copying dgl/nn/tensorflow/conv/ginconv.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/tensorflow/conv
copying dgl/nn/tensorflow/conv/sageconv.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/tensorflow/conv
copying dgl/nn/tensorflow/conv/edgeconv.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/tensorflow/conv
copying dgl/nn/tensorflow/conv/__init__.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/tensorflow/conv
copying dgl/nn/tensorflow/conv/chebconv.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/tensorflow/conv
copying dgl/nn/tensorflow/conv/sgconv.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/tensorflow/conv
copying dgl/nn/tensorflow/conv/gatconv.py -> build/lib.linux-x86_64-cpython-39/dgl/nn/tensorflow/conv
creating build/lib.linux-x86_64-cpython-39/dgl/contrib/sampling
copying dgl/contrib/sampling/sampler.py -> build/lib.linux-x86_64-cpython-39/dgl/contrib/sampling
copying dgl/contrib/sampling/dis_sampler.py -> build/lib.linux-x86_64-cpython-39/dgl/contrib/sampling
copying dgl/contrib/sampling/__init__.py -> build/lib.linux-x86_64-cpython-39/dgl/contrib/sampling
creating build/lib.linux-x86_64-cpython-39/dgl/contrib/data
copying dgl/contrib/data/knowledge_graph.py -> build/lib.linux-x86_64-cpython-39/dgl/contrib/data
copying dgl/contrib/data/__init__.py -> build/lib.linux-x86_64-cpython-39/dgl/contrib/data
creating build/lib.linux-x86_64-cpython-39/dgl/optim/pytorch
copying dgl/optim/pytorch/__init__.py -> build/lib.linux-x86_64-cpython-39/dgl/optim/pytorch
copying dgl/optim/pytorch/sparse_optim.py -> build/lib.linux-x86_64-cpython-39/dgl/optim/pytorch
creating build/lib.linux-x86_64-cpython-39/dgl/optim/mxnet
copying dgl/optim/mxnet/__init__.py -> build/lib.linux-x86_64-cpython-39/dgl/optim/mxnet
creating build/lib.linux-x86_64-cpython-39/dgl/optim/tensorflow
copying dgl/optim/tensorflow/__init__.py -> build/lib.linux-x86_64-cpython-39/dgl/optim/tensorflow
creating build/lib.linux-x86_64-cpython-39/dgl/_dataloading/pytorch
copying dgl/_dataloading/pytorch/dataloader.py -> build/lib.linux-x86_64-cpython-39/dgl/_dataloading/pytorch
copying dgl/_dataloading/pytorch/__init__.py -> build/lib.linux-x86_64-cpython-39/dgl/_dataloading/pytorch
creating build/lib.linux-x86_64-cpython-39/dgl/backend/pytorch
copying dgl/backend/pytorch/sparse.py -> build/lib.linux-x86_64-cpython-39/dgl/backend/pytorch
copying dgl/backend/pytorch/tensor.py -> build/lib.linux-x86_64-cpython-39/dgl/backend/pytorch
copying dgl/backend/pytorch/__init__.py -> build/lib.linux-x86_64-cpython-39/dgl/backend/pytorch
creating build/lib.linux-x86_64-cpython-39/dgl/backend/mxnet
copying dgl/backend/mxnet/sparse.py -> build/lib.linux-x86_64-cpython-39/dgl/backend/mxnet
copying dgl/backend/mxnet/tensor.py -> build/lib.linux-x86_64-cpython-39/dgl/backend/mxnet
copying dgl/backend/mxnet/__init__.py -> build/lib.linux-x86_64-cpython-39/dgl/backend/mxnet
copying dgl/backend/mxnet/sparse_optim.py -> build/lib.linux-x86_64-cpython-39/dgl/backend/mxnet
creating build/lib.linux-x86_64-cpython-39/dgl/backend/tensorflow
copying dgl/backend/tensorflow/sparse.py -> build/lib.linux-x86_64-cpython-39/dgl/backend/tensorflow
copying dgl/backend/tensorflow/tensor.py -> build/lib.linux-x86_64-cpython-39/dgl/backend/tensorflow
copying dgl/backend/tensorflow/__init__.py -> build/lib.linux-x86_64-cpython-39/dgl/backend/tensorflow
copying dgl/backend/tensorflow/sparse_optim.py -> build/lib.linux-x86_64-cpython-39/dgl/backend/tensorflow
creating build/lib.linux-x86_64-cpython-39/dgl/distributed/nn
copying dgl/distributed/nn/__init__.py -> build/lib.linux-x86_64-cpython-39/dgl/distributed/nn
creating build/lib.linux-x86_64-cpython-39/dgl/distributed/optim
copying dgl/distributed/optim/__init__.py -> build/lib.linux-x86_64-cpython-39/dgl/distributed/optim
creating build/lib.linux-x86_64-cpython-39/dgl/distributed/nn/pytorch
copying dgl/distributed/nn/pytorch/sparse_emb.py -> build/lib.linux-x86_64-cpython-39/dgl/distributed/nn/pytorch
copying dgl/distributed/nn/pytorch/__init__.py -> build/lib.linux-x86_64-cpython-39/dgl/distributed/nn/pytorch
creating build/lib.linux-x86_64-cpython-39/dgl/distributed/nn/mxnet
copying dgl/distributed/nn/mxnet/__init__.py -> build/lib.linux-x86_64-cpython-39/dgl/distributed/nn/mxnet
creating build/lib.linux-x86_64-cpython-39/dgl/distributed/nn/tensorflow
copying dgl/distributed/nn/tensorflow/__init__.py -> build/lib.linux-x86_64-cpython-39/dgl/distributed/nn/tensorflow
creating build/lib.linux-x86_64-cpython-39/dgl/distributed/optim/pytorch
copying dgl/distributed/optim/pytorch/utils.py -> build/lib.linux-x86_64-cpython-39/dgl/distributed/optim/pytorch
copying dgl/distributed/optim/pytorch/__init__.py -> build/lib.linux-x86_64-cpython-39/dgl/distributed/optim/pytorch
copying dgl/distributed/optim/pytorch/sparse_optim.py -> build/lib.linux-x86_64-cpython-39/dgl/distributed/optim/pytorch
creating build/lib.linux-x86_64-cpython-39/dgl/distributed/optim/mxnet
copying dgl/distributed/optim/mxnet/__init__.py -> build/lib.linux-x86_64-cpython-39/dgl/distributed/optim/mxnet
creating build/lib.linux-x86_64-cpython-39/dgl/distributed/optim/tensorflow
copying dgl/distributed/optim/tensorflow/__init__.py -> build/lib.linux-x86_64-cpython-39/dgl/distributed/optim/tensorflow
creating build/lib.linux-x86_64-cpython-39/dgl/_deprecate/runtime
copying dgl/_deprecate/runtime/spmv.py -> build/lib.linux-x86_64-cpython-39/dgl/_deprecate/runtime
copying dgl/_deprecate/runtime/degree_bucketing.py -> build/lib.linux-x86_64-cpython-39/dgl/_deprecate/runtime
copying dgl/_deprecate/runtime/__init__.py -> build/lib.linux-x86_64-cpython-39/dgl/_deprecate/runtime
copying dgl/_deprecate/runtime/scheduler.py -> build/lib.linux-x86_64-cpython-39/dgl/_deprecate/runtime
copying dgl/_deprecate/runtime/adapter.py -> build/lib.linux-x86_64-cpython-39/dgl/_deprecate/runtime
copying dgl/_deprecate/runtime/runtime.py -> build/lib.linux-x86_64-cpython-39/dgl/_deprecate/runtime
creating build/lib.linux-x86_64-cpython-39/dgl/_deprecate/runtime/ir
copying dgl/_deprecate/runtime/ir/program.py -> build/lib.linux-x86_64-cpython-39/dgl/_deprecate/runtime/ir
copying dgl/_deprecate/runtime/ir/executor.py -> build/lib.linux-x86_64-cpython-39/dgl/_deprecate/runtime/ir
copying dgl/_deprecate/runtime/ir/var.py -> build/lib.linux-x86_64-cpython-39/dgl/_deprecate/runtime/ir
copying dgl/_deprecate/runtime/ir/registry.py -> build/lib.linux-x86_64-cpython-39/dgl/_deprecate/runtime/ir
copying dgl/_deprecate/runtime/ir/__init__.py -> build/lib.linux-x86_64-cpython-39/dgl/_deprecate/runtime/ir
creating build/lib.linux-x86_64-cpython-39/dgl/distgnn/partition
copying dgl/distgnn/partition/libra_partition.py -> build/lib.linux-x86_64-cpython-39/dgl/distgnn/partition
copying dgl/distgnn/partition/__init__.py -> build/lib.linux-x86_64-cpython-39/dgl/distgnn/partition
creating build/lib.linux-x86_64-cpython-39/dgl/distgnn/tools
copying dgl/distgnn/tools/__init__.py -> build/lib.linux-x86_64-cpython-39/dgl/distgnn/tools
copying dgl/distgnn/tools/tools.py -> build/lib.linux-x86_64-cpython-39/dgl/distgnn/tools
creating build/lib.linux-x86_64-cpython-39/dgl/_ffi/_ctypes
copying dgl/_ffi/_ctypes/function.py -> build/lib.linux-x86_64-cpython-39/dgl/_ffi/_ctypes
copying dgl/_ffi/_ctypes/types.py -> build/lib.linux-x86_64-cpython-39/dgl/_ffi/_ctypes
copying dgl/_ffi/_ctypes/ndarray.py -> build/lib.linux-x86_64-cpython-39/dgl/_ffi/_ctypes
copying dgl/_ffi/_ctypes/__init__.py -> build/lib.linux-x86_64-cpython-39/dgl/_ffi/_ctypes
copying dgl/_ffi/_ctypes/object.py -> build/lib.linux-x86_64-cpython-39/dgl/_ffi/_ctypes
creating build/lib.linux-x86_64-cpython-39/dgl/_ffi/_cy3
copying dgl/_ffi/_cy3/__init__.py -> build/lib.linux-x86_64-cpython-39/dgl/_ffi/_cy3
creating build/lib.linux-x86_64-cpython-39/dgl/_ffi/_cy2
copying dgl/_ffi/_cy2/__init__.py -> build/lib.linux-x86_64-cpython-39/dgl/_ffi/_cy2
running build_ext
creating build/bdist.linux-x86_64
creating build/bdist.linux-x86_64/egg
creating build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/mock_sparse
copying build/lib.linux-x86_64-cpython-39/dgl/mock_sparse/sp_matrix.py -> build/bdist.linux-x86_64/egg/dgl/mock_sparse
copying build/lib.linux-x86_64-cpython-39/dgl/mock_sparse/diag_matrix.py -> build/bdist.linux-x86_64/egg/dgl/mock_sparse
copying build/lib.linux-x86_64-cpython-39/dgl/mock_sparse/reduction.py -> build/bdist.linux-x86_64/egg/dgl/mock_sparse
copying build/lib.linux-x86_64-cpython-39/dgl/mock_sparse/__init__.py -> build/bdist.linux-x86_64/egg/dgl/mock_sparse
copying build/lib.linux-x86_64-cpython-39/dgl/mock_sparse/elementwise_op_sp.py -> build/bdist.linux-x86_64/egg/dgl/mock_sparse
creating build/bdist.linux-x86_64/egg/dgl/geometry
copying build/lib.linux-x86_64-cpython-39/dgl/geometry/edge_coarsening.py -> build/bdist.linux-x86_64/egg/dgl/geometry
copying build/lib.linux-x86_64-cpython-39/dgl/geometry/fps.py -> build/bdist.linux-x86_64/egg/dgl/geometry
copying build/lib.linux-x86_64-cpython-39/dgl/geometry/capi.py -> build/bdist.linux-x86_64/egg/dgl/geometry
copying build/lib.linux-x86_64-cpython-39/dgl/geometry/__init__.py -> build/bdist.linux-x86_64/egg/dgl/geometry
creating build/bdist.linux-x86_64/egg/dgl/function
copying build/lib.linux-x86_64-cpython-39/dgl/function/base.py -> build/bdist.linux-x86_64/egg/dgl/function
copying build/lib.linux-x86_64-cpython-39/dgl/function/message.py -> build/bdist.linux-x86_64/egg/dgl/function
copying build/lib.linux-x86_64-cpython-39/dgl/function/reducer.py -> build/bdist.linux-x86_64/egg/dgl/function
copying build/lib.linux-x86_64-cpython-39/dgl/function/__init__.py -> build/bdist.linux-x86_64/egg/dgl/function
copying build/lib.linux-x86_64-cpython-39/dgl/partition.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/nn
creating build/bdist.linux-x86_64/egg/dgl/nn/pytorch
copying build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/glob.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch
creating build/bdist.linux-x86_64/egg/dgl/nn/pytorch/link
copying build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/link/transe.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/link
copying build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/link/edgepred.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/link
copying build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/link/transr.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/link
copying build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/link/__init__.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/link
copying build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/hetero.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch
copying build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/sparse_emb.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch
copying build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/linear.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch
copying build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/utils.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch
creating build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/conv/agnnconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/conv/relgraphconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/conv/appnpconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/conv/dgnconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/conv/grouprevres.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/conv/twirlsconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/conv/egatconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/conv/graphconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/conv/densechebconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/conv/gcn2conv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/conv/ginconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/conv/dotgatconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/conv/cfconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/conv/gmmconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/conv/sageconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/conv/densesageconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/conv/gineconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/conv/edgeconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/conv/nnconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/conv/__init__.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/conv/gatv2conv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/conv/hgtconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/conv/pnaconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/conv/gatedgraphconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/conv/atomicconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/conv/tagconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/conv/densegraphconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/conv/egnnconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/conv/chebconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/conv/sgconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/conv/gatconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv
copying build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/__init__.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch
creating build/bdist.linux-x86_64/egg/dgl/nn/pytorch/explain
copying build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/explain/gnnexplainer.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/explain
copying build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/explain/__init__.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch/explain
copying build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/factory.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch
copying build/lib.linux-x86_64-cpython-39/dgl/nn/pytorch/softmax.py -> build/bdist.linux-x86_64/egg/dgl/nn/pytorch
creating build/bdist.linux-x86_64/egg/dgl/nn/mxnet
copying build/lib.linux-x86_64-cpython-39/dgl/nn/mxnet/glob.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet
copying build/lib.linux-x86_64-cpython-39/dgl/nn/mxnet/hetero.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet
copying build/lib.linux-x86_64-cpython-39/dgl/nn/mxnet/utils.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet
creating build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-cpython-39/dgl/nn/mxnet/conv/agnnconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-cpython-39/dgl/nn/mxnet/conv/relgraphconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-cpython-39/dgl/nn/mxnet/conv/appnpconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-cpython-39/dgl/nn/mxnet/conv/graphconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-cpython-39/dgl/nn/mxnet/conv/densechebconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-cpython-39/dgl/nn/mxnet/conv/ginconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-cpython-39/dgl/nn/mxnet/conv/gmmconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-cpython-39/dgl/nn/mxnet/conv/sageconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-cpython-39/dgl/nn/mxnet/conv/densesageconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-cpython-39/dgl/nn/mxnet/conv/edgeconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-cpython-39/dgl/nn/mxnet/conv/nnconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-cpython-39/dgl/nn/mxnet/conv/__init__.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-cpython-39/dgl/nn/mxnet/conv/gatedgraphconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-cpython-39/dgl/nn/mxnet/conv/tagconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-cpython-39/dgl/nn/mxnet/conv/densegraphconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-cpython-39/dgl/nn/mxnet/conv/chebconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-cpython-39/dgl/nn/mxnet/conv/sgconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-cpython-39/dgl/nn/mxnet/conv/gatconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv
copying build/lib.linux-x86_64-cpython-39/dgl/nn/mxnet/__init__.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet
copying build/lib.linux-x86_64-cpython-39/dgl/nn/mxnet/softmax.py -> build/bdist.linux-x86_64/egg/dgl/nn/mxnet
copying build/lib.linux-x86_64-cpython-39/dgl/nn/__init__.py -> build/bdist.linux-x86_64/egg/dgl/nn
creating build/bdist.linux-x86_64/egg/dgl/nn/tensorflow
copying build/lib.linux-x86_64-cpython-39/dgl/nn/tensorflow/glob.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow
copying build/lib.linux-x86_64-cpython-39/dgl/nn/tensorflow/hetero.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow
copying build/lib.linux-x86_64-cpython-39/dgl/nn/tensorflow/utils.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow
creating build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-cpython-39/dgl/nn/tensorflow/conv/relgraphconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-cpython-39/dgl/nn/tensorflow/conv/appnpconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-cpython-39/dgl/nn/tensorflow/conv/graphconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-cpython-39/dgl/nn/tensorflow/conv/densechebconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-cpython-39/dgl/nn/tensorflow/conv/ginconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-cpython-39/dgl/nn/tensorflow/conv/sageconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-cpython-39/dgl/nn/tensorflow/conv/edgeconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-cpython-39/dgl/nn/tensorflow/conv/__init__.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-cpython-39/dgl/nn/tensorflow/conv/chebconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-cpython-39/dgl/nn/tensorflow/conv/sgconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-cpython-39/dgl/nn/tensorflow/conv/gatconv.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv
copying build/lib.linux-x86_64-cpython-39/dgl/nn/tensorflow/__init__.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow
copying build/lib.linux-x86_64-cpython-39/dgl/nn/tensorflow/softmax.py -> build/bdist.linux-x86_64/egg/dgl/nn/tensorflow
creating build/bdist.linux-x86_64/egg/dgl/nn/functional
copying build/lib.linux-x86_64-cpython-39/dgl/nn/functional/__init__.py -> build/bdist.linux-x86_64/egg/dgl/nn/functional
copying build/lib.linux-x86_64-cpython-39/dgl/core.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-cpython-39/dgl/subgraph.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/contrib
copying build/lib.linux-x86_64-cpython-39/dgl/contrib/dis_kvstore.py -> build/bdist.linux-x86_64/egg/dgl/contrib
copying build/lib.linux-x86_64-cpython-39/dgl/contrib/unified_tensor.py -> build/bdist.linux-x86_64/egg/dgl/contrib
copying build/lib.linux-x86_64-cpython-39/dgl/contrib/graph_store.py -> build/bdist.linux-x86_64/egg/dgl/contrib
creating build/bdist.linux-x86_64/egg/dgl/contrib/sampling
copying build/lib.linux-x86_64-cpython-39/dgl/contrib/sampling/sampler.py -> build/bdist.linux-x86_64/egg/dgl/contrib/sampling
copying build/lib.linux-x86_64-cpython-39/dgl/contrib/sampling/dis_sampler.py -> build/bdist.linux-x86_64/egg/dgl/contrib/sampling
copying build/lib.linux-x86_64-cpython-39/dgl/contrib/sampling/__init__.py -> build/bdist.linux-x86_64/egg/dgl/contrib/sampling
creating build/bdist.linux-x86_64/egg/dgl/contrib/data
copying build/lib.linux-x86_64-cpython-39/dgl/contrib/data/knowledge_graph.py -> build/bdist.linux-x86_64/egg/dgl/contrib/data
copying build/lib.linux-x86_64-cpython-39/dgl/contrib/data/__init__.py -> build/bdist.linux-x86_64/egg/dgl/contrib/data
copying build/lib.linux-x86_64-cpython-39/dgl/contrib/__init__.py -> build/bdist.linux-x86_64/egg/dgl/contrib
copying build/lib.linux-x86_64-cpython-39/dgl/traversal.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/utils
copying build/lib.linux-x86_64-cpython-39/dgl/utils/checks.py -> build/bdist.linux-x86_64/egg/dgl/utils
copying build/lib.linux-x86_64-cpython-39/dgl/utils/exception.py -> build/bdist.linux-x86_64/egg/dgl/utils
copying build/lib.linux-x86_64-cpython-39/dgl/utils/pin_memory.py -> build/bdist.linux-x86_64/egg/dgl/utils
copying build/lib.linux-x86_64-cpython-39/dgl/utils/internal.py -> build/bdist.linux-x86_64/egg/dgl/utils
copying build/lib.linux-x86_64-cpython-39/dgl/utils/filter.py -> build/bdist.linux-x86_64/egg/dgl/utils
copying build/lib.linux-x86_64-cpython-39/dgl/utils/data.py -> build/bdist.linux-x86_64/egg/dgl/utils
copying build/lib.linux-x86_64-cpython-39/dgl/utils/__init__.py -> build/bdist.linux-x86_64/egg/dgl/utils
copying build/lib.linux-x86_64-cpython-39/dgl/utils/shared_mem.py -> build/bdist.linux-x86_64/egg/dgl/utils
creating build/bdist.linux-x86_64/egg/dgl/multiprocessing
copying build/lib.linux-x86_64-cpython-39/dgl/multiprocessing/pytorch.py -> build/bdist.linux-x86_64/egg/dgl/multiprocessing
copying build/lib.linux-x86_64-cpython-39/dgl/multiprocessing/__init__.py -> build/bdist.linux-x86_64/egg/dgl/multiprocessing
copying build/lib.linux-x86_64-cpython-39/dgl/base.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/optim
creating build/bdist.linux-x86_64/egg/dgl/optim/pytorch
copying build/lib.linux-x86_64-cpython-39/dgl/optim/pytorch/__init__.py -> build/bdist.linux-x86_64/egg/dgl/optim/pytorch
copying build/lib.linux-x86_64-cpython-39/dgl/optim/pytorch/sparse_optim.py -> build/bdist.linux-x86_64/egg/dgl/optim/pytorch
creating build/bdist.linux-x86_64/egg/dgl/optim/mxnet
copying build/lib.linux-x86_64-cpython-39/dgl/optim/mxnet/__init__.py -> build/bdist.linux-x86_64/egg/dgl/optim/mxnet
copying build/lib.linux-x86_64-cpython-39/dgl/optim/__init__.py -> build/bdist.linux-x86_64/egg/dgl/optim
creating build/bdist.linux-x86_64/egg/dgl/optim/tensorflow
copying build/lib.linux-x86_64-cpython-39/dgl/optim/tensorflow/__init__.py -> build/bdist.linux-x86_64/egg/dgl/optim/tensorflow
copying build/lib.linux-x86_64-cpython-39/dgl/heterograph_index.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-cpython-39/dgl/sparse.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-cpython-39/dgl/global_config.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/dataloading
copying build/lib.linux-x86_64-cpython-39/dgl/dataloading/base.py -> build/bdist.linux-x86_64/egg/dgl/dataloading
copying build/lib.linux-x86_64-cpython-39/dgl/dataloading/dataloader.py -> build/bdist.linux-x86_64/egg/dgl/dataloading
copying build/lib.linux-x86_64-cpython-39/dgl/dataloading/graphsaint.py -> build/bdist.linux-x86_64/egg/dgl/dataloading
copying build/lib.linux-x86_64-cpython-39/dgl/dataloading/shadow.py -> build/bdist.linux-x86_64/egg/dgl/dataloading
copying build/lib.linux-x86_64-cpython-39/dgl/dataloading/negative_sampler.py -> build/bdist.linux-x86_64/egg/dgl/dataloading
copying build/lib.linux-x86_64-cpython-39/dgl/dataloading/__init__.py -> build/bdist.linux-x86_64/egg/dgl/dataloading
copying build/lib.linux-x86_64-cpython-39/dgl/dataloading/cluster_gcn.py -> build/bdist.linux-x86_64/egg/dgl/dataloading
copying build/lib.linux-x86_64-cpython-39/dgl/dataloading/neighbor_sampler.py -> build/bdist.linux-x86_64/egg/dgl/dataloading
copying build/lib.linux-x86_64-cpython-39/dgl/dataloading/dist_dataloader.py -> build/bdist.linux-x86_64/egg/dgl/dataloading
creating build/bdist.linux-x86_64/egg/dgl/_dataloading
copying build/lib.linux-x86_64-cpython-39/dgl/_dataloading/neighbor.py -> build/bdist.linux-x86_64/egg/dgl/_dataloading
copying build/lib.linux-x86_64-cpython-39/dgl/_dataloading/dataloader.py -> build/bdist.linux-x86_64/egg/dgl/_dataloading
creating build/bdist.linux-x86_64/egg/dgl/_dataloading/pytorch
copying build/lib.linux-x86_64-cpython-39/dgl/_dataloading/pytorch/dataloader.py -> build/bdist.linux-x86_64/egg/dgl/_dataloading/pytorch
copying build/lib.linux-x86_64-cpython-39/dgl/_dataloading/pytorch/__init__.py -> build/bdist.linux-x86_64/egg/dgl/_dataloading/pytorch
copying build/lib.linux-x86_64-cpython-39/dgl/_dataloading/shadow.py -> build/bdist.linux-x86_64/egg/dgl/_dataloading
copying build/lib.linux-x86_64-cpython-39/dgl/_dataloading/negative_sampler.py -> build/bdist.linux-x86_64/egg/dgl/_dataloading
copying build/lib.linux-x86_64-cpython-39/dgl/_dataloading/__init__.py -> build/bdist.linux-x86_64/egg/dgl/_dataloading
copying build/lib.linux-x86_64-cpython-39/dgl/_dataloading/cluster_gcn.py -> build/bdist.linux-x86_64/egg/dgl/_dataloading
copying build/lib.linux-x86_64-cpython-39/dgl/_api_internal.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-cpython-39/dgl/heterograph.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/backend
copying build/lib.linux-x86_64-cpython-39/dgl/backend/backend.py -> build/bdist.linux-x86_64/egg/dgl/backend
creating build/bdist.linux-x86_64/egg/dgl/backend/pytorch
copying build/lib.linux-x86_64-cpython-39/dgl/backend/pytorch/sparse.py -> build/bdist.linux-x86_64/egg/dgl/backend/pytorch
copying build/lib.linux-x86_64-cpython-39/dgl/backend/pytorch/tensor.py -> build/bdist.linux-x86_64/egg/dgl/backend/pytorch
copying build/lib.linux-x86_64-cpython-39/dgl/backend/pytorch/__init__.py -> build/bdist.linux-x86_64/egg/dgl/backend/pytorch
creating build/bdist.linux-x86_64/egg/dgl/backend/mxnet
copying build/lib.linux-x86_64-cpython-39/dgl/backend/mxnet/sparse.py -> build/bdist.linux-x86_64/egg/dgl/backend/mxnet
copying build/lib.linux-x86_64-cpython-39/dgl/backend/mxnet/tensor.py -> build/bdist.linux-x86_64/egg/dgl/backend/mxnet
copying build/lib.linux-x86_64-cpython-39/dgl/backend/mxnet/__init__.py -> build/bdist.linux-x86_64/egg/dgl/backend/mxnet
copying build/lib.linux-x86_64-cpython-39/dgl/backend/mxnet/sparse_optim.py -> build/bdist.linux-x86_64/egg/dgl/backend/mxnet
copying build/lib.linux-x86_64-cpython-39/dgl/backend/__init__.py -> build/bdist.linux-x86_64/egg/dgl/backend
creating build/bdist.linux-x86_64/egg/dgl/backend/tensorflow
copying build/lib.linux-x86_64-cpython-39/dgl/backend/tensorflow/sparse.py -> build/bdist.linux-x86_64/egg/dgl/backend/tensorflow
copying build/lib.linux-x86_64-cpython-39/dgl/backend/tensorflow/tensor.py -> build/bdist.linux-x86_64/egg/dgl/backend/tensorflow
copying build/lib.linux-x86_64-cpython-39/dgl/backend/tensorflow/__init__.py -> build/bdist.linux-x86_64/egg/dgl/backend/tensorflow
copying build/lib.linux-x86_64-cpython-39/dgl/backend/tensorflow/sparse_optim.py -> build/bdist.linux-x86_64/egg/dgl/backend/tensorflow
copying build/lib.linux-x86_64-cpython-39/dgl/backend/set_default_backend.py -> build/bdist.linux-x86_64/egg/dgl/backend
copying build/lib.linux-x86_64-cpython-39/dgl/network.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/ops
copying build/lib.linux-x86_64-cpython-39/dgl/ops/edge_softmax.py -> build/bdist.linux-x86_64/egg/dgl/ops
copying build/lib.linux-x86_64-cpython-39/dgl/ops/sddmm.py -> build/bdist.linux-x86_64/egg/dgl/ops
copying build/lib.linux-x86_64-cpython-39/dgl/ops/gather_mm.py -> build/bdist.linux-x86_64/egg/dgl/ops
copying build/lib.linux-x86_64-cpython-39/dgl/ops/segment.py -> build/bdist.linux-x86_64/egg/dgl/ops
copying build/lib.linux-x86_64-cpython-39/dgl/ops/__init__.py -> build/bdist.linux-x86_64/egg/dgl/ops
copying build/lib.linux-x86_64-cpython-39/dgl/ops/spmm.py -> build/bdist.linux-x86_64/egg/dgl/ops
copying build/lib.linux-x86_64-cpython-39/dgl/logging.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-cpython-39/dgl/graph_index.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/cuda
copying build/lib.linux-x86_64-cpython-39/dgl/cuda/nccl.py -> build/bdist.linux-x86_64/egg/dgl/cuda
copying build/lib.linux-x86_64-cpython-39/dgl/cuda/__init__.py -> build/bdist.linux-x86_64/egg/dgl/cuda
copying build/lib.linux-x86_64-cpython-39/dgl/udf.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/storages
copying build/lib.linux-x86_64-cpython-39/dgl/storages/pytorch_tensor.py -> build/bdist.linux-x86_64/egg/dgl/storages
copying build/lib.linux-x86_64-cpython-39/dgl/storages/base.py -> build/bdist.linux-x86_64/egg/dgl/storages
copying build/lib.linux-x86_64-cpython-39/dgl/storages/tensor.py -> build/bdist.linux-x86_64/egg/dgl/storages
copying build/lib.linux-x86_64-cpython-39/dgl/storages/__init__.py -> build/bdist.linux-x86_64/egg/dgl/storages
copying build/lib.linux-x86_64-cpython-39/dgl/storages/numpy.py -> build/bdist.linux-x86_64/egg/dgl/storages
copying build/lib.linux-x86_64-cpython-39/dgl/ndarray.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-cpython-39/dgl/init.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-cpython-39/dgl/view.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/sampling
copying build/lib.linux-x86_64-cpython-39/dgl/sampling/neighbor.py -> build/bdist.linux-x86_64/egg/dgl/sampling
copying build/lib.linux-x86_64-cpython-39/dgl/sampling/pinsage.py -> build/bdist.linux-x86_64/egg/dgl/sampling
copying build/lib.linux-x86_64-cpython-39/dgl/sampling/negative.py -> build/bdist.linux-x86_64/egg/dgl/sampling
copying build/lib.linux-x86_64-cpython-39/dgl/sampling/randomwalks.py -> build/bdist.linux-x86_64/egg/dgl/sampling
copying build/lib.linux-x86_64-cpython-39/dgl/sampling/utils.py -> build/bdist.linux-x86_64/egg/dgl/sampling
copying build/lib.linux-x86_64-cpython-39/dgl/sampling/node2vec_randomwalk.py -> build/bdist.linux-x86_64/egg/dgl/sampling
copying build/lib.linux-x86_64-cpython-39/dgl/sampling/__init__.py -> build/bdist.linux-x86_64/egg/dgl/sampling
copying build/lib.linux-x86_64-cpython-39/dgl/generators.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-cpython-39/dgl/data/qm7b.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-cpython-39/dgl/data/graph_serialize.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-cpython-39/dgl/data/dgl_dataset.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-cpython-39/dgl/data/csv_dataset_base.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-cpython-39/dgl/data/knowledge_graph.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-cpython-39/dgl/data/qm9_edge.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-cpython-39/dgl/data/gnn_benchmark.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-cpython-39/dgl/data/qm9.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-cpython-39/dgl/data/icews18.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-cpython-39/dgl/data/gdelt.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-cpython-39/dgl/data/utils.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-cpython-39/dgl/data/sbm.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-cpython-39/dgl/data/csv_dataset.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-cpython-39/dgl/data/karate.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-cpython-39/dgl/data/citation_graph.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-cpython-39/dgl/data/rdf.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-cpython-39/dgl/data/bitcoinotc.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-cpython-39/dgl/data/ppi.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-cpython-39/dgl/data/tensor_serialize.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-cpython-39/dgl/data/flickr.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-cpython-39/dgl/data/wikics.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-cpython-39/dgl/data/heterograph_serialize.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-cpython-39/dgl/data/tu.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-cpython-39/dgl/data/__init__.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-cpython-39/dgl/data/gindt.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-cpython-39/dgl/data/tree.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-cpython-39/dgl/data/synthetic.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-cpython-39/dgl/data/adapter.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-cpython-39/dgl/data/fakenews.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-cpython-39/dgl/data/fraud.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-cpython-39/dgl/data/minigc.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-cpython-39/dgl/data/yelp.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-cpython-39/dgl/data/reddit.py -> build/bdist.linux-x86_64/egg/dgl/data
copying build/lib.linux-x86_64-cpython-39/dgl/convert.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-cpython-39/dgl/container.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-cpython-39/dgl/merge.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-cpython-39/dgl/readout.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-cpython-39/dgl/__init__.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-cpython-39/dgl/distributed/dist_graph.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-cpython-39/dgl/distributed/graph_partition_book.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-cpython-39/dgl/distributed/partition.py -> build/bdist.linux-x86_64/egg/dgl/distributed
creating build/bdist.linux-x86_64/egg/dgl/distributed/nn
creating build/bdist.linux-x86_64/egg/dgl/distributed/nn/pytorch
copying build/lib.linux-x86_64-cpython-39/dgl/distributed/nn/pytorch/sparse_emb.py -> build/bdist.linux-x86_64/egg/dgl/distributed/nn/pytorch
copying build/lib.linux-x86_64-cpython-39/dgl/distributed/nn/pytorch/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distributed/nn/pytorch
creating build/bdist.linux-x86_64/egg/dgl/distributed/nn/mxnet
copying build/lib.linux-x86_64-cpython-39/dgl/distributed/nn/mxnet/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distributed/nn/mxnet
copying build/lib.linux-x86_64-cpython-39/dgl/distributed/nn/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distributed/nn
creating build/bdist.linux-x86_64/egg/dgl/distributed/nn/tensorflow
copying build/lib.linux-x86_64-cpython-39/dgl/distributed/nn/tensorflow/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distributed/nn/tensorflow
creating build/bdist.linux-x86_64/egg/dgl/distributed/optim
creating build/bdist.linux-x86_64/egg/dgl/distributed/optim/pytorch
copying build/lib.linux-x86_64-cpython-39/dgl/distributed/optim/pytorch/utils.py -> build/bdist.linux-x86_64/egg/dgl/distributed/optim/pytorch
copying build/lib.linux-x86_64-cpython-39/dgl/distributed/optim/pytorch/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distributed/optim/pytorch
copying build/lib.linux-x86_64-cpython-39/dgl/distributed/optim/pytorch/sparse_optim.py -> build/bdist.linux-x86_64/egg/dgl/distributed/optim/pytorch
creating build/bdist.linux-x86_64/egg/dgl/distributed/optim/mxnet
copying build/lib.linux-x86_64-cpython-39/dgl/distributed/optim/mxnet/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distributed/optim/mxnet
copying build/lib.linux-x86_64-cpython-39/dgl/distributed/optim/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distributed/optim
creating build/bdist.linux-x86_64/egg/dgl/distributed/optim/tensorflow
copying build/lib.linux-x86_64-cpython-39/dgl/distributed/optim/tensorflow/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distributed/optim/tensorflow
copying build/lib.linux-x86_64-cpython-39/dgl/distributed/standalone_kvstore.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-cpython-39/dgl/distributed/graph_services.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-cpython-39/dgl/distributed/role.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-cpython-39/dgl/distributed/constants.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-cpython-39/dgl/distributed/rpc_server.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-cpython-39/dgl/distributed/kvstore.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-cpython-39/dgl/distributed/dist_context.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-cpython-39/dgl/distributed/rpc.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-cpython-39/dgl/distributed/rpc_client.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-cpython-39/dgl/distributed/id_map.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-cpython-39/dgl/distributed/server_state.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-cpython-39/dgl/distributed/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-cpython-39/dgl/distributed/shared_mem_utils.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-cpython-39/dgl/distributed/dist_tensor.py -> build/bdist.linux-x86_64/egg/dgl/distributed
copying build/lib.linux-x86_64-cpython-39/dgl/distributed/dist_dataloader.py -> build/bdist.linux-x86_64/egg/dgl/distributed
creating build/bdist.linux-x86_64/egg/dgl/_deprecate
copying build/lib.linux-x86_64-cpython-39/dgl/_deprecate/nodeflow.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate
copying build/lib.linux-x86_64-cpython-39/dgl/_deprecate/udf.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate
copying build/lib.linux-x86_64-cpython-39/dgl/_deprecate/view.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate
copying build/lib.linux-x86_64-cpython-39/dgl/_deprecate/kernel.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate
copying build/lib.linux-x86_64-cpython-39/dgl/_deprecate/__init__.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate
creating build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime
copying build/lib.linux-x86_64-cpython-39/dgl/_deprecate/runtime/spmv.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime
copying build/lib.linux-x86_64-cpython-39/dgl/_deprecate/runtime/degree_bucketing.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime
copying build/lib.linux-x86_64-cpython-39/dgl/_deprecate/runtime/__init__.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime
copying build/lib.linux-x86_64-cpython-39/dgl/_deprecate/runtime/scheduler.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime
copying build/lib.linux-x86_64-cpython-39/dgl/_deprecate/runtime/adapter.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime
copying build/lib.linux-x86_64-cpython-39/dgl/_deprecate/runtime/runtime.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime
creating build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/ir
copying build/lib.linux-x86_64-cpython-39/dgl/_deprecate/runtime/ir/program.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/ir
copying build/lib.linux-x86_64-cpython-39/dgl/_deprecate/runtime/ir/executor.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/ir
copying build/lib.linux-x86_64-cpython-39/dgl/_deprecate/runtime/ir/var.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/ir
copying build/lib.linux-x86_64-cpython-39/dgl/_deprecate/runtime/ir/registry.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/ir
copying build/lib.linux-x86_64-cpython-39/dgl/_deprecate/runtime/ir/__init__.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/ir
copying build/lib.linux-x86_64-cpython-39/dgl/_deprecate/graph.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate
copying build/lib.linux-x86_64-cpython-39/dgl/_deprecate/frame.py -> build/bdist.linux-x86_64/egg/dgl/_deprecate
copying build/lib.linux-x86_64-cpython-39/dgl/propagate.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-cpython-39/dgl/batch.py -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/transforms
copying build/lib.linux-x86_64-cpython-39/dgl/transforms/functional.py -> build/bdist.linux-x86_64/egg/dgl/transforms
copying build/lib.linux-x86_64-cpython-39/dgl/transforms/__init__.py -> build/bdist.linux-x86_64/egg/dgl/transforms
copying build/lib.linux-x86_64-cpython-39/dgl/transforms/module.py -> build/bdist.linux-x86_64/egg/dgl/transforms
creating build/bdist.linux-x86_64/egg/dgl/distgnn
creating build/bdist.linux-x86_64/egg/dgl/distgnn/partition
copying build/lib.linux-x86_64-cpython-39/dgl/distgnn/partition/libra_partition.py -> build/bdist.linux-x86_64/egg/dgl/distgnn/partition
copying build/lib.linux-x86_64-cpython-39/dgl/distgnn/partition/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distgnn/partition
creating build/bdist.linux-x86_64/egg/dgl/distgnn/tools
copying build/lib.linux-x86_64-cpython-39/dgl/distgnn/tools/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distgnn/tools
copying build/lib.linux-x86_64-cpython-39/dgl/distgnn/tools/tools.py -> build/bdist.linux-x86_64/egg/dgl/distgnn/tools
copying build/lib.linux-x86_64-cpython-39/dgl/distgnn/__init__.py -> build/bdist.linux-x86_64/egg/dgl/distgnn
creating build/bdist.linux-x86_64/egg/dgl/_ffi
copying build/lib.linux-x86_64-cpython-39/dgl/_ffi/runtime_ctypes.py -> build/bdist.linux-x86_64/egg/dgl/_ffi
copying build/lib.linux-x86_64-cpython-39/dgl/_ffi/function.py -> build/bdist.linux-x86_64/egg/dgl/_ffi
copying build/lib.linux-x86_64-cpython-39/dgl/_ffi/libinfo.py -> build/bdist.linux-x86_64/egg/dgl/_ffi
copying build/lib.linux-x86_64-cpython-39/dgl/_ffi/streams.py -> build/bdist.linux-x86_64/egg/dgl/_ffi
copying build/lib.linux-x86_64-cpython-39/dgl/_ffi/base.py -> build/bdist.linux-x86_64/egg/dgl/_ffi
creating build/bdist.linux-x86_64/egg/dgl/_ffi/_ctypes
copying build/lib.linux-x86_64-cpython-39/dgl/_ffi/_ctypes/function.py -> build/bdist.linux-x86_64/egg/dgl/_ffi/_ctypes
copying build/lib.linux-x86_64-cpython-39/dgl/_ffi/_ctypes/types.py -> build/bdist.linux-x86_64/egg/dgl/_ffi/_ctypes
copying build/lib.linux-x86_64-cpython-39/dgl/_ffi/_ctypes/ndarray.py -> build/bdist.linux-x86_64/egg/dgl/_ffi/_ctypes
copying build/lib.linux-x86_64-cpython-39/dgl/_ffi/_ctypes/__init__.py -> build/bdist.linux-x86_64/egg/dgl/_ffi/_ctypes
copying build/lib.linux-x86_64-cpython-39/dgl/_ffi/_ctypes/object.py -> build/bdist.linux-x86_64/egg/dgl/_ffi/_ctypes
copying build/lib.linux-x86_64-cpython-39/dgl/_ffi/ndarray.py -> build/bdist.linux-x86_64/egg/dgl/_ffi
copying build/lib.linux-x86_64-cpython-39/dgl/_ffi/object_generic.py -> build/bdist.linux-x86_64/egg/dgl/_ffi
copying build/lib.linux-x86_64-cpython-39/dgl/_ffi/__init__.py -> build/bdist.linux-x86_64/egg/dgl/_ffi
copying build/lib.linux-x86_64-cpython-39/dgl/_ffi/object.py -> build/bdist.linux-x86_64/egg/dgl/_ffi
creating build/bdist.linux-x86_64/egg/dgl/_ffi/_cy3
copying build/lib.linux-x86_64-cpython-39/dgl/_ffi/_cy3/__init__.py -> build/bdist.linux-x86_64/egg/dgl/_ffi/_cy3
creating build/bdist.linux-x86_64/egg/dgl/_ffi/_cy2
copying build/lib.linux-x86_64-cpython-39/dgl/_ffi/_cy2/__init__.py -> build/bdist.linux-x86_64/egg/dgl/_ffi/_cy2
copying build/lib.linux-x86_64-cpython-39/dgl/frame.py -> build/bdist.linux-x86_64/egg/dgl
copying build/lib.linux-x86_64-cpython-39/dgl/random.py -> build/bdist.linux-x86_64/egg/dgl
byte-compiling build/bdist.linux-x86_64/egg/dgl/mock_sparse/sp_matrix.py to sp_matrix.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/mock_sparse/diag_matrix.py to diag_matrix.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/mock_sparse/reduction.py to reduction.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/mock_sparse/__init__.py to __init__.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/mock_sparse/elementwise_op_sp.py to elementwise_op_sp.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/geometry/edge_coarsening.py to edge_coarsening.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/geometry/fps.py to fps.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/geometry/capi.py to capi.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/geometry/__init__.py to __init__.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/function/base.py to base.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/function/message.py to message.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/function/reducer.py to reducer.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/function/__init__.py to __init__.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/partition.py to partition.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/glob.py to glob.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/link/transe.py to transe.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/link/edgepred.py to edgepred.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/link/transr.py to transr.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/link/__init__.py to __init__.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/hetero.py to hetero.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/sparse_emb.py to sparse_emb.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/linear.py to linear.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/utils.py to utils.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/agnnconv.py to agnnconv.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/relgraphconv.py to relgraphconv.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/appnpconv.py to appnpconv.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/dgnconv.py to dgnconv.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/grouprevres.py to grouprevres.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/twirlsconv.py to twirlsconv.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/egatconv.py to egatconv.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/graphconv.py to graphconv.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/densechebconv.py to densechebconv.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/gcn2conv.py to gcn2conv.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/ginconv.py to ginconv.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/dotgatconv.py to dotgatconv.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/cfconv.py to cfconv.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/gmmconv.py to gmmconv.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/sageconv.py to sageconv.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/densesageconv.py to densesageconv.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/gineconv.py to gineconv.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/edgeconv.py to edgeconv.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/nnconv.py to nnconv.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/__init__.py to __init__.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/gatv2conv.py to gatv2conv.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/hgtconv.py to hgtconv.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/pnaconv.py to pnaconv.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/gatedgraphconv.py to gatedgraphconv.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/atomicconv.py to atomicconv.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/tagconv.py to tagconv.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/densegraphconv.py to densegraphconv.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/egnnconv.py to egnnconv.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/chebconv.py to chebconv.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/sgconv.py to sgconv.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/conv/gatconv.py to gatconv.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/__init__.py to __init__.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/explain/gnnexplainer.py to gnnexplainer.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/explain/__init__.py to __init__.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/factory.py to factory.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/pytorch/softmax.py to softmax.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/glob.py to glob.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/hetero.py to hetero.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/utils.py to utils.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/agnnconv.py to agnnconv.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/relgraphconv.py to relgraphconv.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/appnpconv.py to appnpconv.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/graphconv.py to graphconv.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/densechebconv.py to densechebconv.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/ginconv.py to ginconv.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/gmmconv.py to gmmconv.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/sageconv.py to sageconv.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/densesageconv.py to densesageconv.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/edgeconv.py to edgeconv.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/nnconv.py to nnconv.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/__init__.py to __init__.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/gatedgraphconv.py to gatedgraphconv.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/tagconv.py to tagconv.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/densegraphconv.py to densegraphconv.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/chebconv.py to chebconv.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/sgconv.py to sgconv.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/conv/gatconv.py to gatconv.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/__init__.py to __init__.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/mxnet/softmax.py to softmax.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/__init__.py to __init__.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/glob.py to glob.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/hetero.py to hetero.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/utils.py to utils.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv/relgraphconv.py to relgraphconv.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv/appnpconv.py to appnpconv.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv/graphconv.py to graphconv.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv/densechebconv.py to densechebconv.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv/ginconv.py to ginconv.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv/sageconv.py to sageconv.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv/edgeconv.py to edgeconv.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv/__init__.py to __init__.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv/chebconv.py to chebconv.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv/sgconv.py to sgconv.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/conv/gatconv.py to gatconv.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/__init__.py to __init__.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/tensorflow/softmax.py to softmax.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/nn/functional/__init__.py to __init__.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/core.py to core.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/subgraph.py to subgraph.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/contrib/dis_kvstore.py to dis_kvstore.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/contrib/unified_tensor.py to unified_tensor.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/contrib/graph_store.py to graph_store.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/contrib/sampling/sampler.py to sampler.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/contrib/sampling/dis_sampler.py to dis_sampler.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/contrib/sampling/__init__.py to __init__.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/contrib/data/knowledge_graph.py to knowledge_graph.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/contrib/data/__init__.py to __init__.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/contrib/__init__.py to __init__.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/traversal.py to traversal.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/utils/checks.py to checks.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/utils/exception.py to exception.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/utils/pin_memory.py to pin_memory.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/utils/internal.py to internal.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/utils/filter.py to filter.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/utils/data.py to data.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/utils/__init__.py to __init__.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/utils/shared_mem.py to shared_mem.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/multiprocessing/pytorch.py to pytorch.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/multiprocessing/__init__.py to __init__.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/base.py to base.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/optim/pytorch/__init__.py to __init__.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/optim/pytorch/sparse_optim.py to sparse_optim.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/optim/mxnet/__init__.py to __init__.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/optim/__init__.py to __init__.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/optim/tensorflow/__init__.py to __init__.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/heterograph_index.py to heterograph_index.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/sparse.py to sparse.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/global_config.py to global_config.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/dataloading/base.py to base.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/dataloading/dataloader.py to dataloader.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/dataloading/graphsaint.py to graphsaint.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/dataloading/shadow.py to shadow.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/dataloading/negative_sampler.py to negative_sampler.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/dataloading/__init__.py to __init__.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/dataloading/cluster_gcn.py to cluster_gcn.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/dataloading/neighbor_sampler.py to neighbor_sampler.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/dataloading/dist_dataloader.py to dist_dataloader.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_dataloading/neighbor.py to neighbor.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_dataloading/dataloader.py to dataloader.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_dataloading/pytorch/dataloader.py to dataloader.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_dataloading/pytorch/__init__.py to __init__.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_dataloading/shadow.py to shadow.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_dataloading/negative_sampler.py to negative_sampler.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_dataloading/__init__.py to __init__.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_dataloading/cluster_gcn.py to cluster_gcn.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_api_internal.py to _api_internal.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/heterograph.py to heterograph.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/backend.py to backend.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/pytorch/sparse.py to sparse.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/pytorch/tensor.py to tensor.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/pytorch/__init__.py to __init__.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/mxnet/sparse.py to sparse.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/mxnet/tensor.py to tensor.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/mxnet/__init__.py to __init__.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/mxnet/sparse_optim.py to sparse_optim.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/__init__.py to __init__.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/tensorflow/sparse.py to sparse.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/tensorflow/tensor.py to tensor.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/tensorflow/__init__.py to __init__.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/tensorflow/sparse_optim.py to sparse_optim.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/backend/set_default_backend.py to set_default_backend.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/network.py to network.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/ops/edge_softmax.py to edge_softmax.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/ops/sddmm.py to sddmm.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/ops/gather_mm.py to gather_mm.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/ops/segment.py to segment.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/ops/__init__.py to __init__.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/ops/spmm.py to spmm.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/logging.py to logging.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/graph_index.py to graph_index.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/cuda/nccl.py to nccl.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/cuda/__init__.py to __init__.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/udf.py to udf.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/storages/pytorch_tensor.py to pytorch_tensor.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/storages/base.py to base.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/storages/tensor.py to tensor.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/storages/__init__.py to __init__.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/storages/numpy.py to numpy.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/ndarray.py to ndarray.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/init.py to init.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/view.py to view.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/sampling/neighbor.py to neighbor.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/sampling/pinsage.py to pinsage.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/sampling/negative.py to negative.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/sampling/randomwalks.py to randomwalks.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/sampling/utils.py to utils.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/sampling/node2vec_randomwalk.py to node2vec_randomwalk.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/sampling/__init__.py to __init__.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/generators.py to generators.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/qm7b.py to qm7b.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/graph_serialize.py to graph_serialize.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/dgl_dataset.py to dgl_dataset.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/csv_dataset_base.py to csv_dataset_base.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/knowledge_graph.py to knowledge_graph.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/qm9_edge.py to qm9_edge.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/gnn_benchmark.py to gnn_benchmark.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/qm9.py to qm9.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/icews18.py to icews18.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/gdelt.py to gdelt.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/utils.py to utils.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/sbm.py to sbm.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/csv_dataset.py to csv_dataset.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/karate.py to karate.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/citation_graph.py to citation_graph.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/rdf.py to rdf.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/bitcoinotc.py to bitcoinotc.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/ppi.py to ppi.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/tensor_serialize.py to tensor_serialize.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/flickr.py to flickr.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/wikics.py to wikics.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/heterograph_serialize.py to heterograph_serialize.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/tu.py to tu.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/__init__.py to __init__.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/gindt.py to gindt.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/tree.py to tree.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/synthetic.py to synthetic.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/adapter.py to adapter.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/fakenews.py to fakenews.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/fraud.py to fraud.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/minigc.py to minigc.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/yelp.py to yelp.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/data/reddit.py to reddit.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/convert.py to convert.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/container.py to container.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/merge.py to merge.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/readout.py to readout.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/__init__.py to __init__.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/dist_graph.py to dist_graph.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/graph_partition_book.py to graph_partition_book.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/partition.py to partition.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/nn/pytorch/sparse_emb.py to sparse_emb.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/nn/pytorch/__init__.py to __init__.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/nn/mxnet/__init__.py to __init__.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/nn/__init__.py to __init__.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/nn/tensorflow/__init__.py to __init__.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/optim/pytorch/utils.py to utils.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/optim/pytorch/__init__.py to __init__.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/optim/pytorch/sparse_optim.py to sparse_optim.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/optim/mxnet/__init__.py to __init__.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/optim/__init__.py to __init__.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/optim/tensorflow/__init__.py to __init__.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/standalone_kvstore.py to standalone_kvstore.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/graph_services.py to graph_services.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/role.py to role.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/constants.py to constants.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/rpc_server.py to rpc_server.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/kvstore.py to kvstore.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/dist_context.py to dist_context.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/rpc.py to rpc.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/rpc_client.py to rpc_client.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/id_map.py to id_map.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/server_state.py to server_state.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/__init__.py to __init__.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/shared_mem_utils.py to shared_mem_utils.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/dist_tensor.py to dist_tensor.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distributed/dist_dataloader.py to dist_dataloader.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/nodeflow.py to nodeflow.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/udf.py to udf.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/view.py to view.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/kernel.py to kernel.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/__init__.py to __init__.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/spmv.py to spmv.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/degree_bucketing.py to degree_bucketing.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/__init__.py to __init__.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/scheduler.py to scheduler.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/adapter.py to adapter.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/runtime.py to runtime.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/ir/program.py to program.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/ir/executor.py to executor.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/ir/var.py to var.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/ir/registry.py to registry.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/runtime/ir/__init__.py to __init__.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/graph.py to graph.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_deprecate/frame.py to frame.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/propagate.py to propagate.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/batch.py to batch.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/transforms/functional.py to functional.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/transforms/__init__.py to __init__.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/transforms/module.py to module.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distgnn/partition/libra_partition.py to libra_partition.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distgnn/partition/__init__.py to __init__.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distgnn/tools/__init__.py to __init__.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distgnn/tools/tools.py to tools.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/distgnn/__init__.py to __init__.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/runtime_ctypes.py to runtime_ctypes.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/function.py to function.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/libinfo.py to libinfo.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/streams.py to streams.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/base.py to base.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/_ctypes/function.py to function.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/_ctypes/types.py to types.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/_ctypes/ndarray.py to ndarray.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/_ctypes/__init__.py to __init__.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/_ctypes/object.py to object.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/ndarray.py to ndarray.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/object_generic.py to object_generic.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/__init__.py to __init__.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/object.py to object.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/_cy3/__init__.py to __init__.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/_ffi/_cy2/__init__.py to __init__.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/frame.py to frame.cpython-39.pyc
byte-compiling build/bdist.linux-x86_64/egg/dgl/random.py to random.cpython-39.pyc
installing package data to build/bdist.linux-x86_64/egg
running install_data
copying ../build/libdgl.so -> build/bdist.linux-x86_64/egg/dgl
creating build/bdist.linux-x86_64/egg/dgl/tensoradapter
creating build/bdist.linux-x86_64/egg/dgl/tensoradapter/pytorch
creating build/bdist.linux-x86_64/egg/EGG-INFO
copying dgl.egg-info/PKG-INFO -> build/bdist.linux-x86_64/egg/EGG-INFO
copying dgl.egg-info/SOURCES.txt -> build/bdist.linux-x86_64/egg/EGG-INFO
copying dgl.egg-info/dependency_links.txt -> build/bdist.linux-x86_64/egg/EGG-INFO
copying dgl.egg-info/not-zip-safe -> build/bdist.linux-x86_64/egg/EGG-INFO
copying dgl.egg-info/requires.txt -> build/bdist.linux-x86_64/egg/EGG-INFO
copying dgl.egg-info/top_level.txt -> build/bdist.linux-x86_64/egg/EGG-INFO
creating dist
creating 'dist/dgl-0.9-py3.9-linux-x86_64.egg' and adding 'build/bdist.linux-x86_64/egg' to it
removing 'build/bdist.linux-x86_64/egg' (and everything under it)
Processing dgl-0.9-py3.9-linux-x86_64.egg
creating /opt/conda/lib/python3.9/site-packages/dgl-0.9-py3.9-linux-x86_64.egg
Extracting dgl-0.9-py3.9-linux-x86_64.egg to /opt/conda/lib/python3.9/site-packages
Adding dgl 0.9 to easy-install.pth file

Installed /opt/conda/lib/python3.9/site-packages/dgl-0.9-py3.9-linux-x86_64.egg
Processing dependencies for dgl==0.9
Searching for certifi>=2017.4.17
Reading https://pypi.org/simple/certifi/
/opt/conda/lib/python3.9/site-packages/pkg_resources/__init__.py:123: PkgResourcesDeprecationWarning:  is an invalid version and will not be supported in a future release
  warnings.warn(
Downloading https://files.pythonhosted.org/packages/1d/38/fa96a426e0c0e68aabc68e896584b83ad1eec779265a028e156ce509630e/certifi-2022.9.24-py3-none-any.whl#sha256=90c1a32f1d68f940488354e36370f6cca89f0f106db09518524c88d6ed83f382
Best match: certifi 2022.9.24
Processing certifi-2022.9.24-py3-none-any.whl
Installing certifi-2022.9.24-py3-none-any.whl to /opt/conda/lib/python3.9/site-packages
Adding certifi 2022.9.24 to easy-install.pth file

Installed /opt/conda/lib/python3.9/site-packages/certifi-2022.9.24-py3.9.egg
Searching for psutil==5.9.2
Best match: psutil 5.9.2
Adding psutil 5.9.2 to easy-install.pth file

Using /opt/conda/lib/python3.9/site-packages
Searching for tqdm==4.64.1
Best match: tqdm 4.64.1
Adding tqdm 4.64.1 to easy-install.pth file
Installing tqdm script to /opt/conda/bin

Using /opt/conda/lib/python3.9/site-packages
Searching for requests==2.28.1
Best match: requests 2.28.1
Adding requests 2.28.1 to easy-install.pth file

Using /opt/conda/lib/python3.9/site-packages
Searching for networkx==2.8.6
Best match: networkx 2.8.6
Adding networkx 2.8.6 to easy-install.pth file

Using /opt/conda/lib/python3.9/site-packages
Searching for scipy==1.9.1
Best match: scipy 1.9.1
Adding scipy 1.9.1 to easy-install.pth file

Using /opt/conda/lib/python3.9/site-packages
Searching for numpy==1.22.4
Best match: numpy 1.22.4
Adding numpy 1.22.4 to easy-install.pth file
Installing f2py script to /opt/conda/bin
Installing f2py3 script to /opt/conda/bin
Installing f2py3.9 script to /opt/conda/bin

Using /opt/conda/lib/python3.9/site-packages
Searching for urllib3==1.26.11
Best match: urllib3 1.26.11
Adding urllib3 1.26.11 to easy-install.pth file

Using /opt/conda/lib/python3.9/site-packages
Searching for idna==3.3
Best match: idna 3.3
Adding idna 3.3 to easy-install.pth file

Using /opt/conda/lib/python3.9/site-packages
Searching for charset-normalizer==2.1.1
Best match: charset-normalizer 2.1.1
Adding charset-normalizer 2.1.1 to easy-install.pth file
Installing normalizer script to /opt/conda/bin

Using /opt/conda/lib/python3.9/site-packages
Finished processing dependencies for dgl==0.9
WARNING: Cython is not installed, will compile without cython module
running build_ext
~/jenkins/workspace/dgl_PR-4648@3
[Pipeline] sh
+ ls -lh /usr/lib/x86_64-linux-gnu/
total 507M
-rw-r--r--  1 root root  496 May  3 10:19 Mcrt1.o
-rw-r--r--  1 root root 1.8K May  3 10:19 Scrt1.o
drwxr-xr-x  1 root root 4.0K Sep 19 18:34 audit
drwxr-xr-x  2 root root 4.0K Apr 28 00:03 coreutils
-rw-r--r--  1 root root 1.9K May  3 10:19 crt1.o
-rw-r--r--  1 root root 1.2K May  3 10:19 crti.o
-rw-r--r--  1 root root  648 May  3 10:19 crtn.o
drwxr-xr-x  1 root root 4.0K Sep 19 18:34 engines-1.1
drwxr-xr-x  1 root root  12K Sep 19 18:34 gconv
-rw-r--r--  1 root root 2.5K May  3 10:19 gcrt1.o
-rw-r--r--  1 root root 2.3K May  3 10:19 grcrt1.o
drwxr-xr-x  3 root root 4.0K Sep 19 18:34 krb5
drwxr-xr-x  2 root root 4.0K May 20 08:07 ldscripts
-rw-r--r--  1 root root 2.0K May  3 10:19 libBrokenLocale.a
lrwxrwxrwx  1 root root   42 May  3 10:19 libBrokenLocale.so -> /lib/x86_64-linux-gnu/libBrokenLocale.so.1
-rw-r--r--  1 root root  23K May  3 10:19 libanl.a
lrwxrwxrwx  1 root root   33 May  3 10:19 libanl.so -> /lib/x86_64-linux-gnu/libanl.so.1
lrwxrwxrwx  1 root root   19 Jun 15  2021 libapt-pkg.so.5.0 -> libapt-pkg.so.5.0.2
-rw-r--r--  1 root root 1.8M Jun 15  2021 libapt-pkg.so.5.0.2
lrwxrwxrwx  1 root root   23 Jun 15  2021 libapt-private.so.0.0 -> libapt-private.so.0.0.0
-rw-r--r--  1 root root 415K Jun 15  2021 libapt-private.so.0.0.0
lrwxrwxrwx  1 root root   16 Dec  4  2019 libasan.so.4 -> libasan.so.4.0.0
-rw-r--r--  1 root root 1.4M Dec  4  2019 libasan.so.4.0.0
lrwxrwxrwx  1 root root   16 Dec 15  2017 libasn1.so.8 -> libasn1.so.8.0.0
-rw-r--r--  1 root root 647K Dec 15  2017 libasn1.so.8.0.0
lrwxrwxrwx  1 root root   18 Feb  6  2018 libassuan.so.0 -> libassuan.so.0.8.1
-rw-r--r--  1 root root  75K Feb  6  2018 libassuan.so.0.8.1
lrwxrwxrwx  1 root root   18 Mar 10  2020 libatomic.so.1 -> libatomic.so.1.2.0
-rw-r--r--  1 root root  27K Mar 10  2020 libatomic.so.1.2.0
-rw-r--r--  1 root root 1.3M Oct 20  2021 libbfd-2.30-system.so
-rw-r--r--  1 root root 5.3M May  3 10:19 libc.a
-rw-r--r--  1 root root  298 May  3 10:19 libc.so
-rw-r--r--  1 root root  20K May  3 10:19 libc_nonshared.a
lrwxrwxrwx  1 root root   15 Mar 10  2020 libcc1.so.0 -> libcc1.so.0.0.0
-rw-r--r--  1 root root 116K Mar 10  2020 libcc1.so.0.0.0
lrwxrwxrwx  1 root root   34 May  3 10:19 libcidn.so -> /lib/x86_64-linux-gnu/libcidn.so.1
lrwxrwxrwx  1 root root   19 Dec  4  2019 libcilkrts.so.5 -> libcilkrts.so.5.0.0
-rw-r--r--  1 root root 117K Dec  4  2019 libcilkrts.so.5.0.0
-rw-r--r--  1 root root  61K May  3 10:19 libcrypt.a
lrwxrwxrwx  1 root root   35 May  3 10:19 libcrypt.so -> /lib/x86_64-linux-gnu/libcrypt.so.1
-rw-r--r--  1 root root 2.8M Jul  4 11:25 libcrypto.so.1.1
lrwxrwxrwx  1 root root   19 Aug 31 19:18 libcurl-gnutls.so.3 -> libcurl-gnutls.so.4
lrwxrwxrwx  1 root root   23 Aug 31 19:18 libcurl-gnutls.so.4 -> libcurl-gnutls.so.4.5.0
-rw-r--r--  1 root root 515K Aug 31 19:18 libcurl-gnutls.so.4.5.0
lrwxrwxrwx  1 root root   16 Aug 31 19:18 libcurl.so.4 -> libcurl.so.4.5.0
-rw-r--r--  1 root root 519K Aug 31 19:18 libcurl.so.4.5.0
-rw-r--r--  1 root root 1.7M Jun  3  2019 libdb-5.3.so
lrwxrwxrwx  1 root root   25 Jun  8  2016 libdebconfclient.so.0 -> libdebconfclient.so.0.0.0
-rw-r--r--  1 root root  11K Jun  8  2016 libdebconfclient.so.0.0.0
-rw-r--r--  1 root root  12K May  3 10:19 libdl.a
lrwxrwxrwx  1 root root   32 May  3 10:19 libdl.so -> /lib/x86_64-linux-gnu/libdl.so.2
lrwxrwxrwx  1 root root   18 Mar  8  2022 libexpatw.so.1 -> libexpatw.so.1.6.7
-rw-r--r--  1 root root 203K Mar  8  2022 libexpatw.so.1.6.7
lrwxrwxrwx  1 root root   15 Jan  7  2018 libffi.so.6 -> libffi.so.6.0.4
-rw-r--r--  1 root root  31K Jan  7  2018 libffi.so.6.0.4
lrwxrwxrwx  1 root root   14 May 23  2018 libform.so.5 -> libform.so.5.9
-rw-r--r--  1 root root  60K May 23  2018 libform.so.5.9
lrwxrwxrwx  1 root root   15 May 23  2018 libformw.so.5 -> libformw.so.5.9
-rw-r--r--  1 root root  68K May 23  2018 libformw.so.5.9
-rw-r--r--  1 root root 1.2K May  3 10:19 libg.a
lrwxrwxrwx  1 root root   16 Mar 14  2018 libgdbm.so.5 -> libgdbm.so.5.0.0
-rw-r--r--  1 root root  51K Mar 14  2018 libgdbm.so.5.0.0
lrwxrwxrwx  1 root root   23 Mar 14  2018 libgdbm_compat.so.4 -> libgdbm_compat.so.4.0.0
-rw-r--r--  1 root root  14K Mar 14  2018 libgdbm_compat.so.4.0.0
lrwxrwxrwx  1 root root   16 Jan 24  2018 libgmp.so.10 -> libgmp.so.10.3.2
-rw-r--r--  1 root root 515K Jan 24  2018 libgmp.so.10.3.2
lrwxrwxrwx  1 root root   21 Aug  2 12:58 libgnutls.so.30 -> libgnutls.so.30.14.10
-rw-r--r--  1 root root 1.4M Aug  2 12:58 libgnutls.so.30.14.10
lrwxrwxrwx  1 root root   16 Mar 10  2020 libgomp.so.1 -> libgomp.so.1.0.0
-rw-r--r--  1 root root 188K Mar 10  2020 libgomp.so.1.0.0
lrwxrwxrwx  1 root root   18 Dec 15  2017 libgssapi.so.3 -> libgssapi.so.3.0.0
-rw-r--r--  1 root root 260K Dec 15  2017 libgssapi.so.3.0.0
lrwxrwxrwx  1 root root   21 Nov 11  2020 libgssapi_krb5.so.2 -> libgssapi_krb5.so.2.2
-rw-r--r--  1 root root 299K Nov 11  2020 libgssapi_krb5.so.2.2
lrwxrwxrwx  1 root root   19 Dec 15  2017 libhcrypto.so.4 -> libhcrypto.so.4.1.0
-rw-r--r--  1 root root 213K Dec 15  2017 libhcrypto.so.4.1.0
lrwxrwxrwx  1 root root   20 Dec 15  2017 libheimbase.so.1 -> libheimbase.so.1.0.0
-rw-r--r--  1 root root  59K Dec 15  2017 libheimbase.so.1.0.0
lrwxrwxrwx  1 root root   20 Dec 15  2017 libheimntlm.so.0 -> libheimntlm.so.0.1.0
-rw-r--r--  1 root root  35K Dec 15  2017 libheimntlm.so.0.1.0
lrwxrwxrwx  1 root root   17 Jun 14  2021 libhogweed.so.4 -> libhogweed.so.4.5
-rw-r--r--  1 root root 215K Jun 14  2021 libhogweed.so.4.5
lrwxrwxrwx  1 root root   17 Dec 15  2017 libhx509.so.5 -> libhx509.so.5.0.0
-rw-r--r--  1 root root 294K Dec 15  2017 libhx509.so.5.0.0
lrwxrwxrwx  1 root root   16 Oct 24  2019 libidn2.so.0 -> libidn2.so.0.3.3
-rw-r--r--  1 root root 114K Oct 24  2019 libidn2.so.0.3.3
lrwxrwxrwx  1 root root   16 Mar  9  2018 libisl.so.19 -> libisl.so.19.0.0
-rw-r--r--  1 root root 1.6M Mar  9  2018 libisl.so.19.0.0
lrwxrwxrwx  1 root root   15 Mar 10  2020 libitm.so.1 -> libitm.so.1.0.0
-rw-r--r--  1 root root 111K Mar 10  2020 libitm.so.1.0.0
lrwxrwxrwx  1 root root   18 Nov 11  2020 libk5crypto.so.3 -> libk5crypto.so.3.1
-rw-r--r--  1 root root 195K Nov 11  2020 libk5crypto.so.3.1
lrwxrwxrwx  1 root root   17 Dec 15  2017 libkrb5.so.26 -> libkrb5.so.26.0.0
-rw-r--r--  1 root root 561K Dec 15  2017 libkrb5.so.26.0.0
lrwxrwxrwx  1 root root   14 Nov 11  2020 libkrb5.so.3 -> libkrb5.so.3.3
-rw-r--r--  1 root root 857K Nov 11  2020 libkrb5.so.3.3
lrwxrwxrwx  1 root root   21 Nov 11  2020 libkrb5support.so.0 -> libkrb5support.so.0.1
-rw-r--r--  1 root root  43K Nov 11  2020 libkrb5support.so.0.1
lrwxrwxrwx  1 root root   17 Sep  3  2016 libksba.so.8 -> libksba.so.8.11.6
-rw-r--r--  1 root root 223K Sep  3  2016 libksba.so.8.11.6
lrwxrwxrwx  1 root root   21 May 12 13:52 liblber-2.4.so.2 -> liblber-2.4.so.2.10.8
-rw-r--r--  1 root root  55K May 12 13:52 liblber-2.4.so.2.10.8
lrwxrwxrwx  1 root root   18 May 12 13:52 libldap-2.4.so.2 -> libldap_r-2.4.so.2
lrwxrwxrwx  1 root root   23 May 12 13:52 libldap_r-2.4.so.2 -> libldap_r-2.4.so.2.10.8
-rw-r--r--  1 root root 320K May 12 13:52 libldap_r-2.4.so.2.10.8
lrwxrwxrwx  1 root root   16 Mar 10  2020 liblsan.so.0 -> liblsan.so.0.0.0
-rw-r--r--  1 root root 338K Mar 10  2020 liblsan.so.0.0.0
lrwxrwxrwx  1 root root   15 May 20  2021 liblz4.so.1 -> liblz4.so.1.7.1
-rw-r--r--  1 root root 111K May 20  2021 liblz4.so.1.7.1
-rw-r--r--  1 root root 3.5M May  3 10:19 libm-2.27.a
-rw-r--r--  1 root root  132 May  3 10:19 libm.a
-rw-r--r--  1 root root  186 May  3 10:19 libm.so
-rw-r--r--  1 root root 1.5K May  3 10:19 libmcheck.a
lrwxrwxrwx  1 root root   14 May 23  2018 libmenu.so.5 -> libmenu.so.5.9
-rw-r--r--  1 root root  35K May 23  2018 libmenu.so.5.9
lrwxrwxrwx  1 root root   15 May 23  2018 libmenuw.so.5 -> libmenuw.so.5.9
-rw-r--r--  1 root root  35K May 23  2018 libmenuw.so.5.9
lrwxrwxrwx  1 root root   15 Jan 23  2018 libmpc.so.3 -> libmpc.so.3.1.0
-rw-r--r--  1 root root  96K Jan 23  2018 libmpc.so.3.1.0
lrwxrwxrwx  1 root root   16 Feb  7  2018 libmpfr.so.6 -> libmpfr.so.6.0.1
-rw-r--r--  1 root root 512K Feb  7  2018 libmpfr.so.6.0.1
lrwxrwxrwx  1 root root   15 Mar 10  2020 libmpx.so.2 -> libmpx.so.2.0.1
-rw-r--r--  1 root root  19K Mar 10  2020 libmpx.so.2.0.1
lrwxrwxrwx  1 root root   23 Mar 10  2020 libmpxwrappers.so.2 -> libmpxwrappers.so.2.0.1
-rw-r--r--  1 root root  15K Mar 10  2020 libmpxwrappers.so.2.0.1
-rw-r--r--  1 root root 337K May  3 10:19 libmvec.a
lrwxrwxrwx  1 root root   34 May  3 10:19 libmvec.so -> /lib/x86_64-linux-gnu/libmvec.so.1
-rw-r--r--  1 root root 5.9K May  3 10:19 libmvec_nonshared.a
lrwxrwxrwx  1 root root   12 Oct 19  2021 libnccl.so -> libnccl.so.2
lrwxrwxrwx  1 root root   17 Oct 19  2021 libnccl.so.2 -> libnccl.so.2.11.4
-rw-r--r--  1 root root 220M Oct 19  2021 libnccl.so.2.11.4
-rw-r--r--  1 root root 238M Oct 19  2021 libnccl_static.a
lrwxrwxrwx  1 root root   16 Jun 14  2021 libnettle.so.6 -> libnettle.so.6.5
-rw-r--r--  1 root root 215K Jun 14  2021 libnettle.so.6.5
lrwxrwxrwx  1 root root   21 Apr 10  2018 libnghttp2.so.14 -> libnghttp2.so.14.15.2
-rw-r--r--  1 root root 150K Apr 10  2018 libnghttp2.so.14.15.2
lrwxrwxrwx  1 root root   16 Nov 12  2017 libnpth.so.0 -> libnpth.so.0.1.1
-rw-r--r--  1 root root  15K Nov 12  2017 libnpth.so.0.1.1
-rw-r--r--  1 root root 201K May  3 10:19 libnsl.a
lrwxrwxrwx  1 root root   33 May  3 10:19 libnsl.so -> /lib/x86_64-linux-gnu/libnsl.so.1
lrwxrwxrwx  1 root root   40 May  3 10:19 libnss_compat.so -> /lib/x86_64-linux-gnu/libnss_compat.so.2
lrwxrwxrwx  1 root root   37 May  3 10:19 libnss_dns.so -> /lib/x86_64-linux-gnu/libnss_dns.so.2
lrwxrwxrwx  1 root root   39 May  3 10:19 libnss_files.so -> /lib/x86_64-linux-gnu/libnss_files.so.2
lrwxrwxrwx  1 root root   40 May  3 10:19 libnss_hesiod.so -> /lib/x86_64-linux-gnu/libnss_hesiod.so.2
lrwxrwxrwx  1 root root   37 May  3 10:19 libnss_nis.so -> /lib/x86_64-linux-gnu/libnss_nis.so.2
lrwxrwxrwx  1 root root   41 May  3 10:19 libnss_nisplus.so -> /lib/x86_64-linux-gnu/libnss_nisplus.so.2
-rw-r--r--  1 root root  71K Jun 20  2018 libnuma.a
lrwxrwxrwx  1 root root   16 Jun 20  2018 libnuma.so -> libnuma.so.1.0.0
lrwxrwxrwx  1 root root   16 Jun 20  2018 libnuma.so.1 -> libnuma.so.1.0.0
-rw-r--r--  1 root root  43K Jun 20  2018 libnuma.so.1.0.0
-rw-r--r--  1 root root 1.7M Oct 20  2021 libopcodes-2.30-system.so
lrwxrwxrwx  1 root root   19 Jan  4  2021 libp11-kit.so.0 -> libp11-kit.so.0.3.0
-rw-r--r--  1 root root 1.2M Jan  4  2021 libp11-kit.so.0.3.0
lrwxrwxrwx  1 root root   15 May 23  2018 libpanel.so.5 -> libpanel.so.5.9
-rw-r--r--  1 root root  14K May 23  2018 libpanel.so.5.9
lrwxrwxrwx  1 root root   16 May 23  2018 libpanelw.so.5 -> libpanelw.so.5.9
-rw-r--r--  1 root root  14K May 23  2018 libpanelw.so.5.9
lrwxrwxrwx  1 root root   22 May 17 07:42 libpcreposix.so.3 -> libpcreposix.so.3.13.3
-rw-r--r--  1 root root 9.9K May 17 07:42 libpcreposix.so.3.13.3
lrwxrwxrwx  1 root root   17 Oct 19  2020 libperl.so.5.26 -> libperl.so.5.26.1
-rw-r--r--  1 root root 2.0M Oct 19  2020 libperl.so.5.26.1
lrwxrwxrwx  1 root root   15 Mar  4  2018 libpsl.so.5 -> libpsl.so.5.2.0
-rw-r--r--  1 root root  54K Mar  4  2018 libpsl.so.5.2.0
-rw-r--r--  1 root root 6.0M May  3 10:19 libpthread.a
-rw-r--r--  1 root root  252 May  3 10:19 libpthread.so
-rw-r--r--  1 root root  29K May  3 10:19 libpthread_nonshared.a
lrwxrwxrwx  1 root root   20 Mar 10  2020 libquadmath.so.0 -> libquadmath.so.0.0.0
-rw-r--r--  1 root root 256K Mar 10  2020 libquadmath.so.0.0.0
-rw-r--r--  1 root root 136K May  3 10:19 libresolv.a
lrwxrwxrwx  1 root root   36 May  3 10:19 libresolv.so -> /lib/x86_64-linux-gnu/libresolv.so.2
lrwxrwxrwx  1 root root   18 Dec 15  2017 libroken.so.18 -> libroken.so.18.1.0
-rw-r--r--  1 root root  87K Dec 15  2017 libroken.so.18.1.0
-rw-r--r--  1 root root  52K May  3 10:19 librpcsvc.a
-rw-r--r--  1 root root  76K May  3 10:19 librt.a
lrwxrwxrwx  1 root root   32 May  3 10:19 librt.so -> /lib/x86_64-linux-gnu/librt.so.1
-rw-r--r--  1 root root 111K Apr 27  2016 librtmp.so.1
lrwxrwxrwx  1 root root   18 Feb 15  2022 libsasl2.so.2 -> libsasl2.so.2.0.25
-rw-r--r--  1 root root 107K Feb 15  2022 libsasl2.so.2.0.25
-rw-r--r--  1 root root 251K Mar  1  2018 libsemanage.so.1
lrwxrwxrwx  1 root root   19 Feb  5  2018 libsigsegv.so.2 -> libsigsegv.so.2.0.5
-rw-r--r--  1 root root  15K Feb  5  2018 libsigsegv.so.2.0.5
lrwxrwxrwx  1 root root   19 Sep 14 17:02 libsqlite3.so.0 -> libsqlite3.so.0.8.6
-rw-r--r--  1 root root 1.1M Sep 14 17:02 libsqlite3.so.0.8.6
-rw-r--r--  1 root root 564K Jul  4 11:25 libssl.so.1.1
lrwxrwxrwx  1 root root   19 Mar 10  2020 libstdc++.so.6 -> libstdc++.so.6.0.25
-rw-r--r--  1 root root 1.6M Mar 10  2020 libstdc++.so.6.0.25
lrwxrwxrwx  1 root root   17 Jan 21  2018 libtasn1.so.6 -> libtasn1.so.6.5.5
-rw-r--r--  1 root root  74K Jan 21  2018 libtasn1.so.6.5.5
lrwxrwxrwx  1 root root   39 May  3 10:19 libthread_db.so -> /lib/x86_64-linux-gnu/libthread_db.so.1
lrwxrwxrwx  1 root root   13 May 23  2018 libtic.so.5 -> libtic.so.5.9
-rw-r--r--  1 root root  62K May 23  2018 libtic.so.5.9
lrwxrwxrwx  1 root root   16 Mar 10  2020 libtsan.so.0 -> libtsan.so.0.0.0
-rw-r--r--  1 root root 942K Mar 10  2020 libtsan.so.0.0.0
-rw-r--r--  1 root root 3.6K Mar 10  2020 libtsan_preinit.o
lrwxrwxrwx  1 root root   17 Dec  4  2019 libubsan.so.0 -> libubsan.so.0.0.0
-rw-r--r--  1 root root 322K Dec  4  2019 libubsan.so.0.0.0
lrwxrwxrwx  1 root root   21 Mar 21  2019 libunistring.so.2 -> libunistring.so.2.1.0
-rw-r--r--  1 root root 1.5M Mar 21  2019 libunistring.so.2.1.0
lrwxrwxrwx  1 root root   19 May  1  2020 liburcu-bp.so.6 -> liburcu-bp.so.6.0.0
-rw-r--r--  1 root root  31K May  1  2020 liburcu-bp.so.6.0.0
lrwxrwxrwx  1 root root   20 May  1  2020 liburcu-cds.so.6 -> liburcu-cds.so.6.0.0
-rw-r--r--  1 root root  38K May  1  2020 liburcu-cds.so.6.0.0
lrwxrwxrwx  1 root root   23 May  1  2020 liburcu-common.so.6 -> liburcu-common.so.6.0.0
-rw-r--r--  1 root root  18K May  1  2020 liburcu-common.so.6.0.0
lrwxrwxrwx  1 root root   19 May  1  2020 liburcu-mb.so.6 -> liburcu-mb.so.6.0.0
-rw-r--r--  1 root root  27K May  1  2020 liburcu-mb.so.6.0.0
lrwxrwxrwx  1 root root   21 May  1  2020 liburcu-qsbr.so.6 -> liburcu-qsbr.so.6.0.0
-rw-r--r--  1 root root  27K May  1  2020 liburcu-qsbr.so.6.0.0
lrwxrwxrwx  1 root root   23 May  1  2020 liburcu-signal.so.6 -> liburcu-signal.so.6.0.0
-rw-r--r--  1 root root  31K May  1  2020 liburcu-signal.so.6.0.0
lrwxrwxrwx  1 root root   16 May  1  2020 liburcu.so.6 -> liburcu.so.6.0.0
-rw-r--r--  1 root root  31K May  1  2020 liburcu.so.6.0.0
-rw-r--r--  1 root root  15K May  3 10:19 libutil.a
lrwxrwxrwx  1 root root   34 May  3 10:19 libutil.so -> /lib/x86_64-linux-gnu/libutil.so.1
lrwxrwxrwx  1 root root   16 Dec 15  2017 libwind.so.0 -> libwind.so.0.0.0
-rw-r--r--  1 root root 162K Dec 15  2017 libwind.so.0.0.0
lrwxrwxrwx  1 root root   16 Mar  3  2021 libzstd.so.1 -> libzstd.so.1.3.3
-rw-r--r--  1 root root 490K Mar  3  2021 libzstd.so.1.3.3
drwxr-xr-x  1 root root 4.0K May 20 08:07 perl
drwxr-xr-x 16 root root 4.0K Apr 28 00:03 perl-base
-rw-r--r--  1 root root 1.8K May  3 10:19 rcrt1.o
drwxr-xr-x  2 root root 4.0K May 20 07:57 sasl2
[Pipeline] echo
Packing build/libdgl.so, build/runUnitTests, python/dgl/_ffi/_cy3/core.cpython-*-x86_64-linux-gnu.so, build/tensoradapter/pytorch/*.so into dgl-cugraph-linux
[Pipeline] stash
Stashed 2 file(s)
Post stage
[Pipeline] cleanWs
[WS-CLEANUP] Deleting project workspace...
[WS-CLEANUP] Deferred wipeout is disabled by the job configuration...
[WS-CLEANUP] done
[Pipeline] }
$ docker stop --time=1 4e33dd6e703ed0dd95f590a393973aab145ae4e5b4395a56a04d85836079138d
$ docker rm -f 4e33dd6e703ed0dd95f590a393973aab145ae4e5b4395a56a04d85836079138d
[Pipeline] // withDockerContainer
[Pipeline] }
[Pipeline] // withEnv
[Pipeline] }
[Pipeline] // node
[Pipeline] }
[Pipeline] // stage
[Pipeline] }
[Pipeline] // parallel
[Pipeline] }
[Pipeline] // stage
[Pipeline] stage
[Pipeline] { (Test)
[Pipeline] parallel
[Pipeline] { (Branch: C++ CPU)
[Pipeline] { (Branch: C++ GPU)
[Pipeline] { (Branch: C++ CPU (Win64))
[Pipeline] { (Branch: Tensorflow CPU)
[Pipeline] { (Branch: Tensorflow GPU)
[Pipeline] { (Branch: Torch CPU)
[Pipeline] { (Branch: Torch CPU (Win64))
[Pipeline] { (Branch: Torch GPU)
[Pipeline] { (Branch: Distributed)
[Pipeline] { (Branch: PyTorch Cugraph GPU)
[Pipeline] { (Branch: MXNet CPU)
[Pipeline] { (Branch: MXNet GPU)
[Pipeline] { (Branch: DGL-Go)
[Pipeline] stage
[Pipeline] { (C++ CPU)
[Pipeline] stage
[Pipeline] { (C++ GPU)
[Pipeline] stage
[Pipeline] { (C++ CPU (Win64))
[Pipeline] stage
[Pipeline] { (Tensorflow CPU)
[Pipeline] stage
[Pipeline] { (Tensorflow GPU)
[Pipeline] stage
[Pipeline] { (Torch CPU)
[Pipeline] stage
[Pipeline] { (Torch CPU (Win64))
[Pipeline] stage
[Pipeline] { (Torch GPU)
[Pipeline] stage
[Pipeline] { (Distributed)
[Pipeline] stage
[Pipeline] { (PyTorch Cugraph GPU)
[Pipeline] stage
[Pipeline] { (MXNet CPU)
[Pipeline] stage
[Pipeline] { (MXNet GPU)
[Pipeline] stage
[Pipeline] { (DGL-Go)
[Pipeline] node
[Pipeline] node
[Pipeline] node
[Pipeline] node
[Pipeline] node
[Pipeline] node
[Pipeline] node
[Pipeline] node
[Pipeline] node
[Pipeline] node
[Pipeline] node
[Pipeline] node
[Pipeline] node
Still waiting to schedule task
Waiting for next available executor on ‘windows’
Still waiting to schedule task
Waiting for next available executor on ‘windows’
Still waiting to schedule task
Waiting for next available executor on ‘dgl-manual-large-cpu’
Still waiting to schedule task
Waiting for next available executor on ‘dglci-manual-gpu-worker’
Still waiting to schedule task
Waiting for next available executor on ‘dgl-manual-large-cpu’
Still waiting to schedule task
Waiting for next available executor on ‘dglci-manual-gpu-worker’
Still waiting to schedule task
Waiting for next available executor on ‘dgl-manual-large-cpu’
Still waiting to schedule task
Waiting for next available executor on ‘dglci-manual-gpu-worker’
Still waiting to schedule task
Waiting for next available executor on ‘dgl-manual-large-cpu’
Still waiting to schedule task
Waiting for next available executor on ‘dglci-manual-gpu-worker’
Still waiting to schedule task
Waiting for next available executor on ‘dgl-manual-large-cpu’
Still waiting to schedule task
Waiting for next available executor on ‘dglci-manual-gpu-worker’
Still waiting to schedule task
Waiting for next available executor on ‘dgl-manual-large-cpu’
Running on dglci-windows in C:\Jenkins\workspace\dgl_PR-4648
[Pipeline] {
[Pipeline] checkout
The recommended git tool is: git
using credential 150de63f-189c-4717-bcaf-010460d2f51a
Cloning the remote Git repository
Cloning with configured refspecs honoured and without tags
Cloning repository https://github.com/dmlc/dgl.git
 > git init C:\Jenkins\workspace\dgl_PR-4648 # timeout=10
Fetching upstream changes from https://github.com/dmlc/dgl.git
 > git --version # timeout=10
 > git --version # 'git version 2.20.0.windows.1'
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git fetch --no-tags --force --progress -- https://github.com/dmlc/dgl.git +refs/pull/4648/head:refs/remotes/origin/PR-4648 +refs/heads/master:refs/remotes/origin/master # timeout=10
Cleaning workspace
Fetching without tags
Merging remotes/origin/master commit d78a3a4baf611b90871a849f58647160c7cd9ab4 into PR head commit 37ed78035f406940ed88a4b60cadd2b7cfde5fea
 > git config remote.origin.url https://github.com/dmlc/dgl.git # timeout=10
 > git config --add remote.origin.fetch +refs/pull/4648/head:refs/remotes/origin/PR-4648 # timeout=10
 > git config --add remote.origin.fetch +refs/heads/master:refs/remotes/origin/master # timeout=10
 > git config remote.origin.url https://github.com/dmlc/dgl.git # timeout=10
 > git rev-parse --verify HEAD # timeout=10
No valid HEAD. Skipping the resetting
 > git clean -fdx # timeout=10
Fetching upstream changes from https://github.com/dmlc/dgl.git
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git fetch --no-tags --force --progress -- https://github.com/dmlc/dgl.git +refs/pull/4648/head:refs/remotes/origin/PR-4648 +refs/heads/master:refs/remotes/origin/master # timeout=10
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
Merge succeeded, producing 37ed78035f406940ed88a4b60cadd2b7cfde5fea
Checking out Revision 37ed78035f406940ed88a4b60cadd2b7cfde5fea (PR-4648)
Commit message: "fix for pytorch < 1.12"
Cleaning workspace
[Pipeline] withEnv
[Pipeline] {
 > git remote # timeout=10
 > git config --get remote.origin.url # timeout=10
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git merge d78a3a4baf611b90871a849f58647160c7cd9ab4 # timeout=10
 > git rev-parse "HEAD^{commit}" # timeout=10
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git branch -a -v --no-abbrev # timeout=10
 > git checkout -b PR-4648 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git rev-parse --verify HEAD # timeout=10
Resetting working tree
 > git reset --hard # timeout=10
 > git clean -fdx # timeout=10
[Pipeline] checkout
The recommended git tool is: git
using credential 150de63f-189c-4717-bcaf-010460d2f51a
Fetching changes from the remote Git repository
Cleaning workspace
Fetching without tags
Merging remotes/origin/master commit d78a3a4baf611b90871a849f58647160c7cd9ab4 into PR head commit 37ed78035f406940ed88a4b60cadd2b7cfde5fea
Merge succeeded, producing 37ed78035f406940ed88a4b60cadd2b7cfde5fea
Checking out Revision 37ed78035f406940ed88a4b60cadd2b7cfde5fea (PR-4648)
Commit message: "fix for pytorch < 1.12"
Cleaning workspace
 > git rev-parse --resolve-git-dir C:\Jenkins\workspace\dgl_PR-4648\.git # timeout=10
 > git config remote.origin.url https://github.com/dmlc/dgl.git # timeout=10
 > git rev-parse --verify HEAD # timeout=10
Resetting working tree
 > git reset --hard # timeout=10
 > git clean -fdx # timeout=10
Fetching upstream changes from https://github.com/dmlc/dgl.git
 > git --version # timeout=10
 > git --version # 'git version 2.20.0.windows.1'
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git fetch --no-tags --force --progress -- https://github.com/dmlc/dgl.git +refs/pull/4648/head:refs/remotes/origin/PR-4648 +refs/heads/master:refs/remotes/origin/master # timeout=10
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git remote # timeout=10
 > git config --get remote.origin.url # timeout=10
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git merge d78a3a4baf611b90871a849f58647160c7cd9ab4 # timeout=10
 > git rev-parse "HEAD^{commit}" # timeout=10
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git branch -a -v --no-abbrev # timeout=10
 > git branch -D PR-4648 # timeout=10
 > git checkout -b PR-4648 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git rev-parse --verify HEAD # timeout=10
Resetting working tree
 > git reset --hard # timeout=10
 > git clean -fdx # timeout=10
[Pipeline] bat

administrator@WIN-O7QS55HVSDB C:\Jenkins\workspace\dgl_PR-4648>git submodule update --recursive --init 
Submodule 'third_party/METIS' (https://github.com/KarypisLab/METIS.git) registered for path 'third_party/METIS'
Submodule 'third_party/dlpack' (https://github.com/dmlc/dlpack.git) registered for path 'third_party/dlpack'
Submodule 'third_party/dmlc-core' (https://github.com/dmlc/dmlc-core.git) registered for path 'third_party/dmlc-core'
Submodule 'third_party/googletest' (https://github.com/google/googletest.git) registered for path 'third_party/googletest'
Submodule 'third_party/libxsmm' (https://github.com/hfp/libxsmm.git) registered for path 'third_party/libxsmm'
Submodule 'third_party/nanoflann' (https://github.com/jlblancoc/nanoflann) registered for path 'third_party/nanoflann'
Submodule 'third_party/nccl' (https://github.com/nvidia/nccl) registered for path 'third_party/nccl'
Submodule 'third_party/phmap' (https://github.com/greg7mdp/parallel-hashmap.git) registered for path 'third_party/phmap'
Submodule 'third_party/tensorpipe' (https://github.com/pytorch/tensorpipe) registered for path 'third_party/tensorpipe'
Submodule 'third_party/thrust' (https://github.com/NVIDIA/thrust.git) registered for path 'third_party/thrust'
Submodule 'third_party/tvm' (https://github.com/apache/incubator-tvm) registered for path 'third_party/tvm'
Submodule 'third_party/xbyak' (https://github.com/herumi/xbyak) registered for path 'third_party/xbyak'
Cloning into 'C:/Jenkins/workspace/dgl_PR-4648/third_party/METIS'...
Cloning into 'C:/Jenkins/workspace/dgl_PR-4648/third_party/dlpack'...
Cloning into 'C:/Jenkins/workspace/dgl_PR-4648/third_party/nanoflann'...
Cloning into 'C:/Jenkins/workspace/dgl_PR-4648/third_party/dmlc-core'...
Cloning into 'C:/Jenkins/workspace/dgl_PR-4648/third_party/nccl'...
Cloning into 'C:/Jenkins/workspace/dgl_PR-4648/third_party/phmap'...
Cloning into 'C:/Jenkins/workspace/dgl_PR-4648/third_party/googletest'...
Cloning into 'C:/Jenkins/workspace/dgl_PR-4648/third_party/tensorpipe'...
Cloning into 'C:/Jenkins/workspace/dgl_PR-4648/third_party/xbyak'...
Cloning into 'C:/Jenkins/workspace/dgl_PR-4648/third_party/libxsmm'...
Cloning into 'C:/Jenkins/workspace/dgl_PR-4648/third_party/thrust'...
Cloning into 'C:/Jenkins/workspace/dgl_PR-4648/third_party/tvm'...
Submodule path 'third_party/METIS': checked out '10603482cc055626ff99d2ac3ab7e73d1119439d'
Submodule 'GKlib' (https://github.com/KarypisLab/GKlib.git) registered for path 'third_party/METIS/GKlib'
Cloning into 'C:/Jenkins/workspace/dgl_PR-4648/third_party/METIS/GKlib'...
Submodule path 'third_party/METIS/GKlib': checked out '62de20c7f97c719abdc96d7c3e266a1f7cb52bc5'
Submodule path 'third_party/dlpack': checked out 'e2bdd3bee8cb6501558042633fa59144cc8b7f5f'
Submodule path 'third_party/dmlc-core': checked out 'bfad207b448480783a1f428ae3d93d87032d8349'
Submodule path 'third_party/googletest': checked out 'f71fb4f9a912ec945401cc49a287a759b6131026'
Submodule path 'third_party/libxsmm': checked out 'fa687556130b6298430f1c0555a14cf79ab6101c'
Submodule path 'third_party/nanoflann': checked out '4c47ca200209550c5628c89803591f8a753c8181'
Submodule path 'third_party/nccl': checked out 'e11238b3029795d33f958b5868d47c90c4f22628'
Submodule path 'third_party/phmap': checked out '25293cefd8b85491b45600c03fe8edf07647553f'
Submodule path 'third_party/tensorpipe': checked out '6042f1a4cbce8eef997f11ed0012de137b317361'
Submodule 'third_party/googletest' (https://github.com/google/googletest.git) registered for path 'third_party/tensorpipe/third_party/googletest'
Submodule 'third_party/libnop' (https://github.com/google/libnop.git) registered for path 'third_party/tensorpipe/third_party/libnop'
Submodule 'third_party/libuv' (https://github.com/libuv/libuv.git) registered for path 'third_party/tensorpipe/third_party/libuv'
Submodule 'third_party/pybind11' (https://github.com/pybind/pybind11.git) registered for path 'third_party/tensorpipe/third_party/pybind11'
Cloning into 'C:/Jenkins/workspace/dgl_PR-4648/third_party/tensorpipe/third_party/googletest'...
Cloning into 'C:/Jenkins/workspace/dgl_PR-4648/third_party/tensorpipe/third_party/libnop'...
Cloning into 'C:/Jenkins/workspace/dgl_PR-4648/third_party/tensorpipe/third_party/pybind11'...
Cloning into 'C:/Jenkins/workspace/dgl_PR-4648/third_party/tensorpipe/third_party/libuv'...
Submodule path 'third_party/tensorpipe/third_party/googletest': checked out 'aee0f9d9b5b87796ee8a0ab26b7587ec30e8858e'
Submodule path 'third_party/tensorpipe/third_party/libnop': checked out 'aa95422ea8c409e3f078d2ee7708a5f59a8b9fa2'
Submodule path 'third_party/tensorpipe/third_party/libuv': checked out '1dff88e5161cba5c59276d2070d2e304e4dcb242'
Submodule path 'third_party/tensorpipe/third_party/pybind11': checked out 'a23996fce38ff6ccfbcdc09f1e63f2c4be5ea2ef'
Submodule 'tools/clang' (https://github.com/wjakob/clang-cindex-python3) registered for path 'third_party/tensorpipe/third_party/pybind11/tools/clang'
Cloning into 'C:/Jenkins/workspace/dgl_PR-4648/third_party/tensorpipe/third_party/pybind11/tools/clang'...
Submodule path 'third_party/tensorpipe/third_party/pybind11/tools/clang': checked out '6a00cbc4a9b8e68b71caf7f774b3f9c753ae84d5'
Submodule path 'third_party/thrust': checked out '6a3078c64cab0e2f276340fa5dcafa0d758ed890'
Submodule 'cub' (https://github.com/NVIDIA/cub.git) registered for path 'third_party/thrust/dependencies/cub'
Cloning into 'C:/Jenkins/workspace/dgl_PR-4648/third_party/thrust/dependencies/cub'...
Submodule path 'third_party/thrust/dependencies/cub': checked out 'cdaa9558a85e45d849016e5fe7b6e4ee79113f95'
Submodule path 'third_party/tvm': checked out 'b2e418cb109df4cd1f17a2cf2894a1b396a6b838'
Submodule 'dlpack' (https://github.com/dmlc/dlpack) registered for path 'third_party/tvm/3rdparty/dlpack'
Submodule 'dmlc-core' (https://github.com/dmlc/dmlc-core) registered for path 'third_party/tvm/3rdparty/dmlc-core'
Submodule '3rdparty/rang' (https://github.com/agauniyal/rang) registered for path 'third_party/tvm/3rdparty/rang'
Submodule '3rdparty/vta-hw' (https://github.com/apache/incubator-tvm-vta) registered for path 'third_party/tvm/3rdparty/vta-hw'
Cloning into 'C:/Jenkins/workspace/dgl_PR-4648/third_party/tvm/3rdparty/dlpack'...
Cloning into 'C:/Jenkins/workspace/dgl_PR-4648/third_party/tvm/3rdparty/rang'...
Cloning into 'C:/Jenkins/workspace/dgl_PR-4648/third_party/tvm/3rdparty/dmlc-core'...
Cloning into 'C:/Jenkins/workspace/dgl_PR-4648/third_party/tvm/3rdparty/vta-hw'...
Submodule path 'third_party/tvm/3rdparty/dlpack': checked out '3ec04430e89a6834e5a1b99471f415fa939bf642'
Submodule path 'third_party/tvm/3rdparty/dmlc-core': checked out '6c401e242c59a1f4c913918246591bb13fd714e7'
Submodule path 'third_party/tvm/3rdparty/rang': checked out 'cabe04d6d6b05356fa8f9741704924788f0dd762'
Submodule path 'third_party/tvm/3rdparty/vta-hw': checked out '87ce9acfae550d1a487746e9d06c2e250076e54c'
Submodule path 'third_party/xbyak': checked out '757e4063f6464740b8ff4a2cae9136d2f8458020'
[Pipeline] unstash
[Pipeline] echo
Unpacked build\dgl.dll, build\runUnitTests.exe, build\tensoradapter\pytorch\*.dll from dgl-cpu-win64
[Pipeline] bat

administrator@WIN-O7QS55HVSDB C:\Jenkins\workspace\dgl_PR-4648>CALL tests\scripts\task_cpp_unit_test.bat 
Running main() from C:\Jenkins\workspace\dgl_PR-4648\third_party\googletest\googletest\src\gtest_main.cc
[==========] Running 89 tests from 21 test suites.
[----------] Global test environment set-up.
[----------] 1 test from GraphTest
[ RUN      ] GraphTest.TestNumVertices
[       OK ] GraphTest.TestNumVertices (0 ms)
[----------] 1 test from GraphTest (0 ms total)

[----------] 3 tests from MessageQueueTest
[ RUN      ] MessageQueueTest.AddRemove
[05:02:13] C:\Jenkins\workspace\dgl_PR-4648\src\rpc\network\msg_queue.cc:27: Message is larger than the queue.
[       OK ] MessageQueueTest.AddRemove (0 ms)
[ RUN      ] MessageQueueTest.EmptyAndNoMoreAdd
[       OK ] MessageQueueTest.EmptyAndNoMoreAdd (0 ms)
[ RUN      ] MessageQueueTest.MultiThread
[       OK ] MessageQueueTest.MultiThread (9 ms)
[----------] 3 tests from MessageQueueTest (9 ms total)

[----------] 1 test from SocketCommunicatorTest
[ RUN      ] SocketCommunicatorTest.SendAndRecv
[       OK ] SocketCommunicatorTest.SendAndRecv (6006 ms)
[----------] 1 test from SocketCommunicatorTest (6006 ms total)

[----------] 3 tests from SplitStringTest
[ RUN      ] SplitStringTest.SplitStringUsingCompoundDelim
[       OK ] SplitStringTest.SplitStringUsingCompoundDelim (0 ms)
[ RUN      ] SplitStringTest.testSplitStringUsingSingleDelim
[       OK ] SplitStringTest.testSplitStringUsingSingleDelim (0 ms)
[ RUN      ] SplitStringTest.testSplitingNoDelimString
[       OK ] SplitStringTest.testSplitingNoDelimString (0 ms)
[----------] 3 tests from SplitStringTest (0 ms total)

[----------] 1 test from StringPrintf
[ RUN      ] StringPrintf.normal
[       OK ] StringPrintf.normal (0 ms)
[----------] 1 test from StringPrintf (0 ms total)

[----------] 13 tests from ArrayTest
[ RUN      ] ArrayTest.TestCreate
[       OK ] ArrayTest.TestCreate (0 ms)
[ RUN      ] ArrayTest.TestRange
[       OK ] ArrayTest.TestRange (0 ms)
[ RUN      ] ArrayTest.TestFull
[       OK ] ArrayTest.TestFull (0 ms)
[ RUN      ] ArrayTest.TestClone
[       OK ] ArrayTest.TestClone (0 ms)
[ RUN      ] ArrayTest.TestAsNumBits
[       OK ] ArrayTest.TestAsNumBits (0 ms)
[ RUN      ] ArrayTest.Arith
[       OK ] ArrayTest.Arith (1 ms)
[ RUN      ] ArrayTest.HStack
[       OK ] ArrayTest.HStack (0 ms)
[ RUN      ] ArrayTest.TestIndexSelect
[       OK ] ArrayTest.TestIndexSelect (0 ms)
[ RUN      ] ArrayTest.TestRelabel_
[       OK ] ArrayTest.TestRelabel_ (0 ms)
[ RUN      ] ArrayTest.CumSum
[       OK ] ArrayTest.CumSum (0 ms)
[ RUN      ] ArrayTest.Scatter_
[       OK ] ArrayTest.Scatter_ (0 ms)
[ RUN      ] ArrayTest.NonZero
[       OK ] ArrayTest.NonZero (0 ms)
[ RUN      ] ArrayTest.Sort
[       OK ] ArrayTest.Sort (0 ms)
[----------] 13 tests from ArrayTest (1 ms total)

[----------] 2 tests from MatrixTest
[ RUN      ] MatrixTest.TestToSimpleCsr
[       OK ] MatrixTest.TestToSimpleCsr (1 ms)
[ RUN      ] MatrixTest.TestToSimpleCoo
[       OK ] MatrixTest.TestToSimpleCoo (0 ms)
[----------] 2 tests from MatrixTest (1 ms total)

[----------] 2 tests from DisjointUnionTest
[ RUN      ] DisjointUnionTest.TestDisjointUnionPartitionCoo
[       OK ] DisjointUnionTest.TestDisjointUnionPartitionCoo (0 ms)
[ RUN      ] DisjointUnionTest.TestDisjointUnionPartitionCsr
[       OK ] DisjointUnionTest.TestDisjointUnionPartitionCsr (1 ms)
[----------] 2 tests from DisjointUnionTest (1 ms total)

[----------] 2 tests from SliceContiguousChunk
[ RUN      ] SliceContiguousChunk.TestSliceContiguousChunkCoo
[       OK ] SliceContiguousChunk.TestSliceContiguousChunkCoo (0 ms)
[ RUN      ] SliceContiguousChunk.TestSliceContiguousChunkCsr
[       OK ] SliceContiguousChunk.TestSliceContiguousChunkCsr (0 ms)
[----------] 2 tests from SliceContiguousChunk (0 ms total)

[----------] 2 tests from MatrixUnionTest
[ RUN      ] MatrixUnionTest.TestMatrixUnionCsr
[       OK ] MatrixUnionTest.TestMatrixUnionCsr (0 ms)
[ RUN      ] MatrixUnionTest.TestMatrixUnionCoo
[       OK ] MatrixUnionTest.TestMatrixUnionCoo (0 ms)
[----------] 2 tests from MatrixUnionTest (0 ms total)

[----------] 1 test from LineGraphTest
[ RUN      ] LineGraphTest.LineGraphCOO
[       OK ] LineGraphTest.LineGraphCOO (0 ms)
[----------] 1 test from LineGraphTest (0 ms total)

[----------] 3 tests from CsrmmTest
[ RUN      ] CsrmmTest.TestCsrmm
[       OK ] CsrmmTest.TestCsrmm (1 ms)
[ RUN      ] CsrmmTest.TestCsrsum
[       OK ] CsrmmTest.TestCsrsum (1 ms)
[ RUN      ] CsrmmTest.TestCsrmask
[       OK ] CsrmmTest.TestCsrmask (2 ms)
[----------] 3 tests from CsrmmTest (4 ms total)

[----------] 2 tests from PartitionTest
[ RUN      ] PartitionTest.TestRemainderPartition
[       OK ] PartitionTest.TestRemainderPartition (0 ms)
[ RUN      ] PartitionTest.TestRangePartition
[       OK ] PartitionTest.TestRangePartition (0 ms)
[----------] 2 tests from PartitionTest (0 ms total)

[----------] 11 tests from RowwiseTest
[ RUN      ] RowwiseTest.TestCSRSampling
[       OK ] RowwiseTest.TestCSRSampling (22 ms)
[ RUN      ] RowwiseTest.TestCSRSamplingUniform
[       OK ] RowwiseTest.TestCSRSamplingUniform (2 ms)
[ RUN      ] RowwiseTest.TestCSRPerEtypeSampling
[       OK ] RowwiseTest.TestCSRPerEtypeSampling (11 ms)
[ RUN      ] RowwiseTest.TestCSRPerEtypeSamplingUniform
[       OK ] RowwiseTest.TestCSRPerEtypeSamplingUniform (6 ms)
[ RUN      ] RowwiseTest.TestCOOSampling
[       OK ] RowwiseTest.TestCOOSampling (1855 ms)
[ RUN      ] RowwiseTest.TestCOOSamplingUniform
[       OK ] RowwiseTest.TestCOOSamplingUniform (213 ms)
[ RUN      ] RowwiseTest.TestCOOerEtypeSampling
[       OK ] RowwiseTest.TestCOOerEtypeSampling (4322 ms)
[ RUN      ] RowwiseTest.TestCOOPerEtypeSamplingUniform
[       OK ] RowwiseTest.TestCOOPerEtypeSamplingUniform (2814 ms)
[ RUN      ] RowwiseTest.TestCSRTopk
[       OK ] RowwiseTest.TestCSRTopk (2 ms)
[ RUN      ] RowwiseTest.TestCOOTopk
[       OK ] RowwiseTest.TestCOOTopk (21 ms)
[ RUN      ] RowwiseTest.TestCSRSamplingBiased
[       OK ] RowwiseTest.TestCSRSamplingBiased (2 ms)
[----------] 11 tests from RowwiseTest (9270 ms total)

[----------] 3 tests from SampleUtilsTest
[ RUN      ] SampleUtilsTest.TestWithReplacement
[       OK ] SampleUtilsTest.TestWithReplacement (2230 ms)
[ RUN      ] SampleUtilsTest.TestWithoutReplacementOrder
[       OK ] SampleUtilsTest.TestWithoutReplacementOrder (0 ms)
[ RUN      ] SampleUtilsTest.TestWithoutReplacementUnique
[       OK ] SampleUtilsTest.TestWithoutReplacementUnique (6263 ms)
[----------] 3 tests from SampleUtilsTest (8493 ms total)

[----------] 3 tests from RandomTest
[ RUN      ] RandomTest.TestChoice
[       OK ] RandomTest.TestChoice (1 ms)
[ RUN      ] RandomTest.TestUniformChoice
[       OK ] RandomTest.TestUniformChoice (0 ms)
[ RUN      ] RandomTest.TestBiasedChoice
[       OK ] RandomTest.TestBiasedChoice (34 ms)
[----------] 3 tests from RandomTest (36 ms total)

[----------] 4 tests from Serialize
[ RUN      ] Serialize.UnitGraph_COO
[       OK ] Serialize.UnitGraph_COO (0 ms)
[ RUN      ] Serialize.UnitGraph_CSR
[       OK ] Serialize.UnitGraph_CSR (0 ms)
[ RUN      ] Serialize.ImmutableGraph
[       OK ] Serialize.ImmutableGraph (1 ms)
[ RUN      ] Serialize.HeteroGraph
[       OK ] Serialize.HeteroGraph (2 ms)
[----------] 4 tests from Serialize (3 ms total)

[----------] 3 tests from SmartPtrTest/0, where TypeParam = class std::shared_ptr<class MyClass>
[ RUN      ] SmartPtrTest/0.Obj_Test
[       OK ] SmartPtrTest/0.Obj_Test (0 ms)
[ RUN      ] SmartPtrTest/0.Vector_Test1
[       OK ] SmartPtrTest/0.Vector_Test1 (0 ms)
[ RUN      ] SmartPtrTest/0.Vector_Test2
[       OK ] SmartPtrTest/0.Vector_Test2 (0 ms)
[----------] 3 tests from SmartPtrTest/0 (0 ms total)

[----------] 3 tests from SmartPtrTest/1, where TypeParam = class std::unique_ptr<class MyClass,struct std::default_delete<class MyClass> >
[ RUN      ] SmartPtrTest/1.Obj_Test
[       OK ] SmartPtrTest/1.Obj_Test (0 ms)
[ RUN      ] SmartPtrTest/1.Vector_Test1
[       OK ] SmartPtrTest/1.Vector_Test1 (0 ms)
[ RUN      ] SmartPtrTest/1.Vector_Test2
[       OK ] SmartPtrTest/1.Vector_Test2 (0 ms)
[----------] 3 tests from SmartPtrTest/1 (0 ms total)

[----------] 19 tests from SpmatTest
[ RUN      ] SpmatTest.COOToCSR
[       OK ] SpmatTest.COOToCSR (0 ms)
[ RUN      ] SpmatTest.TestCOOHasDuplicate
[       OK ] SpmatTest.TestCOOHasDuplicate (0 ms)
[ RUN      ] SpmatTest.COOSort
[       OK ] SpmatTest.COOSort (0 ms)
[ RUN      ] SpmatTest.TestCOOReorder
[       OK ] SpmatTest.TestCOOReorder (2 ms)
[ RUN      ] SpmatTest.COOGetData
[       OK ] SpmatTest.COOGetData (7 ms)
[ RUN      ] SpmatTest.COOGetDataAndIndices
[       OK ] SpmatTest.COOGetDataAndIndices (0 ms)
[ RUN      ] SpmatTest.TestCSRIsNonZero
[       OK ] SpmatTest.TestCSRIsNonZero (0 ms)
[ RUN      ] SpmatTest.TestCSRGetRowNNZ
[       OK ] SpmatTest.TestCSRGetRowNNZ (0 ms)
[ RUN      ] SpmatTest.TestCSRGetRowColumnIndices
[       OK ] SpmatTest.TestCSRGetRowColumnIndices (0 ms)
[ RUN      ] SpmatTest.TestCSRGetRowData
[       OK ] SpmatTest.TestCSRGetRowData (0 ms)
[ RUN      ] SpmatTest.CSRGetData
[       OK ] SpmatTest.CSRGetData (1 ms)
[ RUN      ] SpmatTest.CSRGetDataAndIndices
[       OK ] SpmatTest.CSRGetDataAndIndices (0 ms)
[ RUN      ] SpmatTest.CSRTranspose
[       OK ] SpmatTest.CSRTranspose (0 ms)
[ RUN      ] SpmatTest.CSRToCOO
[       OK ] SpmatTest.CSRToCOO (1 ms)
[ RUN      ] SpmatTest.TestCSRSliceRows
[       OK ] SpmatTest.TestCSRSliceRows (30 ms)
[ RUN      ] SpmatTest.CSRSliceMatrix
[       OK ] SpmatTest.CSRSliceMatrix (13 ms)
[ RUN      ] SpmatTest.CSRHasDuplicate
[       OK ] SpmatTest.CSRHasDuplicate (0 ms)
[ RUN      ] SpmatTest.CSRSort
[       OK ] SpmatTest.CSRSort (1 ms)
[ RUN      ] SpmatTest.TestCSRReorder
[       OK ] SpmatTest.TestCSRReorder (0 ms)
[----------] 19 tests from SpmatTest (55 ms total)

[----------] 7 tests from UniGraphTest
[ RUN      ] UniGraphTest.TestUnitGraph_CopyTo
[       OK ] UniGraphTest.TestUnitGraph_CopyTo (0 ms)
[ RUN      ] UniGraphTest.TestUnitGraph_InOutDegrees
[       OK ] UniGraphTest.TestUnitGraph_InOutDegrees (4 ms)
[ RUN      ] UniGraphTest.TestUnitGraph_Create
[       OK ] UniGraphTest.TestUnitGraph_Create (2 ms)
[ RUN      ] UniGraphTest.TestUnitGraph_GetInCSR
[       OK ] UniGraphTest.TestUnitGraph_GetInCSR (4 ms)
[ RUN      ] UniGraphTest.TestUnitGraph_GetOutCSR
[       OK ] UniGraphTest.TestUnitGraph_GetOutCSR (5 ms)
[ RUN      ] UniGraphTest.TestUnitGraph_GetCOO
[       OK ] UniGraphTest.TestUnitGraph_GetCOO (1 ms)
[ RUN      ] UniGraphTest.TestUnitGraph_Reserve
[       OK ] UniGraphTest.TestUnitGraph_Reserve (5 ms)
[----------] 7 tests from UniGraphTest (21 ms total)

[----------] Global test environment tear-down
[==========] 89 tests from 21 test suites ran. (23901 ms total)
[  PASSED  ] 89 tests.
Post stage
[Pipeline] cleanWs
[WS-CLEANUP] Deleting project workspace...
[WS-CLEANUP] Deferred wipeout is disabled by the job configuration...
[WS-CLEANUP] done
[Pipeline] }
[Pipeline] // withEnv
[Pipeline] }
Running on dglci-windows in C:\Jenkins\workspace\dgl_PR-4648
[Pipeline] // node
[Pipeline] {
[Pipeline] }
[Pipeline] // stage
[Pipeline] }
[Pipeline] checkout
The recommended git tool is: git
using credential 150de63f-189c-4717-bcaf-010460d2f51a
Cloning the remote Git repository
Cloning with configured refspecs honoured and without tags
Cloning repository https://github.com/dmlc/dgl.git
 > git init C:\Jenkins\workspace\dgl_PR-4648 # timeout=10
Fetching upstream changes from https://github.com/dmlc/dgl.git
 > git --version # timeout=10
 > git --version # 'git version 2.20.0.windows.1'
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git fetch --no-tags --force --progress -- https://github.com/dmlc/dgl.git +refs/pull/4648/head:refs/remotes/origin/PR-4648 +refs/heads/master:refs/remotes/origin/master # timeout=10
Cleaning workspace
 > git config remote.origin.url https://github.com/dmlc/dgl.git # timeout=10
 > git config --add remote.origin.fetch +refs/pull/4648/head:refs/remotes/origin/PR-4648 # timeout=10
 > git config --add remote.origin.fetch +refs/heads/master:refs/remotes/origin/master # timeout=10
 > git config remote.origin.url https://github.com/dmlc/dgl.git # timeout=10
 > git rev-parse --verify HEAD # timeout=10
No valid HEAD. Skipping the resetting
 > git clean -fdx # timeout=10
Fetching without tags
Merging remotes/origin/master commit d78a3a4baf611b90871a849f58647160c7cd9ab4 into PR head commit 37ed78035f406940ed88a4b60cadd2b7cfde5fea
Merge succeeded, producing 37ed78035f406940ed88a4b60cadd2b7cfde5fea
Checking out Revision 37ed78035f406940ed88a4b60cadd2b7cfde5fea (PR-4648)
Fetching upstream changes from https://github.com/dmlc/dgl.git
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git fetch --no-tags --force --progress -- https://github.com/dmlc/dgl.git +refs/pull/4648/head:refs/remotes/origin/PR-4648 +refs/heads/master:refs/remotes/origin/master # timeout=10
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git remote # timeout=10
 > git config --get remote.origin.url # timeout=10
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git merge d78a3a4baf611b90871a849f58647160c7cd9ab4 # timeout=10
 > git rev-parse "HEAD^{commit}" # timeout=10
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
Commit message: "fix for pytorch < 1.12"
Cleaning workspace
[Pipeline] withEnv
[Pipeline] {
[Pipeline] stage
[Pipeline] { (Torch CPU (Win64) Unit test)
[Pipeline] checkout
The recommended git tool is: git
using credential 150de63f-189c-4717-bcaf-010460d2f51a
Fetching changes from the remote Git repository
Cleaning workspace
 > git branch -a -v --no-abbrev # timeout=10
 > git checkout -b PR-4648 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git rev-parse --verify HEAD # timeout=10
Resetting working tree
 > git reset --hard # timeout=10
 > git clean -fdx # timeout=10
Fetching without tags
Merging remotes/origin/master commit d78a3a4baf611b90871a849f58647160c7cd9ab4 into PR head commit 37ed78035f406940ed88a4b60cadd2b7cfde5fea
Merge succeeded, producing 37ed78035f406940ed88a4b60cadd2b7cfde5fea
Checking out Revision 37ed78035f406940ed88a4b60cadd2b7cfde5fea (PR-4648)
Commit message: "fix for pytorch < 1.12"
Cleaning workspace
 > git rev-parse --resolve-git-dir C:\Jenkins\workspace\dgl_PR-4648\.git # timeout=10
 > git config remote.origin.url https://github.com/dmlc/dgl.git # timeout=10
 > git rev-parse --verify HEAD # timeout=10
Resetting working tree
 > git reset --hard # timeout=10
 > git clean -fdx # timeout=10
Fetching upstream changes from https://github.com/dmlc/dgl.git
 > git --version # timeout=10
 > git --version # 'git version 2.20.0.windows.1'
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git fetch --no-tags --force --progress -- https://github.com/dmlc/dgl.git +refs/pull/4648/head:refs/remotes/origin/PR-4648 +refs/heads/master:refs/remotes/origin/master # timeout=10
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git remote # timeout=10
 > git config --get remote.origin.url # timeout=10
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git merge d78a3a4baf611b90871a849f58647160c7cd9ab4 # timeout=10
 > git rev-parse "HEAD^{commit}" # timeout=10
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git branch -a -v --no-abbrev # timeout=10
 > git branch -D PR-4648 # timeout=10
 > git checkout -b PR-4648 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git rev-parse --verify HEAD # timeout=10
Resetting working tree
 > git reset --hard # timeout=10
 > git clean -fdx # timeout=10
[Pipeline] bat

administrator@WIN-O7QS55HVSDB C:\Jenkins\workspace\dgl_PR-4648>git submodule update --recursive --init 
Submodule 'third_party/METIS' (https://github.com/KarypisLab/METIS.git) registered for path 'third_party/METIS'
Submodule 'third_party/dlpack' (https://github.com/dmlc/dlpack.git) registered for path 'third_party/dlpack'
Submodule 'third_party/dmlc-core' (https://github.com/dmlc/dmlc-core.git) registered for path 'third_party/dmlc-core'
Submodule 'third_party/googletest' (https://github.com/google/googletest.git) registered for path 'third_party/googletest'
Submodule 'third_party/libxsmm' (https://github.com/hfp/libxsmm.git) registered for path 'third_party/libxsmm'
Submodule 'third_party/nanoflann' (https://github.com/jlblancoc/nanoflann) registered for path 'third_party/nanoflann'
Submodule 'third_party/nccl' (https://github.com/nvidia/nccl) registered for path 'third_party/nccl'
Submodule 'third_party/phmap' (https://github.com/greg7mdp/parallel-hashmap.git) registered for path 'third_party/phmap'
Submodule 'third_party/tensorpipe' (https://github.com/pytorch/tensorpipe) registered for path 'third_party/tensorpipe'
Submodule 'third_party/thrust' (https://github.com/NVIDIA/thrust.git) registered for path 'third_party/thrust'
Submodule 'third_party/tvm' (https://github.com/apache/incubator-tvm) registered for path 'third_party/tvm'
Submodule 'third_party/xbyak' (https://github.com/herumi/xbyak) registered for path 'third_party/xbyak'
Cloning into 'C:/Jenkins/workspace/dgl_PR-4648/third_party/METIS'...
Cloning into 'C:/Jenkins/workspace/dgl_PR-4648/third_party/dlpack'...
Cloning into 'C:/Jenkins/workspace/dgl_PR-4648/third_party/dmlc-core'...
Cloning into 'C:/Jenkins/workspace/dgl_PR-4648/third_party/googletest'...
Cloning into 'C:/Jenkins/workspace/dgl_PR-4648/third_party/phmap'...
Cloning into 'C:/Jenkins/workspace/dgl_PR-4648/third_party/nanoflann'...
Cloning into 'C:/Jenkins/workspace/dgl_PR-4648/third_party/nccl'...
Cloning into 'C:/Jenkins/workspace/dgl_PR-4648/third_party/tensorpipe'...
Cloning into 'C:/Jenkins/workspace/dgl_PR-4648/third_party/xbyak'...
Cloning into 'C:/Jenkins/workspace/dgl_PR-4648/third_party/libxsmm'...
Cloning into 'C:/Jenkins/workspace/dgl_PR-4648/third_party/thrust'...
Cloning into 'C:/Jenkins/workspace/dgl_PR-4648/third_party/tvm'...
Submodule path 'third_party/METIS': checked out '10603482cc055626ff99d2ac3ab7e73d1119439d'
Submodule 'GKlib' (https://github.com/KarypisLab/GKlib.git) registered for path 'third_party/METIS/GKlib'
Cloning into 'C:/Jenkins/workspace/dgl_PR-4648/third_party/METIS/GKlib'...
Submodule path 'third_party/METIS/GKlib': checked out '62de20c7f97c719abdc96d7c3e266a1f7cb52bc5'
Submodule path 'third_party/dlpack': checked out 'e2bdd3bee8cb6501558042633fa59144cc8b7f5f'
Submodule path 'third_party/dmlc-core': checked out 'bfad207b448480783a1f428ae3d93d87032d8349'
Submodule path 'third_party/googletest': checked out 'f71fb4f9a912ec945401cc49a287a759b6131026'
Submodule path 'third_party/libxsmm': checked out 'fa687556130b6298430f1c0555a14cf79ab6101c'
Submodule path 'third_party/nanoflann': checked out '4c47ca200209550c5628c89803591f8a753c8181'
Submodule path 'third_party/nccl': checked out 'e11238b3029795d33f958b5868d47c90c4f22628'
Submodule path 'third_party/phmap': checked out '25293cefd8b85491b45600c03fe8edf07647553f'
Submodule path 'third_party/tensorpipe': checked out '6042f1a4cbce8eef997f11ed0012de137b317361'
Submodule 'third_party/googletest' (https://github.com/google/googletest.git) registered for path 'third_party/tensorpipe/third_party/googletest'
Submodule 'third_party/libnop' (https://github.com/google/libnop.git) registered for path 'third_party/tensorpipe/third_party/libnop'
Submodule 'third_party/libuv' (https://github.com/libuv/libuv.git) registered for path 'third_party/tensorpipe/third_party/libuv'
Submodule 'third_party/pybind11' (https://github.com/pybind/pybind11.git) registered for path 'third_party/tensorpipe/third_party/pybind11'
Cloning into 'C:/Jenkins/workspace/dgl_PR-4648/third_party/tensorpipe/third_party/googletest'...
Cloning into 'C:/Jenkins/workspace/dgl_PR-4648/third_party/tensorpipe/third_party/libnop'...
Cloning into 'C:/Jenkins/workspace/dgl_PR-4648/third_party/tensorpipe/third_party/libuv'...
Cloning into 'C:/Jenkins/workspace/dgl_PR-4648/third_party/tensorpipe/third_party/pybind11'...
Submodule path 'third_party/tensorpipe/third_party/googletest': checked out 'aee0f9d9b5b87796ee8a0ab26b7587ec30e8858e'
Submodule path 'third_party/tensorpipe/third_party/libnop': checked out 'aa95422ea8c409e3f078d2ee7708a5f59a8b9fa2'
Submodule path 'third_party/tensorpipe/third_party/libuv': checked out '1dff88e5161cba5c59276d2070d2e304e4dcb242'
Submodule path 'third_party/tensorpipe/third_party/pybind11': checked out 'a23996fce38ff6ccfbcdc09f1e63f2c4be5ea2ef'
Submodule 'tools/clang' (https://github.com/wjakob/clang-cindex-python3) registered for path 'third_party/tensorpipe/third_party/pybind11/tools/clang'
Cloning into 'C:/Jenkins/workspace/dgl_PR-4648/third_party/tensorpipe/third_party/pybind11/tools/clang'...
Submodule path 'third_party/tensorpipe/third_party/pybind11/tools/clang': checked out '6a00cbc4a9b8e68b71caf7f774b3f9c753ae84d5'
Submodule path 'third_party/thrust': checked out '6a3078c64cab0e2f276340fa5dcafa0d758ed890'
Submodule 'cub' (https://github.com/NVIDIA/cub.git) registered for path 'third_party/thrust/dependencies/cub'
Cloning into 'C:/Jenkins/workspace/dgl_PR-4648/third_party/thrust/dependencies/cub'...
Submodule path 'third_party/thrust/dependencies/cub': checked out 'cdaa9558a85e45d849016e5fe7b6e4ee79113f95'
Submodule path 'third_party/tvm': checked out 'b2e418cb109df4cd1f17a2cf2894a1b396a6b838'
Submodule 'dlpack' (https://github.com/dmlc/dlpack) registered for path 'third_party/tvm/3rdparty/dlpack'
Submodule 'dmlc-core' (https://github.com/dmlc/dmlc-core) registered for path 'third_party/tvm/3rdparty/dmlc-core'
Submodule '3rdparty/rang' (https://github.com/agauniyal/rang) registered for path 'third_party/tvm/3rdparty/rang'
Submodule '3rdparty/vta-hw' (https://github.com/apache/incubator-tvm-vta) registered for path 'third_party/tvm/3rdparty/vta-hw'
Cloning into 'C:/Jenkins/workspace/dgl_PR-4648/third_party/tvm/3rdparty/dlpack'...
Cloning into 'C:/Jenkins/workspace/dgl_PR-4648/third_party/tvm/3rdparty/rang'...
Cloning into 'C:/Jenkins/workspace/dgl_PR-4648/third_party/tvm/3rdparty/dmlc-core'...
Cloning into 'C:/Jenkins/workspace/dgl_PR-4648/third_party/tvm/3rdparty/vta-hw'...
Submodule path 'third_party/tvm/3rdparty/dlpack': checked out '3ec04430e89a6834e5a1b99471f415fa939bf642'
Submodule path 'third_party/tvm/3rdparty/dmlc-core': checked out '6c401e242c59a1f4c913918246591bb13fd714e7'
Submodule path 'third_party/tvm/3rdparty/rang': checked out 'cabe04d6d6b05356fa8f9741704924788f0dd762'
Submodule path 'third_party/tvm/3rdparty/vta-hw': checked out '87ce9acfae550d1a487746e9d06c2e250076e54c'
Submodule path 'third_party/xbyak': checked out '757e4063f6464740b8ff4a2cae9136d2f8458020'
[Pipeline] unstash
[Pipeline] echo
Unpacked build\dgl.dll, build\runUnitTests.exe, build\tensoradapter\pytorch\*.dll from dgl-cpu-win64
[Pipeline] timeout
Timeout set to expire in 30 min
[Pipeline] {
[Pipeline] bat

administrator@WIN-O7QS55HVSDB C:\Jenkins\workspace\dgl_PR-4648>CALL tests\scripts\task_unit_test.bat pytorch 
Requirement already satisfied: pytest in c:\program files\python36\lib\site-packages (7.0.1)
Requirement already satisfied: psutil in c:\users\administrator\envs\jenkins-dgl-pr-4648-2\lib\site-packages (5.9.2)
Requirement already satisfied: pandas in c:\program files\python36\lib\site-packages (1.1.5)
Requirement already satisfied: pyyaml in c:\program files\python36\lib\site-packages (6.0)
Requirement already satisfied: pydantic in c:\program files\python36\lib\site-packages (1.9.0)
Collecting rdflib
  Using cached rdflib-5.0.0-py3-none-any.whl (231 kB)
Requirement already satisfied: colorama in c:\program files\python36\lib\site-packages (from pytest) (0.4.3)
Requirement already satisfied: tomli>=1.0.0 in c:\program files\python36\lib\site-packages (from pytest) (1.2.3)
Requirement already satisfied: py>=1.8.2 in c:\program files\python36\lib\site-packages (from pytest) (1.9.0)
Requirement already satisfied: iniconfig in c:\program files\python36\lib\site-packages (from pytest) (1.0.1)
Requirement already satisfied: pluggy<2.0,>=0.12 in c:\program files\python36\lib\site-packages (from pytest) (0.13.1)
Requirement already satisfied: packaging in c:\program files\python36\lib\site-packages (from pytest) (20.4)
Requirement already satisfied: importlib-metadata>=0.12 in c:\program files\python36\lib\site-packages (from pytest) (1.7.0)
Requirement already satisfied: atomicwrites>=1.0 in c:\program files\python36\lib\site-packages (from pytest) (1.4.0)
Requirement already satisfied: attrs>=19.2.0 in c:\program files\python36\lib\site-packages (from pytest) (19.3.0)
Requirement already satisfied: numpy>=1.15.4 in c:\program files\python36\lib\site-packages (from pandas) (1.19.5)
Requirement already satisfied: pytz>=2017.2 in c:\program files\python36\lib\site-packages (from pandas) (2021.1)
Requirement already satisfied: python-dateutil>=2.7.3 in c:\program files\python36\lib\site-packages (from pandas) (2.8.1)
Requirement already satisfied: dataclasses>=0.6 in c:\program files\python36\lib\site-packages (from pydantic) (0.8)
Requirement already satisfied: typing-extensions>=3.7.4.3 in c:\program files\python36\lib\site-packages (from pydantic) (4.0.1)
Collecting isodate
  Using cached isodate-0.6.1-py2.py3-none-any.whl (41 kB)
Requirement already satisfied: pyparsing in c:\program files\python36\lib\site-packages (from rdflib) (2.4.7)
Requirement already satisfied: six in c:\program files\python36\lib\site-packages (from rdflib) (1.15.0)
Requirement already satisfied: zipp>=0.5 in c:\program files\python36\lib\site-packages (from importlib-metadata>=0.12->pytest) (3.1.0)
Installing collected packages: isodate, rdflib
Successfully installed isodate-0.6.1 rdflib-5.0.0
============================= test session starts =============================
platform win32 -- Python 3.6.7, pytest-7.0.1, pluggy-0.13.1 -- C:\Users\Administrator\Envs\jenkins-dgl-PR-4648-2\Scripts\python.exe
cachedir: .pytest_cache
rootdir: C:\Jenkins\workspace\dgl_PR-4648
collecting ... collected 3978 items

tests/pytorch/test_dataloader.py::test_graph_dataloader PASSED           [  0%]
tests/pytorch/test_dataloader.py::test_cluster_gcn[0] SKIPPED (Do no...) [  0%]
tests/pytorch/test_dataloader.py::test_cluster_gcn[4] SKIPPED (Do no...) [  0%]
tests/pytorch/test_dataloader.py::test_shadow[0] PASSED                  [  0%]
tests/pytorch/test_dataloader.py::test_shadow[4] PASSED                  [  0%]
tests/pytorch/test_dataloader.py::test_saint[node-0] PASSED              [  0%]
tests/pytorch/test_dataloader.py::test_saint[node-4] PASSED              [  0%]
tests/pytorch/test_dataloader.py::test_saint[edge-0] PASSED              [  0%]
tests/pytorch/test_dataloader.py::test_saint[edge-4] PASSED              [  0%]
tests/pytorch/test_dataloader.py::test_saint[walk-0] PASSED              [  0%]
tests/pytorch/test_dataloader.py::test_saint[walk-4] PASSED              [  0%]
tests/pytorch/test_dataloader.py::test_neighbor_nonuniform[False-cpu-idtype0] PASSED [  0%]
tests/pytorch/test_dataloader.py::test_neighbor_nonuniform[False-cpu-idtype1] PASSED [  0%]
tests/pytorch/test_dataloader.py::test_neighbor_nonuniform[False-uva_cuda_indices-idtype0] SKIPPED [  0%]
tests/pytorch/test_dataloader.py::test_neighbor_nonuniform[False-uva_cuda_indices-idtype1] SKIPPED [  0%]
tests/pytorch/test_dataloader.py::test_neighbor_nonuniform[False-uva_cpu_indices-idtype0] SKIPPED [  0%]
tests/pytorch/test_dataloader.py::test_neighbor_nonuniform[False-uva_cpu_indices-idtype1] SKIPPED [  0%]
tests/pytorch/test_dataloader.py::test_neighbor_nonuniform[False-pure_gpu-idtype0] SKIPPED [  0%]
tests/pytorch/test_dataloader.py::test_neighbor_nonuniform[False-pure_gpu-idtype1] SKIPPED [  0%]
tests/pytorch/test_dataloader.py::test_neighbor_nonuniform[True-cpu-idtype0] PASSED [  0%]
tests/pytorch/test_dataloader.py::test_neighbor_nonuniform[True-cpu-idtype1] PASSED [  0%]
tests/pytorch/test_dataloader.py::test_neighbor_nonuniform[True-uva_cuda_indices-idtype0] SKIPPED [  0%]
tests/pytorch/test_dataloader.py::test_neighbor_nonuniform[True-uva_cuda_indices-idtype1] SKIPPED [  0%]
tests/pytorch/test_dataloader.py::test_neighbor_nonuniform[True-uva_cpu_indices-idtype0] SKIPPED [  0%]
tests/pytorch/test_dataloader.py::test_neighbor_nonuniform[True-uva_cpu_indices-idtype1] SKIPPED [  0%]
tests/pytorch/test_dataloader.py::test_neighbor_nonuniform[True-pure_gpu-idtype0] SKIPPED [  0%]
tests/pytorch/test_dataloader.py::test_neighbor_nonuniform[True-pure_gpu-idtype1] SKIPPED [  0%]
tests/pytorch/test_dataloader.py::test_node_dataloader[False-cpu-full-idtype0] PASSED [  0%]
tests/pytorch/test_dataloader.py::test_node_dataloader[False-cpu-full-idtype1] PASSED [  0%]
tests/pytorch/test_dataloader.py::test_node_dataloader[False-cpu-neighbor-idtype0] PASSED [  0%]
tests/pytorch/test_dataloader.py::test_node_dataloader[False-cpu-neighbor-idtype1] PASSED [  0%]
tests/pytorch/test_dataloader.py::test_node_dataloader[False-cpu-neighbor2-idtype0] PASSED [  0%]
tests/pytorch/test_dataloader.py::test_node_dataloader[False-cpu-neighbor2-idtype1] PASSED [  0%]
tests/pytorch/test_dataloader.py::test_node_dataloader[False-uva_cuda_indices-full-idtype0] SKIPPED [  0%]
tests/pytorch/test_dataloader.py::test_node_dataloader[False-uva_cuda_indices-full-idtype1] SKIPPED [  0%]
tests/pytorch/test_dataloader.py::test_node_dataloader[False-uva_cuda_indices-neighbor-idtype0] SKIPPED [  0%]
tests/pytorch/test_dataloader.py::test_node_dataloader[False-uva_cuda_indices-neighbor-idtype1] SKIPPED [  0%]
tests/pytorch/test_dataloader.py::test_node_dataloader[False-uva_cuda_indices-neighbor2-idtype0] SKIPPED [  0%]
tests/pytorch/test_dataloader.py::test_node_dataloader[False-uva_cuda_indices-neighbor2-idtype1] SKIPPED [  0%]
tests/pytorch/test_dataloader.py::test_node_dataloader[False-uva_cpu_indices-full-idtype0] SKIPPED [  1%]
tests/pytorch/test_dataloader.py::test_node_dataloader[False-uva_cpu_indices-full-idtype1] SKIPPED [  1%]
tests/pytorch/test_dataloader.py::test_node_dataloader[False-uva_cpu_indices-neighbor-idtype0] SKIPPED [  1%]
tests/pytorch/test_dataloader.py::test_node_dataloader[False-uva_cpu_indices-neighbor-idtype1] SKIPPED [  1%]
tests/pytorch/test_dataloader.py::test_node_dataloader[False-uva_cpu_indices-neighbor2-idtype0] SKIPPED [  1%]
tests/pytorch/test_dataloader.py::test_node_dataloader[False-uva_cpu_indices-neighbor2-idtype1] SKIPPED [  1%]
tests/pytorch/test_dataloader.py::test_node_dataloader[False-pure_gpu-full-idtype0] SKIPPED [  1%]
tests/pytorch/test_dataloader.py::test_node_dataloader[False-pure_gpu-full-idtype1] SKIPPED [  1%]
tests/pytorch/test_dataloader.py::test_node_dataloader[False-pure_gpu-neighbor-idtype0] SKIPPED [  1%]
tests/pytorch/test_dataloader.py::test_node_dataloader[False-pure_gpu-neighbor-idtype1] SKIPPED [  1%]
tests/pytorch/test_dataloader.py::test_node_dataloader[False-pure_gpu-neighbor2-idtype0] SKIPPED [  1%]
tests/pytorch/test_dataloader.py::test_node_dataloader[False-pure_gpu-neighbor2-idtype1] SKIPPED [  1%]
tests/pytorch/test_dataloader.py::test_node_dataloader[True-cpu-full-idtype0] PASSED [  1%]
tests/pytorch/test_dataloader.py::test_node_dataloader[True-cpu-full-idtype1] PASSED [  1%]
tests/pytorch/test_dataloader.py::test_node_dataloader[True-cpu-neighbor-idtype0] PASSED [  1%]
tests/pytorch/test_dataloader.py::test_node_dataloader[True-cpu-neighbor-idtype1] PASSED [  1%]
tests/pytorch/test_dataloader.py::test_node_dataloader[True-cpu-neighbor2-idtype0] PASSED [  1%]
tests/pytorch/test_dataloader.py::test_node_dataloader[True-cpu-neighbor2-idtype1] PASSED [  1%]
tests/pytorch/test_dataloader.py::test_node_dataloader[True-uva_cuda_indices-full-idtype0] SKIPPED [  1%]
tests/pytorch/test_dataloader.py::test_node_dataloader[True-uva_cuda_indices-full-idtype1] SKIPPED [  1%]
tests/pytorch/test_dataloader.py::test_node_dataloader[True-uva_cuda_indices-neighbor-idtype0] SKIPPED [  1%]
tests/pytorch/test_dataloader.py::test_node_dataloader[True-uva_cuda_indices-neighbor-idtype1] SKIPPED [  1%]
tests/pytorch/test_dataloader.py::test_node_dataloader[True-uva_cuda_indices-neighbor2-idtype0] SKIPPED [  1%]
tests/pytorch/test_dataloader.py::test_node_dataloader[True-uva_cuda_indices-neighbor2-idtype1] SKIPPED [  1%]
tests/pytorch/test_dataloader.py::test_node_dataloader[True-uva_cpu_indices-full-idtype0] SKIPPED [  1%]
tests/pytorch/test_dataloader.py::test_node_dataloader[True-uva_cpu_indices-full-idtype1] SKIPPED [  1%]
tests/pytorch/test_dataloader.py::test_node_dataloader[True-uva_cpu_indices-neighbor-idtype0] SKIPPED [  1%]
tests/pytorch/test_dataloader.py::test_node_dataloader[True-uva_cpu_indices-neighbor-idtype1] SKIPPED [  1%]
tests/pytorch/test_dataloader.py::test_node_dataloader[True-uva_cpu_indices-neighbor2-idtype0] SKIPPED [  1%]
tests/pytorch/test_dataloader.py::test_node_dataloader[True-uva_cpu_indices-neighbor2-idtype1] SKIPPED [  1%]
tests/pytorch/test_dataloader.py::test_node_dataloader[True-pure_gpu-full-idtype0] SKIPPED [  1%]
tests/pytorch/test_dataloader.py::test_node_dataloader[True-pure_gpu-full-idtype1] SKIPPED [  1%]
tests/pytorch/test_dataloader.py::test_node_dataloader[True-pure_gpu-neighbor-idtype0] SKIPPED [  1%]
tests/pytorch/test_dataloader.py::test_node_dataloader[True-pure_gpu-neighbor-idtype1] SKIPPED [  1%]
tests/pytorch/test_dataloader.py::test_node_dataloader[True-pure_gpu-neighbor2-idtype0] SKIPPED [  1%]
tests/pytorch/test_dataloader.py::test_node_dataloader[True-pure_gpu-neighbor2-idtype1] SKIPPED [  1%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[False-cpu-neg_sampler0-full-idtype0] PASSED [  1%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[False-cpu-neg_sampler0-full-idtype1] PASSED [  1%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[False-cpu-neg_sampler0-neighbor-idtype0] PASSED [  1%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[False-cpu-neg_sampler0-neighbor-idtype1] PASSED [  1%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[False-cpu-neg_sampler1-full-idtype0] PASSED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[False-cpu-neg_sampler1-full-idtype1] PASSED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[False-cpu-neg_sampler1-neighbor-idtype0] PASSED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[False-cpu-neg_sampler1-neighbor-idtype1] PASSED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[False-cpu-neg_sampler2-full-idtype0] PASSED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[False-cpu-neg_sampler2-full-idtype1] PASSED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[False-cpu-neg_sampler2-neighbor-idtype0] PASSED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[False-cpu-neg_sampler2-neighbor-idtype1] PASSED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[False-uva-neg_sampler0-full-idtype0] SKIPPED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[False-uva-neg_sampler0-full-idtype1] SKIPPED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[False-uva-neg_sampler0-neighbor-idtype0] SKIPPED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[False-uva-neg_sampler0-neighbor-idtype1] SKIPPED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[False-uva-neg_sampler1-full-idtype0] SKIPPED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[False-uva-neg_sampler1-full-idtype1] SKIPPED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[False-uva-neg_sampler1-neighbor-idtype0] SKIPPED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[False-uva-neg_sampler1-neighbor-idtype1] SKIPPED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[False-uva-neg_sampler2-full-idtype0] SKIPPED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[False-uva-neg_sampler2-full-idtype1] SKIPPED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[False-uva-neg_sampler2-neighbor-idtype0] SKIPPED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[False-uva-neg_sampler2-neighbor-idtype1] SKIPPED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[False-pure_gpu-neg_sampler0-full-idtype0] SKIPPED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[False-pure_gpu-neg_sampler0-full-idtype1] SKIPPED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[False-pure_gpu-neg_sampler0-neighbor-idtype0] SKIPPED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[False-pure_gpu-neg_sampler0-neighbor-idtype1] SKIPPED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[False-pure_gpu-neg_sampler1-full-idtype0] SKIPPED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[False-pure_gpu-neg_sampler1-full-idtype1] SKIPPED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[False-pure_gpu-neg_sampler1-neighbor-idtype0] SKIPPED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[False-pure_gpu-neg_sampler1-neighbor-idtype1] SKIPPED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[False-pure_gpu-neg_sampler2-full-idtype0] SKIPPED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[False-pure_gpu-neg_sampler2-full-idtype1] SKIPPED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[False-pure_gpu-neg_sampler2-neighbor-idtype0] SKIPPED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[False-pure_gpu-neg_sampler2-neighbor-idtype1] SKIPPED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[True-cpu-neg_sampler0-full-idtype0] PASSED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[True-cpu-neg_sampler0-full-idtype1] PASSED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[True-cpu-neg_sampler0-neighbor-idtype0] PASSED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[True-cpu-neg_sampler0-neighbor-idtype1] PASSED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[True-cpu-neg_sampler1-full-idtype0] PASSED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[True-cpu-neg_sampler1-full-idtype1] PASSED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[True-cpu-neg_sampler1-neighbor-idtype0] PASSED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[True-cpu-neg_sampler1-neighbor-idtype1] PASSED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[True-cpu-neg_sampler2-full-idtype0] PASSED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[True-cpu-neg_sampler2-full-idtype1] PASSED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[True-cpu-neg_sampler2-neighbor-idtype0] PASSED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[True-cpu-neg_sampler2-neighbor-idtype1] PASSED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[True-uva-neg_sampler0-full-idtype0] SKIPPED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[True-uva-neg_sampler0-full-idtype1] SKIPPED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[True-uva-neg_sampler0-neighbor-idtype0] SKIPPED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[True-uva-neg_sampler0-neighbor-idtype1] SKIPPED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[True-uva-neg_sampler1-full-idtype0] SKIPPED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[True-uva-neg_sampler1-full-idtype1] SKIPPED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[True-uva-neg_sampler1-neighbor-idtype0] SKIPPED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[True-uva-neg_sampler1-neighbor-idtype1] SKIPPED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[True-uva-neg_sampler2-full-idtype0] SKIPPED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[True-uva-neg_sampler2-full-idtype1] SKIPPED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[True-uva-neg_sampler2-neighbor-idtype0] SKIPPED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[True-uva-neg_sampler2-neighbor-idtype1] SKIPPED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[True-pure_gpu-neg_sampler0-full-idtype0] SKIPPED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[True-pure_gpu-neg_sampler0-full-idtype1] SKIPPED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[True-pure_gpu-neg_sampler0-neighbor-idtype0] SKIPPED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[True-pure_gpu-neg_sampler0-neighbor-idtype1] SKIPPED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[True-pure_gpu-neg_sampler1-full-idtype0] SKIPPED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[True-pure_gpu-neg_sampler1-full-idtype1] SKIPPED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[True-pure_gpu-neg_sampler1-neighbor-idtype0] SKIPPED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[True-pure_gpu-neg_sampler1-neighbor-idtype1] SKIPPED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[True-pure_gpu-neg_sampler2-full-idtype0] SKIPPED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[True-pure_gpu-neg_sampler2-full-idtype1] SKIPPED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[True-pure_gpu-neg_sampler2-neighbor-idtype0] SKIPPED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[True-pure_gpu-neg_sampler2-neighbor-idtype1] SKIPPED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader_excludes[1-sampler0-None-False] PASSED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader_excludes[1-sampler0-None-True] PASSED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader_excludes[1-sampler0-self-False] PASSED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader_excludes[1-sampler0-self-True] PASSED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader_excludes[1-sampler0-reverse_id-False] PASSED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader_excludes[1-sampler0-reverse_id-True] PASSED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader_excludes[1-sampler0-reverse_types-False] PASSED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader_excludes[1-sampler0-reverse_types-True] PASSED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader_excludes[1-sampler1-None-False] PASSED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader_excludes[1-sampler1-None-True] PASSED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader_excludes[1-sampler1-self-False] PASSED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader_excludes[1-sampler1-self-True] PASSED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader_excludes[1-sampler1-reverse_id-False] PASSED [  4%]
tests/pytorch/test_dataloader.py::test_edge_dataloader_excludes[1-sampler1-reverse_id-True] PASSED [  4%]
tests/pytorch/test_dataloader.py::test_edge_dataloader_excludes[1-sampler1-reverse_types-False] PASSED [  4%]
tests/pytorch/test_dataloader.py::test_edge_dataloader_excludes[1-sampler1-reverse_types-True] PASSED [  4%]
tests/pytorch/test_dataloader.py::test_edge_dataloader_excludes[50-sampler0-None-False] PASSED [  4%]
tests/pytorch/test_dataloader.py::test_edge_dataloader_excludes[50-sampler0-None-True] PASSED [  4%]
tests/pytorch/test_dataloader.py::test_edge_dataloader_excludes[50-sampler0-self-False] PASSED [  4%]
tests/pytorch/test_dataloader.py::test_edge_dataloader_excludes[50-sampler0-self-True] PASSED [  4%]
tests/pytorch/test_dataloader.py::test_edge_dataloader_excludes[50-sampler0-reverse_id-False] PASSED [  4%]
tests/pytorch/test_dataloader.py::test_edge_dataloader_excludes[50-sampler0-reverse_id-True] PASSED [  4%]
tests/pytorch/test_dataloader.py::test_edge_dataloader_excludes[50-sampler0-reverse_types-False] PASSED [  4%]
tests/pytorch/test_dataloader.py::test_edge_dataloader_excludes[50-sampler0-reverse_types-True] PASSED [  4%]
tests/pytorch/test_dataloader.py::test_edge_dataloader_excludes[50-sampler1-None-False] PASSED [  4%]
tests/pytorch/test_dataloader.py::test_edge_dataloader_excludes[50-sampler1-None-True] PASSED [  4%]
tests/pytorch/test_dataloader.py::test_edge_dataloader_excludes[50-sampler1-self-False] PASSED [  4%]
tests/pytorch/test_dataloader.py::test_edge_dataloader_excludes[50-sampler1-self-True] PASSED [  4%]
tests/pytorch/test_dataloader.py::test_edge_dataloader_excludes[50-sampler1-reverse_id-False] PASSED [  4%]
tests/pytorch/test_dataloader.py::test_edge_dataloader_excludes[50-sampler1-reverse_id-True] PASSED [  4%]
tests/pytorch/test_dataloader.py::test_edge_dataloader_excludes[50-sampler1-reverse_types-False] PASSED [  4%]
tests/pytorch/test_dataloader.py::test_edge_dataloader_excludes[50-sampler1-reverse_types-True] PASSED [  4%]
tests/pytorch/test_dist_optim.py::test_sparse_opt SKIPPED (Do not su...) [  4%]
tests/pytorch/test_geometry.py::test_fps PASSED                          [  4%]
tests/pytorch/test_geometry.py::test_fps_start_idx PASSED                [  4%]
tests/pytorch/test_geometry.py::test_knn_cpu[False-euclidean-bruteforce-blas] PASSED [  4%]
tests/pytorch/test_geometry.py::test_knn_cpu[False-euclidean-bruteforce] PASSED [  4%]
tests/pytorch/test_geometry.py::test_knn_cpu[False-euclidean-kd-tree] PASSED [  4%]
tests/pytorch/test_geometry.py::test_knn_cpu[False-cosine-bruteforce-blas] PASSED [  4%]
tests/pytorch/test_geometry.py::test_knn_cpu[False-cosine-bruteforce] PASSED [  4%]
tests/pytorch/test_geometry.py::test_knn_cpu[False-cosine-kd-tree] PASSED [  4%]
tests/pytorch/test_geometry.py::test_knn_cpu[True-euclidean-bruteforce-blas] FAILED [  4%]
tests/pytorch/test_geometry.py::test_knn_cpu[True-euclidean-bruteforce] FAILED [  4%]
tests/pytorch/test_geometry.py::test_knn_cpu[True-euclidean-kd-tree] FAILED [  4%]
tests/pytorch/test_geometry.py::test_knn_cpu[True-cosine-bruteforce-blas] FAILED [  4%]
tests/pytorch/test_geometry.py::test_knn_cpu[True-cosine-bruteforce] FAILED [  4%]
tests/pytorch/test_geometry.py::test_knn_cpu[True-cosine-kd-tree] FAILED [  4%]
tests/pytorch/test_geometry.py::test_knn_cuda[False-euclidean-bruteforce-blas] PASSED [  4%]
tests/pytorch/test_geometry.py::test_knn_cuda[False-euclidean-bruteforce] PASSED [  4%]
tests/pytorch/test_geometry.py::test_knn_cuda[False-euclidean-bruteforce-sharemem] PASSED [  4%]
tests/pytorch/test_geometry.py::test_knn_cuda[False-cosine-bruteforce-blas] PASSED [  4%]
tests/pytorch/test_geometry.py::test_knn_cuda[False-cosine-bruteforce] PASSED [  5%]
tests/pytorch/test_geometry.py::test_knn_cuda[False-cosine-bruteforce-sharemem] PASSED [  5%]
tests/pytorch/test_geometry.py::test_knn_cuda[True-euclidean-bruteforce-blas] PASSED [  5%]
tests/pytorch/test_geometry.py::test_knn_cuda[True-euclidean-bruteforce] PASSED [  5%]
tests/pytorch/test_geometry.py::test_knn_cuda[True-euclidean-bruteforce-sharemem] PASSED [  5%]
tests/pytorch/test_geometry.py::test_knn_cuda[True-cosine-bruteforce-blas] PASSED [  5%]
tests/pytorch/test_geometry.py::test_knn_cuda[True-cosine-bruteforce] PASSED [  5%]
tests/pytorch/test_geometry.py::test_knn_cuda[True-cosine-bruteforce-sharemem] PASSED [  5%]
tests/pytorch/test_geometry.py::test_edge_coarsening[True-True-g0-idtype0] PASSED [  5%]
tests/pytorch/test_geometry.py::test_edge_coarsening[True-True-g0-idtype1] PASSED [  5%]
tests/pytorch/test_geometry.py::test_edge_coarsening[True-True-g1-idtype0] PASSED [  5%]
tests/pytorch/test_geometry.py::test_edge_coarsening[True-True-g1-idtype1] PASSED [  5%]
tests/pytorch/test_geometry.py::test_edge_coarsening[True-True-g2-idtype0] PASSED [  5%]
tests/pytorch/test_geometry.py::test_edge_coarsening[True-True-g2-idtype1] PASSED [  5%]
tests/pytorch/test_geometry.py::test_edge_coarsening[True-True-g3-idtype0] PASSED [  5%]
tests/pytorch/test_geometry.py::test_edge_coarsening[True-True-g3-idtype1] PASSED [  5%]
tests/pytorch/test_geometry.py::test_edge_coarsening[True-True-g4-idtype0] PASSED [  5%]
tests/pytorch/test_geometry.py::test_edge_coarsening[True-True-g4-idtype1] PASSED [  5%]
tests/pytorch/test_geometry.py::test_edge_coarsening[True-True-g5-idtype0] PASSED [  5%]
tests/pytorch/test_geometry.py::test_edge_coarsening[True-True-g5-idtype1] PASSED [  5%]
tests/pytorch/test_geometry.py::test_edge_coarsening[True-True-g6-idtype0] PASSED [  5%]
tests/pytorch/test_geometry.py::test_edge_coarsening[True-True-g6-idtype1] PASSED [  5%]
tests/pytorch/test_geometry.py::test_edge_coarsening[True-False-g0-idtype0] PASSED [  5%]
tests/pytorch/test_geometry.py::test_edge_coarsening[True-False-g0-idtype1] PASSED [  5%]
tests/pytorch/test_geometry.py::test_edge_coarsening[True-False-g1-idtype0] PASSED [  5%]
tests/pytorch/test_geometry.py::test_edge_coarsening[True-False-g1-idtype1] PASSED [  5%]
tests/pytorch/test_geometry.py::test_edge_coarsening[True-False-g2-idtype0] PASSED [  5%]
tests/pytorch/test_geometry.py::test_edge_coarsening[True-False-g2-idtype1] PASSED [  5%]
tests/pytorch/test_geometry.py::test_edge_coarsening[True-False-g3-idtype0] PASSED [  5%]
tests/pytorch/test_geometry.py::test_edge_coarsening[True-False-g3-idtype1] PASSED [  5%]
tests/pytorch/test_geometry.py::test_edge_coarsening[True-False-g4-idtype0] PASSED [  5%]
tests/pytorch/test_geometry.py::test_edge_coarsening[True-False-g4-idtype1] PASSED [  5%]
tests/pytorch/test_geometry.py::test_edge_coarsening[True-False-g5-idtype0] PASSED [  5%]
tests/pytorch/test_geometry.py::test_edge_coarsening[True-False-g5-idtype1] PASSED [  5%]
tests/pytorch/test_geometry.py::test_edge_coarsening[True-False-g6-idtype0] PASSED [  5%]
tests/pytorch/test_geometry.py::test_edge_coarsening[True-False-g6-idtype1] PASSED [  5%]
tests/pytorch/test_geometry.py::test_edge_coarsening[False-True-g0-idtype0] PASSED [  5%]
tests/pytorch/test_geometry.py::test_edge_coarsening[False-True-g0-idtype1] PASSED [  5%]
tests/pytorch/test_geometry.py::test_edge_coarsening[False-True-g1-idtype0] PASSED [  5%]
tests/pytorch/test_geometry.py::test_edge_coarsening[False-True-g1-idtype1] PASSED [  5%]
tests/pytorch/test_geometry.py::test_edge_coarsening[False-True-g2-idtype0] PASSED [  6%]
tests/pytorch/test_geometry.py::test_edge_coarsening[False-True-g2-idtype1] PASSED [  6%]
tests/pytorch/test_geometry.py::test_edge_coarsening[False-True-g3-idtype0] PASSED [  6%]
tests/pytorch/test_geometry.py::test_edge_coarsening[False-True-g3-idtype1] PASSED [  6%]
tests/pytorch/test_geometry.py::test_edge_coarsening[False-True-g4-idtype0] PASSED [  6%]
tests/pytorch/test_geometry.py::test_edge_coarsening[False-True-g4-idtype1] PASSED [  6%]
tests/pytorch/test_geometry.py::test_edge_coarsening[False-True-g5-idtype0] PASSED [  6%]
tests/pytorch/test_geometry.py::test_edge_coarsening[False-True-g5-idtype1] PASSED [  6%]
tests/pytorch/test_geometry.py::test_edge_coarsening[False-True-g6-idtype0] PASSED [  6%]
tests/pytorch/test_geometry.py::test_edge_coarsening[False-True-g6-idtype1] PASSED [  6%]
tests/pytorch/test_geometry.py::test_edge_coarsening[False-False-g0-idtype0] PASSED [  6%]
tests/pytorch/test_geometry.py::test_edge_coarsening[False-False-g0-idtype1] PASSED [  6%]
tests/pytorch/test_geometry.py::test_edge_coarsening[False-False-g1-idtype0] PASSED [  6%]
tests/pytorch/test_geometry.py::test_edge_coarsening[False-False-g1-idtype1] PASSED [  6%]
tests/pytorch/test_geometry.py::test_edge_coarsening[False-False-g2-idtype0] PASSED [  6%]
tests/pytorch/test_geometry.py::test_edge_coarsening[False-False-g2-idtype1] PASSED [  6%]
tests/pytorch/test_geometry.py::test_edge_coarsening[False-False-g3-idtype0] PASSED [  6%]
tests/pytorch/test_geometry.py::test_edge_coarsening[False-False-g3-idtype1] PASSED [  6%]
tests/pytorch/test_geometry.py::test_edge_coarsening[False-False-g4-idtype0] PASSED [  6%]
tests/pytorch/test_geometry.py::test_edge_coarsening[False-False-g4-idtype1] PASSED [  6%]
tests/pytorch/test_geometry.py::test_edge_coarsening[False-False-g5-idtype0] PASSED [  6%]
tests/pytorch/test_geometry.py::test_edge_coarsening[False-False-g5-idtype1] PASSED [  6%]
tests/pytorch/test_geometry.py::test_edge_coarsening[False-False-g6-idtype0] PASSED [  6%]
tests/pytorch/test_geometry.py::test_edge_coarsening[False-False-g6-idtype1] PASSED [  6%]
tests/pytorch/test_ipc.py::test_torch_ipc SKIPPED (Do not support wi...) [  6%]
tests/pytorch/test_nn.py::test_graph_conv0[1] FAILED                     [  6%]
tests/pytorch/test_nn.py::test_graph_conv0[2] FAILED                     [  6%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-none-g0-idtype0] FAILED [  6%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-none-g0-idtype1] FAILED [  6%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-none-g1-idtype0] FAILED [  6%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-none-g1-idtype1] FAILED [  6%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-none-g2-idtype0] FAILED [  6%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-none-g2-idtype1] FAILED [  6%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-none-g3-idtype0] FAILED [  6%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-none-g3-idtype1] FAILED [  6%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-none-g4-idtype0] FAILED [  6%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-none-g4-idtype1] FAILED [  6%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-none-g5-idtype0] FAILED [  6%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-none-g5-idtype1] FAILED [  6%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-none-g6-idtype0] FAILED [  6%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-none-g6-idtype1] FAILED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-none-g7-idtype0] FAILED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-none-g7-idtype1] FAILED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-both-g0-idtype0] FAILED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-both-g0-idtype1] FAILED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-both-g1-idtype0] FAILED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-both-g1-idtype1] FAILED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-both-g2-idtype0] FAILED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-both-g2-idtype1] FAILED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-both-g3-idtype0] FAILED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-both-g3-idtype1] FAILED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-both-g4-idtype0] FAILED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-both-g4-idtype1] FAILED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-both-g5-idtype0] FAILED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-both-g5-idtype1] FAILED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-both-g6-idtype0] FAILED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-both-g6-idtype1] FAILED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-both-g7-idtype0] FAILED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-both-g7-idtype1] FAILED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-right-g0-idtype0] FAILED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-right-g0-idtype1] FAILED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-right-g1-idtype0] FAILED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-right-g1-idtype1] FAILED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-right-g2-idtype0] FAILED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-right-g2-idtype1] FAILED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-right-g3-idtype0] FAILED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-right-g3-idtype1] FAILED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-right-g4-idtype0] FAILED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-right-g4-idtype1] FAILED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-right-g5-idtype0] FAILED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-right-g5-idtype1] FAILED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-right-g6-idtype0] FAILED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-right-g6-idtype1] FAILED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-right-g7-idtype0] FAILED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-right-g7-idtype1] FAILED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-left-g0-idtype0] FAILED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-left-g0-idtype1] FAILED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-left-g1-idtype0] FAILED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-left-g1-idtype1] FAILED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-left-g2-idtype0] FAILED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-left-g2-idtype1] FAILED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-left-g3-idtype0] FAILED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-left-g3-idtype1] FAILED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-left-g4-idtype0] FAILED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-left-g4-idtype1] FAILED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-left-g5-idtype0] FAILED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-left-g5-idtype1] FAILED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-left-g6-idtype0] FAILED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-left-g6-idtype1] FAILED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-left-g7-idtype0] FAILED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-left-g7-idtype1] FAILED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-none-g0-idtype0] FAILED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-none-g0-idtype1] FAILED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-none-g1-idtype0] FAILED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-none-g1-idtype1] FAILED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-none-g2-idtype0] FAILED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-none-g2-idtype1] FAILED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-none-g3-idtype0] FAILED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-none-g3-idtype1] FAILED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-none-g4-idtype0] FAILED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-none-g4-idtype1] FAILED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-none-g5-idtype0] FAILED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-none-g5-idtype1] FAILED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-none-g6-idtype0] FAILED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-none-g6-idtype1] FAILED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-none-g7-idtype0] FAILED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-none-g7-idtype1] FAILED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-both-g0-idtype0] FAILED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-both-g0-idtype1] FAILED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-both-g1-idtype0] FAILED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-both-g1-idtype1] FAILED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-both-g2-idtype0] FAILED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-both-g2-idtype1] FAILED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-both-g3-idtype0] FAILED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-both-g3-idtype1] FAILED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-both-g4-idtype0] FAILED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-both-g4-idtype1] FAILED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-both-g5-idtype0] FAILED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-both-g5-idtype1] FAILED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-both-g6-idtype0] FAILED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-both-g6-idtype1] FAILED [  9%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-both-g7-idtype0] FAILED [  9%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-both-g7-idtype1] FAILED [  9%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-right-g0-idtype0] FAILED [  9%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-right-g0-idtype1] FAILED [  9%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-right-g1-idtype0] FAILED [  9%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-right-g1-idtype1] FAILED [  9%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-right-g2-idtype0] FAILED [  9%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-right-g2-idtype1] FAILED [  9%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-right-g3-idtype0] FAILED [  9%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-right-g3-idtype1] FAILED [  9%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-right-g4-idtype0] FAILED [  9%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-right-g4-idtype1] FAILED [  9%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-right-g5-idtype0] FAILED [  9%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-right-g5-idtype1] FAILED [  9%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-right-g6-idtype0] FAILED [  9%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-right-g6-idtype1] FAILED [  9%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-right-g7-idtype0] FAILED [  9%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-right-g7-idtype1] FAILED [  9%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-left-g0-idtype0] FAILED [  9%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-left-g0-idtype1] FAILED [  9%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-left-g1-idtype0] FAILED [  9%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-left-g1-idtype1] FAILED [  9%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-left-g2-idtype0] FAILED [  9%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-left-g2-idtype1] FAILED [  9%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-left-g3-idtype0] FAILED [  9%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-left-g3-idtype1] FAILED [  9%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-left-g4-idtype0] FAILED [  9%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-left-g4-idtype1] FAILED [  9%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-left-g5-idtype0] FAILED [  9%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-left-g5-idtype1] FAILED [  9%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-left-g6-idtype0] FAILED [  9%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-left-g6-idtype1] FAILED [  9%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-left-g7-idtype0] FAILED [  9%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-left-g7-idtype1] FAILED [  9%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-none-g0-idtype0] FAILED [  9%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-none-g0-idtype1] FAILED [  9%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-none-g1-idtype0] FAILED [  9%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-none-g1-idtype1] FAILED [  9%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-none-g2-idtype0] FAILED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-none-g2-idtype1] FAILED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-none-g3-idtype0] FAILED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-none-g3-idtype1] FAILED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-none-g4-idtype0] FAILED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-none-g4-idtype1] FAILED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-none-g5-idtype0] FAILED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-none-g5-idtype1] FAILED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-none-g6-idtype0] FAILED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-none-g6-idtype1] FAILED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-none-g7-idtype0] FAILED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-none-g7-idtype1] FAILED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-both-g0-idtype0] FAILED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-both-g0-idtype1] FAILED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-both-g1-idtype0] FAILED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-both-g1-idtype1] FAILED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-both-g2-idtype0] FAILED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-both-g2-idtype1] FAILED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-both-g3-idtype0] FAILED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-both-g3-idtype1] FAILED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-both-g4-idtype0] FAILED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-both-g4-idtype1] FAILED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-both-g5-idtype0] FAILED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-both-g5-idtype1] FAILED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-both-g6-idtype0] FAILED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-both-g6-idtype1] FAILED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-both-g7-idtype0] FAILED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-both-g7-idtype1] FAILED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-right-g0-idtype0] FAILED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-right-g0-idtype1] FAILED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-right-g1-idtype0] FAILED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-right-g1-idtype1] FAILED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-right-g2-idtype0] FAILED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-right-g2-idtype1] FAILED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-right-g3-idtype0] FAILED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-right-g3-idtype1] FAILED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-right-g4-idtype0] FAILED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-right-g4-idtype1] FAILED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-right-g5-idtype0] FAILED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-right-g5-idtype1] FAILED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-right-g6-idtype0] FAILED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-right-g6-idtype1] FAILED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-right-g7-idtype0] FAILED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-right-g7-idtype1] FAILED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-left-g0-idtype0] FAILED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-left-g0-idtype1] FAILED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-left-g1-idtype0] FAILED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-left-g1-idtype1] FAILED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-left-g2-idtype0] FAILED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-left-g2-idtype1] FAILED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-left-g3-idtype0] FAILED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-left-g3-idtype1] FAILED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-left-g4-idtype0] FAILED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-left-g4-idtype1] FAILED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-left-g5-idtype0] FAILED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-left-g5-idtype1] FAILED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-left-g6-idtype0] FAILED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-left-g6-idtype1] FAILED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-left-g7-idtype0] FAILED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-left-g7-idtype1] FAILED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-none-g0-idtype0] FAILED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-none-g0-idtype1] FAILED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-none-g1-idtype0] FAILED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-none-g1-idtype1] FAILED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-none-g2-idtype0] FAILED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-none-g2-idtype1] FAILED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-none-g3-idtype0] FAILED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-none-g3-idtype1] FAILED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-none-g4-idtype0] FAILED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-none-g4-idtype1] FAILED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-none-g5-idtype0] FAILED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-none-g5-idtype1] FAILED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-none-g6-idtype0] FAILED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-none-g6-idtype1] FAILED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-none-g7-idtype0] FAILED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-none-g7-idtype1] FAILED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-both-g0-idtype0] FAILED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-both-g0-idtype1] FAILED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-both-g1-idtype0] FAILED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-both-g1-idtype1] FAILED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-both-g2-idtype0] FAILED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-both-g2-idtype1] FAILED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-both-g3-idtype0] FAILED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-both-g3-idtype1] FAILED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-both-g4-idtype0] FAILED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-both-g4-idtype1] FAILED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-both-g5-idtype0] FAILED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-both-g5-idtype1] FAILED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-both-g6-idtype0] FAILED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-both-g6-idtype1] FAILED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-both-g7-idtype0] FAILED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-both-g7-idtype1] FAILED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-right-g0-idtype0] FAILED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-right-g0-idtype1] FAILED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-right-g1-idtype0] FAILED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-right-g1-idtype1] FAILED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-right-g2-idtype0] FAILED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-right-g2-idtype1] FAILED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-right-g3-idtype0] FAILED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-right-g3-idtype1] FAILED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-right-g4-idtype0] FAILED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-right-g4-idtype1] FAILED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-right-g5-idtype0] FAILED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-right-g5-idtype1] FAILED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-right-g6-idtype0] FAILED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-right-g6-idtype1] FAILED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-right-g7-idtype0] FAILED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-right-g7-idtype1] FAILED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-left-g0-idtype0] FAILED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-left-g0-idtype1] FAILED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-left-g1-idtype0] FAILED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-left-g1-idtype1] FAILED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-left-g2-idtype0] FAILED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-left-g2-idtype1] FAILED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-left-g3-idtype0] FAILED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-left-g3-idtype1] FAILED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-left-g4-idtype0] FAILED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-left-g4-idtype1] FAILED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-left-g5-idtype0] FAILED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-left-g5-idtype1] FAILED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-left-g6-idtype0] FAILED [ 13%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-left-g6-idtype1] FAILED [ 13%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-left-g7-idtype0] FAILED [ 13%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-left-g7-idtype1] FAILED [ 13%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-none-g0-idtype0] FAILED [ 13%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-none-g0-idtype1] FAILED [ 13%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-none-g1-idtype0] FAILED [ 13%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-none-g1-idtype1] FAILED [ 13%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-none-g2-idtype0] FAILED [ 13%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-none-g2-idtype1] FAILED [ 13%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-none-g3-idtype0] FAILED [ 13%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-none-g3-idtype1] FAILED [ 13%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-none-g4-idtype0] FAILED [ 13%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-none-g4-idtype1] FAILED [ 13%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-none-g5-idtype0] FAILED [ 13%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-none-g5-idtype1] FAILED [ 13%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-none-g6-idtype0] FAILED [ 13%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-none-g6-idtype1] FAILED [ 13%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-none-g7-idtype0] FAILED [ 13%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-none-g7-idtype1] FAILED [ 13%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-both-g0-idtype0] FAILED [ 13%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-both-g0-idtype1] FAILED [ 13%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-both-g1-idtype0] FAILED [ 13%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-both-g1-idtype1] FAILED [ 13%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-both-g2-idtype0] FAILED [ 13%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-both-g2-idtype1] FAILED [ 13%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-both-g3-idtype0] FAILED [ 13%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-both-g3-idtype1] FAILED [ 13%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-both-g4-idtype0] FAILED [ 13%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-both-g4-idtype1] FAILED [ 13%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-both-g5-idtype0] FAILED [ 13%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-both-g5-idtype1] FAILED [ 13%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-both-g6-idtype0] FAILED [ 13%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-both-g6-idtype1] FAILED [ 13%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-both-g7-idtype0] FAILED [ 13%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-both-g7-idtype1] FAILED [ 13%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-right-g0-idtype0] FAILED [ 13%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-right-g0-idtype1] FAILED [ 13%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-right-g1-idtype0] FAILED [ 13%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-right-g1-idtype1] FAILED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-right-g2-idtype0] FAILED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-right-g2-idtype1] FAILED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-right-g3-idtype0] FAILED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-right-g3-idtype1] FAILED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-right-g4-idtype0] FAILED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-right-g4-idtype1] FAILED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-right-g5-idtype0] FAILED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-right-g5-idtype1] FAILED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-right-g6-idtype0] FAILED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-right-g6-idtype1] FAILED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-right-g7-idtype0] FAILED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-right-g7-idtype1] FAILED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-left-g0-idtype0] FAILED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-left-g0-idtype1] FAILED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-left-g1-idtype0] FAILED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-left-g1-idtype1] FAILED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-left-g2-idtype0] FAILED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-left-g2-idtype1] FAILED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-left-g3-idtype0] FAILED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-left-g3-idtype1] FAILED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-left-g4-idtype0] FAILED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-left-g4-idtype1] FAILED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-left-g5-idtype0] FAILED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-left-g5-idtype1] FAILED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-left-g6-idtype0] FAILED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-left-g6-idtype1] FAILED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-left-g7-idtype0] FAILED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-left-g7-idtype1] FAILED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-none-g0-idtype0] FAILED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-none-g0-idtype1] FAILED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-none-g1-idtype0] FAILED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-none-g1-idtype1] FAILED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-none-g2-idtype0] FAILED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-none-g2-idtype1] FAILED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-none-g3-idtype0] FAILED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-none-g3-idtype1] FAILED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-none-g4-idtype0] FAILED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-none-g4-idtype1] FAILED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-none-g5-idtype0] FAILED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-none-g5-idtype1] FAILED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-none-g6-idtype0] FAILED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-none-g6-idtype1] FAILED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-none-g7-idtype0] FAILED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-none-g7-idtype1] FAILED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-both-g0-idtype0] FAILED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-both-g0-idtype1] FAILED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-both-g1-idtype0] FAILED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-both-g1-idtype1] FAILED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-both-g2-idtype0] FAILED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-both-g2-idtype1] FAILED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-both-g3-idtype0] FAILED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-both-g3-idtype1] FAILED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-both-g4-idtype0] FAILED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-both-g4-idtype1] FAILED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-both-g5-idtype0] FAILED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-both-g5-idtype1] FAILED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-both-g6-idtype0] FAILED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-both-g6-idtype1] FAILED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-both-g7-idtype0] FAILED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-both-g7-idtype1] FAILED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-right-g0-idtype0] FAILED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-right-g0-idtype1] FAILED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-right-g1-idtype0] FAILED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-right-g1-idtype1] FAILED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-right-g2-idtype0] FAILED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-right-g2-idtype1] FAILED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-right-g3-idtype0] FAILED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-right-g3-idtype1] FAILED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-right-g4-idtype0] FAILED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-right-g4-idtype1] FAILED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-right-g5-idtype0] FAILED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-right-g5-idtype1] FAILED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-right-g6-idtype0] FAILED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-right-g6-idtype1] FAILED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-right-g7-idtype0] FAILED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-right-g7-idtype1] FAILED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-left-g0-idtype0] FAILED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-left-g0-idtype1] FAILED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-left-g1-idtype0] FAILED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-left-g1-idtype1] FAILED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-left-g2-idtype0] FAILED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-left-g2-idtype1] FAILED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-left-g3-idtype0] FAILED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-left-g3-idtype1] FAILED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-left-g4-idtype0] FAILED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-left-g4-idtype1] FAILED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-left-g5-idtype0] FAILED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-left-g5-idtype1] FAILED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-left-g6-idtype0] FAILED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-left-g6-idtype1] FAILED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-left-g7-idtype0] FAILED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-left-g7-idtype1] FAILED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-none-g0-idtype0] FAILED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-none-g0-idtype1] FAILED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-none-g1-idtype0] FAILED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-none-g1-idtype1] FAILED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-none-g2-idtype0] FAILED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-none-g2-idtype1] FAILED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-none-g3-idtype0] FAILED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-none-g3-idtype1] FAILED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-none-g4-idtype0] FAILED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-none-g4-idtype1] FAILED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-none-g5-idtype0] FAILED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-none-g5-idtype1] FAILED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-none-g6-idtype0] FAILED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-none-g6-idtype1] FAILED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-none-g7-idtype0] FAILED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-none-g7-idtype1] FAILED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-both-g0-idtype0] FAILED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-both-g0-idtype1] FAILED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-both-g1-idtype0] FAILED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-both-g1-idtype1] FAILED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-both-g2-idtype0] FAILED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-both-g2-idtype1] FAILED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-both-g3-idtype0] FAILED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-both-g3-idtype1] FAILED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-both-g4-idtype0] FAILED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-both-g4-idtype1] FAILED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-both-g5-idtype0] FAILED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-both-g5-idtype1] FAILED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-both-g6-idtype0] FAILED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-both-g6-idtype1] FAILED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-both-g7-idtype0] FAILED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-both-g7-idtype1] FAILED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-right-g0-idtype0] FAILED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-right-g0-idtype1] FAILED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-right-g1-idtype0] FAILED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-right-g1-idtype1] FAILED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-right-g2-idtype0] FAILED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-right-g2-idtype1] FAILED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-right-g3-idtype0] FAILED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-right-g3-idtype1] FAILED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-right-g4-idtype0] FAILED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-right-g4-idtype1] FAILED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-right-g5-idtype0] FAILED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-right-g5-idtype1] FAILED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-right-g6-idtype0] FAILED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-right-g6-idtype1] FAILED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-right-g7-idtype0] FAILED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-right-g7-idtype1] FAILED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-left-g0-idtype0] FAILED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-left-g0-idtype1] FAILED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-left-g1-idtype0] FAILED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-left-g1-idtype1] FAILED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-left-g2-idtype0] FAILED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-left-g2-idtype1] FAILED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-left-g3-idtype0] FAILED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-left-g3-idtype1] FAILED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-left-g4-idtype0] FAILED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-left-g4-idtype1] FAILED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-left-g5-idtype0] FAILED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-left-g5-idtype1] FAILED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-left-g6-idtype0] FAILED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-left-g6-idtype1] FAILED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-left-g7-idtype0] FAILED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-left-g7-idtype1] FAILED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-none-g0-idtype0] FAILED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-none-g0-idtype1] FAILED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-none-g1-idtype0] FAILED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-none-g1-idtype1] FAILED [ 18%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-none-g2-idtype0] FAILED [ 18%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-none-g2-idtype1] FAILED [ 18%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-none-g3-idtype0] FAILED [ 18%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-none-g3-idtype1] FAILED [ 18%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-none-g4-idtype0] FAILED [ 18%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-none-g4-idtype1] FAILED [ 18%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-none-g5-idtype0] FAILED [ 18%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-none-g5-idtype1] FAILED [ 18%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-none-g6-idtype0] FAILED [ 18%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-none-g6-idtype1] FAILED [ 18%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-none-g7-idtype0] FAILED [ 18%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-none-g7-idtype1] FAILED [ 18%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-both-g0-idtype0] FAILED [ 18%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-both-g0-idtype1] FAILED [ 18%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-both-g1-idtype0] FAILED [ 18%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-both-g1-idtype1] FAILED [ 18%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-both-g2-idtype0] FAILED [ 18%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-both-g2-idtype1] FAILED [ 18%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-both-g3-idtype0] FAILED [ 18%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-both-g3-idtype1] FAILED [ 18%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-both-g4-idtype0] FAILED [ 18%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-both-g4-idtype1] FAILED [ 18%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-both-g5-idtype0] FAILED [ 18%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-both-g5-idtype1] FAILED [ 18%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-both-g6-idtype0] FAILED [ 18%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-both-g6-idtype1] FAILED [ 18%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-both-g7-idtype0] FAILED [ 18%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-both-g7-idtype1] FAILED [ 18%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-right-g0-idtype0] FAILED [ 18%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-right-g0-idtype1] FAILED [ 18%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-right-g1-idtype0] FAILED [ 18%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-right-g1-idtype1] FAILED [ 18%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-right-g2-idtype0] FAILED [ 18%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-right-g2-idtype1] FAILED [ 18%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-right-g3-idtype0] FAILED [ 18%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-right-g3-idtype1] FAILED [ 18%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-right-g4-idtype0] FAILED [ 18%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-right-g4-idtype1] FAILED [ 18%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-right-g5-idtype0] FAILED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-right-g5-idtype1] FAILED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-right-g6-idtype0] FAILED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-right-g6-idtype1] FAILED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-right-g7-idtype0] FAILED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-right-g7-idtype1] FAILED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-left-g0-idtype0] FAILED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-left-g0-idtype1] FAILED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-left-g1-idtype0] FAILED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-left-g1-idtype1] FAILED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-left-g2-idtype0] FAILED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-left-g2-idtype1] FAILED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-left-g3-idtype0] FAILED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-left-g3-idtype1] FAILED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-left-g4-idtype0] FAILED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-left-g4-idtype1] FAILED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-left-g5-idtype0] FAILED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-left-g5-idtype1] FAILED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-left-g6-idtype0] FAILED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-left-g6-idtype1] FAILED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-left-g7-idtype0] FAILED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-left-g7-idtype1] FAILED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[1-True-True-none-g0-idtype0] FAILED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[1-True-True-none-g0-idtype1] FAILED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[1-True-True-both-g0-idtype0] FAILED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[1-True-True-both-g0-idtype1] FAILED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[1-True-True-right-g0-idtype0] FAILED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[1-True-True-right-g0-idtype1] FAILED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[1-True-False-none-g0-idtype0] FAILED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[1-True-False-none-g0-idtype1] FAILED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[1-True-False-both-g0-idtype0] FAILED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[1-True-False-both-g0-idtype1] FAILED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[1-True-False-right-g0-idtype0] FAILED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[1-True-False-right-g0-idtype1] FAILED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[1-False-True-none-g0-idtype0] FAILED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[1-False-True-none-g0-idtype1] FAILED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[1-False-True-both-g0-idtype0] FAILED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[1-False-True-both-g0-idtype1] FAILED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[1-False-True-right-g0-idtype0] FAILED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[1-False-True-right-g0-idtype1] FAILED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[1-False-False-none-g0-idtype0] FAILED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[1-False-False-none-g0-idtype1] FAILED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[1-False-False-both-g0-idtype0] FAILED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[1-False-False-both-g0-idtype1] FAILED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[1-False-False-right-g0-idtype0] FAILED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[1-False-False-right-g0-idtype1] FAILED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[2-True-True-none-g0-idtype0] FAILED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[2-True-True-none-g0-idtype1] FAILED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[2-True-True-both-g0-idtype0] FAILED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[2-True-True-both-g0-idtype1] FAILED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[2-True-True-right-g0-idtype0] FAILED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[2-True-True-right-g0-idtype1] FAILED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[2-True-False-none-g0-idtype0] FAILED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[2-True-False-none-g0-idtype1] FAILED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[2-True-False-both-g0-idtype0] FAILED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[2-True-False-both-g0-idtype1] FAILED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[2-True-False-right-g0-idtype0] FAILED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[2-True-False-right-g0-idtype1] FAILED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[2-False-True-none-g0-idtype0] FAILED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[2-False-True-none-g0-idtype1] FAILED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[2-False-True-both-g0-idtype0] FAILED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[2-False-True-both-g0-idtype1] FAILED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[2-False-True-right-g0-idtype0] FAILED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[2-False-True-right-g0-idtype1] FAILED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[2-False-False-none-g0-idtype0] FAILED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[2-False-False-none-g0-idtype1] FAILED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[2-False-False-both-g0-idtype0] FAILED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[2-False-False-both-g0-idtype1] FAILED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[2-False-False-right-g0-idtype0] FAILED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[2-False-False-right-g0-idtype1] FAILED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[1-True-True-none-g0-idtype0] FAILED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[1-True-True-none-g0-idtype1] FAILED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[1-True-True-both-g0-idtype0] FAILED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[1-True-True-both-g0-idtype1] FAILED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[1-True-True-right-g0-idtype0] FAILED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[1-True-True-right-g0-idtype1] FAILED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[1-True-False-none-g0-idtype0] FAILED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[1-True-False-none-g0-idtype1] FAILED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[1-True-False-both-g0-idtype0] FAILED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[1-True-False-both-g0-idtype1] FAILED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[1-True-False-right-g0-idtype0] FAILED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[1-True-False-right-g0-idtype1] FAILED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[1-False-True-none-g0-idtype0] FAILED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[1-False-True-none-g0-idtype1] FAILED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[1-False-True-both-g0-idtype0] FAILED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[1-False-True-both-g0-idtype1] FAILED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[1-False-True-right-g0-idtype0] FAILED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[1-False-True-right-g0-idtype1] FAILED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[1-False-False-none-g0-idtype0] FAILED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[1-False-False-none-g0-idtype1] FAILED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[1-False-False-both-g0-idtype0] FAILED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[1-False-False-both-g0-idtype1] FAILED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[1-False-False-right-g0-idtype0] FAILED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[1-False-False-right-g0-idtype1] FAILED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[2-True-True-none-g0-idtype0] FAILED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[2-True-True-none-g0-idtype1] FAILED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[2-True-True-both-g0-idtype0] FAILED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[2-True-True-both-g0-idtype1] FAILED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[2-True-True-right-g0-idtype0] FAILED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[2-True-True-right-g0-idtype1] FAILED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[2-True-False-none-g0-idtype0] FAILED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[2-True-False-none-g0-idtype1] FAILED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[2-True-False-both-g0-idtype0] FAILED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[2-True-False-both-g0-idtype1] FAILED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[2-True-False-right-g0-idtype0] FAILED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[2-True-False-right-g0-idtype1] FAILED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[2-False-True-none-g0-idtype0] FAILED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[2-False-True-none-g0-idtype1] FAILED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[2-False-True-both-g0-idtype0] FAILED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[2-False-True-both-g0-idtype1] FAILED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[2-False-True-right-g0-idtype0] FAILED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[2-False-True-right-g0-idtype1] FAILED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[2-False-False-none-g0-idtype0] FAILED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[2-False-False-none-g0-idtype1] FAILED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[2-False-False-both-g0-idtype0] FAILED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[2-False-False-both-g0-idtype1] FAILED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[2-False-False-right-g0-idtype0] FAILED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[2-False-False-right-g0-idtype1] FAILED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-True-True-none-g0-idtype0] FAILED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-True-True-none-g0-idtype1] FAILED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-True-True-none-g1-idtype0] FAILED [ 22%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-True-True-none-g1-idtype1] FAILED [ 22%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-True-True-both-g0-idtype0] FAILED [ 22%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-True-True-both-g0-idtype1] FAILED [ 22%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-True-True-both-g1-idtype0] FAILED [ 22%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-True-True-both-g1-idtype1] FAILED [ 22%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-True-True-right-g0-idtype0] FAILED [ 22%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-True-True-right-g0-idtype1] FAILED [ 22%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-True-True-right-g1-idtype0] FAILED [ 22%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-True-True-right-g1-idtype1] FAILED [ 22%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-True-False-none-g0-idtype0] FAILED [ 22%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-True-False-none-g0-idtype1] FAILED [ 22%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-True-False-none-g1-idtype0] FAILED [ 22%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-True-False-none-g1-idtype1] FAILED [ 22%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-True-False-both-g0-idtype0] FAILED [ 22%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-True-False-both-g0-idtype1] FAILED [ 22%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-True-False-both-g1-idtype0] FAILED [ 22%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-True-False-both-g1-idtype1] FAILED [ 22%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-True-False-right-g0-idtype0] FAILED [ 22%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-True-False-right-g0-idtype1] FAILED [ 22%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-True-False-right-g1-idtype0] FAILED [ 22%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-True-False-right-g1-idtype1] FAILED [ 22%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-False-True-none-g0-idtype0] FAILED [ 22%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-False-True-none-g0-idtype1] FAILED [ 22%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-False-True-none-g1-idtype0] FAILED [ 22%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-False-True-none-g1-idtype1] FAILED [ 22%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-False-True-both-g0-idtype0] FAILED [ 22%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-False-True-both-g0-idtype1] FAILED [ 22%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-False-True-both-g1-idtype0] FAILED [ 22%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-False-True-both-g1-idtype1] FAILED [ 22%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-False-True-right-g0-idtype0] FAILED [ 22%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-False-True-right-g0-idtype1] FAILED [ 22%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-False-True-right-g1-idtype0] FAILED [ 22%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-False-True-right-g1-idtype1] FAILED [ 22%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-False-False-none-g0-idtype0] FAILED [ 22%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-False-False-none-g0-idtype1] FAILED [ 22%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-False-False-none-g1-idtype0] FAILED [ 22%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-False-False-none-g1-idtype1] FAILED [ 22%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-False-False-both-g0-idtype0] FAILED [ 22%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-False-False-both-g0-idtype1] FAILED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-False-False-both-g1-idtype0] FAILED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-False-False-both-g1-idtype1] FAILED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-False-False-right-g0-idtype0] FAILED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-False-False-right-g0-idtype1] FAILED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-False-False-right-g1-idtype0] FAILED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-False-False-right-g1-idtype1] FAILED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-True-True-none-g0-idtype0] FAILED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-True-True-none-g0-idtype1] FAILED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-True-True-none-g1-idtype0] FAILED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-True-True-none-g1-idtype1] FAILED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-True-True-both-g0-idtype0] FAILED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-True-True-both-g0-idtype1] FAILED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-True-True-both-g1-idtype0] FAILED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-True-True-both-g1-idtype1] FAILED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-True-True-right-g0-idtype0] FAILED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-True-True-right-g0-idtype1] FAILED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-True-True-right-g1-idtype0] FAILED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-True-True-right-g1-idtype1] FAILED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-True-False-none-g0-idtype0] FAILED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-True-False-none-g0-idtype1] FAILED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-True-False-none-g1-idtype0] FAILED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-True-False-none-g1-idtype1] FAILED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-True-False-both-g0-idtype0] FAILED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-True-False-both-g0-idtype1] FAILED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-True-False-both-g1-idtype0] FAILED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-True-False-both-g1-idtype1] FAILED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-True-False-right-g0-idtype0] FAILED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-True-False-right-g0-idtype1] FAILED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-True-False-right-g1-idtype0] FAILED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-True-False-right-g1-idtype1] FAILED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-False-True-none-g0-idtype0] FAILED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-False-True-none-g0-idtype1] FAILED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-False-True-none-g1-idtype0] FAILED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-False-True-none-g1-idtype1] FAILED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-False-True-both-g0-idtype0] FAILED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-False-True-both-g0-idtype1] FAILED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-False-True-both-g1-idtype0] FAILED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-False-True-both-g1-idtype1] FAILED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-False-True-right-g0-idtype0] FAILED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-False-True-right-g0-idtype1] FAILED [ 24%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-False-True-right-g1-idtype0] FAILED [ 24%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-False-True-right-g1-idtype1] FAILED [ 24%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-False-False-none-g0-idtype0] FAILED [ 24%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-False-False-none-g0-idtype1] FAILED [ 24%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-False-False-none-g1-idtype0] FAILED [ 24%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-False-False-none-g1-idtype1] FAILED [ 24%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-False-False-both-g0-idtype0] FAILED [ 24%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-False-False-both-g0-idtype1] FAILED [ 24%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-False-False-both-g1-idtype0] FAILED [ 24%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-False-False-both-g1-idtype1] FAILED [ 24%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-False-False-right-g0-idtype0] FAILED [ 24%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-False-False-right-g0-idtype1] FAILED [ 24%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-False-False-right-g1-idtype0] FAILED [ 24%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-False-False-right-g1-idtype1] FAILED [ 24%]
tests/pytorch/test_nn.py::test_tagconv[1] FAILED                         [ 24%]
tests/pytorch/test_nn.py::test_tagconv[2] FAILED                         [ 24%]
tests/pytorch/test_nn.py::test_set2set FAILED                            [ 24%]
tests/pytorch/test_nn.py::test_glob_att_pool FAILED                      [ 24%]
tests/pytorch/test_nn.py::test_simple_pool FAILED                        [ 24%]
tests/pytorch/test_nn.py::test_set_trans PASSED                          [ 24%]
tests/pytorch/test_nn.py::test_rgcn[1-idtype0] FAILED                    [ 24%]
tests/pytorch/test_nn.py::test_rgcn[1-idtype1] FAILED                    [ 24%]
tests/pytorch/test_nn.py::test_rgcn[8-idtype0] FAILED                    [ 24%]
tests/pytorch/test_nn.py::test_rgcn[8-idtype1] FAILED                    [ 24%]
tests/pytorch/test_nn.py::test_rgcn[32-idtype0] FAILED                   [ 24%]
tests/pytorch/test_nn.py::test_rgcn[32-idtype1] FAILED                   [ 24%]
tests/pytorch/test_nn.py::test_rgcn_default_nbasis[1-idtype0] FAILED     [ 24%]
tests/pytorch/test_nn.py::test_rgcn_default_nbasis[1-idtype1] FAILED     [ 24%]
tests/pytorch/test_nn.py::test_rgcn_default_nbasis[10-idtype0] FAILED    [ 24%]
tests/pytorch/test_nn.py::test_rgcn_default_nbasis[10-idtype1] FAILED    [ 24%]
tests/pytorch/test_nn.py::test_rgcn_default_nbasis[40-idtype0] FAILED    [ 24%]
tests/pytorch/test_nn.py::test_rgcn_default_nbasis[40-idtype1] FAILED    [ 24%]
tests/pytorch/test_nn.py::test_gat_conv[1-1-g0-idtype0] FAILED           [ 24%]
tests/pytorch/test_nn.py::test_gat_conv[1-1-g0-idtype1] FAILED           [ 24%]
tests/pytorch/test_nn.py::test_gat_conv[1-1-g1-idtype0] FAILED           [ 24%]
tests/pytorch/test_nn.py::test_gat_conv[1-1-g1-idtype1] FAILED           [ 24%]
tests/pytorch/test_nn.py::test_gat_conv[1-1-g2-idtype0] FAILED           [ 24%]
tests/pytorch/test_nn.py::test_gat_conv[1-1-g2-idtype1] FAILED           [ 24%]
tests/pytorch/test_nn.py::test_gat_conv[1-1-g3-idtype0] FAILED           [ 24%]
tests/pytorch/test_nn.py::test_gat_conv[1-1-g3-idtype1] FAILED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[1-1-g4-idtype0] FAILED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[1-1-g4-idtype1] FAILED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[1-1-g5-idtype0] FAILED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[1-1-g5-idtype1] FAILED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[1-1-g6-idtype0] FAILED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[1-1-g6-idtype1] FAILED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[1-5-g0-idtype0] FAILED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[1-5-g0-idtype1] FAILED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[1-5-g1-idtype0] FAILED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[1-5-g1-idtype1] FAILED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[1-5-g2-idtype0] FAILED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[1-5-g2-idtype1] FAILED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[1-5-g3-idtype0] FAILED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[1-5-g3-idtype1] FAILED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[1-5-g4-idtype0] FAILED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[1-5-g4-idtype1] FAILED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[1-5-g5-idtype0] FAILED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[1-5-g5-idtype1] FAILED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[1-5-g6-idtype0] FAILED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[1-5-g6-idtype1] FAILED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[4-1-g0-idtype0] FAILED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[4-1-g0-idtype1] FAILED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[4-1-g1-idtype0] FAILED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[4-1-g1-idtype1] FAILED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[4-1-g2-idtype0] FAILED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[4-1-g2-idtype1] FAILED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[4-1-g3-idtype0] FAILED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[4-1-g3-idtype1] FAILED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[4-1-g4-idtype0] FAILED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[4-1-g4-idtype1] FAILED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[4-1-g5-idtype0] FAILED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[4-1-g5-idtype1] FAILED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[4-1-g6-idtype0] FAILED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[4-1-g6-idtype1] FAILED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[4-5-g0-idtype0] FAILED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[4-5-g0-idtype1] FAILED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[4-5-g1-idtype0] FAILED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[4-5-g1-idtype1] FAILED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[4-5-g2-idtype0] FAILED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[4-5-g2-idtype1] FAILED           [ 26%]
tests/pytorch/test_nn.py::test_gat_conv[4-5-g3-idtype0] FAILED           [ 26%]
tests/pytorch/test_nn.py::test_gat_conv[4-5-g3-idtype1] FAILED           [ 26%]
tests/pytorch/test_nn.py::test_gat_conv[4-5-g4-idtype0] FAILED           [ 26%]
tests/pytorch/test_nn.py::test_gat_conv[4-5-g4-idtype1] FAILED           [ 26%]
tests/pytorch/test_nn.py::test_gat_conv[4-5-g5-idtype0] FAILED           [ 26%]
tests/pytorch/test_nn.py::test_gat_conv[4-5-g5-idtype1] FAILED           [ 26%]
tests/pytorch/test_nn.py::test_gat_conv[4-5-g6-idtype0] FAILED           [ 26%]
tests/pytorch/test_nn.py::test_gat_conv[4-5-g6-idtype1] FAILED           [ 26%]
tests/pytorch/test_nn.py::test_gat_conv_bi[1-1-g0-idtype0] FAILED        [ 26%]
tests/pytorch/test_nn.py::test_gat_conv_bi[1-1-g0-idtype1] FAILED        [ 26%]
tests/pytorch/test_nn.py::test_gat_conv_bi[1-1-g1-idtype0] FAILED        [ 26%]
tests/pytorch/test_nn.py::test_gat_conv_bi[1-1-g1-idtype1] FAILED        [ 26%]
tests/pytorch/test_nn.py::test_gat_conv_bi[1-2-g0-idtype0] FAILED        [ 26%]
tests/pytorch/test_nn.py::test_gat_conv_bi[1-2-g0-idtype1] FAILED        [ 26%]
tests/pytorch/test_nn.py::test_gat_conv_bi[1-2-g1-idtype0] FAILED        [ 26%]
tests/pytorch/test_nn.py::test_gat_conv_bi[1-2-g1-idtype1] FAILED        [ 26%]
tests/pytorch/test_nn.py::test_gat_conv_bi[4-1-g0-idtype0] FAILED        [ 26%]
tests/pytorch/test_nn.py::test_gat_conv_bi[4-1-g0-idtype1] FAILED        [ 26%]
tests/pytorch/test_nn.py::test_gat_conv_bi[4-1-g1-idtype0] FAILED        [ 26%]
tests/pytorch/test_nn.py::test_gat_conv_bi[4-1-g1-idtype1] FAILED        [ 26%]
tests/pytorch/test_nn.py::test_gat_conv_bi[4-2-g0-idtype0] FAILED        [ 26%]
tests/pytorch/test_nn.py::test_gat_conv_bi[4-2-g0-idtype1] FAILED        [ 26%]
tests/pytorch/test_nn.py::test_gat_conv_bi[4-2-g1-idtype0] FAILED        [ 26%]
tests/pytorch/test_nn.py::test_gat_conv_bi[4-2-g1-idtype1] FAILED        [ 26%]
tests/pytorch/test_nn.py::test_gatv2_conv[1-1-g0-idtype0] FAILED         [ 26%]
tests/pytorch/test_nn.py::test_gatv2_conv[1-1-g0-idtype1] FAILED         [ 26%]
tests/pytorch/test_nn.py::test_gatv2_conv[1-1-g1-idtype0] FAILED         [ 26%]
tests/pytorch/test_nn.py::test_gatv2_conv[1-1-g1-idtype1] FAILED         [ 26%]
tests/pytorch/test_nn.py::test_gatv2_conv[1-1-g2-idtype0] FAILED         [ 26%]
tests/pytorch/test_nn.py::test_gatv2_conv[1-1-g2-idtype1] FAILED         [ 26%]
tests/pytorch/test_nn.py::test_gatv2_conv[1-1-g3-idtype0] FAILED         [ 26%]
tests/pytorch/test_nn.py::test_gatv2_conv[1-1-g3-idtype1] FAILED         [ 26%]
tests/pytorch/test_nn.py::test_gatv2_conv[1-1-g4-idtype0] FAILED         [ 26%]
tests/pytorch/test_nn.py::test_gatv2_conv[1-1-g4-idtype1] FAILED         [ 26%]
tests/pytorch/test_nn.py::test_gatv2_conv[1-1-g5-idtype0] FAILED         [ 26%]
tests/pytorch/test_nn.py::test_gatv2_conv[1-1-g5-idtype1] FAILED         [ 26%]
tests/pytorch/test_nn.py::test_gatv2_conv[1-1-g6-idtype0] FAILED         [ 26%]
tests/pytorch/test_nn.py::test_gatv2_conv[1-1-g6-idtype1] FAILED         [ 26%]
tests/pytorch/test_nn.py::test_gatv2_conv[1-5-g0-idtype0] FAILED         [ 26%]
tests/pytorch/test_nn.py::test_gatv2_conv[1-5-g0-idtype1] FAILED         [ 27%]
tests/pytorch/test_nn.py::test_gatv2_conv[1-5-g1-idtype0] FAILED         [ 27%]
tests/pytorch/test_nn.py::test_gatv2_conv[1-5-g1-idtype1] FAILED         [ 27%]
tests/pytorch/test_nn.py::test_gatv2_conv[1-5-g2-idtype0] FAILED         [ 27%]
tests/pytorch/test_nn.py::test_gatv2_conv[1-5-g2-idtype1] FAILED         [ 27%]
tests/pytorch/test_nn.py::test_gatv2_conv[1-5-g3-idtype0] FAILED         [ 27%]
tests/pytorch/test_nn.py::test_gatv2_conv[1-5-g3-idtype1] FAILED         [ 27%]
tests/pytorch/test_nn.py::test_gatv2_conv[1-5-g4-idtype0] FAILED         [ 27%]
tests/pytorch/test_nn.py::test_gatv2_conv[1-5-g4-idtype1] FAILED         [ 27%]
tests/pytorch/test_nn.py::test_gatv2_conv[1-5-g5-idtype0] FAILED         [ 27%]
tests/pytorch/test_nn.py::test_gatv2_conv[1-5-g5-idtype1] FAILED         [ 27%]
tests/pytorch/test_nn.py::test_gatv2_conv[1-5-g6-idtype0] FAILED         [ 27%]
tests/pytorch/test_nn.py::test_gatv2_conv[1-5-g6-idtype1] FAILED         [ 27%]
tests/pytorch/test_nn.py::test_gatv2_conv[4-1-g0-idtype0] FAILED         [ 27%]
tests/pytorch/test_nn.py::test_gatv2_conv[4-1-g0-idtype1] FAILED         [ 27%]
tests/pytorch/test_nn.py::test_gatv2_conv[4-1-g1-idtype0] FAILED         [ 27%]
tests/pytorch/test_nn.py::test_gatv2_conv[4-1-g1-idtype1] FAILED         [ 27%]
tests/pytorch/test_nn.py::test_gatv2_conv[4-1-g2-idtype0] FAILED         [ 27%]
tests/pytorch/test_nn.py::test_gatv2_conv[4-1-g2-idtype1] FAILED         [ 27%]
tests/pytorch/test_nn.py::test_gatv2_conv[4-1-g3-idtype0] FAILED         [ 27%]
tests/pytorch/test_nn.py::test_gatv2_conv[4-1-g3-idtype1] FAILED         [ 27%]
tests/pytorch/test_nn.py::test_gatv2_conv[4-1-g4-idtype0] FAILED         [ 27%]
tests/pytorch/test_nn.py::test_gatv2_conv[4-1-g4-idtype1] FAILED         [ 27%]
tests/pytorch/test_nn.py::test_gatv2_conv[4-1-g5-idtype0] FAILED         [ 27%]
tests/pytorch/test_nn.py::test_gatv2_conv[4-1-g5-idtype1] FAILED         [ 27%]
tests/pytorch/test_nn.py::test_gatv2_conv[4-1-g6-idtype0] FAILED         [ 27%]
tests/pytorch/test_nn.py::test_gatv2_conv[4-1-g6-idtype1] FAILED         [ 27%]
tests/pytorch/test_nn.py::test_gatv2_conv[4-5-g0-idtype0] FAILED         [ 27%]
tests/pytorch/test_nn.py::test_gatv2_conv[4-5-g0-idtype1] FAILED         [ 27%]
tests/pytorch/test_nn.py::test_gatv2_conv[4-5-g1-idtype0] FAILED         [ 27%]
tests/pytorch/test_nn.py::test_gatv2_conv[4-5-g1-idtype1] FAILED         [ 27%]
tests/pytorch/test_nn.py::test_gatv2_conv[4-5-g2-idtype0] FAILED         [ 27%]
tests/pytorch/test_nn.py::test_gatv2_conv[4-5-g2-idtype1] FAILED         [ 27%]
tests/pytorch/test_nn.py::test_gatv2_conv[4-5-g3-idtype0] FAILED         [ 27%]
tests/pytorch/test_nn.py::test_gatv2_conv[4-5-g3-idtype1] FAILED         [ 27%]
tests/pytorch/test_nn.py::test_gatv2_conv[4-5-g4-idtype0] FAILED         [ 27%]
tests/pytorch/test_nn.py::test_gatv2_conv[4-5-g4-idtype1] FAILED         [ 27%]
tests/pytorch/test_nn.py::test_gatv2_conv[4-5-g5-idtype0] FAILED         [ 27%]
tests/pytorch/test_nn.py::test_gatv2_conv[4-5-g5-idtype1] FAILED         [ 27%]
tests/pytorch/test_nn.py::test_gatv2_conv[4-5-g6-idtype0] FAILED         [ 28%]
tests/pytorch/test_nn.py::test_gatv2_conv[4-5-g6-idtype1] FAILED         [ 28%]
tests/pytorch/test_nn.py::test_gatv2_conv_bi[1-1-g0-idtype0] FAILED      [ 28%]
tests/pytorch/test_nn.py::test_gatv2_conv_bi[1-1-g0-idtype1] FAILED      [ 28%]
tests/pytorch/test_nn.py::test_gatv2_conv_bi[1-1-g1-idtype0] FAILED      [ 28%]
tests/pytorch/test_nn.py::test_gatv2_conv_bi[1-1-g1-idtype1] FAILED      [ 28%]
tests/pytorch/test_nn.py::test_gatv2_conv_bi[1-2-g0-idtype0] FAILED      [ 28%]
tests/pytorch/test_nn.py::test_gatv2_conv_bi[1-2-g0-idtype1] FAILED      [ 28%]
tests/pytorch/test_nn.py::test_gatv2_conv_bi[1-2-g1-idtype0] FAILED      [ 28%]
tests/pytorch/test_nn.py::test_gatv2_conv_bi[1-2-g1-idtype1] FAILED      [ 28%]
tests/pytorch/test_nn.py::test_gatv2_conv_bi[4-1-g0-idtype0] FAILED      [ 28%]
tests/pytorch/test_nn.py::test_gatv2_conv_bi[4-1-g0-idtype1] FAILED      [ 28%]
tests/pytorch/test_nn.py::test_gatv2_conv_bi[4-1-g1-idtype0] FAILED      [ 28%]
tests/pytorch/test_nn.py::test_gatv2_conv_bi[4-1-g1-idtype1] FAILED      [ 28%]
tests/pytorch/test_nn.py::test_gatv2_conv_bi[4-2-g0-idtype0] FAILED      [ 28%]
tests/pytorch/test_nn.py::test_gatv2_conv_bi[4-2-g0-idtype1] FAILED      [ 28%]
tests/pytorch/test_nn.py::test_gatv2_conv_bi[4-2-g1-idtype0] FAILED      [ 28%]
tests/pytorch/test_nn.py::test_gatv2_conv_bi[4-2-g1-idtype1] FAILED      [ 28%]
tests/pytorch/test_nn.py::test_egat_conv[1-1-1-g0-idtype0] FAILED        [ 28%]
tests/pytorch/test_nn.py::test_egat_conv[1-1-1-g0-idtype1] FAILED        [ 28%]
tests/pytorch/test_nn.py::test_egat_conv[1-1-1-g1-idtype0] FAILED        [ 28%]
tests/pytorch/test_nn.py::test_egat_conv[1-1-1-g1-idtype1] FAILED        [ 28%]
tests/pytorch/test_nn.py::test_egat_conv[1-1-1-g2-idtype0] FAILED        [ 28%]
tests/pytorch/test_nn.py::test_egat_conv[1-1-1-g2-idtype1] FAILED        [ 28%]
tests/pytorch/test_nn.py::test_egat_conv[1-1-1-g3-idtype0] FAILED        [ 28%]
tests/pytorch/test_nn.py::test_egat_conv[1-1-1-g3-idtype1] FAILED        [ 28%]
tests/pytorch/test_nn.py::test_egat_conv[1-1-1-g4-idtype0] FAILED        [ 28%]
tests/pytorch/test_nn.py::test_egat_conv[1-1-1-g4-idtype1] FAILED        [ 28%]
tests/pytorch/test_nn.py::test_egat_conv[1-1-1-g5-idtype0] FAILED        [ 28%]
tests/pytorch/test_nn.py::test_egat_conv[1-1-1-g5-idtype1] FAILED        [ 28%]
tests/pytorch/test_nn.py::test_egat_conv[1-1-5-g0-idtype0] FAILED        [ 28%]
tests/pytorch/test_nn.py::test_egat_conv[1-1-5-g0-idtype1] FAILED        [ 28%]
tests/pytorch/test_nn.py::test_egat_conv[1-1-5-g1-idtype0] FAILED        [ 28%]
tests/pytorch/test_nn.py::test_egat_conv[1-1-5-g1-idtype1] FAILED        [ 28%]
tests/pytorch/test_nn.py::test_egat_conv[1-1-5-g2-idtype0] FAILED        [ 28%]
tests/pytorch/test_nn.py::test_egat_conv[1-1-5-g2-idtype1] FAILED        [ 28%]
tests/pytorch/test_nn.py::test_egat_conv[1-1-5-g3-idtype0] FAILED        [ 28%]
tests/pytorch/test_nn.py::test_egat_conv[1-1-5-g3-idtype1] FAILED        [ 28%]
tests/pytorch/test_nn.py::test_egat_conv[1-1-5-g4-idtype0] FAILED        [ 28%]
tests/pytorch/test_nn.py::test_egat_conv[1-1-5-g4-idtype1] FAILED        [ 28%]
tests/pytorch/test_nn.py::test_egat_conv[1-1-5-g5-idtype0] FAILED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[1-1-5-g5-idtype1] FAILED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[1-5-1-g0-idtype0] FAILED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[1-5-1-g0-idtype1] FAILED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[1-5-1-g1-idtype0] FAILED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[1-5-1-g1-idtype1] FAILED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[1-5-1-g2-idtype0] FAILED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[1-5-1-g2-idtype1] FAILED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[1-5-1-g3-idtype0] FAILED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[1-5-1-g3-idtype1] FAILED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[1-5-1-g4-idtype0] FAILED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[1-5-1-g4-idtype1] FAILED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[1-5-1-g5-idtype0] FAILED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[1-5-1-g5-idtype1] FAILED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[1-5-5-g0-idtype0] FAILED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[1-5-5-g0-idtype1] FAILED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[1-5-5-g1-idtype0] FAILED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[1-5-5-g1-idtype1] FAILED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[1-5-5-g2-idtype0] FAILED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[1-5-5-g2-idtype1] FAILED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[1-5-5-g3-idtype0] FAILED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[1-5-5-g3-idtype1] FAILED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[1-5-5-g4-idtype0] FAILED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[1-5-5-g4-idtype1] FAILED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[1-5-5-g5-idtype0] FAILED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[1-5-5-g5-idtype1] FAILED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[4-1-1-g0-idtype0] FAILED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[4-1-1-g0-idtype1] FAILED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[4-1-1-g1-idtype0] FAILED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[4-1-1-g1-idtype1] FAILED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[4-1-1-g2-idtype0] FAILED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[4-1-1-g2-idtype1] FAILED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[4-1-1-g3-idtype0] FAILED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[4-1-1-g3-idtype1] FAILED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[4-1-1-g4-idtype0] FAILED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[4-1-1-g4-idtype1] FAILED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[4-1-1-g5-idtype0] FAILED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[4-1-1-g5-idtype1] FAILED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[4-1-5-g0-idtype0] FAILED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[4-1-5-g0-idtype1] FAILED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[4-1-5-g1-idtype0] FAILED        [ 30%]
tests/pytorch/test_nn.py::test_egat_conv[4-1-5-g1-idtype1] FAILED        [ 30%]
tests/pytorch/test_nn.py::test_egat_conv[4-1-5-g2-idtype0] FAILED        [ 30%]
tests/pytorch/test_nn.py::test_egat_conv[4-1-5-g2-idtype1] FAILED        [ 30%]
tests/pytorch/test_nn.py::test_egat_conv[4-1-5-g3-idtype0] FAILED        [ 30%]
tests/pytorch/test_nn.py::test_egat_conv[4-1-5-g3-idtype1] FAILED        [ 30%]
tests/pytorch/test_nn.py::test_egat_conv[4-1-5-g4-idtype0] FAILED        [ 30%]
tests/pytorch/test_nn.py::test_egat_conv[4-1-5-g4-idtype1] FAILED        [ 30%]
tests/pytorch/test_nn.py::test_egat_conv[4-1-5-g5-idtype0] FAILED        [ 30%]
tests/pytorch/test_nn.py::test_egat_conv[4-1-5-g5-idtype1] FAILED        [ 30%]
tests/pytorch/test_nn.py::test_egat_conv[4-5-1-g0-idtype0] FAILED        [ 30%]
tests/pytorch/test_nn.py::test_egat_conv[4-5-1-g0-idtype1] FAILED        [ 30%]
tests/pytorch/test_nn.py::test_egat_conv[4-5-1-g1-idtype0] FAILED        [ 30%]
tests/pytorch/test_nn.py::test_egat_conv[4-5-1-g1-idtype1] FAILED        [ 30%]
tests/pytorch/test_nn.py::test_egat_conv[4-5-1-g2-idtype0] FAILED        [ 30%]
tests/pytorch/test_nn.py::test_egat_conv[4-5-1-g2-idtype1] FAILED        [ 30%]
tests/pytorch/test_nn.py::test_egat_conv[4-5-1-g3-idtype0] FAILED        [ 30%]
tests/pytorch/test_nn.py::test_egat_conv[4-5-1-g3-idtype1] FAILED        [ 30%]
tests/pytorch/test_nn.py::test_egat_conv[4-5-1-g4-idtype0] FAILED        [ 30%]
tests/pytorch/test_nn.py::test_egat_conv[4-5-1-g4-idtype1] FAILED        [ 30%]
tests/pytorch/test_nn.py::test_egat_conv[4-5-1-g5-idtype0] FAILED        [ 30%]
tests/pytorch/test_nn.py::test_egat_conv[4-5-1-g5-idtype1] FAILED        [ 30%]
tests/pytorch/test_nn.py::test_egat_conv[4-5-5-g0-idtype0] FAILED        [ 30%]
tests/pytorch/test_nn.py::test_egat_conv[4-5-5-g0-idtype1] FAILED        [ 30%]
tests/pytorch/test_nn.py::test_egat_conv[4-5-5-g1-idtype0] FAILED        [ 30%]
tests/pytorch/test_nn.py::test_egat_conv[4-5-5-g1-idtype1] FAILED        [ 30%]
tests/pytorch/test_nn.py::test_egat_conv[4-5-5-g2-idtype0] FAILED        [ 30%]
tests/pytorch/test_nn.py::test_egat_conv[4-5-5-g2-idtype1] FAILED        [ 30%]
tests/pytorch/test_nn.py::test_egat_conv[4-5-5-g3-idtype0] FAILED        [ 30%]
tests/pytorch/test_nn.py::test_egat_conv[4-5-5-g3-idtype1] FAILED        [ 30%]
tests/pytorch/test_nn.py::test_egat_conv[4-5-5-g4-idtype0] FAILED        [ 30%]
tests/pytorch/test_nn.py::test_egat_conv[4-5-5-g4-idtype1] FAILED        [ 30%]
tests/pytorch/test_nn.py::test_egat_conv[4-5-5-g5-idtype0] FAILED        [ 30%]
tests/pytorch/test_nn.py::test_egat_conv[4-5-5-g5-idtype1] FAILED        [ 30%]
tests/pytorch/test_nn.py::test_egat_conv_bi[1-1-1-g0-idtype0] FAILED     [ 30%]
tests/pytorch/test_nn.py::test_egat_conv_bi[1-1-1-g0-idtype1] FAILED     [ 30%]
tests/pytorch/test_nn.py::test_egat_conv_bi[1-1-1-g1-idtype0] FAILED     [ 30%]
tests/pytorch/test_nn.py::test_egat_conv_bi[1-1-1-g1-idtype1] FAILED     [ 30%]
tests/pytorch/test_nn.py::test_egat_conv_bi[1-1-5-g0-idtype0] FAILED     [ 30%]
tests/pytorch/test_nn.py::test_egat_conv_bi[1-1-5-g0-idtype1] FAILED     [ 30%]
tests/pytorch/test_nn.py::test_egat_conv_bi[1-1-5-g1-idtype0] FAILED     [ 31%]
tests/pytorch/test_nn.py::test_egat_conv_bi[1-1-5-g1-idtype1] FAILED     [ 31%]
tests/pytorch/test_nn.py::test_egat_conv_bi[1-5-1-g0-idtype0] FAILED     [ 31%]
tests/pytorch/test_nn.py::test_egat_conv_bi[1-5-1-g0-idtype1] FAILED     [ 31%]
tests/pytorch/test_nn.py::test_egat_conv_bi[1-5-1-g1-idtype0] FAILED     [ 31%]
tests/pytorch/test_nn.py::test_egat_conv_bi[1-5-1-g1-idtype1] FAILED     [ 31%]
tests/pytorch/test_nn.py::test_egat_conv_bi[1-5-5-g0-idtype0] FAILED     [ 31%]
tests/pytorch/test_nn.py::test_egat_conv_bi[1-5-5-g0-idtype1] FAILED     [ 31%]
tests/pytorch/test_nn.py::test_egat_conv_bi[1-5-5-g1-idtype0] FAILED     [ 31%]
tests/pytorch/test_nn.py::test_egat_conv_bi[1-5-5-g1-idtype1] FAILED     [ 31%]
tests/pytorch/test_nn.py::test_egat_conv_bi[4-1-1-g0-idtype0] FAILED     [ 31%]
tests/pytorch/test_nn.py::test_egat_conv_bi[4-1-1-g0-idtype1] FAILED     [ 31%]
tests/pytorch/test_nn.py::test_egat_conv_bi[4-1-1-g1-idtype0] FAILED     [ 31%]
tests/pytorch/test_nn.py::test_egat_conv_bi[4-1-1-g1-idtype1] FAILED     [ 31%]
tests/pytorch/test_nn.py::test_egat_conv_bi[4-1-5-g0-idtype0] FAILED     [ 31%]
tests/pytorch/test_nn.py::test_egat_conv_bi[4-1-5-g0-idtype1] FAILED     [ 31%]
tests/pytorch/test_nn.py::test_egat_conv_bi[4-1-5-g1-idtype0] FAILED     [ 31%]
tests/pytorch/test_nn.py::test_egat_conv_bi[4-1-5-g1-idtype1] FAILED     [ 31%]
tests/pytorch/test_nn.py::test_egat_conv_bi[4-5-1-g0-idtype0] FAILED     [ 31%]
tests/pytorch/test_nn.py::test_egat_conv_bi[4-5-1-g0-idtype1] FAILED     [ 31%]
tests/pytorch/test_nn.py::test_egat_conv_bi[4-5-1-g1-idtype0] FAILED     [ 31%]
tests/pytorch/test_nn.py::test_egat_conv_bi[4-5-1-g1-idtype1] FAILED     [ 31%]
tests/pytorch/test_nn.py::test_egat_conv_bi[4-5-5-g0-idtype0] FAILED     [ 31%]
tests/pytorch/test_nn.py::test_egat_conv_bi[4-5-5-g0-idtype1] FAILED     [ 31%]
tests/pytorch/test_nn.py::test_egat_conv_bi[4-5-5-g1-idtype0] FAILED     [ 31%]
tests/pytorch/test_nn.py::test_egat_conv_bi[4-5-5-g1-idtype1] FAILED     [ 31%]
tests/pytorch/test_nn.py::test_sage_conv[mean-g0-idtype0] FAILED         [ 31%]
tests/pytorch/test_nn.py::test_sage_conv[mean-g0-idtype1] FAILED         [ 31%]
tests/pytorch/test_nn.py::test_sage_conv[mean-g1-idtype0] FAILED         [ 31%]
tests/pytorch/test_nn.py::test_sage_conv[mean-g1-idtype1] FAILED         [ 31%]
tests/pytorch/test_nn.py::test_sage_conv[mean-g2-idtype0] FAILED         [ 31%]
tests/pytorch/test_nn.py::test_sage_conv[mean-g2-idtype1] FAILED         [ 31%]
tests/pytorch/test_nn.py::test_sage_conv[mean-g3-idtype0] FAILED         [ 31%]
tests/pytorch/test_nn.py::test_sage_conv[mean-g3-idtype1] FAILED         [ 31%]
tests/pytorch/test_nn.py::test_sage_conv[mean-g4-idtype0] FAILED         [ 31%]
tests/pytorch/test_nn.py::test_sage_conv[mean-g4-idtype1] FAILED         [ 31%]
tests/pytorch/test_nn.py::test_sage_conv[mean-g5-idtype0] FAILED         [ 31%]
tests/pytorch/test_nn.py::test_sage_conv[mean-g5-idtype1] FAILED         [ 31%]
tests/pytorch/test_nn.py::test_sage_conv[mean-g6-idtype0] FAILED         [ 31%]
tests/pytorch/test_nn.py::test_sage_conv[mean-g6-idtype1] FAILED         [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[mean-g7-idtype0] FAILED         [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[mean-g7-idtype1] FAILED         [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[pool-g0-idtype0] FAILED         [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[pool-g0-idtype1] FAILED         [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[pool-g1-idtype0] FAILED         [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[pool-g1-idtype1] FAILED         [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[pool-g2-idtype0] FAILED         [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[pool-g2-idtype1] FAILED         [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[pool-g3-idtype0] FAILED         [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[pool-g3-idtype1] FAILED         [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[pool-g4-idtype0] FAILED         [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[pool-g4-idtype1] FAILED         [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[pool-g5-idtype0] FAILED         [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[pool-g5-idtype1] FAILED         [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[pool-g6-idtype0] FAILED         [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[pool-g6-idtype1] FAILED         [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[pool-g7-idtype0] FAILED         [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[pool-g7-idtype1] FAILED         [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[gcn-g0-idtype0] FAILED          [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[gcn-g0-idtype1] FAILED          [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[gcn-g1-idtype0] FAILED          [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[gcn-g1-idtype1] FAILED          [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[gcn-g2-idtype0] FAILED          [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[gcn-g2-idtype1] FAILED          [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[gcn-g3-idtype0] FAILED          [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[gcn-g3-idtype1] FAILED          [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[gcn-g4-idtype0] FAILED          [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[gcn-g4-idtype1] FAILED          [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[gcn-g5-idtype0] FAILED          [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[gcn-g5-idtype1] FAILED          [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[gcn-g6-idtype0] FAILED          [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[gcn-g6-idtype1] FAILED          [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[gcn-g7-idtype0] FAILED          [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[gcn-g7-idtype1] FAILED          [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[lstm-g0-idtype0] FAILED         [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[lstm-g0-idtype1] FAILED         [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[lstm-g1-idtype0] FAILED         [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[lstm-g1-idtype1] FAILED         [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[lstm-g2-idtype0] FAILED         [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[lstm-g2-idtype1] FAILED         [ 33%]
tests/pytorch/test_nn.py::test_sage_conv[lstm-g3-idtype0] FAILED         [ 33%]
tests/pytorch/test_nn.py::test_sage_conv[lstm-g3-idtype1] FAILED         [ 33%]
tests/pytorch/test_nn.py::test_sage_conv[lstm-g4-idtype0] FAILED         [ 33%]
tests/pytorch/test_nn.py::test_sage_conv[lstm-g4-idtype1] FAILED         [ 33%]
tests/pytorch/test_nn.py::test_sage_conv[lstm-g5-idtype0] FAILED         [ 33%]
tests/pytorch/test_nn.py::test_sage_conv[lstm-g5-idtype1] FAILED         [ 33%]
tests/pytorch/test_nn.py::test_sage_conv[lstm-g6-idtype0] FAILED         [ 33%]
tests/pytorch/test_nn.py::test_sage_conv[lstm-g6-idtype1] FAILED         [ 33%]
tests/pytorch/test_nn.py::test_sage_conv[lstm-g7-idtype0] FAILED         [ 33%]
tests/pytorch/test_nn.py::test_sage_conv[lstm-g7-idtype1] FAILED         [ 33%]
tests/pytorch/test_nn.py::test_sage_conv_bi[1-mean-g0-idtype0] FAILED    [ 33%]
tests/pytorch/test_nn.py::test_sage_conv_bi[1-mean-g0-idtype1] FAILED    [ 33%]
tests/pytorch/test_nn.py::test_sage_conv_bi[1-mean-g1-idtype0] FAILED    [ 33%]
tests/pytorch/test_nn.py::test_sage_conv_bi[1-mean-g1-idtype1] FAILED    [ 33%]
tests/pytorch/test_nn.py::test_sage_conv_bi[1-mean-g2-idtype0] FAILED    [ 33%]
tests/pytorch/test_nn.py::test_sage_conv_bi[1-mean-g2-idtype1] FAILED    [ 33%]
tests/pytorch/test_nn.py::test_sage_conv_bi[1-pool-g0-idtype0] FAILED    [ 33%]
tests/pytorch/test_nn.py::test_sage_conv_bi[1-pool-g0-idtype1] FAILED    [ 33%]
tests/pytorch/test_nn.py::test_sage_conv_bi[1-pool-g1-idtype0] FAILED    [ 33%]
tests/pytorch/test_nn.py::test_sage_conv_bi[1-pool-g1-idtype1] FAILED    [ 33%]
tests/pytorch/test_nn.py::test_sage_conv_bi[1-pool-g2-idtype0] FAILED    [ 33%]
tests/pytorch/test_nn.py::test_sage_conv_bi[1-pool-g2-idtype1] FAILED    [ 33%]
tests/pytorch/test_nn.py::test_sage_conv_bi[1-gcn-g0-idtype0] FAILED     [ 33%]
tests/pytorch/test_nn.py::test_sage_conv_bi[1-gcn-g0-idtype1] FAILED     [ 33%]
tests/pytorch/test_nn.py::test_sage_conv_bi[1-gcn-g1-idtype0] FAILED     [ 33%]
tests/pytorch/test_nn.py::test_sage_conv_bi[1-gcn-g1-idtype1] FAILED     [ 33%]
tests/pytorch/test_nn.py::test_sage_conv_bi[1-gcn-g2-idtype0] FAILED     [ 33%]
tests/pytorch/test_nn.py::test_sage_conv_bi[1-gcn-g2-idtype1] FAILED     [ 33%]
tests/pytorch/test_nn.py::test_sage_conv_bi[1-lstm-g0-idtype0] FAILED    [ 33%]
tests/pytorch/test_nn.py::test_sage_conv_bi[1-lstm-g0-idtype1] FAILED    [ 33%]
tests/pytorch/test_nn.py::test_sage_conv_bi[1-lstm-g1-idtype0] FAILED    [ 33%]
tests/pytorch/test_nn.py::test_sage_conv_bi[1-lstm-g1-idtype1] FAILED    [ 33%]
tests/pytorch/test_nn.py::test_sage_conv_bi[1-lstm-g2-idtype0] FAILED    [ 33%]
tests/pytorch/test_nn.py::test_sage_conv_bi[1-lstm-g2-idtype1] FAILED    [ 33%]
tests/pytorch/test_nn.py::test_sage_conv_bi[2-mean-g0-idtype0] FAILED    [ 33%]
tests/pytorch/test_nn.py::test_sage_conv_bi[2-mean-g0-idtype1] FAILED    [ 33%]
tests/pytorch/test_nn.py::test_sage_conv_bi[2-mean-g1-idtype0] FAILED    [ 33%]
tests/pytorch/test_nn.py::test_sage_conv_bi[2-mean-g1-idtype1] FAILED    [ 33%]
tests/pytorch/test_nn.py::test_sage_conv_bi[2-mean-g2-idtype0] FAILED    [ 33%]
tests/pytorch/test_nn.py::test_sage_conv_bi[2-mean-g2-idtype1] FAILED    [ 34%]
tests/pytorch/test_nn.py::test_sage_conv_bi[2-pool-g0-idtype0] FAILED    [ 34%]
tests/pytorch/test_nn.py::test_sage_conv_bi[2-pool-g0-idtype1] FAILED    [ 34%]
tests/pytorch/test_nn.py::test_sage_conv_bi[2-pool-g1-idtype0] FAILED    [ 34%]
tests/pytorch/test_nn.py::test_sage_conv_bi[2-pool-g1-idtype1] FAILED    [ 34%]
tests/pytorch/test_nn.py::test_sage_conv_bi[2-pool-g2-idtype0] FAILED    [ 34%]
tests/pytorch/test_nn.py::test_sage_conv_bi[2-pool-g2-idtype1] FAILED    [ 34%]
tests/pytorch/test_nn.py::test_sage_conv_bi[2-gcn-g0-idtype0] FAILED     [ 34%]
tests/pytorch/test_nn.py::test_sage_conv_bi[2-gcn-g0-idtype1] FAILED     [ 34%]
tests/pytorch/test_nn.py::test_sage_conv_bi[2-gcn-g1-idtype0] FAILED     [ 34%]
tests/pytorch/test_nn.py::test_sage_conv_bi[2-gcn-g1-idtype1] FAILED     [ 34%]
tests/pytorch/test_nn.py::test_sage_conv_bi[2-gcn-g2-idtype0] FAILED     [ 34%]
tests/pytorch/test_nn.py::test_sage_conv_bi[2-gcn-g2-idtype1] FAILED     [ 34%]
tests/pytorch/test_nn.py::test_sage_conv_bi[2-lstm-g0-idtype0] FAILED    [ 34%]
tests/pytorch/test_nn.py::test_sage_conv_bi[2-lstm-g0-idtype1] FAILED    [ 34%]
tests/pytorch/test_nn.py::test_sage_conv_bi[2-lstm-g1-idtype0] FAILED    [ 34%]
tests/pytorch/test_nn.py::test_sage_conv_bi[2-lstm-g1-idtype1] FAILED    [ 34%]
tests/pytorch/test_nn.py::test_sage_conv_bi[2-lstm-g2-idtype0] FAILED    [ 34%]
tests/pytorch/test_nn.py::test_sage_conv_bi[2-lstm-g2-idtype1] FAILED    [ 34%]
tests/pytorch/test_nn.py::test_sage_conv2[1-idtype0] FAILED              [ 34%]
tests/pytorch/test_nn.py::test_sage_conv2[1-idtype1] FAILED              [ 34%]
tests/pytorch/test_nn.py::test_sage_conv2[2-idtype0] FAILED              [ 34%]
tests/pytorch/test_nn.py::test_sage_conv2[2-idtype1] FAILED              [ 34%]
tests/pytorch/test_nn.py::test_sgc_conv[1-g0-idtype0] FAILED             [ 34%]
tests/pytorch/test_nn.py::test_sgc_conv[1-g0-idtype1] FAILED             [ 34%]
tests/pytorch/test_nn.py::test_sgc_conv[1-g1-idtype0] FAILED             [ 34%]
tests/pytorch/test_nn.py::test_sgc_conv[1-g1-idtype1] FAILED             [ 34%]
tests/pytorch/test_nn.py::test_sgc_conv[1-g2-idtype0] FAILED             [ 34%]
tests/pytorch/test_nn.py::test_sgc_conv[1-g2-idtype1] FAILED             [ 34%]
tests/pytorch/test_nn.py::test_sgc_conv[1-g3-idtype0] FAILED             [ 34%]
tests/pytorch/test_nn.py::test_sgc_conv[1-g3-idtype1] FAILED             [ 34%]
tests/pytorch/test_nn.py::test_sgc_conv[1-g4-idtype0] FAILED             [ 34%]
tests/pytorch/test_nn.py::test_sgc_conv[1-g4-idtype1] FAILED             [ 34%]
tests/pytorch/test_nn.py::test_sgc_conv[1-g5-idtype0] FAILED             [ 34%]
tests/pytorch/test_nn.py::test_sgc_conv[1-g5-idtype1] FAILED             [ 34%]
tests/pytorch/test_nn.py::test_sgc_conv[2-g0-idtype0] FAILED             [ 34%]
tests/pytorch/test_nn.py::test_sgc_conv[2-g0-idtype1] FAILED             [ 34%]
tests/pytorch/test_nn.py::test_sgc_conv[2-g1-idtype0] FAILED             [ 34%]
tests/pytorch/test_nn.py::test_sgc_conv[2-g1-idtype1] FAILED             [ 34%]
tests/pytorch/test_nn.py::test_sgc_conv[2-g2-idtype0] FAILED             [ 34%]
tests/pytorch/test_nn.py::test_sgc_conv[2-g2-idtype1] FAILED             [ 35%]
tests/pytorch/test_nn.py::test_sgc_conv[2-g3-idtype0] FAILED             [ 35%]
tests/pytorch/test_nn.py::test_sgc_conv[2-g3-idtype1] FAILED             [ 35%]
tests/pytorch/test_nn.py::test_sgc_conv[2-g4-idtype0] FAILED             [ 35%]
tests/pytorch/test_nn.py::test_sgc_conv[2-g4-idtype1] FAILED             [ 35%]
tests/pytorch/test_nn.py::test_sgc_conv[2-g5-idtype0] FAILED             [ 35%]
tests/pytorch/test_nn.py::test_sgc_conv[2-g5-idtype1] FAILED             [ 35%]
tests/pytorch/test_nn.py::test_appnp_conv[g0-idtype0] FAILED             [ 35%]
tests/pytorch/test_nn.py::test_appnp_conv[g0-idtype1] FAILED             [ 35%]
tests/pytorch/test_nn.py::test_appnp_conv[g1-idtype0] FAILED             [ 35%]
tests/pytorch/test_nn.py::test_appnp_conv[g1-idtype1] FAILED             [ 35%]
tests/pytorch/test_nn.py::test_appnp_conv[g2-idtype0] FAILED             [ 35%]
tests/pytorch/test_nn.py::test_appnp_conv[g2-idtype1] FAILED             [ 35%]
tests/pytorch/test_nn.py::test_appnp_conv[g3-idtype0] FAILED             [ 35%]
tests/pytorch/test_nn.py::test_appnp_conv[g3-idtype1] FAILED             [ 35%]
tests/pytorch/test_nn.py::test_appnp_conv[g4-idtype0] FAILED             [ 35%]
tests/pytorch/test_nn.py::test_appnp_conv[g4-idtype1] FAILED             [ 35%]
tests/pytorch/test_nn.py::test_appnp_conv[g5-idtype0] FAILED             [ 35%]
tests/pytorch/test_nn.py::test_appnp_conv[g5-idtype1] FAILED             [ 35%]
tests/pytorch/test_nn.py::test_appnp_conv_e_weight[g0-idtype0] FAILED    [ 35%]
tests/pytorch/test_nn.py::test_appnp_conv_e_weight[g0-idtype1] FAILED    [ 35%]
tests/pytorch/test_nn.py::test_appnp_conv_e_weight[g1-idtype0] FAILED    [ 35%]
tests/pytorch/test_nn.py::test_appnp_conv_e_weight[g1-idtype1] FAILED    [ 35%]
tests/pytorch/test_nn.py::test_appnp_conv_e_weight[g2-idtype0] FAILED    [ 35%]
tests/pytorch/test_nn.py::test_appnp_conv_e_weight[g2-idtype1] FAILED    [ 35%]
tests/pytorch/test_nn.py::test_appnp_conv_e_weight[g3-idtype0] FAILED    [ 35%]
tests/pytorch/test_nn.py::test_appnp_conv_e_weight[g3-idtype1] FAILED    [ 35%]
tests/pytorch/test_nn.py::test_appnp_conv_e_weight[g4-idtype0] FAILED    [ 35%]
tests/pytorch/test_nn.py::test_appnp_conv_e_weight[g4-idtype1] FAILED    [ 35%]
tests/pytorch/test_nn.py::test_appnp_conv_e_weight[g5-idtype0] FAILED    [ 35%]
tests/pytorch/test_nn.py::test_appnp_conv_e_weight[g5-idtype1] FAILED    [ 35%]
tests/pytorch/test_nn.py::test_gcn2conv_e_weight[True-g0-idtype0] FAILED [ 35%]
tests/pytorch/test_nn.py::test_gcn2conv_e_weight[True-g0-idtype1] FAILED [ 35%]
tests/pytorch/test_nn.py::test_gcn2conv_e_weight[True-g1-idtype0] FAILED [ 35%]
tests/pytorch/test_nn.py::test_gcn2conv_e_weight[True-g1-idtype1] FAILED [ 35%]
tests/pytorch/test_nn.py::test_gcn2conv_e_weight[True-g2-idtype0] FAILED [ 35%]
tests/pytorch/test_nn.py::test_gcn2conv_e_weight[True-g2-idtype1] FAILED [ 35%]
tests/pytorch/test_nn.py::test_gcn2conv_e_weight[True-g3-idtype0] FAILED [ 35%]
tests/pytorch/test_nn.py::test_gcn2conv_e_weight[True-g3-idtype1] FAILED [ 35%]
tests/pytorch/test_nn.py::test_gcn2conv_e_weight[True-g4-idtype0] FAILED [ 35%]
tests/pytorch/test_nn.py::test_gcn2conv_e_weight[True-g4-idtype1] FAILED [ 36%]
tests/pytorch/test_nn.py::test_gcn2conv_e_weight[True-g5-idtype0] FAILED [ 36%]
tests/pytorch/test_nn.py::test_gcn2conv_e_weight[True-g5-idtype1] FAILED [ 36%]
tests/pytorch/test_nn.py::test_gcn2conv_e_weight[False-g0-idtype0] FAILED [ 36%]
tests/pytorch/test_nn.py::test_gcn2conv_e_weight[False-g0-idtype1] FAILED [ 36%]
tests/pytorch/test_nn.py::test_gcn2conv_e_weight[False-g1-idtype0] FAILED [ 36%]
tests/pytorch/test_nn.py::test_gcn2conv_e_weight[False-g1-idtype1] FAILED [ 36%]
tests/pytorch/test_nn.py::test_gcn2conv_e_weight[False-g2-idtype0] FAILED [ 36%]
tests/pytorch/test_nn.py::test_gcn2conv_e_weight[False-g2-idtype1] FAILED [ 36%]
tests/pytorch/test_nn.py::test_gcn2conv_e_weight[False-g3-idtype0] FAILED [ 36%]
tests/pytorch/test_nn.py::test_gcn2conv_e_weight[False-g3-idtype1] FAILED [ 36%]
tests/pytorch/test_nn.py::test_gcn2conv_e_weight[False-g4-idtype0] FAILED [ 36%]
tests/pytorch/test_nn.py::test_gcn2conv_e_weight[False-g4-idtype1] FAILED [ 36%]
tests/pytorch/test_nn.py::test_gcn2conv_e_weight[False-g5-idtype0] FAILED [ 36%]
tests/pytorch/test_nn.py::test_gcn2conv_e_weight[False-g5-idtype1] FAILED [ 36%]
tests/pytorch/test_nn.py::test_sgconv_e_weight[g0-idtype0] FAILED        [ 36%]
tests/pytorch/test_nn.py::test_sgconv_e_weight[g0-idtype1] FAILED        [ 36%]
tests/pytorch/test_nn.py::test_sgconv_e_weight[g1-idtype0] FAILED        [ 36%]
tests/pytorch/test_nn.py::test_sgconv_e_weight[g1-idtype1] FAILED        [ 36%]
tests/pytorch/test_nn.py::test_sgconv_e_weight[g2-idtype0] FAILED        [ 36%]
tests/pytorch/test_nn.py::test_sgconv_e_weight[g2-idtype1] FAILED        [ 36%]
tests/pytorch/test_nn.py::test_sgconv_e_weight[g3-idtype0] FAILED        [ 36%]
tests/pytorch/test_nn.py::test_sgconv_e_weight[g3-idtype1] FAILED        [ 36%]
tests/pytorch/test_nn.py::test_sgconv_e_weight[g4-idtype0] FAILED        [ 36%]
tests/pytorch/test_nn.py::test_sgconv_e_weight[g4-idtype1] FAILED        [ 36%]
tests/pytorch/test_nn.py::test_sgconv_e_weight[g5-idtype0] FAILED        [ 36%]
tests/pytorch/test_nn.py::test_sgconv_e_weight[g5-idtype1] FAILED        [ 36%]
tests/pytorch/test_nn.py::test_tagconv_e_weight[g0-idtype0] FAILED       [ 36%]
tests/pytorch/test_nn.py::test_tagconv_e_weight[g0-idtype1] FAILED       [ 36%]
tests/pytorch/test_nn.py::test_tagconv_e_weight[g1-idtype0] FAILED       [ 36%]
tests/pytorch/test_nn.py::test_tagconv_e_weight[g1-idtype1] FAILED       [ 36%]
tests/pytorch/test_nn.py::test_tagconv_e_weight[g2-idtype0] FAILED       [ 36%]
tests/pytorch/test_nn.py::test_tagconv_e_weight[g2-idtype1] FAILED       [ 36%]
tests/pytorch/test_nn.py::test_tagconv_e_weight[g3-idtype0] FAILED       [ 36%]
tests/pytorch/test_nn.py::test_tagconv_e_weight[g3-idtype1] FAILED       [ 36%]
tests/pytorch/test_nn.py::test_tagconv_e_weight[g4-idtype0] FAILED       [ 36%]
tests/pytorch/test_nn.py::test_tagconv_e_weight[g4-idtype1] FAILED       [ 36%]
tests/pytorch/test_nn.py::test_tagconv_e_weight[g5-idtype0] FAILED       [ 36%]
tests/pytorch/test_nn.py::test_tagconv_e_weight[g5-idtype1] FAILED       [ 36%]
tests/pytorch/test_nn.py::test_gin_conv[mean-g0-idtype0] FAILED          [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[mean-g0-idtype1] FAILED          [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[mean-g1-idtype0] FAILED          [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[mean-g1-idtype1] FAILED          [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[mean-g2-idtype0] FAILED          [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[mean-g2-idtype1] FAILED          [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[mean-g3-idtype0] FAILED          [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[mean-g3-idtype1] FAILED          [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[mean-g4-idtype0] FAILED          [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[mean-g4-idtype1] FAILED          [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[mean-g5-idtype0] FAILED          [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[mean-g5-idtype1] FAILED          [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[mean-g6-idtype0] FAILED          [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[mean-g6-idtype1] FAILED          [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[max-g0-idtype0] FAILED           [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[max-g0-idtype1] FAILED           [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[max-g1-idtype0] FAILED           [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[max-g1-idtype1] FAILED           [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[max-g2-idtype0] FAILED           [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[max-g2-idtype1] FAILED           [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[max-g3-idtype0] FAILED           [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[max-g3-idtype1] FAILED           [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[max-g4-idtype0] FAILED           [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[max-g4-idtype1] FAILED           [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[max-g5-idtype0] FAILED           [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[max-g5-idtype1] FAILED           [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[max-g6-idtype0] FAILED           [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[max-g6-idtype1] FAILED           [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[sum-g0-idtype0] FAILED           [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[sum-g0-idtype1] FAILED           [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[sum-g1-idtype0] FAILED           [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[sum-g1-idtype1] FAILED           [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[sum-g2-idtype0] FAILED           [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[sum-g2-idtype1] FAILED           [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[sum-g3-idtype0] FAILED           [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[sum-g3-idtype1] FAILED           [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[sum-g4-idtype0] FAILED           [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[sum-g4-idtype1] FAILED           [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[sum-g5-idtype0] FAILED           [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[sum-g5-idtype1] FAILED           [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[sum-g6-idtype0] FAILED           [ 38%]
tests/pytorch/test_nn.py::test_gin_conv[sum-g6-idtype1] FAILED           [ 38%]
tests/pytorch/test_nn.py::test_gine_conv[g0-idtype0] FAILED              [ 38%]
tests/pytorch/test_nn.py::test_gine_conv[g0-idtype1] FAILED              [ 38%]
tests/pytorch/test_nn.py::test_gine_conv[g1-idtype0] FAILED              [ 38%]
tests/pytorch/test_nn.py::test_gine_conv[g1-idtype1] FAILED              [ 38%]
tests/pytorch/test_nn.py::test_gine_conv[g2-idtype0] FAILED              [ 38%]
tests/pytorch/test_nn.py::test_gine_conv[g2-idtype1] FAILED              [ 38%]
tests/pytorch/test_nn.py::test_gine_conv[g3-idtype0] FAILED              [ 38%]
tests/pytorch/test_nn.py::test_gine_conv[g3-idtype1] FAILED              [ 38%]
tests/pytorch/test_nn.py::test_gine_conv[g4-idtype0] FAILED              [ 38%]
tests/pytorch/test_nn.py::test_gine_conv[g4-idtype1] FAILED              [ 38%]
tests/pytorch/test_nn.py::test_gine_conv[g5-idtype0] FAILED              [ 38%]
tests/pytorch/test_nn.py::test_gine_conv[g5-idtype1] FAILED              [ 38%]
tests/pytorch/test_nn.py::test_gine_conv[g6-idtype0] FAILED              [ 38%]
tests/pytorch/test_nn.py::test_gine_conv[g6-idtype1] FAILED              [ 38%]
tests/pytorch/test_nn.py::test_gine_conv[g7-idtype0] FAILED              [ 38%]
tests/pytorch/test_nn.py::test_gine_conv[g7-idtype1] FAILED              [ 38%]
tests/pytorch/test_nn.py::test_gin_conv_bi[mean-g0-idtype0] FAILED       [ 38%]
tests/pytorch/test_nn.py::test_gin_conv_bi[mean-g0-idtype1] FAILED       [ 38%]
tests/pytorch/test_nn.py::test_gin_conv_bi[mean-g1-idtype0] FAILED       [ 38%]
tests/pytorch/test_nn.py::test_gin_conv_bi[mean-g1-idtype1] FAILED       [ 38%]
tests/pytorch/test_nn.py::test_gin_conv_bi[max-g0-idtype0] FAILED        [ 38%]
tests/pytorch/test_nn.py::test_gin_conv_bi[max-g0-idtype1] FAILED        [ 38%]
tests/pytorch/test_nn.py::test_gin_conv_bi[max-g1-idtype0] FAILED        [ 38%]
tests/pytorch/test_nn.py::test_gin_conv_bi[max-g1-idtype1] FAILED        [ 38%]
tests/pytorch/test_nn.py::test_gin_conv_bi[sum-g0-idtype0] FAILED        [ 38%]
tests/pytorch/test_nn.py::test_gin_conv_bi[sum-g0-idtype1] FAILED        [ 38%]
tests/pytorch/test_nn.py::test_gin_conv_bi[sum-g1-idtype0] FAILED        [ 38%]
tests/pytorch/test_nn.py::test_gin_conv_bi[sum-g1-idtype1] FAILED        [ 38%]
tests/pytorch/test_nn.py::test_agnn_conv[g0-idtype0] FAILED              [ 38%]
tests/pytorch/test_nn.py::test_agnn_conv[g0-idtype1] FAILED              [ 38%]
tests/pytorch/test_nn.py::test_agnn_conv[g1-idtype0] FAILED              [ 38%]
tests/pytorch/test_nn.py::test_agnn_conv[g1-idtype1] FAILED              [ 38%]
tests/pytorch/test_nn.py::test_agnn_conv[g2-idtype0] FAILED              [ 38%]
tests/pytorch/test_nn.py::test_agnn_conv[g2-idtype1] FAILED              [ 38%]
tests/pytorch/test_nn.py::test_agnn_conv[g3-idtype0] FAILED              [ 38%]
tests/pytorch/test_nn.py::test_agnn_conv[g3-idtype1] FAILED              [ 38%]
tests/pytorch/test_nn.py::test_agnn_conv[g4-idtype0] FAILED              [ 38%]
tests/pytorch/test_nn.py::test_agnn_conv[g4-idtype1] FAILED              [ 38%]
tests/pytorch/test_nn.py::test_agnn_conv[g5-idtype0] FAILED              [ 39%]
tests/pytorch/test_nn.py::test_agnn_conv[g5-idtype1] FAILED              [ 39%]
tests/pytorch/test_nn.py::test_agnn_conv[g6-idtype0] FAILED              [ 39%]
tests/pytorch/test_nn.py::test_agnn_conv[g6-idtype1] FAILED              [ 39%]
tests/pytorch/test_nn.py::test_agnn_conv_bi[g0-idtype0] FAILED           [ 39%]
tests/pytorch/test_nn.py::test_agnn_conv_bi[g0-idtype1] FAILED           [ 39%]
tests/pytorch/test_nn.py::test_agnn_conv_bi[g1-idtype0] FAILED           [ 39%]
tests/pytorch/test_nn.py::test_agnn_conv_bi[g1-idtype1] FAILED           [ 39%]
tests/pytorch/test_nn.py::test_gated_graph_conv[g0-idtype0] FAILED       [ 39%]
tests/pytorch/test_nn.py::test_gated_graph_conv[g0-idtype1] FAILED       [ 39%]
tests/pytorch/test_nn.py::test_gated_graph_conv[g1-idtype0] FAILED       [ 39%]
tests/pytorch/test_nn.py::test_gated_graph_conv[g1-idtype1] FAILED       [ 39%]
tests/pytorch/test_nn.py::test_gated_graph_conv[g2-idtype0] FAILED       [ 39%]
tests/pytorch/test_nn.py::test_gated_graph_conv[g2-idtype1] FAILED       [ 39%]
tests/pytorch/test_nn.py::test_gated_graph_conv[g3-idtype0] FAILED       [ 39%]
tests/pytorch/test_nn.py::test_gated_graph_conv[g3-idtype1] FAILED       [ 39%]
tests/pytorch/test_nn.py::test_gated_graph_conv[g4-idtype0] FAILED       [ 39%]
tests/pytorch/test_nn.py::test_gated_graph_conv[g4-idtype1] FAILED       [ 39%]
tests/pytorch/test_nn.py::test_gated_graph_conv[g5-idtype0] FAILED       [ 39%]
tests/pytorch/test_nn.py::test_gated_graph_conv[g5-idtype1] FAILED       [ 39%]
tests/pytorch/test_nn.py::test_gated_graph_conv_one_etype[g0-idtype0] FAILED [ 39%]
tests/pytorch/test_nn.py::test_gated_graph_conv_one_etype[g0-idtype1] FAILED [ 39%]
tests/pytorch/test_nn.py::test_gated_graph_conv_one_etype[g1-idtype0] FAILED [ 39%]
tests/pytorch/test_nn.py::test_gated_graph_conv_one_etype[g1-idtype1] FAILED [ 39%]
tests/pytorch/test_nn.py::test_gated_graph_conv_one_etype[g2-idtype0] FAILED [ 39%]
tests/pytorch/test_nn.py::test_gated_graph_conv_one_etype[g2-idtype1] FAILED [ 39%]
tests/pytorch/test_nn.py::test_gated_graph_conv_one_etype[g3-idtype0] FAILED [ 39%]
tests/pytorch/test_nn.py::test_gated_graph_conv_one_etype[g3-idtype1] FAILED [ 39%]
tests/pytorch/test_nn.py::test_gated_graph_conv_one_etype[g4-idtype0] FAILED [ 39%]
tests/pytorch/test_nn.py::test_gated_graph_conv_one_etype[g4-idtype1] FAILED [ 39%]
tests/pytorch/test_nn.py::test_gated_graph_conv_one_etype[g5-idtype0] FAILED [ 39%]
tests/pytorch/test_nn.py::test_gated_graph_conv_one_etype[g5-idtype1] FAILED [ 39%]
tests/pytorch/test_nn.py::test_nn_conv[g0-idtype0] FAILED                [ 39%]
tests/pytorch/test_nn.py::test_nn_conv[g0-idtype1] FAILED                [ 39%]
tests/pytorch/test_nn.py::test_nn_conv[g1-idtype0] FAILED                [ 39%]
tests/pytorch/test_nn.py::test_nn_conv[g1-idtype1] FAILED                [ 39%]
tests/pytorch/test_nn.py::test_nn_conv[g2-idtype0] FAILED                [ 39%]
tests/pytorch/test_nn.py::test_nn_conv[g2-idtype1] FAILED                [ 39%]
tests/pytorch/test_nn.py::test_nn_conv[g3-idtype0] FAILED                [ 39%]
tests/pytorch/test_nn.py::test_nn_conv[g3-idtype1] FAILED                [ 39%]
tests/pytorch/test_nn.py::test_nn_conv[g4-idtype0] FAILED                [ 40%]
tests/pytorch/test_nn.py::test_nn_conv[g4-idtype1] FAILED                [ 40%]
tests/pytorch/test_nn.py::test_nn_conv[g5-idtype0] FAILED                [ 40%]
tests/pytorch/test_nn.py::test_nn_conv[g5-idtype1] FAILED                [ 40%]
tests/pytorch/test_nn.py::test_nn_conv[g6-idtype0] FAILED                [ 40%]
tests/pytorch/test_nn.py::test_nn_conv[g6-idtype1] FAILED                [ 40%]
tests/pytorch/test_nn.py::test_nn_conv_bi[g0-idtype0] FAILED             [ 40%]
tests/pytorch/test_nn.py::test_nn_conv_bi[g0-idtype1] FAILED             [ 40%]
tests/pytorch/test_nn.py::test_nn_conv_bi[g1-idtype0] FAILED             [ 40%]
tests/pytorch/test_nn.py::test_nn_conv_bi[g1-idtype1] FAILED             [ 40%]
tests/pytorch/test_nn.py::test_gmm_conv[g0-idtype0] FAILED               [ 40%]
tests/pytorch/test_nn.py::test_gmm_conv[g0-idtype1] FAILED               [ 40%]
tests/pytorch/test_nn.py::test_gmm_conv[g1-idtype0] FAILED               [ 40%]
tests/pytorch/test_nn.py::test_gmm_conv[g1-idtype1] FAILED               [ 40%]
tests/pytorch/test_nn.py::test_gmm_conv[g2-idtype0] FAILED               [ 40%]
tests/pytorch/test_nn.py::test_gmm_conv[g2-idtype1] FAILED               [ 40%]
tests/pytorch/test_nn.py::test_gmm_conv[g3-idtype0] FAILED               [ 40%]
tests/pytorch/test_nn.py::test_gmm_conv[g3-idtype1] FAILED               [ 40%]
tests/pytorch/test_nn.py::test_gmm_conv[g4-idtype0] FAILED               [ 40%]
tests/pytorch/test_nn.py::test_gmm_conv[g4-idtype1] FAILED               [ 40%]
tests/pytorch/test_nn.py::test_gmm_conv[g5-idtype0] FAILED               [ 40%]
tests/pytorch/test_nn.py::test_gmm_conv[g5-idtype1] FAILED               [ 40%]
tests/pytorch/test_nn.py::test_gmm_conv_bi[g0-idtype0] FAILED            [ 40%]
tests/pytorch/test_nn.py::test_gmm_conv_bi[g0-idtype1] FAILED            [ 40%]
tests/pytorch/test_nn.py::test_gmm_conv_bi[g1-idtype0] FAILED            [ 40%]
tests/pytorch/test_nn.py::test_gmm_conv_bi[g1-idtype1] FAILED            [ 40%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g0-both-idtype0] FAILED [ 40%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g0-both-idtype1] FAILED [ 40%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g0-right-idtype0] FAILED [ 40%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g0-right-idtype1] FAILED [ 40%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g0-none-idtype0] FAILED [ 40%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g0-none-idtype1] FAILED [ 40%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g1-both-idtype0] FAILED [ 40%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g1-both-idtype1] FAILED [ 40%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g1-right-idtype0] FAILED [ 40%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g1-right-idtype1] FAILED [ 40%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g1-none-idtype0] FAILED [ 40%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g1-none-idtype1] FAILED [ 40%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g2-both-idtype0] FAILED [ 40%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g2-both-idtype1] FAILED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g2-right-idtype0] FAILED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g2-right-idtype1] FAILED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g2-none-idtype0] FAILED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g2-none-idtype1] FAILED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g3-both-idtype0] FAILED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g3-both-idtype1] FAILED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g3-right-idtype0] FAILED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g3-right-idtype1] FAILED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g3-none-idtype0] FAILED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g3-none-idtype1] FAILED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g4-both-idtype0] FAILED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g4-both-idtype1] FAILED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g4-right-idtype0] FAILED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g4-right-idtype1] FAILED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g4-none-idtype0] FAILED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g4-none-idtype1] FAILED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g5-both-idtype0] FAILED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g5-both-idtype1] FAILED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g5-right-idtype0] FAILED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g5-right-idtype1] FAILED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g5-none-idtype0] FAILED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g5-none-idtype1] FAILED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g6-both-idtype0] FAILED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g6-both-idtype1] FAILED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g6-right-idtype0] FAILED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g6-right-idtype1] FAILED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g6-none-idtype0] FAILED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g6-none-idtype1] FAILED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g7-both-idtype0] FAILED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g7-both-idtype1] FAILED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g7-right-idtype0] FAILED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g7-right-idtype1] FAILED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g7-none-idtype0] FAILED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g7-none-idtype1] FAILED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g0-both-idtype0] FAILED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g0-both-idtype1] FAILED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g0-right-idtype0] FAILED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g0-right-idtype1] FAILED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g0-none-idtype0] FAILED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g0-none-idtype1] FAILED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g1-both-idtype0] FAILED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g1-both-idtype1] FAILED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g1-right-idtype0] FAILED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g1-right-idtype1] FAILED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g1-none-idtype0] FAILED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g1-none-idtype1] FAILED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g2-both-idtype0] FAILED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g2-both-idtype1] FAILED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g2-right-idtype0] FAILED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g2-right-idtype1] FAILED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g2-none-idtype0] FAILED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g2-none-idtype1] FAILED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g3-both-idtype0] FAILED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g3-both-idtype1] FAILED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g3-right-idtype0] FAILED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g3-right-idtype1] FAILED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g3-none-idtype0] FAILED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g3-none-idtype1] FAILED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g4-both-idtype0] FAILED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g4-both-idtype1] FAILED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g4-right-idtype0] FAILED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g4-right-idtype1] FAILED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g4-none-idtype0] FAILED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g4-none-idtype1] FAILED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g5-both-idtype0] FAILED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g5-both-idtype1] FAILED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g5-right-idtype0] FAILED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g5-right-idtype1] FAILED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g5-none-idtype0] FAILED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g5-none-idtype1] FAILED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g6-both-idtype0] FAILED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g6-both-idtype1] FAILED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g6-right-idtype0] FAILED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g6-right-idtype1] FAILED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g6-none-idtype0] FAILED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g6-none-idtype1] FAILED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g7-both-idtype0] FAILED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g7-both-idtype1] FAILED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g7-right-idtype0] FAILED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g7-right-idtype1] FAILED [ 43%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g7-none-idtype0] FAILED [ 43%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g7-none-idtype1] FAILED [ 43%]
tests/pytorch/test_nn.py::test_dense_sage_conv[1-g0-idtype0] FAILED      [ 43%]
tests/pytorch/test_nn.py::test_dense_sage_conv[1-g0-idtype1] FAILED      [ 43%]
tests/pytorch/test_nn.py::test_dense_sage_conv[1-g1-idtype0] FAILED      [ 43%]
tests/pytorch/test_nn.py::test_dense_sage_conv[1-g1-idtype1] FAILED      [ 43%]
tests/pytorch/test_nn.py::test_dense_sage_conv[1-g2-idtype0] FAILED      [ 43%]
tests/pytorch/test_nn.py::test_dense_sage_conv[1-g2-idtype1] FAILED      [ 43%]
tests/pytorch/test_nn.py::test_dense_sage_conv[1-g3-idtype0] FAILED      [ 43%]
tests/pytorch/test_nn.py::test_dense_sage_conv[1-g3-idtype1] FAILED      [ 43%]
tests/pytorch/test_nn.py::test_dense_sage_conv[1-g4-idtype0] FAILED      [ 43%]
tests/pytorch/test_nn.py::test_dense_sage_conv[1-g4-idtype1] FAILED      [ 43%]
tests/pytorch/test_nn.py::test_dense_sage_conv[1-g5-idtype0] FAILED      [ 43%]
tests/pytorch/test_nn.py::test_dense_sage_conv[1-g5-idtype1] FAILED      [ 43%]
tests/pytorch/test_nn.py::test_dense_sage_conv[1-g6-idtype0] FAILED      [ 43%]
tests/pytorch/test_nn.py::test_dense_sage_conv[1-g6-idtype1] FAILED      [ 43%]
tests/pytorch/test_nn.py::test_dense_sage_conv[1-g7-idtype0] FAILED      [ 43%]
tests/pytorch/test_nn.py::test_dense_sage_conv[1-g7-idtype1] FAILED      [ 43%]
tests/pytorch/test_nn.py::test_dense_sage_conv[1-g8-idtype0] FAILED      [ 43%]
tests/pytorch/test_nn.py::test_dense_sage_conv[1-g8-idtype1] FAILED      [ 43%]
tests/pytorch/test_nn.py::test_dense_sage_conv[1-g9-idtype0] FAILED      [ 43%]
tests/pytorch/test_nn.py::test_dense_sage_conv[1-g9-idtype1] FAILED      [ 43%]
tests/pytorch/test_nn.py::test_dense_sage_conv[2-g0-idtype0] FAILED      [ 43%]
tests/pytorch/test_nn.py::test_dense_sage_conv[2-g0-idtype1] FAILED      [ 43%]
tests/pytorch/test_nn.py::test_dense_sage_conv[2-g1-idtype0] FAILED      [ 43%]
tests/pytorch/test_nn.py::test_dense_sage_conv[2-g1-idtype1] FAILED      [ 43%]
tests/pytorch/test_nn.py::test_dense_sage_conv[2-g2-idtype0] FAILED      [ 43%]
tests/pytorch/test_nn.py::test_dense_sage_conv[2-g2-idtype1] FAILED      [ 43%]
tests/pytorch/test_nn.py::test_dense_sage_conv[2-g3-idtype0] FAILED      [ 43%]
tests/pytorch/test_nn.py::test_dense_sage_conv[2-g3-idtype1] FAILED      [ 43%]
tests/pytorch/test_nn.py::test_dense_sage_conv[2-g4-idtype0] FAILED      [ 43%]
tests/pytorch/test_nn.py::test_dense_sage_conv[2-g4-idtype1] FAILED      [ 43%]
tests/pytorch/test_nn.py::test_dense_sage_conv[2-g5-idtype0] FAILED      [ 43%]
tests/pytorch/test_nn.py::test_dense_sage_conv[2-g5-idtype1] FAILED      [ 43%]
tests/pytorch/test_nn.py::test_dense_sage_conv[2-g6-idtype0] FAILED      [ 43%]
tests/pytorch/test_nn.py::test_dense_sage_conv[2-g6-idtype1] FAILED      [ 43%]
tests/pytorch/test_nn.py::test_dense_sage_conv[2-g7-idtype0] FAILED      [ 43%]
tests/pytorch/test_nn.py::test_dense_sage_conv[2-g7-idtype1] FAILED      [ 43%]
tests/pytorch/test_nn.py::test_dense_sage_conv[2-g8-idtype0] FAILED      [ 43%]
tests/pytorch/test_nn.py::test_dense_sage_conv[2-g8-idtype1] FAILED      [ 44%]
tests/pytorch/test_nn.py::test_dense_sage_conv[2-g9-idtype0] FAILED      [ 44%]
tests/pytorch/test_nn.py::test_dense_sage_conv[2-g9-idtype1] FAILED      [ 44%]
tests/pytorch/test_nn.py::test_edge_conv[1-g0-idtype0] FAILED            [ 44%]
tests/pytorch/test_nn.py::test_edge_conv[1-g0-idtype1] FAILED            [ 44%]
tests/pytorch/test_nn.py::test_edge_conv[1-g1-idtype0] FAILED            [ 44%]
tests/pytorch/test_nn.py::test_edge_conv[1-g1-idtype1] FAILED            [ 44%]
tests/pytorch/test_nn.py::test_edge_conv[1-g2-idtype0] FAILED            [ 44%]
tests/pytorch/test_nn.py::test_edge_conv[1-g2-idtype1] FAILED            [ 44%]
tests/pytorch/test_nn.py::test_edge_conv[1-g3-idtype0] FAILED            [ 44%]
tests/pytorch/test_nn.py::test_edge_conv[1-g3-idtype1] FAILED            [ 44%]
tests/pytorch/test_nn.py::test_edge_conv[1-g4-idtype0] FAILED            [ 44%]
tests/pytorch/test_nn.py::test_edge_conv[1-g4-idtype1] FAILED            [ 44%]
tests/pytorch/test_nn.py::test_edge_conv[1-g5-idtype0] FAILED            [ 44%]
tests/pytorch/test_nn.py::test_edge_conv[1-g5-idtype1] FAILED            [ 44%]
tests/pytorch/test_nn.py::test_edge_conv[1-g6-idtype0] FAILED            [ 44%]
tests/pytorch/test_nn.py::test_edge_conv[1-g6-idtype1] FAILED            [ 44%]
tests/pytorch/test_nn.py::test_edge_conv[2-g0-idtype0] FAILED            [ 44%]
tests/pytorch/test_nn.py::test_edge_conv[2-g0-idtype1] FAILED            [ 44%]
tests/pytorch/test_nn.py::test_edge_conv[2-g1-idtype0] FAILED            [ 44%]
tests/pytorch/test_nn.py::test_edge_conv[2-g1-idtype1] FAILED            [ 44%]
tests/pytorch/test_nn.py::test_edge_conv[2-g2-idtype0] FAILED            [ 44%]
tests/pytorch/test_nn.py::test_edge_conv[2-g2-idtype1] FAILED            [ 44%]
tests/pytorch/test_nn.py::test_edge_conv[2-g3-idtype0] FAILED            [ 44%]
tests/pytorch/test_nn.py::test_edge_conv[2-g3-idtype1] FAILED            [ 44%]
tests/pytorch/test_nn.py::test_edge_conv[2-g4-idtype0] FAILED            [ 44%]
tests/pytorch/test_nn.py::test_edge_conv[2-g4-idtype1] FAILED            [ 44%]
tests/pytorch/test_nn.py::test_edge_conv[2-g5-idtype0] FAILED            [ 44%]
tests/pytorch/test_nn.py::test_edge_conv[2-g5-idtype1] FAILED            [ 44%]
tests/pytorch/test_nn.py::test_edge_conv[2-g6-idtype0] FAILED            [ 44%]
tests/pytorch/test_nn.py::test_edge_conv[2-g6-idtype1] FAILED            [ 44%]
tests/pytorch/test_nn.py::test_edge_conv_bi[1-g0-idtype0] FAILED         [ 44%]
tests/pytorch/test_nn.py::test_edge_conv_bi[1-g0-idtype1] FAILED         [ 44%]
tests/pytorch/test_nn.py::test_edge_conv_bi[1-g1-idtype0] FAILED         [ 44%]
tests/pytorch/test_nn.py::test_edge_conv_bi[1-g1-idtype1] FAILED         [ 44%]
tests/pytorch/test_nn.py::test_edge_conv_bi[2-g0-idtype0] FAILED         [ 44%]
tests/pytorch/test_nn.py::test_edge_conv_bi[2-g0-idtype1] FAILED         [ 44%]
tests/pytorch/test_nn.py::test_edge_conv_bi[2-g1-idtype0] FAILED         [ 44%]
tests/pytorch/test_nn.py::test_edge_conv_bi[2-g1-idtype1] FAILED         [ 44%]
tests/pytorch/test_nn.py::test_dotgat_conv[1-1-g0-idtype0] FAILED        [ 44%]
tests/pytorch/test_nn.py::test_dotgat_conv[1-1-g0-idtype1] FAILED        [ 45%]
tests/pytorch/test_nn.py::test_dotgat_conv[1-1-g1-idtype0] FAILED        [ 45%]
tests/pytorch/test_nn.py::test_dotgat_conv[1-1-g1-idtype1] FAILED        [ 45%]
tests/pytorch/test_nn.py::test_dotgat_conv[1-1-g2-idtype0] FAILED        [ 45%]
tests/pytorch/test_nn.py::test_dotgat_conv[1-1-g2-idtype1] FAILED        [ 45%]
tests/pytorch/test_nn.py::test_dotgat_conv[1-1-g3-idtype0] FAILED        [ 45%]
tests/pytorch/test_nn.py::test_dotgat_conv[1-1-g3-idtype1] FAILED        [ 45%]
tests/pytorch/test_nn.py::test_dotgat_conv[1-1-g4-idtype0] FAILED        [ 45%]
tests/pytorch/test_nn.py::test_dotgat_conv[1-1-g4-idtype1] FAILED        [ 45%]
tests/pytorch/test_nn.py::test_dotgat_conv[1-1-g5-idtype0] FAILED        [ 45%]
tests/pytorch/test_nn.py::test_dotgat_conv[1-1-g5-idtype1] FAILED        [ 45%]
tests/pytorch/test_nn.py::test_dotgat_conv[1-1-g6-idtype0] FAILED        [ 45%]
tests/pytorch/test_nn.py::test_dotgat_conv[1-1-g6-idtype1] FAILED        [ 45%]
tests/pytorch/test_nn.py::test_dotgat_conv[1-2-g0-idtype0] FAILED        [ 45%]
tests/pytorch/test_nn.py::test_dotgat_conv[1-2-g0-idtype1] FAILED        [ 45%]
tests/pytorch/test_nn.py::test_dotgat_conv[1-2-g1-idtype0] FAILED        [ 45%]
tests/pytorch/test_nn.py::test_dotgat_conv[1-2-g1-idtype1] FAILED        [ 45%]
tests/pytorch/test_nn.py::test_dotgat_conv[1-2-g2-idtype0] FAILED        [ 45%]
tests/pytorch/test_nn.py::test_dotgat_conv[1-2-g2-idtype1] FAILED        [ 45%]
tests/pytorch/test_nn.py::test_dotgat_conv[1-2-g3-idtype0] FAILED        [ 45%]
tests/pytorch/test_nn.py::test_dotgat_conv[1-2-g3-idtype1] FAILED        [ 45%]
tests/pytorch/test_nn.py::test_dotgat_conv[1-2-g4-idtype0] FAILED        [ 45%]
tests/pytorch/test_nn.py::test_dotgat_conv[1-2-g4-idtype1] FAILED        [ 45%]
tests/pytorch/test_nn.py::test_dotgat_conv[1-2-g5-idtype0] FAILED        [ 45%]
tests/pytorch/test_nn.py::test_dotgat_conv[1-2-g5-idtype1] FAILED        [ 45%]
tests/pytorch/test_nn.py::test_dotgat_conv[1-2-g6-idtype0] FAILED        [ 45%]
tests/pytorch/test_nn.py::test_dotgat_conv[1-2-g6-idtype1] FAILED        [ 45%]
tests/pytorch/test_nn.py::test_dotgat_conv[4-1-g0-idtype0] FAILED        [ 45%]
tests/pytorch/test_nn.py::test_dotgat_conv[4-1-g0-idtype1] FAILED        [ 45%]
tests/pytorch/test_nn.py::test_dotgat_conv[4-1-g1-idtype0] FAILED        [ 45%]
tests/pytorch/test_nn.py::test_dotgat_conv[4-1-g1-idtype1] FAILED        [ 45%]
tests/pytorch/test_nn.py::test_dotgat_conv[4-1-g2-idtype0] FAILED        [ 45%]
tests/pytorch/test_nn.py::test_dotgat_conv[4-1-g2-idtype1] FAILED        [ 45%]
tests/pytorch/test_nn.py::test_dotgat_conv[4-1-g3-idtype0] FAILED        [ 45%]
tests/pytorch/test_nn.py::test_dotgat_conv[4-1-g3-idtype1] FAILED        [ 45%]
tests/pytorch/test_nn.py::test_dotgat_conv[4-1-g4-idtype0] FAILED        [ 45%]
tests/pytorch/test_nn.py::test_dotgat_conv[4-1-g4-idtype1] FAILED        [ 45%]
tests/pytorch/test_nn.py::test_dotgat_conv[4-1-g5-idtype0] FAILED        [ 45%]
tests/pytorch/test_nn.py::test_dotgat_conv[4-1-g5-idtype1] FAILED        [ 45%]
tests/pytorch/test_nn.py::test_dotgat_conv[4-1-g6-idtype0] FAILED        [ 46%]
tests/pytorch/test_nn.py::test_dotgat_conv[4-1-g6-idtype1] FAILED        [ 46%]
tests/pytorch/test_nn.py::test_dotgat_conv[4-2-g0-idtype0] FAILED        [ 46%]
tests/pytorch/test_nn.py::test_dotgat_conv[4-2-g0-idtype1] FAILED        [ 46%]
tests/pytorch/test_nn.py::test_dotgat_conv[4-2-g1-idtype0] FAILED        [ 46%]
tests/pytorch/test_nn.py::test_dotgat_conv[4-2-g1-idtype1] FAILED        [ 46%]
tests/pytorch/test_nn.py::test_dotgat_conv[4-2-g2-idtype0] FAILED        [ 46%]
tests/pytorch/test_nn.py::test_dotgat_conv[4-2-g2-idtype1] FAILED        [ 46%]
tests/pytorch/test_nn.py::test_dotgat_conv[4-2-g3-idtype0] FAILED        [ 46%]
tests/pytorch/test_nn.py::test_dotgat_conv[4-2-g3-idtype1] FAILED        [ 46%]
tests/pytorch/test_nn.py::test_dotgat_conv[4-2-g4-idtype0] FAILED        [ 46%]
tests/pytorch/test_nn.py::test_dotgat_conv[4-2-g4-idtype1] FAILED        [ 46%]
tests/pytorch/test_nn.py::test_dotgat_conv[4-2-g5-idtype0] FAILED        [ 46%]
tests/pytorch/test_nn.py::test_dotgat_conv[4-2-g5-idtype1] FAILED        [ 46%]
tests/pytorch/test_nn.py::test_dotgat_conv[4-2-g6-idtype0] FAILED        [ 46%]
tests/pytorch/test_nn.py::test_dotgat_conv[4-2-g6-idtype1] FAILED        [ 46%]
tests/pytorch/test_nn.py::test_dotgat_conv_bi[1-1-g0-idtype0] FAILED     [ 46%]
tests/pytorch/test_nn.py::test_dotgat_conv_bi[1-1-g0-idtype1] FAILED     [ 46%]
tests/pytorch/test_nn.py::test_dotgat_conv_bi[1-1-g1-idtype0] FAILED     [ 46%]
tests/pytorch/test_nn.py::test_dotgat_conv_bi[1-1-g1-idtype1] FAILED     [ 46%]
tests/pytorch/test_nn.py::test_dotgat_conv_bi[1-2-g0-idtype0] FAILED     [ 46%]
tests/pytorch/test_nn.py::test_dotgat_conv_bi[1-2-g0-idtype1] FAILED     [ 46%]
tests/pytorch/test_nn.py::test_dotgat_conv_bi[1-2-g1-idtype0] FAILED     [ 46%]
tests/pytorch/test_nn.py::test_dotgat_conv_bi[1-2-g1-idtype1] FAILED     [ 46%]
tests/pytorch/test_nn.py::test_dotgat_conv_bi[4-1-g0-idtype0] FAILED     [ 46%]
tests/pytorch/test_nn.py::test_dotgat_conv_bi[4-1-g0-idtype1] FAILED     [ 46%]
tests/pytorch/test_nn.py::test_dotgat_conv_bi[4-1-g1-idtype0] FAILED     [ 46%]
tests/pytorch/test_nn.py::test_dotgat_conv_bi[4-1-g1-idtype1] FAILED     [ 46%]
tests/pytorch/test_nn.py::test_dotgat_conv_bi[4-2-g0-idtype0] FAILED     [ 46%]
tests/pytorch/test_nn.py::test_dotgat_conv_bi[4-2-g0-idtype1] FAILED     [ 46%]
tests/pytorch/test_nn.py::test_dotgat_conv_bi[4-2-g1-idtype0] FAILED     [ 46%]
tests/pytorch/test_nn.py::test_dotgat_conv_bi[4-2-g1-idtype1] FAILED     [ 46%]
tests/pytorch/test_nn.py::test_dense_cheb_conv[1] FAILED                 [ 46%]
tests/pytorch/test_nn.py::test_dense_cheb_conv[2] FAILED                 [ 46%]
tests/pytorch/test_nn.py::test_sequential FAILED                         [ 46%]
tests/pytorch/test_nn.py::test_atomic_conv[g0-idtype0] PASSED            [ 46%]
tests/pytorch/test_nn.py::test_atomic_conv[g0-idtype1] PASSED            [ 46%]
tests/pytorch/test_nn.py::test_atomic_conv[g1-idtype0] PASSED            [ 46%]
tests/pytorch/test_nn.py::test_atomic_conv[g1-idtype1] PASSED            [ 46%]
tests/pytorch/test_nn.py::test_atomic_conv[g2-idtype0] PASSED            [ 46%]
tests/pytorch/test_nn.py::test_atomic_conv[g2-idtype1] PASSED            [ 47%]
tests/pytorch/test_nn.py::test_atomic_conv[g3-idtype0] PASSED            [ 47%]
tests/pytorch/test_nn.py::test_atomic_conv[g3-idtype1] PASSED            [ 47%]
tests/pytorch/test_nn.py::test_atomic_conv[g4-idtype0] PASSED            [ 47%]
tests/pytorch/test_nn.py::test_atomic_conv[g4-idtype1] PASSED            [ 47%]
tests/pytorch/test_nn.py::test_atomic_conv[g5-idtype0] PASSED            [ 47%]
tests/pytorch/test_nn.py::test_atomic_conv[g5-idtype1] PASSED            [ 47%]
tests/pytorch/test_nn.py::test_cf_conv[1-g0-idtype0] FAILED              [ 47%]
tests/pytorch/test_nn.py::test_cf_conv[1-g0-idtype1] FAILED              [ 47%]
tests/pytorch/test_nn.py::test_cf_conv[1-g1-idtype0] FAILED              [ 47%]
tests/pytorch/test_nn.py::test_cf_conv[1-g1-idtype1] FAILED              [ 47%]
tests/pytorch/test_nn.py::test_cf_conv[1-g2-idtype0] FAILED              [ 47%]
tests/pytorch/test_nn.py::test_cf_conv[1-g2-idtype1] FAILED              [ 47%]
tests/pytorch/test_nn.py::test_cf_conv[1-g3-idtype0] FAILED              [ 47%]
tests/pytorch/test_nn.py::test_cf_conv[1-g3-idtype1] FAILED              [ 47%]
tests/pytorch/test_nn.py::test_cf_conv[1-g4-idtype0] FAILED              [ 47%]
tests/pytorch/test_nn.py::test_cf_conv[1-g4-idtype1] FAILED              [ 47%]
tests/pytorch/test_nn.py::test_cf_conv[1-g5-idtype0] FAILED              [ 47%]
tests/pytorch/test_nn.py::test_cf_conv[1-g5-idtype1] FAILED              [ 47%]
tests/pytorch/test_nn.py::test_cf_conv[1-g6-idtype0] FAILED              [ 47%]
tests/pytorch/test_nn.py::test_cf_conv[1-g6-idtype1] FAILED              [ 47%]
tests/pytorch/test_nn.py::test_cf_conv[1-g7-idtype0] FAILED              [ 47%]
tests/pytorch/test_nn.py::test_cf_conv[1-g7-idtype1] FAILED              [ 47%]
tests/pytorch/test_nn.py::test_cf_conv[3-g0-idtype0] FAILED              [ 47%]
tests/pytorch/test_nn.py::test_cf_conv[3-g0-idtype1] FAILED              [ 47%]
tests/pytorch/test_nn.py::test_cf_conv[3-g1-idtype0] FAILED              [ 47%]
tests/pytorch/test_nn.py::test_cf_conv[3-g1-idtype1] FAILED              [ 47%]
tests/pytorch/test_nn.py::test_cf_conv[3-g2-idtype0] FAILED              [ 47%]
tests/pytorch/test_nn.py::test_cf_conv[3-g2-idtype1] FAILED              [ 47%]
tests/pytorch/test_nn.py::test_cf_conv[3-g3-idtype0] FAILED              [ 47%]
tests/pytorch/test_nn.py::test_cf_conv[3-g3-idtype1] FAILED              [ 47%]
tests/pytorch/test_nn.py::test_cf_conv[3-g4-idtype0] FAILED              [ 47%]
tests/pytorch/test_nn.py::test_cf_conv[3-g4-idtype1] FAILED              [ 47%]
tests/pytorch/test_nn.py::test_cf_conv[3-g5-idtype0] FAILED              [ 47%]
tests/pytorch/test_nn.py::test_cf_conv[3-g5-idtype1] FAILED              [ 47%]
tests/pytorch/test_nn.py::test_cf_conv[3-g6-idtype0] FAILED              [ 47%]
tests/pytorch/test_nn.py::test_cf_conv[3-g6-idtype1] FAILED              [ 47%]
tests/pytorch/test_nn.py::test_cf_conv[3-g7-idtype0] FAILED              [ 47%]
tests/pytorch/test_nn.py::test_cf_conv[3-g7-idtype1] FAILED              [ 47%]
tests/pytorch/test_nn.py::test_hetero_conv[False-sum-idtype0] FAILED     [ 47%]
tests/pytorch/test_nn.py::test_hetero_conv[False-sum-idtype1] FAILED     [ 48%]
tests/pytorch/test_nn.py::test_hetero_conv[False-max-idtype0] FAILED     [ 48%]
tests/pytorch/test_nn.py::test_hetero_conv[False-max-idtype1] FAILED     [ 48%]
tests/pytorch/test_nn.py::test_hetero_conv[False-min-idtype0] FAILED     [ 48%]
tests/pytorch/test_nn.py::test_hetero_conv[False-min-idtype1] FAILED     [ 48%]
tests/pytorch/test_nn.py::test_hetero_conv[False-mean-idtype0] FAILED    [ 48%]
tests/pytorch/test_nn.py::test_hetero_conv[False-mean-idtype1] FAILED    [ 48%]
tests/pytorch/test_nn.py::test_hetero_conv[False-stack-idtype0] FAILED   [ 48%]
tests/pytorch/test_nn.py::test_hetero_conv[False-stack-idtype1] FAILED   [ 48%]
tests/pytorch/test_nn.py::test_hetero_conv[False-myagg-idtype0] FAILED   [ 48%]
tests/pytorch/test_nn.py::test_hetero_conv[False-myagg-idtype1] FAILED   [ 48%]
tests/pytorch/test_nn.py::test_hetero_conv[True-sum-idtype0] FAILED      [ 48%]
tests/pytorch/test_nn.py::test_hetero_conv[True-sum-idtype1] FAILED      [ 48%]
tests/pytorch/test_nn.py::test_hetero_conv[True-max-idtype0] FAILED      [ 48%]
tests/pytorch/test_nn.py::test_hetero_conv[True-max-idtype1] FAILED      [ 48%]
tests/pytorch/test_nn.py::test_hetero_conv[True-min-idtype0] FAILED      [ 48%]
tests/pytorch/test_nn.py::test_hetero_conv[True-min-idtype1] FAILED      [ 48%]
tests/pytorch/test_nn.py::test_hetero_conv[True-mean-idtype0] FAILED     [ 48%]
tests/pytorch/test_nn.py::test_hetero_conv[True-mean-idtype1] FAILED     [ 48%]
tests/pytorch/test_nn.py::test_hetero_conv[True-stack-idtype0] FAILED    [ 48%]
tests/pytorch/test_nn.py::test_hetero_conv[True-stack-idtype1] FAILED    [ 48%]
tests/pytorch/test_nn.py::test_hetero_conv[True-myagg-idtype0] FAILED    [ 48%]
tests/pytorch/test_nn.py::test_hetero_conv[True-myagg-idtype1] FAILED    [ 48%]
tests/pytorch/test_nn.py::test_hetero_linear[1] PASSED                   [ 48%]
tests/pytorch/test_nn.py::test_hetero_linear[2] PASSED                   [ 48%]
tests/pytorch/test_nn.py::test_hetero_linear[100] PASSED                 [ 48%]
tests/pytorch/test_nn.py::test_hetero_embedding[1] PASSED                [ 48%]
tests/pytorch/test_nn.py::test_hetero_embedding[2] PASSED                [ 48%]
tests/pytorch/test_nn.py::test_hetero_embedding[100] PASSED              [ 48%]
tests/pytorch/test_nn.py::test_gnnexplainer[1-g0-idtype0] FAILED         [ 48%]
tests/pytorch/test_nn.py::test_gnnexplainer[1-g0-idtype1] FAILED         [ 48%]
tests/pytorch/test_nn.py::test_gnnexplainer[1-g1-idtype0] FAILED         [ 48%]
tests/pytorch/test_nn.py::test_gnnexplainer[1-g1-idtype1] FAILED         [ 48%]
tests/pytorch/test_nn.py::test_gnnexplainer[1-g2-idtype0] FAILED         [ 48%]
tests/pytorch/test_nn.py::test_gnnexplainer[1-g2-idtype1] FAILED         [ 48%]
tests/pytorch/test_nn.py::test_gnnexplainer[1-g3-idtype0] FAILED         [ 48%]
tests/pytorch/test_nn.py::test_gnnexplainer[1-g3-idtype1] FAILED         [ 48%]
tests/pytorch/test_nn.py::test_gnnexplainer[1-g4-idtype0] FAILED         [ 48%]
tests/pytorch/test_nn.py::test_gnnexplainer[1-g4-idtype1] FAILED         [ 48%]
tests/pytorch/test_nn.py::test_gnnexplainer[1-g5-idtype0] FAILED         [ 48%]
tests/pytorch/test_nn.py::test_gnnexplainer[1-g5-idtype1] FAILED         [ 49%]
tests/pytorch/test_nn.py::test_gnnexplainer[2-g0-idtype0] FAILED         [ 49%]
tests/pytorch/test_nn.py::test_gnnexplainer[2-g0-idtype1] FAILED         [ 49%]
tests/pytorch/test_nn.py::test_gnnexplainer[2-g1-idtype0] FAILED         [ 49%]
tests/pytorch/test_nn.py::test_gnnexplainer[2-g1-idtype1] FAILED         [ 49%]
tests/pytorch/test_nn.py::test_gnnexplainer[2-g2-idtype0] FAILED         [ 49%]
tests/pytorch/test_nn.py::test_gnnexplainer[2-g2-idtype1] FAILED         [ 49%]
tests/pytorch/test_nn.py::test_gnnexplainer[2-g3-idtype0] FAILED         [ 49%]
tests/pytorch/test_nn.py::test_gnnexplainer[2-g3-idtype1] FAILED         [ 49%]
tests/pytorch/test_nn.py::test_gnnexplainer[2-g4-idtype0] FAILED         [ 49%]
tests/pytorch/test_nn.py::test_gnnexplainer[2-g4-idtype1] FAILED         [ 49%]
tests/pytorch/test_nn.py::test_gnnexplainer[2-g5-idtype0] FAILED         [ 49%]
tests/pytorch/test_nn.py::test_gnnexplainer[2-g5-idtype1] FAILED         [ 49%]
tests/pytorch/test_nn.py::test_jumping_knowledge PASSED                  [ 49%]
tests/pytorch/test_nn.py::test_edge_predictor[dot] PASSED                [ 49%]
tests/pytorch/test_nn.py::test_edge_predictor[cos] PASSED                [ 49%]
tests/pytorch/test_nn.py::test_edge_predictor[ele] PASSED                [ 49%]
tests/pytorch/test_nn.py::test_edge_predictor[cat] PASSED                [ 49%]
tests/pytorch/test_nn.py::test_ke_score_funcs PASSED                     [ 49%]
tests/pytorch/test_nn.py::test_twirls FAILED                             [ 49%]
tests/pytorch/test_nn.py::test_typed_linear[None-None-4] PASSED          [ 49%]
tests/pytorch/test_nn.py::test_typed_linear[None-None-32] PASSED         [ 49%]
tests/pytorch/test_nn.py::test_typed_linear[basis-4-4] PASSED            [ 49%]
tests/pytorch/test_nn.py::test_typed_linear[basis-4-32] PASSED           [ 49%]
tests/pytorch/test_nn.py::test_typed_linear[bdd-4-4] PASSED              [ 49%]
tests/pytorch/test_nn.py::test_typed_linear[bdd-4-32] PASSED             [ 49%]
tests/pytorch/test_nn.py::test_hgt[1-4-idtype0] FAILED                   [ 49%]
tests/pytorch/test_nn.py::test_hgt[1-4-idtype1] FAILED                   [ 49%]
tests/pytorch/test_nn.py::test_radius_graph[True-True] PASSED            [ 49%]
tests/pytorch/test_nn.py::test_radius_graph[True-False] PASSED           [ 49%]
tests/pytorch/test_nn.py::test_radius_graph[False-True] PASSED           [ 49%]
tests/pytorch/test_nn.py::test_radius_graph[False-False] PASSED          [ 49%]
tests/pytorch/test_nn.py::test_group_rev_res[idtype0] FAILED             [ 49%]
tests/pytorch/test_nn.py::test_group_rev_res[idtype1] FAILED             [ 49%]
tests/pytorch/test_nn.py::test_egnn_conv[16-16-16-16] FAILED             [ 49%]
tests/pytorch/test_nn.py::test_egnn_conv[16-16-16-32] FAILED             [ 49%]
tests/pytorch/test_nn.py::test_egnn_conv[16-16-32-16] FAILED             [ 49%]
tests/pytorch/test_nn.py::test_egnn_conv[16-16-32-32] FAILED             [ 49%]
tests/pytorch/test_nn.py::test_egnn_conv[16-32-16-16] FAILED             [ 49%]
tests/pytorch/test_nn.py::test_egnn_conv[16-32-16-32] FAILED             [ 50%]
tests/pytorch/test_nn.py::test_egnn_conv[16-32-32-16] FAILED             [ 50%]
tests/pytorch/test_nn.py::test_egnn_conv[16-32-32-32] FAILED             [ 50%]
tests/pytorch/test_nn.py::test_egnn_conv[10-16-16-16] FAILED             [ 50%]
tests/pytorch/test_nn.py::test_egnn_conv[10-16-16-32] FAILED             [ 50%]
tests/pytorch/test_nn.py::test_egnn_conv[10-16-32-16] FAILED             [ 50%]
tests/pytorch/test_nn.py::test_egnn_conv[10-16-32-32] FAILED             [ 50%]
tests/pytorch/test_nn.py::test_egnn_conv[10-32-16-16] FAILED             [ 50%]
tests/pytorch/test_nn.py::test_egnn_conv[10-32-16-32] FAILED             [ 50%]
tests/pytorch/test_nn.py::test_egnn_conv[10-32-32-16] FAILED             [ 50%]
tests/pytorch/test_nn.py::test_egnn_conv[10-32-32-32] FAILED             [ 50%]
tests/pytorch/test_nn.py::test_egnn_conv[0-16-16-16] FAILED              [ 50%]
tests/pytorch/test_nn.py::test_egnn_conv[0-16-16-32] FAILED              [ 50%]
tests/pytorch/test_nn.py::test_egnn_conv[0-16-32-16] FAILED              [ 50%]
tests/pytorch/test_nn.py::test_egnn_conv[0-16-32-32] FAILED              [ 50%]
tests/pytorch/test_nn.py::test_egnn_conv[0-32-16-16] FAILED              [ 50%]
tests/pytorch/test_nn.py::test_egnn_conv[0-32-16-32] FAILED              [ 50%]
tests/pytorch/test_nn.py::test_egnn_conv[0-32-32-16] FAILED              [ 50%]
tests/pytorch/test_nn.py::test_egnn_conv[0-32-32-32] FAILED              [ 50%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-2.5-scalers0-aggregators0-16-16] PASSED [ 50%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-2.5-scalers0-aggregators0-16-32] PASSED [ 50%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-2.5-scalers0-aggregators0-32-16] PASSED [ 50%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-2.5-scalers0-aggregators0-32-32] PASSED [ 50%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-2.5-scalers0-aggregators1-16-16] PASSED [ 50%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-2.5-scalers0-aggregators1-16-32] PASSED [ 50%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-2.5-scalers0-aggregators1-32-16] PASSED [ 50%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-2.5-scalers0-aggregators1-32-32] PASSED [ 50%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-2.5-scalers0-aggregators2-16-16] PASSED [ 50%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-2.5-scalers0-aggregators2-16-32] PASSED [ 50%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-2.5-scalers0-aggregators2-32-16] PASSED [ 50%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-2.5-scalers0-aggregators2-32-32] PASSED [ 50%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-2.5-scalers1-aggregators0-16-16] PASSED [ 50%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-2.5-scalers1-aggregators0-16-32] PASSED [ 50%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-2.5-scalers1-aggregators0-32-16] PASSED [ 50%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-2.5-scalers1-aggregators0-32-32] PASSED [ 50%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-2.5-scalers1-aggregators1-16-16] PASSED [ 50%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-2.5-scalers1-aggregators1-16-32] PASSED [ 50%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-2.5-scalers1-aggregators1-32-16] PASSED [ 50%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-2.5-scalers1-aggregators1-32-32] PASSED [ 50%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-2.5-scalers1-aggregators2-16-16] PASSED [ 50%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-2.5-scalers1-aggregators2-16-32] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-2.5-scalers1-aggregators2-32-16] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-2.5-scalers1-aggregators2-32-32] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-7.4-scalers0-aggregators0-16-16] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-7.4-scalers0-aggregators0-16-32] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-7.4-scalers0-aggregators0-32-16] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-7.4-scalers0-aggregators0-32-32] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-7.4-scalers0-aggregators1-16-16] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-7.4-scalers0-aggregators1-16-32] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-7.4-scalers0-aggregators1-32-16] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-7.4-scalers0-aggregators1-32-32] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-7.4-scalers0-aggregators2-16-16] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-7.4-scalers0-aggregators2-16-32] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-7.4-scalers0-aggregators2-32-16] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-7.4-scalers0-aggregators2-32-32] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-7.4-scalers1-aggregators0-16-16] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-7.4-scalers1-aggregators0-16-32] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-7.4-scalers1-aggregators0-32-16] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-7.4-scalers1-aggregators0-32-32] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-7.4-scalers1-aggregators1-16-16] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-7.4-scalers1-aggregators1-16-32] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-7.4-scalers1-aggregators1-32-16] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-7.4-scalers1-aggregators1-32-32] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-7.4-scalers1-aggregators2-16-16] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-7.4-scalers1-aggregators2-16-32] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-7.4-scalers1-aggregators2-32-16] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-7.4-scalers1-aggregators2-32-32] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-2.5-scalers0-aggregators0-16-16] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-2.5-scalers0-aggregators0-16-32] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-2.5-scalers0-aggregators0-32-16] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-2.5-scalers0-aggregators0-32-32] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-2.5-scalers0-aggregators1-16-16] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-2.5-scalers0-aggregators1-16-32] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-2.5-scalers0-aggregators1-32-16] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-2.5-scalers0-aggregators1-32-32] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-2.5-scalers0-aggregators2-16-16] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-2.5-scalers0-aggregators2-16-32] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-2.5-scalers0-aggregators2-32-16] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-2.5-scalers0-aggregators2-32-32] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-2.5-scalers1-aggregators0-16-16] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-2.5-scalers1-aggregators0-16-32] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-2.5-scalers1-aggregators0-32-16] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-2.5-scalers1-aggregators0-32-32] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-2.5-scalers1-aggregators1-16-16] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-2.5-scalers1-aggregators1-16-32] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-2.5-scalers1-aggregators1-32-16] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-2.5-scalers1-aggregators1-32-32] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-2.5-scalers1-aggregators2-16-16] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-2.5-scalers1-aggregators2-16-32] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-2.5-scalers1-aggregators2-32-16] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-2.5-scalers1-aggregators2-32-32] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-7.4-scalers0-aggregators0-16-16] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-7.4-scalers0-aggregators0-16-32] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-7.4-scalers0-aggregators0-32-16] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-7.4-scalers0-aggregators0-32-32] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-7.4-scalers0-aggregators1-16-16] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-7.4-scalers0-aggregators1-16-32] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-7.4-scalers0-aggregators1-32-16] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-7.4-scalers0-aggregators1-32-32] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-7.4-scalers0-aggregators2-16-16] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-7.4-scalers0-aggregators2-16-32] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-7.4-scalers0-aggregators2-32-16] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-7.4-scalers0-aggregators2-32-32] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-7.4-scalers1-aggregators0-16-16] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-7.4-scalers1-aggregators0-16-32] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-7.4-scalers1-aggregators0-32-16] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-7.4-scalers1-aggregators0-32-32] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-7.4-scalers1-aggregators1-16-16] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-7.4-scalers1-aggregators1-16-32] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-7.4-scalers1-aggregators1-32-16] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-7.4-scalers1-aggregators1-32-32] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-7.4-scalers1-aggregators2-16-16] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-7.4-scalers1-aggregators2-16-32] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-7.4-scalers1-aggregators2-32-16] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-7.4-scalers1-aggregators2-32-32] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-2.5-scalers0-aggregators0-16-16] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-2.5-scalers0-aggregators0-16-32] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-2.5-scalers0-aggregators0-32-16] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-2.5-scalers0-aggregators0-32-32] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-2.5-scalers0-aggregators1-16-16] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-2.5-scalers0-aggregators1-16-32] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-2.5-scalers0-aggregators1-32-16] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-2.5-scalers0-aggregators1-32-32] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-2.5-scalers0-aggregators2-16-16] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-2.5-scalers0-aggregators2-16-32] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-2.5-scalers0-aggregators2-32-16] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-2.5-scalers0-aggregators2-32-32] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-2.5-scalers1-aggregators0-16-16] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-2.5-scalers1-aggregators0-16-32] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-2.5-scalers1-aggregators0-32-16] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-2.5-scalers1-aggregators0-32-32] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-2.5-scalers1-aggregators1-16-16] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-2.5-scalers1-aggregators1-16-32] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-2.5-scalers1-aggregators1-32-16] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-2.5-scalers1-aggregators1-32-32] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-2.5-scalers1-aggregators2-16-16] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-2.5-scalers1-aggregators2-16-32] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-2.5-scalers1-aggregators2-32-16] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-2.5-scalers1-aggregators2-32-32] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-7.4-scalers0-aggregators0-16-16] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-7.4-scalers0-aggregators0-16-32] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-7.4-scalers0-aggregators0-32-16] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-7.4-scalers0-aggregators0-32-32] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-7.4-scalers0-aggregators1-16-16] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-7.4-scalers0-aggregators1-16-32] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-7.4-scalers0-aggregators1-32-16] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-7.4-scalers0-aggregators1-32-32] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-7.4-scalers0-aggregators2-16-16] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-7.4-scalers0-aggregators2-16-32] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-7.4-scalers0-aggregators2-32-16] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-7.4-scalers0-aggregators2-32-32] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-7.4-scalers1-aggregators0-16-16] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-7.4-scalers1-aggregators0-16-32] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-7.4-scalers1-aggregators0-32-16] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-7.4-scalers1-aggregators0-32-32] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-7.4-scalers1-aggregators1-16-16] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-7.4-scalers1-aggregators1-16-32] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-7.4-scalers1-aggregators1-32-16] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-7.4-scalers1-aggregators1-32-32] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-7.4-scalers1-aggregators2-16-16] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-7.4-scalers1-aggregators2-16-32] PASSED [ 54%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-7.4-scalers1-aggregators2-32-16] PASSED [ 54%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-7.4-scalers1-aggregators2-32-32] PASSED [ 54%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-2.5-scalers0-aggregators0-16-16] PASSED [ 54%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-2.5-scalers0-aggregators0-16-32] PASSED [ 54%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-2.5-scalers0-aggregators0-32-16] PASSED [ 54%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-2.5-scalers0-aggregators0-32-32] PASSED [ 54%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-2.5-scalers0-aggregators1-16-16] PASSED [ 54%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-2.5-scalers0-aggregators1-16-32] PASSED [ 54%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-2.5-scalers0-aggregators1-32-16] PASSED [ 54%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-2.5-scalers0-aggregators1-32-32] PASSED [ 54%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-2.5-scalers0-aggregators2-16-16] PASSED [ 54%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-2.5-scalers0-aggregators2-16-32] PASSED [ 54%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-2.5-scalers0-aggregators2-32-16] PASSED [ 54%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-2.5-scalers0-aggregators2-32-32] PASSED [ 54%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-2.5-scalers1-aggregators0-16-16] PASSED [ 54%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-2.5-scalers1-aggregators0-16-32] PASSED [ 54%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-2.5-scalers1-aggregators0-32-16] PASSED [ 54%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-2.5-scalers1-aggregators0-32-32] PASSED [ 54%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-2.5-scalers1-aggregators1-16-16] PASSED [ 54%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-2.5-scalers1-aggregators1-16-32] PASSED [ 54%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-2.5-scalers1-aggregators1-32-16] PASSED [ 54%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-2.5-scalers1-aggregators1-32-32] PASSED [ 54%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-2.5-scalers1-aggregators2-16-16] PASSED [ 54%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-2.5-scalers1-aggregators2-16-32] PASSED [ 54%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-2.5-scalers1-aggregators2-32-16] PASSED [ 54%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-2.5-scalers1-aggregators2-32-32] PASSED [ 54%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-7.4-scalers0-aggregators0-16-16] PASSED [ 54%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-7.4-scalers0-aggregators0-16-32] PASSED [ 54%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-7.4-scalers0-aggregators0-32-16] PASSED [ 54%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-7.4-scalers0-aggregators0-32-32] PASSED [ 54%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-7.4-scalers0-aggregators1-16-16] PASSED [ 54%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-7.4-scalers0-aggregators1-16-32] PASSED [ 54%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-7.4-scalers0-aggregators1-32-16] PASSED [ 54%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-7.4-scalers0-aggregators1-32-32] PASSED [ 54%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-7.4-scalers0-aggregators2-16-16] PASSED [ 54%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-7.4-scalers0-aggregators2-16-32] PASSED [ 54%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-7.4-scalers0-aggregators2-32-16] PASSED [ 54%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-7.4-scalers0-aggregators2-32-32] PASSED [ 54%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-7.4-scalers1-aggregators0-16-16] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-7.4-scalers1-aggregators0-16-32] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-7.4-scalers1-aggregators0-32-16] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-7.4-scalers1-aggregators0-32-32] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-7.4-scalers1-aggregators1-16-16] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-7.4-scalers1-aggregators1-16-32] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-7.4-scalers1-aggregators1-32-16] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-7.4-scalers1-aggregators1-32-32] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-7.4-scalers1-aggregators2-16-16] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-7.4-scalers1-aggregators2-16-32] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-7.4-scalers1-aggregators2-32-16] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-7.4-scalers1-aggregators2-32-32] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-2.5-scalers0-aggregators0-16-16] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-2.5-scalers0-aggregators0-16-32] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-2.5-scalers0-aggregators0-32-16] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-2.5-scalers0-aggregators0-32-32] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-2.5-scalers0-aggregators1-16-16] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-2.5-scalers0-aggregators1-16-32] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-2.5-scalers0-aggregators1-32-16] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-2.5-scalers0-aggregators1-32-32] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-2.5-scalers0-aggregators2-16-16] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-2.5-scalers0-aggregators2-16-32] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-2.5-scalers0-aggregators2-32-16] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-2.5-scalers0-aggregators2-32-32] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-2.5-scalers1-aggregators0-16-16] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-2.5-scalers1-aggregators0-16-32] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-2.5-scalers1-aggregators0-32-16] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-2.5-scalers1-aggregators0-32-32] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-2.5-scalers1-aggregators1-16-16] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-2.5-scalers1-aggregators1-16-32] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-2.5-scalers1-aggregators1-32-16] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-2.5-scalers1-aggregators1-32-32] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-2.5-scalers1-aggregators2-16-16] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-2.5-scalers1-aggregators2-16-32] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-2.5-scalers1-aggregators2-32-16] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-2.5-scalers1-aggregators2-32-32] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-7.4-scalers0-aggregators0-16-16] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-7.4-scalers0-aggregators0-16-32] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-7.4-scalers0-aggregators0-32-16] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-7.4-scalers0-aggregators0-32-32] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-7.4-scalers0-aggregators1-16-16] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-7.4-scalers0-aggregators1-16-32] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-7.4-scalers0-aggregators1-32-16] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-7.4-scalers0-aggregators1-32-32] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-7.4-scalers0-aggregators2-16-16] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-7.4-scalers0-aggregators2-16-32] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-7.4-scalers0-aggregators2-32-16] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-7.4-scalers0-aggregators2-32-32] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-7.4-scalers1-aggregators0-16-16] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-7.4-scalers1-aggregators0-16-32] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-7.4-scalers1-aggregators0-32-16] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-7.4-scalers1-aggregators0-32-32] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-7.4-scalers1-aggregators1-16-16] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-7.4-scalers1-aggregators1-16-32] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-7.4-scalers1-aggregators1-32-16] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-7.4-scalers1-aggregators1-32-32] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-7.4-scalers1-aggregators2-16-16] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-7.4-scalers1-aggregators2-16-32] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-7.4-scalers1-aggregators2-32-16] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-7.4-scalers1-aggregators2-32-32] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-2.5-scalers0-aggregators0-16-16] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-2.5-scalers0-aggregators0-16-32] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-2.5-scalers0-aggregators0-32-16] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-2.5-scalers0-aggregators0-32-32] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-2.5-scalers0-aggregators1-16-16] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-2.5-scalers0-aggregators1-16-32] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-2.5-scalers0-aggregators1-32-16] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-2.5-scalers0-aggregators1-32-32] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-2.5-scalers0-aggregators2-16-16] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-2.5-scalers0-aggregators2-16-32] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-2.5-scalers0-aggregators2-32-16] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-2.5-scalers0-aggregators2-32-32] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-2.5-scalers1-aggregators0-16-16] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-2.5-scalers1-aggregators0-16-32] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-2.5-scalers1-aggregators0-32-16] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-2.5-scalers1-aggregators0-32-32] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-2.5-scalers1-aggregators1-16-16] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-2.5-scalers1-aggregators1-16-32] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-2.5-scalers1-aggregators1-32-16] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-2.5-scalers1-aggregators1-32-32] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-2.5-scalers1-aggregators2-16-16] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-2.5-scalers1-aggregators2-16-32] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-2.5-scalers1-aggregators2-32-16] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-2.5-scalers1-aggregators2-32-32] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-7.4-scalers0-aggregators0-16-16] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-7.4-scalers0-aggregators0-16-32] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-7.4-scalers0-aggregators0-32-16] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-7.4-scalers0-aggregators0-32-32] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-7.4-scalers0-aggregators1-16-16] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-7.4-scalers0-aggregators1-16-32] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-7.4-scalers0-aggregators1-32-16] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-7.4-scalers0-aggregators1-32-32] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-7.4-scalers0-aggregators2-16-16] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-7.4-scalers0-aggregators2-16-32] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-7.4-scalers0-aggregators2-32-16] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-7.4-scalers0-aggregators2-32-32] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-7.4-scalers1-aggregators0-16-16] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-7.4-scalers1-aggregators0-16-32] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-7.4-scalers1-aggregators0-32-16] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-7.4-scalers1-aggregators0-32-32] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-7.4-scalers1-aggregators1-16-16] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-7.4-scalers1-aggregators1-16-32] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-7.4-scalers1-aggregators1-32-16] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-7.4-scalers1-aggregators1-32-32] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-7.4-scalers1-aggregators2-16-16] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-7.4-scalers1-aggregators2-16-32] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-7.4-scalers1-aggregators2-32-16] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-7.4-scalers1-aggregators2-32-32] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-2.5-scalers0-aggregators0-16-16] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-2.5-scalers0-aggregators0-16-32] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-2.5-scalers0-aggregators0-32-16] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-2.5-scalers0-aggregators0-32-32] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-2.5-scalers0-aggregators1-16-16] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-2.5-scalers0-aggregators1-16-32] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-2.5-scalers0-aggregators1-32-16] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-2.5-scalers0-aggregators1-32-32] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-2.5-scalers0-aggregators2-16-16] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-2.5-scalers0-aggregators2-16-32] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-2.5-scalers0-aggregators2-32-16] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-2.5-scalers0-aggregators2-32-32] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-2.5-scalers1-aggregators0-16-16] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-2.5-scalers1-aggregators0-16-32] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-2.5-scalers1-aggregators0-32-16] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-2.5-scalers1-aggregators0-32-32] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-2.5-scalers1-aggregators1-16-16] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-2.5-scalers1-aggregators1-16-32] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-2.5-scalers1-aggregators1-32-16] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-2.5-scalers1-aggregators1-32-32] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-2.5-scalers1-aggregators2-16-16] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-2.5-scalers1-aggregators2-16-32] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-2.5-scalers1-aggregators2-32-16] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-2.5-scalers1-aggregators2-32-32] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-7.4-scalers0-aggregators0-16-16] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-7.4-scalers0-aggregators0-16-32] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-7.4-scalers0-aggregators0-32-16] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-7.4-scalers0-aggregators0-32-32] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-7.4-scalers0-aggregators1-16-16] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-7.4-scalers0-aggregators1-16-32] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-7.4-scalers0-aggregators1-32-16] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-7.4-scalers0-aggregators1-32-32] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-7.4-scalers0-aggregators2-16-16] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-7.4-scalers0-aggregators2-16-32] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-7.4-scalers0-aggregators2-32-16] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-7.4-scalers0-aggregators2-32-32] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-7.4-scalers1-aggregators0-16-16] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-7.4-scalers1-aggregators0-16-32] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-7.4-scalers1-aggregators0-32-16] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-7.4-scalers1-aggregators0-32-32] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-7.4-scalers1-aggregators1-16-16] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-7.4-scalers1-aggregators1-16-32] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-7.4-scalers1-aggregators1-32-16] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-7.4-scalers1-aggregators1-32-32] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-7.4-scalers1-aggregators2-16-16] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-7.4-scalers1-aggregators2-16-32] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-7.4-scalers1-aggregators2-32-16] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-7.4-scalers1-aggregators2-32-32] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-2.5-scalers0-aggregators0-16-16] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-2.5-scalers0-aggregators0-16-32] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-2.5-scalers0-aggregators0-32-16] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-2.5-scalers0-aggregators0-32-32] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-2.5-scalers0-aggregators1-16-16] PASSED [ 59%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-2.5-scalers0-aggregators1-16-32] PASSED [ 59%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-2.5-scalers0-aggregators1-32-16] PASSED [ 59%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-2.5-scalers0-aggregators1-32-32] PASSED [ 59%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-2.5-scalers0-aggregators2-16-16] PASSED [ 59%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-2.5-scalers0-aggregators2-16-32] PASSED [ 59%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-2.5-scalers0-aggregators2-32-16] PASSED [ 59%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-2.5-scalers0-aggregators2-32-32] PASSED [ 59%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-2.5-scalers1-aggregators0-16-16] PASSED [ 59%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-2.5-scalers1-aggregators0-16-32] PASSED [ 59%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-2.5-scalers1-aggregators0-32-16] PASSED [ 59%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-2.5-scalers1-aggregators0-32-32] PASSED [ 59%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-2.5-scalers1-aggregators1-16-16] PASSED [ 59%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-2.5-scalers1-aggregators1-16-32] PASSED [ 59%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-2.5-scalers1-aggregators1-32-16] PASSED [ 59%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-2.5-scalers1-aggregators1-32-32] PASSED [ 59%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-2.5-scalers1-aggregators2-16-16] PASSED [ 59%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-2.5-scalers1-aggregators2-16-32] PASSED [ 59%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-2.5-scalers1-aggregators2-32-16] PASSED [ 59%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-2.5-scalers1-aggregators2-32-32] PASSED [ 59%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-7.4-scalers0-aggregators0-16-16] PASSED [ 59%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-7.4-scalers0-aggregators0-16-32] PASSED [ 59%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-7.4-scalers0-aggregators0-32-16] PASSED [ 59%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-7.4-scalers0-aggregators0-32-32] PASSED [ 59%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-7.4-scalers0-aggregators1-16-16] PASSED [ 59%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-7.4-scalers0-aggregators1-16-32] PASSED [ 59%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-7.4-scalers0-aggregators1-32-16] PASSED [ 59%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-7.4-scalers0-aggregators1-32-32] PASSED [ 59%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-7.4-scalers0-aggregators2-16-16] PASSED [ 59%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-7.4-scalers0-aggregators2-16-32] PASSED [ 59%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-7.4-scalers0-aggregators2-32-16] PASSED [ 59%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-7.4-scalers0-aggregators2-32-32] PASSED [ 59%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-7.4-scalers1-aggregators0-16-16] PASSED [ 59%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-7.4-scalers1-aggregators0-16-32] PASSED [ 59%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-7.4-scalers1-aggregators0-32-16] PASSED [ 59%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-7.4-scalers1-aggregators0-32-32] PASSED [ 59%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-7.4-scalers1-aggregators1-16-16] PASSED [ 59%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-7.4-scalers1-aggregators1-16-32] PASSED [ 59%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-7.4-scalers1-aggregators1-32-16] PASSED [ 59%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-7.4-scalers1-aggregators1-32-32] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-7.4-scalers1-aggregators2-16-16] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-7.4-scalers1-aggregators2-16-32] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-7.4-scalers1-aggregators2-32-16] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-7.4-scalers1-aggregators2-32-32] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-2.5-scalers0-aggregators0-16-16] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-2.5-scalers0-aggregators0-16-32] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-2.5-scalers0-aggregators0-32-16] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-2.5-scalers0-aggregators0-32-32] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-2.5-scalers0-aggregators1-16-16] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-2.5-scalers0-aggregators1-16-32] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-2.5-scalers0-aggregators1-32-16] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-2.5-scalers0-aggregators1-32-32] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-2.5-scalers0-aggregators2-16-16] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-2.5-scalers0-aggregators2-16-32] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-2.5-scalers0-aggregators2-32-16] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-2.5-scalers0-aggregators2-32-32] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-2.5-scalers1-aggregators0-16-16] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-2.5-scalers1-aggregators0-16-32] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-2.5-scalers1-aggregators0-32-16] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-2.5-scalers1-aggregators0-32-32] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-2.5-scalers1-aggregators1-16-16] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-2.5-scalers1-aggregators1-16-32] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-2.5-scalers1-aggregators1-32-16] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-2.5-scalers1-aggregators1-32-32] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-2.5-scalers1-aggregators2-16-16] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-2.5-scalers1-aggregators2-16-32] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-2.5-scalers1-aggregators2-32-16] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-2.5-scalers1-aggregators2-32-32] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-7.4-scalers0-aggregators0-16-16] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-7.4-scalers0-aggregators0-16-32] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-7.4-scalers0-aggregators0-32-16] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-7.4-scalers0-aggregators0-32-32] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-7.4-scalers0-aggregators1-16-16] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-7.4-scalers0-aggregators1-16-32] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-7.4-scalers0-aggregators1-32-16] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-7.4-scalers0-aggregators1-32-32] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-7.4-scalers0-aggregators2-16-16] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-7.4-scalers0-aggregators2-16-32] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-7.4-scalers0-aggregators2-32-16] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-7.4-scalers0-aggregators2-32-32] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-7.4-scalers1-aggregators0-16-16] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-7.4-scalers1-aggregators0-16-32] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-7.4-scalers1-aggregators0-32-16] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-7.4-scalers1-aggregators0-32-32] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-7.4-scalers1-aggregators1-16-16] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-7.4-scalers1-aggregators1-16-32] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-7.4-scalers1-aggregators1-32-16] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-7.4-scalers1-aggregators1-32-32] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-7.4-scalers1-aggregators2-16-16] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-7.4-scalers1-aggregators2-16-32] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-7.4-scalers1-aggregators2-32-16] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-7.4-scalers1-aggregators2-32-32] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-2.5-scalers0-aggregators0-16-16] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-2.5-scalers0-aggregators0-16-32] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-2.5-scalers0-aggregators0-32-16] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-2.5-scalers0-aggregators0-32-32] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-2.5-scalers0-aggregators1-16-16] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-2.5-scalers0-aggregators1-16-32] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-2.5-scalers0-aggregators1-32-16] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-2.5-scalers0-aggregators1-32-32] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-2.5-scalers0-aggregators2-16-16] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-2.5-scalers0-aggregators2-16-32] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-2.5-scalers0-aggregators2-32-16] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-2.5-scalers0-aggregators2-32-32] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-2.5-scalers1-aggregators0-16-16] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-2.5-scalers1-aggregators0-16-32] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-2.5-scalers1-aggregators0-32-16] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-2.5-scalers1-aggregators0-32-32] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-2.5-scalers1-aggregators1-16-16] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-2.5-scalers1-aggregators1-16-32] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-2.5-scalers1-aggregators1-32-16] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-2.5-scalers1-aggregators1-32-32] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-2.5-scalers1-aggregators2-16-16] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-2.5-scalers1-aggregators2-16-32] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-2.5-scalers1-aggregators2-32-16] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-2.5-scalers1-aggregators2-32-32] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-7.4-scalers0-aggregators0-16-16] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-7.4-scalers0-aggregators0-16-32] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-7.4-scalers0-aggregators0-32-16] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-7.4-scalers0-aggregators0-32-32] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-7.4-scalers0-aggregators1-16-16] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-7.4-scalers0-aggregators1-16-32] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-7.4-scalers0-aggregators1-32-16] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-7.4-scalers0-aggregators1-32-32] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-7.4-scalers0-aggregators2-16-16] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-7.4-scalers0-aggregators2-16-32] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-7.4-scalers0-aggregators2-32-16] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-7.4-scalers0-aggregators2-32-32] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-7.4-scalers1-aggregators0-16-16] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-7.4-scalers1-aggregators0-16-32] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-7.4-scalers1-aggregators0-32-16] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-7.4-scalers1-aggregators0-32-32] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-7.4-scalers1-aggregators1-16-16] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-7.4-scalers1-aggregators1-16-32] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-7.4-scalers1-aggregators1-32-16] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-7.4-scalers1-aggregators1-32-32] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-7.4-scalers1-aggregators2-16-16] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-7.4-scalers1-aggregators2-16-32] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-7.4-scalers1-aggregators2-32-16] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-7.4-scalers1-aggregators2-32-32] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-2.5-scalers0-aggregators0-16-16] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-2.5-scalers0-aggregators0-16-32] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-2.5-scalers0-aggregators0-32-16] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-2.5-scalers0-aggregators0-32-32] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-2.5-scalers0-aggregators1-16-16] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-2.5-scalers0-aggregators1-16-32] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-2.5-scalers0-aggregators1-32-16] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-2.5-scalers0-aggregators1-32-32] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-2.5-scalers0-aggregators2-16-16] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-2.5-scalers0-aggregators2-16-32] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-2.5-scalers0-aggregators2-32-16] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-2.5-scalers0-aggregators2-32-32] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-2.5-scalers1-aggregators0-16-16] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-2.5-scalers1-aggregators0-16-32] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-2.5-scalers1-aggregators0-32-16] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-2.5-scalers1-aggregators0-32-32] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-2.5-scalers1-aggregators1-16-16] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-2.5-scalers1-aggregators1-16-32] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-2.5-scalers1-aggregators1-32-16] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-2.5-scalers1-aggregators1-32-32] PASSED [ 63%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-2.5-scalers1-aggregators2-16-16] PASSED [ 63%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-2.5-scalers1-aggregators2-16-32] PASSED [ 63%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-2.5-scalers1-aggregators2-32-16] PASSED [ 63%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-2.5-scalers1-aggregators2-32-32] PASSED [ 63%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-7.4-scalers0-aggregators0-16-16] PASSED [ 63%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-7.4-scalers0-aggregators0-16-32] PASSED [ 63%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-7.4-scalers0-aggregators0-32-16] PASSED [ 63%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-7.4-scalers0-aggregators0-32-32] PASSED [ 63%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-7.4-scalers0-aggregators1-16-16] PASSED [ 63%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-7.4-scalers0-aggregators1-16-32] PASSED [ 63%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-7.4-scalers0-aggregators1-32-16] PASSED [ 63%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-7.4-scalers0-aggregators1-32-32] PASSED [ 63%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-7.4-scalers0-aggregators2-16-16] PASSED [ 63%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-7.4-scalers0-aggregators2-16-32] PASSED [ 63%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-7.4-scalers0-aggregators2-32-16] PASSED [ 63%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-7.4-scalers0-aggregators2-32-32] PASSED [ 63%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-7.4-scalers1-aggregators0-16-16] PASSED [ 63%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-7.4-scalers1-aggregators0-16-32] PASSED [ 63%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-7.4-scalers1-aggregators0-32-16] PASSED [ 63%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-7.4-scalers1-aggregators0-32-32] PASSED [ 63%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-7.4-scalers1-aggregators1-16-16] PASSED [ 63%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-7.4-scalers1-aggregators1-16-32] PASSED [ 63%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-7.4-scalers1-aggregators1-32-16] PASSED [ 63%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-7.4-scalers1-aggregators1-32-32] PASSED [ 63%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-7.4-scalers1-aggregators2-16-16] PASSED [ 63%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-7.4-scalers1-aggregators2-16-32] PASSED [ 63%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-7.4-scalers1-aggregators2-32-16] PASSED [ 63%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-7.4-scalers1-aggregators2-32-32] PASSED [ 63%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-2.5-scalers0-aggregators0-16-16] PASSED [ 63%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-2.5-scalers0-aggregators0-16-32] PASSED [ 63%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-2.5-scalers0-aggregators0-32-16] PASSED [ 63%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-2.5-scalers0-aggregators0-32-32] PASSED [ 63%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-2.5-scalers0-aggregators1-16-16] PASSED [ 63%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-2.5-scalers0-aggregators1-16-32] PASSED [ 63%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-2.5-scalers0-aggregators1-32-16] PASSED [ 63%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-2.5-scalers0-aggregators1-32-32] PASSED [ 63%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-2.5-scalers0-aggregators2-16-16] PASSED [ 63%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-2.5-scalers0-aggregators2-16-32] PASSED [ 63%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-2.5-scalers0-aggregators2-32-16] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-2.5-scalers0-aggregators2-32-32] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-2.5-scalers1-aggregators0-16-16] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-2.5-scalers1-aggregators0-16-32] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-2.5-scalers1-aggregators0-32-16] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-2.5-scalers1-aggregators0-32-32] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-2.5-scalers1-aggregators1-16-16] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-2.5-scalers1-aggregators1-16-32] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-2.5-scalers1-aggregators1-32-16] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-2.5-scalers1-aggregators1-32-32] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-2.5-scalers1-aggregators2-16-16] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-2.5-scalers1-aggregators2-16-32] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-2.5-scalers1-aggregators2-32-16] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-2.5-scalers1-aggregators2-32-32] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-7.4-scalers0-aggregators0-16-16] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-7.4-scalers0-aggregators0-16-32] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-7.4-scalers0-aggregators0-32-16] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-7.4-scalers0-aggregators0-32-32] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-7.4-scalers0-aggregators1-16-16] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-7.4-scalers0-aggregators1-16-32] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-7.4-scalers0-aggregators1-32-16] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-7.4-scalers0-aggregators1-32-32] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-7.4-scalers0-aggregators2-16-16] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-7.4-scalers0-aggregators2-16-32] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-7.4-scalers0-aggregators2-32-16] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-7.4-scalers0-aggregators2-32-32] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-7.4-scalers1-aggregators0-16-16] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-7.4-scalers1-aggregators0-16-32] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-7.4-scalers1-aggregators0-32-16] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-7.4-scalers1-aggregators0-32-32] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-7.4-scalers1-aggregators1-16-16] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-7.4-scalers1-aggregators1-16-32] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-7.4-scalers1-aggregators1-32-16] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-7.4-scalers1-aggregators1-32-32] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-7.4-scalers1-aggregators2-16-16] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-7.4-scalers1-aggregators2-16-32] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-7.4-scalers1-aggregators2-32-16] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-7.4-scalers1-aggregators2-32-32] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-2.5-scalers0-aggregators0-16-16] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-2.5-scalers0-aggregators0-16-32] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-2.5-scalers0-aggregators0-32-16] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-2.5-scalers0-aggregators0-32-32] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-2.5-scalers0-aggregators1-16-16] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-2.5-scalers0-aggregators1-16-32] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-2.5-scalers0-aggregators1-32-16] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-2.5-scalers0-aggregators1-32-32] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-2.5-scalers0-aggregators2-16-16] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-2.5-scalers0-aggregators2-16-32] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-2.5-scalers0-aggregators2-32-16] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-2.5-scalers0-aggregators2-32-32] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-2.5-scalers1-aggregators0-16-16] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-2.5-scalers1-aggregators0-16-32] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-2.5-scalers1-aggregators0-32-16] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-2.5-scalers1-aggregators0-32-32] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-2.5-scalers1-aggregators1-16-16] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-2.5-scalers1-aggregators1-16-32] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-2.5-scalers1-aggregators1-32-16] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-2.5-scalers1-aggregators1-32-32] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-2.5-scalers1-aggregators2-16-16] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-2.5-scalers1-aggregators2-16-32] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-2.5-scalers1-aggregators2-32-16] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-2.5-scalers1-aggregators2-32-32] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-7.4-scalers0-aggregators0-16-16] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-7.4-scalers0-aggregators0-16-32] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-7.4-scalers0-aggregators0-32-16] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-7.4-scalers0-aggregators0-32-32] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-7.4-scalers0-aggregators1-16-16] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-7.4-scalers0-aggregators1-16-32] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-7.4-scalers0-aggregators1-32-16] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-7.4-scalers0-aggregators1-32-32] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-7.4-scalers0-aggregators2-16-16] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-7.4-scalers0-aggregators2-16-32] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-7.4-scalers0-aggregators2-32-16] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-7.4-scalers0-aggregators2-32-32] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-7.4-scalers1-aggregators0-16-16] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-7.4-scalers1-aggregators0-16-32] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-7.4-scalers1-aggregators0-32-16] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-7.4-scalers1-aggregators0-32-32] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-7.4-scalers1-aggregators1-16-16] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-7.4-scalers1-aggregators1-16-32] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-7.4-scalers1-aggregators1-32-16] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-7.4-scalers1-aggregators1-32-32] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-7.4-scalers1-aggregators2-16-16] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-7.4-scalers1-aggregators2-16-32] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-7.4-scalers1-aggregators2-32-16] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-7.4-scalers1-aggregators2-32-32] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-2.5-scalers0-aggregators0-16-16] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-2.5-scalers0-aggregators0-16-32] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-2.5-scalers0-aggregators0-32-16] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-2.5-scalers0-aggregators0-32-32] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-2.5-scalers0-aggregators1-16-16] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-2.5-scalers0-aggregators1-16-32] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-2.5-scalers0-aggregators1-32-16] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-2.5-scalers0-aggregators1-32-32] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-2.5-scalers0-aggregators2-16-16] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-2.5-scalers0-aggregators2-16-32] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-2.5-scalers0-aggregators2-32-16] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-2.5-scalers0-aggregators2-32-32] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-2.5-scalers1-aggregators0-16-16] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-2.5-scalers1-aggregators0-16-32] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-2.5-scalers1-aggregators0-32-16] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-2.5-scalers1-aggregators0-32-32] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-2.5-scalers1-aggregators1-16-16] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-2.5-scalers1-aggregators1-16-32] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-2.5-scalers1-aggregators1-32-16] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-2.5-scalers1-aggregators1-32-32] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-2.5-scalers1-aggregators2-16-16] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-2.5-scalers1-aggregators2-16-32] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-2.5-scalers1-aggregators2-32-16] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-2.5-scalers1-aggregators2-32-32] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-7.4-scalers0-aggregators0-16-16] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-7.4-scalers0-aggregators0-16-32] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-7.4-scalers0-aggregators0-32-16] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-7.4-scalers0-aggregators0-32-32] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-7.4-scalers0-aggregators1-16-16] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-7.4-scalers0-aggregators1-16-32] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-7.4-scalers0-aggregators1-32-16] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-7.4-scalers0-aggregators1-32-32] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-7.4-scalers0-aggregators2-16-16] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-7.4-scalers0-aggregators2-16-32] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-7.4-scalers0-aggregators2-32-16] PASSED [ 67%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-7.4-scalers0-aggregators2-32-32] PASSED [ 67%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-7.4-scalers1-aggregators0-16-16] PASSED [ 67%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-7.4-scalers1-aggregators0-16-32] PASSED [ 67%]

tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-7.4-scalers1-aggregators0-32-32] PASSED [ 67%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-7.4-scalers1-aggregators1-16-16] PASSED [ 67%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-7.4-scalers1-aggregators1-16-32] PASSED [ 67%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-7.4-scalers1-aggregators1-32-16] PASSED [ 67%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-7.4-scalers1-aggregators1-32-32] PASSED [ 67%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-7.4-scalers1-aggregators2-16-16] PASSED [ 67%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-7.4-scalers1-aggregators2-16-32] PASSED [ 67%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-7.4-scalers1-aggregators2-32-16] PASSED [ 67%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-7.4-scalers1-aggregators2-32-32] PASSED [ 67%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-2.5-scalers0-aggregators0-16-16] PASSED [ 67%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-2.5-scalers0-aggregators0-16-32] PASSED [ 67%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-2.5-scalers0-aggregators0-32-16] PASSED [ 67%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-2.5-scalers0-aggregators0-32-32] PASSED [ 67%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-2.5-scalers0-aggregators1-16-16] PASSED [ 67%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-2.5-scalers0-aggregators1-16-32] PASSED [ 67%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-2.5-scalers0-aggregators1-32-16] PASSED [ 67%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-2.5-scalers0-aggregators1-32-32] PASSED [ 67%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-2.5-scalers0-aggregators2-16-16] PASSED [ 67%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-2.5-scalers0-aggregators2-16-32] PASSED [ 67%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-2.5-scalers0-aggregators2-32-16] PASSED [ 67%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-2.5-scalers0-aggregators2-32-32] PASSED [ 67%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-2.5-scalers1-aggregators0-16-16] PASSED [ 67%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-2.5-scalers1-aggregators0-16-32] PASSED [ 67%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-2.5-scalers1-aggregators0-32-16] PASSED [ 67%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-2.5-scalers1-aggregators0-32-32] PASSED [ 67%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-2.5-scalers1-aggregators1-16-16] PASSED [ 67%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-2.5-scalers1-aggregators1-16-32] PASSED [ 67%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-2.5-scalers1-aggregators1-32-16] PASSED [ 67%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-2.5-scalers1-aggregators1-32-32] PASSED [ 67%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-2.5-scalers1-aggregators2-16-16] PASSED [ 67%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-2.5-scalers1-aggregators2-16-32] PASSED [ 67%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-2.5-scalers1-aggregators2-32-16] PASSED [ 67%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-2.5-scalers1-aggregators2-32-32] PASSED [ 67%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-7.4-scalers0-aggregators0-16-16] PASSED [ 67%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-7.4-scalers0-aggregators0-16-32] PASSED [ 67%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-7.4-scalers0-aggregators0-32-16] PASSED [ 68%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-7.4-scalers0-aggregators0-32-32] PASSED [ 68%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-7.4-scalers0-aggregators1-16-16] PASSED [ 68%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-7.4-scalers0-aggregators1-16-32] PASSED [ 68%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-7.4-scalers0-aggregators1-32-16] PASSED [ 68%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-7.4-scalers0-aggregators1-32-32] PASSED [ 68%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-7.4-scalers0-aggregators2-16-16] PASSED [ 68%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-7.4-scalers0-aggregators2-16-32] PASSED [ 68%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-7.4-scalers0-aggregators2-32-16] PASSED [ 68%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-7.4-scalers0-aggregators2-32-32] PASSED [ 68%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-7.4-scalers1-aggregators0-16-16] PASSED [ 68%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-7.4-scalers1-aggregators0-16-32] PASSED [ 68%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-7.4-scalers1-aggregators0-32-16] PASSED [ 68%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-7.4-scalers1-aggregators0-32-32] PASSED [ 68%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-7.4-scalers1-aggregators1-16-16] PASSED [ 68%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-7.4-scalers1-aggregators1-16-32] PASSED [ 68%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-7.4-scalers1-aggregators1-32-16] PASSED [ 68%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-7.4-scalers1-aggregators1-32-32] PASSED [ 68%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-7.4-scalers1-aggregators2-16-16] PASSED [ 68%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-7.4-scalers1-aggregators2-16-32] PASSED [ 68%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-7.4-scalers1-aggregators2-32-16] PASSED [ 68%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-7.4-scalers1-aggregators2-32-32] PASSED [ 68%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-2.5-scalers0-aggregators0-16-16] PASSED [ 68%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-2.5-scalers0-aggregators0-16-32] PASSED [ 68%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-2.5-scalers0-aggregators0-32-16] PASSED [ 68%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-2.5-scalers0-aggregators0-32-32] PASSED [ 68%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-2.5-scalers0-aggregators1-16-16] PASSED [ 68%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-2.5-scalers0-aggregators1-16-32] PASSED [ 68%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-2.5-scalers0-aggregators1-32-16] PASSED [ 68%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-2.5-scalers0-aggregators1-32-32] PASSED [ 68%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-2.5-scalers0-aggregators2-16-16] PASSED [ 68%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-2.5-scalers0-aggregators2-16-32] PASSED [ 68%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-2.5-scalers0-aggregators2-32-16] PASSED [ 68%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-2.5-scalers0-aggregators2-32-32] PASSED [ 68%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-2.5-scalers1-aggregators0-16-16] PASSED [ 68%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-2.5-scalers1-aggregators0-16-32] PASSED [ 68%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-2.5-scalers1-aggregators0-32-16] PASSED [ 68%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-2.5-scalers1-aggregators0-32-32] PASSED [ 68%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-2.5-scalers1-aggregators1-16-16] PASSED [ 68%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-2.5-scalers1-aggregators1-16-32] PASSED [ 69%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-2.5-scalers1-aggregators1-32-16] PASSED [ 69%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-2.5-scalers1-aggregators1-32-32] PASSED [ 69%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-2.5-scalers1-aggregators2-16-16] PASSED [ 69%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-2.5-scalers1-aggregators2-16-32] PASSED [ 69%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-2.5-scalers1-aggregators2-32-16] PASSED [ 69%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-2.5-scalers1-aggregators2-32-32] PASSED [ 69%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-7.4-scalers0-aggregators0-16-16] PASSED [ 69%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-7.4-scalers0-aggregators0-16-32] PASSED [ 69%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-7.4-scalers0-aggregators0-32-16] PASSED [ 69%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-7.4-scalers0-aggregators0-32-32] PASSED [ 69%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-7.4-scalers0-aggregators1-16-16] PASSED [ 69%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-7.4-scalers0-aggregators1-16-32] PASSED [ 69%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-7.4-scalers0-aggregators1-32-16] PASSED [ 69%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-7.4-scalers0-aggregators1-32-32] PASSED [ 69%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-7.4-scalers0-aggregators2-16-16] PASSED [ 69%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-7.4-scalers0-aggregators2-16-32] PASSED [ 69%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-7.4-scalers0-aggregators2-32-16] PASSED [ 69%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-7.4-scalers0-aggregators2-32-32] PASSED [ 69%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-7.4-scalers1-aggregators0-16-16] PASSED [ 69%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-7.4-scalers1-aggregators0-16-32] PASSED [ 69%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-7.4-scalers1-aggregators0-32-16] PASSED [ 69%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-7.4-scalers1-aggregators0-32-32] PASSED [ 69%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-7.4-scalers1-aggregators1-16-16] PASSED [ 69%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-7.4-scalers1-aggregators1-16-32] PASSED [ 69%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-7.4-scalers1-aggregators1-32-16] PASSED [ 69%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-7.4-scalers1-aggregators1-32-32] PASSED [ 69%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-7.4-scalers1-aggregators2-16-16] PASSED [ 69%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-7.4-scalers1-aggregators2-16-32] PASSED [ 69%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-7.4-scalers1-aggregators2-32-16] PASSED [ 69%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-7.4-scalers1-aggregators2-32-32] PASSED [ 69%]
tests/pytorch/test_nn.py::test_label_prop[True-True-True-sym-0.0-3] FAILED [ 69%]
tests/pytorch/test_nn.py::test_label_prop[True-True-True-sym-0.0-5] FAILED [ 69%]
tests/pytorch/test_nn.py::test_label_prop[True-True-True-sym-0.5-3] FAILED [ 69%]
tests/pytorch/test_nn.py::test_label_prop[True-True-True-sym-0.5-5] FAILED [ 69%]
tests/pytorch/test_nn.py::test_label_prop[True-True-True-sym-1.0-3] FAILED [ 69%]
tests/pytorch/test_nn.py::test_label_prop[True-True-True-sym-1.0-5] FAILED [ 69%]
tests/pytorch/test_nn.py::test_label_prop[True-True-True-row-0.0-3] FAILED [ 69%]
tests/pytorch/test_nn.py::test_label_prop[True-True-True-row-0.0-5] FAILED [ 69%]
tests/pytorch/test_nn.py::test_label_prop[True-True-True-row-0.5-3] FAILED [ 69%]
tests/pytorch/test_nn.py::test_label_prop[True-True-True-row-0.5-5] FAILED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[True-True-True-row-1.0-3] FAILED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[True-True-True-row-1.0-5] FAILED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[True-True-False-sym-0.0-3] FAILED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[True-True-False-sym-0.0-5] FAILED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[True-True-False-sym-0.5-3] FAILED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[True-True-False-sym-0.5-5] FAILED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[True-True-False-sym-1.0-3] FAILED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[True-True-False-sym-1.0-5] FAILED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[True-True-False-row-0.0-3] FAILED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[True-True-False-row-0.0-5] FAILED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[True-True-False-row-0.5-3] FAILED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[True-True-False-row-0.5-5] FAILED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[True-True-False-row-1.0-3] FAILED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[True-True-False-row-1.0-5] FAILED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[True-False-True-sym-0.0-3] FAILED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[True-False-True-sym-0.0-5] FAILED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[True-False-True-sym-0.5-3] FAILED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[True-False-True-sym-0.5-5] FAILED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[True-False-True-sym-1.0-3] FAILED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[True-False-True-sym-1.0-5] FAILED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[True-False-True-row-0.0-3] FAILED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[True-False-True-row-0.0-5] FAILED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[True-False-True-row-0.5-3] FAILED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[True-False-True-row-0.5-5] FAILED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[True-False-True-row-1.0-3] FAILED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[True-False-True-row-1.0-5] FAILED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[True-False-False-sym-0.0-3] FAILED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[True-False-False-sym-0.0-5] FAILED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[True-False-False-sym-0.5-3] FAILED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[True-False-False-sym-0.5-5] FAILED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[True-False-False-sym-1.0-3] FAILED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[True-False-False-sym-1.0-5] FAILED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[True-False-False-row-0.0-3] FAILED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[True-False-False-row-0.0-5] FAILED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[True-False-False-row-0.5-3] FAILED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[True-False-False-row-0.5-5] FAILED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[True-False-False-row-1.0-3] FAILED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[True-False-False-row-1.0-5] FAILED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[False-True-True-sym-0.0-3] FAILED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[False-True-True-sym-0.0-5] FAILED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-True-True-sym-0.5-3] FAILED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-True-True-sym-0.5-5] FAILED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-True-True-sym-1.0-3] FAILED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-True-True-sym-1.0-5] FAILED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-True-True-row-0.0-3] FAILED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-True-True-row-0.0-5] FAILED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-True-True-row-0.5-3] FAILED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-True-True-row-0.5-5] FAILED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-True-True-row-1.0-3] FAILED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-True-True-row-1.0-5] FAILED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-True-False-sym-0.0-3] FAILED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-True-False-sym-0.0-5] FAILED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-True-False-sym-0.5-3] FAILED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-True-False-sym-0.5-5] FAILED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-True-False-sym-1.0-3] FAILED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-True-False-sym-1.0-5] FAILED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-True-False-row-0.0-3] FAILED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-True-False-row-0.0-5] FAILED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-True-False-row-0.5-3] FAILED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-True-False-row-0.5-5] FAILED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-True-False-row-1.0-3] FAILED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-True-False-row-1.0-5] FAILED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-False-True-sym-0.0-3] FAILED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-False-True-sym-0.0-5] FAILED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-False-True-sym-0.5-3] FAILED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-False-True-sym-0.5-5] FAILED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-False-True-sym-1.0-3] FAILED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-False-True-sym-1.0-5] FAILED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-False-True-row-0.0-3] FAILED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-False-True-row-0.0-5] FAILED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-False-True-row-0.5-3] FAILED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-False-True-row-0.5-5] FAILED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-False-True-row-1.0-3] FAILED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-False-True-row-1.0-5] FAILED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-False-False-sym-0.0-3] FAILED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-False-False-sym-0.0-5] FAILED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-False-False-sym-0.5-3] FAILED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-False-False-sym-0.5-5] FAILED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-False-False-sym-1.0-3] FAILED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-False-False-sym-1.0-5] FAILED [ 72%]
tests/pytorch/test_nn.py::test_label_prop[False-False-False-row-0.0-3] FAILED [ 72%]
tests/pytorch/test_nn.py::test_label_prop[False-False-False-row-0.0-5] FAILED [ 72%]
tests/pytorch/test_nn.py::test_label_prop[False-False-False-row-0.5-3] FAILED [ 72%]
tests/pytorch/test_nn.py::test_label_prop[False-False-False-row-0.5-5] FAILED [ 72%]
tests/pytorch/test_nn.py::test_label_prop[False-False-False-row-1.0-3] FAILED [ 72%]
tests/pytorch/test_nn.py::test_label_prop[False-False-False-row-1.0-5] FAILED [ 72%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-2.5-scalers0-aggregators0-16-16] PASSED [ 72%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-2.5-scalers0-aggregators0-16-32] PASSED [ 72%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-2.5-scalers0-aggregators0-32-16] PASSED [ 72%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-2.5-scalers0-aggregators0-32-32] PASSED [ 72%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-2.5-scalers0-aggregators1-16-16] PASSED [ 72%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-2.5-scalers0-aggregators1-16-32] PASSED [ 72%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-2.5-scalers0-aggregators1-32-16] PASSED [ 72%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-2.5-scalers0-aggregators1-32-32] PASSED [ 72%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-2.5-scalers0-aggregators2-16-16] PASSED [ 72%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-2.5-scalers0-aggregators2-16-32] PASSED [ 72%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-2.5-scalers0-aggregators2-32-16] PASSED [ 72%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-2.5-scalers0-aggregators2-32-32] PASSED [ 72%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-2.5-scalers1-aggregators0-16-16] PASSED [ 72%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-2.5-scalers1-aggregators0-16-32] PASSED [ 72%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-2.5-scalers1-aggregators0-32-16] PASSED [ 72%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-2.5-scalers1-aggregators0-32-32] PASSED [ 72%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-2.5-scalers1-aggregators1-16-16] PASSED [ 72%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-2.5-scalers1-aggregators1-16-32] PASSED [ 72%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-2.5-scalers1-aggregators1-32-16] PASSED [ 72%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-2.5-scalers1-aggregators1-32-32] PASSED [ 72%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-2.5-scalers1-aggregators2-16-16] PASSED [ 72%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-2.5-scalers1-aggregators2-16-32] PASSED [ 72%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-2.5-scalers1-aggregators2-32-16] PASSED [ 72%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-2.5-scalers1-aggregators2-32-32] PASSED [ 72%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-7.4-scalers0-aggregators0-16-16] PASSED [ 72%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-7.4-scalers0-aggregators0-16-32] PASSED [ 72%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-7.4-scalers0-aggregators0-32-16] PASSED [ 72%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-7.4-scalers0-aggregators0-32-32] PASSED [ 72%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-7.4-scalers0-aggregators1-16-16] PASSED [ 72%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-7.4-scalers0-aggregators1-16-32] PASSED [ 72%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-7.4-scalers0-aggregators1-32-16] PASSED [ 72%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-7.4-scalers0-aggregators1-32-32] PASSED [ 72%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-7.4-scalers0-aggregators2-16-16] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-7.4-scalers0-aggregators2-16-32] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-7.4-scalers0-aggregators2-32-16] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-7.4-scalers0-aggregators2-32-32] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-7.4-scalers1-aggregators0-16-16] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-7.4-scalers1-aggregators0-16-32] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-7.4-scalers1-aggregators0-32-16] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-7.4-scalers1-aggregators0-32-32] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-7.4-scalers1-aggregators1-16-16] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-7.4-scalers1-aggregators1-16-32] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-7.4-scalers1-aggregators1-32-16] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-7.4-scalers1-aggregators1-32-32] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-7.4-scalers1-aggregators2-16-16] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-7.4-scalers1-aggregators2-16-32] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-7.4-scalers1-aggregators2-32-16] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-7.4-scalers1-aggregators2-32-32] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-2.5-scalers0-aggregators0-16-16] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-2.5-scalers0-aggregators0-16-32] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-2.5-scalers0-aggregators0-32-16] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-2.5-scalers0-aggregators0-32-32] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-2.5-scalers0-aggregators1-16-16] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-2.5-scalers0-aggregators1-16-32] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-2.5-scalers0-aggregators1-32-16] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-2.5-scalers0-aggregators1-32-32] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-2.5-scalers0-aggregators2-16-16] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-2.5-scalers0-aggregators2-16-32] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-2.5-scalers0-aggregators2-32-16] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-2.5-scalers0-aggregators2-32-32] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-2.5-scalers1-aggregators0-16-16] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-2.5-scalers1-aggregators0-16-32] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-2.5-scalers1-aggregators0-32-16] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-2.5-scalers1-aggregators0-32-32] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-2.5-scalers1-aggregators1-16-16] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-2.5-scalers1-aggregators1-16-32] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-2.5-scalers1-aggregators1-32-16] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-2.5-scalers1-aggregators1-32-32] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-2.5-scalers1-aggregators2-16-16] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-2.5-scalers1-aggregators2-16-32] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-2.5-scalers1-aggregators2-32-16] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-2.5-scalers1-aggregators2-32-32] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-7.4-scalers0-aggregators0-16-16] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-7.4-scalers0-aggregators0-16-32] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-7.4-scalers0-aggregators0-32-16] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-7.4-scalers0-aggregators0-32-32] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-7.4-scalers0-aggregators1-16-16] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-7.4-scalers0-aggregators1-16-32] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-7.4-scalers0-aggregators1-32-16] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-7.4-scalers0-aggregators1-32-32] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-7.4-scalers0-aggregators2-16-16] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-7.4-scalers0-aggregators2-16-32] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-7.4-scalers0-aggregators2-32-16] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-7.4-scalers0-aggregators2-32-32] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-7.4-scalers1-aggregators0-16-16] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-7.4-scalers1-aggregators0-16-32] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-7.4-scalers1-aggregators0-32-16] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-7.4-scalers1-aggregators0-32-32] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-7.4-scalers1-aggregators1-16-16] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-7.4-scalers1-aggregators1-16-32] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-7.4-scalers1-aggregators1-32-16] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-7.4-scalers1-aggregators1-32-32] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-7.4-scalers1-aggregators2-16-16] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-7.4-scalers1-aggregators2-16-32] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-7.4-scalers1-aggregators2-32-16] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-7.4-scalers1-aggregators2-32-32] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-2.5-scalers0-aggregators0-16-16] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-2.5-scalers0-aggregators0-16-32] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-2.5-scalers0-aggregators0-32-16] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-2.5-scalers0-aggregators0-32-32] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-2.5-scalers0-aggregators1-16-16] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-2.5-scalers0-aggregators1-16-32] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-2.5-scalers0-aggregators1-32-16] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-2.5-scalers0-aggregators1-32-32] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-2.5-scalers0-aggregators2-16-16] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-2.5-scalers0-aggregators2-16-32] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-2.5-scalers0-aggregators2-32-16] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-2.5-scalers0-aggregators2-32-32] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-2.5-scalers1-aggregators0-16-16] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-2.5-scalers1-aggregators0-16-32] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-2.5-scalers1-aggregators0-32-16] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-2.5-scalers1-aggregators0-32-32] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-2.5-scalers1-aggregators1-16-16] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-2.5-scalers1-aggregators1-16-32] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-2.5-scalers1-aggregators1-32-16] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-2.5-scalers1-aggregators1-32-32] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-2.5-scalers1-aggregators2-16-16] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-2.5-scalers1-aggregators2-16-32] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-2.5-scalers1-aggregators2-32-16] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-2.5-scalers1-aggregators2-32-32] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-7.4-scalers0-aggregators0-16-16] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-7.4-scalers0-aggregators0-16-32] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-7.4-scalers0-aggregators0-32-16] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-7.4-scalers0-aggregators0-32-32] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-7.4-scalers0-aggregators1-16-16] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-7.4-scalers0-aggregators1-16-32] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-7.4-scalers0-aggregators1-32-16] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-7.4-scalers0-aggregators1-32-32] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-7.4-scalers0-aggregators2-16-16] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-7.4-scalers0-aggregators2-16-32] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-7.4-scalers0-aggregators2-32-16] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-7.4-scalers0-aggregators2-32-32] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-7.4-scalers1-aggregators0-16-16] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-7.4-scalers1-aggregators0-16-32] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-7.4-scalers1-aggregators0-32-16] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-7.4-scalers1-aggregators0-32-32] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-7.4-scalers1-aggregators1-16-16] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-7.4-scalers1-aggregators1-16-32] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-7.4-scalers1-aggregators1-32-16] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-7.4-scalers1-aggregators1-32-32] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-7.4-scalers1-aggregators2-16-16] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-7.4-scalers1-aggregators2-16-32] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-7.4-scalers1-aggregators2-32-16] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-7.4-scalers1-aggregators2-32-32] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-2.5-scalers0-aggregators0-16-16] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-2.5-scalers0-aggregators0-16-32] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-2.5-scalers0-aggregators0-32-16] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-2.5-scalers0-aggregators0-32-32] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-2.5-scalers0-aggregators1-16-16] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-2.5-scalers0-aggregators1-16-32] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-2.5-scalers0-aggregators1-32-16] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-2.5-scalers0-aggregators1-32-32] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-2.5-scalers0-aggregators2-16-16] PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-2.5-scalers0-aggregators2-16-32] PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-2.5-scalers0-aggregators2-32-16] PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-2.5-scalers0-aggregators2-32-32] PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-2.5-scalers1-aggregators0-16-16] PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-2.5-scalers1-aggregators0-16-32] PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-2.5-scalers1-aggregators0-32-16] PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-2.5-scalers1-aggregators0-32-32] PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-2.5-scalers1-aggregators1-16-16] PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-2.5-scalers1-aggregators1-16-32] PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-2.5-scalers1-aggregators1-32-16] PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-2.5-scalers1-aggregators1-32-32] PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-2.5-scalers1-aggregators2-16-16] PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-2.5-scalers1-aggregators2-16-32] PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-2.5-scalers1-aggregators2-32-16] PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-2.5-scalers1-aggregators2-32-32] PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-7.4-scalers0-aggregators0-16-16] PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-7.4-scalers0-aggregators0-16-32] PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-7.4-scalers0-aggregators0-32-16] PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-7.4-scalers0-aggregators0-32-32] PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-7.4-scalers0-aggregators1-16-16] PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-7.4-scalers0-aggregators1-16-32] PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-7.4-scalers0-aggregators1-32-16] PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-7.4-scalers0-aggregators1-32-32] PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-7.4-scalers0-aggregators2-16-16] PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-7.4-scalers0-aggregators2-16-32] PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-7.4-scalers0-aggregators2-32-16] PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-7.4-scalers0-aggregators2-32-32] PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-7.4-scalers1-aggregators0-16-16] PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-7.4-scalers1-aggregators0-16-32] PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-7.4-scalers1-aggregators0-32-16] Running on dglci-manual-gpu-worker in /root/jenkins/workspace/dgl_PR-4648
[Pipeline] {
[Pipeline] checkout
The recommended git tool is: git
using credential 150de63f-189c-4717-bcaf-010460d2f51a
Cloning the remote Git repository
Cloning with configured refspecs honoured and without tags
Cloning repository https://github.com/dmlc/dgl.git
 > git init /root/jenkins/workspace/dgl_PR-4648 # timeout=10
Fetching upstream changes from https://github.com/dmlc/dgl.git
 > git --version # timeout=10
 > git --version # 'git version 2.17.1'
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git fetch --no-tags --progress -- https://github.com/dmlc/dgl.git +refs/pull/4648/head:refs/remotes/origin/PR-4648 +refs/heads/master:refs/remotes/origin/master # timeout=10
Cleaning workspace
Fetching without tags
Merging remotes/origin/master commit d78a3a4baf611b90871a849f58647160c7cd9ab4 into PR head commit 37ed78035f406940ed88a4b60cadd2b7cfde5fea
Merge succeeded, producing 37ed78035f406940ed88a4b60cadd2b7cfde5fea
Checking out Revision 37ed78035f406940ed88a4b60cadd2b7cfde5fea (PR-4648)
 > git config remote.origin.url https://github.com/dmlc/dgl.git # timeout=10
 > git config --add remote.origin.fetch +refs/pull/4648/head:refs/remotes/origin/PR-4648 # timeout=10
 > git config --add remote.origin.fetch +refs/heads/master:refs/remotes/origin/master # timeout=10
 > git config remote.origin.url https://github.com/dmlc/dgl.git # timeout=10
 > git rev-parse --verify HEAD # timeout=10
No valid HEAD. Skipping the resetting
 > git clean -fdx # timeout=10
Fetching upstream changes from https://github.com/dmlc/dgl.git
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git fetch --no-tags --progress -- https://github.com/dmlc/dgl.git +refs/pull/4648/head:refs/remotes/origin/PR-4648 +refs/heads/master:refs/remotes/origin/master # timeout=10
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git remote # timeout=10
 > git config --get remote.origin.url # timeout=10
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git merge d78a3a4baf611b90871a849f58647160c7cd9ab4 # timeout=10
 > git rev-parse HEAD^{commit} # timeout=10
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
Commit message: "fix for pytorch < 1.12"
Cleaning workspace
PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-7.4-scalers1-aggregators0-32-32] [Pipeline] withEnv
[Pipeline] {
[Pipeline] isUnix
[Pipeline] withEnv
[Pipeline] {
[Pipeline] sh
 > git branch -a -v --no-abbrev # timeout=10
 > git checkout -b PR-4648 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git rev-parse --verify HEAD # timeout=10
Resetting working tree
 > git reset --hard # timeout=10
 > git clean -fdx # timeout=10
+ docker pull dgllib/dgl-ci-gpu:cu101_v220816
cu101_v220816: Pulling from dgllib/dgl-ci-gpu
Digest: sha256:ca40fc52876a2563a4e904d0c271d658c1acc8e6a4f8611b578bb49f8c7fd925
Status: Image is up to date for dgllib/dgl-ci-gpu:cu101_v220816
docker.io/dgllib/dgl-ci-gpu:cu101_v220816
[Pipeline] }
PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-7.4-scalers1-aggregators1-16-16] [Pipeline] // withEnv
[Pipeline] isUnix
[Pipeline] withEnv
[Pipeline] {
[Pipeline] sh
+ docker inspect -f . dgllib/dgl-ci-gpu:cu101_v220816
.
[Pipeline] }
[Pipeline] // withEnv
[Pipeline] withDockerContainer
dglci-manual-gpu-worker does not seem to be running inside a container
$ docker run -t -d -u 0:0 --runtime nvidia -w /root/jenkins/workspace/dgl_PR-4648 -v /root/jenkins/workspace/dgl_PR-4648:/root/jenkins/workspace/dgl_PR-4648:rw,z -v /root/jenkins/workspace/dgl_PR-4648@tmp:/root/jenkins/workspace/dgl_PR-4648@tmp:rw,z -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** dgllib/dgl-ci-gpu:cu101_v220816 cat
$ docker top b2202f61772e10290ef30a280d6bba21fca8f51e5ea4851aba487f9ad5c84e0e -eo pid,comm
[Pipeline] {
[Pipeline] sh
+ rm -rf CMakeLists.txt CONTRIBUTORS.md Jenkinsfile LICENSE NEWS.md README.md apps benchmarks cmake conda dglgo docker docs examples featgraph include pyproject.toml python readthedocs.yml src tensoradapter tests third_party tools tutorials
[Pipeline] checkout
The recommended git tool is: git
using credential 150de63f-189c-4717-bcaf-010460d2f51a
Warning: JENKINS-30600: special launcher org.jenkinsci.plugins.docker.workflow.WithContainerStep$Decorator$1@37856046; decorates RemoteLauncher[hudson.remoting.Channel@41a64267:dglci-manual-gpu-worker] will be ignored (a typical symptom is the Git executable not being run inside a designated container)
Fetching changes from the remote Git repository
Cleaning workspace
Fetching without tags
PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-7.4-scalers1-aggregators1-16-32] Merging remotes/origin/master commit d78a3a4baf611b90871a849f58647160c7cd9ab4 into PR head commit 37ed78035f406940ed88a4b60cadd2b7cfde5fea
Merge succeeded, producing 37ed78035f406940ed88a4b60cadd2b7cfde5fea
Checking out Revision 37ed78035f406940ed88a4b60cadd2b7cfde5fea (PR-4648)
Commit message: "fix for pytorch < 1.12"
Cleaning workspace
 > git rev-parse --resolve-git-dir /root/jenkins/workspace/dgl_PR-4648/.git # timeout=10
 > git config remote.origin.url https://github.com/dmlc/dgl.git # timeout=10
 > git rev-parse --verify HEAD # timeout=10
Resetting working tree
 > git reset --hard # timeout=10
 > git clean -fdx # timeout=10
Fetching upstream changes from https://github.com/dmlc/dgl.git
 > git --version # timeout=10
 > git --version # 'git version 2.17.1'
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git fetch --no-tags --progress -- https://github.com/dmlc/dgl.git +refs/pull/4648/head:refs/remotes/origin/PR-4648 +refs/heads/master:refs/remotes/origin/master # timeout=10
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git remote # timeout=10
 > git config --get remote.origin.url # timeout=10
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git merge d78a3a4baf611b90871a849f58647160c7cd9ab4 # timeout=10
 > git rev-parse HEAD^{commit} # timeout=10
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git branch -a -v --no-abbrev # timeout=10
 > git branch -D PR-4648 # timeout=10
 > git checkout -b PR-4648 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git rev-parse --verify HEAD # timeout=10
Resetting working tree
 > git reset --hard # timeout=10
 > git clean -fdx # timeout=10
[Pipeline] sh
+ git submodule update --recursive --init
Submodule 'third_party/METIS' (https://github.com/KarypisLab/METIS.git) registered for path 'third_party/METIS'
Submodule 'third_party/dlpack' (https://github.com/dmlc/dlpack.git) registered for path 'third_party/dlpack'
Submodule 'third_party/dmlc-core' (https://github.com/dmlc/dmlc-core.git) registered for path 'third_party/dmlc-core'
Submodule 'third_party/googletest' (https://github.com/google/googletest.git) registered for path 'third_party/googletest'
Submodule 'third_party/libxsmm' (https://github.com/hfp/libxsmm.git) registered for path 'third_party/libxsmm'
Submodule 'third_party/nanoflann' (https://github.com/jlblancoc/nanoflann) registered for path 'third_party/nanoflann'
Submodule 'third_party/nccl' (https://github.com/nvidia/nccl) registered for path 'third_party/nccl'
Submodule 'third_party/phmap' (https://github.com/greg7mdp/parallel-hashmap.git) registered for path 'third_party/phmap'
Submodule 'third_party/tensorpipe' (https://github.com/pytorch/tensorpipe) registered for path 'third_party/tensorpipe'
Submodule 'third_party/thrust' (https://github.com/NVIDIA/thrust.git) registered for path 'third_party/thrust'
Submodule 'third_party/tvm' (https://github.com/apache/incubator-tvm) registered for path 'third_party/tvm'
Submodule 'third_party/xbyak' (https://github.com/herumi/xbyak) registered for path 'third_party/xbyak'
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/METIS'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/dlpack'...
PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-7.4-scalers1-aggregators1-32-16] Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/dmlc-core'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/googletest'...
PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-7.4-scalers1-aggregators1-32-32] Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/libxsmm'...
PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-7.4-scalers1-aggregators2-16-16] PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-7.4-scalers1-aggregators2-16-32] PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-7.4-scalers1-aggregators2-32-16] PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-7.4-scalers1-aggregators2-32-32] PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-2.5-scalers0-aggregators0-16-16] PASSED [ 77%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-2.5-scalers0-aggregators0-16-32] PASSED [ 77%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-2.5-scalers0-aggregators0-32-16] PASSED [ 77%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-2.5-scalers0-aggregators0-32-32] PASSED [ 77%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-2.5-scalers0-aggregators1-16-16] PASSED [ 77%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-2.5-scalers0-aggregators1-16-32] PASSED [ 77%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-2.5-scalers0-aggregators1-32-16] PASSED [ 77%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-2.5-scalers0-aggregators1-32-32] PASSED [ 77%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-2.5-scalers0-aggregators2-16-16] PASSED [ 77%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-2.5-scalers0-aggregators2-16-32] PASSED [ 77%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-2.5-scalers0-aggregators2-32-16] PASSED [ 77%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-2.5-scalers0-aggregators2-32-32] PASSED [ 77%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-2.5-scalers1-aggregators0-16-16] PASSED [ 77%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-2.5-scalers1-aggregators0-16-32] PASSED [ 77%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-2.5-scalers1-aggregators0-32-16] PASSED [ 77%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-2.5-scalers1-aggregators0-32-32] Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/nanoflann'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/nccl'...
PASSED [ 77%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-2.5-scalers1-aggregators1-16-16] PASSED [ 77%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-2.5-scalers1-aggregators1-16-32] Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/phmap'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tensorpipe'...
PASSED [ 77%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-2.5-scalers1-aggregators1-32-16] PASSED [ 77%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-2.5-scalers1-aggregators1-32-32] Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/thrust'...
PASSED [ 77%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-2.5-scalers1-aggregators2-16-16] PASSED [ 77%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-2.5-scalers1-aggregators2-16-32] PASSED [ 77%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-2.5-scalers1-aggregators2-32-16] PASSED [ 77%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-2.5-scalers1-aggregators2-32-32] PASSED [ 77%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-7.4-scalers0-aggregators0-16-16] Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tvm'...
PASSED [ 77%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-7.4-scalers0-aggregators0-16-32] PASSED [ 77%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-7.4-scalers0-aggregators0-32-16] PASSED [ 77%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-7.4-scalers0-aggregators0-32-32] PASSED [ 77%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-7.4-scalers0-aggregators1-16-16] PASSED [ 77%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-7.4-scalers0-aggregators1-16-32] PASSED [ 77%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-7.4-scalers0-aggregators1-32-16] PASSED [ 77%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-7.4-scalers0-aggregators1-32-32] PASSED [ 77%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-7.4-scalers0-aggregators2-16-16] PASSED [ 77%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-7.4-scalers0-aggregators2-16-32] PASSED [ 77%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-7.4-scalers0-aggregators2-32-16] Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/xbyak'...
Submodule path 'third_party/METIS': checked out '10603482cc055626ff99d2ac3ab7e73d1119439d'
Submodule 'GKlib' (https://github.com/KarypisLab/GKlib.git) registered for path 'third_party/METIS/GKlib'
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/METIS/GKlib'...
PASSED [ 77%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-7.4-scalers0-aggregators2-32-32] Submodule path 'third_party/METIS/GKlib': checked out '62de20c7f97c719abdc96d7c3e266a1f7cb52bc5'
Submodule path 'third_party/dlpack': checked out 'e2bdd3bee8cb6501558042633fa59144cc8b7f5f'
Submodule path 'third_party/dmlc-core': checked out 'bfad207b448480783a1f428ae3d93d87032d8349'
Submodule path 'third_party/googletest': checked out 'f71fb4f9a912ec945401cc49a287a759b6131026'
PASSED [ 77%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-7.4-scalers1-aggregators0-16-16] Submodule path 'third_party/libxsmm': checked out 'fa687556130b6298430f1c0555a14cf79ab6101c'
Submodule path 'third_party/nanoflann': checked out '4c47ca200209550c5628c89803591f8a753c8181'
Submodule path 'third_party/nccl': checked out 'e11238b3029795d33f958b5868d47c90c4f22628'
Submodule path 'third_party/phmap': checked out '25293cefd8b85491b45600c03fe8edf07647553f'
PASSED [ 77%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-7.4-scalers1-aggregators0-16-32] Submodule path 'third_party/tensorpipe': checked out '6042f1a4cbce8eef997f11ed0012de137b317361'
Submodule 'third_party/googletest' (https://github.com/google/googletest.git) registered for path 'third_party/tensorpipe/third_party/googletest'
Submodule 'third_party/libnop' (https://github.com/google/libnop.git) registered for path 'third_party/tensorpipe/third_party/libnop'
Submodule 'third_party/libuv' (https://github.com/libuv/libuv.git) registered for path 'third_party/tensorpipe/third_party/libuv'
Submodule 'third_party/pybind11' (https://github.com/pybind/pybind11.git) registered for path 'third_party/tensorpipe/third_party/pybind11'
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tensorpipe/third_party/googletest'...
PASSED [ 77%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-7.4-scalers1-aggregators0-32-16] Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tensorpipe/third_party/libnop'...
PASSED [ 77%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-7.4-scalers1-aggregators0-32-32] Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tensorpipe/third_party/libuv'...
PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-7.4-scalers1-aggregators1-16-16] PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-7.4-scalers1-aggregators1-16-32] Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tensorpipe/third_party/pybind11'...
PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-7.4-scalers1-aggregators1-32-16] Submodule path 'third_party/tensorpipe/third_party/googletest': checked out 'aee0f9d9b5b87796ee8a0ab26b7587ec30e8858e'
Submodule path 'third_party/tensorpipe/third_party/libnop': checked out 'aa95422ea8c409e3f078d2ee7708a5f59a8b9fa2'
Submodule path 'third_party/tensorpipe/third_party/libuv': checked out '1dff88e5161cba5c59276d2070d2e304e4dcb242'
PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-7.4-scalers1-aggregators1-32-32] Submodule path 'third_party/tensorpipe/third_party/pybind11': checked out 'a23996fce38ff6ccfbcdc09f1e63f2c4be5ea2ef'
Submodule 'tools/clang' (https://github.com/wjakob/clang-cindex-python3) registered for path 'third_party/tensorpipe/third_party/pybind11/tools/clang'
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tensorpipe/third_party/pybind11/tools/clang'...
Submodule path 'third_party/tensorpipe/third_party/pybind11/tools/clang': checked out '6a00cbc4a9b8e68b71caf7f774b3f9c753ae84d5'
Submodule path 'third_party/thrust': checked out '6a3078c64cab0e2f276340fa5dcafa0d758ed890'
Submodule 'cub' (https://github.com/NVIDIA/cub.git) registered for path 'third_party/thrust/dependencies/cub'
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/thrust/dependencies/cub'...
PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-7.4-scalers1-aggregators2-16-16] PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-7.4-scalers1-aggregators2-16-32] PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-7.4-scalers1-aggregators2-32-16] PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-7.4-scalers1-aggregators2-32-32] PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-2.5-scalers0-aggregators0-16-16] Submodule path 'third_party/thrust/dependencies/cub': checked out 'cdaa9558a85e45d849016e5fe7b6e4ee79113f95'
PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-2.5-scalers0-aggregators0-16-32] PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-2.5-scalers0-aggregators0-32-16] Submodule path 'third_party/tvm': checked out 'b2e418cb109df4cd1f17a2cf2894a1b396a6b838'
Submodule 'dlpack' (https://github.com/dmlc/dlpack) registered for path 'third_party/tvm/3rdparty/dlpack'
Submodule 'dmlc-core' (https://github.com/dmlc/dmlc-core) registered for path 'third_party/tvm/3rdparty/dmlc-core'
Submodule '3rdparty/rang' (https://github.com/agauniyal/rang) registered for path 'third_party/tvm/3rdparty/rang'
Submodule '3rdparty/vta-hw' (https://github.com/apache/incubator-tvm-vta) registered for path 'third_party/tvm/3rdparty/vta-hw'
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tvm/3rdparty/dlpack'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tvm/3rdparty/dmlc-core'...
PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-2.5-scalers0-aggregators0-32-32] Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tvm/3rdparty/rang'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tvm/3rdparty/vta-hw'...
PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-2.5-scalers0-aggregators1-16-16] PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-2.5-scalers0-aggregators1-16-32] Submodule path 'third_party/tvm/3rdparty/dlpack': checked out '3ec04430e89a6834e5a1b99471f415fa939bf642'
Submodule path 'third_party/tvm/3rdparty/dmlc-core': checked out '6c401e242c59a1f4c913918246591bb13fd714e7'
Submodule path 'third_party/tvm/3rdparty/rang': checked out 'cabe04d6d6b05356fa8f9741704924788f0dd762'
Submodule path 'third_party/tvm/3rdparty/vta-hw': checked out '87ce9acfae550d1a487746e9d06c2e250076e54c'
Submodule path 'third_party/xbyak': checked out '757e4063f6464740b8ff4a2cae9136d2f8458020'
[Pipeline] unstash
PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-2.5-scalers0-aggregators1-32-16] PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-2.5-scalers0-aggregators1-32-32] PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-2.5-scalers0-aggregators2-16-16] PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-2.5-scalers0-aggregators2-16-32] PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-2.5-scalers0-aggregators2-32-16] PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-2.5-scalers0-aggregators2-32-32] PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-2.5-scalers1-aggregators0-16-16] PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-2.5-scalers1-aggregators0-16-32] PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-2.5-scalers1-aggregators0-32-16] PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-2.5-scalers1-aggregators0-32-32] [Pipeline] echo
Unpacked build/libdgl.so, build/runUnitTests, python/dgl/_ffi/_cy3/core.cpython-*-x86_64-linux-gnu.so, build/tensoradapter/pytorch/*.so from dgl-gpu-linux
[Pipeline] sh
PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-2.5-scalers1-aggregators1-16-16] + bash tests/scripts/task_cpp_unit_test.sh
/root/jenkins/workspace/dgl_PR-4648
~/jenkins/workspace/dgl_PR-4648/build ~/jenkins/workspace/dgl_PR-4648
total 284M
-rwxr-xr-x 1 root root 282M Sep 27 04:50 libdgl.so
-rwxr-xr-x 1 root root 2.5M Sep 27 04:50 runUnitTests
drwxr-xr-x 3 root root 4.0K Sep 27 05:17 tensoradapter
Running main() from /root/jenkins/workspace/dgl_PR-4648@2/third_party/googletest/googletest/src/gtest_main.cc
[==========] Running 101 tests from 23 test suites.
[----------] Global test environment set-up.
[----------] 1 test from GraphTest
[ RUN      ] GraphTest.TestNumVertices
[       OK ] GraphTest.TestNumVertices (0 ms)
[----------] 1 test from GraphTest (0 ms total)

[----------] 3 tests from MessageQueueTest
[ RUN      ] MessageQueueTest.AddRemove
[05:17:45] /root/jenkins/workspace/dgl_PR-4648@2/src/rpc/network/msg_queue.cc:27: Message is larger than the queue.
[       OK ] MessageQueueTest.AddRemove (1 ms)
[ RUN      ] MessageQueueTest.EmptyAndNoMoreAdd
[       OK ] MessageQueueTest.EmptyAndNoMoreAdd (0 ms)
[ RUN      ] MessageQueueTest.MultiThread
[       OK ] MessageQueueTest.MultiThread (6 ms)
[----------] 3 tests from MessageQueueTest (7 ms total)

[----------] 3 tests from SocketCommunicatorTest
[ RUN      ] SocketCommunicatorTest.SendAndRecv
PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-2.5-scalers1-aggregators1-16-32] PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-2.5-scalers1-aggregators1-32-16] PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-2.5-scalers1-aggregators1-32-32] PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-2.5-scalers1-aggregators2-16-16] PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-2.5-scalers1-aggregators2-16-32] PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-2.5-scalers1-aggregators2-32-16] PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-2.5-scalers1-aggregators2-32-32] PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-7.4-scalers0-aggregators0-16-16] PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-7.4-scalers0-aggregators0-16-32] PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-7.4-scalers0-aggregators0-32-16] PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-7.4-scalers0-aggregators0-32-32] PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-7.4-scalers0-aggregators1-16-16] PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-7.4-scalers0-aggregators1-16-32] PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-7.4-scalers0-aggregators1-32-16] PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-7.4-scalers0-aggregators1-32-32] PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-7.4-scalers0-aggregators2-16-16] PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-7.4-scalers0-aggregators2-16-32] PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-7.4-scalers0-aggregators2-32-16] PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-7.4-scalers0-aggregators2-32-32] PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-7.4-scalers1-aggregators0-16-16] PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-7.4-scalers1-aggregators0-16-32] PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-7.4-scalers1-aggregators0-32-16] PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-7.4-scalers1-aggregators0-32-32] PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-7.4-scalers1-aggregators1-16-16] PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-7.4-scalers1-aggregators1-16-32] PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-7.4-scalers1-aggregators1-32-16] PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-7.4-scalers1-aggregators1-32-32] PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-7.4-scalers1-aggregators2-16-16] PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-7.4-scalers1-aggregators2-16-32] PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-7.4-scalers1-aggregators2-32-16] PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-7.4-scalers1-aggregators2-32-32] PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-2.5-scalers0-aggregators0-16-16] PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-2.5-scalers0-aggregators0-16-32] PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-2.5-scalers0-aggregators0-32-16] PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-2.5-scalers0-aggregators0-32-32] [       OK ] SocketCommunicatorTest.SendAndRecv (7002 ms)
[ RUN      ] SocketCommunicatorTest.SendAndRecvTimeout
PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-2.5-scalers0-aggregators1-16-16] PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-2.5-scalers0-aggregators1-16-32] PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-2.5-scalers0-aggregators1-32-16] PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-2.5-scalers0-aggregators1-32-32] PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-2.5-scalers0-aggregators2-16-16] PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-2.5-scalers0-aggregators2-16-32] PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-2.5-scalers0-aggregators2-32-16] PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-2.5-scalers0-aggregators2-32-32] PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-2.5-scalers1-aggregators0-16-16] PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-2.5-scalers1-aggregators0-16-32] PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-2.5-scalers1-aggregators0-32-16] PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-2.5-scalers1-aggregators0-32-32] PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-2.5-scalers1-aggregators1-16-16] PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-2.5-scalers1-aggregators1-16-32] PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-2.5-scalers1-aggregators1-32-16] [       OK ] SocketCommunicatorTest.SendAndRecvTimeout (5000 ms)
[ RUN      ] SocketCommunicatorTest.TCPSocketBind
[       OK ] SocketCommunicatorTest.TCPSocketBind (0 ms)
[----------] 3 tests from SocketCommunicatorTest (12002 ms total)

[----------] 3 tests from SplitStringTest
[ RUN      ] SplitStringTest.SplitStringUsingCompoundDelim
[       OK ] SplitStringTest.SplitStringUsingCompoundDelim (0 ms)
[ RUN      ] SplitStringTest.testSplitStringUsingSingleDelim
[       OK ] SplitStringTest.testSplitStringUsingSingleDelim (0 ms)
[ RUN      ] SplitStringTest.testSplitingNoDelimString
[       OK ] SplitStringTest.testSplitingNoDelimString (0 ms)
[----------] 3 tests from SplitStringTest (0 ms total)

[----------] 1 test from StringPrintf
[ RUN      ] StringPrintf.normal
[       OK ] StringPrintf.normal (0 ms)
[----------] 1 test from StringPrintf (0 ms total)

[----------] 13 tests from ArrayTest
[ RUN      ] ArrayTest.TestCreate
[       OK ] ArrayTest.TestCreate (0 ms)
[ RUN      ] ArrayTest.TestRange
PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-2.5-scalers1-aggregators1-32-32] [       OK ] ArrayTest.TestRange (701 ms)
[ RUN      ] ArrayTest.TestFull
[       OK ] ArrayTest.TestFull (0 ms)
[ RUN      ] ArrayTest.TestClone
[       OK ] ArrayTest.TestClone (0 ms)
[ RUN      ] ArrayTest.TestAsNumBits
[       OK ] ArrayTest.TestAsNumBits (0 ms)
[ RUN      ] ArrayTest.Arith
[       OK ] ArrayTest.Arith (1 ms)
[ RUN      ] ArrayTest.HStack
[       OK ] ArrayTest.HStack (1 ms)
[ RUN      ] ArrayTest.TestIndexSelect
[       OK ] ArrayTest.TestIndexSelect (0 ms)
[ RUN      ] ArrayTest.TestRelabel_
[       OK ] ArrayTest.TestRelabel_ (1 ms)
[ RUN      ] ArrayTest.CumSum
[       OK ] ArrayTest.CumSum (0 ms)
[ RUN      ] ArrayTest.Scatter_
[       OK ] ArrayTest.Scatter_ (0 ms)
[ RUN      ] ArrayTest.NonZero
[       OK ] ArrayTest.NonZero (1 ms)
[ RUN      ] ArrayTest.Sort
[       OK ] ArrayTest.Sort (0 ms)
[----------] 13 tests from ArrayTest (706 ms total)

[----------] 2 tests from MatrixTest
[ RUN      ] MatrixTest.TestToSimpleCsr
[       OK ] MatrixTest.TestToSimpleCsr (1 ms)
[ RUN      ] MatrixTest.TestToSimpleCoo
[       OK ] MatrixTest.TestToSimpleCoo (0 ms)
[----------] 2 tests from MatrixTest (1 ms total)

[----------] 2 tests from DisjointUnionTest
[ RUN      ] DisjointUnionTest.TestDisjointUnionPartitionCoo
[       OK ] DisjointUnionTest.TestDisjointUnionPartitionCoo (0 ms)
[ RUN      ] DisjointUnionTest.TestDisjointUnionPartitionCsr
[       OK ] DisjointUnionTest.TestDisjointUnionPartitionCsr (0 ms)
[----------] 2 tests from DisjointUnionTest (0 ms total)

[----------] 2 tests from SliceContiguousChunk
[ RUN      ] SliceContiguousChunk.TestSliceContiguousChunkCoo
[       OK ] SliceContiguousChunk.TestSliceContiguousChunkCoo (0 ms)
[ RUN      ] SliceContiguousChunk.TestSliceContiguousChunkCsr
[       OK ] SliceContiguousChunk.TestSliceContiguousChunkCsr (1 ms)
[----------] 2 tests from SliceContiguousChunk (1 ms total)

[----------] 2 tests from MatrixUnionTest
[ RUN      ] MatrixUnionTest.TestMatrixUnionCsr
[       OK ] MatrixUnionTest.TestMatrixUnionCsr (0 ms)
[ RUN      ] MatrixUnionTest.TestMatrixUnionCoo
[       OK ] MatrixUnionTest.TestMatrixUnionCoo (0 ms)
[----------] 2 tests from MatrixUnionTest (0 ms total)

[----------] 1 test from LineGraphTest
[ RUN      ] LineGraphTest.LineGraphCOO
[       OK ] LineGraphTest.LineGraphCOO (0 ms)
[----------] 1 test from LineGraphTest (0 ms total)

[----------] 3 tests from CsrmmTest
[ RUN      ] CsrmmTest.TestCsrmm
PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-2.5-scalers1-aggregators2-16-16] [       OK ] CsrmmTest.TestCsrmm (337 ms)
[ RUN      ] CsrmmTest.TestCsrsum
[       OK ] CsrmmTest.TestCsrsum (5 ms)
[ RUN      ] CsrmmTest.TestCsrmask
[       OK ] CsrmmTest.TestCsrmask (0 ms)
[----------] 3 tests from CsrmmTest (342 ms total)

[----------] 2 tests from PartitionTest
[ RUN      ] PartitionTest.TestRemainderPartition
[       OK ] PartitionTest.TestRemainderPartition (19 ms)
[ RUN      ] PartitionTest.TestRangePartition
[       OK ] PartitionTest.TestRangePartition (23 ms)
[----------] 2 tests from PartitionTest (42 ms total)

[----------] 11 tests from RowwiseTest
[ RUN      ] RowwiseTest.TestCSRSampling
[       OK ] RowwiseTest.TestCSRSampling (11 ms)
[ RUN      ] RowwiseTest.TestCSRSamplingUniform
[       OK ] RowwiseTest.TestCSRSamplingUniform (2 ms)
[ RUN      ] RowwiseTest.TestCSRPerEtypeSampling
[       OK ] RowwiseTest.TestCSRPerEtypeSampling (10 ms)
[ RUN      ] RowwiseTest.TestCSRPerEtypeSamplingUniform
[       OK ] RowwiseTest.TestCSRPerEtypeSamplingUniform (6 ms)
[ RUN      ] RowwiseTest.TestCOOSampling
PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-2.5-scalers1-aggregators2-16-32] [       OK ] RowwiseTest.TestCOOSampling (201 ms)
[ RUN      ] RowwiseTest.TestCOOSamplingUniform
[       OK ] RowwiseTest.TestCOOSamplingUniform (33 ms)
[ RUN      ] RowwiseTest.TestCOOerEtypeSampling
PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-2.5-scalers1-aggregators2-32-16] PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-2.5-scalers1-aggregators2-32-32] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-7.4-scalers0-aggregators0-16-16] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-7.4-scalers0-aggregators0-16-32] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-7.4-scalers0-aggregators0-32-16] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-7.4-scalers0-aggregators0-32-32] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-7.4-scalers0-aggregators1-16-16] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-7.4-scalers0-aggregators1-16-32] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-7.4-scalers0-aggregators1-32-16] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-7.4-scalers0-aggregators1-32-32] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-7.4-scalers0-aggregators2-16-16] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-7.4-scalers0-aggregators2-16-32] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-7.4-scalers0-aggregators2-32-16] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-7.4-scalers0-aggregators2-32-32] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-7.4-scalers1-aggregators0-16-16] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-7.4-scalers1-aggregators0-16-32] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-7.4-scalers1-aggregators0-32-16] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-7.4-scalers1-aggregators0-32-32] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-7.4-scalers1-aggregators1-16-16] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-7.4-scalers1-aggregators1-16-32] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-7.4-scalers1-aggregators1-32-16] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-7.4-scalers1-aggregators1-32-32] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-7.4-scalers1-aggregators2-16-16] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-7.4-scalers1-aggregators2-16-32] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-7.4-scalers1-aggregators2-32-16] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-7.4-scalers1-aggregators2-32-32] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-2.5-scalers0-aggregators0-16-16] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-2.5-scalers0-aggregators0-16-32] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-2.5-scalers0-aggregators0-32-16] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-2.5-scalers0-aggregators0-32-32] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-2.5-scalers0-aggregators1-16-16] [       OK ] RowwiseTest.TestCOOerEtypeSampling (3897 ms)
[ RUN      ] RowwiseTest.TestCOOPerEtypeSamplingUniform
PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-2.5-scalers0-aggregators1-16-32] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-2.5-scalers0-aggregators1-32-16] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-2.5-scalers0-aggregators1-32-32] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-2.5-scalers0-aggregators2-16-16] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-2.5-scalers0-aggregators2-16-32] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-2.5-scalers0-aggregators2-32-16] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-2.5-scalers0-aggregators2-32-32] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-2.5-scalers1-aggregators0-16-16] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-2.5-scalers1-aggregators0-16-32] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-2.5-scalers1-aggregators0-32-16] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-2.5-scalers1-aggregators0-32-32] PASSED [ 81%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-2.5-scalers1-aggregators1-16-16] PASSED [ 81%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-2.5-scalers1-aggregators1-16-32] PASSED [ 81%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-2.5-scalers1-aggregators1-32-16] PASSED [ 81%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-2.5-scalers1-aggregators1-32-32] PASSED [ 81%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-2.5-scalers1-aggregators2-16-16] PASSED [ 81%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-2.5-scalers1-aggregators2-16-32] PASSED [ 81%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-2.5-scalers1-aggregators2-32-16] PASSED [ 81%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-2.5-scalers1-aggregators2-32-32] [       OK ] RowwiseTest.TestCOOPerEtypeSamplingUniform (2637 ms)
[ RUN      ] RowwiseTest.TestCSRTopk
[       OK ] RowwiseTest.TestCSRTopk (6 ms)
[ RUN      ] RowwiseTest.TestCOOTopk
[       OK ] RowwiseTest.TestCOOTopk (4 ms)
[ RUN      ] RowwiseTest.TestCSRSamplingBiased
[       OK ] RowwiseTest.TestCSRSamplingBiased (2 ms)
[----------] 11 tests from RowwiseTest (6809 ms total)

[----------] 3 tests from SampleUtilsTest
[ RUN      ] SampleUtilsTest.TestWithReplacement
PASSED [ 81%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-7.4-scalers0-aggregators0-16-16] PASSED [ 81%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-7.4-scalers0-aggregators0-16-32] PASSED [ 81%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-7.4-scalers0-aggregators0-32-16] PASSED [ 81%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-7.4-scalers0-aggregators0-32-32] PASSED [ 81%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-7.4-scalers0-aggregators1-16-16] PASSED [ 81%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-7.4-scalers0-aggregators1-16-32] PASSED [ 81%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-7.4-scalers0-aggregators1-32-16] PASSED [ 81%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-7.4-scalers0-aggregators1-32-32] PASSED [ 81%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-7.4-scalers0-aggregators2-16-16] PASSED [ 81%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-7.4-scalers0-aggregators2-16-32] PASSED [ 81%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-7.4-scalers0-aggregators2-32-16] PASSED [ 81%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-7.4-scalers0-aggregators2-32-32] PASSED [ 81%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-7.4-scalers1-aggregators0-16-16] PASSED [ 81%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-7.4-scalers1-aggregators0-16-32] PASSED [ 81%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-7.4-scalers1-aggregators0-32-16] PASSED [ 81%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-7.4-scalers1-aggregators0-32-32] PASSED [ 81%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-7.4-scalers1-aggregators1-16-16] PASSED [ 81%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-7.4-scalers1-aggregators1-16-32] PASSED [ 81%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-7.4-scalers1-aggregators1-32-16] PASSED [ 81%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-7.4-scalers1-aggregators1-32-32] PASSED [ 81%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-7.4-scalers1-aggregators2-16-16] PASSED [ 81%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-7.4-scalers1-aggregators2-16-32] PASSED [ 81%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-7.4-scalers1-aggregators2-32-16] [       OK ] SampleUtilsTest.TestWithReplacement (3302 ms)
[ RUN      ] SampleUtilsTest.TestWithoutReplacementOrder
[       OK ] SampleUtilsTest.TestWithoutReplacementOrder (0 ms)
[ RUN      ] SampleUtilsTest.TestWithoutReplacementUnique
PASSED [ 81%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-7.4-scalers1-aggregators2-32-32] PASSED [ 81%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-2.5-scalers0-aggregators0-16-16] PASSED [ 81%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-2.5-scalers0-aggregators0-16-32] PASSED [ 81%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-2.5-scalers0-aggregators0-32-16] PASSED [ 81%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-2.5-scalers0-aggregators0-32-32] PASSED [ 81%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-2.5-scalers0-aggregators1-16-16] PASSED [ 81%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-2.5-scalers0-aggregators1-16-32] PASSED [ 81%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-2.5-scalers0-aggregators1-32-16] PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-2.5-scalers0-aggregators1-32-32] PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-2.5-scalers0-aggregators2-16-16] PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-2.5-scalers0-aggregators2-16-32] PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-2.5-scalers0-aggregators2-32-16] PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-2.5-scalers0-aggregators2-32-32] PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-2.5-scalers1-aggregators0-16-16] PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-2.5-scalers1-aggregators0-16-32] PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-2.5-scalers1-aggregators0-32-16] PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-2.5-scalers1-aggregators0-32-32] PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-2.5-scalers1-aggregators1-16-16] PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-2.5-scalers1-aggregators1-16-32] PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-2.5-scalers1-aggregators1-32-16] PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-2.5-scalers1-aggregators1-32-32] PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-2.5-scalers1-aggregators2-16-16] PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-2.5-scalers1-aggregators2-16-32] PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-2.5-scalers1-aggregators2-32-16] PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-2.5-scalers1-aggregators2-32-32] PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-7.4-scalers0-aggregators0-16-16] PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-7.4-scalers0-aggregators0-16-32] PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-7.4-scalers0-aggregators0-32-16] PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-7.4-scalers0-aggregators0-32-32] [       OK ] SampleUtilsTest.TestWithoutReplacementUnique (4103 ms)
[----------] 3 tests from SampleUtilsTest (7405 ms total)

[----------] 3 tests from RandomTest
[ RUN      ] RandomTest.TestChoice
[       OK ] RandomTest.TestChoice (1 ms)
[ RUN      ] RandomTest.TestUniformChoice
[       OK ] RandomTest.TestUniformChoice (0 ms)
[ RUN      ] RandomTest.TestBiasedChoice
[       OK ] RandomTest.TestBiasedChoice (16 ms)
[----------] 3 tests from RandomTest (17 ms total)

[----------] 4 tests from Serialize
[ RUN      ] Serialize.UnitGraph_COO
[       OK ] Serialize.UnitGraph_COO (0 ms)
[ RUN      ] Serialize.UnitGraph_CSR
[       OK ] Serialize.UnitGraph_CSR (0 ms)
[ RUN      ] Serialize.ImmutableGraph
[       OK ] Serialize.ImmutableGraph (0 ms)
[ RUN      ] Serialize.HeteroGraph
[       OK ] Serialize.HeteroGraph (0 ms)
[----------] 4 tests from Serialize (0 ms total)

[----------] 3 tests from SmartPtrTest/0, where TypeParam = std::shared_ptr<MyClass>
[ RUN      ] SmartPtrTest/0.Obj_Test
[       OK ] SmartPtrTest/0.Obj_Test (0 ms)
[ RUN      ] SmartPtrTest/0.Vector_Test1
[       OK ] SmartPtrTest/0.Vector_Test1 (0 ms)
[ RUN      ] SmartPtrTest/0.Vector_Test2
[       OK ] SmartPtrTest/0.Vector_Test2 (0 ms)
[----------] 3 tests from SmartPtrTest/0 (0 ms total)

[----------] 3 tests from SmartPtrTest/1, where TypeParam = std::unique_ptr<MyClass, std::default_delete<MyClass> >
[ RUN      ] SmartPtrTest/1.Obj_Test
[       OK ] SmartPtrTest/1.Obj_Test (0 ms)
[ RUN      ] SmartPtrTest/1.Vector_Test1
[       OK ] SmartPtrTest/1.Vector_Test1 (0 ms)
[ RUN      ] SmartPtrTest/1.Vector_Test2
[       OK ] SmartPtrTest/1.Vector_Test2 (0 ms)
[----------] 3 tests from SmartPtrTest/1 (0 ms total)

[----------] 19 tests from SpmatTest
[ RUN      ] SpmatTest.COOToCSR
[       OK ] SpmatTest.COOToCSR (4 ms)
[ RUN      ] SpmatTest.TestCOOHasDuplicate
[       OK ] SpmatTest.TestCOOHasDuplicate (0 ms)
[ RUN      ] SpmatTest.COOSort
[       OK ] SpmatTest.COOSort (2 ms)
[ RUN      ] SpmatTest.TestCOOReorder
[       OK ] SpmatTest.TestCOOReorder (0 ms)
[ RUN      ] SpmatTest.COOGetData
[       OK ] SpmatTest.COOGetData (10 ms)
[ RUN      ] SpmatTest.COOGetDataAndIndices
[       OK ] SpmatTest.COOGetDataAndIndices (0 ms)
[ RUN      ] SpmatTest.TestCSRIsNonZero
[       OK ] SpmatTest.TestCSRIsNonZero (3 ms)
[ RUN      ] SpmatTest.TestCSRGetRowNNZ
[       OK ] SpmatTest.TestCSRGetRowNNZ (0 ms)
[ RUN      ] SpmatTest.TestCSRGetRowColumnIndices
[       OK ] SpmatTest.TestCSRGetRowColumnIndices (1 ms)
[ RUN      ] SpmatTest.TestCSRGetRowData
[       OK ] SpmatTest.TestCSRGetRowData (0 ms)
[ RUN      ] SpmatTest.CSRGetData
[       OK ] SpmatTest.CSRGetData (2 ms)
[ RUN      ] SpmatTest.CSRGetDataAndIndices
[       OK ] SpmatTest.CSRGetDataAndIndices (0 ms)
[ RUN      ] SpmatTest.CSRTranspose
[       OK ] SpmatTest.CSRTranspose (1 ms)
[ RUN      ] SpmatTest.CSRToCOO
[       OK ] SpmatTest.CSRToCOO (1 ms)
[ RUN      ] SpmatTest.TestCSRSliceRows
[       OK ] SpmatTest.TestCSRSliceRows (56 ms)
[ RUN      ] SpmatTest.CSRSliceMatrix
[       OK ] SpmatTest.CSRSliceMatrix (6 ms)
[ RUN      ] SpmatTest.CSRHasDuplicate
[       OK ] SpmatTest.CSRHasDuplicate (1 ms)
[ RUN      ] SpmatTest.CSRSort
[       OK ] SpmatTest.CSRSort (1 ms)
[ RUN      ] SpmatTest.TestCSRReorder
[       OK ] SpmatTest.TestCSRReorder (0 ms)
[----------] 19 tests from SpmatTest (88 ms total)

[----------] 6 tests from SpmmTest
[ RUN      ] SpmmTest.TestSpmmCopyLhs
[       OK ] SpmmTest.TestSpmmCopyLhs (1 ms)
[ RUN      ] SpmmTest.TestSpmmCopyRhs
[       OK ] SpmmTest.TestSpmmCopyRhs (0 ms)
[ RUN      ] SpmmTest.TestSpmmAdd
[       OK ] SpmmTest.TestSpmmAdd (0 ms)
[ RUN      ] SpmmTest.TestSpmmSub
[       OK ] SpmmTest.TestSpmmSub (1 ms)
[ RUN      ] SpmmTest.TestSpmmMul
[       OK ] SpmmTest.TestSpmmMul (0 ms)
[ RUN      ] SpmmTest.TestSpmmDiv
[       OK ] SpmmTest.TestSpmmDiv (0 ms)
[----------] 6 tests from SpmmTest (2 ms total)

[----------] 7 tests from UniGraphTest
[ RUN      ] UniGraphTest.TestUnitGraph_CopyTo
[       OK ] UniGraphTest.TestUnitGraph_CopyTo (1 ms)
[ RUN      ] UniGraphTest.TestUnitGraph_InOutDegrees
PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-7.4-scalers0-aggregators1-16-16] [       OK ] UniGraphTest.TestUnitGraph_InOutDegrees (14 ms)
[ RUN      ] UniGraphTest.TestUnitGraph_Create
[       OK ] UniGraphTest.TestUnitGraph_Create (0 ms)
[ RUN      ] UniGraphTest.TestUnitGraph_GetInCSR
[       OK ] UniGraphTest.TestUnitGraph_GetInCSR (0 ms)
[ RUN      ] UniGraphTest.TestUnitGraph_GetOutCSR
[       OK ] UniGraphTest.TestUnitGraph_GetOutCSR (1 ms)
[ RUN      ] UniGraphTest.TestUnitGraph_GetCOO
[       OK ] UniGraphTest.TestUnitGraph_GetCOO (0 ms)
[ RUN      ] UniGraphTest.TestUnitGraph_Reserve
[       OK ] UniGraphTest.TestUnitGraph_Reserve (0 ms)
[----------] 7 tests from UniGraphTest (16 ms total)

[----------] 4 tests from ZeroCopySerialize
[ RUN      ] ZeroCopySerialize.NDArray
[       OK ] ZeroCopySerialize.NDArray (0 ms)
[ RUN      ] ZeroCopySerialize.ZeroShapeNDArray
[       OK ] ZeroCopySerialize.ZeroShapeNDArray (0 ms)
[ RUN      ] ZeroCopySerialize.SharedMem
[       OK ] ZeroCopySerialize.SharedMem (0 ms)
[ RUN      ] ZeroCopySerialize.HeteroGraph
[       OK ] ZeroCopySerialize.HeteroGraph (0 ms)
[----------] 4 tests from ZeroCopySerialize (0 ms total)

[----------] Global test environment tear-down
[==========] 101 tests from 23 test suites ran. (27438 ms total)
[  PASSED  ] 101 tests.
~/jenkins/workspace/dgl_PR-4648
Post stage
[Pipeline] cleanWs
[WS-CLEANUP] Deleting project workspace...
[WS-CLEANUP] Deferred wipeout is disabled by the job configuration...
[WS-CLEANUP] done
[Pipeline] }
$ docker stop --time=1 b2202f61772e10290ef30a280d6bba21fca8f51e5ea4851aba487f9ad5c84e0e
PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-7.4-scalers0-aggregators1-16-32] PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-7.4-scalers0-aggregators1-32-16] PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-7.4-scalers0-aggregators1-32-32] PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-7.4-scalers0-aggregators2-16-16] PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-7.4-scalers0-aggregators2-16-32] PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-7.4-scalers0-aggregators2-32-16] PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-7.4-scalers0-aggregators2-32-32] PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-7.4-scalers1-aggregators0-16-16] PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-7.4-scalers1-aggregators0-16-32] PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-7.4-scalers1-aggregators0-32-16] $ docker rm -f b2202f61772e10290ef30a280d6bba21fca8f51e5ea4851aba487f9ad5c84e0e
PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-7.4-scalers1-aggregators0-32-32] PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-7.4-scalers1-aggregators1-16-16] PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-7.4-scalers1-aggregators1-16-32] PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-7.4-scalers1-aggregators1-32-16] PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-7.4-scalers1-aggregators1-32-32] PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-7.4-scalers1-aggregators2-16-16] PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-7.4-scalers1-aggregators2-16-32] [Pipeline] // withDockerContainer
[Pipeline] }
[Pipeline] // withEnv
[Pipeline] }
Running on dglci-manual-gpu-worker in /root/jenkins/workspace/dgl_PR-4648
[Pipeline] // node
[Pipeline] {
[Pipeline] }
[Pipeline] // stage
[Pipeline] }
[Pipeline] checkout
The recommended git tool is: git
using credential 150de63f-189c-4717-bcaf-010460d2f51a
Cloning the remote Git repository
Cloning with configured refspecs honoured and without tags
PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-7.4-scalers1-aggregators2-32-16] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-7.4-scalers1-aggregators2-32-32] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-2.5-scalers0-aggregators0-16-16] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-2.5-scalers0-aggregators0-16-32] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-2.5-scalers0-aggregators0-32-16] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-2.5-scalers0-aggregators0-32-32] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-2.5-scalers0-aggregators1-16-16] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-2.5-scalers0-aggregators1-16-32] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-2.5-scalers0-aggregators1-32-16] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-2.5-scalers0-aggregators1-32-32] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-2.5-scalers0-aggregators2-16-16] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-2.5-scalers0-aggregators2-16-32] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-2.5-scalers0-aggregators2-32-16] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-2.5-scalers0-aggregators2-32-32] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-2.5-scalers1-aggregators0-16-16] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-2.5-scalers1-aggregators0-16-32] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-2.5-scalers1-aggregators0-32-16] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-2.5-scalers1-aggregators0-32-32] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-2.5-scalers1-aggregators1-16-16] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-2.5-scalers1-aggregators1-16-32] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-2.5-scalers1-aggregators1-32-16] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-2.5-scalers1-aggregators1-32-32] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-2.5-scalers1-aggregators2-16-16] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-2.5-scalers1-aggregators2-16-32] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-2.5-scalers1-aggregators2-32-16] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-2.5-scalers1-aggregators2-32-32] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-7.4-scalers0-aggregators0-16-16] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-7.4-scalers0-aggregators0-16-32] Cloning repository https://github.com/dmlc/dgl.git
 > git init /root/jenkins/workspace/dgl_PR-4648 # timeout=10
Fetching upstream changes from https://github.com/dmlc/dgl.git
 > git --version # timeout=10
 > git --version # 'git version 2.17.1'
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git fetch --no-tags --progress -- https://github.com/dmlc/dgl.git +refs/pull/4648/head:refs/remotes/origin/PR-4648 +refs/heads/master:refs/remotes/origin/master # timeout=10
PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-7.4-scalers0-aggregators0-32-16] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-7.4-scalers0-aggregators0-32-32] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-7.4-scalers0-aggregators1-16-16] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-7.4-scalers0-aggregators1-16-32] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-7.4-scalers0-aggregators1-32-16] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-7.4-scalers0-aggregators1-32-32] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-7.4-scalers0-aggregators2-16-16] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-7.4-scalers0-aggregators2-16-32] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-7.4-scalers0-aggregators2-32-16] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-7.4-scalers0-aggregators2-32-32] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-7.4-scalers1-aggregators0-16-16] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-7.4-scalers1-aggregators0-16-32] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-7.4-scalers1-aggregators0-32-16] PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-7.4-scalers1-aggregators0-32-32] PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-7.4-scalers1-aggregators1-16-16] PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-7.4-scalers1-aggregators1-16-32] PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-7.4-scalers1-aggregators1-32-16] PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-7.4-scalers1-aggregators1-32-32] PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-7.4-scalers1-aggregators2-16-16] Cleaning workspace
Fetching without tags
Merging remotes/origin/master commit d78a3a4baf611b90871a849f58647160c7cd9ab4 into PR head commit 37ed78035f406940ed88a4b60cadd2b7cfde5fea
 > git config remote.origin.url https://github.com/dmlc/dgl.git # timeout=10
 > git config --add remote.origin.fetch +refs/pull/4648/head:refs/remotes/origin/PR-4648 # timeout=10
 > git config --add remote.origin.fetch +refs/heads/master:refs/remotes/origin/master # timeout=10
 > git config remote.origin.url https://github.com/dmlc/dgl.git # timeout=10
 > git rev-parse --verify HEAD # timeout=10
No valid HEAD. Skipping the resetting
 > git clean -fdx # timeout=10
Fetching upstream changes from https://github.com/dmlc/dgl.git
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git fetch --no-tags --progress -- https://github.com/dmlc/dgl.git +refs/pull/4648/head:refs/remotes/origin/PR-4648 +refs/heads/master:refs/remotes/origin/master # timeout=10
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
Merge succeeded, producing 37ed78035f406940ed88a4b60cadd2b7cfde5fea
Checking out Revision 37ed78035f406940ed88a4b60cadd2b7cfde5fea (PR-4648)
PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-7.4-scalers1-aggregators2-16-32] Commit message: "fix for pytorch < 1.12"
Cleaning workspace
PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-7.4-scalers1-aggregators2-32-16] PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-7.4-scalers1-aggregators2-32-32] PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-2.5-scalers0-aggregators0-16-16] PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-2.5-scalers0-aggregators0-16-32] PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-2.5-scalers0-aggregators0-32-16] PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-2.5-scalers0-aggregators0-32-32] PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-2.5-scalers0-aggregators1-16-16] [Pipeline] withEnv
[Pipeline] {
[Pipeline] isUnix
[Pipeline] withEnv
[Pipeline] {
[Pipeline] sh
 > git remote # timeout=10
 > git config --get remote.origin.url # timeout=10
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git merge d78a3a4baf611b90871a849f58647160c7cd9ab4 # timeout=10
 > git rev-parse HEAD^{commit} # timeout=10
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git branch -a -v --no-abbrev # timeout=10
 > git checkout -b PR-4648 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git rev-parse --verify HEAD # timeout=10
Resetting working tree
 > git reset --hard # timeout=10
 > git clean -fdx # timeout=10
PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-2.5-scalers0-aggregators1-16-32] PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-2.5-scalers0-aggregators1-32-16] + docker pull dgllib/dgl-ci-gpu:cu101_v220816
PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-2.5-scalers0-aggregators1-32-32] PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-2.5-scalers0-aggregators2-16-16] PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-2.5-scalers0-aggregators2-16-32] PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-2.5-scalers0-aggregators2-32-16] PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-2.5-scalers0-aggregators2-32-32] PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-2.5-scalers1-aggregators0-16-16] PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-2.5-scalers1-aggregators0-16-32] cu101_v220816: Pulling from dgllib/dgl-ci-gpu
Digest: sha256:ca40fc52876a2563a4e904d0c271d658c1acc8e6a4f8611b578bb49f8c7fd925
Status: Image is up to date for dgllib/dgl-ci-gpu:cu101_v220816
docker.io/dgllib/dgl-ci-gpu:cu101_v220816
[Pipeline] }
[Pipeline] // withEnv
[Pipeline] isUnix
[Pipeline] withEnv
[Pipeline] {
[Pipeline] sh
PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-2.5-scalers1-aggregators0-32-16] PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-2.5-scalers1-aggregators0-32-32] + docker inspect -f . dgllib/dgl-ci-gpu:cu101_v220816
.
PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-2.5-scalers1-aggregators1-16-16] PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-2.5-scalers1-aggregators1-16-32] [Pipeline] }
[Pipeline] // withEnv
[Pipeline] withDockerContainer
dglci-manual-gpu-worker does not seem to be running inside a container
$ docker run -t -d -u 0:0 --runtime nvidia -w /root/jenkins/workspace/dgl_PR-4648 -v /root/jenkins/workspace/dgl_PR-4648:/root/jenkins/workspace/dgl_PR-4648:rw,z -v /root/jenkins/workspace/dgl_PR-4648@tmp:/root/jenkins/workspace/dgl_PR-4648@tmp:rw,z -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** dgllib/dgl-ci-gpu:cu101_v220816 cat
PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-2.5-scalers1-aggregators1-32-16] PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-2.5-scalers1-aggregators1-32-32] PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-2.5-scalers1-aggregators2-16-16] PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-2.5-scalers1-aggregators2-16-32] $ docker top accf251fd0b56b1d65207f2fd238e906ccb6fddb8b2532d85c1dfc3329fb23a3 -eo pid,comm
[Pipeline] {
[Pipeline] stage
PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-2.5-scalers1-aggregators2-32-16] PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-2.5-scalers1-aggregators2-32-32] [Pipeline] { (Tensorflow GPU Unit test)
[Pipeline] sh
PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-7.4-scalers0-aggregators0-16-16] PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-7.4-scalers0-aggregators0-16-32] + rm -rf CMakeLists.txt CONTRIBUTORS.md Jenkinsfile LICENSE NEWS.md README.md apps benchmarks cmake conda dglgo docker docs examples featgraph include pyproject.toml python readthedocs.yml src tensoradapter tests third_party tools tutorials
[Pipeline] checkout
The recommended git tool is: git
using credential 150de63f-189c-4717-bcaf-010460d2f51a
Warning: JENKINS-30600: special launcher org.jenkinsci.plugins.docker.workflow.WithContainerStep$Decorator$1@1ed0eff6; decorates RemoteLauncher[hudson.remoting.Channel@41a64267:dglci-manual-gpu-worker] will be ignored (a typical symptom is the Git executable not being run inside a designated container)
Fetching changes from the remote Git repository
Cleaning workspace
Fetching without tags
PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-7.4-scalers0-aggregators0-32-16] PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-7.4-scalers0-aggregators0-32-32] PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-7.4-scalers0-aggregators1-16-16] PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-7.4-scalers0-aggregators1-16-32] Merging remotes/origin/master commit d78a3a4baf611b90871a849f58647160c7cd9ab4 into PR head commit 37ed78035f406940ed88a4b60cadd2b7cfde5fea
Merge succeeded, producing 37ed78035f406940ed88a4b60cadd2b7cfde5fea
Checking out Revision 37ed78035f406940ed88a4b60cadd2b7cfde5fea (PR-4648)
PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-7.4-scalers0-aggregators1-32-16] PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-7.4-scalers0-aggregators1-32-32] Commit message: "fix for pytorch < 1.12"
Cleaning workspace
PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-7.4-scalers0-aggregators2-16-16] PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-7.4-scalers0-aggregators2-16-32]  > git rev-parse --resolve-git-dir /root/jenkins/workspace/dgl_PR-4648/.git # timeout=10
 > git config remote.origin.url https://github.com/dmlc/dgl.git # timeout=10
 > git rev-parse --verify HEAD # timeout=10
Resetting working tree
 > git reset --hard # timeout=10
 > git clean -fdx # timeout=10
Fetching upstream changes from https://github.com/dmlc/dgl.git
 > git --version # timeout=10
 > git --version # 'git version 2.17.1'
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git fetch --no-tags --progress -- https://github.com/dmlc/dgl.git +refs/pull/4648/head:refs/remotes/origin/PR-4648 +refs/heads/master:refs/remotes/origin/master # timeout=10
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git remote # timeout=10
 > git config --get remote.origin.url # timeout=10
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git merge d78a3a4baf611b90871a849f58647160c7cd9ab4 # timeout=10
 > git rev-parse HEAD^{commit} # timeout=10
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git branch -a -v --no-abbrev # timeout=10
 > git branch -D PR-4648 # timeout=10
 > git checkout -b PR-4648 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git rev-parse --verify HEAD # timeout=10
Resetting working tree
 > git reset --hard # timeout=10
 > git clean -fdx # timeout=10
PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-7.4-scalers0-aggregators2-32-16] PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-7.4-scalers0-aggregators2-32-32] [Pipeline] sh
PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-7.4-scalers1-aggregators0-16-16] PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-7.4-scalers1-aggregators0-16-32] + git submodule update --recursive --init
Submodule 'third_party/METIS' (https://github.com/KarypisLab/METIS.git) registered for path 'third_party/METIS'
Submodule 'third_party/dlpack' (https://github.com/dmlc/dlpack.git) registered for path 'third_party/dlpack'
Submodule 'third_party/dmlc-core' (https://github.com/dmlc/dmlc-core.git) registered for path 'third_party/dmlc-core'
Submodule 'third_party/googletest' (https://github.com/google/googletest.git) registered for path 'third_party/googletest'
Submodule 'third_party/libxsmm' (https://github.com/hfp/libxsmm.git) registered for path 'third_party/libxsmm'
Submodule 'third_party/nanoflann' (https://github.com/jlblancoc/nanoflann) registered for path 'third_party/nanoflann'
Submodule 'third_party/nccl' (https://github.com/nvidia/nccl) registered for path 'third_party/nccl'
Submodule 'third_party/phmap' (https://github.com/greg7mdp/parallel-hashmap.git) registered for path 'third_party/phmap'
Submodule 'third_party/tensorpipe' (https://github.com/pytorch/tensorpipe) registered for path 'third_party/tensorpipe'
Submodule 'third_party/thrust' (https://github.com/NVIDIA/thrust.git) registered for path 'third_party/thrust'
Submodule 'third_party/tvm' (https://github.com/apache/incubator-tvm) registered for path 'third_party/tvm'
Submodule 'third_party/xbyak' (https://github.com/herumi/xbyak) registered for path 'third_party/xbyak'
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/METIS'...
PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-7.4-scalers1-aggregators0-32-16] PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-7.4-scalers1-aggregators0-32-32] PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-7.4-scalers1-aggregators1-16-16] PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-7.4-scalers1-aggregators1-16-32] PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-7.4-scalers1-aggregators1-32-16] PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-7.4-scalers1-aggregators1-32-32] PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-7.4-scalers1-aggregators2-16-16] PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-7.4-scalers1-aggregators2-16-32] PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-7.4-scalers1-aggregators2-32-16] Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/dlpack'...
PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-7.4-scalers1-aggregators2-32-32] PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-2.5-scalers0-aggregators0-16-16] Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/dmlc-core'...
PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-2.5-scalers0-aggregators0-16-32] PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-2.5-scalers0-aggregators0-32-16] PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-2.5-scalers0-aggregators0-32-32] PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-2.5-scalers0-aggregators1-16-16] PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-2.5-scalers0-aggregators1-16-32] PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-2.5-scalers0-aggregators1-32-16] PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-2.5-scalers0-aggregators1-32-32] PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-2.5-scalers0-aggregators2-16-16] Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/googletest'...
PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-2.5-scalers0-aggregators2-16-32] PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-2.5-scalers0-aggregators2-32-16] PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-2.5-scalers0-aggregators2-32-32] PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-2.5-scalers1-aggregators0-16-16] PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-2.5-scalers1-aggregators0-16-32] PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-2.5-scalers1-aggregators0-32-16] PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-2.5-scalers1-aggregators0-32-32] PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-2.5-scalers1-aggregators1-16-16] PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-2.5-scalers1-aggregators1-16-32] PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-2.5-scalers1-aggregators1-32-16] PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-2.5-scalers1-aggregators1-32-32] PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-2.5-scalers1-aggregators2-16-16] PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-2.5-scalers1-aggregators2-16-32] PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-2.5-scalers1-aggregators2-32-16] Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/libxsmm'...
PASSED [ 86%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-2.5-scalers1-aggregators2-32-32] PASSED [ 86%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-7.4-scalers0-aggregators0-16-16] PASSED [ 86%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-7.4-scalers0-aggregators0-16-32] PASSED [ 86%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-7.4-scalers0-aggregators0-32-16] PASSED [ 86%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-7.4-scalers0-aggregators0-32-32] PASSED [ 86%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-7.4-scalers0-aggregators1-16-16] PASSED [ 86%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-7.4-scalers0-aggregators1-16-32] PASSED [ 86%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-7.4-scalers0-aggregators1-32-16] PASSED [ 86%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-7.4-scalers0-aggregators1-32-32] PASSED [ 86%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-7.4-scalers0-aggregators2-16-16] PASSED [ 86%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-7.4-scalers0-aggregators2-16-32] PASSED [ 86%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-7.4-scalers0-aggregators2-32-16] PASSED [ 86%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-7.4-scalers0-aggregators2-32-32] PASSED [ 86%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-7.4-scalers1-aggregators0-16-16] PASSED [ 86%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-7.4-scalers1-aggregators0-16-32] PASSED [ 86%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-7.4-scalers1-aggregators0-32-16] PASSED [ 86%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-7.4-scalers1-aggregators0-32-32] PASSED [ 86%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-7.4-scalers1-aggregators1-16-16] PASSED [ 86%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-7.4-scalers1-aggregators1-16-32] PASSED [ 86%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-7.4-scalers1-aggregators1-32-16] PASSED [ 86%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-7.4-scalers1-aggregators1-32-32] PASSED [ 86%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-7.4-scalers1-aggregators2-16-16] PASSED [ 86%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-7.4-scalers1-aggregators2-16-32] PASSED [ 86%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-7.4-scalers1-aggregators2-32-16] PASSED [ 86%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-7.4-scalers1-aggregators2-32-32] PASSED [ 86%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-2.5-scalers0-aggregators0-16-16] PASSED [ 86%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-2.5-scalers0-aggregators0-16-32] PASSED [ 86%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-2.5-scalers0-aggregators0-32-16] PASSED [ 86%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-2.5-scalers0-aggregators0-32-32] PASSED [ 86%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-2.5-scalers0-aggregators1-16-16] PASSED [ 86%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-2.5-scalers0-aggregators1-16-32] PASSED [ 86%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-2.5-scalers0-aggregators1-32-16] PASSED [ 86%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-2.5-scalers0-aggregators1-32-32] PASSED [ 86%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-2.5-scalers0-aggregators2-16-16] PASSED [ 86%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-2.5-scalers0-aggregators2-16-32] PASSED [ 86%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-2.5-scalers0-aggregators2-32-16] PASSED [ 86%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-2.5-scalers0-aggregators2-32-32] PASSED [ 86%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-2.5-scalers1-aggregators0-16-16] PASSED [ 86%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-2.5-scalers1-aggregators0-16-32] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-2.5-scalers1-aggregators0-32-16] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-2.5-scalers1-aggregators0-32-32] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-2.5-scalers1-aggregators1-16-16] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-2.5-scalers1-aggregators1-16-32] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-2.5-scalers1-aggregators1-32-16] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-2.5-scalers1-aggregators1-32-32] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-2.5-scalers1-aggregators2-16-16] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-2.5-scalers1-aggregators2-16-32] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-2.5-scalers1-aggregators2-32-16] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-2.5-scalers1-aggregators2-32-32] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-7.4-scalers0-aggregators0-16-16] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-7.4-scalers0-aggregators0-16-32] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-7.4-scalers0-aggregators0-32-16] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-7.4-scalers0-aggregators0-32-32] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-7.4-scalers0-aggregators1-16-16] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-7.4-scalers0-aggregators1-16-32] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-7.4-scalers0-aggregators1-32-16] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-7.4-scalers0-aggregators1-32-32] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-7.4-scalers0-aggregators2-16-16] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-7.4-scalers0-aggregators2-16-32] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-7.4-scalers0-aggregators2-32-16] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-7.4-scalers0-aggregators2-32-32] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-7.4-scalers1-aggregators0-16-16] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-7.4-scalers1-aggregators0-16-32] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-7.4-scalers1-aggregators0-32-16] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-7.4-scalers1-aggregators0-32-32] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-7.4-scalers1-aggregators1-16-16] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-7.4-scalers1-aggregators1-16-32] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-7.4-scalers1-aggregators1-32-16] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-7.4-scalers1-aggregators1-32-32] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-7.4-scalers1-aggregators2-16-16] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-7.4-scalers1-aggregators2-16-32] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-7.4-scalers1-aggregators2-32-16] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-7.4-scalers1-aggregators2-32-32] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-2.5-scalers0-aggregators0-16-16] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-2.5-scalers0-aggregators0-16-32] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-2.5-scalers0-aggregators0-32-16] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-2.5-scalers0-aggregators0-32-32] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-2.5-scalers0-aggregators1-16-16] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-2.5-scalers0-aggregators1-16-32] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-2.5-scalers0-aggregators1-32-16] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-2.5-scalers0-aggregators1-32-32] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-2.5-scalers0-aggregators2-16-16] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-2.5-scalers0-aggregators2-16-32] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-2.5-scalers0-aggregators2-32-16] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-2.5-scalers0-aggregators2-32-32] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-2.5-scalers1-aggregators0-16-16] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-2.5-scalers1-aggregators0-16-32] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-2.5-scalers1-aggregators0-32-16] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-2.5-scalers1-aggregators0-32-32] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-2.5-scalers1-aggregators1-16-16] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-2.5-scalers1-aggregators1-16-32] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-2.5-scalers1-aggregators1-32-16] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-2.5-scalers1-aggregators1-32-32] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-2.5-scalers1-aggregators2-16-16] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-2.5-scalers1-aggregators2-16-32] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-2.5-scalers1-aggregators2-32-16] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-2.5-scalers1-aggregators2-32-32] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-7.4-scalers0-aggregators0-16-16] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-7.4-scalers0-aggregators0-16-32] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-7.4-scalers0-aggregators0-32-16] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-7.4-scalers0-aggregators0-32-32] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-7.4-scalers0-aggregators1-16-16] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-7.4-scalers0-aggregators1-16-32] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-7.4-scalers0-aggregators1-32-16] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-7.4-scalers0-aggregators1-32-32] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-7.4-scalers0-aggregators2-16-16] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-7.4-scalers0-aggregators2-16-32] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-7.4-scalers0-aggregators2-32-16] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-7.4-scalers0-aggregators2-32-32] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-7.4-scalers1-aggregators0-16-16] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-7.4-scalers1-aggregators0-16-32] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-7.4-scalers1-aggregators0-32-16] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-7.4-scalers1-aggregators0-32-32] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-7.4-scalers1-aggregators1-16-16] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-7.4-scalers1-aggregators1-16-32] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-7.4-scalers1-aggregators1-32-16] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-7.4-scalers1-aggregators1-32-32] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-7.4-scalers1-aggregators2-16-16] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-7.4-scalers1-aggregators2-16-32] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-7.4-scalers1-aggregators2-32-16] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-7.4-scalers1-aggregators2-32-32] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-2.5-scalers0-aggregators0-16-16] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-2.5-scalers0-aggregators0-16-32] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-2.5-scalers0-aggregators0-32-16] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-2.5-scalers0-aggregators0-32-32] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-2.5-scalers0-aggregators1-16-16] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-2.5-scalers0-aggregators1-16-32] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-2.5-scalers0-aggregators1-32-16] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-2.5-scalers0-aggregators1-32-32] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-2.5-scalers0-aggregators2-16-16] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-2.5-scalers0-aggregators2-16-32] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-2.5-scalers0-aggregators2-32-16] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-2.5-scalers0-aggregators2-32-32] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-2.5-scalers1-aggregators0-16-16] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-2.5-scalers1-aggregators0-16-32] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-2.5-scalers1-aggregators0-32-16] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-2.5-scalers1-aggregators0-32-32] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-2.5-scalers1-aggregators1-16-16] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-2.5-scalers1-aggregators1-16-32] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-2.5-scalers1-aggregators1-32-16] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-2.5-scalers1-aggregators1-32-32] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-2.5-scalers1-aggregators2-16-16] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-2.5-scalers1-aggregators2-16-32] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-2.5-scalers1-aggregators2-32-16] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-2.5-scalers1-aggregators2-32-32] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-7.4-scalers0-aggregators0-16-16] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-7.4-scalers0-aggregators0-16-32] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-7.4-scalers0-aggregators0-32-16] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-7.4-scalers0-aggregators0-32-32] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-7.4-scalers0-aggregators1-16-16] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-7.4-scalers0-aggregators1-16-32] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-7.4-scalers0-aggregators1-32-16] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-7.4-scalers0-aggregators1-32-32] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-7.4-scalers0-aggregators2-16-16] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-7.4-scalers0-aggregators2-16-32] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-7.4-scalers0-aggregators2-32-16] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-7.4-scalers0-aggregators2-32-32] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-7.4-scalers1-aggregators0-16-16] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-7.4-scalers1-aggregators0-16-32] PASSED [ 90%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-7.4-scalers1-aggregators0-32-16] PASSED [ 90%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-7.4-scalers1-aggregators0-32-32] PASSED [ 90%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-7.4-scalers1-aggregators1-16-16] PASSED [ 90%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-7.4-scalers1-aggregators1-16-32] PASSED [ 90%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-7.4-scalers1-aggregators1-32-16] PASSED [ 90%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-7.4-scalers1-aggregators1-32-32] PASSED [ 90%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-7.4-scalers1-aggregators2-16-16] PASSED [ 90%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-7.4-scalers1-aggregators2-16-32] PASSED [ 90%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-7.4-scalers1-aggregators2-32-16] PASSED [ 90%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-7.4-scalers1-aggregators2-32-32] PASSED [ 90%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-2.5-scalers0-aggregators0-16-16] PASSED [ 90%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-2.5-scalers0-aggregators0-16-32] PASSED [ 90%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-2.5-scalers0-aggregators0-32-16] PASSED [ 90%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-2.5-scalers0-aggregators0-32-32] PASSED [ 90%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-2.5-scalers0-aggregators1-16-16] PASSED [ 90%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-2.5-scalers0-aggregators1-16-32] PASSED [ 90%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-2.5-scalers0-aggregators1-32-16] PASSED [ 90%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-2.5-scalers0-aggregators1-32-32] PASSED [ 90%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-2.5-scalers0-aggregators2-16-16] PASSED [ 90%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-2.5-scalers0-aggregators2-16-32] PASSED [ 90%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-2.5-scalers0-aggregators2-32-16] PASSED [ 90%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-2.5-scalers0-aggregators2-32-32] Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/nanoflann'...
PASSED [ 90%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-2.5-scalers1-aggregators0-16-16] PASSED [ 90%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-2.5-scalers1-aggregators0-16-32] PASSED [ 90%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-2.5-scalers1-aggregators0-32-16] PASSED [ 90%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-2.5-scalers1-aggregators0-32-32] PASSED [ 90%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-2.5-scalers1-aggregators1-16-16] Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/nccl'...
PASSED [ 90%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-2.5-scalers1-aggregators1-16-32] PASSED [ 90%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-2.5-scalers1-aggregators1-32-16] PASSED [ 90%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-2.5-scalers1-aggregators1-32-32] PASSED [ 90%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-2.5-scalers1-aggregators2-16-16] Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/phmap'...
PASSED [ 90%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-2.5-scalers1-aggregators2-16-32] PASSED [ 90%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-2.5-scalers1-aggregators2-32-16] PASSED [ 90%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-2.5-scalers1-aggregators2-32-32] PASSED [ 90%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-7.4-scalers0-aggregators0-16-16] Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tensorpipe'...
PASSED [ 90%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-7.4-scalers0-aggregators0-16-32] PASSED [ 90%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-7.4-scalers0-aggregators0-32-16] PASSED [ 90%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-7.4-scalers0-aggregators0-32-32] PASSED [ 90%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-7.4-scalers0-aggregators1-16-16] PASSED [ 91%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-7.4-scalers0-aggregators1-16-32] PASSED [ 91%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-7.4-scalers0-aggregators1-32-16] PASSED [ 91%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-7.4-scalers0-aggregators1-32-32] PASSED [ 91%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-7.4-scalers0-aggregators2-16-16] Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/thrust'...
PASSED [ 91%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-7.4-scalers0-aggregators2-16-32] PASSED [ 91%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-7.4-scalers0-aggregators2-32-16] PASSED [ 91%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-7.4-scalers0-aggregators2-32-32] PASSED [ 91%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-7.4-scalers1-aggregators0-16-16] PASSED [ 91%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-7.4-scalers1-aggregators0-16-32] PASSED [ 91%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-7.4-scalers1-aggregators0-32-16] PASSED [ 91%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-7.4-scalers1-aggregators0-32-32] PASSED [ 91%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-7.4-scalers1-aggregators1-16-16] PASSED [ 91%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-7.4-scalers1-aggregators1-16-32] PASSED [ 91%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-7.4-scalers1-aggregators1-32-16] PASSED [ 91%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-7.4-scalers1-aggregators1-32-32] PASSED [ 91%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-7.4-scalers1-aggregators2-16-16] PASSED [ 91%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-7.4-scalers1-aggregators2-16-32] Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tvm'...
PASSED [ 91%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-7.4-scalers1-aggregators2-32-16] PASSED [ 91%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-7.4-scalers1-aggregators2-32-32] PASSED [ 91%]
tests/pytorch/test_optim.py::test_sparse_adam[1] SKIPPED (Do not sup...) [ 91%]
tests/pytorch/test_optim.py::test_sparse_adam[4] SKIPPED (Do not sup...) [ 91%]
tests/pytorch/test_optim.py::test_sparse_adam[101] SKIPPED (Do not s...) [ 91%]
tests/pytorch/test_optim.py::test_sparse_adam[1024] SKIPPED (Do not ...) [ 91%]
tests/pytorch/test_optim.py::test_sparse_adam_uva[1-False] SKIPPED (...) [ 91%]
tests/pytorch/test_optim.py::test_sparse_adam_uva[1-True] SKIPPED (D...) [ 91%]
tests/pytorch/test_optim.py::test_sparse_adam_uva[1-None] SKIPPED (D...) [ 91%]
tests/pytorch/test_optim.py::test_sparse_adam_uva[4-False] SKIPPED (...) [ 91%]
tests/pytorch/test_optim.py::test_sparse_adam_uva[4-True] SKIPPED (D...) [ 91%]
tests/pytorch/test_optim.py::test_sparse_adam_uva[4-None] SKIPPED (D...) [ 91%]
tests/pytorch/test_optim.py::test_sparse_adam_uva[101-False] SKIPPED     [ 91%]
tests/pytorch/test_optim.py::test_sparse_adam_uva[101-True] SKIPPED      [ 91%]
tests/pytorch/test_optim.py::test_sparse_adam_uva[101-None] SKIPPED      [ 91%]
tests/pytorch/test_optim.py::test_sparse_adam_uva[1024-False] SKIPPED    [ 91%]
tests/pytorch/test_optim.py::test_sparse_adam_uva[1024-True] SKIPPED     [ 91%]
tests/pytorch/test_optim.py::test_sparse_adam_uva[1024-None] SKIPPED     [ 91%]
tests/pytorch/test_optim.py::test_sparse_adam_dtype[1-dtype0] SKIPPED    [ 91%]
tests/pytorch/test_optim.py::test_sparse_adam_dtype[1-dtype1] SKIPPED    [ 91%]
tests/pytorch/test_optim.py::test_sparse_adam_dtype[4-dtype0] SKIPPED    [ 91%]
tests/pytorch/test_optim.py::test_sparse_adam_dtype[4-dtype1] SKIPPED    [ 91%]
tests/pytorch/test_optim.py::test_sparse_adam_dtype[101-dtype0] SKIPPED  [ 92%]
tests/pytorch/test_optim.py::test_sparse_adam_dtype[101-dtype1] SKIPPED  [ 92%]
tests/pytorch/test_optim.py::test_sparse_adam_dtype[1024-dtype0] SKIPPED [ 92%]
tests/pytorch/test_optim.py::test_sparse_adam_dtype[1024-dtype1] SKIPPED [ 92%]
tests/pytorch/test_optim.py::test_sparse_adam_zero_step SKIPPED (Do ...) [ 92%]
tests/pytorch/test_optim.py::test_multiprocess_cpu_sparse_adam[2] SKIPPED [ 92%]
tests/pytorch/test_optim.py::test_multiprocess_cpu_sparse_adam[4] SKIPPED [ 92%]
tests/pytorch/test_optim.py::test_multiprocess_sparse_adam[nccl-2] SKIPPED [ 92%]
tests/pytorch/test_optim.py::test_multiprocess_sparse_adam[nccl-4] SKIPPED [ 92%]
tests/pytorch/test_optim.py::test_multiprocess_sparse_adam[nccl-8] SKIPPED [ 92%]
tests/pytorch/test_optim.py::test_multiprocess_sparse_adam[gloo-2] SKIPPED [ 92%]
tests/pytorch/test_optim.py::test_multiprocess_sparse_adam[gloo-4] SKIPPED [ 92%]
tests/pytorch/test_optim.py::test_multiprocess_sparse_adam[gloo-8] SKIPPED [ 92%]
tests/pytorch/test_optim.py::test_multiprocess_sparse_adam_cuda_tensor[2] SKIPPED [ 92%]
tests/pytorch/test_optim.py::test_multiprocess_sparse_adam_cuda_tensor[4] SKIPPED [ 92%]
tests/pytorch/test_optim.py::test_multiprocess_sparse_adam_cuda_tensor[8] SKIPPED [ 92%]
tests/pytorch/test_optim.py::test_multiprocess_sparse_adam_cpu_zero_step[2] SKIPPED [ 92%]
tests/pytorch/test_optim.py::test_multiprocess_sparse_adam_cpu_zero_step[4] SKIPPED [ 92%]
tests/pytorch/test_optim.py::test_multiprocess_sparse_adam_zero_step[nccl-2] SKIPPED [ 92%]
tests/pytorch/test_optim.py::test_multiprocess_sparse_adam_zero_step[nccl-4] SKIPPED [ 92%]
tests/pytorch/test_optim.py::test_multiprocess_sparse_adam_zero_step[nccl-8] SKIPPED [ 92%]
tests/pytorch/test_optim.py::test_multiprocess_sparse_adam_zero_step[gloo-2] SKIPPED [ 92%]
tests/pytorch/test_optim.py::test_multiprocess_sparse_adam_zero_step[gloo-4] SKIPPED [ 92%]
tests/pytorch/test_optim.py::test_multiprocess_sparse_adam_zero_step[gloo-8] SKIPPED [ 92%]
tests/pytorch/test_optim.py::test_multiprocess_sparse_adam_zero_step_cuda_tensor[2] SKIPPED [ 92%]
tests/pytorch/test_optim.py::test_multiprocess_sparse_adam_zero_step_cuda_tensor[4] SKIPPED [ 92%]
tests/pytorch/test_optim.py::test_multiprocess_sparse_adam_zero_step_cuda_tensor[8] SKIPPED [ 92%]
tests/pytorch/test_pickle.py::test_pickling_batched_graph PASSED         [ 92%]
tests/pytorch/test_pin_memory.py::test_pin_noncontiguous SKIPPED (Ne...) [ 92%]
tests/pytorch/test_pin_memory.py::test_pin_view SKIPPED (Need gpu fo...) [ 92%]
tests/pytorch/test_pin_memory.py::test_unpin_automatically SKIPPED (...) [ 92%]
tests/pytorch/test_pin_memory.py::test_pin_unpin_column SKIPPED (Nee...) [ 92%]
tests/pytorch/test_pin_memory.py::test_pin_empty SKIPPED (Need gpu f...) [ 92%]
tests/pytorch/test_sparse_emb.py::test_multiprocess_sparse_emb_get_set[1] SKIPPED [ 92%]
tests/pytorch/test_sparse_emb.py::test_multiprocess_sparse_emb_get_set[2] SKIPPED [ 92%]
tests/pytorch/test_sparse_emb.py::test_multiprocess_sparse_emb_get_set[3] SKIPPED [ 92%]
tests/pytorch/test_stream.py::test_basics SKIPPED (stream only runs ...) [ 92%]
tests/pytorch/test_stream.py::test_set_get_stream SKIPPED (stream on...) [ 92%]
tests/pytorch/test_stream.py::test_record_stream_ndarray SKIPPED (st...) [ 92%]
tests/pytorch/test_stream.py::test_record_stream_graph_positive SKIPPED  [ 92%]
tests/pytorch/test_stream.py::test_record_stream_graph_negative SKIPPED  [ 93%]
tests/pytorch/test_unified_tensor.py::test_unified_tensor SKIPPED (D...) [ 93%]
tests/pytorch/test_unified_tensor.py::test_multi_gpu_unified_tensor[1] SKIPPED [ 93%]
tests/pytorch/test_unified_tensor.py::test_multi_gpu_unified_tensor[2] SKIPPED [ 93%]
tests/pytorch/mock_sparse/test_diag.py::test_diag[None-val_shape0] PASSED [ 93%]
tests/pytorch/mock_sparse/test_diag.py::test_diag[None-val_shape1] PASSED [ 93%]
tests/pytorch/mock_sparse/test_diag.py::test_diag[mat_shape1-val_shape0] PASSED [ 93%]
tests/pytorch/mock_sparse/test_diag.py::test_diag[mat_shape1-val_shape1] PASSED [ 93%]
tests/pytorch/mock_sparse/test_diag.py::test_diag[mat_shape2-val_shape0] PASSED [ 93%]
tests/pytorch/mock_sparse/test_diag.py::test_diag[mat_shape2-val_shape1] PASSED [ 93%]
tests/pytorch/mock_sparse/test_diag.py::test_identity[None-shape0] PASSED [ 93%]
tests/pytorch/mock_sparse/test_diag.py::test_identity[None-shape1] PASSED [ 93%]
tests/pytorch/mock_sparse/test_diag.py::test_identity[None-shape2] PASSED [ 93%]
tests/pytorch/mock_sparse/test_diag.py::test_identity[2-shape0] PASSED   [ 93%]
tests/pytorch/mock_sparse/test_diag.py::test_identity[2-shape1] PASSED   [ 93%]
tests/pytorch/mock_sparse/test_diag.py::test_identity[2-shape2] PASSED   [ 93%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_sparse[add-dtype0-idtype0] PASSED [ 93%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_sparse[add-dtype0-idtype1] PASSED [ 93%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_sparse[add-dtype1-idtype0] PASSED [ 93%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_sparse[add-dtype1-idtype1] PASSED [ 93%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_sparse[sub-dtype0-idtype0] PASSED [ 93%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_sparse[sub-dtype0-idtype1] PASSED [ 93%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_sparse[sub-dtype1-idtype0] PASSED [ 93%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_sparse[sub-dtype1-idtype1] PASSED [ 93%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_sparse[mul-dtype0-idtype0] PASSED [ 93%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_sparse[mul-dtype0-idtype1] PASSED [ 93%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_sparse[mul-dtype1-idtype0] PASSED [ 93%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_sparse[mul-dtype1-idtype1] PASSED [ 93%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_sparse[truediv-dtype0-idtype0] PASSED [ 93%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_sparse[truediv-dtype0-idtype1] PASSED [ 93%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_sparse[truediv-dtype1-idtype0] PASSED [ 93%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_sparse[truediv-dtype1-idtype1] PASSED [ 93%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_diag[add-dtype0-idtype0] PASSED [ 93%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_diag[add-dtype0-idtype1] PASSED [ 93%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_diag[add-dtype1-idtype0] PASSED [ 93%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_diag[add-dtype1-idtype1] PASSED [ 93%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_diag[sub-dtype0-idtype0] PASSED [ 93%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_diag[sub-dtype0-idtype1] PASSED [ 93%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_diag[sub-dtype1-idtype0] PASSED [ 93%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_diag[sub-dtype1-idtype1] PASSED [ 93%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_diag[mul-dtype0-idtype0] PASSED [ 94%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_diag[mul-dtype0-idtype1] PASSED [ 94%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_diag[mul-dtype1-idtype0] PASSED [ 94%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_diag[mul-dtype1-idtype1] PASSED [ 94%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_diag[truediv-dtype0-idtype0] PASSED [ 94%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_diag[truediv-dtype0-idtype1] PASSED [ 94%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_diag[truediv-dtype1-idtype0] PASSED [ 94%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_diag[truediv-dtype1-idtype1] PASSED [ 94%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_scalar[2-dtype0-idtype0] PASSED [ 94%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_scalar[2-dtype0-idtype1] PASSED [ 94%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_scalar[2-dtype1-idtype0] PASSED [ 94%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_scalar[2-dtype1-idtype1] PASSED [ 94%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_scalar[2.5-dtype0-idtype0] PASSED [ 94%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_scalar[2.5-dtype0-idtype1] PASSED [ 94%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_scalar[2.5-dtype1-idtype0] PASSED [ 94%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_scalar[2.5-dtype1-idtype1] PASSED [ 94%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_scalar_op_sparse[2-dtype0-idtype0] PASSED [ 94%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_scalar_op_sparse[2-dtype0-idtype1] PASSED [ 94%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_scalar_op_sparse[2-dtype1-idtype0] PASSED [ 94%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_scalar_op_sparse[2-dtype1-idtype1] PASSED [ 94%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_scalar_op_sparse[2.5-dtype0-idtype0] PASSED [ 94%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_scalar_op_sparse[2.5-dtype0-idtype1] PASSED [ 94%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_scalar_op_sparse[2.5-dtype1-idtype0] PASSED [ 94%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_scalar_op_sparse[2.5-dtype1-idtype1] PASSED [ 94%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_expose_op PASSED [ 94%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-sum-extra_shape0-col0-row0-None] SKIPPED [ 94%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-sum-extra_shape0-col0-row0-2] SKIPPED [ 94%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-sum-extra_shape0-col0-row1-None] SKIPPED [ 94%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-sum-extra_shape0-col0-row1-2] SKIPPED [ 94%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-sum-extra_shape0-col1-row0-None] SKIPPED [ 94%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-sum-extra_shape0-col1-row0-2] SKIPPED [ 94%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-sum-extra_shape0-col1-row1-None] SKIPPED [ 94%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-sum-extra_shape0-col1-row1-2] SKIPPED [ 94%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-sum-extra_shape1-col0-row0-None] SKIPPED [ 94%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-sum-extra_shape1-col0-row0-2] SKIPPED [ 94%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-sum-extra_shape1-col0-row1-None] SKIPPED [ 94%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-sum-extra_shape1-col0-row1-2] SKIPPED [ 94%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-sum-extra_shape1-col1-row0-None] SKIPPED [ 94%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-sum-extra_shape1-col1-row0-2] SKIPPED [ 94%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-sum-extra_shape1-col1-row1-None] SKIPPED [ 94%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-sum-extra_shape1-col1-row1-2] SKIPPED [ 95%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smax-extra_shape0-col0-row0-None] SKIPPED [ 95%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smax-extra_shape0-col0-row0-2] SKIPPED [ 95%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smax-extra_shape0-col0-row1-None] SKIPPED [ 95%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smax-extra_shape0-col0-row1-2] SKIPPED [ 95%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smax-extra_shape0-col1-row0-None] SKIPPED [ 95%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smax-extra_shape0-col1-row0-2] SKIPPED [ 95%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smax-extra_shape0-col1-row1-None] SKIPPED [ 95%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smax-extra_shape0-col1-row1-2] SKIPPED [ 95%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smax-extra_shape1-col0-row0-None] SKIPPED [ 95%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smax-extra_shape1-col0-row0-2] SKIPPED [ 95%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smax-extra_shape1-col0-row1-None] SKIPPED [ 95%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smax-extra_shape1-col0-row1-2] SKIPPED [ 95%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smax-extra_shape1-col1-row0-None] SKIPPED [ 95%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smax-extra_shape1-col1-row0-2] SKIPPED [ 95%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smax-extra_shape1-col1-row1-None] SKIPPED [ 95%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smax-extra_shape1-col1-row1-2] SKIPPED [ 95%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smin-extra_shape0-col0-row0-None] SKIPPED [ 95%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smin-extra_shape0-col0-row0-2] SKIPPED [ 95%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smin-extra_shape0-col0-row1-None] SKIPPED [ 95%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smin-extra_shape0-col0-row1-2] SKIPPED [ 95%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smin-extra_shape0-col1-row0-None] SKIPPED [ 95%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smin-extra_shape0-col1-row0-2] SKIPPED [ 95%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smin-extra_shape0-col1-row1-None] SKIPPED [ 95%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smin-extra_shape0-col1-row1-2] SKIPPED [ 95%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smin-extra_shape1-col0-row0-None] SKIPPED [ 95%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smin-extra_shape1-col0-row0-2] SKIPPED [ 95%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smin-extra_shape1-col0-row1-None] SKIPPED [ 95%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smin-extra_shape1-col0-row1-2] SKIPPED [ 95%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smin-extra_shape1-col1-row0-None] SKIPPED [ 95%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smin-extra_shape1-col1-row0-2] SKIPPED [ 95%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smin-extra_shape1-col1-row1-None] SKIPPED [ 95%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smin-extra_shape1-col1-row1-2] SKIPPED [ 95%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smean-extra_shape0-col0-row0-None] SKIPPED [ 95%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smean-extra_shape0-col0-row0-2] SKIPPED [ 95%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smean-extra_shape0-col0-row1-None] SKIPPED [ 95%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smean-extra_shape0-col0-row1-2] SKIPPED [ 95%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smean-extra_shape0-col1-row0-None] SKIPPED [ 95%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smean-extra_shape0-col1-row0-2] SKIPPED [ 95%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smean-extra_shape0-col1-row1-None] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smean-extra_shape0-col1-row1-2] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smean-extra_shape1-col0-row0-None] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smean-extra_shape1-col0-row0-2] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smean-extra_shape1-col0-row1-None] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smean-extra_shape1-col0-row1-2] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smean-extra_shape1-col1-row0-None] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smean-extra_shape1-col1-row0-2] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smean-extra_shape1-col1-row1-None] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smean-extra_shape1-col1-row1-2] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-sum-extra_shape0-col0-row0-None] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-sum-extra_shape0-col0-row0-2] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-sum-extra_shape0-col0-row1-None] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-sum-extra_shape0-col0-row1-2] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-sum-extra_shape0-col1-row0-None] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-sum-extra_shape0-col1-row0-2] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-sum-extra_shape0-col1-row1-None] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-sum-extra_shape0-col1-row1-2] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-sum-extra_shape1-col0-row0-None] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-sum-extra_shape1-col0-row0-2] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-sum-extra_shape1-col0-row1-None] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-sum-extra_shape1-col0-row1-2] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-sum-extra_shape1-col1-row0-None] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-sum-extra_shape1-col1-row0-2] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-sum-extra_shape1-col1-row1-None] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-sum-extra_shape1-col1-row1-2] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smax-extra_shape0-col0-row0-None] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smax-extra_shape0-col0-row0-2] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smax-extra_shape0-col0-row1-None] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smax-extra_shape0-col0-row1-2] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smax-extra_shape0-col1-row0-None] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smax-extra_shape0-col1-row0-2] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smax-extra_shape0-col1-row1-None] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smax-extra_shape0-col1-row1-2] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smax-extra_shape1-col0-row0-None] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smax-extra_shape1-col0-row0-2] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smax-extra_shape1-col0-row1-None] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smax-extra_shape1-col0-row1-2] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smax-extra_shape1-col1-row0-None] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smax-extra_shape1-col1-row0-2] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smax-extra_shape1-col1-row1-None] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smax-extra_shape1-col1-row1-2] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smin-extra_shape0-col0-row0-None] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smin-extra_shape0-col0-row0-2] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smin-extra_shape0-col0-row1-None] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smin-extra_shape0-col0-row1-2] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smin-extra_shape0-col1-row0-None] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smin-extra_shape0-col1-row0-2] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smin-extra_shape0-col1-row1-None] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smin-extra_shape0-col1-row1-2] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smin-extra_shape1-col0-row0-None] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smin-extra_shape1-col0-row0-2] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smin-extra_shape1-col0-row1-None] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smin-extra_shape1-col0-row1-2] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smin-extra_shape1-col1-row0-None] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smin-extra_shape1-col1-row0-2] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smin-extra_shape1-col1-row1-None] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smin-extra_shape1-col1-row1-2] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smean-extra_shape0-col0-row0-None] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smean-extra_shape0-col0-row0-2] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smean-extra_shape0-col0-row1-None] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smean-extra_shape0-col0-row1-2] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smean-extra_shape0-col1-row0-None] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smean-extra_shape0-col1-row0-2] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smean-extra_shape0-col1-row1-None] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smean-extra_shape0-col1-row1-2] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smean-extra_shape1-col0-row0-None] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smean-extra_shape1-col0-row0-2] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smean-extra_shape1-col0-row1-None] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smean-extra_shape1-col0-row1-2] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smean-extra_shape1-col1-row0-None] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smean-extra_shape1-col1-row0-2] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smean-extra_shape1-col1-row1-None] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smean-extra_shape1-col1-row1-2] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-sum-extra_shape0-col0-row0-None] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-sum-extra_shape0-col0-row0-2] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-sum-extra_shape0-col0-row1-None] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-sum-extra_shape0-col0-row1-2] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-sum-extra_shape0-col1-row0-None] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-sum-extra_shape0-col1-row0-2] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-sum-extra_shape0-col1-row1-None] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-sum-extra_shape0-col1-row1-2] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-sum-extra_shape1-col0-row0-None] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-sum-extra_shape1-col0-row0-2] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-sum-extra_shape1-col0-row1-None] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-sum-extra_shape1-col0-row1-2] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-sum-extra_shape1-col1-row0-None] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-sum-extra_shape1-col1-row0-2] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-sum-extra_shape1-col1-row1-None] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-sum-extra_shape1-col1-row1-2] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smax-extra_shape0-col0-row0-None] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smax-extra_shape0-col0-row0-2] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smax-extra_shape0-col0-row1-None] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smax-extra_shape0-col0-row1-2] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smax-extra_shape0-col1-row0-None] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smax-extra_shape0-col1-row0-2] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smax-extra_shape0-col1-row1-None] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smax-extra_shape0-col1-row1-2] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smax-extra_shape1-col0-row0-None] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smax-extra_shape1-col0-row0-2] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smax-extra_shape1-col0-row1-None] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smax-extra_shape1-col0-row1-2] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smax-extra_shape1-col1-row0-None] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smax-extra_shape1-col1-row0-2] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smax-extra_shape1-col1-row1-None] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smax-extra_shape1-col1-row1-2] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smin-extra_shape0-col0-row0-None] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smin-extra_shape0-col0-row0-2] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smin-extra_shape0-col0-row1-None] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smin-extra_shape0-col0-row1-2] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smin-extra_shape0-col1-row0-None] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smin-extra_shape0-col1-row0-2] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smin-extra_shape0-col1-row1-None] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smin-extra_shape0-col1-row1-2] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smin-extra_shape1-col0-row0-None] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smin-extra_shape1-col0-row0-2] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smin-extra_shape1-col0-row1-None] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smin-extra_shape1-col0-row1-2] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smin-extra_shape1-col1-row0-None] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smin-extra_shape1-col1-row0-2] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smin-extra_shape1-col1-row1-None] SKIPPED [ 99%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smin-extra_shape1-col1-row1-2] SKIPPED [ 99%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smean-extra_shape0-col0-row0-None] SKIPPED [ 99%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smean-extra_shape0-col0-row0-2] SKIPPED [ 99%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smean-extra_shape0-col0-row1-None] SKIPPED [ 99%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smean-extra_shape0-col0-row1-2] SKIPPED [ 99%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smean-extra_shape0-col1-row0-None] SKIPPED [ 99%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smean-extra_shape0-col1-row0-2] SKIPPED [ 99%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smean-extra_shape0-col1-row1-None] SKIPPED [ 99%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smean-extra_shape0-col1-row1-2] SKIPPED [ 99%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smean-extra_shape1-col0-row0-None] SKIPPED [ 99%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smean-extra_shape1-col0-row0-2] SKIPPED [ 99%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smean-extra_shape1-col0-row1-None] SKIPPED [ 99%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smean-extra_shape1-col0-row1-2] SKIPPED [ 99%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smean-extra_shape1-col1-row0-None] SKIPPED [ 99%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smean-extra_shape1-col1-row0-2] SKIPPED [ 99%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smean-extra_shape1-col1-row1-None] SKIPPED [ 99%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smean-extra_shape1-col1-row1-2] SKIPPED [ 99%]
tests/pytorch/mock_sparse/test_transpose.py::test_diag_matrix_transpose[None-val_shape0] PASSED [ 99%]
tests/pytorch/mock_sparse/test_transpose.py::test_diag_matrix_transpose[None-val_shape1] PASSED [ 99%]
tests/pytorch/mock_sparse/test_transpose.py::test_diag_matrix_transpose[mat_shape1-val_shape0] PASSED [ 99%]
tests/pytorch/mock_sparse/test_transpose.py::test_diag_matrix_transpose[mat_shape1-val_shape1] PASSED [ 99%]
tests/pytorch/mock_sparse/test_transpose.py::test_diag_matrix_transpose[mat_shape2-val_shape0] PASSED [ 99%]
tests/pytorch/mock_sparse/test_transpose.py::test_diag_matrix_transpose[mat_shape2-val_shape1] PASSED [ 99%]
tests/pytorch/mock_sparse/test_transpose.py::test_sparse_matrix_transpose[extra_shape0-col0-row0-None] PASSED [ 99%]
tests/pytorch/mock_sparse/test_transpose.py::test_sparse_matrix_transpose[extra_shape0-col0-row0-2] PASSED [ 99%]
tests/pytorch/mock_sparse/test_transpose.py::test_sparse_matrix_transpose[extra_shape0-col0-row1-None] PASSED [ 99%]
tests/pytorch/mock_sparse/test_transpose.py::test_sparse_matrix_transpose[extra_shape0-col0-row1-2] PASSED [ 99%]
tests/pytorch/mock_sparse/test_transpose.py::test_sparse_matrix_transpose[extra_shape0-col1-row0-None] PASSED [ 99%]
tests/pytorch/mock_sparse/test_transpose.py::test_sparse_matrix_transpose[extra_shape0-col1-row0-2] PASSED [ 99%]
tests/pytorch/mock_sparse/test_transpose.py::test_sparse_matrix_transpose[extra_shape0-col1-row1-None] PASSED [ 99%]
tests/pytorch/mock_sparse/test_transpose.py::test_sparse_matrix_transpose[extra_shape0-col1-row1-2] PASSED [ 99%]
tests/pytorch/mock_sparse/test_transpose.py::test_sparse_matrix_transpose[extra_shape1-col0-row0-None] PASSED [ 99%]
tests/pytorch/mock_sparse/test_transpose.py::test_sparse_matrix_transpose[extra_shape1-col0-row0-2] PASSED [ 99%]
tests/pytorch/mock_sparse/test_transpose.py::test_sparse_matrix_transpose[extra_shape1-col0-row1-None] PASSED [ 99%]
tests/pytorch/mock_sparse/test_transpose.py::test_sparse_matrix_transpose[extra_shape1-col0-row1-2] PASSED [ 99%]
tests/pytorch/mock_sparse/test_transpose.py::test_sparse_matrix_transpose[extra_shape1-col1-row0-None] PASSED [ 99%]
tests/pytorch/mock_sparse/test_transpose.py::test_sparse_matrix_transpose[extra_shape1-col1-row0-2] PASSED [ 99%]
tests/pytorch/mock_sparse/test_transpose.py::test_sparse_matrix_transpose[extra_shape1-col1-row1-None] PASSED [ 99%]
tests/pytorch/mock_sparse/test_transpose.py::test_sparse_matrix_transpose[extra_shape1-col1-row1-2] PASSED [100%]

================================== FAILURES ===================================
________________ test_knn_cpu[True-euclidean-bruteforce-blas] _________________

algorithm = 'bruteforce-blas', dist = 'euclidean', exclude_self = True

    @pytest.mark.parametrize('algorithm', ['bruteforce-blas', 'bruteforce', 'kd-tree'])
    @pytest.mark.parametrize('dist', ['euclidean', 'cosine'])
    @pytest.mark.parametrize('exclude_self', [False, True])
    def test_knn_cpu(algorithm, dist, exclude_self):
>       _test_knn_common(F.cpu(), algorithm, dist, exclude_self)

tests\pytorch\test_geometry.py:159: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
tests\pytorch\test_geometry.py:74: in _test_knn_common
    g = kg(x.view(2, 4, 3), algorithm, dist, exclude_self)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\factory.py:128: in forward
    exclude_self=exclude_self)
python\dgl\transforms\functional.py:248: in knn_graph
    result = remove_self_loop(result)
python\dgl\transforms\functional.py:2032: in remove_self_loop
    new_g = remove_edges(g, self_loop_eids, etype=etype)
python\dgl\transforms\functional.py:1755: in remove_edges
    g.remove_edges(eids, etype=etype, store_ids=store_ids)
python\dgl\heterograph.py:648: in remove_edges
    one_hot_removed_edges, reducer='sum')
python\dgl\ops\segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

op = 'sum'
x = tensor([1., 0., 0., 0., 1., 0., 0., 0., 1., 0., 0., 0., 1., 0., 0., 0., 1., 0.,
        0., 0., 1., 0., 0., 0., 1., 0., 0., 0., 1., 0., 0., 0.])
offsets = tensor([ 0, 16, 32])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:780: TypeError
___________________ test_knn_cpu[True-euclidean-bruteforce] ___________________

algorithm = 'bruteforce', dist = 'euclidean', exclude_self = True

    @pytest.mark.parametrize('algorithm', ['bruteforce-blas', 'bruteforce', 'kd-tree'])
    @pytest.mark.parametrize('dist', ['euclidean', 'cosine'])
    @pytest.mark.parametrize('exclude_self', [False, True])
    def test_knn_cpu(algorithm, dist, exclude_self):
>       _test_knn_common(F.cpu(), algorithm, dist, exclude_self)

tests\pytorch\test_geometry.py:159: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
tests\pytorch\test_geometry.py:74: in _test_knn_common
    g = kg(x.view(2, 4, 3), algorithm, dist, exclude_self)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\factory.py:128: in forward
    exclude_self=exclude_self)
python\dgl\transforms\functional.py:248: in knn_graph
    result = remove_self_loop(result)
python\dgl\transforms\functional.py:2032: in remove_self_loop
    new_g = remove_edges(g, self_loop_eids, etype=etype)
python\dgl\transforms\functional.py:1755: in remove_edges
    g.remove_edges(eids, etype=etype, store_ids=store_ids)
python\dgl\heterograph.py:648: in remove_edges
    one_hot_removed_edges, reducer='sum')
python\dgl\ops\segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

op = 'sum'
x = tensor([0., 0., 0., 1., 0., 0., 0., 1., 0., 0., 1., 0., 0., 0., 0., 1., 0., 0.,
        0., 1., 0., 0., 0., 1., 0., 0., 1., 0., 0., 0., 0., 1.])
offsets = tensor([ 0, 16, 32])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:780: TypeError
____________________ test_knn_cpu[True-euclidean-kd-tree] _____________________

algorithm = 'kd-tree', dist = 'euclidean', exclude_self = True

    @pytest.mark.parametrize('algorithm', ['bruteforce-blas', 'bruteforce', 'kd-tree'])
    @pytest.mark.parametrize('dist', ['euclidean', 'cosine'])
    @pytest.mark.parametrize('exclude_self', [False, True])
    def test_knn_cpu(algorithm, dist, exclude_self):
>       _test_knn_common(F.cpu(), algorithm, dist, exclude_self)

tests\pytorch\test_geometry.py:159: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
tests\pytorch\test_geometry.py:74: in _test_knn_common
    g = kg(x.view(2, 4, 3), algorithm, dist, exclude_self)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\factory.py:128: in forward
    exclude_self=exclude_self)
python\dgl\transforms\functional.py:248: in knn_graph
    result = remove_self_loop(result)
python\dgl\transforms\functional.py:2032: in remove_self_loop
    new_g = remove_edges(g, self_loop_eids, etype=etype)
python\dgl\transforms\functional.py:1755: in remove_edges
    g.remove_edges(eids, etype=etype, store_ids=store_ids)
python\dgl\heterograph.py:648: in remove_edges
    one_hot_removed_edges, reducer='sum')
python\dgl\ops\segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

op = 'sum'
x = tensor([1., 0., 0., 0., 1., 0., 0., 0., 1., 0., 0., 0., 1., 0., 0., 0., 1., 0.,
        0., 0., 1., 0., 0., 0., 1., 0., 0., 0., 1., 0., 0., 0.])
offsets = tensor([ 0, 16, 32])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:780: TypeError
__________________ test_knn_cpu[True-cosine-bruteforce-blas] __________________

algorithm = 'bruteforce-blas', dist = 'cosine', exclude_self = True

    @pytest.mark.parametrize('algorithm', ['bruteforce-blas', 'bruteforce', 'kd-tree'])
    @pytest.mark.parametrize('dist', ['euclidean', 'cosine'])
    @pytest.mark.parametrize('exclude_self', [False, True])
    def test_knn_cpu(algorithm, dist, exclude_self):
>       _test_knn_common(F.cpu(), algorithm, dist, exclude_self)

tests\pytorch\test_geometry.py:159: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
tests\pytorch\test_geometry.py:74: in _test_knn_common
    g = kg(x.view(2, 4, 3), algorithm, dist, exclude_self)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\factory.py:128: in forward
    exclude_self=exclude_self)
python\dgl\transforms\functional.py:248: in knn_graph
    result = remove_self_loop(result)
python\dgl\transforms\functional.py:2032: in remove_self_loop
    new_g = remove_edges(g, self_loop_eids, etype=etype)
python\dgl\transforms\functional.py:1755: in remove_edges
    g.remove_edges(eids, etype=etype, store_ids=store_ids)
python\dgl\heterograph.py:648: in remove_edges
    one_hot_removed_edges, reducer='sum')
python\dgl\ops\segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

op = 'sum'
x = tensor([1., 0., 0., 0., 1., 0., 0., 0., 1., 0., 0., 0., 1., 0., 0., 0., 1., 0.,
        0., 0., 1., 0., 0., 0., 1., 0., 0., 0., 1., 0., 0., 0.])
offsets = tensor([ 0, 16, 32])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:780: TypeError
____________________ test_knn_cpu[True-cosine-bruteforce] _____________________

algorithm = 'bruteforce', dist = 'cosine', exclude_self = True

    @pytest.mark.parametrize('algorithm', ['bruteforce-blas', 'bruteforce', 'kd-tree'])
    @pytest.mark.parametrize('dist', ['euclidean', 'cosine'])
    @pytest.mark.parametrize('exclude_self', [False, True])
    def test_knn_cpu(algorithm, dist, exclude_self):
>       _test_knn_common(F.cpu(), algorithm, dist, exclude_self)

tests\pytorch\test_geometry.py:159: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
tests\pytorch\test_geometry.py:74: in _test_knn_common
    g = kg(x.view(2, 4, 3), algorithm, dist, exclude_self)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\factory.py:128: in forward
    exclude_self=exclude_self)
python\dgl\transforms\functional.py:248: in knn_graph
    result = remove_self_loop(result)
python\dgl\transforms\functional.py:2032: in remove_self_loop
    new_g = remove_edges(g, self_loop_eids, etype=etype)
python\dgl\transforms\functional.py:1755: in remove_edges
    g.remove_edges(eids, etype=etype, store_ids=store_ids)
python\dgl\heterograph.py:648: in remove_edges
    one_hot_removed_edges, reducer='sum')
python\dgl\ops\segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

op = 'sum'
x = tensor([0., 0., 0., 1., 0., 0., 0., 1., 0., 0., 1., 0., 0., 0., 0., 1., 0., 0.,
        0., 1., 0., 0., 0., 1., 0., 0., 1., 0., 0., 0., 0., 1.])
offsets = tensor([ 0, 16, 32])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:780: TypeError
______________________ test_knn_cpu[True-cosine-kd-tree] ______________________

algorithm = 'kd-tree', dist = 'cosine', exclude_self = True

    @pytest.mark.parametrize('algorithm', ['bruteforce-blas', 'bruteforce', 'kd-tree'])
    @pytest.mark.parametrize('dist', ['euclidean', 'cosine'])
    @pytest.mark.parametrize('exclude_self', [False, True])
    def test_knn_cpu(algorithm, dist, exclude_self):
>       _test_knn_common(F.cpu(), algorithm, dist, exclude_self)

tests\pytorch\test_geometry.py:159: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
tests\pytorch\test_geometry.py:74: in _test_knn_common
    g = kg(x.view(2, 4, 3), algorithm, dist, exclude_self)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\factory.py:128: in forward
    exclude_self=exclude_self)
python\dgl\transforms\functional.py:248: in knn_graph
    result = remove_self_loop(result)
python\dgl\transforms\functional.py:2032: in remove_self_loop
    new_g = remove_edges(g, self_loop_eids, etype=etype)
python\dgl\transforms\functional.py:1755: in remove_edges
    g.remove_edges(eids, etype=etype, store_ids=store_ids)
python\dgl\heterograph.py:648: in remove_edges
    one_hot_removed_edges, reducer='sum')
python\dgl\ops\segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

op = 'sum'
x = tensor([1., 0., 0., 0., 1., 0., 0., 0., 1., 0., 0., 0., 1., 0., 0., 0., 1., 0.,
        0., 0., 1., 0., 0., 0., 1., 0., 0., 0., 1., 0., 0., 0.])
offsets = tensor([ 0, 16, 32])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:780: TypeError
_____________________________ test_graph_conv0[1] _____________________________

out_dim = 1

    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv0(out_dim):
        g = dgl.DGLGraph(nx.path_graph(3)).to(F.ctx())
        ctx = F.ctx()
        adj = g.adjacency_matrix(transpose=True, ctx=ctx)
    
        conv = nn.GraphConv(5, out_dim, norm='none', bias=True)
        conv = conv.to(ctx)
        print(conv)
    
        # test pickle
        th.save(conv, tmp_buffer)
    
    
        # test#1: basic
        h0 = F.ones((3, 5))
>       h1 = conv(g, h0)

tests\pytorch\test_nn.py:39: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE15183488>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0.9117],
        [0.9117],
        [0.9117]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
---------------------------- Captured stdout call -----------------------------
GraphConv(in=5, out=1, normalization=none, activation=None)
_____________________________ test_graph_conv0[2] _____________________________

out_dim = 2

    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv0(out_dim):
        g = dgl.DGLGraph(nx.path_graph(3)).to(F.ctx())
        ctx = F.ctx()
        adj = g.adjacency_matrix(transpose=True, ctx=ctx)
    
        conv = nn.GraphConv(5, out_dim, norm='none', bias=True)
        conv = conv.to(ctx)
        print(conv)
    
        # test pickle
        th.save(conv, tmp_buffer)
    
    
        # test#1: basic
        h0 = F.ones((3, 5))
>       h1 = conv(g, h0)

tests\pytorch\test_nn.py:39: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14CEBC08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.2014, -1.7951],
        [-0.2014, -1.7951],
        [-0.2014, -1.7951]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
---------------------------- Captured stdout call -----------------------------
GraphConv(in=5, out=2, normalization=none, activation=None)
________________ test_graph_conv[1-True-True-none-g0-idtype0] _________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'none', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE15183A08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.8223],
        [ 0.7276]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-True-none-g0-idtype1] _________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'none', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE121BACC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.0147],
        [-1.4779]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-True-none-g1-idtype0] _________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
norm = 'none', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0304DD48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.3774],
        [-0.4557],
        [-0.3917],
        [ 0.7560],
        [-1.0727],
        [-1.1545],
        [ 0.1115],
        [ 0.3102],
        [-0.1478],
        [ 2.0557]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-True-none-g1-idtype1] _________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
norm = 'none', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF02F268C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.4357],
        [-0.3045],
        [ 0.5054],
        [ 0.1024],
        [ 0.9517],
        [ 0.2717],
        [ 0.0886],
        [-0.6465],
        [-0.0434],
        [ 1.2168]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-True-none-g2-idtype0] _________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'none', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE151F1AC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.2488],
        [ 0.2138],
        [ 1.1414],
        [ 1.1984],
        [ 0.4325],
        [-1.0267],
        [ 0.7766],
        [-0.4516],
        [-0.4094],
        [ 1.3725]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-True-none-g2-idtype1] _________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'none', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE15335F08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.5254],
        [ 1.0542],
        [ 1.3021],
        [ 1.0440],
        [ 0.1947],
        [-0.8581],
        [-0.5537],
        [-1.6341],
        [ 0.6213],
        [ 1.5097]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-True-none-g3-idtype0] _________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'none', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE15269E08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.5068],
        [-0.3208],
        [ 0.6224],
        [ 0.3713],
        [-1.5798],
        [-0.1363],
        [ 1.8579],
        [ 0.7797],
        [ 0.1811],
        [ 1.0436]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-True-none-g3-idtype1] _________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'none', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1524C8C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.8289],
        [-0.0837],
        [-2.3822],
        [-0.7160],
        [-0.5155],
        [ 0.6484],
        [ 1.1284],
        [-0.1840],
        [-1.2974],
        [-3.3558]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-True-none-g4-idtype0] _________________

idtype = torch.int32
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'none', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01AF0D08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 2.0381],
        [ 0.1247],
        [ 0.5124],
        [-1.7189],
        [-1.7077],
        [ 1.7704],
        [-0.4823],
        [-1.9169],
        [ 0.2390]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-True-none-g4-idtype1] _________________

idtype = torch.int64
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'none', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14CA7448>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.5360],
        [-0.1115],
        [ 1.1599],
        [-0.0047],
        [ 0.3746],
        [-0.8104],
        [ 0.2047],
        [ 1.1318],
        [ 0.9987]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-True-none-g5-idtype0] _________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'none', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01AD7988>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.2075],
        [ 0.1210],
        [-0.6167],
        [ 0.8429],
        [ 0.1643],
        [-0.3728],
        [-0.2843],
        [-1.1507],
        [-0.1210],
        [-0.4389]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-True-none-g5-idtype1] _________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'none', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE15319D08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.5894],
        [-0.0116],
        [-0.8686],
        [ 0.1827],
        [ 0.3602],
        [ 0.5652],
        [-0.2060],
        [ 0.6458],
        [ 0.0123],
        [-0.5896]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-True-none-g6-idtype0] _________________

idtype = torch.int32, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'none', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE15341B08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-2.4553],
        [-2.3494],
        [ 0.5549],
        [-1.5514],
        [ 0.3303],
        [-0.6024]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-True-none-g6-idtype1] _________________

idtype = torch.int64, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'none', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE120DBAC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.3246],
        [ 2.3402],
        [-2.6253],
        [-0.1761],
        [ 0.9155],
        [-0.4409]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-True-none-g7-idtype0] _________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'none', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF02FA2D88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.0979],
        [ 0.7488],
        [-0.0341],
        [-0.6230],
        [-0.3343],
        [ 0.7471],
        [ 0.4251],
        [ 0.0942],
        [ 0.5293],
        [-0.1892]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-True-none-g7-idtype1] _________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'none', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01B08808>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.1872],
        [-0.4040],
        [-1.4210],
        [ 1.7123],
        [-1.0717],
        [ 0.1974],
        [-0.6033],
        [ 0.4072],
        [-0.5829],
        [-0.2204]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-True-both-g0-idtype0] _________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'both', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE151DE108>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.8189],
        [ 1.0706]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-True-both-g0-idtype1] _________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'both', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03048F48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.1778],
        [ 0.2324]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-True-both-g1-idtype0] _________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
norm = 'both', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14CD1B08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.3643],
        [-0.8181],
        [ 1.6384],
        [ 1.1556],
        [ 1.4191],
        [ 3.6268],
        [-1.3846],
        [ 0.6336],
        [-1.5143],
        [-0.3379]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-True-both-g1-idtype1] _________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
norm = 'both', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01A1DA08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.9189],
        [ 1.1011],
        [ 0.1610],
        [ 0.6617],
        [-0.1451],
        [ 0.2839],
        [ 0.1247],
        [ 0.6932],
        [ 0.0778],
        [ 1.3752]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-True-both-g2-idtype0] _________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'both', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE15372A48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.6773],
        [-0.2205],
        [ 1.5709],
        [ 1.3530],
        [-1.6610],
        [-2.5343],
        [ 1.1941],
        [ 0.4840],
        [ 0.7863],
        [-0.7006]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-True-both-g2-idtype1] _________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'both', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE15336948>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.4969],
        [-0.1960],
        [ 0.1433],
        [ 0.6285],
        [ 1.2778],
        [ 0.3453],
        [ 0.8179],
        [ 0.6267],
        [ 0.2244],
        [ 1.0523]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-True-both-g3-idtype0] _________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'both', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE121AF648>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.9824],
        [-0.6180],
        [ 0.9852],
        [ 0.4254],
        [-0.3827],
        [ 0.2962],
        [ 0.2570],
        [-0.0637],
        [ 1.0724],
        [ 0.1239]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-True-both-g3-idtype1] _________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'both', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14B4DA08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-5.9205e-04],
        [ 1.2998e+00],
        [ 1.0132e+00],
        [-2.0308e-01],
        [ 5.5357e-01],
   ...01],
        [-1.3187e+00],
        [ 5.9655e-01],
        [-1.7961e+00],
        [-1.3357e+00]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-True-both-g4-idtype0] _________________

idtype = torch.int32
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'both', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE151CF088>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.6708],
        [ 0.3308],
        [-2.0196],
        [-0.7199],
        [-1.0714],
        [ 0.1628],
        [-1.8664],
        [-0.4424],
        [-0.2624]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-True-both-g4-idtype1] _________________

idtype = torch.int64
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'both', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF030A5A48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.3273],
        [ 0.2117],
        [-0.1677],
        [-1.2674],
        [ 1.1400],
        [ 0.3679],
        [ 1.1484],
        [ 0.6590],
        [ 0.6034]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-True-both-g5-idtype0] _________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'both', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF02F93348>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.0594],
        [ 0.2425],
        [ 0.8313],
        [ 0.1795],
        [-0.8446],
        [-0.6349],
        [-0.5330],
        [ 0.7257],
        [-2.5928],
        [ 0.0765]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-True-both-g5-idtype1] _________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'both', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE151F5608>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.3994],
        [ 0.5239],
        [-0.4658],
        [-0.3767],
        [-1.2460],
        [-0.1192],
        [ 0.6714],
        [ 3.3496],
        [ 0.4079],
        [ 1.1584]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-True-both-g6-idtype0] _________________

idtype = torch.int32, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'both', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01AE1288>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.0530],
        [-0.5401],
        [ 1.2201],
        [-2.5792],
        [ 0.2281],
        [ 0.9623]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-True-both-g6-idtype1] _________________

idtype = torch.int64, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'both', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14B4DBC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.5524],
        [-1.0499],
        [-3.3022],
        [-0.1428],
        [ 0.7136],
        [-2.6226]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-True-both-g7-idtype0] _________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'both', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14C77748>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 2.9467],
        [-0.4187],
        [-0.5070],
        [-0.8936],
        [-0.0709],
        [-2.6504],
        [ 1.0916],
        [ 0.6802],
        [-0.5709],
        [ 1.8184]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-True-both-g7-idtype1] _________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'both', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01AD56C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.2574],
        [ 0.2301],
        [-0.2629],
        [-0.8789],
        [ 2.8606],
        [-0.8293],
        [ 0.7272],
        [ 0.0625],
        [-0.9932],
        [-3.1976]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-True-right-g0-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'right', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF02F86B08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[4.7687],
        [3.1110]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-True-right-g0-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'right', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE151CF348>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[1.5203],
        [0.2693]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-True-right-g1-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
norm = 'right', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE15BE0208>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.2542],
        [-0.6458],
        [-1.2538],
        [ 2.2001],
        [ 1.9990],
        [-0.2682],
        [-2.3272],
        [-1.5221],
        [-0.6127],
        [ 1.5270]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-True-right-g1-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
norm = 'right', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF02F88DC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 2.7344],
        [-0.9932],
        [-1.4484],
        [ 1.4421],
        [-0.6340],
        [ 1.3776],
        [-1.5206],
        [-0.0028],
        [ 0.8921],
        [-1.9648]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-True-right-g2-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'right', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF02F0B088>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.5211],
        [ 1.3808],
        [-1.2648],
        [ 1.2036],
        [-3.3805],
        [ 0.1068],
        [-0.4450],
        [-0.9805],
        [-1.5781],
        [-1.9320]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-True-right-g2-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'right', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1518F3C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.4215],
        [-2.6242],
        [-0.2604],
        [ 0.0904],
        [ 1.3921],
        [-2.3748],
        [ 0.5906],
        [-0.6220],
        [ 0.0819],
        [ 0.4416]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-True-right-g3-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'right', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14C9D508>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.7173],
        [-0.6420],
        [-0.0549],
        [-0.2515],
        [ 0.4849],
        [ 1.5220],
        [ 0.9913],
        [ 0.4167],
        [-1.2981],
        [-0.3690]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-True-right-g3-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'right', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE146E0848>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.2959],
        [-2.0179],
        [-1.1616],
        [ 0.7495],
        [ 0.3636],
        [ 0.2828],
        [ 3.2298],
        [ 0.9372],
        [ 0.6765],
        [-1.0290]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-True-right-g4-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'right', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01A2A508>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.6934],
        [-0.1125],
        [ 1.8192],
        [-2.3530],
        [ 2.2126],
        [-0.5675],
        [-0.8449],
        [-1.0943],
        [-0.1088]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-True-right-g4-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'right', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07644988>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.7966],
        [-1.1554],
        [-0.4580],
        [ 1.2094],
        [-1.8406],
        [-0.1825],
        [-0.1467],
        [ 0.8038],
        [ 0.8419]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-True-right-g5-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'right', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE121C5748>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-4.2496],
        [ 0.0835],
        [ 0.6602],
        [-0.6661],
        [-1.5652],
        [-0.8721],
        [ 1.6695],
        [-0.3479],
        [-0.2666],
        [-0.5852]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-True-right-g5-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'right', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE146E56C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.3945],
        [-1.3850],
        [ 0.3137],
        [-0.5863],
        [-1.7720],
        [-1.2231],
        [ 2.0397],
        [ 0.0272],
        [ 2.1933],
        [-1.4832]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-True-right-g6-idtype0] ________________

idtype = torch.int32, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'right', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01A60C08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.9290],
        [ 1.1920],
        [ 0.9226],
        [ 0.6478],
        [-1.2884],
        [-0.6153]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':

            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-True-right-g6-idtype1] ________________

idtype = torch.int64, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'right', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF02FCDC88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.0325],
        [-2.2404],
        [ 0.2782],
        [ 0.2156],
        [-1.1665],
        [ 1.7703]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-True-right-g7-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'right', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07628DC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.1675],
        [-0.5347],
        [ 0.5502],
        [-2.0449],
        [ 0.7123],
        [ 1.2910],
        [ 1.5387],
        [ 1.6971],
        [ 1.3014],
        [ 0.0460]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-True-right-g7-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'right', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01AD5748>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 2.3587e+00],
        [-1.3794e-01],
        [ 8.6305e-01],
        [-8.3138e-01],
        [ 1.2370e-01],
   ...00],
        [ 1.6678e-01],
        [ 8.9375e-01],
        [ 2.7686e-04],
        [ 8.4640e-01]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-True-left-g0-idtype0] _________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'left', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14C7F988>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.2470],
        [-0.2297]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-True-left-g0-idtype1] _________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'left', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01B08E48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.0680],
        [ 0.0469]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-True-left-g1-idtype0] _________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
norm = 'left', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01AC9A08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.4420],
        [-0.3754],
        [ 0.6135],
        [-0.6264],
        [-1.8864],
        [-1.2964],
        [ 0.1367],
        [ 0.7154],
        [-0.1115],
        [-0.1078]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-True-left-g1-idtype1] _________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
norm = 'left', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1438E648>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.8868],
        [-0.8612],
        [ 0.3331],
        [-0.4362],
        [-1.7782],
        [ 0.3539],
        [-0.8969],
        [ 1.0638],
        [ 1.8934],
        [ 0.8213]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-True-left-g2-idtype0] _________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'left', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1513BE48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.1859],
        [-0.9136],
        [-0.4591],
        [-0.4074],
        [-0.2436],
        [-0.9790],
        [ 0.6441],
        [ 2.3218],
        [ 1.6560],
        [-0.8347]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-True-left-g2-idtype1] _________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'left', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1533D348>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.3226],
        [ 0.3242],
        [ 0.3617],
        [-0.0668],
        [ 1.5841],
        [-0.8224],
        [-0.0996],
        [ 0.0072],
        [ 1.0775],
        [ 0.9886]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-True-left-g3-idtype0] _________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'left', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1518FAC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.2247],
        [-0.6381],
        [-0.2281],
        [-0.2151],
        [ 3.4745],
        [ 0.8205],
        [-0.1342],
        [-3.0598],
        [-2.3273],
        [ 1.5261]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-True-left-g3-idtype1] _________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'left', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14B528C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.0523],
        [ 0.5282],
        [-0.2292],
        [-0.1588],
        [ 1.7005],
        [-0.1604],
        [-0.7490],
        [ 0.2056],
        [ 0.5445],
        [-2.7731]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-True-left-g4-idtype0] _________________

idtype = torch.int32
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'left', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01AE1788>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.2416],
        [ 0.0594],
        [ 0.4015],
        [-2.1429],
        [ 2.9310],
        [-0.2389],
        [-0.6514],
        [-0.6388],
        [-0.4177]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-True-left-g4-idtype1] _________________

idtype = torch.int64
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'left', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03F02B08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.5803],
        [ 0.4159],
        [-0.3154],
        [-0.2973],
        [-0.7268],
        [-0.4339],
        [ 0.1537],
        [ 0.2137],
        [ 0.3777]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-True-left-g5-idtype0] _________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'left', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE144E2D88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.2459],
        [-0.6376],
        [ 0.4816],
        [ 0.0243],
        [ 0.5500],
        [-0.0834],
        [-0.1993],
        [-0.1725],
        [ 1.4527],
        [ 0.4886]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-True-left-g5-idtype1] _________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'left', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03018A48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.2644],
        [-0.1707],
        [ 0.1003],
        [-0.3199],
        [ 0.2597],
        [-3.6168],
        [-0.0572],
        [-0.3321],
        [ 1.6833],
        [-1.0310]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-True-left-g6-idtype0] _________________

idtype = torch.int32, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'left', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01A2A808>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.3125],
        [ 0.3997],
        [-0.4831],
        [ 0.8024],
        [ 0.3698],
        [ 0.2594]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-True-left-g6-idtype1] _________________

idtype = torch.int64, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'left', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE144C2388>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.1268],
        [ 3.4421],
        [-0.2081],
        [-0.2640],
        [-0.3184],
        [ 0.9954]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-True-left-g7-idtype0] _________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'left', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01A1DF08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.1241],
        [ 0.1502],
        [ 0.0394],
        [ 0.1257],
        [ 1.3690],
        [-1.8282],
        [-0.3530],
        [ 0.6621],
        [-1.3240],
        [ 1.2551]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-True-left-g7-idtype1] _________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'left', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14CE3FC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.0728],
        [ 0.5581],
        [ 0.1960],
        [-0.3930],
        [-1.8265],
        [ 1.2103],
        [ 0.6508],
        [ 1.0136],
        [ 1.5459],
        [-0.0596]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-False-none-g0-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'none', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE121C59C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[4.6588],
        [3.4320]]), rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-False-none-g0-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'none', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF02F3C688>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-3.2732],
        [ 0.3205]]), rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-False-none-g1-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
norm = 'none', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01AF0748>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.4668],
        [ 0.4833],
        [-0.1228],
        [-1.1125],
        [ 0.4774],
        [ 1.2603],
        [-0.0270],
        [-0.0116],
        [-0.7850],
        [ 0.1690]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-False-none-g1-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
norm = 'none', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE15337C88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-3.4981],
        [ 5.6540],
        [-1.7288],
        [ 0.7222],
        [-0.7087],
        [-3.9235],
        [-1.2105],
        [ 1.0120],
        [ 1.3986],
        [ 1.8229]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-False-none-g2-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'none', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03EE8108>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.2975],
        [-0.7394],
        [-0.7387],
        [-0.1274],
        [-5.0020],
        [ 3.8692],
        [ 1.2384],
        [ 1.6330],
        [-0.3165],
        [ 0.1295]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-False-none-g2-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'none', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE144B2648>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.6597],
        [ 0.8131],
        [-2.0727],
        [ 1.7635],
        [ 0.2682],
        [ 4.6444],
        [-2.2695],
        [ 1.3582],
        [-3.3224],
        [ 3.0026]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-False-none-g3-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'none', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14B6D248>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.4822],
        [ 1.9899],
        [ 3.4946],
        [-3.7145],
        [-2.2030],
        [ 5.1317],
        [-0.2879],
        [ 4.0455],
        [ 0.7684],
        [-0.3467]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-False-none-g3-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'none', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14C98388>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.8591],
        [ 1.2500],
        [-2.9458],
        [-2.3425],
        [-2.0940],
        [ 3.6825],
        [-4.6074],
        [ 2.2862],
        [ 1.5371],
        [ 1.5451]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-False-none-g4-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'none', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE120DCC88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.5715],
        [ 0.2835],
        [ 0.9814],
        [ 0.9026],
        [-1.8968],
        [-0.6352],
        [-0.1040],
        [ 1.8033],
        [ 0.2336]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-False-none-g4-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'none', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01A35A08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.1346],
        [-6.8794],
        [-3.3120],
        [-0.0183],
        [ 1.2790],
        [ 2.2828],
        [-3.6938],
        [ 5.7756],
        [-2.4437]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-False-none-g5-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'none', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14C92488>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 2.6440],
        [-1.2724],
        [-2.4345],
        [ 2.9146],
        [ 1.9439],
        [-1.3245],
        [-3.2645],
        [-0.8071],
        [-4.2782],
        [ 1.3942]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-False-none-g5-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'none', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14C77888>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.5986],
        [ 3.4801],
        [ 0.3119],
        [-1.7241],
        [-3.2270],
        [ 0.5810],
        [ 3.2933],
        [-0.3457],
        [ 1.1472],
        [-0.4184]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-False-none-g6-idtype0] ________________

idtype = torch.int32, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'none', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14C42A48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 3.9281],
        [-4.3147],
        [-3.3764],
        [ 1.2308],
        [ 0.5673],
        [-1.4180]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-False-none-g6-idtype1] ________________

idtype = torch.int64, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'none', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14C30988>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.0560],
        [ 0.8844],
        [-4.5814],
        [ 0.8390],
        [ 2.5591],
        [-1.5845]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-False-none-g7-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'none', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01AE7D08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 3.0857],
        [-1.7225],
        [ 1.6712],
        [-2.2457],
        [-1.2002],
        [-0.3519],
        [ 2.4278],
        [-4.1150],
        [ 0.8644],
        [-1.3881]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-False-none-g7-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'none', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE12032A88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.7220],
        [-2.0146],
        [ 2.4561],
        [-0.8713],
        [-2.1352],
        [-1.2516],
        [-1.9972],
        [ 0.7236],
        [ 1.9621],
        [ 0.5596]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-False-both-g0-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'both', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF02FAA8C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0.2791],
        [0.5825]]), rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-False-both-g0-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'both', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF02F1E7C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.0964],
        [ 2.9500]]), rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-False-both-g1-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
norm = 'both', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE15363E48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.0583],
        [ 2.3470],
        [-0.7866],
        [ 1.3568],
        [ 0.8919],
        [-1.4300],
        [ 0.5106],
        [-0.4744],
        [ 1.0346],
        [-2.2374]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-False-both-g1-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
norm = 'both', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF076A1A08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-3.3297],
        [-2.7886],
        [-0.5809],
        [-1.9665],
        [ 2.5598],
        [ 2.1481],
        [-4.3904],
        [-2.7997],
        [-0.1768],
        [-4.1974]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-False-both-g2-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'both', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE151D5588>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.4439],
        [ 1.3954],
        [ 0.1539],
        [ 1.2657],
        [ 0.8821],
        [-0.5206],
        [-0.3020],
        [ 1.2846],
        [ 1.8222],
        [ 1.1358]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-False-both-g2-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'both', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE15337FC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.4231],
        [ 0.6381],
        [ 0.3237],
        [ 0.4363],
        [ 0.2932],
        [ 1.1918],
        [ 0.0766],
        [ 1.7217],
        [-0.7948],
        [ 1.6200]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-False-both-g3-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'both', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE121AD648>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.0137],
        [ 0.1824],
        [-0.2498],
        [-0.8405],
        [ 0.3014],
        [ 1.3871],
        [ 0.1037],
        [-2.2916],
        [ 0.9056],
        [ 1.1520]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-False-both-g3-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'both', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03036048>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.3257],
        [ 1.3336],
        [ 1.2931],
        [-1.7025],
        [-3.4846],
        [ 0.6389],
        [-0.3032],
        [-2.0030],
        [ 2.8290],
        [ 2.8781]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-False-both-g4-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'both', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14B6D1C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 2.7021],
        [-0.1376],
        [ 3.4731],
        [ 2.9127],
        [-2.6180],
        [ 0.1374],
        [-0.3608],
        [ 0.6077],
        [-2.8880]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-False-both-g4-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'both', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03EFE2C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.1994],
        [-0.9301],
        [ 0.4283],
        [-1.5556],
        [ 1.1317],
        [ 2.6270],
        [-2.5347],
        [-1.1605],
        [ 1.2988]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-False-both-g5-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'both', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07640748>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.4245],
        [ 0.9392],
        [-1.6195],
        [ 0.1446],
        [ 4.7796],
        [-1.9656],
        [ 1.5621],
        [-7.6928],
        [-4.4500],
        [ 4.7836]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-False-both-g5-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'both', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF02F3C588>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.1779],
        [-0.3265],
        [-0.4248],
        [ 2.1647],
        [ 1.3665],
        [-0.5303],
        [ 1.1173],
        [ 3.4854],
        [-4.3648],
        [-3.4345]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-False-both-g6-idtype0] ________________

idtype = torch.int32, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'both', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE15190FC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.8516],
        [-0.4969],
        [ 0.5359],
        [-0.9045],
        [ 0.0444],
        [-0.1248]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-False-both-g6-idtype1] ________________

idtype = torch.int64, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'both', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14C34688>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 3.4533],
        [ 2.8339],
        [ 1.0259],
        [-0.5919],
        [-0.6034],
        [ 0.9965]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-False-both-g7-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'both', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF02F6EC48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 3.1756],
        [ 2.7523],
        [-0.8097],
        [-0.8420],
        [-6.4366],
        [ 2.2159],
        [ 6.4790],
        [ 1.0936],
        [-3.4727],
        [ 5.5505]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-False-both-g7-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'both', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE15379A08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.3269],
        [-0.5033],
        [ 0.1339],
        [ 0.8003],
        [ 1.7074],
        [-0.3194],
        [-0.5959],
        [ 0.0968],
        [ 0.7704],
        [-1.6596]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-True-False-right-g0-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'right', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14CC43C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 5.9925],
        [-3.5196]]), rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-True-False-right-g0-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'right', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE121A7048>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-2.1799],
        [ 0.2628]]), rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-True-False-right-g1-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
norm = 'right', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14C775C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 3.3754],
        [ 3.0911],
        [ 0.1195],
        [ 0.1073],
        [ 1.6961],
        [-1.3906],
        [ 1.4328],
        [ 2.7308],
        [-2.5521],
        [-1.4171]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-True-False-right-g1-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
norm = 'right', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF030032C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 4.1112],
        [ 0.1671],
        [-0.8932],
        [-2.9881],
        [ 1.7521],
        [ 2.3557],
        [ 2.5294],
        [-3.1311],
        [-3.9932],
        [-0.0674]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-True-False-right-g2-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'right', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1533FF48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.2047],
        [ 2.4570],
        [-5.2055],
        [-0.7385],
        [ 3.3523],
        [ 0.4935],
        [-0.8638],
        [ 0.5782],
        [ 2.3556],
        [ 0.8004]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-True-False-right-g2-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'right', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03EA8608>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.7345],
        [-0.4289],
        [-2.5716],
        [-3.3592],
        [-0.6372],
        [ 0.7610],
        [ 0.3459],
        [ 2.3952],
        [ 0.6428],
        [ 0.2351]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-True-False-right-g3-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'right', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14CBCB48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.6866],
        [-0.4350],
        [-0.6574],
        [-1.2182],
        [-2.8592],
        [ 1.7199],
        [-3.1307],
        [-2.7521],
        [ 1.0095],
        [-2.5968]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-True-False-right-g3-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'right', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE15281A08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 2.1535],
        [ 1.9413],
        [-1.2240],
        [ 0.2276],
        [ 6.5071],
        [-3.7753],
        [-2.8655],
        [ 7.9541],
        [-4.9124],
        [ 0.4349]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-True-False-right-g4-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'right', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03F45C88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.2137],
        [-1.1257],
        [ 1.1001],
        [ 0.7518],
        [ 0.0788],
        [-0.8680],
        [ 0.3136],
        [-0.0445],
        [-0.2843]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-True-False-right-g4-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'right', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1526FE88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 3.1516],
        [-1.3416],
        [-1.4448],
        [ 1.5543],
        [ 1.1147],
        [-0.1045],
        [-5.0929],
        [-0.9904],
        [ 2.3762]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-True-False-right-g5-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'right', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF02FDB7C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.5168],
        [ 0.7626],
        [ 0.2875],
        [-1.2311],
        [-1.3024],
        [-0.3283],
        [ 4.3824],
        [-2.5153],
        [-4.1696],
        [-3.4523]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-True-False-right-g5-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'right', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07641848>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.8217],
        [-5.4374],
        [-1.0410],
        [-4.4643],
        [-1.6742],
        [ 0.2210],
        [-2.4247],
        [-3.1442],
        [-3.3403],
        [-3.0721]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-True-False-right-g6-idtype0] ________________

idtype = torch.int32, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'right', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF02F88AC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.7206],
        [ 2.2215],
        [ 1.1442],
        [-1.4627],
        [ 1.3266],
        [-1.4311]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-True-False-right-g6-idtype1] ________________

idtype = torch.int64, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'right', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14C56288>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 3.4069],
        [ 4.5611],
        [-2.7442],
        [ 0.4791],
        [-2.0195],
        [ 4.0099]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-True-False-right-g7-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'right', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14CCFA48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.0743],
        [ 2.1593],
        [ 3.2843],
        [-2.4149],
        [ 0.5721],
        [ 2.2278],
        [ 0.6733],
        [-1.5654],
        [-0.3396],
        [ 3.3128]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-True-False-right-g7-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'right', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE15181A08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.4400],
        [ 1.5243],
        [-0.8774],
        [ 1.3377],
        [-2.6963],
        [ 0.1906],
        [ 0.0647],
        [ 1.0922],
        [-2.9392],
        [-0.5387]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-False-left-g0-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'left', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03E90708>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.8563],
        [-0.3228]]), rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-False-left-g0-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'left', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0768ACC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.3931],
        [-0.6176]]), rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-False-left-g1-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
norm = 'left', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0303C948>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.4678],
        [ 0.5233],
        [ 0.0141],
        [-0.7330],
        [-1.8685],
        [-0.1911],
        [ 0.3722],
        [ 1.8142],
        [-1.1401],
        [-2.4681]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-False-left-g1-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
norm = 'left', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1524C488>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.3560],
        [-1.1815],
        [-0.1181],
        [-1.4934],
        [ 1.6957],
        [-0.6899],
        [ 0.9528],
        [-2.7247],
        [-4.4873],
        [ 2.6908]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-False-left-g2-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'left', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03F166C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.8031],
        [ 1.2550],
        [ 0.6952],
        [-0.1406],
        [ 2.6501],
        [ 0.6183],
        [-2.5808],
        [ 1.0571],
        [-4.1204],
        [-0.1672]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-False-left-g2-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'left', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14CA7108>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.7937],
        [ 0.5374],
        [ 0.2590],
        [ 0.9345],
        [-2.8654],
        [-0.4766],
        [-0.4131],
        [ 1.3669],
        [ 0.6023],
        [ 2.3050]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-False-left-g3-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,

      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'left', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF02FF2A08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.0703],
        [-0.4418],
        [-0.4431],
        [ 0.0808],
        [-1.0919],
        [-1.1628],
        [ 0.5067],
        [ 1.1246],
        [ 0.3915],
        [-0.2787]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-False-left-g3-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'left', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0767CF88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.6062],
        [-0.9527],
        [ 0.0796],
        [-0.5014],
        [ 1.2667],
        [-2.6195],
        [ 0.2569],
        [ 2.0142],
        [-3.8119],
        [-0.4498]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-False-left-g4-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'left', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF02FF2588>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.3015],
        [ 0.1633],
        [ 1.4837],
        [ 0.4720],
        [ 0.9793],
        [-0.5012],
        [ 0.7922],
        [ 2.1291],
        [ 1.6809]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-False-left-g4-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'left', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14C6F648>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.2403],
        [ 0.0828],
        [ 0.3444],
        [ 1.4551],
        [-0.0947],
        [-0.3771],
        [-0.9531],
        [ 0.3497],
        [ 0.5257]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-False-left-g5-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'left', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0760D688>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.5732],
        [-1.0499],
        [ 1.1646],
        [-0.4925],
        [ 0.8523],
        [ 0.8395],
        [-0.3943],
        [ 0.8308],
        [ 4.2855],
        [ 4.0427]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-False-left-g5-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'left', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14B67D48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.4024],
        [-0.1478],
        [ 0.3533],
        [ 0.4127],
        [-0.6223],
        [-0.6949],
        [ 0.4662],
        [ 0.4079],
        [-2.7471],
        [-2.6547]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-False-left-g6-idtype0] ________________

idtype = torch.int32, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'left', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07643B08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.9359],
        [-2.4821],
        [-1.5221],
        [ 0.1265],
        [-1.3538],
        [-0.4199]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-False-left-g6-idtype1] ________________

idtype = torch.int64, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'left', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03042648>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.3507],
        [ 1.2127],
        [ 2.0942],
        [-5.9666],
        [ 4.4595],
        [-3.6851]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-False-left-g7-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'left', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1213FF48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.2757],
        [-0.9710],
        [ 0.1670],
        [ 0.1640],
        [ 1.5066],
        [-3.3212],
        [-0.2272],
        [-1.9565],
        [ 0.7175],
        [ 1.1467]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-True-False-left-g7-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'left', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01A77F88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.6261],
        [ 0.4137],
        [-0.5696],
        [ 1.1727],
        [-1.2565],
        [ 3.0911],
        [ 0.6065],
        [-2.1797],
        [ 1.2948],
        [-0.0312]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-False-True-none-g0-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'none', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03F08688>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.2741],
        [-1.8283]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-False-True-none-g0-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'none', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03010B48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.8816],
        [-0.2937]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-False-True-none-g1-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
norm = 'none', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0304B888>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.1645],
        [ 0.0848],
        [ 0.2466],
        [-1.3641],
        [ 0.1756],
        [ 0.1557],
        [-4.0597],
        [-0.6340],
        [-0.3387],
        [-0.1986]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-False-True-none-g1-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
norm = 'none', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1537A548>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.4090],
        [ 0.0678],
        [-2.1399],
        [ 0.2398],
        [-0.7762],
        [-1.5344],
        [-0.7439],
        [ 0.1899],
        [ 0.8835],
        [-0.6220]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-False-True-none-g2-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'none', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07607708>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.0891],
        [ 2.2365],
        [-0.4261],
        [-1.1589],
        [-0.1985],
        [ 2.1459],
        [-4.7127],
        [-1.2597],
        [ 2.1622],
        [-0.0168]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-False-True-none-g2-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'none', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF02FF0FC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.1519],
        [ 0.9184],
        [-2.2148],
        [-1.3970],
        [-1.8682],
        [-1.0715],
        [-1.4028],
        [-0.4557],
        [-0.6026],
        [ 0.1913]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-False-True-none-g3-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'none', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE151DE288>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-2.7931],
        [ 2.6273],
        [ 0.4962],
        [ 1.8085],
        [ 1.7751],
        [-0.1606],
        [-1.8477],
        [ 1.4651],
        [ 1.4120],
        [ 0.3117]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-False-True-none-g3-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'none', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1525A548>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.5574],
        [-0.0413],
        [-0.0748],
        [ 0.5882],
        [-2.0757],
        [-0.3730],
        [-1.0024],
        [ 1.0030],
        [ 1.4917],
        [ 1.6682]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-False-True-none-g4-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'none', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF02F56948>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.1926],
        [-0.4264],
        [ 1.0866],
        [ 1.5811],
        [ 0.1066],
        [ 0.7125],
        [ 1.7494],
        [ 1.3982],
        [-0.8108]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-False-True-none-g4-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'none', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1526F5C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.5678],
        [ 3.2388],
        [ 0.8013],
        [-0.8872],
        [-0.4117],
        [-0.1659],
        [-0.1865],
        [-1.3458],
        [-0.5751]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-False-True-none-g5-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'none', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03E90308>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.3404],
        [-2.0442],
        [ 1.3129],
        [ 0.6260],
        [ 0.1636],
        [ 0.8803],
        [-0.0936],
        [-0.3224],
        [ 0.4327],
        [-0.2761]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-False-True-none-g5-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'none', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14B4BDC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.3253],
        [ 1.1468],
        [ 0.7463],
        [ 1.6034],
        [-0.1102],
        [-0.7595],
        [-1.2077],
        [-2.6908],
        [ 0.2797],
        [ 0.7313]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-False-True-none-g6-idtype0] ________________

idtype = torch.int32, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'none', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF030A5388>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.1022],
        [-0.6165],
        [-0.1433],
        [-2.4645],
        [ 0.1254],
        [-2.7441]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-False-True-none-g6-idtype1] ________________

idtype = torch.int64, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'none', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE152815C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-2.0103],
        [ 1.5464],
        [-1.5818],
        [ 1.9563],
        [ 1.3320],
        [ 0.8648]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-False-True-none-g7-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'none', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07706A48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.2352],
        [-1.8008],
        [-1.0306],
        [ 1.7420],
        [-0.6986],
        [ 0.7388],
        [-1.4930],
        [-1.0122],
        [-0.0268],
        [-1.4054]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-False-True-none-g7-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'none', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE15336108>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.6487],
        [ 0.0650],
        [ 0.5975],
        [ 1.2566],
        [ 2.1387],
        [ 0.2407],
        [-0.9502],
        [ 0.2135],
        [ 1.4931],
        [ 1.1156]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-False-True-both-g0-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'both', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE12067CC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0.1360],
        [0.0013]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-False-True-both-g0-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'both', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF076F6B08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.1411],
        [-0.0186]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-False-True-both-g1-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
norm = 'both', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14C45488>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.5606],
        [ 0.9242],
        [-0.0739],
        [ 0.6071],
        [ 0.4358],
        [-0.2436],
        [ 0.3397],
        [-1.0675],
        [-0.2124],
        [ 0.1771]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-False-True-both-g1-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
norm = 'both', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF02F8E9C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.8423],
        [-0.3107],
        [ 0.3239],
        [-0.4577],
        [ 2.7726],
        [ 0.2495],
        [-1.5075],
        [-1.5133],
        [ 0.9681],
        [-3.7836]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-False-True-both-g2-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'both', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03EB4988>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.6221],
        [ 0.4521],
        [-0.5972],
        [ 0.6848],
        [-0.4029],
        [ 0.5742],
        [ 0.4854],
        [-0.2073],
        [ 1.4666],
        [ 0.3348]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-False-True-both-g2-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'both', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE15337388>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.1209],
        [-0.3748],
        [ 0.7825],
        [-2.0360],
        [-0.8154],
        [-0.1998],
        [-0.1477],
        [-1.5384],
        [-0.2266],
        [-4.0868]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-False-True-both-g3-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'both', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF076FE848>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.0512],
        [-0.3121],
        [ 1.0154],
        [ 0.0440],
        [-0.4233],
        [-2.1993],
        [-0.6919],
        [-1.9257],
        [-0.1916],
        [ 0.7977]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-False-True-both-g3-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'both', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01A7F448>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.2518],
        [ 0.0537],
        [ 1.0431],
        [-0.3329],
        [-1.2088],
        [ 1.5997],
        [ 2.7069],
        [-0.8020],
        [-1.7164],
        [-0.2705]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-False-True-both-g4-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'both', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF076A7948>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.3441],
        [ 1.0776],
        [-0.0985],
        [ 0.3730],
        [ 1.2595],
        [ 0.6266],
        [-0.9664],
        [-0.7934],
        [ 0.8226]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-False-True-both-g4-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'both', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE152956C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.9301],
        [ 0.0983],
        [-0.4169],
        [ 1.0779],
        [ 0.6027],
        [-0.2937],
        [ 1.0076],
        [-0.0518],
        [ 0.8604]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-False-True-both-g5-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'both', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07617208>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.4837],
        [-1.4548],
        [ 0.1783],
        [-0.9259],
        [-1.5109],
        [ 0.3066],
        [ 0.2705],
        [-1.8205],
        [ 0.0303],
        [-1.4813]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-False-True-both-g5-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'both', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01A52548>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.0921],
        [ 0.8762],
        [ 0.2386],
        [ 0.8030],
        [ 1.6006],
        [ 0.1131],
        [ 0.4768],
        [-2.3033],
        [ 1.8319],
        [ 0.9503]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-False-True-both-g6-idtype0] ________________

idtype = torch.int32, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'both', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03E9E988>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.2742],
        [ 0.1310],
        [ 0.7362],
        [-1.2860],
        [-0.7826],
        [ 0.6269]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-False-True-both-g6-idtype1] ________________

idtype = torch.int64, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'both', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE15295508>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.6872],
        [ 0.1482],
        [ 0.0115],
        [ 0.1228],
        [ 0.9463],
        [-1.5334]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-False-True-both-g7-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'both', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03F75DC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.4085],
        [ 0.1970],
        [ 0.3200],
        [ 0.2534],
        [ 0.9936],
        [ 0.2501],
        [-0.2540],
        [-0.0741],
        [-0.4694],
        [ 1.8145]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-False-True-both-g7-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'both', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1513BE08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.6366],
        [-0.5333],
        [-0.5580],
        [-1.0957],
        [-1.1120],
        [ 0.5713],
        [ 1.6415],
        [ 0.5183],
        [-0.9462],
        [ 2.0458]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-True-right-g0-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'right', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01A34548>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.6191],
        [-0.4900]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-True-right-g0-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'right', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01AB12C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.1299],
        [-1.3853]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-True-right-g1-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
norm = 'right', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF030A5C48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-2.0281],
        [ 1.4903],
        [ 0.2889],
        [ 0.6064],
        [-0.8902],
        [ 1.4737],
        [-1.6110],
        [ 0.5858],
        [ 0.7845],
        [-0.6048]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-True-right-g1-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
norm = 'right', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF076AF488>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-2.1182],
        [ 1.4554],
        [-0.6427],
        [ 0.2496],
        [ 0.8410],
        [ 1.0551],
        [-2.5685],
        [ 2.0532],
        [-2.9279],
        [ 1.4171]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-True-right-g2-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'right', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07632C88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[1.6884],
        [0.8493],
        [0.5604],
        [1.3953],
        [0.8591],
        [0.1608],
        [0.0819],
        [2.0578],
        [0.2067],
        [0.4489]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-True-right-g2-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'right', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0302BF08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.2370],
        [-2.1819],
        [-0.8610],
        [-1.1385],
        [-0.0596],
        [ 0.5057],
        [-0.0055],
        [-1.3397],
        [-1.2025],
        [ 2.7667]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-True-right-g3-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'right', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE151BA1C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.3413],
        [ 0.5122],
        [ 1.0920],
        [ 1.6761],
        [-1.7693],
        [-1.0712],
        [-1.4349],
        [ 0.5785],
        [ 0.2275],
        [ 0.5018]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-True-right-g3-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'right', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF02F86C48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.8762],
        [ 0.5858],
        [-0.8768],
        [ 0.8515],
        [ 1.3028],
        [ 0.4221],
        [-1.0545],
        [-1.7212],
        [ 0.1969],
        [ 0.5078]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-True-right-g4-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'right', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF076D0B08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.5942],
        [-0.9213],
        [-1.8538],
        [-0.1059],
        [ 0.9854],
        [ 0.0384],
        [ 1.3984],
        [-2.0510],
        [-0.3209]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-True-right-g4-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'right', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14B6DE88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.1426],
        [-0.1658],
        [-1.1309],
        [ 0.5738],
        [-0.4126],
        [-0.0160],
        [ 2.0481],
        [-1.2155],
        [ 0.4434]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-True-right-g5-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'right', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03EDF388>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.8335],
        [-1.3160],
        [ 1.6109],
        [-0.3918],
        [-0.8440],
        [-2.2232],
        [-0.3663],
        [ 0.5344],
        [ 0.0066],
        [-0.3780]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-True-right-g5-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'right', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE121ADFC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.0456],
        [ 0.4955],
        [ 0.0271],
        [-0.1516],
        [ 0.3761],
        [-0.1619],
        [ 0.0290],
        [-0.2023],
        [-0.4218],
        [ 0.1164]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-True-right-g6-idtype0] ________________

idtype = torch.int32, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'right', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03E90288>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.1038],
        [-2.3259],
        [ 0.3958],
        [ 1.0196],
        [-0.6899],
        [ 0.8837]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-True-right-g6-idtype1] ________________

idtype = torch.int64, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'right', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE15288288>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.6508],
        [-0.5193],
        [-1.0715],
        [-0.8391],
        [ 0.2794],
        [-0.6693]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-True-right-g7-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'right', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF076FC4C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.6950],
        [-0.2065],
        [-0.4509],
        [-0.6798],
        [ 0.5150],
        [-0.3912],
        [ 0.6206],
        [-0.7494],
        [ 0.2922],
        [-0.3657]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-True-right-g7-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'right', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE15155148>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.1652],
        [-1.5207],
        [-0.6436],
        [-0.0827],
        [-2.0013],
        [ 0.5217],
        [-0.6818],
        [-0.2647],
        [ 1.8513],
        [-2.0264]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-False-True-left-g0-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'left', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03F16448>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.3446],
        [-0.5044]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-False-True-left-g0-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'left', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01A43788>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.0838],
        [-0.2615]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-False-True-left-g1-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
norm = 'left', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF076C2808>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.1745],
        [ 0.3051],
        [ 0.6480],
        [-0.1718],
        [-0.7682],
        [-2.4076],
        [ 0.4774],
        [ 0.2362],
        [ 0.0083],
        [-0.7007]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-False-True-left-g1-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
norm = 'left', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF02FCD2C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.0493],
        [-0.0583],
        [-0.0821],
        [-0.1054],
        [-0.1820],
        [ 0.2335],
        [-0.4249],
        [ 0.0952],
        [ 0.7118],
        [ 0.7998]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-False-True-left-g2-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'left', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF076D9908>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.3018],
        [-0.2125],
        [ 0.7850],
        [ 0.5458],
        [ 0.7459],
        [-1.1136],
        [ 0.5814],
        [ 2.0761],
        [ 0.6062],
        [ 2.4590]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-False-True-left-g2-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'left', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE146E8088>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.1394],
        [-0.0749],
        [ 0.4449],
        [ 0.1151],
        [-0.4869],
        [ 0.3282],
        [-0.2145],
        [-0.2971],
        [ 0.6185],
        [-0.1619]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-False-True-left-g3-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'left', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03E93A88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.2085],
        [ 0.4731],
        [ 0.2680],
        [-0.2533],
        [ 1.7207],
        [-0.4825],
        [ 0.4298],
        [ 0.4842],
        [ 1.6548],
        [-0.1591]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-False-True-left-g3-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'left', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF076D2348>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.7822],
        [-0.2133],
        [ 0.0517],
        [ 0.6987],
        [-0.1314],
        [ 1.6082],
        [ 0.4153],
        [ 0.8002],
        [ 2.3445],
        [ 0.8257]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-False-True-left-g4-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'left', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF02F56B08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.2602],
        [ 0.0735],
        [ 0.0876],
        [-0.3472],
        [-0.2682],
        [ 0.0468],
        [ 0.4900],
        [ 0.1405],
        [-0.6808]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-False-True-left-g4-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'left', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03F34408>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.5235],
        [-1.3580],
        [-1.8963],
        [ 0.9715],
        [ 1.0745],
        [ 0.5673],
        [ 0.7357],
        [-0.9971],
        [-1.5654]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-False-True-left-g5-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'left', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1213F408>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.1855],
        [-0.7748],
        [ 0.1910],
        [ 0.3177],
        [ 0.8657],
        [-0.4116],
        [ 0.9477],
        [-1.2624],
        [-1.3538],
        [ 0.5417]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-False-True-left-g5-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'left', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1514C488>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.5144],
        [ 0.3216],
        [ 0.0543],
        [ 0.0273],
        [-1.1756],
        [-0.7328],
        [-0.0448],
        [-1.1289],
        [ 0.9796],
        [ 0.6761]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-False-True-left-g6-idtype0] ________________

idtype = torch.int32, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'left', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0771D7C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.0050],
        [-0.2580],
        [-0.9398],
        [ 0.2937],
        [-1.2709],
        [ 0.4975]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-False-True-left-g6-idtype1] ________________

idtype = torch.int64, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'left', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1513B488>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.6143],
        [ 1.5721],
        [-1.1196],
        [-2.5264],
        [ 0.1103],
        [ 0.7467]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-False-True-left-g7-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'left', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0304DFC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.3258],
        [ 0.3560],
        [ 0.9492],
        [-0.9855],
        [ 0.7863],
        [-0.9705],
        [ 1.3883],
        [ 1.3993],
        [ 1.3731],
        [-0.2480]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[1-False-True-left-g7-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'left', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14B5A088>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.2886],
        [ 0.0457],
        [ 0.0378],
        [ 0.0595],
        [ 0.0489],
        [-1.8677],
        [ 0.8013],
        [-3.1749],
        [ 0.1628],
        [-1.1777]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-False-none-g0-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'none', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())

            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07716E48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.6256],
        [ 0.0127]]), rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-False-none-g0-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'none', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1537A188>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.2338],
        [ 1.8069]]), rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-False-none-g1-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
norm = 'none', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07726C08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.5880],
        [ 2.4978],
        [-1.4408],
        [-0.1207],
        [-1.6998],
        [-0.9828],
        [ 0.2598],
        [ 2.5164],
        [-2.3908],
        [ 1.8766]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-False-none-g1-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
norm = 'none', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03F09048>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-3.8280],
        [ 5.8569],
        [ 1.2777],
        [ 7.4900],
        [-1.2450],
        [ 0.3720],
        [-4.7981],
        [ 1.1046],
        [-8.3443],
        [-5.1565]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-False-none-g2-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'none', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE121388C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.1779],
        [ 1.3015],
        [-1.4283],
        [ 0.9584],
        [-0.5161],
        [ 0.0413],
        [-0.9351],
        [ 0.4656],
        [ 0.4113],
        [ 0.8690]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-False-none-g2-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'none', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14B4DE48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.6716],
        [ 0.5132],
        [ 0.1603],
        [ 0.3683],
        [ 1.3922],
        [-1.5251],
        [-2.7796],
        [ 0.8821],
        [ 1.6174],
        [-0.2422]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-False-none-g3-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'none', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE153416C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.3135],
        [-0.1689],
        [ 2.0452],
        [ 0.2620],
        [ 2.2767],
        [ 1.1864],
        [-1.6842],
        [-1.0519],
        [ 2.5822],
        [-0.5491]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-False-none-g3-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'none', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0302C448>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 2.9692],
        [ 1.6751],
        [-1.1200],
        [-1.6243],
        [-0.7863],
        [ 3.0730],
        [ 0.3680],
        [ 3.7776],
        [ 2.0418],
        [-3.3161]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-False-none-g4-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'none', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14C6FB88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.9026],
        [-0.6367],
        [-0.2877],
        [-0.9867],
        [-0.1171],
        [-0.2924],
        [ 0.0233],
        [ 0.0684],
        [-0.0197]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-False-none-g4-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'none', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14C569C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.2223],
        [ 0.7592],
        [-0.2608],
        [ 1.0893],
        [-2.1606],
        [ 0.8915],
        [ 1.3990],
        [ 1.0384],
        [-0.6188]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-False-none-g5-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'none', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE12067CC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.4293],
        [-1.3417],
        [-0.3960],
        [-0.1869],
        [-1.5616],
        [-0.2592],
        [ 0.4396],
        [ 2.7242],
        [-1.7930],
        [ 1.5360]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-False-none-g5-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'none', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE121A7F08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.5532],
        [-3.3189],
        [-0.1095],
        [ 0.0684],
        [-4.6149],
        [ 0.7661],
        [-2.4832],
        [ 0.2110],
        [-0.3799],
        [ 0.7147]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-False-none-g6-idtype0] ________________

idtype = torch.int32, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'none', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF076A1C08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.3795],
        [-2.2513],
        [ 0.5122],
        [ 2.1903],
        [ 0.1735],
        [-0.8850]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-False-none-g6-idtype1] ________________

idtype = torch.int64, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'none', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF076F2C08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 3.5292],
        [-0.5981],
        [ 0.0168],
        [-2.7670],
        [-1.5131],
        [ 1.1920]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-False-none-g7-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'none', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03F46D88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 3.5097e+00],
        [ 3.0462e-03],
        [-2.4908e+00],
        [ 1.0338e+00],
        [-7.1890e-01],
        [-6.6865e+00],
        [-6.3324e+00],
        [-2.6962e+00],
        [ 2.7939e+00],
        [-2.0413e+00]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-False-none-g7-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'none', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03046348>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.3616],
        [ 0.7891],
        [ 0.3393],
        [-0.0496],
        [-0.7879],
        [-0.8507],
        [-1.0585],
        [-0.0175],
        [-0.3280],
        [-3.0980]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-False-both-g0-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'both', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF076F6448>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.4128],
        [-0.6501]]), rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-False-both-g0-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'both', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01B0CCC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.1681],
        [ 2.2083]]), rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-False-both-g1-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
norm = 'both', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03F2C748>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.1317],
        [-2.6149],
        [-0.8724],
        [ 0.4107],
        [-0.4706],
        [-0.7456],
        [ 0.7287],
        [ 2.2722],
        [-2.1142],
        [-0.6251]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-False-both-g1-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
norm = 'both', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1528A808>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.9909],
        [-0.3135],
        [-0.6499],
        [ 0.3716],
        [ 0.7700],
        [-1.3601],
        [-0.0643],
        [ 0.1055],
        [ 0.1150],
        [ 0.2310]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-False-both-g2-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'both', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE144B6848>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.9552],
        [-0.1748],
        [ 1.5261],
        [ 1.1721],
        [ 3.9686],
        [ 1.8338],
        [ 0.2056],
        [ 0.4766],
        [-0.7383],
        [-0.5593]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-False-both-g2-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'both', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14C30D88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.1012],
        [ 0.9536],
        [ 1.4806],
        [ 0.6625],
        [-1.8739],
        [-0.0427],
        [ 0.0724],
        [-2.4209],
        [-0.0159],
        [-1.4336]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-False-both-g3-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'both', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE144C2D08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.7105],
        [-0.4184],
        [ 0.6777],
        [-1.9940],
        [-2.0492],
        [-3.3511],
        [ 0.3683],
        [ 4.2539],
        [ 0.4026],
        [-2.3076]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-False-both-g3-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'both', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0767C188>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.4240],
        [ 0.9055],
        [-1.8572],
        [-2.1598],
        [ 0.6725],
        [-0.2615],
        [-1.1722],
        [ 0.2581],
        [ 2.6203],
        [-2.6329]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-False-both-g4-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'both', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE152F8648>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.2060],
        [-0.9581],
        [ 1.6921],
        [ 0.9782],
        [ 0.8795],
        [-1.3817],
        [-0.3687],
        [-0.0086],
        [ 1.3053]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-False-both-g4-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'both', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF02F27108>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.3863],
        [ 0.3606],
        [-0.2192],
        [ 3.7313],
        [ 2.9594],
        [-0.9128],
        [ 2.2394],
        [ 0.5958],
        [-1.8183]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-False-both-g5-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'both', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF030347C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.6196],
        [-0.2119],
        [-0.0526],
        [ 0.8136],
        [ 3.2152],
        [-1.2707],
        [ 1.3921],
        [ 2.3172],
        [-0.2829],
        [-1.4588]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-False-both-g5-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'both', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE120C26C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-2.3181],
        [ 0.1270],
        [-0.4616],
        [ 1.0863],
        [-0.6060],
        [ 1.4466],
        [ 2.7397],
        [-1.5046],
        [-1.4813],
        [-3.9032]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-False-both-g6-idtype0] ________________

idtype = torch.int32, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'both', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14CD1EC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-2.2172],
        [-0.0463],
        [-2.5675],
        [-1.8359],
        [-2.1327],
        [-1.4903]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-False-both-g6-idtype1] ________________

idtype = torch.int64, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'both', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF030BCB08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.7716],
        [ 0.7275],
        [ 0.0731],
        [ 0.7891],
        [-1.7282],
        [-0.1440]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-False-both-g7-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'both', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF030ADEC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.4126],
        [ 0.0344],
        [ 0.9074],
        [-0.2892],
        [-0.2756],
        [-0.8743],
        [ 0.8636],
        [-2.0288],
        [-0.0805],
        [-0.6993]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-False-both-g7-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'both', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0309F188>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.3166],
        [ 0.9375],
        [-0.2964],
        [ 0.7283],
        [-1.2417],
        [-1.9639],
        [-1.6766],
        [ 2.0978],
        [-0.4579],
        [ 0.4364]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-False-right-g0-idtype0] _______________

idtype = torch.int32
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'right', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE144E2D88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.1818],
        [-1.4928]]), rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-False-right-g0-idtype1] _______________

idtype = torch.int64
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'right', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF030B8C48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[1.1667],
        [0.8954]]), rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-False-right-g1-idtype0] _______________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
norm = 'right', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF076D0788>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.3557],
        [-1.9649],
        [ 1.5762],
        [-0.7349],
        [-1.2568],
        [ 3.2033],
        [-1.7478],
        [ 5.2006],
        [ 1.0275],
        [ 4.2759]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-False-right-g1-idtype1] _______________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
norm = 'right', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE15BBA848>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.4009],
        [-1.0127],
        [-0.4741],
        [-0.6988],
        [ 2.2464],
        [-1.7625],
        [ 0.4454],
        [ 0.6608],
        [ 1.4734],
        [ 1.4432]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-False-right-g2-idtype0] _______________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'right', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07643708>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-4.2680],
        [-3.7542],
        [-4.7500],
        [-2.7509],
        [ 5.7743],
        [-0.8923],
        [ 3.3396],
        [-1.0490],
        [-1.6268],
        [-2.2152]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-False-right-g2-idtype1] _______________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'right', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF02FCAF88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.5796],
        [ 1.6930],
        [-2.3008],
        [ 0.3052],
        [ 3.7987],
        [ 3.1215],
        [-3.9673],
        [-4.6248],
        [-1.6991],
        [ 0.7200]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-False-right-g3-idtype0] _______________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'right', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03EEBC08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 3.5131e+00],
        [-3.7254e+00],
        [-1.2404e+00],
        [ 2.1429e-03],
        [-3.8855e+00],
        [-1.1975e+00],
        [-3.3572e+00],
        [-2.8301e+00],
        [-1.4235e+00],
        [ 3.7847e+00]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-False-right-g3-idtype1] _______________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'right', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE151CF608>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 3.8195],
        [ 0.0949],
        [-0.5304],
        [ 1.0590],
        [ 5.3441],
        [ 1.3328],
        [-0.5215],
        [-5.4427],
        [-5.7357],
        [ 0.8810]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-False-right-g4-idtype0] _______________

idtype = torch.int32
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'right', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF076051C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.6542],
        [-1.1747],
        [ 0.8104],
        [ 3.0780],
        [ 0.5219],
        [ 0.4924],
        [ 0.2044],
        [ 1.1309],
        [ 1.7793]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-False-right-g4-idtype1] _______________

idtype = torch.int64
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'right', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03F1BCC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.2994],
        [ 3.1315],
        [-3.5103],
        [ 3.1810],
        [ 1.1301],
        [-1.7526],
        [-2.4679],
        [-0.0324],
        [ 0.6423]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-False-right-g5-idtype0] _______________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'right', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07714C48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.6380],
        [ 2.5675],
        [-0.0847],
        [ 0.8192],
        [-4.7353],
        [-0.4515],
        [ 1.8456],
        [ 1.8532],
        [ 2.5695],
        [-5.2477]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-False-right-g5-idtype1] _______________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'right', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE15281848>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.7983],
        [ 0.3627],
        [ 0.0982],
        [-3.2020],
        [ 0.7554],
        [ 2.3019],
        [-3.0868],
        [-4.8390],
        [ 0.1866],
        [ 0.7385]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-False-right-g6-idtype0] _______________

idtype = torch.int32, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'right', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14B4FD88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.7939],
        [ 0.5944],
        [-0.4132],
        [-1.0590],
        [ 0.8973],
        [-1.8916]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-False-right-g6-idtype1] _______________

idtype = torch.int64, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'right', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE15BCEA88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.5444e+00],
        [-2.4221e-02],
        [ 1.7397e+00],
        [-1.0839e-03],
        [-9.5440e-01],
        [ 2.0388e-01]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-False-right-g7-idtype0] _______________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'right', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF02F8E148>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.3730],
        [ 4.8663],
        [ 0.3712],
        [-5.1312],
        [ 0.1763],
        [-0.9011],
        [ 0.7495],
        [ 2.1125],
        [ 1.6965],
        [ 0.6256]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-False-right-g7-idtype1] _______________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'right', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE144F6FC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.3824],
        [-0.0871],
        [-0.5881],
        [ 2.2911],
        [ 0.5024],
        [-0.7644],
        [ 1.2086],
        [-0.8979],
        [ 0.0146],
        [ 0.0437]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-False-left-g0-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'left', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE15277C48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.1921],
        [-0.4517]]), rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-False-left-g0-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'left', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14CE4D08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.0054],
        [ 0.5092]]), rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-False-left-g1-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
norm = 'left', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07727C88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.0680],
        [ 0.7428],
        [ 1.1853],
        [-1.2889],
        [ 2.0872],
        [ 0.2683],
        [-0.1059],
        [-3.1978],
        [-2.6572],
        [-1.0698]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-False-left-g1-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
norm = 'left', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03EBB308>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.5566],
        [ 1.7554],
        [ 0.4849],
        [-0.6584],
        [ 0.2932],
        [ 2.4827],
        [ 1.3432],
        [-2.1189],
        [-2.2858],
        [ 0.0653]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-False-left-g2-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'left', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14992548>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.0249],
        [-0.1851],
        [ 0.3530],
        [ 0.5434],
        [-1.5023],
        [ 1.1959],
        [-0.6451],
        [-3.4866],
        [-1.0649],
        [ 3.6353]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-False-left-g2-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'left', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01ACE3C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.3193],
        [-1.3189],
        [-0.8487],
        [-0.8350],
        [ 2.6867],
        [-0.3911],
        [ 1.9851],
        [-0.4855],
        [-0.8925],
        [-0.6102]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-False-left-g3-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'left', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14C92388>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.0219],
        [ 0.5216],
        [ 0.1647],
        [-0.2298],
        [-1.8173],
        [ 0.2477],
        [ 0.3092],
        [-0.5891],
        [-0.1341],
        [-1.6789]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-False-left-g3-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'left', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE15BAC888>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.4430],
        [ 0.1538],
        [ 0.0891],
        [-0.2003],
        [-0.6354],
        [ 2.5714],
        [ 1.4727],
        [ 0.1955],
        [-0.1545],
        [-0.5666]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-False-left-g4-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'left', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14505B88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.8142],
        [-1.0586],
        [ 0.6165],
        [-0.6057],
        [ 4.7728],
        [-0.1166],
        [ 1.0948],
        [-0.0617],
        [-0.6601]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-False-left-g4-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'left', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14CBCFC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.1415],
        [ 1.1511],
        [ 1.8738],
        [-1.0308],
        [-0.0978],
        [ 0.3157],
        [ 0.4837],
        [ 1.1207],
        [ 5.5642]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-False-left-g5-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'left', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF02F12CC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.2673],
        [-0.9959],
        [ 0.8056],
        [-0.3054],
        [ 2.8140],
        [-0.6102],
        [ 0.5465],
        [ 0.5067],
        [ 4.2311],
        [-0.5584]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-False-left-g5-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'left', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0767CB08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.8379],
        [ 0.9187],
        [-0.6382],
        [-1.5008],
        [ 2.9517],
        [ 1.9404],
        [ 0.8113],
        [ 3.5415],
        [-0.2680],
        [-5.1692]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-False-left-g6-idtype0] ________________

idtype = torch.int32, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'left', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE15BA5048>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.3851],
        [-0.2083],
        [-0.0678],
        [-3.2051],
        [ 1.6549],
        [ 3.8674]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-False-left-g6-idtype1] ________________

idtype = torch.int64, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'left', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0772D888>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-2.4206],
        [ 3.1417],
        [ 4.1865],
        [ 3.2696],
        [-1.9182],
        [ 0.4619]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-False-left-g7-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'left', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14668CC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.3106],
        [-1.1343],
        [-0.6437],
        [-1.1539],
        [-1.1557],
        [-1.9793],
        [-0.2602],
        [ 1.9469],
        [-6.1781],
        [ 2.3204]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[1-False-False-left-g7-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
norm = 'left', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE121ADA48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.3389],
        [ 0.2032],
        [-0.3421],
        [ 0.4110],
        [-0.2262],
        [-0.1069],
        [-0.1590],
        [ 0.5783],
        [-0.7875],
        [-0.6027]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-True-none-g0-idtype0] _________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'none', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07605408>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.3510, -1.2033],
        [ 2.0941, -0.3792]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-True-none-g0-idtype1] _________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'none', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0775E908>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.1537,  0.8681],
        [-0.1864, -1.1034]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-True-none-g1-idtype0] _________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
norm = 'none', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF030271C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.0504,  0.7085],
        [-3.2649, -0.9280],
        [-0.5854, -0.2258],
        [ 1.9170, -1.3022],
      ...4,  2.0313],
        [ 0.3720,  1.3165],
        [ 1.1159, -1.0212],
        [ 0.1067,  1.9031]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-True-none-g1-idtype1] _________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
norm = 'none', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE15BA2908>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.5379, -0.0469],
        [ 0.1300,  0.8569],
        [-1.8163,  2.3030],
        [ 0.0631,  2.5281],
      ...1,  1.2044],
        [-0.6027,  0.1854],
        [-0.9662,  2.3994],
        [ 0.7636,  0.0089]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-True-none-g2-idtype0] _________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'none', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07667508>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-9.3544e-02,  9.8692e-01],
        [-3.5051e-01, -2.9432e-01],
        [-1.4872e+00, -9.7779e-01],
        [ ....4125e-01, -4.0228e-01],
        [ 5.5287e-04,  2.0340e+00],
        [ 2.2533e+00,  1.2363e+00]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-True-none-g2-idtype1] _________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'none', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF030835C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.2253,  1.5063],
        [ 0.0752, -0.6795],
        [-0.3140, -0.8215],
        [-0.6732,  2.0183],
      ...6,  0.7034],
        [ 1.1371, -0.2608],
        [ 0.6084, -0.6646],
        [-0.0247,  0.2019]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-True-none-g3-idtype0] _________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'none', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14505608>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.4706, -1.1936],
        [ 0.4129, -0.5083],
        [-1.7008, -1.0320],
        [-0.6615, -0.7489],
      ...0,  0.0583],
        [-0.0284, -0.5668],
        [-1.1110, -1.4902],
        [ 0.2536,  2.1908]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-True-none-g3-idtype1] _________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'none', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03099E08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.9247,  0.2337],
        [-0.7319, -0.4169],
        [ 0.5557,  0.7236],
        [ 1.4424,  0.0209],
      ...3,  0.6530],
        [ 2.5194,  0.6835],
        [ 0.5937,  0.6240],
        [-0.3103,  0.3496]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-True-none-g4-idtype0] _________________

idtype = torch.int32
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'none', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03E8E988>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.2845, -0.1045],
        [-0.4104, -0.4079],
        [-0.5282,  0.3112],
        [-1.4066,  0.4456],
      ...5,  1.0008],
        [ 0.0641, -1.2041],
        [-1.2834,  1.1388],
        [-1.0793, -0.3818]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-True-none-g4-idtype1] _________________

idtype = torch.int64
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'none', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)


_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14AC2A88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.1577,  0.5563],
        [-0.0793,  1.5517],
        [-0.1574,  1.6363],
        [ 1.2792,  0.2119],
      ...0, -0.8517],
        [-0.1564,  0.0084],
        [ 1.3731,  1.1603],
        [-2.2466, -0.9019]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-True-none-g5-idtype0] _________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'none', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE149925C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.2626, -0.3043],
        [ 0.3865, -0.4808],
        [-1.5827,  1.1025],
        [ 0.0557, -0.7595],
      ...4,  0.3354],
        [-1.8639,  1.3434],
        [ 0.0773, -0.7414],
        [-1.5164,  1.3808]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-True-none-g5-idtype1] _________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'none', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF075E8848>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.6206,  0.7761],
        [-1.2280, -1.1665],
        [-1.7329,  0.7397],
        [ 0.1520,  2.0149],
      ...0,  2.0764],
        [ 1.0090,  0.9422],
        [-0.3521, -1.4957],
        [ 0.3476, -1.0777]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-True-none-g6-idtype0] _________________

idtype = torch.int32, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'none', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF076D0E88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.9730, -0.2990],
        [-1.7396,  1.9544],
        [ 1.4887, -2.3518],
        [-3.3447, -2.5611],
        [ 2.9909,  0.8722],
        [ 0.3814, -0.3130]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-True-none-g6-idtype1] _________________

idtype = torch.int64, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'none', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1531DD88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-2.8895,  3.4297],
        [-1.4949,  1.9177],
        [ 0.6202, -0.8719],
        [ 0.8381, -1.1781],
        [-1.3992,  2.0032],
        [-1.2639,  1.7271]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-True-none-g7-idtype0] _________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'none', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF02FAAA08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.1506, -0.3841],
        [-0.8161, -0.0083],
        [ 1.3171,  1.5105],
        [ 1.3239, -0.7502],
      ...7, -1.3016],
        [-0.7492, -0.0030],
        [-0.1333,  1.1590],
        [-0.0298, -0.0459]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-True-none-g7-idtype1] _________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'none', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1495D3C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.6887,  0.1179],
        [-0.6401,  0.0622],
        [ 1.3569, -1.1112],
        [ 0.6340, -0.8786],
      ...5,  0.0714],
        [ 0.5468,  0.5908],
        [ 0.7396,  0.6066],
        [-1.4022,  1.6542]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-True-both-g0-idtype0] _________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'both', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF02FB87C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.1977, -0.1035],
        [-1.6342,  0.0061]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-True-both-g0-idtype1] _________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'both', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF02F1CF48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.0740,  1.3758],
        [ 0.1760,  0.3614]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-True-both-g1-idtype0] _________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
norm = 'both', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07604D88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.5623, -0.3357],
        [-0.2300,  0.6438],
        [ 0.0048,  1.3100],
        [ 0.8437, -0.1367],
      ...1, -0.5958],
        [ 2.6356,  1.5825],
        [-0.4216, -0.0176],
        [-1.0709, -1.6192]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-True-both-g1-idtype1] _________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
norm = 'both', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14CCFF08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.5604, -0.0900],
        [-0.4357, -0.2566],
        [-0.8337,  0.7762],
        [ 1.5497, -0.5591],
      ...5,  0.2286],
        [ 0.1773,  0.2756],
        [ 0.6831, -0.4208],
        [ 2.4933, -1.4741]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-True-both-g2-idtype0] _________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'both', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF02FCD708>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.0541, -1.0928],
        [ 0.3927,  0.0262],
        [-0.0376,  0.4305],
        [ 0.3495,  0.1998],
      ...6, -1.3729],
        [-0.7693, -3.1809],
        [ 1.6856,  1.1819],
        [-0.7953, -0.3853]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-True-both-g2-idtype1] _________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'both', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF02F1EA08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.4108,  0.1643],
        [ 0.1074, -0.0914],
        [-0.4958, -1.3997],
        [-0.7228,  0.7366],
      ...9,  0.2833],
        [-0.4704,  1.6335],
        [ 2.1157,  0.2978],
        [-1.7060,  0.6353]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-True-both-g3-idtype0] _________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'both', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07607F08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.6307,  1.5040],
        [-1.5134, -1.2025],
        [-0.3035,  0.5177],
        [-0.8012, -0.4628],
      ...7, -1.1845],
        [ 0.9159, -0.9337],
        [-0.6345, -3.0922],
        [ 0.1970,  1.2617]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-True-both-g3-idtype1] _________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'both', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07706408>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.8174, -0.2862],
        [-0.9613, -0.6586],
        [ 0.1412,  0.1271],
        [-0.2842, -0.7089],
      ...7, -1.7088],
        [-1.4630, -0.5822],
        [-0.1760,  0.8594],
        [-1.3426, -1.6255]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-True-both-g4-idtype0] _________________

idtype = torch.int32
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'both', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE151278C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.5354, -1.4531],
        [ 0.0583,  0.5944],
        [ 1.0574,  1.1329],
        [ 1.0087,  1.0953],
      ...1, -0.2766],
        [-0.0589, -0.9179],
        [-0.3779, -0.6678],
        [-0.4575,  0.3944]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-True-both-g4-idtype1] _________________

idtype = torch.int64
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'both', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03092708>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.5246, -0.5166],
        [ 0.2273,  0.9367],
        [-0.6031, -0.5982],
        [ 0.0890, -1.9857],
      ...3, -0.3360],
        [-2.7264,  0.9253],
        [ 1.4066, -0.1912],
        [ 2.6044,  0.1562]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-True-both-g5-idtype0] _________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'both', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF02F578C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.4157, -0.8940],
        [-0.6024,  0.3532],
        [ 0.4387, -0.0652],
        [-0.8490,  0.7169],
      ...8,  0.9790],
        [ 0.5174, -1.7253],
        [ 1.2080, -1.1855],
        [ 0.3793,  0.4778]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-True-both-g5-idtype1] _________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'both', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1524C4C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.4491,  0.0214],
        [ 2.2138, -0.3248],
        [-0.0865, -0.1793],
        [-0.2736, -0.0441],
      ...7,  0.9216],
        [-2.3840,  1.2223],
        [ 1.7165, -0.7194],
        [ 0.8718, -0.5475]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-True-both-g6-idtype0] _________________

idtype = torch.int32, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'both', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14AAB808>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-6.6117e-01, -3.4301e-01],
        [ 1.1839e+00, -5.2122e-01],
        [-1.3259e+00,  5.3978e-01],
        [ 4.0653e-01, -3.7473e-04],
        [-1.0350e+00,  5.4503e-01],
        [-4.0212e-01,  7.9108e-03]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-True-both-g6-idtype1] _________________

idtype = torch.int64, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'both', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03F08A48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.4732, -1.3405],
        [-0.0411,  1.2524],
        [ 0.8239,  1.5063],
        [ 0.5826, -0.0376],
        [-0.6542, -2.1080],
        [-0.8224, -0.6129]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-True-both-g7-idtype0] _________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'both', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1498F448>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 5.1421e-01, -3.7928e-01],
        [ 9.1219e-01,  6.8574e-01],
        [-8.7118e-01, -1.0845e+00],
        [-....8348e-01, -5.1507e-01],
        [ 7.7130e-01, -1.5213e+00],
        [ 1.4845e-01, -8.2265e-01]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-True-both-g7-idtype1] _________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'both', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14A38148>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.0028,  0.0405],
        [-1.3411,  1.6178],
        [ 0.0222,  0.5147],
        [-0.5206,  2.3182],
      ...9,  0.3672],
        [ 0.8854,  1.7129],
        [-1.1327,  0.5161],
        [-1.3566,  0.5046]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-True-right-g0-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'right', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14C94488>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.2188,  0.0886],
        [ 0.4473, -2.1293]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-True-right-g0-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'right', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1213A148>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.0917, -0.0408],
        [-1.5218, -1.4041]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-True-right-g1-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
norm = 'right', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03023248>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.7112, -0.1551],
        [-0.8724, -3.9883],
        [-0.0189,  0.6087],
        [ 0.3582, -0.5541],
      ...9,  1.8897],
        [ 3.4186, -1.1915],
        [ 0.5779, -0.3301],
        [-0.4252, -2.2369]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-True-right-g1-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
norm = 'right', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE149B7A88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.2917,  1.0198],
        [-0.4070,  1.9053],
        [-0.4620,  0.4893],
        [ 2.3145,  0.1027],
      ...9,  0.9367],
        [ 2.5415, -0.1115],
        [ 2.5031, -1.7738],
        [-0.4075,  1.2470]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-True-right-g2-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'right', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07629448>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.2297,  1.1682],
        [-2.2343, -1.4137],
        [-2.0810,  0.0997],
        [-0.6507, -0.1213],
      ...0,  0.1398],
        [ 0.8965,  1.4546],
        [-0.6434, -1.7759],
        [ 1.2594, -0.5066]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-True-right-g2-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'right', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE149ABD08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.8156, -1.5560],
        [ 0.3263,  0.7608],
        [-0.9443,  0.9623],
        [ 2.0258,  1.9506],
      ...8,  0.1300],
        [-2.4970, -0.4890],
        [ 1.2778,  0.8813],
        [-0.1548, -0.8833]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-True-right-g3-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'right', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03EEBF88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.3569, -0.3814],
        [ 0.1992,  0.8138],
        [ 0.4044,  1.7560],
        [ 0.8973, -1.0824],
      ...1,  0.8195],
        [-0.8567,  0.2225],
        [-0.7326, -0.6934],
        [ 0.3696, -1.4561]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-True-right-g3-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'right', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07641BC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 2.0490, -0.8254],
        [ 0.8750,  1.2634],
        [ 1.3757,  0.6084],
        [ 0.8813, -0.5433],
      ...9, -1.1047],
        [-0.9185,  2.9150],
        [-0.2230,  0.0739],
        [ 2.0066, -0.9647]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-True-right-g4-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'right', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14A3E248>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.9710,  0.9731],
        [ 0.1000, -0.1293],
        [-1.7260, -0.5521],
        [ 1.9648, -1.8164],
      ...3,  0.1450],
        [ 0.3354, -0.6281],
        [-0.5788, -0.4707],
        [-2.5957, -2.6988]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-True-right-g4-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'right', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF011A7948>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.7354, -2.4897],
        [ 0.5452,  1.8880],
        [-0.1445,  0.2828],
        [ 0.1568, -1.5635],
      ...4,  0.4912],
        [ 0.1052,  0.7621],
        [-1.0828, -0.5451],
        [-0.3358,  1.7221]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-True-right-g5-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'right', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF02F45048>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.6734, -2.0755],
        [ 0.9396,  0.0535],
        [ 0.6185,  0.5499],
        [ 1.1905,  0.2336],
      ...9, -1.2314],
        [ 0.7117,  0.9211],
        [-1.4289, -0.5696],
        [-0.3712,  0.2487]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-True-right-g5-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'right', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01AF07C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.3207,  0.4255],
        [-2.2391,  0.3768],
        [-1.3602,  0.7954],
        [ 2.5255, -2.3910],
      ...7,  0.0143],
        [ 0.6623,  0.4566],
        [ 1.0536, -1.0605],
        [-0.1507, -2.1614]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-True-right-g6-idtype0] ________________

idtype = torch.int32, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'right', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14668188>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.2132,  3.1866],
        [ 1.0515, -0.3637],
        [ 0.7496,  1.1329],
        [-0.9516,  1.1370],
        [ 1.2935, -1.1386],
        [ 2.1579,  2.3583]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-True-right-g6-idtype1] ________________

idtype = torch.int64, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'right', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14A1AAC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.1593, -1.7061],
        [ 1.5382,  0.0033],
        [-0.4101, -1.1806],
        [ 0.5399,  0.8544],
        [-0.4154,  0.0623],
        [-0.1891, -0.3187]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-True-right-g7-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'right', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE121CCBC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.4628,  0.3408],
        [ 0.2185, -1.1777],
        [ 1.4324,  0.6008],
        [-0.5232,  0.6096],
      ...6, -0.9130],
        [ 0.8207, -0.8987],
        [-0.3755,  1.1943],
        [ 1.1619, -0.2422]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-True-right-g7-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'right', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01A63B88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.3122, -0.0980],
        [-0.1539, -0.3802],
        [ 0.0056,  1.9926],
        [ 0.1946, -0.3754],
      ...9, -0.8729],
        [-0.5135, -0.2962],
        [ 0.4841, -1.2970],
        [ 0.3075,  0.0866]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-True-left-g0-idtype0] _________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'left', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03EBBA48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.2218, -0.2219],
        [-0.0734,  0.1307]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-True-left-g0-idtype1] _________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'left', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF02F1CE08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.0209,  0.3161],
        [-0.0612,  0.1476]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-True-left-g1-idtype0] _________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
norm = 'left', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE149B28C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.9545, -0.3654],
        [ 0.4692, -0.5705],
        [ 0.2420,  0.0956],
        [ 0.5942, -0.1102],
      ...8,  0.4365],
        [ 0.1804, -1.3885],
        [ 1.6430,  0.2337],
        [ 0.5259, -0.1823]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-True-left-g1-idtype1] _________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
norm = 'left', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF011B4908>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.1867, -0.0226],
        [-0.5136, -0.3010],
        [ 0.0695, -0.1557],
        [ 0.1235,  0.3107],
      ...6, -0.0823],
        [-0.0684,  1.1479],
        [ 0.7049, -0.3347],
        [-0.1789, -0.2288]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-True-left-g2-idtype0] _________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'left', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE15179A88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.1032,  0.0531],
        [ 0.0278, -0.0175],
        [-0.0903,  0.1643],
        [-0.0448,  0.2767],
      ...0,  0.0927],
        [-0.5068,  1.3612],
        [-1.2713,  2.4504],
        [-0.8311,  1.1035]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-True-left-g2-idtype1] _________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'left', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03F1BDC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.3112, -0.2099],
        [-0.1574, -0.0667],
        [ 0.2777,  0.2505],
        [ 0.0402, -0.1744],
      ...4, -0.3779],
        [-1.1050,  0.8716],
        [-0.2265, -0.9129],
        [-1.7188,  0.1669]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-True-left-g3-idtype0] _________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'left', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE149BC8C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.3624, -0.8149],
        [ 0.7222, -1.2998],
        [ 0.2078,  1.1462],
        [ 1.0498,  0.5678],
      ...4, -0.0485],
        [ 1.0299,  1.5393],
        [-1.0896, -0.0586],
        [-0.7867,  1.9295]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-True-left-g3-idtype1] _________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'left', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14A1D4C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.0263, -0.0325],
        [ 0.6689,  0.6167],
        [-0.1908,  0.0759],
        [ 0.2869, -0.2454],
      ...1, -0.3056],
        [-0.0752,  0.8032],
        [-1.5357, -0.7494],
        [ 2.2296, -1.1831]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-True-left-g4-idtype0] _________________

idtype = torch.int32
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'left', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE145ABE48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.7238, -0.1302],
        [ 0.2632, -0.7747],
        [ 0.7217, -0.4355],
        [ 1.3497, -0.1121],
      ...7, -0.1635],
        [ 2.6146,  0.9852],
        [ 0.4407,  0.7338],
        [-0.7386, -0.6700]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-True-left-g4-idtype1] _________________

idtype = torch.int64
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'left', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1524C7C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.4028,  0.3682],
        [-0.1282,  0.5773],
        [-0.5238, -0.3647],
        [-0.3389, -0.5345],
      ...3, -0.3821],
        [ 1.0693, -0.5560],
        [ 0.6188, -0.1521],
        [ 1.0323, -1.3502]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-True-left-g5-idtype0] _________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'left', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14A44E08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.9231, -0.5061],
        [ 0.0911,  0.4708],
        [ 0.5256, -0.3522],
        [ 0.4151,  0.2415],
      ...3, -0.1314],
        [-1.8468, -2.0167],
        [ 0.4128, -0.5724],
        [-3.1471, -2.3906]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-True-left-g5-idtype1] _________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'left', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01115C88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.1435, -0.0499],
        [ 0.4408,  0.2236],
        [-0.2207, -0.0237],
        [ 0.3001, -0.1068],
      ...5,  0.6482],
        [ 2.5116, -1.0981],
        [ 2.9757, -0.8697],
        [-1.4709,  0.9189]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-True-left-g6-idtype0] _________________

idtype = torch.int32, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'left', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14A41DC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.0788, -1.0705],
        [ 0.5486,  2.7754],
        [-0.8486,  1.0115],
        [ 0.1418,  1.5106],
        [-0.4816,  0.1798],
        [-0.3411, -1.4440]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-True-left-g6-idtype1] _________________

idtype = torch.int64, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'left', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14C54588>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.3099, -0.8310],
        [-1.1597, -0.0358],
        [ 0.1113, -0.0022],
        [-1.7612, -1.3956],
        [ 1.5562,  0.3306],
        [ 0.6760,  0.0045]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-True-left-g7-idtype0] _________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'left', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07747F08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.9007,  0.1032],
        [-0.4656, -0.3208],
        [-0.0553, -0.1825],
        [ 0.3899,  0.0471],
      ...5,  0.1033],
        [ 0.8115,  0.2386],
        [ 0.2441,  0.5742],
        [-0.8917,  0.3200]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-True-left-g7-idtype1] _________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'left', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF02FF9508>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.0221,  0.1607],
        [ 0.5454, -0.5471],
        [ 0.6494,  0.6237],
        [ 0.2013, -0.5390],
      ...6,  0.4816],
        [ 1.0953,  1.5182],
        [ 1.0593,  1.7145],
        [-1.3243, -0.3539]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-False-none-g0-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'none', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE144EBF88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-3.1038,  1.4069],
        [ 0.3139, -1.4302]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-False-none-g0-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'none', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE146362C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 2.6935,  2.6284],
        [-0.2618,  4.5503]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-False-none-g1-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
norm = 'none', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0767CB08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.8483,  1.2523],
        [-3.0163,  0.6798],
        [-1.8218,  1.4283],
        [ 0.8501, -0.2874],
      ...8515],
        [ 0.0036,  2.8571],
        [ 0.9199, -1.4428],
        [-2.5141,  1.1327],
        [ 1.2865, -0.8480]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-False-none-g1-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
norm = 'none', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01A6FB08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.6574, -0.4906],
        [-0.8586, -0.1681],
        [-0.9011, -0.5774],
        [-1.9939,  1.3447],
      ...4813],
        [-2.6304,  0.1364],
        [-0.7477, -0.1641],
        [-4.1026,  0.4028],
        [ 1.1489, -0.0139]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-False-none-g2-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'none', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14A16B48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.7921,  3.9901],
        [ 0.5120, -5.0783],
        [-0.9904, -1.9665],
        [-1.2874,  0.7187],
      ...5316],
        [-0.4013,  0.5192],
        [-0.9845, -3.7609],
        [ 0.6635,  0.4240],
        [-1.3116, -3.8873]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-False-none-g2-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'none', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07706C08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 3.7867, -1.1327],
        [-0.6717, -0.6868],
        [ 1.0442,  1.4271],
        [-0.6068, -0.3815],
      ...5785],
        [ 3.2830, -3.3188],
        [-0.9756, -3.0226],
        [ 1.3609, -0.8369],
        [ 0.7050, -3.2077]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-False-none-g3-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'none', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14C5B5C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.5593,  1.0231],
        [ 3.4971,  3.3288],
        [ 1.0157, -1.2880],
        [-0.9406, -2.6839],
      ...6769],
        [-1.8497,  0.0611],
        [-1.9347,  0.7612],
        [ 3.1626,  1.6616],
        [-0.4295, -3.2889]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-False-none-g3-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'none', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03EDA948>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.2379,  2.5240],
        [-1.6867, -1.7532],
        [-1.8167, -5.2403],
        [ 1.0331,  0.8809],
      ...6136],
        [ 0.2817,  2.2451],
        [ 1.0936,  1.3443],
        [-0.2824,  2.5660],
        [-2.0554, -5.0014]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-False-none-g4-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'none', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14C9CA08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.4620,  1.0373],
        [ 1.5582,  0.2610],
        [ 0.8856,  0.6735],
        [-1.4881, -0.1415],
      ...2056],
        [ 2.0506, -1.4575],
        [-3.5042, -0.2549],
        [-1.4503,  1.9610],
        [-0.4739, -0.2639]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-False-none-g4-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'none', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF02F6E288>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.7493, -4.3459],
        [ 1.0714, -0.5704],
        [ 0.2167,  0.7940],
        [ 0.7217,  1.3489],
      ...3441],
        [ 1.3855, -0.8567],
        [-0.7682,  1.5881],
        [ 0.8190, -3.2881],
        [-0.2441, -1.3958]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-False-none-g5-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'none', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14652EC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.1822, -1.6228],
        [-2.3096, -0.4543],
        [-0.8827,  5.0671],
        [ 4.8670,  0.0332],
      ...1868],
        [-0.9538, -2.0910],
        [ 0.7860, -1.1361],
        [ 6.3877,  1.1140],
        [-3.0291, -0.2933]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-False-none-g5-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'none', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE151D6608>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.1064,  0.6334],
        [-0.2531,  0.3046],
        [ 0.2591, -1.1230],
        [-0.3274, -0.4442],
      ...3651],
        [ 1.9288, -1.1200],
        [-0.7842, -0.6901],
        [ 0.0570, -0.2404],
        [-0.3439,  5.1512]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-False-none-g6-idtype0] ________________

idtype = torch.int32, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'none', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE15155948>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.6255,  0.2909],
        [ 1.3529, -2.3293],
        [-1.1255,  0.4386],
        [ 5.7547,  0.0166],
        [ 1.2109,  2.3299],
        [ 0.2132,  3.1363]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-False-none-g6-idtype1] ________________

idtype = torch.int64, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'none', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14C767C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-2.7146, -1.0921],
        [ 4.7896, -0.7883],
        [ 1.6459, -0.3005],
        [-1.7993, -0.3793],
        [ 0.9061,  0.9359],
        [ 1.8140,  0.4497]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-False-none-g7-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'none', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01A7F948>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 3.0171,  2.1117],
        [ 2.3655, -0.9829],
        [-1.3327,  1.0677],
        [ 1.3787,  2.0275],
      ...6115],
        [-2.2035, -0.6916],
        [-4.2630,  4.9994],
        [-0.1320, -0.1727],
        [ 0.6567,  0.1052]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-False-none-g7-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'none', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF011507C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-2.7799, -2.1808],
        [ 3.7893,  0.5430],
        [ 0.5264,  1.9132],
        [ 2.9576, -1.7494],
      ...9770],
        [ 0.7313, -2.3477],
        [ 3.9454, -1.5775],
        [-0.2805,  0.6189],
        [ 4.4893, -3.2523]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-False-both-g0-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'both', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF02FD5B88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-2.5910,  0.0130],
        [ 1.7252,  1.3198]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-False-both-g0-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'both', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03F08BC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.8044, -0.7901],
        [-0.4420,  0.3551]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-False-both-g1-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
norm = 'both', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF02FB2408>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.7062,  0.6063],
        [ 0.3307, -0.2482],
        [-0.6499,  0.4537],
        [-1.3669, -0.1204],
      ...8445],
        [-3.7200,  0.6986],
        [-2.6939,  2.2762],
        [-5.0012,  1.3519],
        [ 3.0317, -0.2037]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-False-both-g1-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
norm = 'both', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1533FE48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.1475,  1.1137],
        [ 1.8885,  0.8081],
        [-0.7010,  0.8699],
        [-1.0247, -0.4397],
      ...6786],
        [ 0.9208, -0.5519],
        [-1.2947,  1.7097],
        [ 2.8764, -3.6228],
        [-4.5886,  2.0769]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-False-both-g2-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'both', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07714FC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.2957, -1.1726],
        [-0.7298, -1.8187],
        [-1.8211, -1.5475],
        [-0.9688, -0.2992],
      ...6347],
        [-2.1066, -1.0577],
        [-1.3008, -2.7194],
        [ 3.1044,  1.2392],
        [-1.0657,  1.2425]])
rhs_data = None


        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-False-both-g2-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'both', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14636C48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.3483,  0.9375],
        [ 1.0088,  2.6694],
        [-0.4922, -1.7295],
        [-0.3963,  0.9942],
      ...7332],
        [ 1.5232,  2.3133],
        [-1.1584, -1.1603],
        [-1.4227, -2.1046],
        [ 2.8352, -0.9651]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-False-both-g3-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'both', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1459A8C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.4533, -0.9787],
        [-1.0018,  2.5059],
        [ 0.1557, -1.5101],
        [ 0.4798,  0.8907],
      ...6367],
        [ 0.6822,  3.0508],
        [ 1.4081, -2.2727],
        [ 0.1043,  0.4562],
        [-0.0588, -2.7561]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-False-both-g3-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'both', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF077061C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.4683, -0.8528],
        [-0.2431, -0.0284],
        [-0.2983, -0.6544],
        [-1.5729,  1.0841],
      ...4255],
        [ 1.3360,  1.6341],
        [-0.6720, -1.2897],
        [-2.6432,  4.6038],
        [ 2.7745,  0.5042]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-False-both-g4-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'both', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF02F149C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.1542, -1.0466],
        [ 0.2589,  0.7990],
        [ 1.1994, -1.0754],
        [ 1.3226,  1.7767],
      ...1705],
        [ 2.7515, -1.6489],
        [-2.5441, -1.4849],
        [-2.9065, -0.1691],
        [ 1.2440, -2.0976]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-False-both-g4-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'both', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14C56148>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.4624, -0.5149],
        [-2.4711, -0.0303],
        [ 1.6209, -0.4715],
        [-0.0861,  0.3758],
      ...5926],
        [-2.2514, -0.2710],
        [ 3.9819, -0.4650],
        [ 4.6782, -0.8425],
        [-2.4292,  0.4130]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-False-both-g5-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'both', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03074108>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.7879,  0.7489],
        [ 1.1638,  2.0731],
        [ 1.4356,  0.5307],
        [ 0.5677, -1.2210],
      ...8896],
        [-2.5641,  0.8320],
        [ 0.2211,  0.0270],
        [-0.4287, -0.2813],
        [-1.0749,  1.2159]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-False-both-g5-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'both', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF077099C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.4740,  0.8370],
        [-1.2922,  1.5800],
        [-0.0455, -0.3925],
        [-0.3020, -1.4478],
      ...4211],
        [-0.6816, -0.7538],
        [ 0.4064, -0.8560],
        [ 3.1878, -1.5870],
        [-0.9618,  1.5481]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-False-both-g6-idtype0] ________________

idtype = torch.int32, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'both', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03023148>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 2.1147, -3.3883],
        [-0.1674, -1.8532],
        [ 0.9396,  0.4774],
        [ 0.4868, -0.1447],
        [ 0.1077, -1.8286],
        [ 1.7686, -1.8581]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-False-both-g6-idtype1] ________________

idtype = torch.int64, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'both', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE15288CC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.2536,  0.6840],
        [-4.7416, -1.9338],
        [ 5.1740,  3.0828],
        [ 4.2719, -2.9212],
        [ 3.3435,  3.2829],
        [ 5.4536,  0.3366]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-False-both-g7-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'both', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF02F4E508>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.6112, -1.0322],
        [-0.9660, -1.2656],
        [-0.5568, -0.0170],
        [-1.0585,  0.0364],
      ...6750],
        [ 1.0281,  1.2264],
        [-1.2922, -0.1678],
        [-0.7419,  3.4018],
        [ 0.9181, -2.0621]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-False-both-g7-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'both', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03F34948>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-2.8728,  2.0561],
        [-2.8980, -0.9221],
        [-1.6157, -1.1708],
        [ 0.1209,  0.1299],
      ...8515],
        [ 2.2443,  3.1604],
        [ 2.2964,  0.3459],
        [-1.6326, -0.9366],
        [ 1.8599, -0.2178]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-True-False-right-g0-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'right', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE145A3A48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.6409, -0.4236],
        [ 2.0835, -1.2922]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-True-False-right-g0-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'right', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE153728C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 2.6761, -3.1534],
        [-0.6356, -2.1361]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-True-False-right-g1-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
norm = 'right', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01A7F048>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.5273, -0.4669],
        [ 2.4999, -0.5867],
        [ 1.9906,  0.0103],
        [-2.2557,  1.9009],
      ...7437],
        [ 0.1476,  1.0344],
        [-0.3726,  0.2260],
        [ 2.0140, -1.5441],
        [ 6.2371, -1.6143]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-True-False-right-g1-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
norm = 'right', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14581A48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-4.2971, -2.6336],
        [-0.7071, -1.9350],
        [ 0.5419,  0.4215],
        [-1.0123, -0.1804],
      ...1184],
        [-1.6244,  2.1048],
        [ 1.8255,  1.1557],
        [ 0.2511, -1.4123],
        [-0.0599,  0.1206]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-True-False-right-g2-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'right', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE152694C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.6189,  0.7960],
        [-2.8833, -0.0803],
        [ 1.5625, -2.1088],
        [-4.8389, -0.5395],
      ...4260],
        [-0.8576,  0.3372],
        [-2.9468,  1.1621],
        [ 0.9221, -2.5095],
        [-2.6276,  0.8042]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-True-False-right-g2-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'right', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1463CC08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.9061, -1.3634],
        [ 1.9882, -1.5510],
        [ 0.9047, -1.5794],
        [-3.6180,  0.5518],
      ...8422],
        [ 0.7737, -1.0510],
        [-0.2739, -2.1356],
        [ 1.9877, -1.3571],
        [ 1.5729,  0.3260]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-True-False-right-g3-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'right', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE15363A88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.6694, -2.2023],
        [-1.0809, -0.3481],
        [ 1.8799,  1.6363],
        [ 0.4183,  2.9646],
      ...6454],
        [ 1.2242, -0.1957],
        [-1.3460,  0.4409],
        [ 1.7706,  6.4919],
        [ 0.9262,  3.8824]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-True-False-right-g3-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'right', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1451A4C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 8.3340e-01, -7.0120e-01],
        [ 2.2028e-01,  2.3017e+00],
        [ 2.2716e-01, -6.9125e-01],
        [-...0870e+00],
        [-1.1036e+00,  5.4663e-01],
        [ 5.8810e-01, -5.8195e-02],
        [-1.7244e+00, -2.2836e+00]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-True-False-right-g4-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'right', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE145A1648>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.4139,  0.5883],
        [-0.0606,  2.2650],
        [ 0.6067, -1.3302],
        [ 0.5640,  0.7494],
      ...1316],
        [-0.5894, -0.1608],
        [ 0.9895,  0.8452],
        [ 1.4778, -0.3236],
        [ 0.5015, -0.9081]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-True-False-right-g4-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'right', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE151A1348>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.4592,  1.8212],
        [ 0.8609,  1.1843],
        [ 0.4434,  1.9594],
        [ 3.4160, -1.6861],
      ...4611],
        [-2.6194, -3.4491],
        [ 3.4010, -3.1073],
        [-2.5688, -4.4059],
        [ 2.8482, -2.4701]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-True-False-right-g5-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'right', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF010F4808>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ -0.3845,  -3.5402],
        [  0.6934, -10.1749],
        [ -0.2345,  -1.1569],
        [ -0.1608,  -3.2896]...       [  3.1565,  -1.2959],
        [  1.6170,   5.8598],
        [ -2.6985,  -1.2090],
        [ -0.1040,   3.3796]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-True-False-right-g5-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'right', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14B4BF88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 2.6372, -2.1168],
        [ 0.9317, -2.3713],
        [ 1.3497, -4.0589],
        [-0.3685,  2.1150],
      ...1984],
        [-0.2971, -0.8895],
        [-2.4547,  0.9126],
        [-0.0194,  0.3901],
        [ 0.7676, -0.5656]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-True-False-right-g6-idtype0] ________________

idtype = torch.int32, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'right', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01A27108>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.8322,  1.1402],
        [ 0.7285, -1.7066],
        [ 0.0521, -1.4431],
        [-1.8780,  0.5513],
        [ 1.4213, -1.9293],
        [ 1.2500, -1.7972]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-True-False-right-g6-idtype1] ________________

idtype = torch.int64, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'right', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF010CCDC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-2.6300e+00, -1.6471e+00],
        [-2.6725e+00, -4.6986e-01],
        [ 1.6736e+00,  2.4942e-03],
        [-1.3288e+00, -1.6191e+00],
        [-2.5534e+00, -1.1186e+00],
        [ 2.8410e+00, -2.0453e-02]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-True-False-right-g7-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'right', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14A5AE88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.6823, -2.2873],
        [ 4.3929, -5.4634],
        [ 2.8950, -2.5847],
        [ 1.4286, -0.2859],
      ...3409],
        [ 0.8874,  0.4650],
        [ 1.4367, -0.9764],
        [-1.0290,  0.4248],
        [-0.7761, -0.9615]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-True-False-right-g7-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'right', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14A39AC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.1978,  0.4316],
        [-3.5302,  2.0848],
        [ 0.7262,  1.2848],
        [-1.7018, -3.2592],
      ...5650],
        [ 1.4861, -1.4744],
        [ 2.4296,  2.7572],
        [ 2.0522,  5.6106],
        [-1.5731, -6.0252]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-False-left-g0-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'left', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01AA6FC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.1612, -1.2319],
        [ 0.0279, -0.1106]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-False-left-g0-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'left', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0760E188>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.3059,  0.6040],
        [ 1.1540, -0.1718]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-False-left-g1-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
norm = 'left', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF010F3048>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.1738, -0.3887],
        [-0.4824,  0.9478],
        [-0.6532, -0.2325],
        [ 0.3343, -0.2724],
      ...4535],
        [-0.8336,  0.5149],
        [-0.1196,  0.0824],
        [-1.5326, -0.3856],
        [-2.1580, -0.5017]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-False-left-g1-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
norm = 'left', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE15BE3848>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.8681, -0.6669],
        [ 1.1011,  0.1897],
        [ 1.0957, -0.5683],
        [ 0.0364,  0.1059],
      ...9878],
        [-0.3603,  0.0948],
        [-0.5775, -3.8863],
        [ 0.2336, -1.3953],
        [-0.9510, -0.2164]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-False-left-g2-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'left', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF010C47C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.4897,  0.5152],
        [-0.9188, -1.1178],
        [-0.7001, -1.2068],
        [ 0.3053,  0.4037],
      ...2412],
        [-0.4948, -0.2743],
        [ 0.5543,  1.1961],
        [-0.2504, -0.6629],
        [-1.5702, -2.3218]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-False-left-g2-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'left', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE149BB248>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.0463,  0.2475],
        [-2.1773,  2.0342],
        [-0.3853, -0.3070],
        [ 0.4544,  0.1622],
      ...6741],
        [ 0.5703, -0.4988],
        [-0.2645, -0.3946],
        [ 0.5758, -0.4398],
        [-3.6519, -1.1515]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-False-left-g3-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'left', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE15179E88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.0906,  0.1576],
        [ 0.2463, -2.6585],
        [-0.0829, -0.2686],
        [-0.1055,  0.0452],
      ...8184],
        [-0.2336, -0.1893],
        [-0.4390, -1.1796],
        [ 2.6992, -2.0916],
        [-0.6708,  2.1548]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-False-left-g3-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'left', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF010FAF88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.1176,  0.5127],
        [ 0.4645, -0.7096],
        [ 0.1789, -0.0114],
        [ 0.0430, -0.7143],
      ...0883],
        [-0.3608,  0.0635],
        [-1.2289,  0.2724],
        [ 0.8642,  0.5051],
        [-2.2392, -0.8838]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-False-left-g4-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'left', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1462AD08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.3894, -0.0277],
        [ 1.0413, -0.7152],
        [-0.0575,  1.2355],
        [ 0.1051,  0.1201],
      ...5244],
        [ 0.0978,  0.5693],
        [ 0.7219, -3.7471],
        [-0.7420,  1.6046],
        [ 0.3740,  3.2628]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-False-left-g4-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'left', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14A34DC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.6297,  0.7194],
        [ 1.6322,  0.2568],
        [-0.2841, -0.4658],
        [-1.5503,  3.0142],
      ...8766],
        [ 0.6531, -1.2134],
        [ 1.3474, -2.4522],
        [-0.2198, -1.4678],
        [-0.7074, -1.7778]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-False-left-g5-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'left', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0774A788>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.3027,  0.3270],
        [ 0.7443, -1.0781],
        [ 0.0990, -0.3217],
        [ 0.2598,  0.2043],
      ...0300],
        [ 0.0600,  0.4867],
        [-2.7074,  1.9382],
        [ 0.9198, -0.6707],
        [ 0.8896, -0.3577]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-False-left-g5-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'left', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF075E1248>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.4634,  1.1836],
        [ 0.4213, -0.0968],
        [ 0.2719, -1.0563],
        [-0.1251, -0.3633],
      ...8906],
        [ 0.9060,  1.4242],
        [-2.9295, -1.7370],
        [ 0.3232, -2.2960],
        [ 0.2429, -0.6271]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-False-left-g6-idtype0] ________________

idtype = torch.int32, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'left', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF010E96C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.0149, -1.8837],
        [-0.5751,  0.7855],
        [ 1.4727, -3.9272],
        [ 3.0657,  2.2302],
        [-5.2185,  1.5652],
        [-2.4637,  3.4981]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-False-left-g6-idtype1] ________________

idtype = torch.int64, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'left', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03EDA948>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.7905,  4.0623],
        [ 0.7270, -2.6030],
        [-0.3949, -2.0875],
        [ 0.8102, -2.0254],
        [-0.6968, -2.1203],
        [ 5.1764, -1.9141]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-False-left-g7-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'left', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07730188>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.4590,  0.0446],
        [ 0.5201,  0.2081],
        [ 1.8642, -1.9827],
        [-0.3731,  0.0275],
      ...3855],
        [-1.8288,  1.5983],
        [-0.2209, -2.1816],
        [-2.2623, -0.2227],
        [ 1.3818, -2.1635]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-True-False-left-g7-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'left', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE121C5C08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.4682,  0.3187],
        [-1.1005, -1.5539],
        [ 0.3365,  0.3485],
        [ 0.1688,  0.4762],
      ...7695],
        [-1.1772,  0.6615],
        [ 2.5630, -3.0408],
        [ 3.0410,  1.3100],
        [-0.3057, -0.9149]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-False-True-none-g0-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'none', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14A49C48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.7215, -0.6009],
        [ 1.9219, -0.0229]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-False-True-none-g0-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'none', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE149D2F48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[3.3094, 1.9528],
        [0.6143, 0.4605]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-False-True-none-g1-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
norm = 'none', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14C5B408>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.8285,  1.4804],
        [ 1.5547,  1.3076],
        [-0.1444,  0.4262],
        [ 0.7904,  1.4070],
      ...4, -0.8732],
        [-0.2363, -0.9765],
        [-0.0720,  0.0690],
        [-0.7991,  2.3597]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-False-True-none-g1-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
norm = 'none', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF02FF0888>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.5415, -1.0146],
        [ 1.1876,  0.1628],
        [-0.4105, -0.3801],
        [-2.7933,  2.1704],
      ...6,  0.6673],
        [-0.5348,  2.9953],
        [ 0.2051,  2.4254],
        [-0.0767,  1.9851]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-False-True-none-g2-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'none', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14957B48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.4003,  0.3144],
        [ 0.2791,  1.1120],
        [-0.1074, -0.9436],
        [ 1.2608,  0.8348],
      ...5, -1.1479],
        [-0.8949, -1.1987],
        [-1.0666, -0.4102],
        [-0.9425, -2.0894]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-False-True-none-g2-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'none', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF076C2FC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.4309,  0.6914],
        [-0.3385, -0.5028],
        [-0.5440, -0.3276],
        [-0.1008, -2.3175],
      ...8,  0.8096],
        [-0.5949, -0.8338],
        [ 0.6148, -1.2220],
        [ 0.1815,  0.1309]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-False-True-none-g3-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'none', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE149D7388>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.8029, -0.6757],
        [ 1.5028, -0.3046],
        [ 0.6050, -0.8295],
        [ 0.8183,  0.6606],
      ...5, -1.7407],
        [ 1.7118, -1.6028],
        [ 2.6061, -2.2420],
        [ 3.2592, -2.0647]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-False-True-none-g3-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'none', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01A19F88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.0547, -1.3965],
        [ 0.6742,  1.0604],
        [ 0.1934,  0.1680],
        [ 0.5959,  1.0136],
      ...9,  1.3723],
        [ 1.3612,  0.2100],
        [ 0.3638,  0.6517],
        [-0.2317, -1.1088]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-False-True-none-g4-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'none', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14957508>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 2.1439, -0.7161],
        [-1.1656,  1.5154],
        [ 0.5351, -0.4519],
        [ 0.5981,  2.1657],
      ...6,  1.9419],
        [ 0.6386, -0.2186],
        [ 0.3984,  0.0559],
        [-1.5289, -0.0538]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-False-True-none-g4-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'none', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF030385C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.4237, -0.2309],
        [-0.0696,  0.1275],
        [-0.7958,  0.7236],
        [ 0.5045, -0.2455],
      ...5,  0.2718],
        [ 0.9369, -0.3930],
        [ 0.1960,  0.1728],
        [ 1.2886,  0.5504]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-False-True-none-g5-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'none', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0109E4C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.1341, -1.6704],
        [ 0.5439, -0.2395],
        [ 1.5001, -1.1185],
        [-1.4062,  1.8230],
      ...9, -0.1774],
        [ 0.0616, -0.0739],
        [ 2.6979, -1.7979],
        [-0.1963, -0.2793]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-False-True-none-g5-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'none', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF030A58C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.2082,  3.1299],
        [-0.8607, -0.1224],
        [ 1.1199, -1.7766],
        [-0.8695,  1.0439],
      ...7, -1.9231],
        [ 0.5083,  2.8360],
        [-0.1257, -1.1029],
        [-0.5343, -0.7671]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-False-True-none-g6-idtype0] ________________

idtype = torch.int32, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'none', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1526F648>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.5500,  1.8999],
        [ 0.2105,  2.2328],
        [ 0.6470,  0.5004],
        [ 0.0872, -0.0024],
        [ 0.4570,  2.3549],
        [-0.4528, -0.6453]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-False-True-none-g6-idtype1] ________________

idtype = torch.int64, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'none', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0761A8C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.3161, -2.7572],
        [-0.2388, -1.6739],
        [-1.0710,  1.2060],
        [ 0.1586, -0.4608],
        [ 0.1750,  1.5626],
        [-0.7186, -0.9135]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-False-True-none-g7-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'none', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF010C4288>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.0305,  1.3547],
        [-3.4451,  1.8640],
        [ 3.0705, -1.0544],
        [ 1.4712, -0.0076],
      ...8, -0.0178],
        [ 3.2313, -0.3028],
        [ 1.8590,  0.4222],
        [ 1.7356, -0.2823]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-False-True-none-g7-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'none', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1467DD48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.6157,  2.5450],
        [-0.3667, -1.1611],
        [ 1.2605,  0.8783],
        [-0.7846, -0.4191],
      ...8, -0.0520],
        [-0.6632, -1.0145],
        [-0.6582, -0.2944],
        [-0.7664, -0.0028]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-False-True-both-g0-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'both', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF02FB8EC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.0389,  0.6278],
        [-0.2035,  0.3963]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-False-True-both-g0-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'both', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0110A7C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.2074, -0.2167],
        [ 0.4653,  0.1210]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-False-True-both-g1-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
norm = 'both', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01106D48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.4648,  0.5279],
        [ 0.5429,  0.0861],
        [ 0.6025, -1.0866],
        [-0.8314,  0.7938],
      ...4, -0.5960],
        [ 0.5076,  0.1413],
        [ 0.6201, -0.0464],
        [ 0.0058, -1.0342]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-False-True-both-g1-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
norm = 'both', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0307C748>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.0051,  1.1937],
        [ 0.6452,  2.1683],
        [ 0.2150, -0.9020],
        [-0.6454, -1.3557],
      ...3,  2.0202],
        [ 0.5537, -1.8213],
        [ 1.3704,  1.6704],
        [ 0.7406,  2.0072]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-False-True-both-g2-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'both', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07730F88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.1193, -0.5407],
        [ 0.3473, -0.7138],
        [ 1.2882, -0.7455],
        [-0.4869, -0.2624],
      ...5, -0.4558],
        [ 2.2607, -1.5854],
        [ 1.8564, -0.5718],
        [-0.4792,  1.1571]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-False-True-both-g2-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'both', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01A19548>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.3290, -1.0120],
        [ 1.1802, -0.4449],
        [ 1.2090,  0.3047],
        [ 0.6122,  1.0983],
      ...6,  0.5165],
        [ 1.2386, -1.9762],
        [ 0.6690, -0.2755],
        [-0.6709, -0.1255]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-False-True-both-g3-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'both', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF030AD448>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.1603,  0.7189],
        [ 0.5542,  1.9043],
        [ 0.0085,  0.2670],
        [-1.8678, -0.9349],
      ...6, -1.8488],
        [-1.1848,  0.3490],
        [ 0.3292, -1.9966],
        [ 0.2795,  0.0484]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-False-True-both-g3-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'both', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01120988>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 4.2825e-01,  4.3939e-01],
        [ 1.3273e-03, -6.4144e-01],
        [ 6.6103e-05, -6.5769e-02],
        [-....9850e-01,  7.4838e-01],
        [-2.6344e-01,  6.1650e-02],
        [-5.4277e-01, -9.2936e-01]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-False-True-both-g4-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'both', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07709148>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.6820, -0.4490],
        [-0.6534,  0.4218],
        [ 0.0755, -0.7775],
        [-1.9106,  0.0538],
      ...2, -0.3121],
        [-0.2377, -0.1086],
        [ 1.1542, -0.0073],
        [-2.8631,  0.0444]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-False-True-both-g4-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'both', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE149EE488>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.1875, -0.4811],
        [ 0.0093, -0.4355],
        [-0.1616,  0.5608],
        [-0.1373, -2.3742],
      ...9,  0.4740],
        [ 0.8071,  0.3131],
        [ 0.5745,  1.2973],
        [-0.0358,  0.7617]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-False-True-both-g5-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'both', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF076D2C08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.1346, -0.0168],
        [-0.4085,  0.1094],
        [-0.4616, -0.0596],
        [ 0.4259, -0.4920],
      ...4,  0.0488],
        [ 0.3606,  0.7240],
        [-0.3028, -0.1124],
        [-0.3392, -0.4966]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-False-True-both-g5-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'both', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07727A48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.8723, -0.1084],
        [ 0.3939,  1.5386],
        [ 0.7282, -0.4870],
        [-0.2768, -0.6656],
      ...9, -0.6047],
        [-0.2017,  1.0673],
        [ 0.1335, -0.2963],
        [ 0.3581, -0.2529]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-False-True-both-g6-idtype0] ________________

idtype = torch.int32, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'both', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE152B6B48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.4348, -0.4335],
        [-0.5490,  1.0289],
        [-1.0261,  0.9606],
        [ 1.2647, -1.1168],
        [ 0.6674,  0.8540],
        [ 2.1294, -1.5822]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-False-True-both-g6-idtype1] ________________

idtype = torch.int64, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'both', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE149AB788>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.1681,  1.7924],
        [ 0.3594,  0.1926],
        [-0.9354,  1.6331],
        [ 1.0434, -0.1649],
        [-0.0536, -0.3749],
        [ 1.6989,  0.3665]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-False-True-both-g7-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'both', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01A6F8C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.8653, -0.5314],
        [-0.2706,  0.2322],
        [-1.1174, -0.1664],
        [-0.4866,  0.5918],
      ...1, -0.3811],
        [ 1.8798,  1.7384],
        [ 0.3685, -1.4455],
        [-0.3096,  1.5833]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-False-True-both-g7-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'both', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):

        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1463C2C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.8377, -0.1740],
        [-0.7186,  0.3387],
        [ 0.3076,  0.4532],
        [-0.0313,  0.3506],
      ...1, -0.5528],
        [ 0.1399,  1.0331],
        [ 1.3664,  2.0834],
        [ 0.0207,  0.7577]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-True-right-g0-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'right', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE145725C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.4240, -1.1366],
        [-1.5470,  0.6913]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-True-right-g0-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'right', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01A67408>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-2.0451,  3.5827],
        [-0.2589,  2.9975]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-True-right-g1-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
norm = 'right', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03F16708>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.6752, -2.5473],
        [ 0.3526,  0.1735],
        [ 0.3501, -0.0681],
        [ 0.2961,  0.7414],
      ...4,  0.5054],
        [-0.6150, -0.6288],
        [ 1.1324,  1.4153],
        [ 0.2962,  1.1950]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-True-right-g1-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
norm = 'right', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF011519C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.7639,  0.4116],
        [-0.2727,  2.5790],
        [-0.2827, -0.6160],
        [ 0.3848, -0.6200],
      ...6, -0.2237],
        [-0.1212,  0.1175],
        [-0.1570, -0.4060],
        [-3.8527,  2.4802]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-True-right-g2-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'right', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF076C2208>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.2529,  1.0187],
        [-1.6230,  0.1570],
        [-0.6154, -0.7657],
        [ 0.5829,  1.0349],
      ...6,  0.7011],
        [ 0.7373, -0.8511],
        [-0.0322,  0.4361],
        [ 0.1401, -0.9457]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-True-right-g2-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'right', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01188588>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.0116,  0.3293],
        [ 0.4266,  0.7164],
        [-2.6472,  0.2417],
        [-1.2765, -0.3309],
      ...9,  0.8889],
        [ 0.2066, -1.2834],
        [-2.2381,  1.9153],
        [ 1.3437, -0.1256]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-True-right-g3-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'right', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1498FFC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.0303, -0.8647],
        [-0.8605, -1.6043],
        [ 0.7242,  0.3198],
        [-0.3977, -0.7955],
      ...9, -0.3032],
        [-0.5269, -1.7994],
        [ 0.1449, -1.4156],
        [-1.7704, -3.1813]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-True-right-g3-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'right', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF010A5308>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.9742, -0.5087],
        [ 2.1963,  0.4683],
        [ 0.2532, -0.0683],
        [ 0.0532,  0.4702],
      ...8, -0.0922],
        [-2.2510, -0.9574],
        [-0.1354,  0.0443],
        [-0.3814, -0.6129]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-True-right-g4-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'right', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14616608>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.5364, -1.2618],
        [-0.5551,  3.1570],
        [-0.1650,  0.9120],
        [ 0.6708, -0.8039],
      ...5,  1.3847],
        [-0.6832,  1.9946],
        [ 0.7778,  0.4979],
        [ 0.3069, -0.8330]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-True-right-g4-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'right', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF076EBC48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.6919, -0.9632],
        [-0.1232,  0.4907],
        [ 1.3205,  0.7169],
        [-1.1889,  1.4029],
      ...3,  1.3755],
        [-0.8241,  0.8603],
        [-0.6761,  0.2074],
        [-0.0864,  0.9357]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-True-right-g5-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'right', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1460F908>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.8144,  0.7840],
        [ 1.1890, -0.6397],
        [ 0.0350,  0.0711],
        [ 1.7231, -1.1851],
      ...6,  0.7389],
        [ 1.5516, -0.3109],
        [-1.1854, -0.0660],
        [ 1.6333, -0.0695]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-True-right-g5-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'right', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE146E4D88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.3191, -0.7308],
        [ 0.0545, -0.4415],
        [ 0.0436, -1.2048],
        [-0.5730, -2.0932],
      ...3, -2.1833],
        [-1.5842, -1.0602],
        [ 0.9245, -0.6170],
        [-2.4345,  1.1601]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-True-right-g6-idtype0] ________________

idtype = torch.int32, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'right', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01AB1488>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.4931,  0.7138],
        [-0.4935, -0.6949],
        [-1.8244, -0.3302],
        [-0.8621,  0.2541],
        [ 0.3985, -0.6081],
        [ 1.4703, -1.0107]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-True-right-g6-idtype1] ________________

idtype = torch.int64, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'right', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE145F7A88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.2967, -0.7216],
        [-0.2002,  1.1152],
        [-1.0998,  0.2813],
        [ 1.0353,  1.8115],
        [ 0.9512,  1.7621],
        [ 0.2625, -0.6149]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-True-right-g7-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'right', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01166E88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.2494, -0.1231],
        [ 0.5814,  0.2966],
        [-1.5415, -1.7939],
        [-0.5797, -0.3784],
      ...9,  0.3132],
        [ 2.2392,  2.7090],
        [-0.0607, -0.3020],
        [ 0.3496, -0.0245]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-True-right-g7-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'right', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14580E88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.5918, -2.1120],
        [ 0.1882,  1.9812],
        [ 0.5627,  1.6168],
        [ 0.0852,  1.0712],
      ...3,  0.5160],
        [ 1.2522,  0.9576],
        [ 0.5239,  0.0407],
        [-0.7875,  1.1625]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-False-True-left-g0-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'left', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03E98808>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.3050,  0.2297],
        [-0.2176,  0.3102]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-False-True-left-g0-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'left', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0768A408>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.1752, -0.1806],
        [-0.6860, -0.0584]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-False-True-left-g1-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
norm = 'left', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14612548>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 4.8981e-02, -4.4104e-01],
        [-3.8759e-01,  4.3823e-01],
        [ 3.1159e-02, -1.5059e-01],
        [ ....1572e+00,  1.1781e+00],
        [-1.2046e-01,  2.9382e-01],
        [ 1.8282e+00, -6.3055e-01]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-False-True-left-g1-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
norm = 'left', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF030AF888>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.1617, -0.2199],
        [ 0.0979,  0.5250],
        [-0.0620, -0.1346],
        [-0.1614,  0.0690],
      ...1,  0.4598],
        [-1.4263, -0.0452],
        [ 0.6341,  0.7389],
        [-0.1900, -0.7948]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-False-True-left-g2-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'left', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14597FC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.3375,  0.4805],
        [-0.4332, -0.3284],
        [ 0.0827,  0.2150],
        [-0.0601, -0.1557],
      ...3, -0.2872],
        [ 1.3127, -0.8356],
        [ 1.4305, -1.2473],
        [-2.5683,  0.9713]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-False-True-left-g2-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'left', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01130088>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.3082, -0.2826],
        [ 0.1624, -0.3350],
        [ 0.3857, -0.5098],
        [-0.3554,  0.3339],
      ...0,  0.4494],
        [-2.1937,  2.1234],
        [ 0.5412, -0.8792],
        [ 0.8849, -1.0831]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-False-True-left-g3-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'left', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE149FE4C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.0160,  0.0772],
        [-0.1781, -0.1339],
        [-0.4327,  0.4704],
        [-0.5369,  0.0239],
      ...0,  0.0135],
        [ 0.3294, -0.7339],
        [-1.0114, -2.6865],
        [ 0.8242,  0.1741]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-False-True-left-g3-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'left', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01182AC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.4314,  0.0689],
        [ 0.0121, -1.4098],
        [ 0.1987,  0.7428],
        [-0.7349, -0.2078],
      ...6,  0.7898],
        [-0.5178,  2.1117],
        [ 2.3365,  0.2057],
        [-1.0874,  0.2481]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-False-True-left-g4-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'left', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE149DAA88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.0633, -0.5293],
        [ 0.0563, -0.1972],
        [-0.3883,  0.5232],
        [-0.1065,  1.4957],
      ...0, -0.1866],
        [-1.2393, -1.7269],
        [ 0.7289, -0.1232],
        [ 1.1433,  0.6491]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-False-True-left-g4-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'left', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE151F0C08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.9231, -0.0979],
        [ 0.7230,  0.1390],
        [ 0.4874, -0.5770],
        [-0.0884, -0.6442],
      ...9,  0.2313],
        [ 0.2447,  2.5220],
        [-0.0604, -1.5192],
        [ 0.4670,  0.4919]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-False-True-left-g5-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'left', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14613A88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.3340, -0.5855],
        [-0.5381, -0.4496],
        [ 0.3098, -0.3713],
        [-0.2592, -0.2249],
      ...2, -0.8831],
        [ 0.7138,  1.9002],
        [ 1.0691, -1.2688],
        [-0.4363,  1.8772]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-False-True-left-g5-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'left', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE152FF888>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.0972,  0.1028],
        [ 0.0324,  0.8732],
        [ 0.0306,  0.0273],
        [ 0.0989, -0.0242],
      ...2,  0.8732],
        [-1.9474, -1.4611],
        [-0.8505,  2.6139],
        [-0.3851,  1.4261]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-False-True-left-g6-idtype0] ________________

idtype = torch.int32, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'left', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE145B0A48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.6894,  0.2529],
        [-0.0468, -0.0838],
        [-2.9570,  1.2054],
        [-1.8181,  0.8335],
        [-1.1607, -0.9860],
        [-1.2959, -1.5112]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-False-True-left-g6-idtype1] ________________

idtype = torch.int64, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'left', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03036148>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.1842, -0.5037],
        [-0.1779,  0.3079],
        [ 0.3610, -0.2555],
        [-1.3150, -0.2715],
        [ 0.8261,  1.6040],
        [ 1.5260,  0.3522]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-False-True-left-g7-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'left', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1463E9C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.0089,  0.1631],
        [-1.1527,  0.2998],
        [ 0.0789,  0.2814],
        [-1.1719, -0.0186],
      ...8, -0.2912],
        [ 0.4564, -1.2632],
        [-0.7116,  0.2205],
        [ 0.7336, -0.1895]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_graph_conv[2-False-True-left-g7-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'left', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
>           h_out = conv(g, h)

tests\pytorch\test_nn.py:97: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14CBD888>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.0175,  0.4816],
        [-1.3499,  1.0579],
        [ 0.0284,  0.2270],
        [-0.1736,  0.8167],
      ...3, -0.4947],
        [ 1.8826, -1.3539],
        [ 0.9728,  0.0144],
        [ 0.0260,  0.5413]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-False-none-g0-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'none', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1526F308>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[2.5473, 0.1615],
        [3.5480, 2.2929]]), rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-False-none-g0-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'none', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14A5D4C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.8419, -1.6728],
        [-4.8046, -1.7068]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-False-none-g1-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
norm = 'none', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE15302A88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.7178,  1.6328],
        [-1.7075,  2.9108],
        [ 0.4732,  0.9238],
        [-0.8957, -0.9904],
      ...2658],
        [ 2.4376, -2.7762],
        [ 2.7264,  3.0805],
        [ 1.0320,  0.3583],
        [-1.6805,  1.8533]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-False-none-g1-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
norm = 'none', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14A49308>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 2.4063, -0.2400],
        [-0.4425, -0.3523],
        [ 2.0322,  0.6887],
        [-0.8212, -1.9006],
      ...5500],
        [ 1.2800,  0.3213],
        [ 3.4586,  0.0717],
        [ 6.3578, -0.8477],
        [-0.6974,  0.3242]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-False-none-g2-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'none', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF011CB408>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.7053,  0.9560],
        [-1.3994, -1.0515],
        [-1.6046,  0.5246],
        [ 1.1110, -1.2590],
      ...8153],
        [-0.6104,  0.2701],
        [ 1.6053, -0.2188],
        [-2.3104, -1.1331],
        [-3.5968, -1.1184]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-False-none-g2-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'none', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03F1DCC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.8718,  2.2248],
        [-0.8780,  1.3736],
        [-0.9603, -0.6239],
        [-2.4316,  0.8264],
      ...5128],
        [ 2.7869,  0.5676],
        [ 1.0615,  1.0202],
        [-0.4034, -0.6049],
        [-0.3782, -2.4210]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-False-none-g3-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'none', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF011A2EC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.7611, -5.2887],
        [ 0.5501, -0.7775],
        [-1.1504, -3.9606],
        [-0.4306,  0.3241],
      ...9489],
        [ 0.1746, -0.0360],
        [-0.9172, -1.8499],
        [ 0.1969, -1.1318],
        [-0.3348, -0.6924]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-False-none-g3-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'none', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03F34048>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-3.6499,  3.3221],
        [-2.4210,  3.2537],
        [ 0.9588, -0.8004],
        [ 1.4358, -2.7597],
      ...4114],
        [-3.1661,  5.0269],
        [ 0.5740,  3.1312],
        [-2.4129,  2.4621],
        [-2.3120, -2.6002]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-False-none-g4-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'none', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07709648>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.7448,  2.1082],
        [ 0.0417,  0.1825],
        [ 0.9927, -0.9574],
        [-2.1265,  2.3450],
      ...0187],
        [ 1.6276, -0.8070],
        [-0.8368, -0.1574],
        [ 0.6187, -0.7437],
        [-0.1563,  0.8012]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-False-none-g4-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'none', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03EC7AC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.0844,  0.3951],
        [ 0.0107,  3.4678],
        [-1.9365,  5.7447],
        [-2.1017, -0.2551],
      ...7859],
        [ 2.7314, -2.6363],
        [-1.3374,  0.1929],
        [-0.1072,  1.3035],
        [ 0.1151,  2.3880]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-False-none-g5-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'none', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE15BBA448>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.1341,  0.1142],
        [ 0.4631,  2.6083],
        [-1.9624,  1.4488],
        [ 0.0698, -0.0227],
      ...3957],
        [ 2.3019, -1.4060],
        [-0.0903,  2.1176],
        [ 0.0330,  0.5572],
        [ 1.4720, -4.0337]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-False-none-g5-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'none', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03F16F08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.4857,  1.8935],
        [-1.9614, -2.7994],
        [ 4.5082,  0.9465],
        [-6.4295,  2.9452],
      ...4784],
        [ 1.3294,  2.8282],
        [ 1.0420,  0.9507],
        [ 5.8315,  0.7371],
        [-4.4010, -1.4021]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-False-none-g6-idtype0] ________________

idtype = torch.int32, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'none', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE144F7688>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-4.1252,  0.0298],
        [-6.8054,  0.4658],
        [ 3.5103, -0.5221],
        [-1.4110,  1.3537],
        [ 3.2295, -1.4304],
        [ 4.4501, -0.3283]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-False-none-g6-idtype1] ________________

idtype = torch.int64, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'none', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0772D048>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.6149,  2.7836],
        [ 0.1117, -0.2028],
        [-2.4095,  0.7754],
        [-0.3239,  1.3242],
        [ 0.0970, -0.3163],
        [ 0.2447, -1.0941]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-False-none-g7-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'none', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0109E988>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.2585,  0.2860],
        [ 0.7690,  0.4950],
        [-0.2350,  1.2129],
        [-0.2653, -0.2115],
      ...7883],
        [-1.3150,  0.5654],
        [-1.2211, -3.8940],
        [-0.0259,  0.0337],
        [ 2.1553,  3.5246]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-False-none-g7-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'none', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14CA7C48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-2.5139,  0.1272],
        [ 2.9486, -0.6920],
        [-2.5253, -3.9679],
        [-1.3013, -1.8208],
      ...2864],
        [ 1.2000, -1.3302],
        [ 4.3518, -0.3953],
        [ 3.2044,  2.2777],
        [-0.6825,  0.4838]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-False-both-g0-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'both', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE15360108>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.3533, -0.4085],
        [ 0.0612, -0.3107]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-False-both-g0-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'both', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03046F48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.0986, -0.4261],
        [ 1.4830,  0.7016]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-False-both-g1-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
norm = 'both', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE146E3B08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.5055, -0.1786],
        [-0.8816,  0.5487],
        [ 0.5423, -0.0873],
        [-0.2318, -1.0297],
      ...3018],
        [-4.4126,  0.8305],
        [-2.1958,  0.8466],
        [ 0.0978,  1.4832],
        [ 1.3188, -1.5229]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-False-both-g1-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
norm = 'both', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14A5E188>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.7161,  0.9719],
        [-3.0068,  1.1356],
        [ 0.1645,  0.1861],
        [-1.2578,  0.8632],
      ...9950],
        [ 1.4026,  0.3198],
        [-3.6763, -0.2235],
        [-2.3301, -1.1614],
        [-0.9197, -1.1399]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-False-both-g2-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'both', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE15BA54C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.8245,  0.5419],
        [-1.0245,  2.5764],
        [-0.2129,  1.1025],
        [ 1.3001,  0.1362],
      ...6042],
        [ 0.7959,  1.6844],
        [-2.7224,  2.8578],
        [ 3.1837,  1.1586],
        [ 1.0437, -0.3545]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-False-both-g2-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'both', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF075E8AC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.0814, -0.1032],
        [ 0.0300, -0.1650],
        [ 0.4687, -0.4336],
        [ 0.3893,  0.1429],
      ...8348],
        [-0.4999,  0.3089],
        [-0.5779,  0.9147],
        [-2.0823,  0.7812],
        [ 0.1672, -0.6784]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-False-both-g3-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'both', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF02F4E588>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.1059,  0.8153],
        [ 0.2476, -1.0040],
        [ 0.4853, -0.5580],
        [-0.5773, -0.3269],
      ...6296],
        [-2.1075, -0.0179],
        [-3.4991, -2.4528],
        [ 0.4553,  0.3813],
        [-2.4641, -0.1436]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-False-both-g3-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'both', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14A79C88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.2843,  0.2721],
        [-0.2377, -0.4289],
        [-0.1065,  0.1037],
        [-0.8728, -0.2120],
      ...6507],
        [ 0.0280,  1.0468],
        [-1.4972,  0.8241],
        [-0.4449, -0.0504],
        [-0.1437, -1.1017]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-False-both-g4-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'both', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE121A7788>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.8701, -1.6601],
        [ 3.0653,  2.0465],
        [-1.1553, -0.2288],
        [-4.5052, -2.9616],
      ...5885],
        [-0.3941,  1.8723],
        [-0.0490,  0.8324],
        [ 1.4819,  0.4298],
        [-2.3221, -0.5813]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-False-both-g4-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'both', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF010AEC48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.7461, -0.6506],
        [ 1.8525, -0.0516],
        [-0.5321,  1.0190],
        [ 0.3271, -1.4910],
      ...4728],
        [-1.5314, -0.1121],
        [-1.8178,  0.1759],
        [-0.4576, -0.4630],
        [ 1.5889,  1.0866]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-False-both-g5-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'both', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14616C48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.0214, -1.1046],
        [-0.5760, -0.0508],
        [-1.8635, -0.4690],
        [-0.5241,  2.0931],
      ...5362],
        [ 1.8254,  1.0294],
        [-1.1921, -3.8580],
        [-1.3183,  1.8634],
        [-1.4641, -0.2206]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-False-both-g5-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'both', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE15BA6488>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.8642,  0.0354],
        [ 0.5945,  0.5043],
        [ 0.4947, -0.0327],
        [ 2.1413, -0.5389],
      ...0137],
        [ 0.4641,  0.5711],
        [-1.0487,  0.7903],
        [ 2.2448, -0.4297],
        [-1.1508,  1.3551]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-False-both-g6-idtype0] ________________

idtype = torch.int32, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'both', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE144F7188>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.7792, -0.9386],
        [ 1.2711, -1.5208],
        [ 0.1815,  1.5126],
        [-0.4756,  3.2999],
        [-0.1464,  1.6148],
        [ 0.1345, -1.7093]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-False-both-g6-idtype1] ________________

idtype = torch.int64, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'both', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE15339E48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.1758,  0.4340],
        [-0.8091,  0.0600],
        [-2.2226,  2.3948],
        [ 1.2565,  0.2738],
        [-4.3683,  0.8923],
        [ 0.0251, -0.3826]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-False-both-g7-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'both', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14B4F848>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.7341, -0.5517],
        [ 3.9555,  1.2566],
        [ 1.2910,  0.5469],
        [-2.6206, -0.5164],
      ...6637],
        [-2.3861,  0.7074],
        [-3.4838,  1.4221],
        [-4.6070, -0.0627],
        [ 0.0761,  2.4854]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-False-both-g7-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'both', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1459A3C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-3.3476,  0.8941],
        [ 0.4025,  2.4219],
        [-2.3476,  0.7798],
        [ 1.0799,  2.2725],
      ...6172],
        [-0.1561, -1.3979],
        [ 0.3715, -1.5823],
        [-2.2653, -0.0685],
        [ 0.7197, -2.4360]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-False-right-g0-idtype0] _______________

idtype = torch.int32
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'right', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14A44288>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.6458, -0.4761],
        [-0.2568,  3.9496]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-False-right-g0-idtype1] _______________

idtype = torch.int64
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'right', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE145EDC88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.6614,  2.3261],
        [-1.1673, -3.7749]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-False-right-g1-idtype0] _______________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
norm = 'right', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14613248>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.2938,  1.4717],
        [ 1.2816,  0.1879],
        [ 2.1415,  1.3154],
        [ 2.3745,  1.1183],
      ...7598],
        [-3.6644, -2.6267],
        [-2.3141, -0.7239],
        [-2.0334, -0.2522],
        [-0.4541,  0.5085]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-False-right-g1-idtype1] _______________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
norm = 'right', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF02F2A848>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.0998,  0.1522],
        [ 2.5539,  0.5601],
        [-2.6954, -0.3529],
        [-1.1412, -0.0461],
      ...6624],
        [ 0.5106, -0.6169],
        [-0.2095, -0.0815],
        [-1.3676, -0.2454],
        [ 1.3715,  0.6018]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-False-right-g2-idtype0] _______________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'right', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE15C16388>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.1656,  2.8663],
        [ 0.1766, -1.8682],
        [-1.1478, -0.4058],
        [-0.7478, -0.5165],
      ...4722],
        [ 0.8284,  0.2661],
        [ 2.9640,  1.0675],
        [ 0.7402, -0.0840],
        [ 0.4492, -0.4228]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-False-right-g2-idtype1] _______________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'right', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF011373C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-2.8395,  0.0984],
        [ 2.2624,  1.6085],
        [-0.4491,  3.3438],
        [-1.4750,  3.2735],
      ...1617],
        [-2.8793,  0.5710],
        [-4.0139, -2.5488],
        [-3.3310, -2.0883],
        [-2.4604, -0.0887]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-False-right-g3-idtype0] _______________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'right', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14AC22C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.2971, -0.8830],
        [-1.9560,  2.3729],
        [ 0.8491,  3.3576],
        [ 0.8262,  0.9050],
      ...2257],
        [ 7.8312, -1.5554],
        [-3.7430,  1.9351],
        [ 1.8630,  1.7098],
        [ 2.6070, -1.3754]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-False-right-g3-idtype1] _______________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'right', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14649588>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.4388,  0.2759],
        [ 1.4506, -2.7567],
        [ 0.9826, -3.2315],
        [-3.1449,  3.2906],
      ...6432],
        [-0.0799, -0.5410],
        [-0.0877, -0.5883],
        [-0.9023,  0.9301],
        [ 0.5053, -0.4283]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-False-right-g4-idtype0] _______________

idtype = torch.int32
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'right', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0109FE08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.0586,  0.2199],
        [ 0.4763,  1.6019],
        [ 1.9474,  0.0352],
        [-0.5688, -0.1214],
      ...9607],
        [ 0.4134, -0.6467],
        [-2.5958, -0.1549],
        [-1.8618, -1.8620],
        [ 1.1341,  0.1197]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-False-right-g4-idtype1] _______________

idtype = torch.int64
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'right', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE149D5B88>
op = 'copy_lhs', reduce_op = 'sum'

        [-0.8282, -4.8018],
        [-0.0452, -0.1485],
        [ 0.9104,  1.2502],
      ...3353],
        [-0.1639, -4.1885],
        [-2.4279,  1.4491],
        [ 0.5827,  2.1371],
        [ 0.2090,  4.7612]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-False-right-g5-idtype0] _______________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'right', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0110A548>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-2.6581, -1.8333],
        [-0.1267,  0.7648],
        [-3.2151, -0.0341],
        [-3.5960, -1.9909],
      ...4440],
        [-2.4528, -3.1929],
        [-4.8574, -1.3313],
        [ 2.0699,  0.4109],
        [ 1.0389,  1.3800]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-False-right-g5-idtype1] _______________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'right', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03F43388>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.3250,  0.4656],
        [ 0.3167, -0.8636],
        [-0.3144,  2.5902],
        [-0.2223,  1.2967],
      ...4113],
        [-1.7959,  3.2753],
        [-0.3183, -0.4363],
        [ 1.0781, -3.3787],
        [ 0.4418, -0.7608]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-False-right-g6-idtype0] _______________

idtype = torch.int32, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'right', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01A7F6C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.3218,  6.7626],
        [ 0.8358,  1.3400],
        [ 1.0052,  7.4336],
        [-1.7378,  1.1728],
        [-0.1023,  0.9432],
        [-0.9876, -1.3587]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-False-right-g6-idtype1] _______________

idtype = torch.int64, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'right', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE145DCF08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.4109, -0.9929],
        [-1.0476, -0.0363],
        [ 0.9144,  1.8973],
        [ 1.4533,  0.5308],
        [-0.5698, -0.1699],
        [-1.0061,  0.6821]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-False-right-g7-idtype0] _______________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'right', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14C34D08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.7628,  4.4383],
        [-2.5359,  1.5889],
        [ 0.8011, -0.6141],
        [-2.8351, -1.9287],
      ...4708],
        [-2.3146, -7.5274],
        [-2.8181, -4.1121],
        [ 0.5941,  0.1342],
        [-0.2040,  5.0609]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-False-right-g7-idtype1] _______________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'right', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF02F77848>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.5384, -0.9701],
        [ 2.6945, -0.0529],
        [-2.8188,  0.5542],
        [-4.3822, -0.9978],
      ...9662],
        [ 0.0588, -1.2434],
        [-0.4687, -1.5735],
        [-1.5605,  0.1412],
        [ 2.9055,  2.3679]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-False-left-g0-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'left', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14957108>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.3850,  0.2387],
        [ 0.4232, -0.1218]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-False-left-g0-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'left', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF010B6388>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.7654, -0.5863],
        [-0.4024, -0.8502]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-False-left-g1-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
norm = 'left', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE145DFE48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.0915, -0.5824],
        [-0.3154, -0.8315],
        [-1.3535,  0.7389],
        [ 0.2907, -0.2688],
      ...7046],
        [-1.1573,  1.1132],
        [ 1.7272, -4.4239],
        [-0.5765,  3.0220],
        [-0.7804, -0.5613]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-False-left-g1-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
norm = 'left', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE149D5948>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.3865,  0.0834],
        [-0.5287, -1.0291],
        [ 1.1623,  0.0241],
        [ 0.7479,  0.5175],
      ...0874],
        [-0.6089,  1.6730],
        [ 0.4392,  1.4046],
        [-2.9566,  2.1303],
        [ 2.5902, -1.5457]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-False-left-g2-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'left', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14B58D88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.9227, -0.6885],
        [-2.0334, -1.5706],
        [ 1.1277,  1.7209],
        [-0.4361, -0.3407],
      ...6836],
        [ 1.4991,  1.2031],
        [-0.9522, -4.6665],
        [-1.2120, -0.0138],
        [ 1.6207, -0.5767]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-False-left-g2-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'left', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1462DA48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.0107e-01, -2.4149e-01],
        [-5.3020e-01,  1.6443e-01],
        [-5.8602e-01, -2.0977e-03],
        [ ...4478e-01],
        [-2.8991e+00, -2.9585e-01],
        [ 1.6705e+00,  8.7000e-02],
        [ 7.0896e-01, -5.2461e-01]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-False-left-g3-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'left', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1535AFC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.6853,  0.1328],
        [-1.7697,  0.5352],
        [-0.4654, -0.2123],
        [ 0.4295,  0.6319],
      ...6685],
        [ 0.4499, -1.5083],
        [ 1.7482, -0.3459],
        [ 5.4792, -2.0370],
        [-3.5047,  1.8903]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-False-left-g3-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'left', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01156648>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.3934,  0.1329],
        [ 0.7780, -0.0773],
        [-0.5195, -0.2756],
        [-0.7194,  0.8215],
      ...8474],
        [-1.0128, -1.4158],
        [-0.6066,  1.1560],
        [ 2.2824,  1.5496],
        [ 3.3647,  0.5634]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-False-left-g4-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'left', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14CE3F48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 2.1560,  0.1863],
        [-0.3192,  0.1991],
        [-0.4113, -0.0415],
        [ 0.5920, -0.0412],
      ...2683],
        [-0.5136,  0.2055],
        [-1.1226, -1.9548],
        [ 0.3002, -0.3637],
        [-3.4102,  3.2545]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-False-left-g4-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'left', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03EA9048>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.3099, -2.4330],
        [ 0.6095, -0.0643],
        [-0.1694,  1.5040],
        [-0.4595, -4.0571],
      ...8606],
        [-1.1805, -1.0198],
        [ 2.5866,  0.7669],
        [-0.1388,  0.9306],
        [-2.6197, -2.8633]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-False-left-g5-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'left', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03E87D08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.3822, -0.5067],
        [-0.6998,  1.3621],
        [ 1.5377,  0.3518],
        [ 1.2337, -0.0359],
      ...2095],
        [-0.9856,  1.8726],
        [ 1.4337,  2.7913],
        [-0.8470,  0.9216],
        [-2.0844,  1.2601]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-False-left-g5-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'left', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01156388>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-3.7128e-02, -2.2927e-01],
        [ 2.2885e-01,  3.2371e-01],
        [ 1.0472e-01, -5.1541e-01],
        [-...7833e-02],
        [ 1.0966e+00, -1.0675e+00],
        [-3.6253e+00,  5.1798e+00],
        [ 3.3043e+00,  2.0403e+00]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-False-left-g6-idtype0] ________________

idtype = torch.int32, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'left', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01169E48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.4307, -0.3076],
        [ 2.0112,  2.4887],
        [-0.1668, -0.6595],
        [ 0.8482,  0.1906],
        [ 1.9205,  3.8002],
        [ 1.1632, -0.2879]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-False-left-g6-idtype1] ________________

idtype = torch.int64, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'left', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14A24288>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-2.5379, -3.8132],
        [-1.6588, -1.1660],
        [ 0.3637, -1.8675],
        [-0.6216, -0.0696],
        [ 3.9410,  1.7853],
        [-0.2589, -0.4759]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-False-left-g7-idtype0] ________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'left', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE15BA51C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.3085,  1.2705],
        [ 0.0701, -1.2509],
        [ 0.6480, -0.6933],
        [-0.1034, -0.0729],
      ...5920],
        [ 1.4725,  0.7000],
        [ 2.1536, -5.1331],
        [ 1.9001, -0.4455],
        [-2.9530,  3.3822]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv[2-False-False-left-g7-idtype1] ________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
norm = 'left', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right', 'left'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv(idtype, g, norm, weight, bias, out_dim):
        # Test one tensor input
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        if weight:
            h_out = conv(g, h)
        else:
>           h_out = conv(g, h, weight=ext_w)

tests\pytorch\test_nn.py:99: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14594B88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.2622, -0.1059],
        [-0.4994,  0.4790],
        [ 0.3466, -0.3163],
        [-0.2021,  0.3678],
      ...3892],
        [ 0.2216, -0.4702],
        [-0.1176,  0.4874],
        [ 0.0395,  0.2552],
        [-0.6466, -0.1235]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
____________ test_graph_conv_e_weight[1-True-True-none-g0-idtype0] ____________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32)})
norm = 'none', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        e_w = g.edata['scalar_w']
        if weight:
>           h_out = conv(g, h, edge_weight=e_w)

tests\pytorch\test_nn.py:117: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF076EB6C8>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[ 0.4877],
        [ 0.0529],
        [-0.3414],
        [ 1.2382],
        [-0.1430],
        [ 0.3928],
        [ 1.2473],
        [ 0.8224],
        [-0.9995],
        [ 0.4580]], grad_fn=<ViewBackward>)
rhs_data = tensor([[0.2636],
        [0.6412],
        [0.6275],
        [1.5607],
        [1.2491],
        [0.7272],
        [0... [1.4196],
        [2.5846],
        [0.5729],
        [1.4330],
        [0.6490],
        [0.3398],
        [0.1452]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
____________ test_graph_conv_e_weight[1-True-True-none-g0-idtype1] ____________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32)})
norm = 'none', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        e_w = g.edata['scalar_w']
        if weight:
>           h_out = conv(g, h, edge_weight=e_w)

tests\pytorch\test_nn.py:117: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0115A508>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[-0.3279],
        [-0.3252],
        [-0.8284],
        [-0.3574],
        [ 0.1058],
        [ 0.2615],
        [-0.0346],
        [ 0.1276],
        [ 0.3270],
        [ 0.0133]], grad_fn=<ViewBackward>)
rhs_data = tensor([[0.2636],
        [0.6412],
        [0.6275],
        [1.5607],
        [1.2491],
        [0.7272],
        [0... [1.4196],
        [2.5846],
        [0.5729],
        [1.4330],
        [0.6490],
        [0.3398],
        [0.1452]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
____________ test_graph_conv_e_weight[1-True-True-both-g0-idtype0] ____________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32)})
norm = 'both', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        e_w = g.edata['scalar_w']
        if weight:
>           h_out = conv(g, h, edge_weight=e_w)

tests\pytorch\test_nn.py:117: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03F34288>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[-0.3330],
        [ 1.1824],
        [-1.4395],
        [ 0.1145],
        [ 1.0097],
        [ 1.4174],
        [ 0.7604],
        [-0.7124],
        [ 1.0790],
        [-0.9422]], grad_fn=<ViewBackward>)
rhs_data = tensor([[0.2636],
        [0.6412],
        [0.6275],
        [1.5607],
        [1.2491],
        [0.7272],
        [0... [1.4196],
        [2.5846],
        [0.5729],
        [1.4330],
        [0.6490],
        [0.3398],
        [0.1452]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
____________ test_graph_conv_e_weight[1-True-True-both-g0-idtype1] ____________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32)})
norm = 'both', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        e_w = g.edata['scalar_w']
        if weight:
>           h_out = conv(g, h, edge_weight=e_w)

tests\pytorch\test_nn.py:117: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01AF58C8>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[-0.7415],
        [ 1.2812],
        [-0.9311],
        [-0.4101],
        [-0.0374],
        [ 0.2462],
        [ 1.0377],
        [ 0.7239],
        [ 2.0212],
        [ 0.3814]], grad_fn=<ViewBackward>)
rhs_data = tensor([[0.2636],
        [0.6412],
        [0.6275],
        [1.5607],
        [1.2491],
        [0.7272],
        [0... [1.4196],
        [2.5846],
        [0.5729],
        [1.4330],
        [0.6490],
        [0.3398],
        [0.1452]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
___________ test_graph_conv_e_weight[1-True-True-right-g0-idtype0] ____________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32)})
norm = 'right', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        e_w = g.edata['scalar_w']
        if weight:
>           h_out = conv(g, h, edge_weight=e_w)

tests\pytorch\test_nn.py:117: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF030346C8>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[-1.5052],
        [ 1.3803],
        [-1.1570],
        [-1.1289],
        [-0.5888],
        [-1.0042],
        [ 0.8109],
        [ 0.2467],
        [ 1.1267],
        [ 0.0591]], grad_fn=<ViewBackward>)
rhs_data = tensor([[0.2636],
        [0.6412],
        [0.6275],
        [1.5607],
        [1.2491],
        [0.7272],
        [0... [1.4196],
        [2.5846],
        [0.5729],
        [1.4330],
        [0.6490],
        [0.3398],
        [0.1452]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
___________ test_graph_conv_e_weight[1-True-True-right-g0-idtype1] ____________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32)})
norm = 'right', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        e_w = g.edata['scalar_w']
        if weight:
>           h_out = conv(g, h, edge_weight=e_w)

tests\pytorch\test_nn.py:117: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE152AF308>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[-1.0505],
        [ 0.0183],
        [ 0.0168],
        [ 0.8952],
        [-1.0532],
        [ 1.4961],
        [-1.0853],
        [ 0.6983],
        [-0.5974],
        [-0.1409]], grad_fn=<ViewBackward>)
rhs_data = tensor([[0.2636],
        [0.6412],
        [0.6275],
        [1.5607],
        [1.2491],
        [0.7272],
        [0... [1.4196],
        [2.5846],
        [0.5729],
        [1.4330],
        [0.6490],
        [0.3398],
        [0.1452]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
___________ test_graph_conv_e_weight[1-True-False-none-g0-idtype0] ____________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32)})
norm = 'none', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        e_w = g.edata['scalar_w']
        if weight:
            h_out = conv(g, h, edge_weight=e_w)
        else:
>           h_out = conv(g, h, weight=ext_w, edge_weight=e_w)

tests\pytorch\test_nn.py:119: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14686448>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[-0.7455],
        [-5.2148],
        [-1.7958],
        [ 4.3237],
        [ 2.0967],
        [-0.6206],
        [ 2.2373],
        [-5.3644],
        [ 2.3700],
        [-5.5190]])
rhs_data = tensor([[0.2636],
        [0.6412],
        [0.6275],
        [1.5607],
        [1.2491],
        [0.7272],
        [0... [1.4196],
        [2.5846],
        [0.5729],
        [1.4330],
        [0.6490],
        [0.3398],
        [0.1452]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
___________ test_graph_conv_e_weight[1-True-False-none-g0-idtype1] ____________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32)})
norm = 'none', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        e_w = g.edata['scalar_w']
        if weight:
            h_out = conv(g, h, edge_weight=e_w)
        else:
>           h_out = conv(g, h, weight=ext_w, edge_weight=e_w)

tests\pytorch\test_nn.py:119: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14A49C08>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[-4.6659],
        [ 0.6509],
        [ 0.4283],
        [ 0.0513],
        [-5.1421],
        [-0.8258],
        [-1.4878],
        [ 0.8601],
        [ 1.3723],
        [-0.5258]])
rhs_data = tensor([[0.2636],
        [0.6412],
        [0.6275],
        [1.5607],
        [1.2491],
        [0.7272],
        [0... [1.4196],
        [2.5846],
        [0.5729],
        [1.4330],
        [0.6490],
        [0.3398],
        [0.1452]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
___________ test_graph_conv_e_weight[1-True-False-both-g0-idtype0] ____________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32)})
norm = 'both', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        e_w = g.edata['scalar_w']
        if weight:
            h_out = conv(g, h, edge_weight=e_w)
        else:
>           h_out = conv(g, h, weight=ext_w, edge_weight=e_w)

tests\pytorch\test_nn.py:119: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07632E48>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[-1.7335],
        [-2.2185],
        [ 0.7887],
        [ 0.0138],
        [-1.7915],
        [ 1.1615],
        [ 1.1113],
        [ 0.5282],
        [ 1.6656],
        [ 0.0082]])
rhs_data = tensor([[0.2636],
        [0.6412],
        [0.6275],
        [1.5607],
        [1.2491],
        [0.7272],
        [0... [1.4196],
        [2.5846],
        [0.5729],
        [1.4330],
        [0.6490],
        [0.3398],
        [0.1452]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
___________ test_graph_conv_e_weight[1-True-False-both-g0-idtype1] ____________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32)})
norm = 'both', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        e_w = g.edata['scalar_w']
        if weight:
            h_out = conv(g, h, edge_weight=e_w)
        else:
>           h_out = conv(g, h, weight=ext_w, edge_weight=e_w)

tests\pytorch\test_nn.py:119: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03EE82C8>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[ 1.6404],
        [-1.2775],
        [-0.7442],
        [-0.2700],
        [-2.0586],
        [-1.2875],
        [-0.6615],
        [ 1.0923],
        [-1.3595],
        [-2.4035]])
rhs_data = tensor([[0.2636],
        [0.6412],
        [0.6275],
        [1.5607],
        [1.2491],
        [0.7272],
        [0... [1.4196],
        [2.5846],
        [0.5729],
        [1.4330],
        [0.6490],
        [0.3398],
        [0.1452]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
___________ test_graph_conv_e_weight[1-True-False-right-g0-idtype0] ___________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32)})
norm = 'right', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        e_w = g.edata['scalar_w']
        if weight:
            h_out = conv(g, h, edge_weight=e_w)
        else:
>           h_out = conv(g, h, weight=ext_w, edge_weight=e_w)

tests\pytorch\test_nn.py:119: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14A0A888>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[ 2.2432],
        [ 2.9197],
        [ 1.4772],
        [ 1.2968],
        [ 4.6916],
        [ 1.6431],
        [-2.4121],
        [ 5.2912],
        [-1.0122],
        [ 1.7813]])
rhs_data = tensor([[0.2636],
        [0.6412],
        [0.6275],
        [1.5607],
        [1.2491],
        [0.7272],
        [0... [1.4196],
        [2.5846],
        [0.5729],
        [1.4330],
        [0.6490],
        [0.3398],
        [0.1452]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
___________ test_graph_conv_e_weight[1-True-False-right-g0-idtype1] ___________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32)})
norm = 'right', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        e_w = g.edata['scalar_w']
        if weight:
            h_out = conv(g, h, edge_weight=e_w)
        else:
>           h_out = conv(g, h, weight=ext_w, edge_weight=e_w)

tests\pytorch\test_nn.py:119: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14559E88>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[ 0.6073],
        [ 0.3081],
        [-0.1906],
        [ 0.0041],
        [ 0.5309],
        [-1.8029],
        [ 0.0913],
        [ 0.0575],
        [-0.6231],
        [-1.5754]])
rhs_data = tensor([[0.2636],
        [0.6412],
        [0.6275],
        [1.5607],
        [1.2491],
        [0.7272],
        [0... [1.4196],
        [2.5846],
        [0.5729],
        [1.4330],
        [0.6490],
        [0.3398],
        [0.1452]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
___________ test_graph_conv_e_weight[1-False-True-none-g0-idtype0] ____________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32)})
norm = 'none', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        e_w = g.edata['scalar_w']
        if weight:
>           h_out = conv(g, h, edge_weight=e_w)

tests\pytorch\test_nn.py:117: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14A66D88>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[ 0.1043],
        [-0.0625],
        [-0.5570],
        [-0.3847],
        [ 0.2742],
        [ 0.1069],
        [-0.6647],
        [-0.5342],
        [-0.2113],
        [-0.6410]], grad_fn=<ViewBackward>)
rhs_data = tensor([[0.2636],
        [0.6412],
        [0.6275],
        [1.5607],
        [1.2491],
        [0.7272],
        [0... [1.4196],
        [2.5846],
        [0.5729],
        [1.4330],
        [0.6490],
        [0.3398],
        [0.1452]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
___________ test_graph_conv_e_weight[1-False-True-none-g0-idtype1] ____________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32)})
norm = 'none', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        e_w = g.edata['scalar_w']
        if weight:
>           h_out = conv(g, h, edge_weight=e_w)

tests\pytorch\test_nn.py:117: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF010AC348>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[-7.6125e-01],
        [ 5.4765e-02],
        [-1.2756e+00],
        [ 1.2723e+00],
        [-1.4818e+00],
   ...],
        [-8.4316e-01],
        [ 1.3385e-03],
        [ 1.5277e+00],
        [-1.8785e-01]], grad_fn=<ViewBackward>)
rhs_data = tensor([[0.2636],
        [0.6412],
        [0.6275],
        [1.5607],
        [1.2491],
        [0.7272],
        [0... [1.4196],
        [2.5846],
        [0.5729],
        [1.4330],
        [0.6490],
        [0.3398],
        [0.1452]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
___________ test_graph_conv_e_weight[1-False-True-both-g0-idtype0] ____________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32)})
norm = 'both', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        e_w = g.edata['scalar_w']
        if weight:
>           h_out = conv(g, h, edge_weight=e_w)

tests\pytorch\test_nn.py:117: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14537988>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[-0.5332],
        [-0.5773],
        [ 0.5188],
        [ 1.2162],
        [ 0.3919],
        [-1.8548],
        [-1.2726],
        [-1.0423],
        [-0.6045],
        [-2.6664]], grad_fn=<ViewBackward>)
rhs_data = tensor([[0.2636],
        [0.6412],
        [0.6275],
        [1.5607],
        [1.2491],
        [0.7272],
        [0... [1.4196],
        [2.5846],
        [0.5729],
        [1.4330],
        [0.6490],
        [0.3398],
        [0.1452]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
___________ test_graph_conv_e_weight[1-False-True-both-g0-idtype1] ____________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32)})
norm = 'both', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        e_w = g.edata['scalar_w']
        if weight:
>           h_out = conv(g, h, edge_weight=e_w)

tests\pytorch\test_nn.py:117: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01158308>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[ 0.2765],
        [ 0.9435],
        [ 0.0752],
        [-0.5624],
        [ 1.9305],
        [-0.1518],
        [-1.7705],
        [-0.5975],
        [-2.2535],
        [-1.5242]], grad_fn=<ViewBackward>)
rhs_data = tensor([[0.2636],
        [0.6412],
        [0.6275],
        [1.5607],
        [1.2491],
        [0.7272],
        [0... [1.4196],
        [2.5846],
        [0.5729],
        [1.4330],
        [0.6490],
        [0.3398],
        [0.1452]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
___________ test_graph_conv_e_weight[1-False-True-right-g0-idtype0] ___________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32)})
norm = 'right', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        e_w = g.edata['scalar_w']
        if weight:
>           h_out = conv(g, h, edge_weight=e_w)

tests\pytorch\test_nn.py:117: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1469EB48>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[-0.4121],
        [-1.5410],
        [ 0.4339],
        [ 0.9566],
        [-1.8335],
        [ 0.4357],
        [ 0.2167],
        [ 1.7359],
        [ 1.5783],
        [ 0.7778]], grad_fn=<ViewBackward>)
rhs_data = tensor([[0.2636],
        [0.6412],
        [0.6275],
        [1.5607],
        [1.2491],
        [0.7272],
        [0... [1.4196],
        [2.5846],
        [0.5729],
        [1.4330],
        [0.6490],
        [0.3398],
        [0.1452]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
___________ test_graph_conv_e_weight[1-False-True-right-g0-idtype1] ___________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32)})
norm = 'right', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        e_w = g.edata['scalar_w']
        if weight:
>           h_out = conv(g, h, edge_weight=e_w)

tests\pytorch\test_nn.py:117: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1463CF88>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[-0.8243],
        [ 2.3516],
        [-1.1463],
        [-0.7258],
        [-0.9015],
        [ 0.7196],
        [ 1.7132],
        [ 0.0803],
        [ 0.3772],
        [-0.7832]], grad_fn=<ViewBackward>)
rhs_data = tensor([[0.2636],
        [0.6412],
        [0.6275],
        [1.5607],
        [1.2491],
        [0.7272],
        [0... [1.4196],
        [2.5846],
        [0.5729],
        [1.4330],
        [0.6490],
        [0.3398],
        [0.1452]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
___________ test_graph_conv_e_weight[1-False-False-none-g0-idtype0] ___________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32)})
norm = 'none', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        e_w = g.edata['scalar_w']
        if weight:
            h_out = conv(g, h, edge_weight=e_w)
        else:
>           h_out = conv(g, h, weight=ext_w, edge_weight=e_w)

tests\pytorch\test_nn.py:119: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF010AC188>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[ 0.7385],
        [ 0.8706],
        [ 0.8200],
        [ 1.8062],
        [-3.5475],
        [-3.3429],
        [-2.2859],
        [-1.8538],
        [-0.0371],
        [ 0.0308]])
rhs_data = tensor([[0.2636],
        [0.6412],
        [0.6275],
        [1.5607],
        [1.2491],
        [0.7272],
        [0... [1.4196],
        [2.5846],
        [0.5729],
        [1.4330],
        [0.6490],
        [0.3398],
        [0.1452]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
___________ test_graph_conv_e_weight[1-False-False-none-g0-idtype1] ___________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32)})
norm = 'none', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        e_w = g.edata['scalar_w']
        if weight:
            h_out = conv(g, h, edge_weight=e_w)
        else:
>           h_out = conv(g, h, weight=ext_w, edge_weight=e_w)

tests\pytorch\test_nn.py:119: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF030BC888>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[-5.4672],
        [-4.7355],
        [-0.8634],
        [-0.9149],
        [ 0.8999],
        [-1.5965],
        [-0.4160],
        [ 3.4646],
        [-2.0384],
        [-3.3900]])
rhs_data = tensor([[0.2636],
        [0.6412],
        [0.6275],
        [1.5607],
        [1.2491],
        [0.7272],
        [0... [1.4196],
        [2.5846],
        [0.5729],
        [1.4330],
        [0.6490],
        [0.3398],
        [0.1452]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
___________ test_graph_conv_e_weight[1-False-False-both-g0-idtype0] ___________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32)})
norm = 'both', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        e_w = g.edata['scalar_w']
        if weight:
            h_out = conv(g, h, edge_weight=e_w)
        else:
>           h_out = conv(g, h, weight=ext_w, edge_weight=e_w)

tests\pytorch\test_nn.py:119: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14541A08>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[ 0.8375],
        [ 3.4503],
        [-1.0103],
        [ 2.1868],
        [-0.8784],
        [-1.6144],
        [ 1.7861],
        [ 3.9775],
        [-0.4818],
        [ 3.5261]])
rhs_data = tensor([[0.2636],
        [0.6412],
        [0.6275],
        [1.5607],
        [1.2491],
        [0.7272],
        [0... [1.4196],
        [2.5846],
        [0.5729],
        [1.4330],
        [0.6490],
        [0.3398],
        [0.1452]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
___________ test_graph_conv_e_weight[1-False-False-both-g0-idtype1] ___________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32)})
norm = 'both', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        e_w = g.edata['scalar_w']
        if weight:
            h_out = conv(g, h, edge_weight=e_w)
        else:
>           h_out = conv(g, h, weight=ext_w, edge_weight=e_w)

tests\pytorch\test_nn.py:119: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01A19308>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[ 0.4770],
        [ 1.2469],
        [-0.9251],
        [-1.6392],
        [ 0.4527],
        [ 1.5740],
        [ 0.8871],
        [ 2.5137],
        [ 0.6273],
        [ 1.5126]])
rhs_data = tensor([[0.2636],
        [0.6412],
        [0.6275],
        [1.5607],
        [1.2491],
        [0.7272],
        [0... [1.4196],
        [2.5846],
        [0.5729],
        [1.4330],
        [0.6490],
        [0.3398],
        [0.1452]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________ test_graph_conv_e_weight[1-False-False-right-g0-idtype0] ___________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32)})
norm = 'right', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        e_w = g.edata['scalar_w']
        if weight:
            h_out = conv(g, h, edge_weight=e_w)
        else:
>           h_out = conv(g, h, weight=ext_w, edge_weight=e_w)

tests\pytorch\test_nn.py:119: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1460FBC8>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[-1.3293],
        [-0.1422],
        [ 1.0020],
        [ 0.7395],
        [ 0.0505],
        [ 0.3893],
        [-0.3602],
        [-0.3616],
        [-0.8091],
        [-0.1110]])
rhs_data = tensor([[0.2636],
        [0.6412],
        [0.6275],
        [1.5607],
        [1.2491],
        [0.7272],
        [0... [1.4196],
        [2.5846],
        [0.5729],
        [1.4330],
        [0.6490],
        [0.3398],
        [0.1452]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________ test_graph_conv_e_weight[1-False-False-right-g0-idtype1] ___________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32)})
norm = 'right', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        e_w = g.edata['scalar_w']
        if weight:
            h_out = conv(g, h, edge_weight=e_w)
        else:
>           h_out = conv(g, h, weight=ext_w, edge_weight=e_w)

tests\pytorch\test_nn.py:119: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14572288>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[-1.8943],
        [ 1.8461],
        [-2.6342],
        [ 2.2855],
        [-0.2479],
        [ 0.3282],
        [-1.0664],
        [-1.2323],
        [ 2.0722],
        [ 1.3859]])
rhs_data = tensor([[0.2636],
        [0.6412],
        [0.6275],
        [1.5607],
        [1.2491],
        [0.7272],
        [0... [1.4196],
        [2.5846],
        [0.5729],
        [1.4330],
        [0.6490],
        [0.3398],
        [0.1452]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
____________ test_graph_conv_e_weight[2-True-True-none-g0-idtype0] ____________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32)})
norm = 'none', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        e_w = g.edata['scalar_w']
        if weight:
>           h_out = conv(g, h, edge_weight=e_w)

tests\pytorch\test_nn.py:117: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE145C2208>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[-0.3852, -0.3228],
        [ 1.2578, -1.5226],
        [ 2.8948, -1.4933],
        [-2.4005,  0.8043],
      ... -0.4166],
        [ 0.1285,  0.0803],
        [ 0.7049, -0.1192],
        [-0.3140,  0.2866]], grad_fn=<ViewBackward>)
rhs_data = tensor([[0.2636],
        [0.6412],
        [0.6275],
        [1.5607],
        [1.2491],
        [0.7272],
        [0... [1.4196],
        [2.5846],
        [0.5729],
        [1.4330],
        [0.6490],
        [0.3398],
        [0.1452]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
____________ test_graph_conv_e_weight[2-True-True-none-g0-idtype1] ____________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32)})
norm = 'none', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        e_w = g.edata['scalar_w']
        if weight:
>           h_out = conv(g, h, edge_weight=e_w)

tests\pytorch\test_nn.py:117: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1462DBC8>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[ 0.1567, -0.7134],
        [ 1.4694,  1.6862],
        [-0.8307, -1.7097],
        [-0.8201, -0.0612],
      ... -0.1440],
        [ 0.2534, -0.3844],
        [-0.2288,  0.7127],
        [-1.1561,  0.8944]], grad_fn=<ViewBackward>)
rhs_data = tensor([[0.2636],
        [0.6412],
        [0.6275],
        [1.5607],
        [1.2491],
        [0.7272],
        [0... [1.4196],
        [2.5846],
        [0.5729],
        [1.4330],
        [0.6490],
        [0.3398],
        [0.1452]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
____________ test_graph_conv_e_weight[2-True-True-both-g0-idtype0] ____________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32)})
norm = 'both', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        e_w = g.edata['scalar_w']
        if weight:
>           h_out = conv(g, h, edge_weight=e_w)

tests\pytorch\test_nn.py:117: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14661D08>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[-1.2172,  0.9689],
        [ 0.9453, -0.9239],
        [-0.9046,  0.4045],
        [-1.0842,  0.5299],
      ... -0.4291],
        [-0.4016,  0.0277],
        [ 1.0378, -1.4233],
        [ 0.4934, -0.3972]], grad_fn=<ViewBackward>)
rhs_data = tensor([[0.2636],
        [0.6412],
        [0.6275],
        [1.5607],
        [1.2491],
        [0.7272],
        [0... [1.4196],
        [2.5846],
        [0.5729],
        [1.4330],
        [0.6490],
        [0.3398],
        [0.1452]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
____________ test_graph_conv_e_weight[2-True-True-both-g0-idtype1] ____________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32)})
norm = 'both', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        e_w = g.edata['scalar_w']
        if weight:
>           h_out = conv(g, h, edge_weight=e_w)

tests\pytorch\test_nn.py:117: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03F2FE48>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[-0.8267, -0.0437],
        [ 0.8411,  0.7800],
        [ 0.6452,  0.0302],
        [-0.1216, -0.1410],
      ...  0.7056],
        [-0.0614,  0.1294],
        [-1.3993, -0.9046],
        [ 0.0971,  0.1511]], grad_fn=<ViewBackward>)
rhs_data = tensor([[0.2636],
        [0.6412],
        [0.6275],
        [1.5607],
        [1.2491],
        [0.7272],
        [0... [1.4196],
        [2.5846],
        [0.5729],
        [1.4330],
        [0.6490],
        [0.3398],
        [0.1452]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
___________ test_graph_conv_e_weight[2-True-True-right-g0-idtype0] ____________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32)})
norm = 'right', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        e_w = g.edata['scalar_w']
        if weight:
>           h_out = conv(g, h, edge_weight=e_w)

tests\pytorch\test_nn.py:117: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE149BA708>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[-0.6204,  0.2217],
        [-1.4120, -0.7790],
        [ 3.0222,  1.3503],
        [-0.4580, -1.0752],
      ... -1.6044],
        [ 0.2720, -0.7981],
        [-0.4857, -0.5967],
        [ 0.0428, -1.0461]], grad_fn=<ViewBackward>)
rhs_data = tensor([[0.2636],
        [0.6412],
        [0.6275],
        [1.5607],
        [1.2491],
        [0.7272],
        [0... [1.4196],
        [2.5846],
        [0.5729],
        [1.4330],
        [0.6490],
        [0.3398],
        [0.1452]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
___________ test_graph_conv_e_weight[2-True-True-right-g0-idtype1] ____________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32)})
norm = 'right', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        e_w = g.edata['scalar_w']
        if weight:
>           h_out = conv(g, h, edge_weight=e_w)

tests\pytorch\test_nn.py:117: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1453FC08>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[-1.0895, -1.5226],
        [-0.0627,  0.5230],
        [ 1.6304,  2.2730],
        [ 0.7728, -0.6397],
      ... -0.7152],
        [-0.5429, -1.2627],
        [-0.6952, -1.0449],
        [ 0.8115, -0.7641]], grad_fn=<ViewBackward>)
rhs_data = tensor([[0.2636],
        [0.6412],
        [0.6275],
        [1.5607],
        [1.2491],
        [0.7272],
        [0... [1.4196],
        [2.5846],
        [0.5729],
        [1.4330],
        [0.6490],
        [0.3398],
        [0.1452]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
___________ test_graph_conv_e_weight[2-True-False-none-g0-idtype0] ____________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32)})
norm = 'none', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        e_w = g.edata['scalar_w']
        if weight:
            h_out = conv(g, h, edge_weight=e_w)
        else:
>           h_out = conv(g, h, weight=ext_w, edge_weight=e_w)

tests\pytorch\test_nn.py:119: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF011CB408>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[-2.0786, -0.1997],
        [ 0.9699,  1.0757],
        [-2.8533, -1.9335],
        [-3.0849, -3.3806],
      ...4443],
        [-3.2135, -4.3409],
        [-1.8638, -2.4936],
        [ 2.2427, -0.7416],
        [-0.9404, -1.0373]])
rhs_data = tensor([[0.2636],
        [0.6412],
        [0.6275],
        [1.5607],
        [1.2491],
        [0.7272],
        [0... [1.4196],
        [2.5846],
        [0.5729],
        [1.4330],
        [0.6490],
        [0.3398],
        [0.1452]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
___________ test_graph_conv_e_weight[2-True-False-none-g0-idtype1] ____________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32)})
norm = 'none', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        e_w = g.edata['scalar_w']
        if weight:
            h_out = conv(g, h, edge_weight=e_w)
        else:
>           h_out = conv(g, h, weight=ext_w, edge_weight=e_w)

tests\pytorch\test_nn.py:119: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE145ABF88>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[-0.1988,  0.1522],
        [ 1.2731, -1.7170],
        [-2.5432, -2.3599],
        [ 1.8377,  1.1629],
      ...5060],
        [ 0.9371, -0.9900],
        [-0.1461,  0.4561],
        [-0.9634,  1.4348],
        [ 2.1179, -1.4838]])
rhs_data = tensor([[0.2636],
        [0.6412],
        [0.6275],
        [1.5607],
        [1.2491],
        [0.7272],
        [0... [1.4196],
        [2.5846],
        [0.5729],
        [1.4330],
        [0.6490],
        [0.3398],
        [0.1452]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
___________ test_graph_conv_e_weight[2-True-False-both-g0-idtype0] ____________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32)})
norm = 'both', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        e_w = g.edata['scalar_w']
        if weight:
            h_out = conv(g, h, edge_weight=e_w)
        else:
>           h_out = conv(g, h, weight=ext_w, edge_weight=e_w)

tests\pytorch\test_nn.py:119: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01188548>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[-0.9450,  0.6088],
        [-2.5757, -1.8900],
        [ 2.9722, -0.0756],
        [ 2.9765,  0.0258],
      ...2333],
        [ 4.2602,  1.0731],
        [ 0.7015, -1.5368],
        [ 2.9042,  1.7220],
        [-1.5094, -0.1385]])
rhs_data = tensor([[0.2636],
        [0.6412],
        [0.6275],
        [1.5607],
        [1.2491],
        [0.7272],
        [0... [1.4196],
        [2.5846],
        [0.5729],
        [1.4330],
        [0.6490],
        [0.3398],
        [0.1452]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
___________ test_graph_conv_e_weight[2-True-False-both-g0-idtype1] ____________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32)})
norm = 'both', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        e_w = g.edata['scalar_w']
        if weight:
            h_out = conv(g, h, edge_weight=e_w)
        else:
>           h_out = conv(g, h, weight=ext_w, edge_weight=e_w)

tests\pytorch\test_nn.py:119: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1451E288>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[ 0.1950,  0.5774],
        [-1.1712, -1.1387],
        [ 0.1357,  0.5841],
        [-0.4665, -0.2378],
      ...2669],
        [ 1.8534,  1.3137],
        [-1.5385, -1.2876],
        [-1.8800, -1.8955],
        [ 3.2703, -1.0233]])
rhs_data = tensor([[0.2636],
        [0.6412],
        [0.6275],
        [1.5607],
        [1.2491],
        [0.7272],
        [0... [1.4196],
        [2.5846],
        [0.5729],
        [1.4330],
        [0.6490],
        [0.3398],
        [0.1452]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
___________ test_graph_conv_e_weight[2-True-False-right-g0-idtype0] ___________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32)})
norm = 'right', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        e_w = g.edata['scalar_w']
        if weight:
            h_out = conv(g, h, edge_weight=e_w)
        else:
>           h_out = conv(g, h, weight=ext_w, edge_weight=e_w)

tests\pytorch\test_nn.py:119: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF011B4F88>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[ 0.6285,  0.0319],
        [-1.8856, -1.1028],
        [ 1.0134,  0.5464],
        [-4.8132, -1.0679],
      ...0325],
        [-0.9835, -0.2027],
        [ 0.6625,  0.1174],
        [-0.6421,  0.5773],
        [-0.4712, -1.3254]])
rhs_data = tensor([[0.2636],
        [0.6412],
        [0.6275],
        [1.5607],
        [1.2491],
        [0.7272],
        [0... [1.4196],
        [2.5846],
        [0.5729],
        [1.4330],
        [0.6490],
        [0.3398],
        [0.1452]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
___________ test_graph_conv_e_weight[2-True-False-right-g0-idtype1] ___________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32)})
norm = 'right', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        e_w = g.edata['scalar_w']
        if weight:
            h_out = conv(g, h, edge_weight=e_w)
        else:
>           h_out = conv(g, h, weight=ext_w, edge_weight=e_w)

tests\pytorch\test_nn.py:119: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01061E48>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[ 1.1648e-03,  1.2402e+00],
        [ 4.3735e+00,  3.7361e-01],
        [-1.4894e+00,  2.0280e+00],
        [-...4810e+00],
        [-2.4275e+00,  3.1603e+00],
        [-2.5286e+00,  1.1351e+00],
        [ 3.2538e+00,  1.9230e+00]])
rhs_data = tensor([[0.2636],
        [0.6412],
        [0.6275],
        [1.5607],
        [1.2491],
        [0.7272],
        [0... [1.4196],
        [2.5846],
        [0.5729],
        [1.4330],
        [0.6490],
        [0.3398],
        [0.1452]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
___________ test_graph_conv_e_weight[2-False-True-none-g0-idtype0] ____________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32)})
norm = 'none', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        e_w = g.edata['scalar_w']
        if weight:
>           h_out = conv(g, h, edge_weight=e_w)

tests\pytorch\test_nn.py:117: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14537A88>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[-0.3432, -2.0368],
        [ 0.8998, -0.4630],
        [-1.3543, -0.8712],
        [-1.0895, -0.2765],
      ... -0.6825],
        [-0.5466, -0.9366],
        [-1.1563, -0.0949],
        [-0.1367,  1.1262]], grad_fn=<ViewBackward>)
rhs_data = tensor([[0.2636],
        [0.6412],
        [0.6275],
        [1.5607],
        [1.2491],
        [0.7272],
        [0... [1.4196],
        [2.5846],
        [0.5729],
        [1.4330],
        [0.6490],
        [0.3398],
        [0.1452]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
___________ test_graph_conv_e_weight[2-False-True-none-g0-idtype1] ____________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32)})
norm = 'none', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        e_w = g.edata['scalar_w']
        if weight:
>           h_out = conv(g, h, edge_weight=e_w)

tests\pytorch\test_nn.py:117: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03E91708>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[-1.3143,  0.1327],
        [-1.5208, -0.2701],
        [-0.5013, -0.4939],
        [-1.0701, -1.7511],
      ... -1.7177],
        [-1.2160,  0.4835],
        [-0.3627,  2.2071],
        [-1.8008, -2.3090]], grad_fn=<ViewBackward>)
rhs_data = tensor([[0.2636],
        [0.6412],
        [0.6275],
        [1.5607],
        [1.2491],
        [0.7272],
        [0... [1.4196],
        [2.5846],
        [0.5729],
        [1.4330],
        [0.6490],
        [0.3398],
        [0.1452]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
___________ test_graph_conv_e_weight[2-False-True-both-g0-idtype0] ____________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32)})
norm = 'both', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        e_w = g.edata['scalar_w']
        if weight:
>           h_out = conv(g, h, edge_weight=e_w)

tests\pytorch\test_nn.py:117: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01089C88>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[-1.0308,  0.2073],
        [ 0.5260, -0.9092],
        [ 0.2764, -0.1170],
        [-1.9801,  0.4831],
      ... -0.1997],
        [ 1.7295, -0.2561],
        [ 1.4155,  0.4343],
        [-0.0364,  0.4197]], grad_fn=<ViewBackward>)
rhs_data = tensor([[0.2636],
        [0.6412],
        [0.6275],
        [1.5607],
        [1.2491],
        [0.7272],
        [0... [1.4196],
        [2.5846],
        [0.5729],
        [1.4330],
        [0.6490],
        [0.3398],
        [0.1452]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
___________ test_graph_conv_e_weight[2-False-True-both-g0-idtype1] ____________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32)})
norm = 'both', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        e_w = g.edata['scalar_w']
        if weight:
>           h_out = conv(g, h, edge_weight=e_w)

tests\pytorch\test_nn.py:117: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE146AB348>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[-0.1727,  0.0463],
        [ 0.3649, -1.1719],
        [-0.1976,  1.0955],
        [-0.6406, -0.1066],
      ... -2.1732],
        [ 1.4573,  0.2099],
        [ 1.0147, -2.0249],
        [ 0.3322, -1.0338]], grad_fn=<ViewBackward>)
rhs_data = tensor([[0.2636],
        [0.6412],
        [0.6275],
        [1.5607],
        [1.2491],
        [0.7272],
        [0... [1.4196],
        [2.5846],
        [0.5729],
        [1.4330],
        [0.6490],
        [0.3398],
        [0.1452]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
___________ test_graph_conv_e_weight[2-False-True-right-g0-idtype0] ___________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32)})
norm = 'right', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        e_w = g.edata['scalar_w']
        if weight:
>           h_out = conv(g, h, edge_weight=e_w)

tests\pytorch\test_nn.py:117: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0107F688>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[-0.0594,  0.2220],
        [ 0.9740, -1.4039],
        [ 0.6694,  1.1824],
        [-0.5620,  1.9080],
      ... -0.1092],
        [-0.0281,  2.0562],
        [ 1.3107,  0.5666],
        [ 0.6256,  0.4200]], grad_fn=<ViewBackward>)
rhs_data = tensor([[0.2636],
        [0.6412],
        [0.6275],
        [1.5607],
        [1.2491],
        [0.7272],
        [0... [1.4196],
        [2.5846],
        [0.5729],
        [1.4330],
        [0.6490],
        [0.3398],
        [0.1452]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
___________ test_graph_conv_e_weight[2-False-True-right-g0-idtype1] ___________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32)})
norm = 'right', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        e_w = g.edata['scalar_w']
        if weight:
>           h_out = conv(g, h, edge_weight=e_w)

tests\pytorch\test_nn.py:117: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE145A1D88>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[-1.6243,  1.2418],
        [-0.6975,  0.4106],
        [-1.2084,  0.4609],
        [-0.5399,  1.4375],
      ...  0.8950],
        [-0.1846, -0.6907],
        [ 1.3306,  0.8185],
        [ 0.7482,  0.6952]], grad_fn=<ViewBackward>)
rhs_data = tensor([[0.2636],
        [0.6412],
        [0.6275],
        [1.5607],
        [1.2491],
        [0.7272],
        [0... [1.4196],
        [2.5846],
        [0.5729],
        [1.4330],
        [0.6490],
        [0.3398],
        [0.1452]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
___________ test_graph_conv_e_weight[2-False-False-none-g0-idtype0] ___________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32)})
norm = 'none', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        e_w = g.edata['scalar_w']
        if weight:
            h_out = conv(g, h, edge_weight=e_w)
        else:
>           h_out = conv(g, h, weight=ext_w, edge_weight=e_w)

tests\pytorch\test_nn.py:119: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE145C0EC8>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[-6.3929e-01, -2.1269e+00],
        [-2.6556e+00, -6.0399e-01],
        [ 2.4145e+00,  1.2796e+00],
        [-...7185e+00],
        [ 2.4583e+00, -1.3883e-03],
        [-2.8224e+00, -2.2268e+00],
        [ 1.6536e+00, -2.0760e+00]])
rhs_data = tensor([[0.2636],
        [0.6412],
        [0.6275],
        [1.5607],
        [1.2491],
        [0.7272],
        [0... [1.4196],
        [2.5846],
        [0.5729],
        [1.4330],
        [0.6490],
        [0.3398],
        [0.1452]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':

            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
___________ test_graph_conv_e_weight[2-False-False-none-g0-idtype1] ___________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32)})
norm = 'none', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        e_w = g.edata['scalar_w']
        if weight:
            h_out = conv(g, h, edge_weight=e_w)
        else:
>           h_out = conv(g, h, weight=ext_w, edge_weight=e_w)

tests\pytorch\test_nn.py:119: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14541888>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[-0.7281, -0.0294],
        [ 1.9730, -1.4881],
        [-1.2524,  0.2038],
        [ 2.8942,  2.1111],
      ...8065],
        [-1.1226,  2.8057],
        [-1.8537,  1.9522],
        [-0.7855,  0.1242],
        [-1.6980,  2.4839]])
rhs_data = tensor([[0.2636],
        [0.6412],
        [0.6275],
        [1.5607],
        [1.2491],
        [0.7272],
        [0... [1.4196],
        [2.5846],
        [0.5729],
        [1.4330],
        [0.6490],
        [0.3398],
        [0.1452]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
___________ test_graph_conv_e_weight[2-False-False-both-g0-idtype0] ___________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32)})
norm = 'both', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        e_w = g.edata['scalar_w']
        if weight:
            h_out = conv(g, h, edge_weight=e_w)
        else:
>           h_out = conv(g, h, weight=ext_w, edge_weight=e_w)

tests\pytorch\test_nn.py:119: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE149433C8>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[ 0.6095,  1.1709],
        [-2.7083,  4.1018],
        [ 1.9607, -1.4561],
        [-0.4401, -1.2685],
      ...6913],
        [-0.0910,  2.1820],
        [-1.7649, -0.9610],
        [ 0.3822, -1.2950],
        [ 1.2023,  2.8610]])
rhs_data = tensor([[0.2636],
        [0.6412],
        [0.6275],
        [1.5607],
        [1.2491],
        [0.7272],
        [0... [1.4196],
        [2.5846],
        [0.5729],
        [1.4330],
        [0.6490],
        [0.3398],
        [0.1452]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
___________ test_graph_conv_e_weight[2-False-False-both-g0-idtype1] ___________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32)})
norm = 'both', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        e_w = g.edata['scalar_w']
        if weight:
            h_out = conv(g, h, edge_weight=e_w)
        else:
>           h_out = conv(g, h, weight=ext_w, edge_weight=e_w)

tests\pytorch\test_nn.py:119: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0107EC48>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[ 0.2520,  0.5590],
        [ 1.8649,  0.1462],
        [ 0.6869,  0.2136],
        [ 0.2423,  1.4714],
      ...1079],
        [-1.8201, -0.4105],
        [-2.7583, -2.3234],
        [-4.7321,  0.9350],
        [ 1.6715, -1.8228]])
rhs_data = tensor([[0.2636],
        [0.6412],
        [0.6275],
        [1.5607],
        [1.2491],
        [0.7272],
        [0... [1.4196],
        [2.5846],
        [0.5729],
        [1.4330],
        [0.6490],
        [0.3398],
        [0.1452]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________ test_graph_conv_e_weight[2-False-False-right-g0-idtype0] ___________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32)})
norm = 'right', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        e_w = g.edata['scalar_w']
        if weight:
            h_out = conv(g, h, edge_weight=e_w)
        else:
>           h_out = conv(g, h, weight=ext_w, edge_weight=e_w)

tests\pytorch\test_nn.py:119: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01682D08>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[ 3.7128,  2.3379],
        [ 1.9219,  1.2300],
        [-2.6823, -3.2687],
        [-1.0375,  0.9667],
      ...7087],
        [-1.5644, -2.3919],
        [ 1.4282,  0.3831],
        [ 2.6540,  0.4025],
        [-3.1763, -1.5815]])
rhs_data = tensor([[0.2636],
        [0.6412],
        [0.6275],
        [1.5607],
        [1.2491],
        [0.7272],
        [0... [1.4196],
        [2.5846],
        [0.5729],
        [1.4330],
        [0.6490],
        [0.3398],
        [0.1452]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________ test_graph_conv_e_weight[2-False-False-right-g0-idtype1] ___________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32)})
norm = 'right', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        e_w = g.edata['scalar_w']
        if weight:
            h_out = conv(g, h, edge_weight=e_w)
        else:
>           h_out = conv(g, h, weight=ext_w, edge_weight=e_w)

tests\pytorch\test_nn.py:119: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14A71E48>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[-2.6783, -0.7400],
        [ 1.5034, -1.0513],
        [-0.8267, -2.6617],
        [-0.0163,  3.2188],
      ...2458],
        [-2.2590,  2.5459],
        [ 1.9112, -2.2476],
        [-2.1824,  3.0716],
        [ 4.2991, -1.5237]])
rhs_data = tensor([[0.2636],
        [0.6412],
        [0.6275],
        [1.5607],
        [1.2491],
        [0.7272],
        [0... [1.4196],
        [2.5846],
        [0.5729],
        [1.4330],
        [0.6490],
        [0.3398],
        [0.1452]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________ test_graph_conv_e_weight_norm[1-True-True-none-g0-idtype0] __________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32)})
norm = 'none', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight_norm(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        edgenorm = nn.EdgeWeightNorm(norm=norm)
        norm_weight = edgenorm(g, g.edata['scalar_w'])
        if weight:
>           h_out = conv(g, h, edge_weight=norm_weight)

tests\pytorch\test_nn.py:142: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14943FC8>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[-0.3693],
        [ 0.7781],
        [-1.8341],
        [-0.2539],
        [ 0.1112],
        [ 1.1728],
        [-0.4980],
        [-0.5293],
        [ 1.2855],
        [-0.7784]], grad_fn=<ViewBackward>)
rhs_data = tensor([[0.6491],
        [1.8020],
        [0.0416],
        [3.2244],
        [0.1902],
        [0.6207],
        [1... [0.6860],
        [0.3175],
        [2.0501],
        [0.8401],
        [2.2833],
        [1.1770],
        [0.2148]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________ test_graph_conv_e_weight_norm[1-True-True-none-g0-idtype1] __________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32)})
norm = 'none', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight_norm(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        edgenorm = nn.EdgeWeightNorm(norm=norm)
        norm_weight = edgenorm(g, g.edata['scalar_w'])
        if weight:
>           h_out = conv(g, h, edge_weight=norm_weight)

tests\pytorch\test_nn.py:142: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF011C5508>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[-1.0819],
        [-0.3482],
        [-0.2069],
        [ 1.8858],
        [ 0.9585],
        [ 1.1091],
        [-0.7812],
        [ 0.8048],
        [-0.5682],
        [-0.4612]], grad_fn=<ViewBackward>)
rhs_data = tensor([[0.6491],
        [1.8020],
        [0.0416],
        [3.2244],
        [0.1902],
        [0.6207],
        [1... [0.6860],
        [0.3175],
        [2.0501],
        [0.8401],
        [2.2833],
        [1.1770],
        [0.2148]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________ test_graph_conv_e_weight_norm[1-True-True-both-g0-idtype0] __________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32), '_src_out_w': Sch...rch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32), '_edge_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'both', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight_norm(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        edgenorm = nn.EdgeWeightNorm(norm=norm)
>       norm_weight = edgenorm(g, g.edata['scalar_w'])

tests\pytorch\test_nn.py:140: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:117: in forward
    reversed_g.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'out_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0107A108>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([0.6491, 1.8020, 0.0416, 3.2244, 0.1902, 0.6207, 1.4090, 0.3025, 1.7357,
        0.6755, 0.6860, 0.3175, 2.0501, 0.8401, 2.2833, 1.1770, 0.2148])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________ test_graph_conv_e_weight_norm[1-True-True-both-g0-idtype1] __________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32), '_src_out_w': Sch...rch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32), '_edge_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'both', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight_norm(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        edgenorm = nn.EdgeWeightNorm(norm=norm)
>       norm_weight = edgenorm(g, g.edata['scalar_w'])

tests\pytorch\test_nn.py:140: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:117: in forward
    reversed_g.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'out_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0309A708>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([0.6491, 1.8020, 0.0416, 3.2244, 0.1902, 0.6207, 1.4090, 0.3025, 1.7357,
        0.6755, 0.6860, 0.3175, 2.0501, 0.8401, 2.2833, 1.1770, 0.2148])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________ test_graph_conv_e_weight_norm[1-True-True-right-g0-idtype0] _________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32), '_src_out_w': Sch...rch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32), '_edge_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'right', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight_norm(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        edgenorm = nn.EdgeWeightNorm(norm=norm)
>       norm_weight = edgenorm(g, g.edata['scalar_w'])

tests\pytorch\test_nn.py:140: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:123: in forward
    graph.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'in_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE149267C8>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([0.6491, 1.8020, 0.0416, 3.2244, 0.1902, 0.6207, 1.4090, 0.3025, 1.7357,
        0.6755, 0.6860, 0.3175, 2.0501, 0.8401, 2.2833, 1.1770, 0.2148])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________ test_graph_conv_e_weight_norm[1-True-True-right-g0-idtype1] _________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32), '_src_out_w': Sch...rch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32), '_edge_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'right', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight_norm(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        edgenorm = nn.EdgeWeightNorm(norm=norm)
>       norm_weight = edgenorm(g, g.edata['scalar_w'])

tests\pytorch\test_nn.py:140: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:123: in forward
    graph.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'in_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF010760C8>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([0.6491, 1.8020, 0.0416, 3.2244, 0.1902, 0.6207, 1.4090, 0.3025, 1.7357,
        0.6755, 0.6860, 0.3175, 2.0501, 0.8401, 2.2833, 1.1770, 0.2148])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________ test_graph_conv_e_weight_norm[1-True-False-none-g0-idtype0] _________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32), '_src_out_w': Sch...rch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32), '_edge_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'none', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight_norm(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        edgenorm = nn.EdgeWeightNorm(norm=norm)
        norm_weight = edgenorm(g, g.edata['scalar_w'])
        if weight:
            h_out = conv(g, h, edge_weight=norm_weight)
        else:
>           h_out = conv(g, h, weight=ext_w, edge_weight=norm_weight)

tests\pytorch\test_nn.py:144: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01075B48>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[ 1.2051],
        [-0.4916],
        [-1.0503],
        [ 3.5286],
        [-0.4467],
        [ 0.5251],
        [-2.0260],
        [-1.8995],
        [-5.0165],
        [ 0.6717]])
rhs_data = tensor([[0.6491],
        [1.8020],
        [0.0416],
        [3.2244],
        [0.1902],
        [0.6207],
        [1... [0.6860],
        [0.3175],
        [2.0501],
        [0.8401],
        [2.2833],
        [1.1770],
        [0.2148]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________ test_graph_conv_e_weight_norm[1-True-False-none-g0-idtype1] _________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32), '_src_out_w': Sch...rch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32), '_edge_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'none', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight_norm(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        edgenorm = nn.EdgeWeightNorm(norm=norm)
        norm_weight = edgenorm(g, g.edata['scalar_w'])
        if weight:
            h_out = conv(g, h, edge_weight=norm_weight)
        else:
>           h_out = conv(g, h, weight=ext_w, edge_weight=norm_weight)

tests\pytorch\test_nn.py:144: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1493AF88>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[ 1.8975],
        [ 0.9999],
        [-0.8757],
        [-2.3468],
        [-1.3300],
        [ 1.9645],
        [-1.6490],
        [-2.4944],
        [-1.2059],
        [-3.4967]])
rhs_data = tensor([[0.6491],
        [1.8020],
        [0.0416],
        [3.2244],
        [0.1902],
        [0.6207],
        [1... [0.6860],
        [0.3175],
        [2.0501],
        [0.8401],
        [2.2833],
        [1.1770],
        [0.2148]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________ test_graph_conv_e_weight_norm[1-True-False-both-g0-idtype0] _________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32), '_src_out_w': Sch...rch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32), '_edge_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'both', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight_norm(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        edgenorm = nn.EdgeWeightNorm(norm=norm)
>       norm_weight = edgenorm(g, g.edata['scalar_w'])

tests\pytorch\test_nn.py:140: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:117: in forward
    reversed_g.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'out_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0110A808>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([0.6491, 1.8020, 0.0416, 3.2244, 0.1902, 0.6207, 1.4090, 0.3025, 1.7357,
        0.6755, 0.6860, 0.3175, 2.0501, 0.8401, 2.2833, 1.1770, 0.2148])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________ test_graph_conv_e_weight_norm[1-True-False-both-g0-idtype1] _________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32), '_src_out_w': Sch...rch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32), '_edge_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'both', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight_norm(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        edgenorm = nn.EdgeWeightNorm(norm=norm)
>       norm_weight = edgenorm(g, g.edata['scalar_w'])

tests\pytorch\test_nn.py:140: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:117: in forward
    reversed_g.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'out_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1458FC48>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([0.6491, 1.8020, 0.0416, 3.2244, 0.1902, 0.6207, 1.4090, 0.3025, 1.7357,
        0.6755, 0.6860, 0.3175, 2.0501, 0.8401, 2.2833, 1.1770, 0.2148])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________ test_graph_conv_e_weight_norm[1-True-False-right-g0-idtype0] _________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32), '_src_out_w': Sch...rch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32), '_edge_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'right', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight_norm(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        edgenorm = nn.EdgeWeightNorm(norm=norm)
>       norm_weight = edgenorm(g, g.edata['scalar_w'])

tests\pytorch\test_nn.py:140: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:123: in forward
    graph.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'in_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE145DF8C8>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([0.6491, 1.8020, 0.0416, 3.2244, 0.1902, 0.6207, 1.4090, 0.3025, 1.7357,
        0.6755, 0.6860, 0.3175, 2.0501, 0.8401, 2.2833, 1.1770, 0.2148])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________ test_graph_conv_e_weight_norm[1-True-False-right-g0-idtype1] _________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32), '_src_out_w': Sch...rch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32), '_edge_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'right', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight_norm(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        edgenorm = nn.EdgeWeightNorm(norm=norm)
>       norm_weight = edgenorm(g, g.edata['scalar_w'])

tests\pytorch\test_nn.py:140: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:123: in forward
    graph.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'in_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE149CFDC8>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([0.6491, 1.8020, 0.0416, 3.2244, 0.1902, 0.6207, 1.4090, 0.3025, 1.7357,
        0.6755, 0.6860, 0.3175, 2.0501, 0.8401, 2.2833, 1.1770, 0.2148])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________ test_graph_conv_e_weight_norm[1-False-True-none-g0-idtype0] _________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32), '_src_out_w': Sch...rch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32), '_edge_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'none', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight_norm(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        edgenorm = nn.EdgeWeightNorm(norm=norm)
        norm_weight = edgenorm(g, g.edata['scalar_w'])
        if weight:
>           h_out = conv(g, h, edge_weight=norm_weight)

tests\pytorch\test_nn.py:142: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1493C288>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[ 0.8146],
        [ 2.1042],
        [ 3.1334],
        [ 0.0720],
        [ 0.2803],
        [-1.7314],
        [ 1.5446],
        [-2.0058],
        [ 1.4075],
        [ 1.4717]], grad_fn=<ViewBackward>)
rhs_data = tensor([[0.6491],
        [1.8020],
        [0.0416],
        [3.2244],
        [0.1902],
        [0.6207],
        [1... [0.6860],
        [0.3175],
        [2.0501],
        [0.8401],
        [2.2833],
        [1.1770],
        [0.2148]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________ test_graph_conv_e_weight_norm[1-False-True-none-g0-idtype1] _________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32), '_src_out_w': Sch...rch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32), '_edge_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'none', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight_norm(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        edgenorm = nn.EdgeWeightNorm(norm=norm)
        norm_weight = edgenorm(g, g.edata['scalar_w'])
        if weight:
>           h_out = conv(g, h, edge_weight=norm_weight)

tests\pytorch\test_nn.py:142: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE15BD75C8>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[-2.4526],
        [-0.1364],
        [ 0.6217],
        [-1.0290],
        [ 2.4242],
        [-1.2578],
        [-0.0596],
        [ 0.4482],
        [-0.4004],
        [-1.5342]], grad_fn=<ViewBackward>)
rhs_data = tensor([[0.6491],
        [1.8020],
        [0.0416],
        [3.2244],
        [0.1902],
        [0.6207],
        [1... [0.6860],
        [0.3175],
        [2.0501],
        [0.8401],
        [2.2833],
        [1.1770],
        [0.2148]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________ test_graph_conv_e_weight_norm[1-False-True-both-g0-idtype0] _________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32), '_src_out_w': Sch...rch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32), '_edge_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'both', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight_norm(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        edgenorm = nn.EdgeWeightNorm(norm=norm)
>       norm_weight = edgenorm(g, g.edata['scalar_w'])

tests\pytorch\test_nn.py:140: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:117: in forward
    reversed_g.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'out_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01573688>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([0.6491, 1.8020, 0.0416, 3.2244, 0.1902, 0.6207, 1.4090, 0.3025, 1.7357,
        0.6755, 0.6860, 0.3175, 2.0501, 0.8401, 2.2833, 1.1770, 0.2148])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________ test_graph_conv_e_weight_norm[1-False-True-both-g0-idtype1] _________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32), '_src_out_w': Sch...rch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32), '_edge_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'both', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight_norm(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        edgenorm = nn.EdgeWeightNorm(norm=norm)
>       norm_weight = edgenorm(g, g.edata['scalar_w'])

tests\pytorch\test_nn.py:140: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:117: in forward
    reversed_g.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'out_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE15216E08>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([0.6491, 1.8020, 0.0416, 3.2244, 0.1902, 0.6207, 1.4090, 0.3025, 1.7357,
        0.6755, 0.6860, 0.3175, 2.0501, 0.8401, 2.2833, 1.1770, 0.2148])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________ test_graph_conv_e_weight_norm[1-False-True-right-g0-idtype0] _________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32), '_src_out_w': Sch...rch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32), '_edge_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'right', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight_norm(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        edgenorm = nn.EdgeWeightNorm(norm=norm)
>       norm_weight = edgenorm(g, g.edata['scalar_w'])

tests\pytorch\test_nn.py:140: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:123: in forward
    graph.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'in_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF017AEB08>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([0.6491, 1.8020, 0.0416, 3.2244, 0.1902, 0.6207, 1.4090, 0.3025, 1.7357,
        0.6755, 0.6860, 0.3175, 2.0501, 0.8401, 2.2833, 1.1770, 0.2148])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________ test_graph_conv_e_weight_norm[1-False-True-right-g0-idtype1] _________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32), '_src_out_w': Sch...rch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32), '_edge_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'right', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight_norm(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        edgenorm = nn.EdgeWeightNorm(norm=norm)
>       norm_weight = edgenorm(g, g.edata['scalar_w'])

tests\pytorch\test_nn.py:140: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:123: in forward
    graph.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'in_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF011EA908>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([0.6491, 1.8020, 0.0416, 3.2244, 0.1902, 0.6207, 1.4090, 0.3025, 1.7357,
        0.6755, 0.6860, 0.3175, 2.0501, 0.8401, 2.2833, 1.1770, 0.2148])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________ test_graph_conv_e_weight_norm[1-False-False-none-g0-idtype0] _________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32), '_src_out_w': Sch...rch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32), '_edge_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'none', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight_norm(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        edgenorm = nn.EdgeWeightNorm(norm=norm)
        norm_weight = edgenorm(g, g.edata['scalar_w'])
        if weight:
            h_out = conv(g, h, edge_weight=norm_weight)
        else:
>           h_out = conv(g, h, weight=ext_w, edge_weight=norm_weight)

tests\pytorch\test_nn.py:144: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE15289D48>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[ 0.9182],
        [-2.5942],
        [ 3.7762],
        [-0.8477],
        [-1.2193],
        [ 1.2860],
        [ 1.5264],
        [-0.2031],
        [-1.1550],
        [ 0.9802]])
rhs_data = tensor([[0.6491],
        [1.8020],
        [0.0416],
        [3.2244],
        [0.1902],
        [0.6207],
        [1... [0.6860],
        [0.3175],
        [2.0501],
        [0.8401],
        [2.2833],
        [1.1770],
        [0.2148]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________ test_graph_conv_e_weight_norm[1-False-False-none-g0-idtype1] _________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32), '_src_out_w': Sch...rch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32), '_edge_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'none', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight_norm(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        edgenorm = nn.EdgeWeightNorm(norm=norm)
        norm_weight = edgenorm(g, g.edata['scalar_w'])
        if weight:
            h_out = conv(g, h, edge_weight=norm_weight)
        else:
>           h_out = conv(g, h, weight=ext_w, edge_weight=norm_weight)

tests\pytorch\test_nn.py:144: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14A0A148>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[-2.7913],
        [-2.9537],
        [ 5.2398],
        [ 2.5936],
        [ 2.1148],
        [ 1.9507],
        [-1.4690],
        [ 3.1176],
        [ 1.6857],
        [-0.2630]])
rhs_data = tensor([[0.6491],
        [1.8020],
        [0.0416],
        [3.2244],
        [0.1902],
        [0.6207],
        [1... [0.6860],
        [0.3175],
        [2.0501],
        [0.8401],
        [2.2833],
        [1.1770],
        [0.2148]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________ test_graph_conv_e_weight_norm[1-False-False-both-g0-idtype0] _________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32), '_src_out_w': Sch...rch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32), '_edge_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'both', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight_norm(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        edgenorm = nn.EdgeWeightNorm(norm=norm)
>       norm_weight = edgenorm(g, g.edata['scalar_w'])

tests\pytorch\test_nn.py:140: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:117: in forward
    reversed_g.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'out_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01A27188>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([0.6491, 1.8020, 0.0416, 3.2244, 0.1902, 0.6207, 1.4090, 0.3025, 1.7357,
        0.6755, 0.6860, 0.3175, 2.0501, 0.8401, 2.2833, 1.1770, 0.2148])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________ test_graph_conv_e_weight_norm[1-False-False-both-g0-idtype1] _________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32), '_src_out_w': Sch...rch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32), '_edge_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'both', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight_norm(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        edgenorm = nn.EdgeWeightNorm(norm=norm)
>       norm_weight = edgenorm(g, g.edata['scalar_w'])

tests\pytorch\test_nn.py:140: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:117: in forward
    reversed_g.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'out_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01AF5648>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([0.6491, 1.8020, 0.0416, 3.2244, 0.1902, 0.6207, 1.4090, 0.3025, 1.7357,
        0.6755, 0.6860, 0.3175, 2.0501, 0.8401, 2.2833, 1.1770, 0.2148])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________ test_graph_conv_e_weight_norm[1-False-False-right-g0-idtype0] ________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32), '_src_out_w': Sch...rch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32), '_edge_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'right', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight_norm(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        edgenorm = nn.EdgeWeightNorm(norm=norm)
>       norm_weight = edgenorm(g, g.edata['scalar_w'])

tests\pytorch\test_nn.py:140: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:123: in forward
    graph.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'in_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0151F408>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([0.6491, 1.8020, 0.0416, 3.2244, 0.1902, 0.6207, 1.4090, 0.3025, 1.7357,
        0.6755, 0.6860, 0.3175, 2.0501, 0.8401, 2.2833, 1.1770, 0.2148])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________ test_graph_conv_e_weight_norm[1-False-False-right-g0-idtype1] ________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32), '_src_out_w': Sch...rch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32), '_edge_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'right', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight_norm(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        edgenorm = nn.EdgeWeightNorm(norm=norm)
>       norm_weight = edgenorm(g, g.edata['scalar_w'])

tests\pytorch\test_nn.py:140: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:123: in forward
    graph.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'in_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF076EB748>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([0.6491, 1.8020, 0.0416, 3.2244, 0.1902, 0.6207, 1.4090, 0.3025, 1.7357,
        0.6755, 0.6860, 0.3175, 2.0501, 0.8401, 2.2833, 1.1770, 0.2148])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________ test_graph_conv_e_weight_norm[2-True-True-none-g0-idtype0] __________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), '_src_out_w': Sch...rch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32), '_edge_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'none', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight_norm(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        edgenorm = nn.EdgeWeightNorm(norm=norm)
        norm_weight = edgenorm(g, g.edata['scalar_w'])
        if weight:
>           h_out = conv(g, h, edge_weight=norm_weight)

tests\pytorch\test_nn.py:142: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14680D48>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[-0.6632, -0.5378],
        [ 2.6249,  2.1222],
        [-0.0746,  1.8017],
        [ 2.2846,  1.3546],
      ... -1.6471],
        [-0.9610, -0.2798],
        [-0.2984, -2.8773],
        [ 0.9076, -1.1007]], grad_fn=<ViewBackward>)
rhs_data = tensor([[0.6491],
        [1.8020],
        [0.0416],
        [3.2244],
        [0.1902],
        [0.6207],
        [1... [0.6860],
        [0.3175],
        [2.0501],
        [0.8401],
        [2.2833],
        [1.1770],
        [0.2148]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________ test_graph_conv_e_weight_norm[2-True-True-none-g0-idtype1] __________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), '_src_out_w': Sch...rch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32), '_edge_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'none', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight_norm(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        edgenorm = nn.EdgeWeightNorm(norm=norm)
        norm_weight = edgenorm(g, g.edata['scalar_w'])
        if weight:
>           h_out = conv(g, h, edge_weight=norm_weight)

tests\pytorch\test_nn.py:142: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF02FCDBC8>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[ 0.6765, -0.1248],
        [-0.1027,  1.0637],
        [ 0.4872, -0.0853],
        [ 0.6262,  0.3190],
      ... -0.1995],
        [-0.3957,  0.0632],
        [-0.4042, -0.2038],
        [ 0.5874, -0.5352]], grad_fn=<ViewBackward>)
rhs_data = tensor([[0.6491],
        [1.8020],
        [0.0416],
        [3.2244],
        [0.1902],
        [0.6207],
        [1... [0.6860],
        [0.3175],
        [2.0501],
        [0.8401],
        [2.2833],
        [1.1770],
        [0.2148]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________ test_graph_conv_e_weight_norm[2-True-True-both-g0-idtype0] __________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), '_src_out_w': Sch...rch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32), '_edge_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'both', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight_norm(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        edgenorm = nn.EdgeWeightNorm(norm=norm)
>       norm_weight = edgenorm(g, g.edata['scalar_w'])

tests\pytorch\test_nn.py:140: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:117: in forward
    reversed_g.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'out_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14626B08>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([0.6491, 1.8020, 0.0416, 3.2244, 0.1902, 0.6207, 1.4090, 0.3025, 1.7357,
        0.6755, 0.6860, 0.3175, 2.0501, 0.8401, 2.2833, 1.1770, 0.2148])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________ test_graph_conv_e_weight_norm[2-True-True-both-g0-idtype1] __________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), '_src_out_w': Sch...rch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32), '_edge_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'both', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight_norm(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        edgenorm = nn.EdgeWeightNorm(norm=norm)
>       norm_weight = edgenorm(g, g.edata['scalar_w'])

tests\pytorch\test_nn.py:140: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:117: in forward
    reversed_g.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'out_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF011E0AC8>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([0.6491, 1.8020, 0.0416, 3.2244, 0.1902, 0.6207, 1.4090, 0.3025, 1.7357,
        0.6755, 0.6860, 0.3175, 2.0501, 0.8401, 2.2833, 1.1770, 0.2148])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________ test_graph_conv_e_weight_norm[2-True-True-right-g0-idtype0] _________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), '_src_out_w': Sch...rch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32), '_edge_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'right', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight_norm(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        edgenorm = nn.EdgeWeightNorm(norm=norm)
>       norm_weight = edgenorm(g, g.edata['scalar_w'])

tests\pytorch\test_nn.py:140: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:123: in forward
    graph.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'in_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF011EB448>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([0.6491, 1.8020, 0.0416, 3.2244, 0.1902, 0.6207, 1.4090, 0.3025, 1.7357,
        0.6755, 0.6860, 0.3175, 2.0501, 0.8401, 2.2833, 1.1770, 0.2148])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________ test_graph_conv_e_weight_norm[2-True-True-right-g0-idtype1] _________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), '_src_out_w': Sch...rch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32), '_edge_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'right', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight_norm(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        edgenorm = nn.EdgeWeightNorm(norm=norm)
>       norm_weight = edgenorm(g, g.edata['scalar_w'])

tests\pytorch\test_nn.py:140: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:123: in forward
    graph.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'in_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF011D07C8>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([0.6491, 1.8020, 0.0416, 3.2244, 0.1902, 0.6207, 1.4090, 0.3025, 1.7357,
        0.6755, 0.6860, 0.3175, 2.0501, 0.8401, 2.2833, 1.1770, 0.2148])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________ test_graph_conv_e_weight_norm[2-True-False-none-g0-idtype0] _________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), '_src_out_w': Sch...rch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32), '_edge_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'none', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight_norm(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        edgenorm = nn.EdgeWeightNorm(norm=norm)
        norm_weight = edgenorm(g, g.edata['scalar_w'])
        if weight:
            h_out = conv(g, h, edge_weight=norm_weight)
        else:
>           h_out = conv(g, h, weight=ext_w, edge_weight=norm_weight)

tests\pytorch\test_nn.py:144: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE15289F08>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[-1.8320, -2.9323],
        [ 1.0977,  0.8555],
        [ 1.4111,  0.6233],
        [-0.9969,  0.0581],
      ...3791],
        [-0.9958, -1.1627],
        [-1.4568,  0.0343],
        [-1.2527, -0.9189],
        [-0.2323,  0.4465]])
rhs_data = tensor([[0.6491],
        [1.8020],
        [0.0416],
        [3.2244],
        [0.1902],
        [0.6207],
        [1... [0.6860],
        [0.3175],
        [2.0501],
        [0.8401],
        [2.2833],
        [1.1770],
        [0.2148]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________ test_graph_conv_e_weight_norm[2-True-False-none-g0-idtype1] _________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), '_src_out_w': Sch...rch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32), '_edge_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'none', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight_norm(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        edgenorm = nn.EdgeWeightNorm(norm=norm)
        norm_weight = edgenorm(g, g.edata['scalar_w'])
        if weight:
            h_out = conv(g, h, edge_weight=norm_weight)
        else:
>           h_out = conv(g, h, weight=ext_w, edge_weight=norm_weight)

tests\pytorch\test_nn.py:144: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14A795C8>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[ 3.5334, -2.9222],
        [ 0.0948, -1.5191],
        [-0.1692,  0.9799],
        [-2.5485,  1.2069],
      ...3509],
        [-2.6636,  1.6560],
        [-0.7663,  3.8942],
        [ 1.3865, -3.3800],
        [-1.9038,  3.9313]])
rhs_data = tensor([[0.6491],
        [1.8020],
        [0.0416],
        [3.2244],
        [0.1902],
        [0.6207],
        [1... [0.6860],
        [0.3175],
        [2.0501],
        [0.8401],
        [2.2833],
        [1.1770],
        [0.2148]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________ test_graph_conv_e_weight_norm[2-True-False-both-g0-idtype0] _________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), '_src_out_w': Sch...rch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32), '_edge_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'both', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight_norm(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        edgenorm = nn.EdgeWeightNorm(norm=norm)
>       norm_weight = edgenorm(g, g.edata['scalar_w'])

tests\pytorch\test_nn.py:140: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:117: in forward
    reversed_g.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'out_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14626E88>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([0.6491, 1.8020, 0.0416, 3.2244, 0.1902, 0.6207, 1.4090, 0.3025, 1.7357,
        0.6755, 0.6860, 0.3175, 2.0501, 0.8401, 2.2833, 1.1770, 0.2148])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________ test_graph_conv_e_weight_norm[2-True-False-both-g0-idtype1] _________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), '_src_out_w': Sch...rch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32), '_edge_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'both', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight_norm(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        edgenorm = nn.EdgeWeightNorm(norm=norm)
>       norm_weight = edgenorm(g, g.edata['scalar_w'])

tests\pytorch\test_nn.py:140: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:117: in forward
    reversed_g.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'out_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0109CFC8>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([0.6491, 1.8020, 0.0416, 3.2244, 0.1902, 0.6207, 1.4090, 0.3025, 1.7357,
        0.6755, 0.6860, 0.3175, 2.0501, 0.8401, 2.2833, 1.1770, 0.2148])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________ test_graph_conv_e_weight_norm[2-True-False-right-g0-idtype0] _________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), '_src_out_w': Sch...rch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32), '_edge_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'right', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight_norm(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        edgenorm = nn.EdgeWeightNorm(norm=norm)
>       norm_weight = edgenorm(g, g.edata['scalar_w'])

tests\pytorch\test_nn.py:140: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:123: in forward
    graph.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'in_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14628248>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([0.6491, 1.8020, 0.0416, 3.2244, 0.1902, 0.6207, 1.4090, 0.3025, 1.7357,
        0.6755, 0.6860, 0.3175, 2.0501, 0.8401, 2.2833, 1.1770, 0.2148])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________ test_graph_conv_e_weight_norm[2-True-False-right-g0-idtype1] _________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), '_src_out_w': Sch...rch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32), '_edge_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'right', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight_norm(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        edgenorm = nn.EdgeWeightNorm(norm=norm)
>       norm_weight = edgenorm(g, g.edata['scalar_w'])

tests\pytorch\test_nn.py:140: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:123: in forward
    graph.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'in_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01682DC8>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([0.6491, 1.8020, 0.0416, 3.2244, 0.1902, 0.6207, 1.4090, 0.3025, 1.7357,
        0.6755, 0.6860, 0.3175, 2.0501, 0.8401, 2.2833, 1.1770, 0.2148])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________ test_graph_conv_e_weight_norm[2-False-True-none-g0-idtype0] _________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), '_src_out_w': Sch...rch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32), '_edge_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'none', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight_norm(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        edgenorm = nn.EdgeWeightNorm(norm=norm)
        norm_weight = edgenorm(g, g.edata['scalar_w'])
        if weight:
>           h_out = conv(g, h, edge_weight=norm_weight)

tests\pytorch\test_nn.py:142: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1493C248>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[-0.8804,  0.9705],
        [-0.5709,  0.0417],
        [-0.4043,  0.4932],
        [-1.0464, -0.6110],
      ...  0.8335],
        [-1.9237,  2.2635],
        [ 1.5175, -1.8448],
        [ 1.2880, -0.9630]], grad_fn=<ViewBackward>)
rhs_data = tensor([[0.6491],
        [1.8020],
        [0.0416],
        [3.2244],
        [0.1902],
        [0.6207],
        [1... [0.6860],
        [0.3175],
        [2.0501],
        [0.8401],
        [2.2833],
        [1.1770],
        [0.2148]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________ test_graph_conv_e_weight_norm[2-False-True-none-g0-idtype1] _________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), '_src_out_w': Sch...rch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32), '_edge_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'none', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight_norm(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        edgenorm = nn.EdgeWeightNorm(norm=norm)
        norm_weight = edgenorm(g, g.edata['scalar_w'])
        if weight:
>           h_out = conv(g, h, edge_weight=norm_weight)

tests\pytorch\test_nn.py:142: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE146167C8>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[ 0.1116, -0.9953],
        [ 0.1379, -0.7153],
        [ 0.4194,  1.2042],
        [ 0.4357, -0.7326],
      ...  0.1649],
        [-0.7040, -0.0801],
        [-1.7342,  0.7475],
        [-0.0522,  0.5195]], grad_fn=<ViewBackward>)
rhs_data = tensor([[0.6491],
        [1.8020],
        [0.0416],
        [3.2244],
        [0.1902],
        [0.6207],
        [1... [0.6860],
        [0.3175],
        [2.0501],
        [0.8401],
        [2.2833],
        [1.1770],
        [0.2148]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________ test_graph_conv_e_weight_norm[2-False-True-both-g0-idtype0] _________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), '_src_out_w': Sch...rch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32), '_edge_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'both', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight_norm(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        edgenorm = nn.EdgeWeightNorm(norm=norm)
>       norm_weight = edgenorm(g, g.edata['scalar_w'])

tests\pytorch\test_nn.py:140: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:117: in forward
    reversed_g.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'out_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF010C4F48>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([0.6491, 1.8020, 0.0416, 3.2244, 0.1902, 0.6207, 1.4090, 0.3025, 1.7357,
        0.6755, 0.6860, 0.3175, 2.0501, 0.8401, 2.2833, 1.1770, 0.2148])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________ test_graph_conv_e_weight_norm[2-False-True-both-g0-idtype1] _________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), '_src_out_w': Sch...rch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32), '_edge_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'both', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight_norm(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        edgenorm = nn.EdgeWeightNorm(norm=norm)
>       norm_weight = edgenorm(g, g.edata['scalar_w'])

tests\pytorch\test_nn.py:140: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:117: in forward
    reversed_g.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'out_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01119D48>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([0.6491, 1.8020, 0.0416, 3.2244, 0.1902, 0.6207, 1.4090, 0.3025, 1.7357,
        0.6755, 0.6860, 0.3175, 2.0501, 0.8401, 2.2833, 1.1770, 0.2148])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________ test_graph_conv_e_weight_norm[2-False-True-right-g0-idtype0] _________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), '_src_out_w': Sch...rch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32), '_edge_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'right', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight_norm(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        edgenorm = nn.EdgeWeightNorm(norm=norm)
>       norm_weight = edgenorm(g, g.edata['scalar_w'])

tests\pytorch\test_nn.py:140: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:123: in forward
    graph.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'in_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14651A08>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([0.6491, 1.8020, 0.0416, 3.2244, 0.1902, 0.6207, 1.4090, 0.3025, 1.7357,
        0.6755, 0.6860, 0.3175, 2.0501, 0.8401, 2.2833, 1.1770, 0.2148])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________ test_graph_conv_e_weight_norm[2-False-True-right-g0-idtype1] _________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), '_src_out_w': Sch...rch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32), '_edge_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'right', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight_norm(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        edgenorm = nn.EdgeWeightNorm(norm=norm)
>       norm_weight = edgenorm(g, g.edata['scalar_w'])

tests\pytorch\test_nn.py:140: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:123: in forward
    graph.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'in_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0109A6C8>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([0.6491, 1.8020, 0.0416, 3.2244, 0.1902, 0.6207, 1.4090, 0.3025, 1.7357,
        0.6755, 0.6860, 0.3175, 2.0501, 0.8401, 2.2833, 1.1770, 0.2148])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________ test_graph_conv_e_weight_norm[2-False-False-none-g0-idtype0] _________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), '_src_out_w': Sch...rch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32), '_edge_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'none', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight_norm(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        edgenorm = nn.EdgeWeightNorm(norm=norm)
        norm_weight = edgenorm(g, g.edata['scalar_w'])
        if weight:
            h_out = conv(g, h, edge_weight=norm_weight)
        else:
>           h_out = conv(g, h, weight=ext_w, edge_weight=norm_weight)

tests\pytorch\test_nn.py:144: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE145B4648>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[ 0.2091, -2.2793],
        [-4.5884, -2.2817],
        [-2.0145,  2.3447],
        [-2.9322,  1.0785],
      ...8693],
        [ 0.3129,  2.1681],
        [ 1.2107,  1.8013],
        [-1.6702, -2.0279],
        [ 0.0305,  0.9249]])
rhs_data = tensor([[0.6491],
        [1.8020],
        [0.0416],
        [3.2244],
        [0.1902],
        [0.6207],
        [1... [0.6860],
        [0.3175],
        [2.0501],
        [0.8401],
        [2.2833],
        [1.1770],
        [0.2148]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________ test_graph_conv_e_weight_norm[2-False-False-none-g0-idtype1] _________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), '_src_out_w': Sch...rch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32), '_edge_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'none', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight_norm(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        edgenorm = nn.EdgeWeightNorm(norm=norm)
        norm_weight = edgenorm(g, g.edata['scalar_w'])
        if weight:
            h_out = conv(g, h, edge_weight=norm_weight)
        else:
>           h_out = conv(g, h, weight=ext_w, edge_weight=norm_weight)

tests\pytorch\test_nn.py:144: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1456E808>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[-4.3074, -2.7269],
        [-0.9254, -0.4562],
        [-2.4350,  0.6966],
        [-3.8353,  0.3013],
      ...9098],
        [-2.9923, -2.0561],
        [ 1.0857,  1.0475],
        [ 0.2443, -2.8310],
        [-2.3665,  2.0807]])
rhs_data = tensor([[0.6491],
        [1.8020],
        [0.0416],
        [3.2244],
        [0.1902],
        [0.6207],
        [1... [0.6860],
        [0.3175],
        [2.0501],
        [0.8401],
        [2.2833],
        [1.1770],
        [0.2148]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________ test_graph_conv_e_weight_norm[2-False-False-both-g0-idtype0] _________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), '_src_out_w': Sch...rch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32), '_edge_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'both', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight_norm(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        edgenorm = nn.EdgeWeightNorm(norm=norm)
>       norm_weight = edgenorm(g, g.edata['scalar_w'])

tests\pytorch\test_nn.py:140: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:117: in forward
    reversed_g.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'out_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF015F4308>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([0.6491, 1.8020, 0.0416, 3.2244, 0.1902, 0.6207, 1.4090, 0.3025, 1.7357,
        0.6755, 0.6860, 0.3175, 2.0501, 0.8401, 2.2833, 1.1770, 0.2148])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________ test_graph_conv_e_weight_norm[2-False-False-both-g0-idtype1] _________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), '_src_out_w': Sch...rch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32), '_edge_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'both', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight_norm(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        edgenorm = nn.EdgeWeightNorm(norm=norm)
>       norm_weight = edgenorm(g, g.edata['scalar_w'])

tests\pytorch\test_nn.py:140: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:117: in forward
    reversed_g.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'out_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE15277F08>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([0.6491, 1.8020, 0.0416, 3.2244, 0.1902, 0.6207, 1.4090, 0.3025, 1.7357,
        0.6755, 0.6860, 0.3175, 2.0501, 0.8401, 2.2833, 1.1770, 0.2148])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________ test_graph_conv_e_weight_norm[2-False-False-right-g0-idtype0] ________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), '_src_out_w': Sch...rch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32), '_edge_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'right', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight_norm(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        edgenorm = nn.EdgeWeightNorm(norm=norm)
>       norm_weight = edgenorm(g, g.edata['scalar_w'])

tests\pytorch\test_nn.py:140: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:123: in forward
    graph.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'in_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14552088>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([0.6491, 1.8020, 0.0416, 3.2244, 0.1902, 0.6207, 1.4090, 0.3025, 1.7357,
        0.6755, 0.6860, 0.3175, 2.0501, 0.8401, 2.2833, 1.1770, 0.2148])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________ test_graph_conv_e_weight_norm[2-False-False-right-g0-idtype1] ________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), '_src_out_w': Sch...rch.float32), '_edge_weight': Scheme(shape=(), dtype=torch.float32), '_edge_w': Scheme(shape=(), dtype=torch.float32)})
norm = 'right', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_e_weight_norm(idtype, g, norm, weight, bias, out_dim):
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        edgenorm = nn.EdgeWeightNorm(norm=norm)
>       norm_weight = edgenorm(g, g.edata['scalar_w'])

tests\pytorch\test_nn.py:140: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:123: in forward
    graph.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'in_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0105B2C8>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([0.6491, 1.8020, 0.0416, 3.2244, 0.1902, 0.6207, 1.4090, 0.3025, 1.7357,
        0.6755, 0.6860, 0.3175, 2.0501, 0.8401, 2.2833, 1.1770, 0.2148])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv_bi[1-True-True-none-g0-idtype0] _______________

idtype = torch.int32
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'none', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
>           h_out = conv(g, (h, h_dst))

tests\pytorch\test_nn.py:167: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03F34D08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-2.7172],
        [ 1.6776]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv_bi[1-True-True-none-g0-idtype1] _______________

idtype = torch.int64
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'none', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
>           h_out = conv(g, (h, h_dst))

tests\pytorch\test_nn.py:167: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01AA6748>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.8676],
        [-0.1864]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv_bi[1-True-True-none-g1-idtype0] _______________

idtype = torch.int32, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'none', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
>           h_out = conv(g, (h, h_dst))

tests\pytorch\test_nn.py:167: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF011F0DC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.1473],
        [-0.7679],
        [-1.1437],
        [ 1.4817],
        [ 0.1744],
        [-0.0350]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv_bi[1-True-True-none-g1-idtype1] _______________

idtype = torch.int64, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'none', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
>           h_out = conv(g, (h, h_dst))

tests\pytorch\test_nn.py:167: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1458FD08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.0365],
        [-0.4701],
        [-0.3467],
        [ 1.4001],
        [-0.6322],
        [-0.2797]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv_bi[1-True-True-both-g0-idtype0] _______________

idtype = torch.int32
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'both', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
>           h_out = conv(g, (h, h_dst))

tests\pytorch\test_nn.py:167: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF011EBE88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.0853],
        [ 0.4658]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv_bi[1-True-True-both-g0-idtype1] _______________

idtype = torch.int64
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'both', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
>           h_out = conv(g, (h, h_dst))

tests\pytorch\test_nn.py:167: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE15BCE3C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.1521],
        [ 0.0932]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv_bi[1-True-True-both-g1-idtype0] _______________

idtype = torch.int32, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'both', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
>           h_out = conv(g, (h, h_dst))

tests\pytorch\test_nn.py:167: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE146AB888>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.4724],
        [-0.8164],
        [ 1.0302],
        [-0.0519],
        [-0.2556],
        [-4.5618]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv_bi[1-True-True-both-g1-idtype1] _______________

idtype = torch.int64, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'both', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
>           h_out = conv(g, (h, h_dst))

tests\pytorch\test_nn.py:167: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF02F86508>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.7729],
        [-0.8407],
        [ 0.8924],
        [ 0.3327],
        [-0.6871],
        [-0.0722]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________ test_graph_conv_bi[1-True-True-right-g0-idtype0] _______________

idtype = torch.int32
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'right', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
>           h_out = conv(g, (h, h_dst))

tests\pytorch\test_nn.py:167: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF077267C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.5966],
        [ 0.4681]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________ test_graph_conv_bi[1-True-True-right-g0-idtype1] _______________

idtype = torch.int64
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'right', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
>           h_out = conv(g, (h, h_dst))

tests\pytorch\test_nn.py:167: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE15BA6F88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.1633],
        [-2.1853]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________ test_graph_conv_bi[1-True-True-right-g1-idtype0] _______________

idtype = torch.int32, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'right', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
>           h_out = conv(g, (h, h_dst))

tests\pytorch\test_nn.py:167: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE145DED48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.7069],
        [ 0.0694],
        [-1.1921],
        [-0.6330],
        [-0.3881],
        [ 0.8595]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________ test_graph_conv_bi[1-True-True-right-g1-idtype1] _______________

idtype = torch.int64, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'right', weight = True, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
>           h_out = conv(g, (h, h_dst))

tests\pytorch\test_nn.py:167: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF02F27AC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.2839],
        [ 1.7440],
        [ 0.4472],
        [ 0.0149],
        [ 2.3965],
        [-1.2578]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________ test_graph_conv_bi[1-True-False-none-g0-idtype0] _______________

idtype = torch.int32
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'none', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
            h_out = conv(g, (h, h_dst))
        else:
>           h_out = conv(g, (h, h_dst), weight=ext_w)

tests\pytorch\test_nn.py:169: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03EEBD88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.4256],
        [-1.2956]]), rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________ test_graph_conv_bi[1-True-False-none-g0-idtype1] _______________

idtype = torch.int64
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'none', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
            h_out = conv(g, (h, h_dst))
        else:
>           h_out = conv(g, (h, h_dst), weight=ext_w)

tests\pytorch\test_nn.py:169: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0178DA88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-2.4073],
        [-5.0885]]), rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________ test_graph_conv_bi[1-True-False-none-g1-idtype0] _______________

idtype = torch.int32, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'none', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
            h_out = conv(g, (h, h_dst))
        else:
>           h_out = conv(g, (h, h_dst), weight=ext_w)

tests\pytorch\test_nn.py:169: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE149D2988>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.3708],
        [-0.6346],
        [ 2.5032],
        [-0.9559],
        [-1.5705],
        [ 1.0833]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________ test_graph_conv_bi[1-True-False-none-g1-idtype1] _______________

idtype = torch.int64, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'none', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
            h_out = conv(g, (h, h_dst))
        else:
>           h_out = conv(g, (h, h_dst), weight=ext_w)

tests\pytorch\test_nn.py:169: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01769D08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 3.9922],
        [-1.9594],
        [-1.3310],
        [ 0.3653],
        [-2.8011],
        [-3.7338]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________ test_graph_conv_bi[1-True-False-both-g0-idtype0] _______________

idtype = torch.int32
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'both', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
            h_out = conv(g, (h, h_dst))
        else:
>           h_out = conv(g, (h, h_dst), weight=ext_w)

tests\pytorch\test_nn.py:169: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07632B48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-2.6280],
        [-0.5512]]), rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________ test_graph_conv_bi[1-True-False-both-g0-idtype1] _______________

idtype = torch.int64
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'both', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
            h_out = conv(g, (h, h_dst))
        else:
>           h_out = conv(g, (h, h_dst), weight=ext_w)

tests\pytorch\test_nn.py:169: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF015B4308>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.7553],
        [ 0.9498]]), rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________ test_graph_conv_bi[1-True-False-both-g1-idtype0] _______________

idtype = torch.int32, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'both', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
            h_out = conv(g, (h, h_dst))
        else:
>           h_out = conv(g, (h, h_dst), weight=ext_w)


_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0177D9C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.1192],
        [ 0.3118],
        [-0.9258],
        [ 0.4978],
        [-0.3038],
        [ 0.8208]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________ test_graph_conv_bi[1-True-False-both-g1-idtype1] _______________

idtype = torch.int64, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'both', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
            h_out = conv(g, (h, h_dst))
        else:
>           h_out = conv(g, (h, h_dst), weight=ext_w)

tests\pytorch\test_nn.py:169: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE149BA388>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 6.6321],
        [-0.4705],
        [-2.7860],
        [ 0.7577],
        [ 2.6282],
        [ 0.7754]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________ test_graph_conv_bi[1-True-False-right-g0-idtype0] ______________

idtype = torch.int32
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'right', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
            h_out = conv(g, (h, h_dst))
        else:
>           h_out = conv(g, (h, h_dst), weight=ext_w)

tests\pytorch\test_nn.py:169: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0177D7C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 3.0247],
        [-1.0844]]), rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________ test_graph_conv_bi[1-True-False-right-g0-idtype1] ______________

idtype = torch.int64
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'right', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
            h_out = conv(g, (h, h_dst))
        else:
>           h_out = conv(g, (h, h_dst), weight=ext_w)

tests\pytorch\test_nn.py:169: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01076088>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.3362],
        [-0.3181]]), rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________ test_graph_conv_bi[1-True-False-right-g1-idtype0] ______________

idtype = torch.int32, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'right', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
            h_out = conv(g, (h, h_dst))
        else:
>           h_out = conv(g, (h, h_dst), weight=ext_w)

tests\pytorch\test_nn.py:169: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF015F4208>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.0132],
        [ 0.5334],
        [-5.4003],
        [-0.1352],
        [ 0.6653],
        [-0.5733]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________ test_graph_conv_bi[1-True-False-right-g1-idtype1] ______________

idtype = torch.int64, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'right', weight = False, bias = True, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
            h_out = conv(g, (h, h_dst))
        else:
>           h_out = conv(g, (h, h_dst), weight=ext_w)

tests\pytorch\test_nn.py:169: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14AA4748>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.1734],
        [-2.4628],
        [-0.2079],
        [-2.8298],
        [ 1.3201],
        [-0.3578]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________ test_graph_conv_bi[1-False-True-none-g0-idtype0] _______________

idtype = torch.int32
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'none', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
>           h_out = conv(g, (h, h_dst))

tests\pytorch\test_nn.py:167: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF076CF488>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.2656],
        [ 0.0528]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________ test_graph_conv_bi[1-False-True-none-g0-idtype1] _______________

idtype = torch.int64
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'none', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
>           h_out = conv(g, (h, h_dst))

tests\pytorch\test_nn.py:167: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14AAEAC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.9401],
        [-1.0459]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________ test_graph_conv_bi[1-False-True-none-g1-idtype0] _______________

idtype = torch.int32, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'none', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
>           h_out = conv(g, (h, h_dst))

tests\pytorch\test_nn.py:167: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01054D88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.5821],
        [ 2.1218],
        [ 0.2685],
        [ 1.2784],
        [-0.4486],
        [ 0.7418]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________ test_graph_conv_bi[1-False-True-none-g1-idtype1] _______________

idtype = torch.int64, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'none', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
>           h_out = conv(g, (h, h_dst))

tests\pytorch\test_nn.py:167: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14AA4508>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.2911],
        [ 1.1131],
        [-1.0148],
        [-1.0035],
        [ 2.1734],
        [ 1.1257]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________ test_graph_conv_bi[1-False-True-both-g0-idtype0] _______________

idtype = torch.int32
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'both', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
>           h_out = conv(g, (h, h_dst))

tests\pytorch\test_nn.py:167: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0176AA88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.6304],
        [ 0.6561]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________ test_graph_conv_bi[1-False-True-both-g0-idtype1] _______________

idtype = torch.int64
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'both', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
>           h_out = conv(g, (h, h_dst))

tests\pytorch\test_nn.py:167: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14A63F08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0.6441],
        [0.5059]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________ test_graph_conv_bi[1-False-True-both-g1-idtype0] _______________

idtype = torch.int32, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'both', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
>           h_out = conv(g, (h, h_dst))

tests\pytorch\test_nn.py:167: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14AA48C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.8914],
        [ 0.6981],
        [-0.6616],
        [-1.3130],
        [ 0.9533],
        [ 0.9548]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________ test_graph_conv_bi[1-False-True-both-g1-idtype1] _______________

idtype = torch.int64, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'both', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
>           h_out = conv(g, (h, h_dst))

tests\pytorch\test_nn.py:167: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0761A2C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.0274],
        [-1.8962],
        [-1.3701],
        [-1.1161],
        [ 2.3365],
        [-0.8380]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________ test_graph_conv_bi[1-False-True-right-g0-idtype0] ______________

idtype = torch.int32
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'right', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
>           h_out = conv(g, (h, h_dst))

tests\pytorch\test_nn.py:167: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14A66EC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.4212],
        [-0.9779]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________ test_graph_conv_bi[1-False-True-right-g0-idtype1] ______________

idtype = torch.int64
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'right', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
>           h_out = conv(g, (h, h_dst))

tests\pytorch\test_nn.py:167: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14AA85C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[1.1444],
        [0.8866]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________ test_graph_conv_bi[1-False-True-right-g1-idtype0] ______________

idtype = torch.int32, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'right', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
>           h_out = conv(g, (h, h_dst))

tests\pytorch\test_nn.py:167: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03F04608>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-2.0315],
        [-0.3794],
        [-0.1204],
        [-1.6265],
        [ 0.5929],
        [ 1.9186]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________ test_graph_conv_bi[1-False-True-right-g1-idtype1] ______________

idtype = torch.int64, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'right', weight = True, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
>           h_out = conv(g, (h, h_dst))

tests\pytorch\test_nn.py:167: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01A76088>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.4248],
        [-1.2117],
        [-1.2799],
        [-0.1726],
        [ 0.1885],
        [-0.0256]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________ test_graph_conv_bi[1-False-False-none-g0-idtype0] ______________

idtype = torch.int32
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'none', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
            h_out = conv(g, (h, h_dst))
        else:
>           h_out = conv(g, (h, h_dst), weight=ext_w)

tests\pytorch\test_nn.py:169: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14594488>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.6835],
        [10.3995]]), rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________ test_graph_conv_bi[1-False-False-none-g0-idtype1] ______________

idtype = torch.int64
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'none', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
            h_out = conv(g, (h, h_dst))
        else:
>           h_out = conv(g, (h, h_dst), weight=ext_w)

tests\pytorch\test_nn.py:169: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03018188>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-2.4053],
        [ 0.9820]]), rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________ test_graph_conv_bi[1-False-False-none-g1-idtype0] ______________

idtype = torch.int32, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'none', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
            h_out = conv(g, (h, h_dst))
        else:
>           h_out = conv(g, (h, h_dst), weight=ext_w)

tests\pytorch\test_nn.py:169: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14AB9D88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-2.2426],
        [ 2.1040],
        [ 3.4066],
        [ 1.9417],
        [ 1.3830],
        [ 0.2790]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________ test_graph_conv_bi[1-False-False-none-g1-idtype1] ______________

idtype = torch.int64, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'none', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
            h_out = conv(g, (h, h_dst))
        else:
>           h_out = conv(g, (h, h_dst), weight=ext_w)

tests\pytorch\test_nn.py:169: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF011EB908>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.4491],
        [-1.8606],
        [ 2.4502],
        [ 0.3883],
        [ 0.2064],
        [-1.0713]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________ test_graph_conv_bi[1-False-False-both-g0-idtype0] ______________

idtype = torch.int32
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'both', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
            h_out = conv(g, (h, h_dst))
        else:
>           h_out = conv(g, (h, h_dst), weight=ext_w)

tests\pytorch\test_nn.py:169: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01088048>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.8584],
        [ 3.1270]]), rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________ test_graph_conv_bi[1-False-False-both-g0-idtype1] ______________

idtype = torch.int64
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'both', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
            h_out = conv(g, (h, h_dst))
        else:
>           h_out = conv(g, (h, h_dst), weight=ext_w)

tests\pytorch\test_nn.py:169: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0142CE88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.3032],
        [ 0.2018]]), rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________ test_graph_conv_bi[1-False-False-both-g1-idtype0] ______________

idtype = torch.int32, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'both', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
            h_out = conv(g, (h, h_dst))
        else:
>           h_out = conv(g, (h, h_dst), weight=ext_w)

tests\pytorch\test_nn.py:169: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01158648>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.4774],
        [-1.6989],
        [ 0.4459],
        [ 1.2344],
        [ 1.5566],
        [-1.2642]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________ test_graph_conv_bi[1-False-False-both-g1-idtype1] ______________

idtype = torch.int64, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'both', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
            h_out = conv(g, (h, h_dst))
        else:
>           h_out = conv(g, (h, h_dst), weight=ext_w)

tests\pytorch\test_nn.py:169: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01620288>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.8424],
        [-1.3221],
        [ 1.0961],
        [-0.1399],
        [ 0.4930],
        [-2.1929]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_____________ test_graph_conv_bi[1-False-False-right-g0-idtype0] ______________

idtype = torch.int32
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'right', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
            h_out = conv(g, (h, h_dst))
        else:
>           h_out = conv(g, (h, h_dst), weight=ext_w)

tests\pytorch\test_nn.py:169: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF010C4248>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0.1409],
        [5.0546]]), rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_____________ test_graph_conv_bi[1-False-False-right-g0-idtype1] ______________

idtype = torch.int64
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'right', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
            h_out = conv(g, (h, h_dst))
        else:
>           h_out = conv(g, (h, h_dst), weight=ext_w)

tests\pytorch\test_nn.py:169: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0146E408>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-2.1464],
        [-0.3391]]), rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_____________ test_graph_conv_bi[1-False-False-right-g1-idtype0] ______________

idtype = torch.int32, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'right', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
            h_out = conv(g, (h, h_dst))
        else:
>           h_out = conv(g, (h, h_dst), weight=ext_w)

tests\pytorch\test_nn.py:169: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01059508>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.0038],
        [ 1.9149],
        [-0.5750],
        [ 2.8045],
        [ 2.7590],
        [ 0.1439]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_____________ test_graph_conv_bi[1-False-False-right-g1-idtype1] ______________

idtype = torch.int64, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'right', weight = False, bias = False, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
            h_out = conv(g, (h, h_dst))
        else:
>           h_out = conv(g, (h, h_dst), weight=ext_w)

tests\pytorch\test_nn.py:169: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01432108>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.5053],
        [-0.2679],
        [ 0.4639],
        [ 1.3614],
        [-0.8184],
        [ 1.0998]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv_bi[2-True-True-none-g0-idtype0] _______________

idtype = torch.int32
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'none', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
>           h_out = conv(g, (h, h_dst))

tests\pytorch\test_nn.py:167: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0162FE08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.0930,  0.2305],
        [-2.0785, -1.7662]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv_bi[2-True-True-none-g0-idtype1] _______________

idtype = torch.int64
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'none', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
>           h_out = conv(g, (h, h_dst))

tests\pytorch\test_nn.py:167: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF010B6A88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.8775, -1.6434],
        [ 0.8330, -1.3285]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv_bi[2-True-True-none-g1-idtype0] _______________

idtype = torch.int32, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'none', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
>           h_out = conv(g, (h, h_dst))

tests\pytorch\test_nn.py:167: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF011D81C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-4.1399e-01, -2.2294e+00],
        [-1.9270e-03, -2.0184e+00],
        [ 8.0488e-01,  2.3826e-01],
        [ 9.4166e-01,  1.3730e+00],
        [ 4.4610e-01,  2.0979e+00],
        [-5.2363e-01, -5.9245e-02]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv_bi[2-True-True-none-g1-idtype1] _______________

idtype = torch.int64, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'none', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
>           h_out = conv(g, (h, h_dst))

tests\pytorch\test_nn.py:167: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0775FE08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.6771e+00,  5.5007e-01],
        [-8.4047e-01, -5.5681e-01],
        [-1.1456e+00,  4.1625e-04],
        [-7.0751e-01, -2.2716e-01],
        [-2.3134e+00,  2.7290e+00],
        [ 2.6249e-01, -1.1318e+00]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv_bi[2-True-True-both-g0-idtype0] _______________

idtype = torch.int32
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'both', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
>           h_out = conv(g, (h, h_dst))

tests\pytorch\test_nn.py:167: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF030933C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.9775, -0.6751],
        [-0.9400, -0.4242]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv_bi[2-True-True-both-g0-idtype1] _______________

idtype = torch.int64
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'both', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
>           h_out = conv(g, (h, h_dst))

tests\pytorch\test_nn.py:167: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03EA9D08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.1069, -0.4628],
        [ 0.2631, -0.8641]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv_bi[2-True-True-both-g1-idtype0] _______________

idtype = torch.int32, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'both', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
>           h_out = conv(g, (h, h_dst))

tests\pytorch\test_nn.py:167: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0165A9C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.1465, -1.9445],
        [-0.4044,  1.4554],
        [ 0.0235,  0.4235],
        [ 1.2292,  1.6411],
        [ 3.4319, -0.4370],
        [-0.0541,  0.7545]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________ test_graph_conv_bi[2-True-True-both-g1-idtype1] _______________

idtype = torch.int64, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'both', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
>           h_out = conv(g, (h, h_dst))

tests\pytorch\test_nn.py:167: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0107A548>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.8098, -1.7640],
        [-0.1075,  0.7422],
        [ 0.1020, -0.6448],
        [-1.3473, -2.6438],
        [ 0.4806,  1.2444],
        [-0.8714, -1.7638]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________ test_graph_conv_bi[2-True-True-right-g0-idtype0] _______________

idtype = torch.int32
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'right', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
>           h_out = conv(g, (h, h_dst))

tests\pytorch\test_nn.py:167: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF030AFB08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.8104, -1.8433],
        [-1.2139,  0.8408]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________ test_graph_conv_bi[2-True-True-right-g0-idtype1] _______________

idtype = torch.int64
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'right', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
>           h_out = conv(g, (h, h_dst))

tests\pytorch\test_nn.py:167: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE149D4248>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.8547, -1.9630],
        [-1.2771,  1.1714]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________ test_graph_conv_bi[2-True-True-right-g1-idtype0] _______________

idtype = torch.int32, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'right', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
>           h_out = conv(g, (h, h_dst))

tests\pytorch\test_nn.py:167: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01586E08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-2.2864, -1.4859],
        [-0.8474, -1.1991],
        [ 0.9095, -0.7096],
        [-0.6140,  0.0676],
        [-0.6512,  1.6396],
        [ 0.9444, -0.6784]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________ test_graph_conv_bi[2-True-True-right-g1-idtype1] _______________

idtype = torch.int64, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'right', weight = True, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
>           h_out = conv(g, (h, h_dst))

tests\pytorch\test_nn.py:167: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE146ABC88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.3514,  1.9938],
        [-0.1824, -0.8924],
        [ 0.5439, -0.6511],
        [ 2.4835, -2.1830],
        [ 0.5989,  1.1135],
        [-2.7976,  2.6334]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________ test_graph_conv_bi[2-True-False-none-g0-idtype0] _______________

idtype = torch.int32
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'none', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
            h_out = conv(g, (h, h_dst))
        else:
>           h_out = conv(g, (h, h_dst), weight=ext_w)

tests\pytorch\test_nn.py:169: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0162BCC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.0237, -1.8177],
        [ 4.5065, -0.8349]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________ test_graph_conv_bi[2-True-False-none-g0-idtype1] _______________

idtype = torch.int64
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'none', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
            h_out = conv(g, (h, h_dst))
        else:
>           h_out = conv(g, (h, h_dst), weight=ext_w)

tests\pytorch\test_nn.py:169: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0164B288>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.6596, -0.2684],
        [ 2.0552,  0.7941]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________ test_graph_conv_bi[2-True-False-none-g1-idtype0] _______________

idtype = torch.int32, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'none', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
            h_out = conv(g, (h, h_dst))
        else:
>           h_out = conv(g, (h, h_dst), weight=ext_w)

tests\pytorch\test_nn.py:169: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01088388>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.2156,  3.3238],
        [ 0.2938, -2.3145],
        [ 0.7559, -1.2751],
        [-1.2290,  2.2463],
        [ 0.0164,  1.7052],
        [-1.9956,  4.5691]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________ test_graph_conv_bi[2-True-False-none-g1-idtype1] _______________

idtype = torch.int64, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'none', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
            h_out = conv(g, (h, h_dst))
        else:
>           h_out = conv(g, (h, h_dst), weight=ext_w)

tests\pytorch\test_nn.py:169: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF014A4588>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.6501,  2.1923],
        [-5.3344, -4.8047],
        [-1.6959,  0.7860],
        [ 2.6615,  2.1679],
        [-3.0327, -2.2036],
        [-0.1518, -0.4141]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________ test_graph_conv_bi[2-True-False-both-g0-idtype0] _______________

idtype = torch.int32
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'both', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
            h_out = conv(g, (h, h_dst))
        else:
>           h_out = conv(g, (h, h_dst), weight=ext_w)

tests\pytorch\test_nn.py:169: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0109C108>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.0053,  0.6787],
        [-0.7160, -0.4141]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________ test_graph_conv_bi[2-True-False-both-g0-idtype1] _______________

idtype = torch.int64
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'both', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
            h_out = conv(g, (h, h_dst))
        else:
>           h_out = conv(g, (h, h_dst), weight=ext_w)

tests\pytorch\test_nn.py:169: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF02F77CC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.1294, -1.4264],
        [-0.4919, -1.0274]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________ test_graph_conv_bi[2-True-False-both-g1-idtype0] _______________

idtype = torch.int32, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'both', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
            h_out = conv(g, (h, h_dst))
        else:
>           h_out = conv(g, (h, h_dst), weight=ext_w)

tests\pytorch\test_nn.py:169: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE149D2448>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-2.0916,  0.8384],
        [-0.5047, -1.8220],
        [-2.7802,  0.1383],
        [-0.6142,  1.0951],
        [ 3.2490, -0.9822],
        [-2.6834,  0.9883]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________ test_graph_conv_bi[2-True-False-both-g1-idtype1] _______________

idtype = torch.int64, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'both', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
            h_out = conv(g, (h, h_dst))
        else:
>           h_out = conv(g, (h, h_dst), weight=ext_w)

tests\pytorch\test_nn.py:169: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14608088>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.5246, -1.2162],
        [-1.3252, -0.7198],
        [ 0.7371,  3.0659],
        [ 2.7491,  0.3246],
        [ 0.8689,  2.4123],
        [ 0.7081, -4.2143]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________ test_graph_conv_bi[2-True-False-right-g0-idtype0] ______________

idtype = torch.int32
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'right', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
            h_out = conv(g, (h, h_dst))
        else:
>           h_out = conv(g, (h, h_dst), weight=ext_w)

tests\pytorch\test_nn.py:169: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14A7A3C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.8338,  1.1383],
        [-1.9977, -2.8742]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________ test_graph_conv_bi[2-True-False-right-g0-idtype1] ______________

idtype = torch.int64
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'right', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
            h_out = conv(g, (h, h_dst))
        else:
>           h_out = conv(g, (h, h_dst), weight=ext_w)

tests\pytorch\test_nn.py:169: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01089388>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-3.0286, -1.9180],
        [ 1.3014,  2.1747]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________ test_graph_conv_bi[2-True-False-right-g1-idtype0] ______________

idtype = torch.int32, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'right', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
            h_out = conv(g, (h, h_dst))
        else:
>           h_out = conv(g, (h, h_dst), weight=ext_w)

tests\pytorch\test_nn.py:169: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0142C088>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.7682, -2.0738],
        [ 1.1994,  1.6682],
        [ 1.3137,  1.6035],
        [ 0.7510, -0.5819],
        [-0.4705, -1.7249],
        [-0.6571,  3.6041]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________ test_graph_conv_bi[2-True-False-right-g1-idtype1] ______________

idtype = torch.int64, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'right', weight = False, bias = True, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
            h_out = conv(g, (h, h_dst))
        else:
>           h_out = conv(g, (h, h_dst), weight=ext_w)

tests\pytorch\test_nn.py:169: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01075808>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 2.2966, -4.1993],
        [-0.2066, -2.5190],
        [ 0.9365,  3.0353],
        [-1.1778, -3.0554],
        [ 1.3121, -1.3380],
        [-0.0823,  1.8038]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________ test_graph_conv_bi[2-False-True-none-g0-idtype0] _______________

idtype = torch.int32
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'none', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
>           h_out = conv(g, (h, h_dst))

tests\pytorch\test_nn.py:167: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF02FB2288>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.8490, -1.5220],
        [-1.5899,  0.6255]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________ test_graph_conv_bi[2-False-True-none-g0-idtype1] _______________

idtype = torch.int64
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'none', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
>           h_out = conv(g, (h, h_dst))

tests\pytorch\test_nn.py:167: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1456FA08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0.8833, 0.1405],
        [0.7712, 0.7087]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________ test_graph_conv_bi[2-False-True-none-g1-idtype0] _______________

idtype = torch.int32, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'none', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
>           h_out = conv(g, (h, h_dst))

tests\pytorch\test_nn.py:167: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF011E0188>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.0339, -0.1905],
        [ 1.1707, -0.6298],
        [ 2.6241, -0.6557],
        [ 2.4237, -0.6214],
        [-2.5278,  1.8460],
        [-1.5449,  0.9779]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________ test_graph_conv_bi[2-False-True-none-g1-idtype1] _______________

idtype = torch.int64, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'none', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
>           h_out = conv(g, (h, h_dst))

tests\pytorch\test_nn.py:167: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01731B88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.7071, -0.8390],
        [ 2.1976, -0.2198],
        [-1.6608, -0.5138],
        [ 0.5952,  0.6917],
        [-0.4227, -1.8006],
        [ 1.7586,  0.6120]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________ test_graph_conv_bi[2-False-True-both-g0-idtype0] _______________

idtype = torch.int32
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'both', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
>           h_out = conv(g, (h, h_dst))

tests\pytorch\test_nn.py:167: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01432788>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.4692,  0.9583],
        [-0.1725,  0.5508]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________ test_graph_conv_bi[2-False-True-both-g0-idtype1] _______________

idtype = torch.int64
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'both', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
>           h_out = conv(g, (h, h_dst))

tests\pytorch\test_nn.py:167: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF014F4CC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.1676, -1.1259],
        [ 0.3591,  0.7361]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________ test_graph_conv_bi[2-False-True-both-g1-idtype0] _______________

idtype = torch.int32, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'both', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
>           h_out = conv(g, (h, h_dst))

tests\pytorch\test_nn.py:167: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03083588>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.1212, -0.4925],
        [ 0.9728,  1.5921],
        [-1.6108,  2.4999],
        [-0.5508, -0.0093],
        [-0.2067, -0.4359],
        [ 1.1792,  0.4420]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________ test_graph_conv_bi[2-False-True-both-g1-idtype1] _______________

idtype = torch.int64, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'both', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
>           h_out = conv(g, (h, h_dst))

tests\pytorch\test_nn.py:167: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01598DC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.1557,  0.2599],
        [-0.2249, -0.3826],
        [-0.9367, -1.2368],
        [ 0.8575,  0.9160],
        [ 1.6676,  1.8103],
        [ 0.2975,  0.6746]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________ test_graph_conv_bi[2-False-True-right-g0-idtype0] ______________

idtype = torch.int32
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'right', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
>           h_out = conv(g, (h, h_dst))

tests\pytorch\test_nn.py:167: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0162B108>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.6120, -0.3789],
        [-0.2921,  2.8216]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________ test_graph_conv_bi[2-False-True-right-g0-idtype1] ______________

idtype = torch.int64
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'right', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
>           h_out = conv(g, (h, h_dst))

tests\pytorch\test_nn.py:167: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF010AB6C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.5314,  0.0304],
        [-0.5313, -2.2308]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________ test_graph_conv_bi[2-False-True-right-g1-idtype0] ______________

idtype = torch.int32, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'right', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
>           h_out = conv(g, (h, h_dst))

tests\pytorch\test_nn.py:167: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01563BC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-2.7557, -1.0420],
        [-1.6258, -2.3598],
        [ 0.0874,  0.9857],
        [ 1.4396,  0.6630],
        [ 1.1299,  1.5311],
        [-0.2805,  1.1451]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________ test_graph_conv_bi[2-False-True-right-g1-idtype1] ______________

idtype = torch.int64, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'right', weight = True, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
>           h_out = conv(g, (h, h_dst))

tests\pytorch\test_nn.py:167: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14A54CC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.7852,  0.1437],
        [-0.0886, -0.9989],
        [ 0.9741, -0.3672],
        [-0.4484, -0.0274],
        [ 1.8418, -0.2340],
        [ 0.0712,  0.7711]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________ test_graph_conv_bi[2-False-False-none-g0-idtype0] ______________

idtype = torch.int32
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'none', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
            h_out = conv(g, (h, h_dst))
        else:
>           h_out = conv(g, (h, h_dst), weight=ext_w)

tests\pytorch\test_nn.py:169: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE149480C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-2.1033,  3.0279],
        [ 0.0574,  1.9923]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________ test_graph_conv_bi[2-False-False-none-g0-idtype1] ______________

idtype = torch.int64
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'none', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
            h_out = conv(g, (h, h_dst))
        else:
>           h_out = conv(g, (h, h_dst), weight=ext_w)

tests\pytorch\test_nn.py:169: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1462E588>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.5429, -1.5642],
        [ 3.7061,  1.0933]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________ test_graph_conv_bi[2-False-False-none-g1-idtype0] ______________

idtype = torch.int32, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'none', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
            h_out = conv(g, (h, h_dst))
        else:
>           h_out = conv(g, (h, h_dst), weight=ext_w)

tests\pytorch\test_nn.py:169: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF011E1AC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.6038,  1.2058],
        [-0.3917,  1.0704],
        [ 0.5374,  0.2239],
        [-0.1427, -1.2547],
        [ 0.8091, -1.2874],
        [-1.3453,  0.2471]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________ test_graph_conv_bi[2-False-False-none-g1-idtype1] ______________

idtype = torch.int64, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'none', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
            h_out = conv(g, (h, h_dst))
        else:
>           h_out = conv(g, (h, h_dst), weight=ext_w)

tests\pytorch\test_nn.py:169: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF015BF508>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.0483,  1.4158],
        [ 2.3433, -0.7039],
        [ 0.2960,  0.2901],
        [-0.0928,  0.7765],
        [-0.1737,  0.3759],
        [ 1.3935, -1.5423]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________ test_graph_conv_bi[2-False-False-both-g0-idtype0] ______________

idtype = torch.int32
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'both', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
            h_out = conv(g, (h, h_dst))
        else:
>           h_out = conv(g, (h, h_dst), weight=ext_w)

tests\pytorch\test_nn.py:169: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0176C788>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.0818,  0.3127],
        [-0.8253, -1.2334]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________ test_graph_conv_bi[2-False-False-both-g0-idtype1] ______________

idtype = torch.int64
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'both', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
            h_out = conv(g, (h, h_dst))
        else:
>           h_out = conv(g, (h, h_dst), weight=ext_w)

tests\pytorch\test_nn.py:169: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14920EC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.2622, -0.6107],
        [-0.8719, -0.7810]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________ test_graph_conv_bi[2-False-False-both-g1-idtype0] ______________

idtype = torch.int32, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'both', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
            h_out = conv(g, (h, h_dst))
        else:
>           h_out = conv(g, (h, h_dst), weight=ext_w)

tests\pytorch\test_nn.py:169: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03077148>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-2.8052, -0.4997],
        [ 4.9514,  0.7642],
        [-2.9086, -0.9738],
        [ 1.2751, -2.7894],
        [-1.1706, -0.5089],
        [-5.1705, -5.3062]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________ test_graph_conv_bi[2-False-False-both-g1-idtype1] ______________

idtype = torch.int64, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'both', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
            h_out = conv(g, (h, h_dst))
        else:
>           h_out = conv(g, (h, h_dst), weight=ext_w)

tests\pytorch\test_nn.py:169: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF015AAA08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.2583,  1.7689],
        [ 2.7066, -4.8905],
        [-1.5381,  2.0689],
        [-1.6922,  0.7714],
        [-1.7100,  0.7687],
        [-0.8533,  0.5723]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_____________ test_graph_conv_bi[2-False-False-right-g0-idtype0] ______________

idtype = torch.int32
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'right', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
            h_out = conv(g, (h, h_dst))
        else:
>           h_out = conv(g, (h, h_dst), weight=ext_w)

tests\pytorch\test_nn.py:169: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all

python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE145DCDC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 4.5708, -5.4147],
        [-0.9800,  0.9646]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_____________ test_graph_conv_bi[2-False-False-right-g0-idtype1] ______________

idtype = torch.int64
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
norm = 'right', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
            h_out = conv(g, (h, h_dst))
        else:
>           h_out = conv(g, (h, h_dst), weight=ext_w)

tests\pytorch\test_nn.py:169: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF015A1A48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.4051,  1.3072],
        [-0.8023, -0.3428]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_____________ test_graph_conv_bi[2-False-False-right-g1-idtype0] ______________

idtype = torch.int32, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'right', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
            h_out = conv(g, (h, h_dst))
        else:
>           h_out = conv(g, (h, h_dst), weight=ext_w)

tests\pytorch\test_nn.py:169: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14AA2B08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.3338, -1.4458],
        [ 2.0507, -1.7456],
        [ 3.2306, -4.2193],
        [ 0.2923,  0.9065],
        [-3.6755, -2.4477],
        [-0.4563, -1.0939]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_____________ test_graph_conv_bi[2-False-False-right-g1-idtype1] ______________

idtype = torch.int64, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
norm = 'right', weight = False, bias = False, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree', 'dglgraph']))
    @pytest.mark.parametrize('norm', ['none', 'both', 'right'])
    @pytest.mark.parametrize('weight', [True, False])
    @pytest.mark.parametrize('bias', [True, False])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_graph_conv_bi(idtype, g, norm, weight, bias, out_dim):
        # Test a pair of tensor inputs
        g = g.astype(idtype).to(F.ctx())
        conv = nn.GraphConv(5, out_dim, norm=norm, weight=weight, bias=bias).to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        ext_w = F.randn((5, out_dim)).to(F.ctx())
        nsrc = g.number_of_src_nodes()
        ndst = g.number_of_dst_nodes()
        h = F.randn((nsrc, 5)).to(F.ctx())
        h_dst = F.randn((ndst, out_dim)).to(F.ctx())
        if weight:
            h_out = conv(g, (h, h_dst))
        else:
>           h_out = conv(g, (h, h_dst), weight=ext_w)

tests\pytorch\test_nn.py:169: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01630688>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.7275, -0.8995],
        [-1.5300, -1.6453],
        [ 0.5779,  1.8393],
        [ 0.2549, -0.2081],
        [-0.1928, -1.0716],
        [-0.5187,  0.1314]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________________ test_tagconv[1] _______________________________

out_dim = 1

    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_tagconv(out_dim):
        g = dgl.DGLGraph(nx.path_graph(3))
        g = g.to(F.ctx())
        ctx = F.ctx()
        adj = g.adjacency_matrix(transpose=True, ctx=ctx)
        norm = th.pow(g.in_degrees().float(), -0.5)
    
        conv = nn.TAGConv(5, out_dim, bias=True)
        conv = conv.to(ctx)
        print(conv)
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        # test#1: basic
        h0 = F.ones((3, 5))
>       h1 = conv(g, h0)

tests\pytorch\test_nn.py:201: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\tagconv.py:138: in forward
    fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF076CF748>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[1.0000, 1.0000, 1.0000, 1.0000, 1.0000],
        [0.7071, 0.7071, 0.7071, 0.7071, 0.7071],
        [1.0000, 1.0000, 1.0000, 1.0000, 1.0000]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
---------------------------- Captured stdout call -----------------------------
TAGConv(
  (lin): Linear(in_features=15, out_features=1, bias=True)
)
_______________________________ test_tagconv[2] _______________________________

out_dim = 2

    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_tagconv(out_dim):
        g = dgl.DGLGraph(nx.path_graph(3))
        g = g.to(F.ctx())
        ctx = F.ctx()
        adj = g.adjacency_matrix(transpose=True, ctx=ctx)
        norm = th.pow(g.in_degrees().float(), -0.5)
    
        conv = nn.TAGConv(5, out_dim, bias=True)
        conv = conv.to(ctx)
        print(conv)
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        # test#1: basic
        h0 = F.ones((3, 5))
>       h1 = conv(g, h0)

tests\pytorch\test_nn.py:201: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\tagconv.py:138: in forward
    fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF011DC2C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[1.0000, 1.0000, 1.0000, 1.0000, 1.0000],
        [0.7071, 0.7071, 0.7071, 0.7071, 0.7071],
        [1.0000, 1.0000, 1.0000, 1.0000, 1.0000]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
---------------------------- Captured stdout call -----------------------------
TAGConv(
  (lin): Linear(in_features=15, out_features=2, bias=True)
)
________________________________ test_set2set _________________________________

    def test_set2set():
        ctx = F.ctx()
        g = dgl.DGLGraph(nx.path_graph(10))
        g = g.to(F.ctx())
    
        s2s = nn.Set2Set(5, 3, 3) # hidden size 5, 3 iters, 3 layers
        s2s = s2s.to(ctx)
        print(s2s)
    
        # test#1: basic
        h0 = F.randn((g.number_of_nodes(), 5))
>       h1 = s2s(g, h0)

tests\pytorch\test_nn.py:234: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\glob.py:587: in forward
    alpha = softmax_nodes(graph, 'e')
python\dgl\readout.py:284: in softmax_nodes
    return segment.segment_softmax(graph.batch_num_nodes(ntype), x)
python\dgl\ops\segment.py:98: in segment_softmax
    value_max = segment_reduce(seglen, value, reducer='max')
python\dgl\ops\segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

op = 'max'
x = tensor([[-0.3325],
        [ 0.3624],
        [-0.1815],
        [-0.0951],
        [-0.3081],
        [-0.4183],
        [ 0.2572],
        [-0.0201],
        [-0.0996],
        [ 0.3648]], grad_fn=<SumBackward1>)
offsets = tensor([ 0, 10])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:780: TypeError
---------------------------- Captured stdout call -----------------------------
Set2Set(
  n_iters=3
  (lstm): LSTM(10, 5, num_layers=3)
)
_____________________________ test_glob_att_pool ______________________________

    def test_glob_att_pool():
        ctx = F.ctx()
        g = dgl.DGLGraph(nx.path_graph(10))
        g = g.to(F.ctx())
    
        gap = nn.GlobalAttentionPooling(th.nn.Linear(5, 1), th.nn.Linear(5, 10))
        gap = gap.to(ctx)
        print(gap)
    
        # test pickle
        th.save(gap, tmp_buffer)
    
        # test#1: basic
        h0 = F.randn((g.number_of_nodes(), 5))
>       h1 = gap(g, h0)

tests\pytorch\test_nn.py:259: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\glob.py:452: in forward
    gate = softmax_nodes(graph, 'gate')
python\dgl\readout.py:284: in softmax_nodes
    return segment.segment_softmax(graph.batch_num_nodes(ntype), x)
python\dgl\ops\segment.py:98: in segment_softmax
    value_max = segment_reduce(seglen, value, reducer='max')
python\dgl\ops\segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

op = 'max'
x = tensor([[0.5741],
        [0.0933],
        [0.2595],
        [0.6300],
        [0.8454],
        [0.2226],
        [0.3971],
        [0.1155],
        [0.0340],
        [0.3834]], grad_fn=<AddmmBackward>)
offsets = tensor([ 0, 10])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:780: TypeError
---------------------------- Captured stdout call -----------------------------
GlobalAttentionPooling(
  (gate_nn): Linear(in_features=5, out_features=1, bias=True)
  (feat_nn): Linear(in_features=5, out_features=10, bias=True)
)
______________________________ test_simple_pool _______________________________

    def test_simple_pool():
        ctx = F.ctx()
        g = dgl.DGLGraph(nx.path_graph(15))
        g = g.to(F.ctx())
    
        sum_pool = nn.SumPooling()
        avg_pool = nn.AvgPooling()
        max_pool = nn.MaxPooling()
        sort_pool = nn.SortPooling(10) # k = 10
        print(sum_pool, avg_pool, max_pool, sort_pool)
    
        # test#1: basic
        h0 = F.randn((g.number_of_nodes(), 5))
        sum_pool = sum_pool.to(ctx)
        avg_pool = avg_pool.to(ctx)
        max_pool = max_pool.to(ctx)
        sort_pool = sort_pool.to(ctx)
>       h1 = sum_pool(g, h0)

tests\pytorch\test_nn.py:285: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\glob.py:94: in forward
    readout = sum_nodes(graph, 'h')
python\dgl\readout.py:180: in sum_nodes
    return readout_nodes(graph, feat, weight, ntype=ntype, op='sum')
python\dgl\readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python\dgl\ops\segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

op = 'sum'
x = tensor([[-0.1293, -2.0777, -0.0795, -0.9244,  0.4134],
        [ 0.9438, -0.3004,  1.1719,  0.0088,  0.7673],
        ...0.3537],
        [ 0.2018,  0.1386,  0.7063, -0.7236, -0.6151],
        [ 0.3169,  0.4938, -1.0419, -0.9172,  0.1263]])
offsets = tensor([ 0, 15])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:780: TypeError
---------------------------- Captured stdout call -----------------------------
SumPooling() AvgPooling() MaxPooling() SortPooling()
____________________________ test_rgcn[1-idtype0] _____________________________

idtype = torch.int32, O = 1

    @parametrize_idtype
    @pytest.mark.parametrize('O', [1, 8, 32])
    def test_rgcn(idtype, O):
        ctx = F.ctx()
        etype = []
        g = dgl.from_scipy(sp.sparse.random(100, 100, density=0.1))
        g = g.astype(idtype).to(F.ctx())
        # 5 etypes
        R = 5
        for i in range(g.number_of_edges()):
            etype.append(i % 5)
        B = 2
        I = 10
    
        h = th.randn((100, I)).to(ctx)
        r = th.tensor(etype).to(ctx)
        norm = th.rand((g.number_of_edges(), 1)).to(ctx)
        sorted_r, idx = th.sort(r)
        sorted_g = dgl.reorder_graph(g, edge_permute_algo='custom', permute_config={'edges_perm' : idx.to(idtype)})
        sorted_norm = norm[idx]
    
        rgc = nn.RelGraphConv(I, O, R).to(ctx)
        th.save(rgc, tmp_buffer)  # test pickle
        rgc_basis = nn.RelGraphConv(I, O, R, "basis", B).to(ctx)
        th.save(rgc_basis, tmp_buffer)  # test pickle
        if O % B == 0:
            rgc_bdd = nn.RelGraphConv(I, O, R, "bdd", B).to(ctx)
            th.save(rgc_bdd, tmp_buffer)  # test pickle
    
        # basic usage
>       h_new = rgc(g, h, r)

tests\pytorch\test_nn.py:389: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\relgraphconv.py:173: in forward
    g.update_all(self.message, fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:369: in message_passing
    ndata = invoke_gspmm(g, fn.copy_e(msg, msg), rfunc, edata=msgdata)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0179D108>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([[-1.2240],
        [-0.6973],
        [ 0.1714],
        [ 0.7233],
        [ 0.7863],
        [ 0.0979],
    ... [ 0.0148],
        [ 0.5405],
        [ 0.1714],
        [-0.7942],
        [ 0.6637]], grad_fn=<IndexSelectBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
____________________________ test_rgcn[1-idtype1] _____________________________

idtype = torch.int64, O = 1

    @parametrize_idtype
    @pytest.mark.parametrize('O', [1, 8, 32])
    def test_rgcn(idtype, O):
        ctx = F.ctx()
        etype = []
        g = dgl.from_scipy(sp.sparse.random(100, 100, density=0.1))
        g = g.astype(idtype).to(F.ctx())
        # 5 etypes
        R = 5
        for i in range(g.number_of_edges()):
            etype.append(i % 5)
        B = 2
        I = 10
    
        h = th.randn((100, I)).to(ctx)
        r = th.tensor(etype).to(ctx)
        norm = th.rand((g.number_of_edges(), 1)).to(ctx)
        sorted_r, idx = th.sort(r)
        sorted_g = dgl.reorder_graph(g, edge_permute_algo='custom', permute_config={'edges_perm' : idx.to(idtype)})
        sorted_norm = norm[idx]
    
        rgc = nn.RelGraphConv(I, O, R).to(ctx)
        th.save(rgc, tmp_buffer)  # test pickle
        rgc_basis = nn.RelGraphConv(I, O, R, "basis", B).to(ctx)
        th.save(rgc_basis, tmp_buffer)  # test pickle
        if O % B == 0:
            rgc_bdd = nn.RelGraphConv(I, O, R, "bdd", B).to(ctx)
            th.save(rgc_bdd, tmp_buffer)  # test pickle
    
        # basic usage
>       h_new = rgc(g, h, r)

tests\pytorch\test_nn.py:389: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\relgraphconv.py:173: in forward
    g.update_all(self.message, fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:369: in message_passing
    ndata = invoke_gspmm(g, fn.copy_e(msg, msg), rfunc, edata=msgdata)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE145EF2C8>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([[ 0.0338],
        [ 0.4640],
        [-0.8780],
        [ 0.0303],
        [ 0.6427],
        [-0.2076],
    ... [-0.1495],
        [-1.0149],
        [ 0.4825],
        [ 0.1272],
        [ 0.0723]], grad_fn=<IndexSelectBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
____________________________ test_rgcn[8-idtype0] _____________________________

idtype = torch.int32, O = 8

    @parametrize_idtype
    @pytest.mark.parametrize('O', [1, 8, 32])
    def test_rgcn(idtype, O):
        ctx = F.ctx()
        etype = []
        g = dgl.from_scipy(sp.sparse.random(100, 100, density=0.1))
        g = g.astype(idtype).to(F.ctx())
        # 5 etypes
        R = 5
        for i in range(g.number_of_edges()):
            etype.append(i % 5)
        B = 2
        I = 10
    
        h = th.randn((100, I)).to(ctx)
        r = th.tensor(etype).to(ctx)
        norm = th.rand((g.number_of_edges(), 1)).to(ctx)
        sorted_r, idx = th.sort(r)
        sorted_g = dgl.reorder_graph(g, edge_permute_algo='custom', permute_config={'edges_perm' : idx.to(idtype)})
        sorted_norm = norm[idx]
    
        rgc = nn.RelGraphConv(I, O, R).to(ctx)
        th.save(rgc, tmp_buffer)  # test pickle
        rgc_basis = nn.RelGraphConv(I, O, R, "basis", B).to(ctx)
        th.save(rgc_basis, tmp_buffer)  # test pickle
        if O % B == 0:
            rgc_bdd = nn.RelGraphConv(I, O, R, "bdd", B).to(ctx)
            th.save(rgc_bdd, tmp_buffer)  # test pickle
    
        # basic usage
>       h_new = rgc(g, h, r)

tests\pytorch\test_nn.py:389: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\relgraphconv.py:173: in forward
    g.update_all(self.message, fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:369: in message_passing
    ndata = invoke_gspmm(g, fn.copy_e(msg, msg), rfunc, edata=msgdata)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01650C48>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([[-0.5004,  0.5150,  0.0837,  ..., -0.2301, -0.5909, -0.4255],
        [ 0.4280,  0.5575,  0.4563,  ..., -0.318...  1.2758],
        [ 0.0272,  0.8308, -0.1578,  ...,  0.8969, -0.2559, -0.7068]],
       grad_fn=<IndexSelectBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
____________________________ test_rgcn[8-idtype1] _____________________________

idtype = torch.int64, O = 8

    @parametrize_idtype
    @pytest.mark.parametrize('O', [1, 8, 32])
    def test_rgcn(idtype, O):
        ctx = F.ctx()
        etype = []
        g = dgl.from_scipy(sp.sparse.random(100, 100, density=0.1))
        g = g.astype(idtype).to(F.ctx())
        # 5 etypes
        R = 5
        for i in range(g.number_of_edges()):
            etype.append(i % 5)
        B = 2
        I = 10
    
        h = th.randn((100, I)).to(ctx)
        r = th.tensor(etype).to(ctx)
        norm = th.rand((g.number_of_edges(), 1)).to(ctx)
        sorted_r, idx = th.sort(r)
        sorted_g = dgl.reorder_graph(g, edge_permute_algo='custom', permute_config={'edges_perm' : idx.to(idtype)})
        sorted_norm = norm[idx]
    
        rgc = nn.RelGraphConv(I, O, R).to(ctx)
        th.save(rgc, tmp_buffer)  # test pickle
        rgc_basis = nn.RelGraphConv(I, O, R, "basis", B).to(ctx)
        th.save(rgc_basis, tmp_buffer)  # test pickle
        if O % B == 0:
            rgc_bdd = nn.RelGraphConv(I, O, R, "bdd", B).to(ctx)
            th.save(rgc_bdd, tmp_buffer)  # test pickle
    
        # basic usage
>       h_new = rgc(g, h, r)

tests\pytorch\test_nn.py:389: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\relgraphconv.py:173: in forward
    g.update_all(self.message, fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:369: in message_passing
    ndata = invoke_gspmm(g, fn.copy_e(msg, msg), rfunc, edata=msgdata)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF011D0A48>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([[-0.4904, -0.4048,  0.1054,  ...,  0.8263, -0.0182, -0.3286],
        [ 0.1295,  0.1673,  0.3674,  ...,  0.168...  0.0509],
        [ 0.1634, -0.2912, -0.2321,  ...,  0.4551,  0.7200,  0.0128]],
       grad_fn=<IndexSelectBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
____________________________ test_rgcn[32-idtype0] ____________________________

idtype = torch.int32, O = 32

    @parametrize_idtype
    @pytest.mark.parametrize('O', [1, 8, 32])
    def test_rgcn(idtype, O):
        ctx = F.ctx()
        etype = []
        g = dgl.from_scipy(sp.sparse.random(100, 100, density=0.1))
        g = g.astype(idtype).to(F.ctx())
        # 5 etypes
        R = 5
        for i in range(g.number_of_edges()):
            etype.append(i % 5)
        B = 2
        I = 10
    
        h = th.randn((100, I)).to(ctx)
        r = th.tensor(etype).to(ctx)
        norm = th.rand((g.number_of_edges(), 1)).to(ctx)
        sorted_r, idx = th.sort(r)
        sorted_g = dgl.reorder_graph(g, edge_permute_algo='custom', permute_config={'edges_perm' : idx.to(idtype)})
        sorted_norm = norm[idx]
    
        rgc = nn.RelGraphConv(I, O, R).to(ctx)
        th.save(rgc, tmp_buffer)  # test pickle
        rgc_basis = nn.RelGraphConv(I, O, R, "basis", B).to(ctx)
        th.save(rgc_basis, tmp_buffer)  # test pickle
        if O % B == 0:
            rgc_bdd = nn.RelGraphConv(I, O, R, "bdd", B).to(ctx)
            th.save(rgc_bdd, tmp_buffer)  # test pickle
    
        # basic usage
>       h_new = rgc(g, h, r)

tests\pytorch\test_nn.py:389: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\relgraphconv.py:173: in forward
    g.update_all(self.message, fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:369: in message_passing
    ndata = invoke_gspmm(g, fn.copy_e(msg, msg), rfunc, edata=msgdata)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF015FC2C8>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([[-0.5956, -1.5377, -0.9770,  ..., -0.3251,  0.6088, -0.7434],
        [-0.5486, -0.7755, -0.4398,  ..., -0.234... -0.4531],
        [-0.4102, -0.0849,  0.3616,  ...,  0.0827,  0.4221, -0.3024]],
       grad_fn=<IndexSelectBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
____________________________ test_rgcn[32-idtype1] ____________________________

idtype = torch.int64, O = 32

    @parametrize_idtype
    @pytest.mark.parametrize('O', [1, 8, 32])
    def test_rgcn(idtype, O):
        ctx = F.ctx()
        etype = []
        g = dgl.from_scipy(sp.sparse.random(100, 100, density=0.1))
        g = g.astype(idtype).to(F.ctx())
        # 5 etypes
        R = 5
        for i in range(g.number_of_edges()):
            etype.append(i % 5)
        B = 2
        I = 10
    
        h = th.randn((100, I)).to(ctx)
        r = th.tensor(etype).to(ctx)
        norm = th.rand((g.number_of_edges(), 1)).to(ctx)
        sorted_r, idx = th.sort(r)
        sorted_g = dgl.reorder_graph(g, edge_permute_algo='custom', permute_config={'edges_perm' : idx.to(idtype)})
        sorted_norm = norm[idx]
    
        rgc = nn.RelGraphConv(I, O, R).to(ctx)
        th.save(rgc, tmp_buffer)  # test pickle
        rgc_basis = nn.RelGraphConv(I, O, R, "basis", B).to(ctx)
        th.save(rgc_basis, tmp_buffer)  # test pickle
        if O % B == 0:
            rgc_bdd = nn.RelGraphConv(I, O, R, "bdd", B).to(ctx)
            th.save(rgc_bdd, tmp_buffer)  # test pickle
    
        # basic usage
>       h_new = rgc(g, h, r)

tests\pytorch\test_nn.py:389: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\relgraphconv.py:173: in forward
    g.update_all(self.message, fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:369: in message_passing
    ndata = invoke_gspmm(g, fn.copy_e(msg, msg), rfunc, edata=msgdata)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1469E188>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([[-0.5183, -0.7084,  0.7880,  ...,  0.0194,  0.1953, -0.7940],
        [ 0.3973, -0.3417,  0.0471,  ...,  0.442...  0.5010],
        [ 0.7327, -0.2675, -0.8149,  ..., -0.9808, -0.4003, -0.1540]],
       grad_fn=<IndexSelectBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_____________________ test_rgcn_default_nbasis[1-idtype0] _____________________

idtype = torch.int32, O = 1

    @parametrize_idtype
    @pytest.mark.parametrize('O', [1, 10, 40])
    def test_rgcn_default_nbasis(idtype, O):
        ctx = F.ctx()
        etype = []
        g = dgl.from_scipy(sp.sparse.random(100, 100, density=0.1))
        g = g.astype(idtype).to(F.ctx())
        # 5 etypes
        R = 5
        for i in range(g.number_of_edges()):
            etype.append(i % 5)
        I = 10
    
        h = th.randn((100, I)).to(ctx)
        r = th.tensor(etype).to(ctx)
        norm = th.rand((g.number_of_edges(), 1)).to(ctx)
        sorted_r, idx = th.sort(r)
        sorted_g = dgl.reorder_graph(g, edge_permute_algo='custom', permute_config={'edges_perm' : idx.to(idtype)})
        sorted_norm = norm[idx]
    
        rgc = nn.RelGraphConv(I, O, R).to(ctx)
        th.save(rgc, tmp_buffer)  # test pickle
        rgc_basis = nn.RelGraphConv(I, O, R, "basis").to(ctx)
        th.save(rgc_basis, tmp_buffer)  # test pickle
        if O % R == 0:
            rgc_bdd = nn.RelGraphConv(I, O, R, "bdd").to(ctx)
            th.save(rgc_bdd, tmp_buffer)  # test pickle
    
        # basic usage
>       h_new = rgc(g, h, r)

tests\pytorch\test_nn.py:444: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\relgraphconv.py:173: in forward
    g.update_all(self.message, fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:369: in message_passing
    ndata = invoke_gspmm(g, fn.copy_e(msg, msg), rfunc, edata=msgdata)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03EEEC88>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([[ 0.1932],
        [-0.4442],
        [ 1.2024],
        [-0.4945],
        [-0.8663],
        [ 0.6029],
    ... [ 0.0407],
        [ 0.0574],
        [ 0.3677],
        [ 0.0793],
        [ 0.3107]], grad_fn=<IndexSelectBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_____________________ test_rgcn_default_nbasis[1-idtype1] _____________________

idtype = torch.int64, O = 1

    @parametrize_idtype
    @pytest.mark.parametrize('O', [1, 10, 40])
    def test_rgcn_default_nbasis(idtype, O):
        ctx = F.ctx()
        etype = []
        g = dgl.from_scipy(sp.sparse.random(100, 100, density=0.1))
        g = g.astype(idtype).to(F.ctx())
        # 5 etypes
        R = 5
        for i in range(g.number_of_edges()):
            etype.append(i % 5)
        I = 10
    
        h = th.randn((100, I)).to(ctx)
        r = th.tensor(etype).to(ctx)
        norm = th.rand((g.number_of_edges(), 1)).to(ctx)
        sorted_r, idx = th.sort(r)
        sorted_g = dgl.reorder_graph(g, edge_permute_algo='custom', permute_config={'edges_perm' : idx.to(idtype)})
        sorted_norm = norm[idx]
    
        rgc = nn.RelGraphConv(I, O, R).to(ctx)
        th.save(rgc, tmp_buffer)  # test pickle
        rgc_basis = nn.RelGraphConv(I, O, R, "basis").to(ctx)
        th.save(rgc_basis, tmp_buffer)  # test pickle
        if O % R == 0:
            rgc_bdd = nn.RelGraphConv(I, O, R, "bdd").to(ctx)
            th.save(rgc_bdd, tmp_buffer)  # test pickle
    
        # basic usage
>       h_new = rgc(g, h, r)

tests\pytorch\test_nn.py:444: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\relgraphconv.py:173: in forward
    g.update_all(self.message, fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:369: in message_passing
    ndata = invoke_gspmm(g, fn.copy_e(msg, msg), rfunc, edata=msgdata)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01585C48>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([[ 0.6487],
        [-0.3070],
        [-0.7505],
        [ 0.7301],
        [ 0.1716],
        [ 1.0842],
    ... [-0.6335],
        [-0.5028],
        [-0.4268],
        [-0.0053],
        [ 0.5330]], grad_fn=<IndexSelectBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
____________________ test_rgcn_default_nbasis[10-idtype0] _____________________

idtype = torch.int32, O = 10

    @parametrize_idtype
    @pytest.mark.parametrize('O', [1, 10, 40])
    def test_rgcn_default_nbasis(idtype, O):
        ctx = F.ctx()
        etype = []
        g = dgl.from_scipy(sp.sparse.random(100, 100, density=0.1))
        g = g.astype(idtype).to(F.ctx())
        # 5 etypes
        R = 5
        for i in range(g.number_of_edges()):
            etype.append(i % 5)
        I = 10
    
        h = th.randn((100, I)).to(ctx)
        r = th.tensor(etype).to(ctx)
        norm = th.rand((g.number_of_edges(), 1)).to(ctx)
        sorted_r, idx = th.sort(r)
        sorted_g = dgl.reorder_graph(g, edge_permute_algo='custom', permute_config={'edges_perm' : idx.to(idtype)})
        sorted_norm = norm[idx]
    
        rgc = nn.RelGraphConv(I, O, R).to(ctx)
        th.save(rgc, tmp_buffer)  # test pickle
        rgc_basis = nn.RelGraphConv(I, O, R, "basis").to(ctx)
        th.save(rgc_basis, tmp_buffer)  # test pickle
        if O % R == 0:
            rgc_bdd = nn.RelGraphConv(I, O, R, "bdd").to(ctx)
            th.save(rgc_bdd, tmp_buffer)  # test pickle
    
        # basic usage
>       h_new = rgc(g, h, r)

tests\pytorch\test_nn.py:444: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\relgraphconv.py:173: in forward
    g.update_all(self.message, fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:369: in message_passing
    ndata = invoke_gspmm(g, fn.copy_e(msg, msg), rfunc, edata=msgdata)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14AC8908>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([[-0.8781, -0.8904, -0.8871,  ...,  0.3654,  0.7610,  0.2420],
        [-0.8465,  0.0241, -0.5828,  ..., -0.308...  0.2724],
        [-0.6384, -0.3284,  0.7624,  ..., -1.0435,  0.3147,  0.2131]],
       grad_fn=<IndexSelectBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
____________________ test_rgcn_default_nbasis[10-idtype1] _____________________

idtype = torch.int64, O = 10

    @parametrize_idtype
    @pytest.mark.parametrize('O', [1, 10, 40])
    def test_rgcn_default_nbasis(idtype, O):
        ctx = F.ctx()
        etype = []
        g = dgl.from_scipy(sp.sparse.random(100, 100, density=0.1))
        g = g.astype(idtype).to(F.ctx())
        # 5 etypes
        R = 5
        for i in range(g.number_of_edges()):
            etype.append(i % 5)
        I = 10
    
        h = th.randn((100, I)).to(ctx)
        r = th.tensor(etype).to(ctx)
        norm = th.rand((g.number_of_edges(), 1)).to(ctx)
        sorted_r, idx = th.sort(r)
        sorted_g = dgl.reorder_graph(g, edge_permute_algo='custom', permute_config={'edges_perm' : idx.to(idtype)})
        sorted_norm = norm[idx]
    
        rgc = nn.RelGraphConv(I, O, R).to(ctx)
        th.save(rgc, tmp_buffer)  # test pickle
        rgc_basis = nn.RelGraphConv(I, O, R, "basis").to(ctx)
        th.save(rgc_basis, tmp_buffer)  # test pickle
        if O % R == 0:
            rgc_bdd = nn.RelGraphConv(I, O, R, "bdd").to(ctx)
            th.save(rgc_bdd, tmp_buffer)  # test pickle
    
        # basic usage
>       h_new = rgc(g, h, r)

tests\pytorch\test_nn.py:444: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\relgraphconv.py:173: in forward
    g.update_all(self.message, fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:369: in message_passing
    ndata = invoke_gspmm(g, fn.copy_e(msg, msg), rfunc, edata=msgdata)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14580A88>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([[-1.0661,  1.2825, -0.6266,  ...,  0.3066, -0.9195, -0.7139],
        [ 0.5909,  0.1753,  0.1612,  ..., -0.780...  0.0345],
        [-0.8277,  0.4179, -1.4531,  ..., -1.2665,  1.1368, -0.1736]],
       grad_fn=<IndexSelectBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
____________________ test_rgcn_default_nbasis[40-idtype0] _____________________

idtype = torch.int32, O = 40

    @parametrize_idtype
    @pytest.mark.parametrize('O', [1, 10, 40])
    def test_rgcn_default_nbasis(idtype, O):
        ctx = F.ctx()
        etype = []
        g = dgl.from_scipy(sp.sparse.random(100, 100, density=0.1))
        g = g.astype(idtype).to(F.ctx())
        # 5 etypes
        R = 5
        for i in range(g.number_of_edges()):
            etype.append(i % 5)
        I = 10
    
        h = th.randn((100, I)).to(ctx)
        r = th.tensor(etype).to(ctx)
        norm = th.rand((g.number_of_edges(), 1)).to(ctx)
        sorted_r, idx = th.sort(r)
        sorted_g = dgl.reorder_graph(g, edge_permute_algo='custom', permute_config={'edges_perm' : idx.to(idtype)})
        sorted_norm = norm[idx]
    
        rgc = nn.RelGraphConv(I, O, R).to(ctx)
        th.save(rgc, tmp_buffer)  # test pickle
        rgc_basis = nn.RelGraphConv(I, O, R, "basis").to(ctx)
        th.save(rgc_basis, tmp_buffer)  # test pickle
        if O % R == 0:
            rgc_bdd = nn.RelGraphConv(I, O, R, "bdd").to(ctx)
            th.save(rgc_bdd, tmp_buffer)  # test pickle
    
        # basic usage
>       h_new = rgc(g, h, r)

tests\pytorch\test_nn.py:444: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\relgraphconv.py:173: in forward
    g.update_all(self.message, fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:369: in message_passing
    ndata = invoke_gspmm(g, fn.copy_e(msg, msg), rfunc, edata=msgdata)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14680E08>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([[ 0.7001, -0.2555,  0.3475,  ..., -0.1695,  0.9420, -0.3143],
        [-0.4761,  0.0753,  0.1837,  ..., -0.200...  0.7230],
        [-1.5536,  0.9215, -0.9472,  ...,  0.8587, -0.6231,  0.3237]],
       grad_fn=<IndexSelectBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
____________________ test_rgcn_default_nbasis[40-idtype1] _____________________

idtype = torch.int64, O = 40

    @parametrize_idtype
    @pytest.mark.parametrize('O', [1, 10, 40])
    def test_rgcn_default_nbasis(idtype, O):
        ctx = F.ctx()
        etype = []
        g = dgl.from_scipy(sp.sparse.random(100, 100, density=0.1))
        g = g.astype(idtype).to(F.ctx())
        # 5 etypes
        R = 5
        for i in range(g.number_of_edges()):
            etype.append(i % 5)
        I = 10
    
        h = th.randn((100, I)).to(ctx)
        r = th.tensor(etype).to(ctx)
        norm = th.rand((g.number_of_edges(), 1)).to(ctx)
        sorted_r, idx = th.sort(r)
        sorted_g = dgl.reorder_graph(g, edge_permute_algo='custom', permute_config={'edges_perm' : idx.to(idtype)})
        sorted_norm = norm[idx]
    
        rgc = nn.RelGraphConv(I, O, R).to(ctx)
        th.save(rgc, tmp_buffer)  # test pickle
        rgc_basis = nn.RelGraphConv(I, O, R, "basis").to(ctx)
        th.save(rgc_basis, tmp_buffer)  # test pickle
        if O % R == 0:
            rgc_bdd = nn.RelGraphConv(I, O, R, "bdd").to(ctx)
            th.save(rgc_bdd, tmp_buffer)  # test pickle
    
        # basic usage
>       h_new = rgc(g, h, r)

tests\pytorch\test_nn.py:444: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\relgraphconv.py:173: in forward
    g.update_all(self.message, fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:369: in message_passing
    ndata = invoke_gspmm(g, fn.copy_e(msg, msg), rfunc, edata=msgdata)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF014EAE48>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([[ 1.0039, -0.5955, -0.0490,  ...,  0.8440,  0.4723,  0.3927],
        [-0.3716, -0.6820, -0.5605,  ...,  0.048...  0.0618],
        [ 0.5527, -0.2110,  0.3849,  ...,  0.3952, -0.8336, -0.1499]],
       grad_fn=<IndexSelectBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________________ test_gat_conv[1-1-g0-idtype0] ________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'ft': Scheme(shap...), 'er': Scheme(shape=(1, 1), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
idtype = torch.int32, out_dim = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:480: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatconv.py:303: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF010F3C88>
op = 'add'
lhs_data = tensor([[[-0.1650]],

        [[-0.6172]],

        [[-0.1875]],

        [[ 0.6192]],

        [[ 0.3698]],

        ...

        [[ 0.8059]],

        [[-0.3322]],

        [[ 0.2771]],

        [[ 0.1082]]], grad_fn=<UnsqueezeBackward0>)
rhs_data = tensor([[[ 0.0200]],

        [[ 0.0749]],

        [[ 0.0228]],

        [[-0.0751]],

        [[-0.0449]],

        ...

        [[-0.0978]],

        [[ 0.0403]],

        [[-0.0336]],

        [[-0.0131]]], grad_fn=<UnsqueezeBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
________________________ test_gat_conv[1-1-g0-idtype1] ________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'ft': Scheme(shap...), 'er': Scheme(shape=(1, 1), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
idtype = torch.int64, out_dim = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:480: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatconv.py:303: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF011F0D48>
op = 'add'
lhs_data = tensor([[[ 2.4681]],

        [[-1.6144]],

        [[ 0.9395]],

        [[ 1.2816]],

        [[-0.3932]],

        ...

        [[ 0.1351]],

        [[-0.5771]],

        [[-2.7710]],

        [[ 1.8190]]], grad_fn=<UnsqueezeBackward0>)
rhs_data = tensor([[[-5.9841]],

        [[ 3.9142]],

        [[-2.2780]],

        [[-3.1072]],

        [[ 0.9534]],

        ...

        [[-0.3275]],

        [[ 1.3992]],

        [[ 6.7184]],

        [[-4.4103]]], grad_fn=<UnsqueezeBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
________________________ test_gat_conv[1-1-g1-idtype0] ________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'ft': Scheme(shape=(1, 1), dtype=torch.float32), 'el': Scheme(shape=(1, 1), dtype=torch.float32), 'er': Scheme(shape=(1, 1), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:480: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatconv.py:303: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01731688>
op = 'add'
lhs_data = tensor([[[ 0.3166]],

        [[-2.2701]],

        [[-1.8705]],

        [[ 0.1864]],

        [[-0.7116]],

        ...

        [[-3.4579]],

        [[-0.0101]],

        [[-1.4921]],

        [[-2.8930]]], grad_fn=<UnsqueezeBackward0>)
rhs_data = tensor([[[ 0.0792]],

        [[-0.5676]],

        [[-0.4677]],

        [[ 0.0466]],

        [[-0.1779]],

        ...

        [[-0.8646]],

        [[-0.0025]],

        [[-0.3731]],

        [[-0.7233]]], grad_fn=<UnsqueezeBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
________________________ test_gat_conv[1-1-g1-idtype1] ________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'ft': Scheme(shape=(1, 1), dtype=torch.float32), 'el': Scheme(shape=(1, 1), dtype=torch.float32), 'er': Scheme(shape=(1, 1), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:480: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatconv.py:303: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01539188>
op = 'add'
lhs_data = tensor([[[ 0.4859]],

        [[ 5.3679]],

        [[-1.4604]],

        [[-1.9768]],

        [[ 1.9826]],

        ...

        [[-4.8519]],

        [[-2.7468]],

        [[ 4.5642]],

        [[ 0.4012]]], grad_fn=<UnsqueezeBackward0>)
rhs_data = tensor([[[-0.3106]],

        [[-3.4314]],

        [[ 0.9335]],

        [[ 1.2636]],

        [[-1.2674]],

        ...

        [[ 3.1015]],

        [[ 1.7559]],

        [[-2.9176]],

        [[-0.2565]]], grad_fn=<UnsqueezeBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
________________________ test_gat_conv[1-1-g2-idtype0] ________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'ft': Scheme(shap...r': Scheme(shape=(1, 1), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int32, out_dim = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:480: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatconv.py:303: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01059108>
op = 'add'
lhs_data = tensor([[[ 0.4571]],

        [[ 0.1209]],

        [[-0.6261]],

        [[ 0.0500]],

        [[ 1.0809]],

        ...

        [[ 0.4801]],

        [[ 0.2702]],

        [[ 0.2558]],

        [[-0.2243]]], grad_fn=<UnsqueezeBackward0>)
rhs_data = tensor([[[-2.2373]],

        [[-0.5918]],

        [[ 3.0648]],

        [[-0.2447]],

        [[-5.2909]],

        ...

        [[-2.3500]],

        [[-1.3227]],

        [[-1.2522]],

        [[ 1.0981]]], grad_fn=<UnsqueezeBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
________________________ test_gat_conv[1-1-g2-idtype1] ________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'ft': Scheme(shap...r': Scheme(shape=(1, 1), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int64, out_dim = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:480: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatconv.py:303: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE145F7808>
op = 'add'
lhs_data = tensor([[[ 0.0070]],

        [[-0.5195]],

        [[ 1.7941]],

        [[-0.5468]],

        [[ 1.7042]],

        ...

        [[ 1.1612]],

        [[-1.2101]],

        [[-0.7298]],

        [[ 2.4521]]], grad_fn=<UnsqueezeBackward0>)
rhs_data = tensor([[[-0.0245]],

        [[ 1.8113]],

        [[-6.2548]],

        [[ 1.9064]],

        [[-5.9413]],

        ...

        [[-4.0482]],

        [[ 4.2187]],

        [[ 2.5444]],

        [[-8.5487]]], grad_fn=<UnsqueezeBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
________________________ test_gat_conv[1-1-g3-idtype0] ________________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'ft': Scheme(shape=(1, 1), dtype=torch.float32), 'el': Scheme(shape=(1, 1), dtype=torch.float32), 'er': Scheme(shape=(1, 1), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:480: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatconv.py:303: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01196B88>
op = 'add'
lhs_data = tensor([[[ 0.8349]],

        [[ 1.4136]],

        [[ 0.7201]],

        [[ 0.8119]],

        [[-2.4114]],

        [[-1.8802]],

        [[ 0.1493]],

        [[-0.7703]],

        [[-0.6132]]], grad_fn=<UnsqueezeBackward0>)
rhs_data = tensor([[[-0.9068]],

        [[-1.5353]],

        [[-0.7821]],

        [[-0.8818]],

        [[ 2.6191]],

        [[ 2.0421]],

        [[-0.1622]],

        [[ 0.8367]],

        [[ 0.6660]]], grad_fn=<UnsqueezeBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
________________________ test_gat_conv[1-1-g3-idtype1] ________________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'ft': Scheme(shape=(1, 1), dtype=torch.float32), 'el': Scheme(shape=(1, 1), dtype=torch.float32), 'er': Scheme(shape=(1, 1), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:480: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatconv.py:303: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0144FD08>
op = 'add'
lhs_data = tensor([[[ 0.7742]],

        [[ 0.0970]],

        [[ 0.8368]],

        [[ 1.0546]],

        [[ 0.2960]],

        [[ 0.6417]],

        [[ 0.2710]],

        [[-0.0930]],

        [[-0.2901]]], grad_fn=<UnsqueezeBackward0>)
rhs_data = tensor([[[-1.3556]],

        [[-0.1699]],

        [[-1.4652]],

        [[-1.8465]],

        [[-0.5183]],

        [[-1.1235]],

        [[-0.4745]],

        [[ 0.1628]],

        [[ 0.5079]]], grad_fn=<UnsqueezeBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
________________________ test_gat_conv[1-1-g4-idtype0] ________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'ft': Scheme(shape=(1, 1), dtype=torch.float32), 'el': Scheme(shape=(1, 1), dtype=torch.float32), 'er': Scheme(shape=(1, 1), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:480: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatconv.py:303: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03018988>
op = 'add'
lhs_data = tensor([[[-0.5293]],

        [[ 0.0380]],

        [[ 0.6663]],

        [[-0.7647]],

        [[ 0.4598]],

        ...

        [[-0.1009]],

        [[ 0.6384]],

        [[ 0.1764]],

        [[ 0.0025]]], grad_fn=<UnsqueezeBackward0>)
rhs_data = tensor([[[ -7.3919]],

        [[  0.5308]],

        [[  9.3042]],

        [[-10.6788]],

        [[  6.4209]],

   ...      [[ -1.4087]],

        [[  8.9148]],

        [[  2.4640]],

        [[  0.0354]]], grad_fn=<UnsqueezeBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
________________________ test_gat_conv[1-1-g4-idtype1] ________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'ft': Scheme(shape=(1, 1), dtype=torch.float32), 'el': Scheme(shape=(1, 1), dtype=torch.float32), 'er': Scheme(shape=(1, 1), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:480: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatconv.py:303: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE145A38C8>
op = 'add'
lhs_data = tensor([[[-2.1451]],

        [[-2.8286]],

        [[ 5.6773]],

        [[ 0.3103]],

        [[ 7.2222]],

        ...

        [[-3.6402]],

        [[ 0.4824]],

        [[ 0.0498]],

        [[ 2.0053]]], grad_fn=<UnsqueezeBackward0>)
rhs_data = tensor([[[-1.7711]],

        [[-2.3355]],

        [[ 4.6875]],

        [[ 0.2562]],

        [[ 5.9631]],

        ...

        [[-3.0055]],

        [[ 0.3983]],

        [[ 0.0411]],

        [[ 1.6557]]], grad_fn=<UnsqueezeBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
________________________ test_gat_conv[1-1-g5-idtype0] ________________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int32
out_dim = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:480: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatconv.py:303: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01487348>
op = 'add'
lhs_data = tensor([[[-0.6303]],

        [[-0.3682]],

        [[ 0.4352]],

        [[ 0.5459]],

        [[-1.2812]],

        [[ 0.6484]]], grad_fn=<UnsqueezeBackward0>)
rhs_data = tensor([[[-0.9593]],

        [[-0.5604]],

        [[ 0.6624]]], grad_fn=<UnsqueezeBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
________________________ test_gat_conv[1-1-g5-idtype1] ________________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int64
out_dim = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:480: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatconv.py:303: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF017ED9C8>
op = 'add'
lhs_data = tensor([[[ 2.5130]],

        [[ 0.3896]],

        [[ 2.3373]],

        [[ 3.8646]],

        [[-0.9286]],

        [[-0.0705]]], grad_fn=<UnsqueezeBackward0>)
rhs_data = tensor([[[-4.2474]],

        [[-0.6584]],

        [[-3.9504]]], grad_fn=<UnsqueezeBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
________________________ test_gat_conv[1-1-g6-idtype0] ________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'ft': Scheme(shape=(1, 1), dtype=torch.float32), 'el': Scheme(shape=(1, 1), dtype=torch.float32), 'er': Scheme(shape=(1, 1), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:480: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatconv.py:303: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07614A88>
op = 'add'
lhs_data = tensor([[[ 0.0098]],

        [[ 0.1097]],

        [[ 0.0095]],

        [[-0.0190]],

        [[ 0.0505]],

        ...

        [[-0.0564]],

        [[-0.0897]],

        [[ 0.1960]],

        [[-0.1524]]], grad_fn=<UnsqueezeBackward0>)
rhs_data = tensor([[[-0.1806]],

        [[-2.0224]],

        [[-0.1748]],

        [[ 0.3509]],

        [[-0.9307]],

        ...

        [[ 1.0401]],

        [[ 1.6534]],

        [[-3.6147]],

        [[ 2.8102]]], grad_fn=<UnsqueezeBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
________________________ test_gat_conv[1-1-g6-idtype1] ________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'ft': Scheme(shape=(1, 1), dtype=torch.float32), 'el': Scheme(shape=(1, 1), dtype=torch.float32), 'er': Scheme(shape=(1, 1), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:480: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatconv.py:303: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01851108>
op = 'add'
lhs_data = tensor([[[ 1.9403]],

        [[ 1.9842]],

        [[ 0.4303]],

        [[ 1.8369]],

        [[-1.9152]],

        ...

        [[ 0.2403]],

        [[-3.4911]],

        [[-0.6532]],

        [[-7.9902]]], grad_fn=<UnsqueezeBackward0>)
rhs_data = tensor([[[-0.5821]],

        [[-0.5952]],

        [[-0.1291]],

        [[-0.5510]],

        [[ 0.5745]],

        ...

        [[-0.0721]],

        [[ 1.0473]],

        [[ 0.1960]],

        [[ 2.3969]]], grad_fn=<UnsqueezeBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
________________________ test_gat_conv[1-5-g0-idtype0] ________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'ft': Scheme(shap...), 'er': Scheme(shape=(1, 1), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
idtype = torch.int32, out_dim = 5, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:480: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatconv.py:303: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01518988>
op = 'add'
lhs_data = tensor([[[ 2.0756]],

        [[-2.1026]],

        [[-1.7871]],

        [[ 3.6761]],

        [[ 1.7978]],

        ...

        [[ 1.4999]],

        [[-2.1914]],

        [[ 4.5788]],

        [[-2.2304]]], grad_fn=<UnsqueezeBackward0>)
rhs_data = tensor([[[ 0.3871]],

        [[ 0.0901]],

        [[-0.6526]],

        [[ 2.5571]],

        [[ 0.5008]],

        ...

        [[-0.3593]],

        [[-0.1909]],

        [[ 1.0692]],

        [[-0.0610]]], grad_fn=<UnsqueezeBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
________________________ test_gat_conv[1-5-g0-idtype1] ________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'ft': Scheme(shap...), 'er': Scheme(shape=(1, 1), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
idtype = torch.int64, out_dim = 5, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:480: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatconv.py:303: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0182E188>
op = 'add'
lhs_data = tensor([[[-1.1076]],

        [[ 1.5069]],

        [[ 0.2092]],

        [[ 1.3420]],

        [[ 0.0207]],

        ...

        [[-0.9363]],

        [[ 1.0478]],

        [[ 0.7604]],

        [[-1.2209]]], grad_fn=<UnsqueezeBackward0>)
rhs_data = tensor([[[ 0.5252]],

        [[-0.8738]],

        [[ 1.1782]],

        [[ 0.2439]],

        [[-0.0819]],

        ...

        [[ 0.6176]],

        [[-0.7654]],

        [[-0.5823]],

        [[ 0.3991]]], grad_fn=<UnsqueezeBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
________________________ test_gat_conv[1-5-g1-idtype0] ________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'ft': Scheme(shape=(1, 5), dtype=torch.float32), 'el': Scheme(shape=(1, 1), dtype=torch.float32), 'er': Scheme(shape=(1, 1), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 5, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:480: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatconv.py:303: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14526DC8>
op = 'add'
lhs_data = tensor([[[-1.5317]],

        [[ 0.1660]],

        [[ 2.0731]],

        [[ 1.1465]],

        [[-0.6028]],

        ...

        [[ 0.7517]],

        [[-0.4361]],

        [[ 0.6501]],

        [[ 0.3267]]], grad_fn=<UnsqueezeBackward0>)
rhs_data = tensor([[[ 6.8904]],

        [[ 2.1512]],

        [[-4.8666]],

        [[-4.2607]],

        [[ 3.7643]],

        ...

        [[-2.1651]],

        [[ 3.7591]],

        [[-3.2777]],

        [[-1.9009]]], grad_fn=<UnsqueezeBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
________________________ test_gat_conv[1-5-g1-idtype1] ________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'ft': Scheme(shape=(1, 5), dtype=torch.float32), 'el': Scheme(shape=(1, 1), dtype=torch.float32), 'er': Scheme(shape=(1, 1), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 5, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:480: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatconv.py:303: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01993FC8>
op = 'add'
lhs_data = tensor([[[ 0.1592]],

        [[-1.1662]],

        [[-2.7068]],

        [[-1.1107]],

        [[ 2.2456]],

        ...

        [[-3.4301]],

        [[ 0.5366]],

        [[-0.6935]],

        [[ 0.4184]]], grad_fn=<UnsqueezeBackward0>)
rhs_data = tensor([[[ 2.1154]],

        [[-1.5085]],

        [[-4.2839]],

        [[-7.9195]],

        [[ 3.5641]],

        ...

        [[-5.3141]],

        [[-0.4902]],

        [[ 0.7335]],

        [[ 3.3115]]], grad_fn=<UnsqueezeBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
________________________ test_gat_conv[1-5-g2-idtype0] ________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'ft': Scheme(shap...r': Scheme(shape=(1, 1), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int32, out_dim = 5, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:480: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatconv.py:303: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE15BA2588>
op = 'add'
lhs_data = tensor([[[ 1.1522]],

        [[-0.3416]],

        [[ 0.1758]],

        [[-0.6830]],

        [[ 0.3654]],

        ...

        [[ 0.1315]],

        [[-0.8625]],

        [[-0.2937]],

        [[ 0.5077]]], grad_fn=<UnsqueezeBackward0>)
rhs_data = tensor([[[ 0.6708]],

        [[-0.0107]],

        [[-0.2822]],

        [[ 0.3850]],

        [[ 0.6594]],

        ...

        [[-1.2101]],

        [[ 0.0383]],

        [[-0.5565]],

        [[-0.5272]]], grad_fn=<UnsqueezeBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
________________________ test_gat_conv[1-5-g2-idtype1] ________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'ft': Scheme(shap...r': Scheme(shape=(1, 1), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int64, out_dim = 5, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:480: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatconv.py:303: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01555588>
op = 'add'
lhs_data = tensor([[[-0.1790]],

        [[-0.7814]],

        [[ 4.9874]],

        [[-0.6802]],

        [[-2.2669]],

        ...

        [[ 1.8695]],

        [[-0.0702]],

        [[-0.5041]],

        [[ 1.3621]]], grad_fn=<UnsqueezeBackward0>)
rhs_data = tensor([[[-1.8432]],

        [[ 0.8166]],

        [[ 1.9445]],

        [[ 0.6126]],

        [[ 0.3049]],

        ...

        [[-1.2361]],

        [[-0.4389]],

        [[-0.8158]],

        [[ 1.4580]]], grad_fn=<UnsqueezeBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
________________________ test_gat_conv[1-5-g3-idtype0] ________________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'ft': Scheme(shape=(1, 5), dtype=torch.float32), 'el': Scheme(shape=(1, 1), dtype=torch.float32), 'er': Scheme(shape=(1, 1), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 5, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:480: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatconv.py:303: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14948B08>
op = 'add'
lhs_data = tensor([[[ 0.4389]],

        [[-7.3465]],

        [[ 4.4955]],

        [[-5.5898]],

        [[-1.6659]],

        [[-0.2334]],

        [[-4.2718]],

        [[-2.1993]],

        [[ 1.3735]]], grad_fn=<UnsqueezeBackward0>)
rhs_data = tensor([[[ 0.0936]],

        [[-2.3284]],

        [[ 1.2039]],

        [[-4.2287]],

        [[ 0.5842]],

        [[ 1.4835]],

        [[-1.4408]],

        [[-0.8756]],

        [[ 2.0756]]], grad_fn=<UnsqueezeBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
________________________ test_gat_conv[1-5-g3-idtype1] ________________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'ft': Scheme(shape=(1, 5), dtype=torch.float32), 'el': Scheme(shape=(1, 1), dtype=torch.float32), 'er': Scheme(shape=(1, 1), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 5, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:480: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatconv.py:303: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14ABE808>
op = 'add'
lhs_data = tensor([[[-0.2230]],

        [[-0.1510]],

        [[-0.1680]],

        [[-0.4649]],

        [[ 0.5044]],

        [[-0.2980]],

        [[ 0.5686]],

        [[ 0.2636]],

        [[ 0.2717]]], grad_fn=<UnsqueezeBackward0>)
rhs_data = tensor([[[-0.9247]],

        [[-0.1859]],

        [[ 0.0740]],

        [[ 0.5798]],

        [[-0.1194]],

        [[ 0.3852]],

        [[-0.1186]],

        [[-0.2958]],

        [[-0.2655]]], grad_fn=<UnsqueezeBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
________________________ test_gat_conv[1-5-g4-idtype0] ________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'ft': Scheme(shape=(1, 5), dtype=torch.float32), 'el': Scheme(shape=(1, 1), dtype=torch.float32), 'er': Scheme(shape=(1, 1), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 5, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:480: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatconv.py:303: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF015C06C8>
op = 'add'
lhs_data = tensor([[[ 0.6896]],

        [[-1.1888]],

        [[-1.2367]],

        [[-0.2142]],

        [[ 1.8823]],

        ...

        [[ 2.8191]],

        [[-2.4613]],

        [[ 1.8531]],

        [[ 1.8106]]], grad_fn=<UnsqueezeBackward0>)
rhs_data = tensor([[[ 0.3881]],

        [[ 0.2867]],

        [[-0.0917]],

        [[-0.3590]],

        [[-1.1796]],

        ...

        [[-1.1787]],

        [[ 1.2113]],

        [[ 0.2954]],

        [[-1.2517]]], grad_fn=<UnsqueezeBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
________________________ test_gat_conv[1-5-g4-idtype1] ________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'ft': Scheme(shape=(1, 5), dtype=torch.float32), 'el': Scheme(shape=(1, 1), dtype=torch.float32), 'er': Scheme(shape=(1, 1), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 5, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:480: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatconv.py:303: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF011CDE08>
op = 'add'
lhs_data = tensor([[[ 5.3988]],

        [[ 4.1892]],

        [[ 1.7641]],

        [[-1.7360]],

        [[-7.1028]],

        ...

        [[-1.3296]],

        [[ 3.9674]],

        [[ 1.4202]],

        [[-3.7856]]], grad_fn=<UnsqueezeBackward0>)
rhs_data = tensor([[[ 0.9456]],

        [[-0.2282]],

        [[-2.2591]],

        [[ 2.3813]],

        [[-0.0644]],

        ...

        [[-2.0209]],

        [[ 1.6289]],

        [[ 0.1251]],

        [[ 0.7595]]], grad_fn=<UnsqueezeBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
________________________ test_gat_conv[1-5-g5-idtype0] ________________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int32
out_dim = 5, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:480: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatconv.py:303: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01582D08>
op = 'add'
lhs_data = tensor([[[ 0.9991]],

        [[ 0.7176]],

        [[-1.6138]],

        [[ 1.5913]],

        [[ 0.1723]],

        [[ 0.2503]]], grad_fn=<UnsqueezeBackward0>)
rhs_data = tensor([[[ 5.1101]],

        [[-2.5817]],

        [[-2.6035]]], grad_fn=<UnsqueezeBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
________________________ test_gat_conv[1-5-g5-idtype1] ________________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int64
out_dim = 5, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:480: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatconv.py:303: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01068608>
op = 'add'
lhs_data = tensor([[[-1.6483]],

        [[-2.5216]],

        [[-1.3903]],

        [[-1.7610]],

        [[-1.6450]],

        [[-2.9325]]], grad_fn=<UnsqueezeBackward0>)
rhs_data = tensor([[[ 0.2066]],

        [[-0.4937]],

        [[ 1.7869]]], grad_fn=<UnsqueezeBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
________________________ test_gat_conv[1-5-g6-idtype0] ________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'ft': Scheme(shape=(1, 5), dtype=torch.float32), 'el': Scheme(shape=(1, 1), dtype=torch.float32), 'er': Scheme(shape=(1, 1), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 5, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:480: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatconv.py:303: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1469B348>
op = 'add'
lhs_data = tensor([[[ 0.3829]],

        [[ 1.2021]],

        [[-0.9324]],

        [[-1.0357]],

        [[ 0.3120]],

        ...

        [[ 2.0654]],

        [[-0.1161]],

        [[-0.1563]],

        [[ 0.7114]]], grad_fn=<UnsqueezeBackward0>)
rhs_data = tensor([[[-1.5390]],

        [[-0.3047]],

        [[ 0.5342]],

        [[-1.9946]],

        [[-1.3498]],

        ...

        [[ 2.4686]],

        [[-0.0199]],

        [[ 0.6054]],

        [[ 0.8698]]], grad_fn=<UnsqueezeBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
________________________ test_gat_conv[1-5-g6-idtype1] ________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'ft': Scheme(shape=(1, 5), dtype=torch.float32), 'el': Scheme(shape=(1, 1), dtype=torch.float32), 'er': Scheme(shape=(1, 1), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 5, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:480: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatconv.py:303: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF016BD8C8>
op = 'add'
lhs_data = tensor([[[-1.3547]],

        [[ 3.5401]],

        [[-5.2967]],

        [[-3.1144]],

        [[-0.1683]],

        ...

        [[ 4.8277]],

        [[-1.5376]],

        [[ 3.4097]],

        [[-1.4697]]], grad_fn=<UnsqueezeBackward0>)
rhs_data = tensor([[[-0.6725]],

        [[-3.3961]],

        [[ 6.5490]],

        [[ 4.5219]],

        [[-1.6624]],

        ...

        [[-3.8969]],

        [[ 2.6684]],

        [[-3.0123]],

        [[ 1.3172]]], grad_fn=<UnsqueezeBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
________________________ test_gat_conv[4-1-g0-idtype0] ________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'ft': Scheme(shap...), 'er': Scheme(shape=(4, 1), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
idtype = torch.int32, out_dim = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:480: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatconv.py:303: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE145EF6C8>
op = 'add'
lhs_data = tensor([[[-1.8391],
         [-3.5985],
         [-0.0669],
         [-0.3181]],

        [[-0.2745],
         [ 2.097...346]],

        [[-1.1638],
         [-3.4928],
         [ 0.0262],
         [-0.1353]]], grad_fn=<UnsqueezeBackward0>)
rhs_data = tensor([[[-1.4685],
         [ 3.1456],
         [-0.2243],
         [-0.1951]],

        [[-0.2192],
         [-1.833...665]],

        [[-0.9293],
         [ 3.0533],
         [ 0.0877],
         [-0.0830]]], grad_fn=<UnsqueezeBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
________________________ test_gat_conv[4-1-g0-idtype1] ________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'ft': Scheme(shap...), 'er': Scheme(shape=(4, 1), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
idtype = torch.int64, out_dim = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:480: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatconv.py:303: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14ACEB88>
op = 'add'
lhs_data = tensor([[[ 4.2121e-03],
         [ 1.1274e-02],
         [-7.4956e-01],
         [ 3.5079e-03]],

        [[ 2.9449e-0...[[-1.3677e-02],
         [ 1.0243e-01],
         [-1.4760e-01],
         [-3.4140e-01]]], grad_fn=<UnsqueezeBackward0>)
rhs_data = tensor([[[ 2.0521e-01],
         [-6.0208e-03],
         [ 8.1934e-01],
         [-1.2023e-04]],

        [[ 1.4347e+0...[[-6.6632e-01],
         [-5.4703e-02],
         [ 1.6135e-01],
         [ 1.1702e-02]]], grad_fn=<UnsqueezeBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
________________________ test_gat_conv[4-1-g1-idtype0] ________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'ft': Scheme(shape=(4, 1), dtype=torch.float32), 'el': Scheme(shape=(4, 1), dtype=torch.float32), 'er': Scheme(shape=(4, 1), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:480: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatconv.py:303: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01598888>
op = 'add'
lhs_data = tensor([[[ 0.0731],
         [ 0.0368],
         [ 0.1330],
         [ 0.7415]],

        [[ 0.2326],
         [ 0.123...379]],

        [[-0.2907],
         [ 0.2626],
         [ 0.0558],
         [ 0.9798]]], grad_fn=<UnsqueezeBackward0>)
rhs_data = tensor([[[ 0.0448],
         [ 0.1246],
         [ 0.2100],
         [ 0.6173]],

        [[ 0.1424],
         [ 0.417...967]],

        [[-0.1780],
         [ 0.8899],
         [ 0.0881],
         [ 0.8158]]], grad_fn=<UnsqueezeBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
________________________ test_gat_conv[4-1-g1-idtype1] ________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'ft': Scheme(shape=(4, 1), dtype=torch.float32), 'el': Scheme(shape=(4, 1), dtype=torch.float32), 'er': Scheme(shape=(4, 1), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:480: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatconv.py:303: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF014EB8C8>
op = 'add'
lhs_data = tensor([[[-0.2194],
         [-0.4381],
         [ 1.5392],
         [-0.1517]],

        [[-0.0310],
         [ 4.048...580]],

        [[ 0.1122],
         [-0.8929],
         [ 1.2413],
         [-0.2934]]], grad_fn=<UnsqueezeBackward0>)
rhs_data = tensor([[[-0.2222],
         [-0.3361],
         [-1.2055],
         [-1.7196]],

        [[-0.0313],
         [ 3.106...904]],

        [[ 0.1136],
         [-0.6851],
         [-0.9721],
         [-3.3247]]], grad_fn=<UnsqueezeBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
________________________ test_gat_conv[4-1-g2-idtype0] ________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'ft': Scheme(shap...r': Scheme(shape=(4, 1), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int32, out_dim = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:480: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatconv.py:303: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07747F08>
op = 'add'
lhs_data = tensor([[[ 0.2129],
         [-0.1651],
         [ 0.3476],
         [-0.5390]],

        [[ 1.1404],
         [-0.342...130]],

        [[-0.7229],
         [-0.0210],
         [ 0.5719],
         [ 0.8942]]], grad_fn=<UnsqueezeBackward0>)
rhs_data = tensor([[[ 6.2385e-03],
         [ 7.9481e-02],
         [-5.4620e-01],
         [ 1.7512e+00]],

        [[ 3.3411e-0...[[-2.1181e-02],
         [ 1.0127e-02],
         [-8.9883e-01],
         [-2.9054e+00]]], grad_fn=<UnsqueezeBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
________________________ test_gat_conv[4-1-g2-idtype1] ________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'ft': Scheme(shap...r': Scheme(shape=(4, 1), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int64, out_dim = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:480: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatconv.py:303: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF016AC6C8>
op = 'add'
lhs_data = tensor([[[ 0.1612],
         [-0.6846],
         [ 0.1214],
         [-0.1106]],

        [[-1.6178],
         [-0.275...878]],

        [[-1.2491],
         [ 0.4058],
         [-0.6443],
         [ 0.1180]]], grad_fn=<UnsqueezeBackward0>)
rhs_data = tensor([[[ 0.1458],
         [ 1.6086],
         [-0.0196],
         [-0.2803]],

        [[-1.4632],
         [ 0.648...226]],

        [[-1.1298],
         [-0.9535],
         [ 0.1042],
         [ 0.2991]]], grad_fn=<UnsqueezeBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
________________________ test_gat_conv[4-1-g3-idtype0] ________________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'ft': Scheme(shape=(4, 1), dtype=torch.float32), 'el': Scheme(shape=(4, 1), dtype=torch.float32), 'er': Scheme(shape=(4, 1), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:480: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatconv.py:303: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF016C4B08>
op = 'add'
lhs_data = tensor([[[ 0.8408],
         [ 0.9831],
         [ 0.5273],
         [-0.2994]],

        [[ 0.4827],
         [-1.089...334]],

        [[ 1.8180],
         [ 0.5948],
         [ 0.5420],
         [ 0.2276]]], grad_fn=<UnsqueezeBackward0>)
rhs_data = tensor([[[-0.4165],
         [-0.7754],
         [-0.7050],
         [-1.0075]],

        [[-0.2391],
         [ 0.859...123]],

        [[-0.9005],
         [-0.4691],
         [-0.7247],
         [ 0.7660]]], grad_fn=<UnsqueezeBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
________________________ test_gat_conv[4-1-g3-idtype1] ________________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'ft': Scheme(shape=(4, 1), dtype=torch.float32), 'el': Scheme(shape=(4, 1), dtype=torch.float32), 'er': Scheme(shape=(4, 1), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:480: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatconv.py:303: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF015AAA08>
op = 'add'
lhs_data = tensor([[[ 3.7230e-01],
         [ 1.5135e+00],
         [ 8.8428e-01],
         [-5.5128e-02]],

        [[-5.9381e-0...[[ 4.9533e-01],
         [ 9.3543e-01],
         [-2.7316e-01],
         [ 9.1078e-03]]], grad_fn=<UnsqueezeBackward0>)
rhs_data = tensor([[[-0.2332],
         [ 1.5187],
         [-0.9573],
         [-0.2131]],

        [[ 0.3720],
         [-1.269...134]],

        [[-0.3103],
         [ 0.9386],
         [ 0.2957],
         [ 0.0352]]], grad_fn=<UnsqueezeBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
________________________ test_gat_conv[4-1-g4-idtype0] ________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'ft': Scheme(shape=(4, 1), dtype=torch.float32), 'el': Scheme(shape=(4, 1), dtype=torch.float32), 'er': Scheme(shape=(4, 1), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:480: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatconv.py:303: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0148B408>
op = 'add'
lhs_data = tensor([[[ 0.0263],
         [ 0.3737],
         [ 0.1586],
         [ 0.0754]],

        [[ 0.0586],
         [-0.173...732]],

        [[-0.0661],
         [-0.2219],
         [ 0.0486],
         [-0.0187]]], grad_fn=<UnsqueezeBackward0>)
rhs_data = tensor([[[ 0.0509],
         [-0.7647],
         [-0.8217],
         [-1.6832]],

        [[ 0.1132],
         [ 0.355...344]],

        [[-0.1277],
         [ 0.4542],
         [-0.2520],
         [ 0.4168]]], grad_fn=<UnsqueezeBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
________________________ test_gat_conv[4-1-g4-idtype1] ________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'ft': Scheme(shape=(4, 1), dtype=torch.float32), 'el': Scheme(shape=(4, 1), dtype=torch.float32), 'er': Scheme(shape=(4, 1), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:480: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatconv.py:303: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE145AB388>
op = 'add'
lhs_data = tensor([[[-3.1385e+00],
         [ 4.6503e-01],
         [ 2.0344e-01],
         [ 4.1876e-02]],

        [[ 1.8386e-0...[[-3.4285e+00],
         [ 7.6762e-02],
         [ 3.6119e-01],
         [ 5.1487e+00]]], grad_fn=<UnsqueezeBackward0>)
rhs_data = tensor([[[-2.6840e-02],
         [ 3.5388e-01],
         [ 3.0025e+00],
         [-2.6629e-02]],

        [[ 1.5724e-0...[[-2.9320e-02],
         [ 5.8415e-02],
         [ 5.3307e+00],
         [-3.2740e+00]]], grad_fn=<UnsqueezeBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
________________________ test_gat_conv[4-1-g5-idtype0] ________________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int32
out_dim = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:480: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatconv.py:303: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01076848>
op = 'add'
lhs_data = tensor([[[ 4.5034],
         [ 0.2751],
         [-1.3171],
         [ 0.7076]],

        [[ 0.1624],
         [ 0.993...406]],

        [[-1.5987],
         [-2.0014],
         [ 1.2782],
         [-0.5551]]], grad_fn=<UnsqueezeBackward0>)
rhs_data = tensor([[[ 6.4658],
         [-0.2133],
         [-0.8975],
         [-4.6700]],

        [[ 0.2331],
         [-0.770...106]],

        [[ 3.2671],
         [ 0.5872],
         [-0.5402],
         [-2.5438]]], grad_fn=<UnsqueezeBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
________________________ test_gat_conv[4-1-g5-idtype1] ________________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int64
out_dim = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:480: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatconv.py:303: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF010F4B48>
op = 'add'
lhs_data = tensor([[[ 0.1368],
         [-0.4337],
         [ 0.5140],
         [-2.4391]],

        [[-1.5370],
         [ 1.569...015]],

        [[-0.2784],
         [-1.1632],
         [ 0.9167],
         [ 4.2419]]], grad_fn=<UnsqueezeBackward0>)
rhs_data = tensor([[[-0.4961],
         [ 0.0128],
         [ 0.0567],
         [ 0.7522]],

        [[ 5.5758],
         [-0.046...344]],

        [[-2.8843],
         [-0.0208],
         [-0.0876],
         [ 0.1101]]], grad_fn=<UnsqueezeBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
________________________ test_gat_conv[4-1-g6-idtype0] ________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'ft': Scheme(shape=(4, 1), dtype=torch.float32), 'el': Scheme(shape=(4, 1), dtype=torch.float32), 'er': Scheme(shape=(4, 1), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:480: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatconv.py:303: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF010CC9C8>
op = 'add'
lhs_data = tensor([[[ 0.0218],
         [ 0.5992],
         [ 0.6557],
         [ 0.1496]],

        [[-0.9409],
         [-0.168...959]],

        [[-0.0408],
         [ 0.2047],
         [ 0.2937],
         [ 0.0450]]], grad_fn=<UnsqueezeBackward0>)
rhs_data = tensor([[[ 8.2178e-02],
         [ 3.6789e+00],
         [ 2.0945e+00],
         [ 2.1408e-02]],

        [[-3.5414e+0...[[-1.5367e-01],
         [ 1.2570e+00],
         [ 9.3822e-01],
         [ 6.4324e-03]]], grad_fn=<UnsqueezeBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
________________________ test_gat_conv[4-1-g6-idtype1] ________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'ft': Scheme(shape=(4, 1), dtype=torch.float32), 'el': Scheme(shape=(4, 1), dtype=torch.float32), 'er': Scheme(shape=(4, 1), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:480: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatconv.py:303: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01129EC8>
op = 'add'
lhs_data = tensor([[[ 2.0087e-01],
         [-4.6334e-01],
         [-7.3687e-02],
         [ 1.6846e+00]],

        [[ 3.4328e-0...[[ 1.9192e-01],
         [-6.4101e-01],
         [-4.8252e-02],
         [ 1.6413e-01]]], grad_fn=<UnsqueezeBackward0>)
rhs_data = tensor([[[ 0.1795],
         [-0.2064],
         [ 1.5393],
         [ 1.4017]],

        [[ 0.3068],
         [-0.025...926]],

        [[ 0.1715],
         [-0.2855],
         [ 1.0079],
         [ 0.1366]]], grad_fn=<UnsqueezeBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
________________________ test_gat_conv[4-5-g0-idtype0] ________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'ft': Scheme(shap...), 'er': Scheme(shape=(4, 1), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
idtype = torch.int32, out_dim = 5, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:480: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatconv.py:303: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1531DC88>
op = 'add'
lhs_data = tensor([[[ 0.0133],
         [ 0.1136],
         [-0.1534],
         [-0.5244]],

        [[ 0.0412],
         [ 0.138...536]],

        [[-0.0864],
         [ 0.2272],
         [ 0.2013],
         [ 0.1085]]], grad_fn=<UnsqueezeBackward0>)
rhs_data = tensor([[[-0.1387],
         [-1.3265],
         [-0.4190],
         [ 0.4093]],

        [[ 0.3888],
         [-0.678...647]],

        [[ 0.5214],
         [-0.4421],
         [ 0.7064],
         [ 0.2876]]], grad_fn=<UnsqueezeBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
________________________ test_gat_conv[4-5-g0-idtype1] ________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'ft': Scheme(shap...), 'er': Scheme(shape=(4, 1), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
idtype = torch.int64, out_dim = 5, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:480: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatconv.py:303: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01488D08>
op = 'add'
lhs_data = tensor([[[-3.2400e-01],
         [-3.3318e-01],
         [ 7.7738e-02],
         [-3.8318e-01]],

        [[ 1.4592e+0...[[ 1.3317e+00],
         [ 9.8326e-01],
         [-9.5699e-01],
         [ 3.7058e-01]]], grad_fn=<UnsqueezeBackward0>)
rhs_data = tensor([[[-0.3254],
         [-0.7533],
         [-0.1201],
         [ 0.4705]],

        [[ 1.0818],
         [ 1.296...022]],

        [[ 0.1501],
         [ 0.5917],
         [ 0.3687],
         [-0.4503]]], grad_fn=<UnsqueezeBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
________________________ test_gat_conv[4-5-g1-idtype0] ________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'ft': Scheme(shape=(4, 5), dtype=torch.float32), 'el': Scheme(shape=(4, 1), dtype=torch.float32), 'er': Scheme(shape=(4, 1), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 5, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:480: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatconv.py:303: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0147B1C8>
op = 'add'
lhs_data = tensor([[[-1.0578e+00],
         [-8.0735e-01],
         [-3.0872e-01],
         [-3.8922e-01]],

        [[ 4.8907e-0...[[-2.5894e+00],
         [-4.7954e-01],
         [-1.1894e+00],
         [-1.8314e-01]]], grad_fn=<UnsqueezeBackward0>)
rhs_data = tensor([[[-0.0611],
         [ 1.5512],
         [ 0.3311],
         [-0.0716]],

        [[ 0.3535],
         [-0.301...298]],

        [[-0.5186],
         [ 1.9115],
         [ 0.6141],
         [ 0.1359]]], grad_fn=<UnsqueezeBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
________________________ test_gat_conv[4-5-g1-idtype1] ________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'ft': Scheme(shape=(4, 5), dtype=torch.float32), 'el': Scheme(shape=(4, 1), dtype=torch.float32), 'er': Scheme(shape=(4, 1), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 5, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:480: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatconv.py:303: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01100AC8>
op = 'add'
lhs_data = tensor([[[ 0.6509],
         [-0.3179],
         [ 0.5977],
         [ 0.3514]],

        [[-0.5883],
         [-0.156...830]],

        [[-0.1152],
         [ 1.5886],
         [ 0.4192],
         [ 0.3899]]], grad_fn=<UnsqueezeBackward0>)
rhs_data = tensor([[[ 1.2406],
         [ 1.1054],
         [ 0.3353],
         [ 1.1703]],

        [[ 0.0576],
         [-0.414...642]],

        [[ 0.7881],
         [ 0.0257],
         [ 0.4227],
         [ 2.0394]]], grad_fn=<UnsqueezeBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
________________________ test_gat_conv[4-5-g2-idtype0] ________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'ft': Scheme(shap...r': Scheme(shape=(4, 1), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int32, out_dim = 5, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:480: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatconv.py:303: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0185E588>
op = 'add'
lhs_data = tensor([[[ 2.7734],
         [ 2.3463],
         [ 0.3434],
         [-0.3500]],

        [[ 0.8232],
         [-0.377...111]],

        [[ 1.1143],
         [ 0.6319],
         [ 0.4170],
         [ 0.2499]]], grad_fn=<UnsqueezeBackward0>)
rhs_data = tensor([[[ 0.4398],
         [ 0.9114],
         [ 0.1064],
         [ 1.5394]],

        [[ 0.1279],
         [ 0.160...280]],

        [[-0.3495],
         [ 1.0082],
         [-0.2303],
         [-0.4494]]], grad_fn=<UnsqueezeBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
________________________ test_gat_conv[4-5-g2-idtype1] ________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'ft': Scheme(shap...r': Scheme(shape=(4, 1), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int64, out_dim = 5, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:480: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatconv.py:303: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF011C5A08>
op = 'add'
lhs_data = tensor([[[ 1.3400],
         [-0.2702],
         [ 0.2603],
         [-0.1598]],

        [[-0.7320],
         [-0.729...275]],

        [[ 0.4009],
         [-1.3191],
         [-0.7519],
         [ 0.5272]]], grad_fn=<UnsqueezeBackward0>)
rhs_data = tensor([[[ 1.4410],
         [ 0.3191],
         [-1.6505],
         [ 0.5480]],

        [[-1.5270],
         [-0.229...487]],

        [[-0.0774],
         [-1.5952],
         [ 0.5087],
         [ 1.1512]]], grad_fn=<UnsqueezeBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
________________________ test_gat_conv[4-5-g3-idtype0] ________________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'ft': Scheme(shape=(4, 5), dtype=torch.float32), 'el': Scheme(shape=(4, 1), dtype=torch.float32), 'er': Scheme(shape=(4, 1), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 5, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:480: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatconv.py:303: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03E98AC8>
op = 'add'
lhs_data = tensor([[[-0.7510],
         [ 0.4597],
         [ 1.4093],
         [ 0.4462]],

        [[-0.0161],
         [-0.631...693]],

        [[ 0.3405],
         [-0.5441],
         [-0.1689],
         [-0.7062]]], grad_fn=<UnsqueezeBackward0>)
rhs_data = tensor([[[-0.3594],
         [ 0.8532],
         [-0.6392],
         [ 0.3471]],

        [[-0.3738],
         [-0.155...196]],

        [[ 0.0779],
         [-0.1490],
         [ 0.4963],
         [-0.0118]]], grad_fn=<UnsqueezeBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
________________________ test_gat_conv[4-5-g3-idtype1] ________________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'ft': Scheme(shape=(4, 5), dtype=torch.float32), 'el': Scheme(shape=(4, 1), dtype=torch.float32), 'er': Scheme(shape=(4, 1), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 5, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:480: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatconv.py:303: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01120948>
op = 'add'
lhs_data = tensor([[[ 7.2095e-02],
         [-3.4801e-01],
         [-6.3903e-01],
         [ 3.2008e-01]],

        [[-2.0732e-0...[[-3.5903e-01],
         [ 1.6846e-01],
         [-8.8319e-01],
         [-1.0961e+00]]], grad_fn=<UnsqueezeBackward0>)
rhs_data = tensor([[[-0.2491],
         [ 0.4868],
         [-0.4591],
         [ 0.5683]],

        [[ 0.0620],
         [-0.061...696]],

        [[-1.2519],
         [-0.0431],
         [-0.0397],
         [-0.9564]]], grad_fn=<UnsqueezeBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
________________________ test_gat_conv[4-5-g4-idtype0] ________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'ft': Scheme(shape=(4, 5), dtype=torch.float32), 'el': Scheme(shape=(4, 1), dtype=torch.float32), 'er': Scheme(shape=(4, 1), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 5, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:480: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatconv.py:303: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01538A08>
op = 'add'
lhs_data = tensor([[[ 2.3259],
         [ 0.2847],
         [ 1.1350],
         [ 2.6880]],

        [[ 0.3409],
         [ 0.530...607]],

        [[ 0.1240],
         [ 1.2985],
         [ 0.1866],
         [-0.0744]]], grad_fn=<UnsqueezeBackward0>)
rhs_data = tensor([[[ 1.2000],
         [-1.0307],
         [-1.5316],
         [ 1.3525]],

        [[-0.6362],
         [ 1.063...840]],

        [[-0.8485],
         [ 1.8330],
         [-0.8987],
         [ 2.0917]]], grad_fn=<UnsqueezeBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
________________________ test_gat_conv[4-5-g4-idtype1] ________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'ft': Scheme(shape=(4, 5), dtype=torch.float32), 'el': Scheme(shape=(4, 1), dtype=torch.float32), 'er': Scheme(shape=(4, 1), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 5, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:480: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatconv.py:303: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03D338C8>
op = 'add'
lhs_data = tensor([[[-9.8859e-01],
         [-2.0779e-01],
         [-5.3789e-01],
         [-6.5061e-01]],

        [[ 1.3655e+0...[[-2.2603e+00],
         [ 1.7601e-01],
         [ 1.8798e-01],
         [-8.6441e-02]]], grad_fn=<UnsqueezeBackward0>)
rhs_data = tensor([[[-0.9091],
         [ 0.4617],
         [ 0.3319],
         [ 0.1775]],

        [[ 1.1608],
         [-2.194...720]],

        [[-1.2865],
         [-0.2895],
         [-0.1303],
         [ 0.3112]]], grad_fn=<UnsqueezeBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
________________________ test_gat_conv[4-5-g5-idtype0] ________________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int32
out_dim = 5, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:480: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatconv.py:303: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1451E108>
op = 'add'
lhs_data = tensor([[[-0.5932],
         [ 0.9503],
         [-0.4168],
         [-0.6529]],

        [[-0.4913],
         [ 1.561...315]],

        [[-0.7327],
         [ 1.4806],
         [-0.9935],
         [ 0.6997]]], grad_fn=<UnsqueezeBackward0>)
rhs_data = tensor([[[-0.3195],
         [ 0.1482],
         [ 0.6857],
         [-0.1038]],

        [[-0.2848],
         [-0.150...359]],

        [[ 0.0979],
         [ 0.4512],
         [ 0.3087],
         [-0.6100]]], grad_fn=<UnsqueezeBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
________________________ test_gat_conv[4-5-g5-idtype1] ________________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int64
out_dim = 5, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:480: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatconv.py:303: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF017AA288>
op = 'add'
lhs_data = tensor([[[ 0.5947],
         [ 0.3126],
         [-0.2544],
         [-0.7775]],

        [[ 0.0212],
         [-0.094...340]],

        [[ 0.5675],
         [ 0.4992],
         [-0.6576],
         [ 0.5452]]], grad_fn=<UnsqueezeBackward0>)
rhs_data = tensor([[[ 2.2880],
         [-1.6414],
         [-0.0386],
         [ 1.1530]],

        [[-0.1992],
         [-0.396...713]],

        [[ 0.1821],
         [ 1.5181],
         [-0.8703],
         [-0.7521]]], grad_fn=<UnsqueezeBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
________________________ test_gat_conv[4-5-g6-idtype0] ________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'ft': Scheme(shape=(4, 5), dtype=torch.float32), 'el': Scheme(shape=(4, 1), dtype=torch.float32), 'er': Scheme(shape=(4, 1), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 5, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:480: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatconv.py:303: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01569188>
op = 'add'
lhs_data = tensor([[[ 0.3711],
         [ 0.3426],
         [-0.5760],
         [ 0.1867]],

        [[ 0.6399],
         [ 0.111...294]],

        [[-1.2394],
         [-0.5844],
         [-0.6362],
         [ 0.6868]]], grad_fn=<UnsqueezeBackward0>)
rhs_data = tensor([[[-0.1041],
         [-0.1129],
         [ 0.6220],
         [ 0.1311]],

        [[-0.6691],
         [-0.116...912]],

        [[ 0.8713],
         [ 0.2302],
         [-0.5589],
         [ 0.1271]]], grad_fn=<UnsqueezeBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
________________________ test_gat_conv[4-5-g6-idtype1] ________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'ft': Scheme(shape=(4, 5), dtype=torch.float32), 'el': Scheme(shape=(4, 1), dtype=torch.float32), 'er': Scheme(shape=(4, 1), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 5, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:480: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatconv.py:303: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01196F08>
op = 'add'
lhs_data = tensor([[[ 5.8502e-01],
         [-8.2176e-01],
         [ 1.7468e+00],
         [-8.4681e-02]],

        [[ 2.9200e-0...[[-1.3753e-01],
         [-1.5352e+00],
         [ 9.7422e-01],
         [-8.6121e-02]]], grad_fn=<UnsqueezeBackward0>)
rhs_data = tensor([[[ 0.4669],
         [ 0.6859],
         [ 0.1342],
         [-1.0924]],

        [[ 1.5372],
         [-2.057...086]],

        [[-1.2751],
         [ 1.8732],
         [ 0.3767],
         [-1.9647]]], grad_fn=<UnsqueezeBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_gat_conv_bi[1-1-g0-idtype0] _______________________

g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int32, out_dim = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gat_conv_bi(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATConv(5, out_dim, num_heads)
        feat = (F.randn((g.number_of_src_nodes(), 5)), F.randn((g.number_of_dst_nodes(), 5)))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:504: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatconv.py:303: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03D33488>
op = 'add'
lhs_data = tensor([[[-1.4320]],

        [[ 1.0467]]], grad_fn=<UnsqueezeBackward0>)
rhs_data = tensor([[[ 0.7716]],

        [[-1.3917]],

        [[-1.9567]],

        [[ 2.0927]]], grad_fn=<UnsqueezeBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_gat_conv_bi[1-1-g0-idtype1] _______________________

g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int64, out_dim = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gat_conv_bi(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATConv(5, out_dim, num_heads)
        feat = (F.randn((g.number_of_src_nodes(), 5)), F.randn((g.number_of_dst_nodes(), 5)))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:504: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatconv.py:303: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF018480C8>
op = 'add'
lhs_data = tensor([[[0.2925]],

        [[1.0149]]], grad_fn=<UnsqueezeBackward0>)
rhs_data = tensor([[[-3.9312]],

        [[-1.4319]],

        [[-1.7173]],

        [[ 2.4287]]], grad_fn=<UnsqueezeBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_gat_conv_bi[1-1-g1-idtype0] _______________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int32
out_dim = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gat_conv_bi(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATConv(5, out_dim, num_heads)
        feat = (F.randn((g.number_of_src_nodes(), 5)), F.randn((g.number_of_dst_nodes(), 5)))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:504: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatconv.py:303: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14616C88>
op = 'add'
lhs_data = tensor([[[ 0.0786]],

        [[-0.6760]],

        [[ 0.6412]],

        [[ 1.7543]],

        [[-1.5492]],

        [[ 0.7744]]], grad_fn=<UnsqueezeBackward0>)
rhs_data = tensor([[[-3.1858]],

        [[-0.2868]],

        [[-8.2755]]], grad_fn=<UnsqueezeBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_gat_conv_bi[1-1-g1-idtype1] _______________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int64
out_dim = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gat_conv_bi(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATConv(5, out_dim, num_heads)
        feat = (F.randn((g.number_of_src_nodes(), 5)), F.randn((g.number_of_dst_nodes(), 5)))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:504: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatconv.py:303: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01429FC8>
op = 'add'
lhs_data = tensor([[[ 0.5071]],

        [[ 0.4418]],

        [[-0.5130]],

        [[-1.2088]],

        [[-0.8907]],

        [[-0.9454]]], grad_fn=<UnsqueezeBackward0>)
rhs_data = tensor([[[-5.8926]],

        [[-3.6759]],

        [[-2.7233]]], grad_fn=<UnsqueezeBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_gat_conv_bi[1-2-g0-idtype0] _______________________

g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int32, out_dim = 2, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gat_conv_bi(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATConv(5, out_dim, num_heads)
        feat = (F.randn((g.number_of_src_nodes(), 5)), F.randn((g.number_of_dst_nodes(), 5)))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:504: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatconv.py:303: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14920CC8>
op = 'add'
lhs_data = tensor([[[-0.3434]],

        [[ 0.4823]]], grad_fn=<UnsqueezeBackward0>)
rhs_data = tensor([[[-1.4087]],

        [[ 1.1793]],

        [[-0.1260]],

        [[ 0.4495]]], grad_fn=<UnsqueezeBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_gat_conv_bi[1-2-g0-idtype1] _______________________

g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int64, out_dim = 2, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gat_conv_bi(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATConv(5, out_dim, num_heads)
        feat = (F.randn((g.number_of_src_nodes(), 5)), F.randn((g.number_of_dst_nodes(), 5)))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:504: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatconv.py:303: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF015F4BC8>
op = 'add'
lhs_data = tensor([[[-1.6297]],

        [[-2.3387]]], grad_fn=<UnsqueezeBackward0>)
rhs_data = tensor([[[ 1.5073]],

        [[-4.1799]],

        [[-4.8029]],

        [[ 2.9013]]], grad_fn=<UnsqueezeBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_gat_conv_bi[1-2-g1-idtype0] _______________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int32
out_dim = 2, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gat_conv_bi(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATConv(5, out_dim, num_heads)
        feat = (F.randn((g.number_of_src_nodes(), 5)), F.randn((g.number_of_dst_nodes(), 5)))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:504: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatconv.py:303: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01836F88>
op = 'add'
lhs_data = tensor([[[ 1.8374]],

        [[-1.9623]],

        [[ 2.3261]],

        [[-1.8526]],

        [[ 2.7206]],

        [[ 0.9523]]], grad_fn=<UnsqueezeBackward0>)
rhs_data = tensor([[[0.5371]],

        [[0.2202]],

        [[0.0986]]], grad_fn=<UnsqueezeBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_gat_conv_bi[1-2-g1-idtype1] _______________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int64
out_dim = 2, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gat_conv_bi(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATConv(5, out_dim, num_heads)
        feat = (F.randn((g.number_of_src_nodes(), 5)), F.randn((g.number_of_dst_nodes(), 5)))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:504: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatconv.py:303: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0105B548>
op = 'add'
lhs_data = tensor([[[-0.1147]],

        [[ 0.3016]],

        [[-0.2579]],

        [[-0.9147]],

        [[-0.3819]],

        [[-0.6094]]], grad_fn=<UnsqueezeBackward0>)
rhs_data = tensor([[[-0.0954]],

        [[-0.4489]],

        [[ 1.9191]]], grad_fn=<UnsqueezeBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_gat_conv_bi[4-1-g0-idtype0] _______________________

g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int32, out_dim = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gat_conv_bi(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATConv(5, out_dim, num_heads)
        feat = (F.randn((g.number_of_src_nodes(), 5)), F.randn((g.number_of_dst_nodes(), 5)))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:504: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatconv.py:303: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14A7D648>
op = 'add'
lhs_data = tensor([[[ 0.2676],
         [ 0.1643],
         [ 0.1074],
         [-0.2957]],

        [[ 0.0728],
         [-0.0875],
         [ 0.0986],
         [-0.5724]]], grad_fn=<UnsqueezeBackward0>)
rhs_data = tensor([[[-1.7734],
         [ 1.4922],
         [ 0.3974],
         [ 0.0229]],

        [[ 5.3069],
         [-4.322...133]],

        [[-5.9294],
         [ 4.4256],
         [-0.5443],
         [ 0.0385]]], grad_fn=<UnsqueezeBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_gat_conv_bi[4-1-g0-idtype1] _______________________

g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int64, out_dim = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gat_conv_bi(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATConv(5, out_dim, num_heads)
        feat = (F.randn((g.number_of_src_nodes(), 5)), F.randn((g.number_of_dst_nodes(), 5)))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:504: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatconv.py:303: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1460F408>
op = 'add'
lhs_data = tensor([[[ 0.0444],
         [ 0.4088],
         [-0.1875],
         [-1.3407]],

        [[ 0.0212],
         [ 0.1467],
         [-0.3226],
         [-1.3909]]], grad_fn=<UnsqueezeBackward0>)
rhs_data = tensor([[[-0.1697],
         [-0.6530],
         [ 0.8314],
         [-0.2115]],

        [[ 0.0469],
         [-1.626...456]],

        [[ 0.1050],
         [-0.6981],
         [ 0.0377],
         [ 0.0473]]], grad_fn=<UnsqueezeBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_gat_conv_bi[4-1-g1-idtype0] _______________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int32
out_dim = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gat_conv_bi(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATConv(5, out_dim, num_heads)
        feat = (F.randn((g.number_of_src_nodes(), 5)), F.randn((g.number_of_dst_nodes(), 5)))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:504: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatconv.py:303: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14A92788>
op = 'add'
lhs_data = tensor([[[-0.0083],
         [ 0.1242],
         [ 0.3526],
         [ 0.6818]],

        [[-0.1275],
         [-0.043...193]],

        [[ 0.3681],
         [ 0.3535],
         [-0.4101],
         [ 0.8809]]], grad_fn=<UnsqueezeBackward0>)
rhs_data = tensor([[[-1.5370],
         [ 1.3722],
         [ 0.5454],
         [-0.9175]],

        [[-2.0455],
         [ 2.171...585]],

        [[ 0.9192],
         [-0.3628],
         [-0.0500],
         [ 0.2580]]], grad_fn=<UnsqueezeBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_gat_conv_bi[4-1-g1-idtype1] _______________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int64
out_dim = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gat_conv_bi(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATConv(5, out_dim, num_heads)
        feat = (F.randn((g.number_of_src_nodes(), 5)), F.randn((g.number_of_dst_nodes(), 5)))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:504: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatconv.py:303: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF017ED588>
op = 'add'
lhs_data = tensor([[[ 1.5533],
         [-0.0980],
         [-0.1912],
         [ 0.2753]],

        [[ 1.3518],
         [-0.020...808]],

        [[-1.2337],
         [-0.0756],
         [-0.1983],
         [ 0.4026]]], grad_fn=<UnsqueezeBackward0>)
rhs_data = tensor([[[ 7.8029e-01],
         [-1.1792e+00],
         [ 4.4077e-04],
         [ 1.3406e-01]],

        [[-5.3115e-0...[[ 1.0161e+00],
         [ 1.8174e+00],
         [-3.5647e+00],
         [-1.1046e+00]]], grad_fn=<UnsqueezeBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_gat_conv_bi[4-2-g0-idtype0] _______________________

g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int32, out_dim = 2, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gat_conv_bi(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATConv(5, out_dim, num_heads)
        feat = (F.randn((g.number_of_src_nodes(), 5)), F.randn((g.number_of_dst_nodes(), 5)))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:504: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatconv.py:303: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01075788>
op = 'add'
lhs_data = tensor([[[ 0.3133],
         [-1.1405],
         [-1.2005],
         [-0.4874]],

        [[ 0.0423],
         [-0.0979],
         [ 1.0660],
         [ 0.8075]]], grad_fn=<UnsqueezeBackward0>)
rhs_data = tensor([[[ 0.3773],
         [-3.5723],
         [-0.5786],
         [-0.4897]],

        [[ 0.0703],
         [-0.086...194]],

        [[-0.1587],
         [-4.0715],
         [ 0.5773],
         [ 0.0537]]], grad_fn=<UnsqueezeBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_gat_conv_bi[4-2-g0-idtype1] _______________________

g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int64, out_dim = 2, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gat_conv_bi(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATConv(5, out_dim, num_heads)
        feat = (F.randn((g.number_of_src_nodes(), 5)), F.randn((g.number_of_dst_nodes(), 5)))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:504: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatconv.py:303: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF014A4C48>
op = 'add'
lhs_data = tensor([[[ 1.2032],
         [ 0.2111],
         [ 0.1349],
         [-0.2698]],

        [[-1.0306],
         [ 2.3088],
         [-0.0388],
         [ 0.4123]]], grad_fn=<UnsqueezeBackward0>)
rhs_data = tensor([[[-0.1317],
         [ 0.8064],
         [-0.2262],
         [ 1.5803]],

        [[ 0.8038],
         [-0.087...906]],

        [[-3.7972],
         [-0.4466],
         [ 0.3719],
         [ 1.8697]]], grad_fn=<UnsqueezeBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_gat_conv_bi[4-2-g1-idtype0] _______________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int32
out_dim = 2, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gat_conv_bi(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATConv(5, out_dim, num_heads)
        feat = (F.randn((g.number_of_src_nodes(), 5)), F.randn((g.number_of_dst_nodes(), 5)))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:504: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatconv.py:303: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF010BB408>
op = 'add'
lhs_data = tensor([[[ 0.4655],
         [-0.5376],
         [ 0.0758],
         [ 0.2456]],

        [[ 0.4037],
         [ 0.262...758]],

        [[ 0.4262],
         [-0.2143],
         [ 0.1921],
         [-0.5283]]], grad_fn=<UnsqueezeBackward0>)
rhs_data = tensor([[[ 1.3178],
         [-0.6492],
         [ 0.5210],
         [-0.8872]],

        [[-0.5974],
         [-0.803...880]],

        [[ 0.8223],
         [ 0.1198],
         [-0.1189],
         [ 0.1537]]], grad_fn=<UnsqueezeBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_gat_conv_bi[4-2-g1-idtype1] _______________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int64
out_dim = 2, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gat_conv_bi(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATConv(5, out_dim, num_heads)
        feat = (F.randn((g.number_of_src_nodes(), 5)), F.randn((g.number_of_dst_nodes(), 5)))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:504: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatconv.py:303: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF017871C8>
op = 'add'
lhs_data = tensor([[[ 1.0931],
         [ 0.6261],
         [-1.9108],
         [ 0.7142]],

        [[-0.9165],
         [ 2.080...444]],

        [[-1.3186],
         [ 1.6897],
         [ 1.2442],
         [-1.9202]]], grad_fn=<UnsqueezeBackward0>)
rhs_data = tensor([[[ 0.5002],
         [-0.2068],
         [ 0.8156],
         [ 0.5811]],

        [[-0.2113],
         [-2.469...318]],

        [[ 0.1673],
         [ 0.1176],
         [ 0.7110],
         [ 0.7620]]], grad_fn=<UnsqueezeBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_______________________ test_gatv2_conv[1-1-g0-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'el': Scheme(shap...), 'er': Scheme(shape=(1, 1), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
idtype = torch.int32, out_dim = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gatv2_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATv2Conv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:519: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatv2conv.py:288: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF017E6DC8>
op = 'add'
lhs_data = tensor([[[ 3.4006]],

        [[-1.2331]],

        [[ 1.5457]],

        [[ 4.0673]],

        [[ 1.1407]],

        ...187]],

        [[-3.1077]],

        [[-1.2028]],

        [[-4.0908]],

        [[ 1.2259]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[ 0.8317]],

        [[ 3.8842]],

        [[ 1.0626]],

        [[-1.9911]],

        [[ 1.7514]],

        ...592]],

        [[ 0.3502]],

        [[-1.5231]],

        [[-0.5106]],

        [[ 0.0689]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_______________________ test_gatv2_conv[1-1-g0-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'el': Scheme(shap...), 'er': Scheme(shape=(1, 1), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
idtype = torch.int64, out_dim = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gatv2_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATv2Conv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:519: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatv2conv.py:288: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14A71288>
op = 'add'
lhs_data = tensor([[[ 0.4397]],

        [[ 1.4534]],

        [[ 2.5390]],

        [[ 1.7433]],

        [[-2.9063]],

        ...991]],

        [[ 5.0230]],

        [[ 0.6680]],

        [[-2.0967]],

        [[ 1.1920]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[ 0.8855]],

        [[-0.2862]],

        [[ 1.3963]],

        [[ 1.7140]],

        [[ 3.0068]],

        ...758]],

        [[ 0.3029]],

        [[ 0.9372]],

        [[-2.3731]],

        [[ 0.3232]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_______________________ test_gatv2_conv[1-1-g1-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'el': Scheme(shape=(1, 1), dtype=torch.float32), 'er': Scheme(shape=(1, 1), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gatv2_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATv2Conv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:519: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatv2conv.py:288: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01487548>
op = 'add'
lhs_data = tensor([[[ 3.0653]],

        [[ 1.0486]],

        [[-1.8175]],

        [[-0.7190]],

        [[-0.2430]],

        ...167]],

        [[ 0.1716]],

        [[ 0.5762]],

        [[-0.7160]],

        [[ 1.0905]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[-0.4232]],

        [[ 0.9032]],

        [[ 0.2506]],

        [[ 0.0391]],

        [[ 0.5624]],

        ...410]],

        [[-2.0975]],

        [[ 0.7346]],

        [[ 0.8059]],

        [[-1.0117]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_______________________ test_gatv2_conv[1-1-g1-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'el': Scheme(shape=(1, 1), dtype=torch.float32), 'er': Scheme(shape=(1, 1), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gatv2_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATv2Conv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:519: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatv2conv.py:288: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0107EC08>
op = 'add'
lhs_data = tensor([[[-4.2231]],

        [[ 1.6339]],

        [[ 1.0617]],

        [[-1.4770]],

        [[ 5.8954]],

        ...221]],

        [[-1.6784]],

        [[-1.2556]],

        [[-1.0745]],

        [[-4.5031]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[-0.7220]],

        [[ 0.4323]],

        [[-0.7422]],

        [[-0.2063]],

        [[-1.1746]],

        ...792]],

        [[ 0.0992]],

        [[ 0.7902]],

        [[-0.0793]],

        [[ 0.0973]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_______________________ test_gatv2_conv[1-1-g2-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'el': Scheme(shap...r': Scheme(shape=(1, 1), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int32, out_dim = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gatv2_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATv2Conv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:519: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatv2conv.py:288: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01479548>
op = 'add'
lhs_data = tensor([[[-1.8902]],

        [[ 2.1492]],

        [[ 1.0958]],

        [[ 2.8659]],

        [[ 0.2726]],

        ...862]],

        [[ 1.1245]],

        [[-2.4504]],

        [[-0.3411]],

        [[-3.9278]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[ 1.1436]],

        [[-2.7407]],

        [[-2.1043]],

        [[ 2.6611]],

        [[-1.7747]],

        ...388]],

        [[-1.6698]],

        [[-0.2374]],

        [[ 0.5936]],

        [[ 0.1277]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_______________________ test_gatv2_conv[1-1-g2-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'el': Scheme(shap...r': Scheme(shape=(1, 1), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int64, out_dim = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gatv2_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATv2Conv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:519: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatv2conv.py:288: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF014A4888>
op = 'add'
lhs_data = tensor([[[-1.0911]],

        [[-3.0789]],

        [[ 0.0673]],

        [[-2.1958]],

        [[ 2.7108]],

        ...033]],

        [[-0.3025]],

        [[ 1.7746]],

        [[-0.1191]],

        [[ 1.4300]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[-4.7716]],

        [[ 3.4543]],

        [[ 0.6603]],

        [[ 1.5550]],

        [[-0.1754]],

        ...590]],

        [[ 0.2426]],

        [[ 0.0713]],

        [[-0.3467]],

        [[ 2.1244]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_______________________ test_gatv2_conv[1-1-g3-idtype0] _______________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'el': Scheme(shape=(1, 1), dtype=torch.float32), 'er': Scheme(shape=(1, 1), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gatv2_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATv2Conv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:519: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatv2conv.py:288: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01082388>
op = 'add'
lhs_data = tensor([[[ 0.7359]],

        [[ 1.2847]],

        [[ 0.6146]],

        [[-1.5344]],

        [[-0.8204]],

        [[-2.9005]],

        [[-0.3618]],

        [[-3.1979]],

        [[ 1.6292]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[ 1.3635]],

        [[-0.2256]],

        [[-0.0303]],

        [[-0.6372]],

        [[-1.2608]],

        [[-0.1977]],

        [[ 1.5722]],

        [[-2.0326]],

        [[ 0.2560]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_______________________ test_gatv2_conv[1-1-g3-idtype1] _______________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'el': Scheme(shape=(1, 1), dtype=torch.float32), 'er': Scheme(shape=(1, 1), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gatv2_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATv2Conv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:519: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatv2conv.py:288: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14AC9D48>
op = 'add'
lhs_data = tensor([[[-1.5605]],

        [[-1.6850]],

        [[ 0.5212]],

        [[-1.2256]],

        [[ 0.3028]],

        [[-0.4965]],

        [[-2.2771]],

        [[ 0.1697]],

        [[-1.2633]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[-1.1682]],

        [[-1.1677]],

        [[ 0.3886]],

        [[-0.6189]],

        [[ 1.6004]],

        [[ 0.8832]],

        [[-1.8164]],

        [[-0.3867]],

        [[-2.3792]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_______________________ test_gatv2_conv[1-1-g4-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'el': Scheme(shape=(1, 1), dtype=torch.float32), 'er': Scheme(shape=(1, 1), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gatv2_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATv2Conv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:519: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatv2conv.py:288: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF019C16C8>
op = 'add'
lhs_data = tensor([[[ 3.2027]],

        [[ 2.5726]],

        [[-4.1354]],

        [[-0.2476]],

        [[-0.5240]],

        ...838]],

        [[-3.5967]],

        [[ 1.1413]],

        [[ 1.1737]],

        [[-1.7776]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[-0.1346]],

        [[-1.6616]],

        [[ 2.6730]],

        [[ 0.5437]],

        [[ 1.1903]],

        ...775]],

        [[ 0.3275]],

        [[-1.1535]],

        [[-0.4655]],

        [[ 0.5846]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_______________________ test_gatv2_conv[1-1-g4-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'el': Scheme(shape=(1, 1), dtype=torch.float32), 'er': Scheme(shape=(1, 1), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gatv2_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATv2Conv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:519: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatv2conv.py:288: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01568988>
op = 'add'
lhs_data = tensor([[[ 0.2378]],

        [[-1.9865]],

        [[-2.2731]],

        [[-3.6583]],

        [[-1.2145]],

        ...959]],

        [[-4.0923]],

        [[-2.0996]],

        [[ 1.1122]],

        [[ 0.7369]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[-0.3954]],

        [[ 0.2837]],

        [[ 1.8011]],

        [[ 0.8587]],

        [[ 1.4266]],

        ...033]],

        [[ 2.4676]],

        [[ 1.3969]],

        [[-0.2243]],

        [[-1.2974]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_______________________ test_gatv2_conv[1-1-g5-idtype0] _______________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int32
out_dim = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gatv2_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATv2Conv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:519: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatv2conv.py:288: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE145DD908>
op = 'add'
lhs_data = tensor([[[-0.0665]],

        [[-0.2400]],

        [[-1.8651]],

        [[ 1.0007]],

        [[ 1.2387]],

        [[ 1.3013]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[-0.0665]],

        [[-0.2400]],

        [[-1.8651]]], grad_fn=<SliceBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_______________________ test_gatv2_conv[1-1-g5-idtype1] _______________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int64
out_dim = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gatv2_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATv2Conv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:519: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatv2conv.py:288: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01563E48>
op = 'add'
lhs_data = tensor([[[-1.8393]],

        [[ 1.9085]],

        [[-0.2135]],

        [[-1.7120]],

        [[-0.7175]],

        [[-0.2323]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[-1.8393]],

        [[ 1.9085]],

        [[-0.2135]]], grad_fn=<SliceBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_______________________ test_gatv2_conv[1-1-g6-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'el': Scheme(shape=(1, 1), dtype=torch.float32), 'er': Scheme(shape=(1, 1), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gatv2_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATv2Conv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:519: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatv2conv.py:288: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF014C9448>
op = 'add'
lhs_data = tensor([[[ 0.6823]],

        [[-2.3584]],

        [[ 0.0413]],

        [[ 0.8772]],

        [[ 2.3237]],

        ...211]],

        [[-0.6354]],

        [[-0.4140]],

        [[ 0.0647]],

        [[-1.5931]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[ 0.2263]],

        [[-3.5048]],

        [[-0.2677]],

        [[ 0.4227]],

        [[ 0.3877]],

        ...524]],

        [[-0.5669]],

        [[-2.2883]],

        [[ 1.2446]],

        [[ 0.2460]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_______________________ test_gatv2_conv[1-1-g6-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'el': Scheme(shape=(1, 1), dtype=torch.float32), 'er': Scheme(shape=(1, 1), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gatv2_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATv2Conv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:519: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatv2conv.py:288: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01513588>
op = 'add'
lhs_data = tensor([[[-3.0107]],

        [[ 1.2382]],

        [[ 0.4750]],

        [[ 1.7366]],

        [[-0.8193]],

        ...037]],

        [[-1.4020]],

        [[-0.3603]],

        [[ 3.6699]],

        [[ 3.9827]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[-3.4804]],

        [[-0.4598]],

        [[ 0.8298]],

        [[ 1.1570]],

        [[-0.2940]],

        ...298]],

        [[-0.4173]],

        [[-1.6607]],

        [[ 3.3611]],

        [[ 2.1099]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_______________________ test_gatv2_conv[1-5-g0-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'el': Scheme(shap...), 'er': Scheme(shape=(1, 5), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
idtype = torch.int32, out_dim = 5, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gatv2_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATv2Conv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:519: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatv2conv.py:288: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01054D48>
op = 'add'
lhs_data = tensor([[[ 1.5148e+00,  2.5888e+00, -2.8531e+00, -2.1871e+00,  3.9441e+00]],

        [[-9.1285e-01, -7.4840e-01,  5.3...00e-01]],

        [[ 6.1948e-01, -5.0324e-02, -9.4104e-01, -1.0029e-01,  1.5055e+00]]],
       grad_fn=<ViewBackward>)
rhs_data = tensor([[[ 2.5246, -2.2849,  1.5732, -0.0083, -0.1820]],

        [[-1.4085,  1.1460, -1.5986,  0.0417,  0.4138]],

  ... -0.0667, -0.7912,  0.1692]],

        [[-0.1236,  0.8247, -1.3329, -0.6027, -0.2373]]],
       grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_______________________ test_gatv2_conv[1-5-g0-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'el': Scheme(shap...), 'er': Scheme(shape=(1, 5), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
idtype = torch.int64, out_dim = 5, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gatv2_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATv2Conv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:519: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatv2conv.py:288: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF016BBB88>
op = 'add'
lhs_data = tensor([[[ 0.5727,  0.4000,  0.8717, -0.4968, -1.4469]],

        [[ 0.4046, -0.2565, -0.4604, -1.2137, -1.3510]],

  ... -0.1572,  0.3949,  0.0574]],

        [[ 0.2527,  0.3987, -0.5381, -0.5574, -0.0133]]],
       grad_fn=<ViewBackward>)
rhs_data = tensor([[[-1.7061,  0.7093, -0.2583, -0.5549,  0.7747]],

        [[ 0.6093,  0.3031, -0.1591, -0.0443, -0.2775]],

  ... -1.5955, -0.2667,  0.6728]],

        [[ 0.5848,  0.0063,  0.2395,  0.3745, -0.7827]]],
       grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_______________________ test_gatv2_conv[1-5-g1-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'el': Scheme(shape=(1, 5), dtype=torch.float32), 'er': Scheme(shape=(1, 5), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 5, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gatv2_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATv2Conv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:519: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatv2conv.py:288: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0107A508>
op = 'add'
lhs_data = tensor([[[ 1.2381e-01, -1.1706e-01,  1.2634e+00,  3.9159e-01, -1.4498e+00]],

        [[-4.5678e-01, -4.4236e-01,  1.3...92e-01]],

        [[-1.0501e+00,  2.6102e-01,  1.0067e+00,  4.1708e-01, -1.8376e-01]]],
       grad_fn=<ViewBackward>)
rhs_data = tensor([[[ 0.5722, -0.5137,  0.7577, -1.2610,  0.0872]],

        [[ 0.7465, -0.4289, -0.7702,  0.2183,  0.6519]],

  ...  0.7434, -0.2745, -0.6597]],

        [[ 0.2686, -0.7324,  0.1841,  0.3042,  0.6393]]],
       grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_______________________ test_gatv2_conv[1-5-g1-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'el': Scheme(shape=(1, 5), dtype=torch.float32), 'er': Scheme(shape=(1, 5), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 5, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gatv2_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATv2Conv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:519: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatv2conv.py:288: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF014C8548>
op = 'add'
lhs_data = tensor([[[ 0.7329,  0.1476,  0.3643,  0.8091,  0.5560]],

        [[-1.3230,  0.8121, -1.4328, -0.0263, -0.4984]],

  ... -0.6759, -0.6629,  0.4691]],

        [[ 0.2818, -1.0143,  1.2156, -3.1878,  0.8440]]],
       grad_fn=<ViewBackward>)
rhs_data = tensor([[[-0.7300, -1.6988, -1.9720,  0.7772, -0.7080]],

        [[ 1.2889,  2.0599,  1.7396, -1.9057,  0.4122]],

  ... -0.8099, -1.0399, -1.4252]],

        [[-2.8105, -2.4611, -1.4127, -0.5273, -3.0439]]],
       grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_______________________ test_gatv2_conv[1-5-g2-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'el': Scheme(shap...r': Scheme(shape=(1, 5), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int32, out_dim = 5, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gatv2_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATv2Conv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:519: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatv2conv.py:288: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE149DE9C8>
op = 'add'
lhs_data = tensor([[[-0.6291,  0.8548, -2.4163,  1.3109, -0.6682]],

        [[ 3.5862, -1.2499, -0.5756, -1.5179, -1.7485]],

  ...  2.9952, -1.8093,  0.4392]],

        [[-0.4763,  0.5867,  0.6701, -0.4600, -0.4346]]],
       grad_fn=<ViewBackward>)
rhs_data = tensor([[[ 1.1074,  0.5343, -0.1144, -1.0100,  0.5554]],

        [[-4.0283,  0.2205, -2.8170,  1.4045, -1.3161]],

  ...  1.7036, -0.8401,  0.8875]],

        [[ 0.1147, -0.8348,  0.7037, -0.7234,  1.0058]]],
       grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_______________________ test_gatv2_conv[1-5-g2-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'el': Scheme(shap...r': Scheme(shape=(1, 5), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int64, out_dim = 5, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gatv2_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATv2Conv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:519: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatv2conv.py:288: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01551AC8>
op = 'add'
lhs_data = tensor([[[ 1.4268, -2.9529,  0.5167,  0.5541, -2.7901]],

        [[ 0.4666,  2.8637,  0.1399, -0.7225, -0.2497]],

  ... -1.0349,  1.5425, -1.6701]],

        [[ 0.2791,  0.7481, -1.4440, -0.2677,  0.7414]]],
       grad_fn=<ViewBackward>)
rhs_data = tensor([[[-1.2443,  5.9451,  0.9736,  2.8319, -0.8730]],

        [[ 0.9384, -1.8799, -1.1468, -0.7729, -0.2988]],

  ...  2.2856,  2.1638,  0.3114]],

        [[ 0.1327, -3.6788,  0.0395, -0.4713, -0.1138]]],
       grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_______________________ test_gatv2_conv[1-5-g3-idtype0] _______________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'el': Scheme(shape=(1, 5), dtype=torch.float32), 'er': Scheme(shape=(1, 5), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 5, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gatv2_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATv2Conv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:519: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatv2conv.py:288: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03CE0588>
op = 'add'
lhs_data = tensor([[[-0.7852, -0.7929,  0.0372,  0.3858, -0.7274]],

        [[-1.8366, -0.8200, -2.1524, -1.1223, -0.6887]],

  ...  0.8607,  0.7988,  1.3602]],

        [[ 1.4845,  1.4996,  1.1831,  1.7277,  1.5745]]],
       grad_fn=<ViewBackward>)
rhs_data = tensor([[[ 0.5907, -1.2105,  0.0297, -0.2240, -0.7962]],

        [[ 2.2359, -0.9497, -0.7823, -2.4867,  0.1249]],

  ...  1.6151,  3.1935,  0.4308]],

        [[-2.6489,  1.4154,  1.2612,  3.2913, -0.1927]]],
       grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_______________________ test_gatv2_conv[1-5-g3-idtype1] _______________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'el': Scheme(shape=(1, 5), dtype=torch.float32), 'er': Scheme(shape=(1, 5), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 5, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gatv2_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATv2Conv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:519: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatv2conv.py:288: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0177FC48>
op = 'add'
lhs_data = tensor([[[ 5.9275e-01, -1.2774e+00, -7.4747e-02,  1.8935e+00, -3.9118e-01]],

        [[ 2.0656e+00, -1.8989e+00, -5.2...87e-01]],

        [[ 1.2896e+00, -2.3212e+00,  3.2567e-04, -2.6042e+00,  6.0513e-01]]],
       grad_fn=<ViewBackward>)
rhs_data = tensor([[[-2.1752e-02, -5.6488e-01, -1.2923e+00, -8.7637e-02, -8.0199e-01]],

        [[ 4.8230e-01, -4.8182e-01,  1.7...60e-01]],

        [[-2.6137e-01,  2.5633e-01,  2.7876e+00,  1.8108e+00, -5.3170e-01]]],
       grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_______________________ test_gatv2_conv[1-5-g4-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'el': Scheme(shape=(1, 5), dtype=torch.float32), 'er': Scheme(shape=(1, 5), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 5, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gatv2_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATv2Conv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:519: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatv2conv.py:288: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1453F408>
op = 'add'
lhs_data = tensor([[[ 0.9970,  0.0828,  0.9195, -0.7405, -0.4092]],

        [[-0.6179,  0.3945,  0.5464,  1.8047, -1.2336]],

  ... -1.1554, -2.2575,  2.1639]],

        [[ 1.1258,  1.2439,  3.2631, -2.1868,  0.8612]]],
       grad_fn=<ViewBackward>)
rhs_data = tensor([[[-0.0748, -0.5715,  0.0551, -1.0049, -0.0556]],

        [[ 0.5954, -0.7703,  3.5665, -0.2745,  1.4391]],

  ...  0.3603,  1.3518, -1.1579]],

        [[-1.0414,  0.4542, -1.2743, -1.4739, -1.1903]]],
       grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_______________________ test_gatv2_conv[1-5-g4-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'el': Scheme(shape=(1, 5), dtype=torch.float32), 'er': Scheme(shape=(1, 5), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 5, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gatv2_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATv2Conv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:519: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatv2conv.py:288: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01585F48>
op = 'add'
lhs_data = tensor([[[-1.5380, -0.9819,  0.2317, -1.3171,  1.3701]],

        [[ 0.2318,  0.7418,  0.7520,  1.1458,  0.1644]],

  ... -0.3337,  1.8191, -0.5714]],

        [[ 0.7866,  0.5468, -0.2027,  1.2150, -0.2311]]],
       grad_fn=<ViewBackward>)
rhs_data = tensor([[[ 1.4102e+00, -5.4912e-01, -1.0784e+00, -1.7385e+00,  1.1818e+00]],

        [[-2.6268e-02, -5.8894e-01,  2.8...68e+00]],

        [[-3.4879e-01,  1.7133e-01,  9.1761e-01, -2.0260e-01, -1.6320e-01]]],
       grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_______________________ test_gatv2_conv[1-5-g5-idtype0] _______________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int32
out_dim = 5, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gatv2_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATv2Conv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:519: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatv2conv.py:288: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01551488>
op = 'add'
lhs_data = tensor([[[ 0.2705,  1.9891,  0.6123,  1.1502, -1.0621]],

        [[-0.8657, -1.5141,  0.6779,  1.6001, -1.9514]],

  ...  1.3996, -1.3605, -0.2892]],

        [[ 1.4117, -1.4259,  1.1109, -0.7142,  0.9708]]],
       grad_fn=<ViewBackward>)
rhs_data = tensor([[[ 0.2705,  1.9891,  0.6123,  1.1502, -1.0621]],

        [[-0.8657, -1.5141,  0.6779,  1.6001, -1.9514]],

        [[-0.6522,  1.4272, -0.0565,  1.0412, -0.3532]]],
       grad_fn=<SliceBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_______________________ test_gatv2_conv[1-5-g5-idtype1] _______________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int64
out_dim = 5, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gatv2_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATv2Conv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:519: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatv2conv.py:288: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE144F71C8>
op = 'add'
lhs_data = tensor([[[-0.8379, -0.5576, -2.4463, -0.0118, -0.1861]],

        [[ 1.4168, -0.7050,  2.4124, -1.1660, -0.7984]],

  ... -2.3412,  1.3344,  0.1921]],

        [[ 0.0425,  1.3176, -1.1753,  1.2575, -0.6275]]],
       grad_fn=<ViewBackward>)
rhs_data = tensor([[[-0.8379, -0.5576, -2.4463, -0.0118, -0.1861]],

        [[ 1.4168, -0.7050,  2.4124, -1.1660, -0.7984]],

        [[ 0.1903,  0.7414,  1.0749,  0.2642,  1.8135]]],
       grad_fn=<SliceBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_______________________ test_gatv2_conv[1-5-g6-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'el': Scheme(shape=(1, 5), dtype=torch.float32), 'er': Scheme(shape=(1, 5), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 5, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gatv2_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATv2Conv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:519: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatv2conv.py:288: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01846B88>
op = 'add'
lhs_data = tensor([[[-0.7497, -0.1481, -0.8256, -0.1280,  0.8366]],

        [[-0.4532, -0.3371, -0.1108, -0.1293,  0.4619]],

  ...  0.4651,  1.1399, -1.3687]],

        [[-1.0048, -0.1007, -0.9975,  0.1151,  0.4724]]],
       grad_fn=<ViewBackward>)
rhs_data = tensor([[[-0.5395, -0.4127,  0.6390,  0.8460,  0.6513]],

        [[ 0.2591, -0.4340,  0.3210, -0.7326,  0.9633]],

  ... -0.7714, -0.2703, -0.0835]],

        [[-0.1452, -0.6599,  0.6557,  0.2352,  1.1718]]],
       grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_______________________ test_gatv2_conv[1-5-g6-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'el': Scheme(shape=(1, 5), dtype=torch.float32), 'er': Scheme(shape=(1, 5), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 5, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gatv2_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATv2Conv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:519: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatv2conv.py:288: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03D5EB88>
op = 'add'
lhs_data = tensor([[[ 2.1320, -1.8957,  1.0553, -0.0278,  1.5026]],

        [[ 3.2985,  0.4263,  0.5315, -0.7771,  0.7964]],

  ...  0.0385,  0.7594,  0.4975]],

        [[ 0.3184,  0.5912, -0.3036, -0.5452, -0.4118]]],
       grad_fn=<ViewBackward>)
rhs_data = tensor([[[-3.7486, -2.1635,  1.3521,  0.4901, -0.2648]],

        [[-0.5974, -0.5042,  0.8162,  1.5071,  0.7871]],

  ...  0.3711, -0.4932, -0.1110]],

        [[ 0.9065,  0.2612,  0.7785,  0.5052,  0.5206]]],
       grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_______________________ test_gatv2_conv[4-1-g0-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'el': Scheme(shap...), 'er': Scheme(shape=(4, 1), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
idtype = torch.int32, out_dim = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gatv2_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATv2Conv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:519: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatv2conv.py:288: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF011EB988>
op = 'add'
lhs_data = tensor([[[-1.5248],
         [ 0.9784],
         [ 1.1883],
         [-0.9808]],

        [[ 1.0447],
         [ 1.542... [ 0.3052]],

        [[-0.0335],
         [ 1.7272],
         [ 1.7558],
         [ 0.4224]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[ 0.4546],
         [ 1.1677],
         [-0.7659],
         [ 1.0554]],

        [[-0.9779],
         [ 1.750... [-0.2529]],

        [[-0.4500],
         [-1.2631],
         [-0.3577],
         [ 0.2579]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_______________________ test_gatv2_conv[4-1-g0-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'el': Scheme(shap...), 'er': Scheme(shape=(4, 1), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
idtype = torch.int64, out_dim = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gatv2_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATv2Conv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:519: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatv2conv.py:288: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0177F288>
op = 'add'
lhs_data = tensor([[[-4.6381],
         [-1.6009],
         [-0.9538],
         [ 0.9968]],

        [[ 4.3718],
         [ 1.643... [ 2.7943]],

        [[-0.3910],
         [-0.5088],
         [ 1.4960],
         [-3.0555]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[ 2.2707],
         [-2.5914],
         [ 2.1770],
         [-3.3759]],

        [[-0.6990],
         [ 2.325... [ 0.4315]],

        [[ 2.0483],
         [-0.4581],
         [ 2.5667],
         [-0.2146]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_______________________ test_gatv2_conv[4-1-g1-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'el': Scheme(shape=(4, 1), dtype=torch.float32), 'er': Scheme(shape=(4, 1), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gatv2_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATv2Conv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:519: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatv2conv.py:288: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE146ACE88>
op = 'add'
lhs_data = tensor([[[ 0.8599],
         [ 0.1408],
         [-0.1989],
         [ 0.6758]],

        [[-0.5660],
         [ 0.099... [ 0.0580]],

        [[-0.4466],
         [-0.0613],
         [-1.3729],
         [ 0.6770]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[ 1.0362],
         [-0.4959],
         [-0.1021],
         [-0.8820]],

        [[ 0.3601],
         [-1.286... [ 1.2704]],

        [[-0.2611],
         [-0.1512],
         [-0.5369],
         [ 0.0809]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_______________________ test_gatv2_conv[4-1-g1-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'el': Scheme(shape=(4, 1), dtype=torch.float32), 'er': Scheme(shape=(4, 1), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gatv2_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATv2Conv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:519: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatv2conv.py:288: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14626208>
op = 'add'
lhs_data = tensor([[[-0.9467],
         [ 2.4755],
         [-1.2045],
         [-0.0873]],

        [[ 1.2653],
         [ 0.896... [ 0.7211]],

        [[-1.3514],
         [-1.5800],
         [ 0.5715],
         [ 0.7125]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[-1.5102e+00],
         [ 6.1391e+00],
         [ 4.4997e+00],
         [ 3.7628e+00]],

        [[ 3.8590e-0...      [[-1.4777e+00],
         [-3.3381e+00],
         [-1.9469e+00],
         [ 1.8677e+00]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_______________________ test_gatv2_conv[4-1-g2-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'el': Scheme(shap...r': Scheme(shape=(4, 1), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int32, out_dim = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gatv2_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATv2Conv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:519: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatv2conv.py:288: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03D17248>
op = 'add'
lhs_data = tensor([[[-0.2340],
         [-1.3897],
         [-2.0290],
         [-2.3646]],

        [[-0.5522],
         [ 0.385... [-0.6346]],

        [[ 1.2220],
         [ 1.2213],
         [-1.6070],
         [ 4.4519]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[-1.4340],
         [-0.3572],
         [-0.1966],
         [ 3.0217]],

        [[ 0.1319],
         [ 0.145... [ 4.5702]],

        [[-0.3333],
         [ 0.1228],
         [ 0.9334],
         [ 0.2400]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_______________________ test_gatv2_conv[4-1-g2-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'el': Scheme(shap...r': Scheme(shape=(4, 1), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int64, out_dim = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gatv2_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATv2Conv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:519: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatv2conv.py:288: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF016C4408>
op = 'add'
lhs_data = tensor([[[ 0.8077],
         [ 0.5897],
         [-0.9183],
         [-0.4936]],

        [[-0.9822],
         [-0.182... [-1.5296]],

        [[ 0.3541],
         [-0.1692],
         [ 1.9746],
         [ 0.0685]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[ 1.0482],
         [-1.9302],
         [ 2.2859],
         [ 3.3255]],

        [[-0.8709],
         [ 0.995... [-0.4375]],

        [[-1.4299],
         [ 0.4609],
         [-1.2054],
         [-1.3472]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_______________________ test_gatv2_conv[4-1-g3-idtype0] _______________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'el': Scheme(shape=(4, 1), dtype=torch.float32), 'er': Scheme(shape=(4, 1), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gatv2_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATv2Conv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:519: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatv2conv.py:288: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF016451C8>
op = 'add'
lhs_data = tensor([[[-0.7441],
         [ 0.7704],
         [-0.2934],
         [-0.5103]],

        [[ 0.2458],
         [-0.843... [ 2.4058]],

        [[ 0.6144],
         [ 0.1958],
         [-0.4387],
         [ 0.0476]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[ 1.1983],
         [ 0.3469],
         [ 0.8223],
         [-0.6551]],

        [[-0.0572],
         [-0.573... [-0.5080]],

        [[-0.6873],
         [ 0.3625],
         [-1.5189],
         [-0.1705]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_______________________ test_gatv2_conv[4-1-g3-idtype1] _______________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'el': Scheme(shape=(4, 1), dtype=torch.float32), 'er': Scheme(shape=(4, 1), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gatv2_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATv2Conv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:519: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatv2conv.py:288: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF015BDE88>
op = 'add'
lhs_data = tensor([[[ 2.4006],
         [-3.9147],
         [-0.3165],
         [ 0.6101]],

        [[ 4.2568],
         [-3.100... [-0.5855]],

        [[-0.8803],
         [ 2.0410],
         [-0.6470],
         [-1.8181]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[-2.4270],
         [ 4.3647],
         [-1.4281],
         [ 0.0511]],

        [[-0.5143],
         [-0.792... [ 1.2132]],

        [[-1.3485],
         [ 1.1077],
         [-0.8536],
         [ 0.6802]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_______________________ test_gatv2_conv[4-1-g4-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'el': Scheme(shape=(4, 1), dtype=torch.float32), 'er': Scheme(shape=(4, 1), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gatv2_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATv2Conv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:519: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatv2conv.py:288: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01137A48>
op = 'add'
lhs_data = tensor([[[ 0.6834],
         [ 1.6342],
         [ 0.7665],
         [-1.7392]],

        [[-2.7071],
         [-0.737... [-0.7066]],

        [[-0.6001],
         [-0.5506],
         [ 0.3813],
         [-1.0512]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[ 1.2763],
         [ 3.2095],
         [-0.6509],
         [-0.9248]],

        [[-0.3592],
         [-1.073... [ 0.7559]],

        [[-0.2009],
         [-0.4086],
         [ 0.7775],
         [ 0.8661]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_______________________ test_gatv2_conv[4-1-g4-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'el': Scheme(shape=(4, 1), dtype=torch.float32), 'er': Scheme(shape=(4, 1), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gatv2_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATv2Conv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:519: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatv2conv.py:288: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0171E208>
op = 'add'
lhs_data = tensor([[[-1.0918],
         [-1.7975],
         [-0.7189],
         [-0.1123]],

        [[ 1.1350],
         [-3.334... [ 1.0172]],

        [[ 0.1344],
         [ 2.7822],
         [ 2.3481],
         [-3.3021]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[-0.7589],
         [-0.2461],
         [ 0.4738],
         [ 0.1189]],

        [[-0.4025],
         [ 1.351... [ 1.2861]],

        [[ 0.8375],
         [ 1.4122],
         [-1.1770],
         [ 0.7791]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_______________________ test_gatv2_conv[4-1-g5-idtype0] _______________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int32
out_dim = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gatv2_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATv2Conv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:519: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatv2conv.py:288: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF011963C8>
op = 'add'
lhs_data = tensor([[[ 0.0148],
         [-0.8263],
         [-0.9582],
         [-1.2691]],

        [[ 0.2637],
         [-1.662... [ 1.4016]],

        [[ 1.4606],
         [ 1.4965],
         [-1.1859],
         [ 0.2376]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[ 0.0148],
         [-0.8263],
         [-0.9582],
         [-1.2691]],

        [[ 0.2637],
         [-1.662...[-1.8439]],

        [[-0.5716],
         [ 1.0494],
         [-0.3590],
         [-2.3777]]], grad_fn=<SliceBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_______________________ test_gatv2_conv[4-1-g5-idtype1] _______________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int64
out_dim = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gatv2_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATv2Conv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:519: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatv2conv.py:288: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1498F548>
op = 'add'
lhs_data = tensor([[[-1.1904],
         [ 0.6946],
         [ 0.9320],
         [-0.3088]],

        [[-0.9415],
         [ 0.141... [-2.3715]],

        [[ 3.6450],
         [-0.3156],
         [-0.3904],
         [-1.0310]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[-1.1904],
         [ 0.6946],
         [ 0.9320],
         [-0.3088]],

        [[-0.9415],
         [ 0.141...[ 2.4069]],

        [[ 1.0316],
         [-0.3703],
         [-1.1103],
         [ 0.4166]]], grad_fn=<SliceBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_______________________ test_gatv2_conv[4-1-g6-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'el': Scheme(shape=(4, 1), dtype=torch.float32), 'er': Scheme(shape=(4, 1), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gatv2_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATv2Conv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:519: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatv2conv.py:288: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01848288>
op = 'add'
lhs_data = tensor([[[ 0.0593],
         [ 0.2462],
         [ 0.0334],
         [-3.8041]],

        [[-0.7625],
         [ 1.966... [ 0.7571]],

        [[-0.2741],
         [-0.1052],
         [-1.5081],
         [-0.0471]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[ 0.9796],
         [ 0.1691],
         [ 1.7341],
         [-0.9623]],

        [[ 1.3113],
         [-0.460... [-0.8449]],

        [[ 0.2877],
         [ 0.0187],
         [-1.1389],
         [-0.7043]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):

            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_______________________ test_gatv2_conv[4-1-g6-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'el': Scheme(shape=(4, 1), dtype=torch.float32), 'er': Scheme(shape=(4, 1), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gatv2_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATv2Conv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:519: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatv2conv.py:288: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE149BA748>
op = 'add'
lhs_data = tensor([[[ 1.0022],
         [-0.5406],
         [ 0.0723],
         [-1.0037]],

        [[ 1.0301],
         [-0.958... [-1.1433]],

        [[ 0.0571],
         [ 0.3238],
         [-0.8948],
         [-1.6934]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[ 7.3543e-01],
         [-7.7134e-01],
         [ 3.8357e+00],
         [-1.8321e+00]],

        [[ 6.0420e-0...      [[-1.4584e-01],
         [-1.8161e+00],
         [-5.3821e-01],
         [-4.2008e-01]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_______________________ test_gatv2_conv[4-5-g0-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'el': Scheme(shap...), 'er': Scheme(shape=(4, 5), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
idtype = torch.int32, out_dim = 5, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gatv2_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATv2Conv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:519: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatv2conv.py:288: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01082788>
op = 'add'
lhs_data = tensor([[[ 0.0613, -0.6849, -0.1974,  0.1149, -0.2255],
         [ 0.8051, -0.1104, -0.3701,  0.0440,  0.4101],
      ...4, -0.6162, -1.0289,  0.8139],
         [-0.9716, -0.1513, -0.6588,  0.7798,  0.6332]]],
       grad_fn=<ViewBackward>)
rhs_data = tensor([[[ 6.2159e-01, -7.8696e-01,  9.8142e-01, -5.5864e-01,  1.4125e-02],
         [-1.8441e-02,  1.6137e+00,  4.140...8434e-01],
         [ 6.6451e-03, -2.4498e-01,  1.1806e+00, -8.6186e-01, -2.2997e-01]]],
       grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_______________________ test_gatv2_conv[4-5-g0-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'el': Scheme(shap...), 'er': Scheme(shape=(4, 5), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
idtype = torch.int64, out_dim = 5, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gatv2_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATv2Conv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:519: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatv2conv.py:288: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE149D61C8>
op = 'add'
lhs_data = tensor([[[-6.7886e-01,  1.9101e-01, -1.3684e+00, -5.3866e-01,  4.0147e-01],
         [ 1.4634e+00,  1.3319e+00,  1.200...7418e+00],
         [ 1.2091e+00,  2.9386e-01, -8.6307e-01, -1.5170e+00, -8.0632e-01]]],
       grad_fn=<ViewBackward>)
rhs_data = tensor([[[-1.6448, -0.3274,  0.2648,  1.7123, -0.7531],
         [ 2.1541, -1.4751,  1.9203, -1.6337, -2.5995],
      ...6,  0.7806,  0.8621,  0.5614],
         [ 2.6895,  0.2919,  1.0267,  0.0368, -0.5082]]],
       grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_______________________ test_gatv2_conv[4-5-g1-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'el': Scheme(shape=(4, 5), dtype=torch.float32), 'er': Scheme(shape=(4, 5), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 5, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gatv2_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATv2Conv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:519: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatv2conv.py:288: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0183B1C8>
op = 'add'
lhs_data = tensor([[[-0.5743,  0.2410, -0.2388,  0.1787, -0.1523],
         [ 0.3622,  0.5593,  0.3596,  0.6419,  0.4960],
      ...4, -0.3340, -0.5994, -0.4656],
         [-0.8027,  0.5380, -0.0445,  0.2114,  0.4394]]],
       grad_fn=<ViewBackward>)
rhs_data = tensor([[[ 0.6225,  0.0318, -0.9630,  0.9272, -0.6198],
         [-0.4992,  0.4227,  0.1694, -0.0806, -0.0693],
      ...3,  0.3578, -0.6495, -0.0965],
         [-0.2865, -0.2395, -0.7947,  0.3286, -0.1745]]],
       grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_______________________ test_gatv2_conv[4-5-g1-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'el': Scheme(shape=(4, 5), dtype=torch.float32), 'er': Scheme(shape=(4, 5), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 5, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gatv2_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATv2Conv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:519: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatv2conv.py:288: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14A66F48>
op = 'add'
lhs_data = tensor([[[-0.3582,  0.6004, -0.4412, -0.7930,  0.6868],
         [ 0.8057, -0.1199, -0.5953, -0.4729, -0.4331],
      ...5, -0.9140, -0.4654, -0.4907],
         [ 0.0962, -0.7285, -0.8933, -0.0422,  0.7941]]],
       grad_fn=<ViewBackward>)
rhs_data = tensor([[[ 0.0247,  0.5473,  0.1720, -0.5313, -0.0664],
         [-0.3336, -0.9624,  0.2008, -0.9940,  0.6640],
      ...2, -1.2648,  0.9912,  0.2205],
         [-0.2759,  1.0096, -1.2643, -0.5249,  0.2208]]],
       grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_______________________ test_gatv2_conv[4-5-g2-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'el': Scheme(shap...r': Scheme(shape=(4, 5), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int32, out_dim = 5, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gatv2_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATv2Conv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:519: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatv2conv.py:288: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14993048>
op = 'add'
lhs_data = tensor([[[-1.5317e-01,  3.6508e-01,  4.8938e-02, -6.3920e-01,  4.2037e-01],
         [-6.2535e-01,  5.7798e-01, -2.148...4254e-02],
         [-1.5052e-01, -7.7757e-01,  2.1689e-03, -4.5364e-01, -1.0242e+00]]],
       grad_fn=<ViewBackward>)
rhs_data = tensor([[[-0.4871,  0.2181, -0.8171, -0.0991, -0.9024],
         [ 0.2800,  0.0137, -0.3549,  0.6533,  0.3550],
      ...4,  0.2325,  0.1351, -0.0360],
         [ 0.2115,  1.2571,  0.5301,  0.2603,  0.1958]]],
       grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_______________________ test_gatv2_conv[4-5-g2-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'el': Scheme(shap...r': Scheme(shape=(4, 5), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int64, out_dim = 5, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gatv2_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATv2Conv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:519: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatv2conv.py:288: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01166EC8>
op = 'add'
lhs_data = tensor([[[-0.2758,  0.1867,  2.0641, -0.4528,  0.3045],
         [ 0.3652,  1.5035, -0.3147, -1.8376, -0.2442],
      ...5, -1.3744, -0.6990,  0.0457],
         [ 0.5195,  0.7881,  1.7572,  0.1334, -0.2182]]],
       grad_fn=<ViewBackward>)
rhs_data = tensor([[[-8.0919e-02,  3.4624e-01, -1.2255e+00,  1.2169e+00, -8.6059e-01],
         [ 5.4277e-01, -4.2840e-01, -1.261...7956e-01],
         [ 8.3811e-01, -1.6032e+00,  3.1493e+00, -5.6948e-01,  2.9192e-01]]],
       grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_______________________ test_gatv2_conv[4-5-g3-idtype0] _______________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'el': Scheme(shape=(4, 5), dtype=torch.float32), 'er': Scheme(shape=(4, 5), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 5, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gatv2_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATv2Conv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:519: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatv2conv.py:288: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0775EB08>
op = 'add'
lhs_data = tensor([[[-8.0491e-03, -2.3519e-01,  1.8379e-01,  2.4737e-01, -4.8001e-03],
         [ 2.9900e-01,  4.8019e-01, -5.735...3996e-01],
         [ 1.4392e-01,  1.3692e-01, -9.4073e-02, -1.0845e+00,  1.2711e+00]]],
       grad_fn=<ViewBackward>)
rhs_data = tensor([[[ 0.1073, -0.0926,  0.2591,  0.1143, -0.1976],
         [ 0.1019,  0.2044,  0.4441,  0.2075, -0.2089],
      ...1, -0.7045,  0.2833,  1.4744],
         [ 0.2011,  0.6529, -0.8639, -1.3477, -1.1950]]],
       grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_______________________ test_gatv2_conv[4-5-g3-idtype1] _______________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'el': Scheme(shape=(4, 5), dtype=torch.float32), 'er': Scheme(shape=(4, 5), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 5, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gatv2_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATv2Conv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:519: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatv2conv.py:288: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01799F88>
op = 'add'
lhs_data = tensor([[[-0.2016, -0.5120,  0.0536,  0.6993, -1.5782],
         [ 0.3968, -0.7288,  0.5348,  0.8288,  1.7559],
      ...1,  0.6992,  1.3581,  1.3295],
         [ 1.0578,  0.3520, -0.9984, -1.1056, -0.5247]]],
       grad_fn=<ViewBackward>)
rhs_data = tensor([[[-3.4696e-01, -2.1659e-01,  5.5323e-01, -1.8265e-02,  2.1405e-01],
         [ 2.9571e-01, -6.2857e-01, -8.337...9140e-02],
         [-1.9953e-01, -2.2892e-01, -6.2482e-01, -2.5903e-01, -1.5734e+00]]],
       grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_______________________ test_gatv2_conv[4-5-g4-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'el': Scheme(shape=(4, 5), dtype=torch.float32), 'er': Scheme(shape=(4, 5), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 5, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gatv2_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATv2Conv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:519: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatv2conv.py:288: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF018B1088>
op = 'add'
lhs_data = tensor([[[-0.0816,  1.1359,  0.5987, -1.3428,  1.3491],
         [ 0.9460,  0.4619,  0.5998,  1.0034,  0.1466],
      ...1, -2.8400, -0.8883, -1.9857],
         [ 0.1402, -0.4429,  0.7958,  0.8786, -0.2275]]],
       grad_fn=<ViewBackward>)
rhs_data = tensor([[[ 2.6757, -1.1637,  0.7641, -0.5051, -0.6453],
         [ 2.3930, -1.7506,  2.7822, -0.7707,  1.1098],
      ...3, -3.7213,  0.1611,  0.1470],
         [-0.3553,  0.9276, -0.8301, -0.7092,  1.3590]]],
       grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_______________________ test_gatv2_conv[4-5-g4-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'el': Scheme(shape=(4, 5), dtype=torch.float32), 'er': Scheme(shape=(4, 5), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 5, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gatv2_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATv2Conv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:519: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatv2conv.py:288: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE145DCD08>
op = 'add'
lhs_data = tensor([[[-1.6312e-01,  4.7990e-01, -5.3793e-01, -3.8500e-01, -2.7367e-01],
         [-4.8571e-01,  2.4375e-01,  5.235...1877e-01],
         [-1.0897e+00,  4.4055e-01, -4.0708e-01, -9.0817e-02, -4.5936e-01]]],
       grad_fn=<ViewBackward>)
rhs_data = tensor([[[-0.7214, -0.3112, -0.0996,  0.6509,  0.0337],
         [-0.5666, -0.4122, -0.3727, -0.3229,  0.3087],
      ...4, -0.3223, -0.3083,  0.9503],
         [ 0.2059,  0.0338,  0.1221, -0.3473,  0.5555]]],
       grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_______________________ test_gatv2_conv[4-5-g5-idtype0] _______________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int32
out_dim = 5, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gatv2_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATv2Conv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:519: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatv2conv.py:288: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF014C5D48>
op = 'add'
lhs_data = tensor([[[-0.2378, -0.4710, -0.7136,  0.9966, -1.6844],
         [-0.9419,  0.6565, -0.2314, -0.1451,  1.5155],
      ...8,  0.1321, -0.5014,  0.4384],
         [ 0.2207, -0.3146,  0.2370, -0.9194,  0.5117]]],
       grad_fn=<ViewBackward>)
rhs_data = tensor([[[-0.2378, -0.4710, -0.7136,  0.9966, -1.6844],
         [-0.9419,  0.6565, -0.2314, -0.1451,  1.5155],
      ..., -1.1368, -0.3990,  0.2568],
         [-0.7300,  0.0355, -1.9954, -0.4114,  0.2230]]],
       grad_fn=<SliceBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_______________________ test_gatv2_conv[4-5-g5-idtype1] _______________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int64
out_dim = 5, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gatv2_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATv2Conv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:519: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatv2conv.py:288: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01624148>
op = 'add'
lhs_data = tensor([[[-0.0448, -0.5802,  0.5853,  1.5337, -0.0938],
         [-0.5366, -0.3445,  0.8199,  0.1596, -0.0936],
      ...7,  0.7534,  1.3606, -1.3052],
         [-2.3090, -0.4686,  0.6032, -1.3555, -3.0084]]],
       grad_fn=<ViewBackward>)
rhs_data = tensor([[[-0.0448, -0.5802,  0.5853,  1.5337, -0.0938],
         [-0.5366, -0.3445,  0.8199,  0.1596, -0.0936],
      ...,  1.4217, -0.2796, -0.8359],
         [-0.9235, -0.2591,  0.3156, -0.4141, -1.4557]]],
       grad_fn=<SliceBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_______________________ test_gatv2_conv[4-5-g6-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'el': Scheme(shape=(4, 5), dtype=torch.float32), 'er': Scheme(shape=(4, 5), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 5, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gatv2_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATv2Conv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:519: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatv2conv.py:288: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF016C50C8>
op = 'add'
lhs_data = tensor([[[-0.8917,  0.9453,  0.1839, -2.4546,  0.3208],
         [-1.3204,  1.9003,  0.3560,  0.4615,  1.2873],
      ...5,  1.2446, -1.1236,  0.3823],
         [ 0.3088, -1.6394, -1.4025,  0.4820,  0.1566]]],
       grad_fn=<ViewBackward>)
rhs_data = tensor([[[ 0.7318, -0.4621,  0.4455,  1.0447, -1.5532],
         [ 2.2933,  0.8608,  1.0303,  0.8239,  0.3713],
      ...7, -0.4553,  2.0909, -0.8320],
         [ 0.6100,  0.9100,  1.4306, -0.2873,  0.6267]]],
       grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_______________________ test_gatv2_conv[4-5-g6-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'el': Scheme(shape=(4, 5), dtype=torch.float32), 'er': Scheme(shape=(4, 5), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 5, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gatv2_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATv2Conv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:519: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatv2conv.py:288: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF018D0D88>
op = 'add'
lhs_data = tensor([[[ 2.7309,  0.1083,  0.9902, -0.2375,  1.5097],
         [ 0.5685, -1.3686, -1.8885, -0.0718,  0.5320],
      ...6,  0.3441, -0.3591, -0.2271],
         [ 0.3779, -0.6857, -0.8396, -0.0035, -0.2674]]],
       grad_fn=<ViewBackward>)
rhs_data = tensor([[[ 2.4341e+00, -7.9033e-01,  3.4665e+00, -1.1424e+00, -1.7548e-01],
         [ 1.6833e+00,  5.3114e-01,  2.301...5984e+00],
         [ 6.3434e-01,  1.2100e-01,  3.0803e-01,  6.8915e-01, -1.0221e+00]]],
       grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_____________________ test_gatv2_conv_bi[1-1-g0-idtype0] ______________________

g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int32, out_dim = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gatv2_conv_bi(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATv2Conv(5, out_dim, num_heads)
        feat = (F.randn((g.number_of_src_nodes(), 5)), F.randn((g.number_of_dst_nodes(), 5)))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:543: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatv2conv.py:288: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF018C4248>
op = 'add'
lhs_data = tensor([[[0.0133]],

        [[2.2480]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[-1.7623]],

        [[ 0.7648]],

        [[-2.9369]],

        [[ 0.5808]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_____________________ test_gatv2_conv_bi[1-1-g0-idtype1] ______________________

g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int64, out_dim = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gatv2_conv_bi(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATv2Conv(5, out_dim, num_heads)
        feat = (F.randn((g.number_of_src_nodes(), 5)), F.randn((g.number_of_dst_nodes(), 5)))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:543: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatv2conv.py:288: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03F0D308>
op = 'add'
lhs_data = tensor([[[-0.9302]],

        [[ 2.4272]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[-0.8574]],

        [[ 1.3571]],

        [[ 1.0753]],

        [[-2.6492]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_____________________ test_gatv2_conv_bi[1-1-g1-idtype0] ______________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int32
out_dim = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gatv2_conv_bi(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATv2Conv(5, out_dim, num_heads)
        feat = (F.randn((g.number_of_src_nodes(), 5)), F.randn((g.number_of_dst_nodes(), 5)))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:543: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatv2conv.py:288: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0109F8C8>
op = 'add'
lhs_data = tensor([[[ 1.2686]],

        [[ 0.4787]],

        [[ 0.7578]],

        [[ 0.5825]],

        [[ 0.8260]],

        [[-2.4887]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[-2.7969]],

        [[-0.7482]],

        [[-0.9773]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_____________________ test_gatv2_conv_bi[1-1-g1-idtype1] ______________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int64
out_dim = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gatv2_conv_bi(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATv2Conv(5, out_dim, num_heads)
        feat = (F.randn((g.number_of_src_nodes(), 5)), F.randn((g.number_of_dst_nodes(), 5)))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:543: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatv2conv.py:288: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03CF37C8>
op = 'add'
lhs_data = tensor([[[-1.5194]],

        [[ 0.1584]],

        [[ 0.3115]],

        [[ 1.6498]],

        [[-2.8148]],

        [[ 0.1842]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[ 2.7564]],

        [[ 3.7364]],

        [[-1.3609]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_____________________ test_gatv2_conv_bi[1-2-g0-idtype0] ______________________

g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int32, out_dim = 2, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gatv2_conv_bi(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATv2Conv(5, out_dim, num_heads)
        feat = (F.randn((g.number_of_src_nodes(), 5)), F.randn((g.number_of_dst_nodes(), 5)))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:543: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatv2conv.py:288: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF016A8908>
op = 'add'
lhs_data = tensor([[[-0.6889, -0.2673]],

        [[-0.7297,  1.5759]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[-0.0778,  1.1332]],

        [[ 1.6522, -0.1728]],

        [[ 0.6892, -1.2536]],

        [[-0.8748,  0.0889]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_____________________ test_gatv2_conv_bi[1-2-g0-idtype1] ______________________

g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int64, out_dim = 2, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gatv2_conv_bi(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATv2Conv(5, out_dim, num_heads)
        feat = (F.randn((g.number_of_src_nodes(), 5)), F.randn((g.number_of_dst_nodes(), 5)))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:543: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatv2conv.py:288: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF030A6448>
op = 'add'
lhs_data = tensor([[[-1.8266,  3.5786]],

        [[ 0.6260, -2.3363]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[-1.0211, -1.3839]],

        [[ 0.5972, -1.4325]],

        [[-2.2996, -1.1856]],

        [[-1.9642,  0.6281]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_____________________ test_gatv2_conv_bi[1-2-g1-idtype0] ______________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int32
out_dim = 2, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gatv2_conv_bi(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATv2Conv(5, out_dim, num_heads)
        feat = (F.randn((g.number_of_src_nodes(), 5)), F.randn((g.number_of_dst_nodes(), 5)))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:543: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatv2conv.py:288: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01514108>
op = 'add'
lhs_data = tensor([[[ 2.7550e-01,  3.7824e+00]],

        [[ 1.4122e+00, -2.6246e+00]],

        [[-1.6688e-03, -1.6573e-01]],

 ...,  1.0081e+00]],

        [[-1.0737e+00,  9.0269e-01]],

        [[ 3.2830e-01,  1.0958e+00]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[-1.2388, -0.9786]],

        [[ 1.7366,  1.6057]],

        [[ 0.2182,  0.3972]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_____________________ test_gatv2_conv_bi[1-2-g1-idtype1] ______________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int64
out_dim = 2, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gatv2_conv_bi(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATv2Conv(5, out_dim, num_heads)
        feat = (F.randn((g.number_of_src_nodes(), 5)), F.randn((g.number_of_dst_nodes(), 5)))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:543: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatv2conv.py:288: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14628988>
op = 'add'
lhs_data = tensor([[[-0.0693,  1.9331]],

        [[-2.2001, -3.9850]],

        [[ 0.7933,  1.0276]],

        [[ 2.6458,  0.9442]],

        [[ 5.0171,  6.0013]],

        [[-2.1558, -1.8875]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[ 1.4751,  0.9986]],

        [[-0.1330,  0.3691]],

        [[ 1.6986,  2.0056]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_____________________ test_gatv2_conv_bi[4-1-g0-idtype0] ______________________

g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int32, out_dim = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gatv2_conv_bi(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATv2Conv(5, out_dim, num_heads)
        feat = (F.randn((g.number_of_src_nodes(), 5)), F.randn((g.number_of_dst_nodes(), 5)))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:543: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatv2conv.py:288: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01106888>
op = 'add'
lhs_data = tensor([[[ 0.2608],
         [-0.4148],
         [ 0.9681],
         [-1.8057]],

        [[-0.7771],
         [ 3.3052],
         [ 0.6150],
         [-1.4448]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[-0.7026],
         [-0.7191],
         [ 2.0442],
         [-1.2088]],

        [[ 0.6318],
         [-1.269... [-2.1726]],

        [[-1.0498],
         [ 0.2173],
         [ 1.4098],
         [ 2.7376]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_____________________ test_gatv2_conv_bi[4-1-g0-idtype1] ______________________

g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int64, out_dim = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gatv2_conv_bi(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATv2Conv(5, out_dim, num_heads)
        feat = (F.randn((g.number_of_src_nodes(), 5)), F.randn((g.number_of_dst_nodes(), 5)))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:543: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatv2conv.py:288: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE145DC2C8>
op = 'add'
lhs_data = tensor([[[-0.0259],
         [-1.6666],
         [ 2.8537],
         [-0.8805]],

        [[-1.0537],
         [-0.8144],
         [ 1.3610],
         [-0.2527]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[ 1.1424],
         [ 0.7351],
         [ 1.0310],
         [-0.0480]],

        [[-2.4787],
         [-0.478... [ 0.1927]],

        [[-2.0536],
         [-0.6037],
         [ 0.7965],
         [ 0.1536]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_____________________ test_gatv2_conv_bi[4-1-g1-idtype0] ______________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int32
out_dim = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gatv2_conv_bi(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATv2Conv(5, out_dim, num_heads)
        feat = (F.randn((g.number_of_src_nodes(), 5)), F.randn((g.number_of_dst_nodes(), 5)))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:543: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatv2conv.py:288: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF016BD588>
op = 'add'
lhs_data = tensor([[[-0.5725],
         [ 2.8325],
         [-1.4587],
         [ 0.4731]],

        [[-0.3568],
         [-0.701... [ 1.5426]],

        [[ 0.5490],
         [-0.4877],
         [ 0.0740],
         [-1.8241]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[-0.6791],
         [ 2.0268],
         [ 1.3207],
         [ 1.1879]],

        [[ 0.5816],
         [ 0.796... [-0.5743]],

        [[-0.6853],
         [-1.1905],
         [ 2.2661],
         [-2.3645]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_____________________ test_gatv2_conv_bi[4-1-g1-idtype1] ______________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int64
out_dim = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gatv2_conv_bi(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATv2Conv(5, out_dim, num_heads)
        feat = (F.randn((g.number_of_src_nodes(), 5)), F.randn((g.number_of_dst_nodes(), 5)))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:543: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatv2conv.py:288: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE146A7648>
op = 'add'
lhs_data = tensor([[[ 3.2302],
         [ 1.0438],
         [-2.8922],
         [-1.0383]],

        [[-0.9124],
         [-0.309... [ 0.3991]],

        [[ 1.6514],
         [-1.1843],
         [ 1.1989],
         [-0.3002]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[ 0.5667],
         [-0.4020],
         [-0.6449],
         [-0.1821]],

        [[-2.7037],
         [ 1.015... [ 0.9937]],

        [[-1.1280],
         [ 0.4077],
         [ 0.6489],
         [-0.4894]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_____________________ test_gatv2_conv_bi[4-2-g0-idtype0] ______________________

g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int32, out_dim = 2, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gatv2_conv_bi(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATv2Conv(5, out_dim, num_heads)
        feat = (F.randn((g.number_of_src_nodes(), 5)), F.randn((g.number_of_dst_nodes(), 5)))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:543: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatv2conv.py:288: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF018A01C8>
op = 'add'
lhs_data = tensor([[[-0.2637, -0.8810],
         [-0.2206, -0.1826],
         [ 0.0475,  0.5577],
         [ 1.3546, -2.1613]],

...1825],
         [ 1.8888, -0.3567],
         [-0.4560,  1.8093],
         [ 0.7429,  0.6310]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[ 0.0463,  0.6432],
         [-0.6608, -0.3124],
         [ 1.3438,  0.3285],
         [ 1.5787,  0.6057]],

...5552],
         [ 0.8361, -0.0426],
         [-1.7327,  0.7641],
         [-1.4477,  0.9004]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_____________________ test_gatv2_conv_bi[4-2-g0-idtype1] ______________________

g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int64, out_dim = 2, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gatv2_conv_bi(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATv2Conv(5, out_dim, num_heads)
        feat = (F.randn((g.number_of_src_nodes(), 5)), F.randn((g.number_of_dst_nodes(), 5)))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:543: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatv2conv.py:288: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF015BF888>
op = 'add'
lhs_data = tensor([[[-1.0723,  1.3507],
         [-0.2587,  0.1481],
         [ 0.1638, -0.0823],
         [ 1.0906, -0.1256]],

...2272],
         [-2.7643, -0.7742],
         [ 1.6863,  1.1412],
         [ 0.8595, -1.9046]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[-2.0025,  0.7664],
         [-0.7697, -0.1440],
         [-1.0659,  1.0813],
         [ 1.2918,  1.1124]],

...1185],
         [ 0.4473,  0.5248],
         [ 0.2229, -1.7197],
         [ 1.5281,  2.9379]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_____________________ test_gatv2_conv_bi[4-2-g1-idtype0] ______________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int32
out_dim = 2, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gatv2_conv_bi(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATv2Conv(5, out_dim, num_heads)
        feat = (F.randn((g.number_of_src_nodes(), 5)), F.randn((g.number_of_dst_nodes(), 5)))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:543: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatv2conv.py:288: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01158BC8>
op = 'add'
lhs_data = tensor([[[ 3.1099, -0.4278],
         [-0.0066, -1.1153],
         [-2.0509, -0.3534],
         [ 0.5579, -1.2069]],

...5524],
         [-0.0546,  1.4579],
         [ 0.4669,  0.1831],
         [ 0.6552, -1.0917]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[-2.0707, -0.0367],
         [-0.6861, -1.9129],
         [-3.4629,  1.5843],
         [ 1.0048, -0.5094]],

...0521],
         [ 0.0683,  0.3065],
         [ 0.2424, -0.8365],
         [ 0.2386, -0.0861]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_____________________ test_gatv2_conv_bi[4-2-g1-idtype1] ______________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int64
out_dim = 2, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_gatv2_conv_bi(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gat = nn.GATv2Conv(5, out_dim, num_heads)
        feat = (F.randn((g.number_of_src_nodes(), 5)), F.randn((g.number_of_dst_nodes(), 5)))
        gat = gat.to(ctx)
>       h = gat(g, feat)

tests\pytorch\test_nn.py:543: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatv2conv.py:288: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE149BA0C8>
op = 'add'
lhs_data = tensor([[[ 0.8851, -1.0254],
         [-0.1103, -1.0770],
         [ 0.5852, -2.0093],
         [-1.2819, -0.4825]],

...4015],
         [ 0.7568, -0.6128],
         [ 1.2711,  0.9593],
         [ 1.4093, -0.6879]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[ 3.2303,  0.2029],
         [-0.6922,  4.7644],
         [ 2.1568, -0.4187],
         [ 0.6550,  3.5110]],

...0094],
         [ 1.2285,  4.7648],
         [-0.4397, -1.8284],
         [ 0.3382,  3.4759]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[1-1-1-g0-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'f_ni': Scheme(sh...), 'f_nj': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
idtype = torch.int32, out_node_feats = 1, out_edge_feats = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF015BA188>
op = 'add'
lhs_data = tensor([[-1.9837],
        [-1.1596],
        [-1.2099],
        [ 2.1841],
        [ 1.0622],
        [-2.0678],
        [ 1.0281],
        [-0.2626],
        [-2.3762],
        [-1.4083]], grad_fn=<MmBackward>)
rhs_data = tensor([[ 1.2065],
        [-1.3186],
        [ 0.5730],
        [ 0.7216],
        [-2.4841],
        [ 1.2998],
        [ 0.4063],
        [-0.8652],
        [-1.0416],
        [-0.8486]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[1-1-1-g0-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'f_ni': Scheme(sh...), 'f_nj': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
idtype = torch.int64, out_node_feats = 1, out_edge_feats = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF018D0548>
op = 'add'
lhs_data = tensor([[-1.7643],
        [ 1.1458],
        [-3.6263],
        [-0.6129],
        [-0.8150],
        [-0.6306],
        [-2.8359],
        [-3.0163],
        [ 1.6360],
        [-4.0968]], grad_fn=<MmBackward>)
rhs_data = tensor([[-0.2899],
        [-1.6179],
        [ 1.5335],
        [-0.7805],
        [ 0.6817],
        [-0.5212],
        [ 0.6766],
        [ 1.2876],
        [ 0.3656],
        [-0.0324]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[1-1-1-g1-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'f_ni': Scheme(shape=(1,), dtype=torch.float32), 'f_nj': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_node_feats = 1, out_edge_feats = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF015BA288>
op = 'add'
lhs_data = tensor([[ 3.4316],
        [ 1.7854],
        [ 0.4916],
        [-1.4800],
        [-3.0656],
        [ 0.9585],
        [-0.5771],
        [ 1.3040],
        [ 1.7081],
        [-0.6167]], grad_fn=<MmBackward>)
rhs_data = tensor([[ 0.4872],
        [-4.4221],
        [ 0.2213],
        [-0.1164],
        [ 2.1821],
        [-0.5149],
        [ 2.4895],
        [-2.8139],
        [-1.5421],
        [ 2.6373]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[1-1-1-g1-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'f_ni': Scheme(shape=(1,), dtype=torch.float32), 'f_nj': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_node_feats = 1, out_edge_feats = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF019B7C08>
op = 'add'
lhs_data = tensor([[-0.3001],
        [ 0.9268],
        [-1.9912],
        [-0.3428],
        [-1.5073],
        [-0.6777],
        [ 0.5952],
        [ 1.7083],
        [ 1.0207],
        [-0.6972]], grad_fn=<MmBackward>)
rhs_data = tensor([[ 1.5802],
        [ 1.4154],
        [-3.8826],
        [-2.8975],
        [-1.0542],
        [-0.6293],
        [-1.8504],
        [-2.3198],
        [ 2.1504],
        [-0.8563]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[1-1-1-g2-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'f_ni': Scheme(sh..._nj': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int32, out_node_feats = 1, out_edge_feats = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14AB00C8>
op = 'add'
lhs_data = tensor([[ 1.3689],
        [ 1.6992],
        [-0.3913],
        [-0.6812],
        [ 1.5617],
        [-0.1367],
        [-1.4977],
        [-0.3283],
        [-1.8440],
        [-1.2188]], grad_fn=<MmBackward>)
rhs_data = tensor([[-0.4436],
        [ 0.3662],
        [ 0.8571],
        [ 5.8676],
        [-2.7601],
        [ 1.8752],
        [-1.3885],
        [-1.5361],
        [-2.3169],
        [ 0.7906]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[1-1-1-g2-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'f_ni': Scheme(sh..._nj': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int64, out_node_feats = 1, out_edge_feats = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0179D608>
op = 'add'
lhs_data = tensor([[-0.0210],
        [-1.3353],
        [-2.5912],
        [-3.9179],
        [ 1.4500],
        [ 1.4419],
        [ 0.1789],
        [ 0.1028],
        [ 3.1465],
        [ 3.4479]], grad_fn=<MmBackward>)
rhs_data = tensor([[ 1.5270],
        [-3.1317],
        [-0.3421],
        [-1.7025],
        [-1.8789],
        [ 3.1518],
        [-0.5934],
        [ 1.8909],
        [ 0.5270],
        [ 2.5669]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[1-1-1-g3-idtype0] _______________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'f_ni': Scheme(shape=(1,), dtype=torch.float32), 'f_nj': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_node_feats = 1, out_edge_feats = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01948148>
op = 'add'
lhs_data = tensor([[-1.8084],
        [ 1.4252],
        [-2.5860],
        [ 0.9591],
        [ 0.7286],
        [ 0.2962],
        [-1.4975],
        [ 1.4559],
        [ 0.6721]], grad_fn=<MmBackward>)
rhs_data = tensor([[ 0.1702],
        [ 2.0008],
        [-0.0319],
        [-1.7652],
        [ 0.2168],
        [ 2.8688],
        [-3.2648],
        [-2.5955],
        [ 0.5352]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[1-1-1-g3-idtype1] _______________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'f_ni': Scheme(shape=(1,), dtype=torch.float32), 'f_nj': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_node_feats = 1, out_edge_feats = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01993CC8>
op = 'add'
lhs_data = tensor([[ 1.6223],
        [ 0.2371],
        [-0.4917],
        [ 0.7207],
        [ 2.4616],
        [ 3.6629],
        [-1.3165],
        [-1.7199],
        [ 1.6290]], grad_fn=<MmBackward>)
rhs_data = tensor([[ 0.2893],
        [ 1.1330],
        [ 0.0593],
        [ 2.1957],
        [-2.2938],
        [-1.3900],
        [ 0.7139],
        [ 3.9108],
        [ 1.2891]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[1-1-1-g4-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'f_ni': Scheme(shape=(1,), dtype=torch.float32), 'f_nj': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_node_feats = 1, out_edge_feats = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF018B4CC8>
op = 'add'
lhs_data = tensor([[-2.2489],
        [ 0.6977],
        [-0.6810],
        [-0.3992],
        [-0.9549],
        [ 0.0957],
        [ 0.3528],
        [ 0.3547],
        [ 2.3431],
        [-1.2437]], grad_fn=<MmBackward>)
rhs_data = tensor([[ 0.6134],
        [ 0.6300],
        [ 0.4434],
        [-1.4351],
        [-2.4716],
        [ 0.5775],
        [-0.1511],
        [-0.5965],
        [-0.7978],
        [ 0.7124]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[1-1-1-g4-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'f_ni': Scheme(shape=(1,), dtype=torch.float32), 'f_nj': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_node_feats = 1, out_edge_feats = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01854D08>
op = 'add'
lhs_data = tensor([[ 0.3889],
        [ 1.7746],
        [-2.5886],
        [ 0.3332],
        [ 1.4764],
        [ 0.5246],
        [-0.4633],
        [ 3.0800],
        [-3.0837],
        [ 0.5896]], grad_fn=<MmBackward>)
rhs_data = tensor([[ 0.5251],
        [-1.6353],
        [ 3.8769],
        [ 0.3921],
        [ 0.5063],
        [ 0.3722],
        [-0.1645],
        [-3.1691],
        [-0.3126],
        [ 0.6427]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[1-1-1-g5-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'f_ni': Scheme(shape=(1,), dtype=torch.float32), 'f_nj': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_node_feats = 1, out_edge_feats = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01769988>
op = 'add'
lhs_data = tensor([[ 2.5590],
        [ 3.4209],
        [ 2.7960],
        [-1.5737],
        [ 1.3153],
        [ 1.7566],
        [-0.2015],
        [-1.0274],
        [ 0.8919],
        [-1.8365]], grad_fn=<MmBackward>)
rhs_data = tensor([[-1.3269],
        [-1.5947],
        [-1.4248],
        [ 2.1636],
        [-0.7930],
        [ 0.0675],
        [-1.2462],
        [ 1.1313],
        [-1.5790],
        [ 2.4162]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[1-1-1-g5-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'f_ni': Scheme(shape=(1,), dtype=torch.float32), 'f_nj': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_node_feats = 1, out_edge_feats = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01720708>
op = 'add'
lhs_data = tensor([[-1.4828],
        [ 0.0515],
        [ 0.5216],
        [ 2.0078],
        [ 1.7605],
        [-0.2434],
        [-2.0837],
        [ 2.1624],
        [ 0.6064],
        [ 0.2911]], grad_fn=<MmBackward>)
rhs_data = tensor([[-2.3120],
        [-2.7919],
        [-3.0364],
        [-0.3088],
        [-1.4967],
        [-0.4539],
        [ 0.3483],
        [-0.9393],
        [-2.0719],
        [ 4.1374]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[1-1-5-g0-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'f_ni': Scheme(sh...), 'f_nj': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
idtype = torch.int32, out_node_feats = 5, out_edge_feats = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF014A7588>
op = 'add'
lhs_data = tensor([[ 2.2232],
        [ 0.8939],
        [ 0.2908],
        [-0.6565],
        [ 0.0358],
        [-1.6915],
        [ 1.0600],
        [-1.1864],
        [-1.2970],
        [-2.6229]], grad_fn=<MmBackward>)
rhs_data = tensor([[ 2.7841],
        [ 0.0039],
        [ 1.1798],
        [ 0.3332],
        [-0.6660],
        [ 0.0334],
        [ 1.5234],
        [-2.8464],
        [-1.5105],
        [-2.3193]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[1-1-5-g0-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'f_ni': Scheme(sh...), 'f_nj': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
idtype = torch.int64, out_node_feats = 5, out_edge_feats = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF015E8B88>
op = 'add'
lhs_data = tensor([[ 0.3848],
        [ 3.6361],
        [ 0.5994],
        [ 0.0802],
        [ 2.1109],
        [ 0.6610],
        [-1.3027],
        [ 5.7181],
        [ 0.2420],
        [-0.0644]], grad_fn=<MmBackward>)
rhs_data = tensor([[ 2.9723e+00],
        [ 2.0588e+00],
        [-1.5392e+00],
        [-5.2600e-01],
        [-1.6400e-03],
   ...01],
        [-3.1768e+00],
        [ 4.5365e+00],
        [ 7.5092e-01],
        [-2.2005e+00]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[1-1-5-g1-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'f_ni': Scheme(shape=(1,), dtype=torch.float32), 'f_nj': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_node_feats = 5, out_edge_feats = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF011D3C48>
op = 'add'
lhs_data = tensor([[ 0.8083],
        [-0.6315],
        [ 1.5591],
        [-1.5418],
        [-1.3874],
        [ 0.2553],
        [-0.5505],
        [ 0.2332],
        [-1.0462],
        [ 3.4172]], grad_fn=<MmBackward>)
rhs_data = tensor([[ 0.1259],
        [ 0.4014],
        [ 0.8101],
        [ 2.7692],
        [ 1.9802],
        [ 0.4524],
        [ 0.3151],
        [-2.6044],
        [ 0.1397],
        [ 0.2772]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[1-1-5-g1-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'f_ni': Scheme(shape=(1,), dtype=torch.float32), 'f_nj': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_node_feats = 5, out_edge_feats = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14581D88>
op = 'add'
lhs_data = tensor([[2.1847],
        [0.8934],
        [1.5143],
        [1.9313],
        [2.5140],
        [0.1342],
        [2.2550],
        [0.9844],
        [2.3754],
        [1.0901]], grad_fn=<MmBackward>)
rhs_data = tensor([[-0.7138],
        [-0.1913],
        [ 0.9808],
        [ 1.9521],
        [ 1.5255],
        [-0.1336],
        [ 2.1051],
        [-0.0150],
        [ 0.2672],
        [ 2.2029]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[1-1-5-g2-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'f_ni': Scheme(sh..._nj': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int32, out_node_feats = 5, out_edge_feats = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0107A508>
op = 'add'
lhs_data = tensor([[ 3.6684],
        [-3.8246],
        [-2.2607],
        [-1.6092],
        [-2.1457],
        [-2.4016],
        [ 0.8714],
        [-2.6216],
        [ 1.2264],
        [-1.1432]], grad_fn=<MmBackward>)
rhs_data = tensor([[-1.2083],
        [ 1.3820],
        [-1.2841],
        [-1.9813],
        [-1.9693],
        [-0.5169],
        [-2.3614],
        [ 0.1031],
        [-1.7166],
        [-0.6641]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[1-1-5-g2-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'f_ni': Scheme(sh..._nj': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int64, out_node_feats = 5, out_edge_feats = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF017C5548>
op = 'add'
lhs_data = tensor([[ 2.1769],
        [-1.2392],
        [-0.7810],
        [-0.1015],
        [ 0.9861],
        [ 0.7646],
        [ 1.4994],
        [-0.4446],
        [ 2.1092],
        [ 1.1567]], grad_fn=<MmBackward>)
rhs_data = tensor([[ 0.6555],
        [-1.0613],
        [-0.9363],
        [ 0.4830],
        [ 2.6376],
        [ 2.1240],
        [ 3.3062],
        [ 1.1973],
        [ 5.2595],
        [ 2.3797]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[1-1-5-g3-idtype0] _______________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'f_ni': Scheme(shape=(1,), dtype=torch.float32), 'f_nj': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_node_feats = 5, out_edge_feats = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF011EA3C8>
op = 'add'
lhs_data = tensor([[-1.5390],
        [ 1.5751],
        [ 0.8392],
        [ 0.7174],
        [ 2.2216],
        [ 2.0319],
        [-3.7562],
        [ 0.2812],
        [ 2.7828]], grad_fn=<MmBackward>)
rhs_data = tensor([[-2.1967],
        [ 3.1941],
        [ 0.3549],
        [ 0.9047],
        [ 0.2886],
        [ 0.3208],
        [ 2.3009],
        [ 0.7998],
        [ 3.7961]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[1-1-5-g3-idtype1] _______________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'f_ni': Scheme(shape=(1,), dtype=torch.float32), 'f_nj': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_node_feats = 5, out_edge_feats = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0172CFC8>
op = 'add'
lhs_data = tensor([[-1.5730],
        [ 1.3183],
        [-3.3253],
        [-0.7844],
        [-2.7867],
        [ 0.9051],
        [-2.0811],
        [ 0.0180],
        [ 0.9482]], grad_fn=<MmBackward>)
rhs_data = tensor([[ 0.6680],
        [ 2.1042],
        [-0.9548],
        [ 0.1887],
        [-1.5646],
        [-2.5504],
        [-0.4785],
        [ 0.2376],
        [-0.1559]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[1-1-5-g4-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'f_ni': Scheme(shape=(1,), dtype=torch.float32), 'f_nj': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_node_feats = 5, out_edge_feats = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE149F4F08>
op = 'add'
lhs_data = tensor([[ 0.1369],
        [ 0.9024],
        [-0.7537],
        [ 0.5417],
        [ 1.0035],
        [-1.0498],
        [ 1.4059],
        [-0.8821],
        [ 1.1979],
        [ 1.7486]], grad_fn=<MmBackward>)
rhs_data = tensor([[-0.2269],
        [ 2.6755],
        [ 0.3550],
        [ 2.6903],
        [ 4.0163],
        [-0.8691],
        [-1.3069],
        [-1.3130],
        [-0.1179],
        [ 1.2068]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[1-1-5-g4-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'f_ni': Scheme(shape=(1,), dtype=torch.float32), 'f_nj': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_node_feats = 5, out_edge_feats = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03D11A88>
op = 'add'
lhs_data = tensor([[-2.5050],
        [-1.3629],
        [-1.3069],
        [-2.7562],
        [ 1.2322],
        [ 1.8122],
        [-0.8600],
        [ 1.0151],
        [-0.6798],
        [-1.5950]], grad_fn=<MmBackward>)
rhs_data = tensor([[-1.2059],
        [-1.9523],
        [ 2.2092],
        [ 2.3935],
        [ 2.2850],
        [ 0.1742],
        [-5.0623],
        [-2.0806],
        [-1.2818],
        [ 3.9557]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[1-1-5-g5-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'f_ni': Scheme(shape=(1,), dtype=torch.float32), 'f_nj': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_node_feats = 5, out_edge_feats = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07632B88>
op = 'add'
lhs_data = tensor([[-0.0983],
        [-0.6363],
        [ 0.8019],
        [-1.7780],
        [-0.9122],
        [ 3.2995],
        [ 4.0411],
        [ 1.0527],
        [-3.4736],
        [ 0.0752]], grad_fn=<MmBackward>)
rhs_data = tensor([[-0.3303],
        [ 0.1051],
        [ 1.4269],
        [-2.4812],
        [-0.9091],
        [ 1.1765],
        [ 1.5323],
        [-1.7500],
        [ 1.0788],
        [-2.0021]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[1-1-5-g5-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'f_ni': Scheme(shape=(1,), dtype=torch.float32), 'f_nj': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_node_feats = 5, out_edge_feats = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0172C848>
op = 'add'
lhs_data = tensor([[ 0.4877],
        [ 2.3753],
        [-3.1280],
        [ 1.7633],
        [ 0.5442],
        [ 1.4534],
        [-1.0089],
        [ 1.0435],
        [ 0.2633],
        [ 0.9303]], grad_fn=<MmBackward>)
rhs_data = tensor([[ 0.7791],
        [ 1.2760],
        [-0.2289],
        [ 2.0642],
        [-2.2068],
        [ 4.2850],
        [ 0.1671],
        [ 0.8219],
        [ 4.3436],
        [ 1.5679]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[1-5-1-g0-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'f_ni': Scheme(sh...), 'f_nj': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
idtype = torch.int32, out_node_feats = 1, out_edge_feats = 5, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03CF8288>
op = 'add'
lhs_data = tensor([[ 2.2292,  0.7761,  0.8729,  0.0677, -0.0351],
        [ 1.7661,  1.0958,  0.8817,  0.0134, -0.8056],
        ...962, -0.4775,  0.5063, -1.8566, -1.2691],
        [-0.1253, -1.9480, -1.6616, -0.8573,  0.5076]], grad_fn=<MmBackward>)
rhs_data = tensor([[ 0.2724, -0.3118, -1.1077,  1.1291, -0.2934],
        [-0.7067,  1.0285, -0.3682, -1.4555, -0.3965],
        ...785, -1.3536,  1.1341,  2.1100,  0.1725],
        [ 1.6862, -2.6581,  2.0313, -0.7307,  0.5329]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[1-5-1-g0-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'f_ni': Scheme(sh...), 'f_nj': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
idtype = torch.int64, out_node_feats = 1, out_edge_feats = 5, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03D15D08>
op = 'add'
lhs_data = tensor([[ 1.5567e+00,  3.7043e-01, -1.1112e+00,  5.4604e-01,  9.7542e-01],
        [ 2.0248e-01,  4.5994e-01, -1.7311e...  1.1999e+00],
        [ 4.6394e-01,  1.0419e+00, -1.4428e+00, -2.8518e-01, -1.8103e+00]],
       grad_fn=<MmBackward>)
rhs_data = tensor([[ 3.0364, -1.0318,  0.0287,  1.1455, -0.2633],
        [-1.0490, -3.4290, -0.2606, -3.2877,  0.9691],
        ...434,  1.6149, -0.1309, -0.0900,  2.1924],
        [ 1.2451, -2.0974, -1.1284, -0.4601,  2.0588]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[1-5-1-g1-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'f_ni': Scheme(shape=(5,), dtype=torch.float32), 'f_nj': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_node_feats = 1, out_edge_feats = 5, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01505448>
op = 'add'
lhs_data = tensor([[-1.1543, -0.3406,  1.3935, -3.1570,  0.6227],
        [ 2.4716,  1.3599,  0.7560, -0.0414,  1.7562],
        ...964,  0.1458, -3.7151, -0.2945,  0.7375],
        [-1.0108, -0.3073,  1.6134, -0.3309, -0.3012]], grad_fn=<MmBackward>)
rhs_data = tensor([[-0.2096, -1.0450,  1.7865,  0.5910,  1.8447],
        [-1.0185,  0.2223,  2.3722, -2.3820, -2.1388],
        ...615,  1.6159, -0.1511, -1.5254, -1.9828],
        [-0.4477, -0.1013, -1.2512, -0.2377,  0.3427]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[1-5-1-g1-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'f_ni': Scheme(shape=(5,), dtype=torch.float32), 'f_nj': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_node_feats = 1, out_edge_feats = 5, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01134D08>
op = 'add'
lhs_data = tensor([[ 2.1198,  0.2209,  1.4118,  0.9327, -0.1623],
        [ 0.7169, -1.4005,  3.0707,  0.6256, -0.4798],
        ...518,  0.9069,  0.2913,  2.2662,  0.1914],
        [-3.2028, -1.0832, -0.2766,  0.1172,  1.9696]], grad_fn=<MmBackward>)
rhs_data = tensor([[ 2.5442e-01,  4.7672e-02, -1.9061e+00, -2.3700e+00, -2.1104e-01],
        [-2.2052e-03,  1.0820e+00,  5.3269e... -1.1084e+00],
        [-2.4736e+00,  7.9400e-01,  1.6422e+00,  7.8995e-01, -1.2234e-01]],
       grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[1-5-1-g2-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'f_ni': Scheme(sh..._nj': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int32, out_node_feats = 1, out_edge_feats = 5, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03CDB908>
op = 'add'
lhs_data = tensor([[ 0.8698,  0.4731, -0.6279, -0.7516, -0.2538],
        [-0.4324, -1.1567, -1.4843, -1.0208,  0.0378],
        ...554,  1.5762,  2.2341, -3.4180,  0.5337],
        [-2.8927, -2.4243,  0.3537,  0.9134, -1.8519]], grad_fn=<MmBackward>)
rhs_data = tensor([[ 2.9725,  0.2351,  2.9036, -4.4582,  1.2636],
        [-3.4791,  2.4749, -3.3202,  2.0570, -0.5733],
        ...851,  1.4939,  0.7504, -1.6213, -0.3278],
        [ 1.4869, -3.5392,  2.5625,  0.5170, -3.3768]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[1-5-1-g2-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'f_ni': Scheme(sh..._nj': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int64, out_node_feats = 1, out_edge_feats = 5, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01585E48>
op = 'add'
lhs_data = tensor([[ 0.0135,  0.1967,  1.3619, -0.7399,  2.4161],
        [-2.3796,  1.6751, -2.0963,  0.4456, -1.9764],
        ...969, -1.0809, -1.1263, -0.8281, -1.8139],
        [-0.4464, -0.1177,  0.6837,  2.9737, -0.8409]], grad_fn=<MmBackward>)
rhs_data = tensor([[-1.7881, -2.5947,  1.8418, -2.2188,  1.7566],
        [-0.5611,  0.1087, -1.3058,  3.4624, -1.5013],
        ...867, -1.1388, -3.6475,  2.4635, -2.4500],
        [ 0.1178, -3.8976, -0.5162,  3.8410, -3.1021]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[1-5-1-g3-idtype0] _______________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'f_ni': Scheme(shape=(5,), dtype=torch.float32), 'f_nj': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_node_feats = 1, out_edge_feats = 5, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14C6F5C8>
op = 'add'
lhs_data = tensor([[ 1.6274, -1.9813,  0.3693, -2.2988, -2.2915],
        [-1.9376, -0.8767, -2.7906, -0.6635, -1.5620],
        ...615,  2.5475,  0.0974,  0.0206,  1.8997],
        [-0.2841,  2.2899, -0.3244,  0.7462,  1.1603]], grad_fn=<MmBackward>)
rhs_data = tensor([[ 2.2075,  0.2019,  0.2367,  3.6682, -3.2571],
        [-2.9066,  0.3186, -2.2495, -0.1979,  2.9868],
        ...704,  0.4992, -0.2510,  0.0846, -0.1753],
        [-0.8731,  0.4930, -0.8623,  0.1239,  3.2514]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[1-5-1-g3-idtype1] _______________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'f_ni': Scheme(shape=(5,), dtype=torch.float32), 'f_nj': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_node_feats = 1, out_edge_feats = 5, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF019AD148>
op = 'add'
lhs_data = tensor([[ 0.2502, -0.1276,  0.5703, -0.1527,  0.9007],
        [ 2.6106, -0.8256, -0.9256, -1.0412,  0.7527],
        ...357,  0.5633,  0.2226, -0.0466, -1.1762],
        [ 2.7168, -3.0784, -0.0506, -0.4492, -0.7324]], grad_fn=<MmBackward>)
rhs_data = tensor([[-2.1059,  2.4533, -1.0715, -0.6698,  0.3795],
        [-1.7206,  4.0144, -2.1319,  0.0607, -0.2874],
        ...299, -0.2342, -0.4059, -1.9755, -1.2225],
        [-0.2999,  4.6867, -1.3491, -0.0919, -1.8703]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[1-5-1-g4-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'f_ni': Scheme(shape=(5,), dtype=torch.float32), 'f_nj': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_node_feats = 1, out_edge_feats = 5, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE145C2348>
op = 'add'
lhs_data = tensor([[ 0.2136, -0.9638,  1.4458, -0.9056,  0.7961],
        [ 2.6831, -1.7321,  1.9454, -0.5555,  0.3980],
        ...582, -0.0589,  0.3639,  0.1005, -1.9299],
        [-3.4502, -1.2031, -0.8054, -1.2209,  1.4334]], grad_fn=<MmBackward>)
rhs_data = tensor([[-1.5482, -1.3303, -1.1250, -0.8845, -0.0371],
        [-0.2660, -0.0963, -1.1495,  0.3533,  0.6082],
        ...484,  2.2112,  0.8943,  1.2655, -0.3795],
        [-0.8039,  1.7557, -0.2629,  1.2672, -1.4617]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[1-5-1-g4-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'f_ni': Scheme(shape=(5,), dtype=torch.float32), 'f_nj': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_node_feats = 1, out_edge_feats = 5, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01680A48>
op = 'add'
lhs_data = tensor([[-1.9538,  1.9636, -1.0473,  1.5812, -1.1933],
        [ 3.1597,  1.6426, -1.6081, -4.2960, -0.1847],
        ...205, -0.2807, -0.7947, -0.1322, -0.2584],
        [ 1.3606,  0.4565, -2.1084, -2.9611, -0.0272]], grad_fn=<MmBackward>)
rhs_data = tensor([[-4.1490e-01,  2.7696e-04,  1.4479e+00, -5.5889e-01,  6.2391e-01],
        [ 5.8932e-01,  2.2507e-01, -1.0255e...  5.9090e-01],
        [-5.3113e-01, -1.4547e+00, -7.5951e-01,  2.7943e+00,  9.5287e-01]],
       grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[1-5-1-g5-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'f_ni': Scheme(shape=(5,), dtype=torch.float32), 'f_nj': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_node_feats = 1, out_edge_feats = 5, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01688188>
op = 'add'
lhs_data = tensor([[-1.5821, -1.9691, -1.5370, -1.9131,  1.7537],
        [ 0.5591,  1.3083,  0.6558,  1.3594, -0.4754],
        ...859,  2.0641, -2.5095, -0.0132,  0.5487],
        [ 0.9767,  2.5717, -0.3369, -3.3207, -0.3065]], grad_fn=<MmBackward>)
rhs_data = tensor([[ 0.3420,  0.2810, -1.0181,  0.1686,  1.6525],
        [-0.8873, -1.8001, -0.1218,  0.7587,  1.9681],
        ...586,  0.8394,  2.0915,  0.4825, -0.5250],
        [-0.8588,  0.5579,  0.7878, -0.7125,  4.4885]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[1-5-1-g5-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'f_ni': Scheme(shape=(5,), dtype=torch.float32), 'f_nj': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_node_feats = 1, out_edge_feats = 5, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE145DFDC8>
op = 'add'
lhs_data = tensor([[-1.3711e+00,  5.1826e-02,  1.5582e+00, -2.8094e+00, -9.9020e-01],
        [ 6.0076e-01,  1.8137e+00,  6.1193e...  1.3008e-01],
        [-6.9144e-01,  3.3420e+00, -5.8137e-01, -1.5156e+00, -8.9822e-01]],
       grad_fn=<MmBackward>)
rhs_data = tensor([[ 0.5989, -1.7439,  0.4382, -0.0057,  2.6945],
        [ 2.5224, -0.1026, -1.0784,  1.8687, -0.5587],
        ...985,  1.7506, -0.6615,  0.5553, -0.8469],
        [ 0.8139, -0.5157, -3.8765,  1.2383, -0.1366]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[1-5-5-g0-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'f_ni': Scheme(sh...), 'f_nj': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
idtype = torch.int32, out_node_feats = 5, out_edge_feats = 5, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14ABE708>
op = 'add'
lhs_data = tensor([[-1.5224,  4.0198, -0.9121,  0.8066,  0.2421],
        [ 0.2687,  0.0524,  0.4975,  0.3253, -0.0616],
        ...524,  0.5998, -2.1082, -0.2585, -0.6619],
        [ 1.9605, -1.9432,  0.8023, -0.7697, -0.0879]], grad_fn=<MmBackward>)
rhs_data = tensor([[ 0.1521, -2.7563,  2.1220,  1.0328,  0.5755],
        [-0.3067, -0.8002,  0.9515,  1.0783, -0.0851],
        ...759, -0.8033,  2.0626,  0.6030, -1.4578],
        [-0.1893,  0.3429, -2.9925, -0.9350,  4.1221]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[1-5-5-g0-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'f_ni': Scheme(sh...), 'f_nj': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
idtype = torch.int64, out_node_feats = 5, out_edge_feats = 5, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF018D0888>
op = 'add'
lhs_data = tensor([[ 0.3066, -0.5182,  0.1193, -0.0776,  1.4858],
        [ 0.7725, -2.2485,  1.4806,  0.7869,  0.3996],
        ...206, -1.1703,  0.7422, -0.1161,  1.4291],
        [ 1.7498, -0.0964,  3.3670,  1.3568, -0.0539]], grad_fn=<MmBackward>)
rhs_data = tensor([[ 2.0710e+00, -1.3243e+00, -5.3452e-01,  1.6118e+00, -5.7696e-01],
        [-4.0600e-01,  1.1775e+00,  1.8890e... -3.8951e-01],
        [-1.2197e-02,  3.8464e+00,  1.7388e+00,  3.3590e-01,  9.4638e-01]],
       grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[1-5-5-g1-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'f_ni': Scheme(shape=(5,), dtype=torch.float32), 'f_nj': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_node_feats = 5, out_edge_feats = 5, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01690948>
op = 'add'
lhs_data = tensor([[-2.2770e-03,  1.5243e+00,  1.9323e-01,  2.1749e-01,  4.1734e-01],
        [ 2.9220e-01,  3.6485e-01,  2.5542e...  4.1600e-01],
        [-1.8994e+00,  2.4025e+00, -2.3199e+00,  7.5394e-01,  1.8846e-01]],
       grad_fn=<MmBackward>)
rhs_data = tensor([[ 1.2815e+00, -2.4189e+00, -1.2301e-01, -6.5389e-01,  1.0223e+00],
        [-1.8534e+00,  4.3350e+00,  6.7586e... -3.8844e-01],
        [ 8.4119e-02,  2.6608e+00, -7.7496e-01, -2.1621e+00, -4.7734e-01]],
       grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[1-5-5-g1-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'f_ni': Scheme(shape=(5,), dtype=torch.float32), 'f_nj': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_node_feats = 5, out_edge_feats = 5, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF077308C8>
op = 'add'
lhs_data = tensor([[-2.6950,  2.2951, -0.1309, -0.0055, -3.1580],
        [-4.5825,  1.6883, -2.1557, -2.5061,  0.4735],
        ...427, -1.3015, -0.9402,  3.7236,  1.7030],
        [ 0.7454, -2.9531,  0.4406,  2.2932,  0.2160]], grad_fn=<MmBackward>)
rhs_data = tensor([[-1.1289,  0.4693,  0.0560, -0.1606, -2.6036],
        [-0.0786,  1.6601, -0.6393,  0.4642,  0.0210],
        ...776,  1.0611,  0.4576, -0.1061,  1.4631],
        [ 0.5337,  1.9598,  0.9898,  1.9486, -0.5580]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[1-5-5-g2-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'f_ni': Scheme(sh..._nj': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int32, out_node_feats = 5, out_edge_feats = 5, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF030E8A08>
op = 'add'
lhs_data = tensor([[-0.7537,  2.0143,  0.3658,  0.7582, -1.8089],
        [-0.4673,  1.1702, -3.6322,  1.0810,  1.5954],
        ...586, -0.1455, -1.6464, -0.2126, -1.6582],
        [ 1.4705,  1.4972,  0.2438,  0.7064, -2.4887]], grad_fn=<MmBackward>)
rhs_data = tensor([[-0.0246,  0.3147,  1.2331, -1.4804,  1.1563],
        [ 0.2305,  1.2871,  3.1029, -3.2249,  0.7716],
        ...642,  1.3884, -0.4639,  0.5447,  0.7121],
        [ 1.3656, -1.5995, -0.4328,  0.7571, -3.4159]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[1-5-5-g2-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'f_ni': Scheme(sh..._nj': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int64, out_node_feats = 5, out_edge_feats = 5, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14AA4648>
op = 'add'
lhs_data = tensor([[ 1.8379,  0.7170, -0.8860, -0.9213, -1.5779],
        [ 0.5741,  1.7654,  1.3555, -1.0251,  0.3889],
        ...383,  0.2383, -0.8433,  0.2941, -0.2822],
        [-3.5883, -2.0470, -0.7100,  0.0262,  0.3010]], grad_fn=<MmBackward>)
rhs_data = tensor([[ 0.2643, -0.6982,  3.4455, -1.7912,  1.5166],
        [-2.1576, -1.8441,  0.1659, -1.9829, -0.1315],
        ...737,  0.1600,  0.3329, -0.6188, -0.8410],
        [ 1.4112, -0.1187, -2.8850,  0.2129, -0.9516]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'


______________________ test_egat_conv[1-5-5-g3-idtype0] _______________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'f_ni': Scheme(shape=(5,), dtype=torch.float32), 'f_nj': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_node_feats = 5, out_edge_feats = 5, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF030E87C8>
op = 'add'
lhs_data = tensor([[-0.5511,  0.3179,  1.7453, -2.7484,  1.2316],
        [-2.4131,  0.8740, -1.9306,  0.6628, -1.4297],
        ...656,  1.8283,  0.5266, -4.1062, -0.6205],
        [ 1.7387,  2.5635,  1.0679,  0.2465, -0.2020]], grad_fn=<MmBackward>)
rhs_data = tensor([[-0.3100, -0.2909, -0.2811,  0.2584,  2.4082],
        [ 0.1045, -1.5974,  1.4981,  0.8283,  1.8680],
        ...759, -2.3608,  1.4576,  1.1587,  1.3148],
        [-0.0707, -0.7638, -0.2511, -1.1084, -2.9884]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[1-5-5-g3-idtype1] _______________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'f_ni': Scheme(shape=(5,), dtype=torch.float32), 'f_nj': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_node_feats = 5, out_edge_feats = 5, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF011FDD08>
op = 'add'
lhs_data = tensor([[ 0.7389, -2.9631, -0.3714,  0.3155, -2.1281],
        [ 0.9205,  3.3359, -1.3430, -0.4578,  0.0119],
        ...834, -0.7278,  1.7913, -0.0206, -0.1821],
        [ 0.5815, -0.3880, -2.6453, -0.3300, -1.9916]], grad_fn=<MmBackward>)
rhs_data = tensor([[-0.2530, -1.9193,  1.0921,  1.6466,  2.9917],
        [ 0.0329,  2.8371,  0.4122, -1.7736, -2.8467],
        ...098,  0.7569, -0.9536,  0.0831, -3.9703],
        [ 2.0233, -0.0954,  0.8983,  1.0181,  0.0321]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[1-5-5-g4-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'f_ni': Scheme(shape=(5,), dtype=torch.float32), 'f_nj': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_node_feats = 5, out_edge_feats = 5, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE149F4F08>
op = 'add'
lhs_data = tensor([[-0.7091, -1.9608, -1.0627,  1.5727, -0.7033],
        [ 0.2077, -0.1939, -1.2146, -2.7953, -1.4495],
        ...293,  0.0235,  0.4523,  0.2061, -0.1833],
        [-0.6810, -0.5318, -0.5044, -1.6564, -0.3337]], grad_fn=<MmBackward>)
rhs_data = tensor([[ 0.8952,  1.5233,  2.0281, -3.0786,  0.1367],
        [ 1.5246,  1.3504, -1.0947, -0.0165,  1.2030],
        ...460, -0.8525,  0.7980, -0.6268,  0.0291],
        [-0.6504,  0.5439,  0.0781,  1.1774, -1.1338]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[1-5-5-g4-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'f_ni': Scheme(shape=(5,), dtype=torch.float32), 'f_nj': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_node_feats = 5, out_edge_feats = 5, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF017C2FC8>
op = 'add'
lhs_data = tensor([[-1.8086,  0.9245,  0.5835,  0.9028,  0.9542],
        [ 0.8189, -0.3348, -1.7175,  0.9434, -0.1541],
        ...599,  1.1003, -0.9780,  1.2732, -0.1804],
        [-3.2987,  1.4191,  0.4780,  0.7067, -0.4373]], grad_fn=<MmBackward>)
rhs_data = tensor([[-0.7585,  0.5401, -1.2831, -1.1542, -0.6545],
        [-0.0938, -1.0621, -0.6654,  0.6071,  0.2818],
        ...408, -0.5371, -0.9386, -0.3707, -0.4968],
        [-0.0840, -0.1149, -0.2563, -0.3345, -1.7704]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[1-5-5-g5-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'f_ni': Scheme(shape=(5,), dtype=torch.float32), 'f_nj': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_node_feats = 5, out_edge_feats = 5, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0171EC48>
op = 'add'
lhs_data = tensor([[ 5.2904, -5.6345,  2.3054, -1.2926,  0.2312],
        [-2.3719, -2.9285, -1.6052,  2.7780, -2.6947],
        ...764, -6.8275, -0.1479, -1.6348, -0.2559],
        [ 0.0123,  1.3724, -0.2375, -0.2877,  0.0654]], grad_fn=<MmBackward>)
rhs_data = tensor([[ 1.8065, -1.0380, -2.2696,  1.5252, -0.7376],
        [ 1.8640, -2.1299,  3.9823,  1.5301,  0.6907],
        ...146,  1.7678, -3.5091,  0.7423, -3.0150],
        [-0.0657, -0.4329,  1.2101, -0.7717,  0.6540]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[1-5-5-g5-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'f_ni': Scheme(shape=(5,), dtype=torch.float32), 'f_nj': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_node_feats = 5, out_edge_feats = 5, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01692A88>
op = 'add'
lhs_data = tensor([[ 1.7367,  2.2677,  2.3784, -2.1027, -0.2515],
        [-1.5536,  0.5589, -1.2308, -0.6574, -1.0994],
        ...716,  1.8955, -1.0991,  2.4814,  1.3006],
        [ 3.5422, -0.1395,  1.7210, -2.8993, -3.5538]], grad_fn=<MmBackward>)
rhs_data = tensor([[-0.4696, -0.7178, -0.1673,  1.2336,  1.8181],
        [-0.4927,  1.5876, -0.9288, -0.8321,  0.7550],
        ...456, -0.9458,  0.9793, -2.3248, -0.0052],
        [-2.2309,  3.6036, -1.1320,  2.7364,  1.4736]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[4-1-1-g0-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'f_ni': Scheme(sh...), 'f_nj': Scheme(shape=(4,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
idtype = torch.int32, out_node_feats = 1, out_edge_feats = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF030ED548>
op = 'add'
lhs_data = tensor([[-0.4705,  5.6044,  3.2821,  0.4232],
        [ 1.0011, -0.2358,  2.1416, -1.0516],
        [-2.2868,  2.6409,...17],
        [-0.1693,  2.1667,  1.2084,  1.1679],
        [-1.9563,  2.1514,  1.5423,  0.6475]], grad_fn=<MmBackward>)
rhs_data = tensor([[-1.7161, -0.1187, -3.8107,  4.7277],
        [-1.8614, -0.1279, -0.9785, -0.0991],
        [ 0.6916, -0.6875,...66],
        [-0.4105,  1.0508, -1.4779,  2.9405],
        [ 1.5260, -1.5907, -2.8946,  1.4005]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[4-1-1-g0-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'f_ni': Scheme(sh...), 'f_nj': Scheme(shape=(4,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
idtype = torch.int64, out_node_feats = 1, out_edge_feats = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14A66EC8>
op = 'add'
lhs_data = tensor([[-1.4309, -1.6022,  1.2219, -0.5968],
        [ 0.7681,  1.4112, -1.5525,  1.1722],
        [ 1.7247,  2.5229,...18],
        [-1.6049, -3.0740,  1.7911, -1.2084],
        [-0.9855, -1.2082, -0.1131, -0.2469]], grad_fn=<MmBackward>)
rhs_data = tensor([[-0.4095, -1.5472, -1.2292,  0.4058],
        [ 0.0172,  0.8372,  1.7636, -1.2318],
        [-1.0580, -3.2101,...97],
        [ 1.7067,  3.8112, -0.2493,  0.2017],
        [ 1.4787,  2.0141, -0.9155,  0.9174]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[4-1-1-g1-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'f_ni': Scheme(shape=(4,), dtype=torch.float32), 'f_nj': Scheme(shape=(4,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_node_feats = 1, out_edge_feats = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03CF3948>
op = 'add'
lhs_data = tensor([[ 2.1911, -0.4616,  1.2143,  1.1113],
        [ 4.9592, -1.4097, -1.1115, -2.7368],
        [-0.7983, -0.7609,...22],
        [-1.7657,  1.0233, -0.0615,  0.9180],
        [ 0.4996,  0.9434,  1.0623,  0.2781]], grad_fn=<MmBackward>)
rhs_data = tensor([[ 0.0516,  3.8297,  1.5290, -1.2061],
        [-1.1254, -2.4673, -4.0393,  0.9669],
        [ 0.3320,  1.3885,...78],
        [-0.3420, -2.2822, -0.8262,  0.5787],
        [ 0.8054,  0.0079,  1.7885, -0.7706]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[4-1-1-g1-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'f_ni': Scheme(shape=(4,), dtype=torch.float32), 'f_nj': Scheme(shape=(4,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_node_feats = 1, out_edge_feats = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14A9DF88>
op = 'add'
lhs_data = tensor([[ 0.0770, -0.0182,  2.7876, -0.2004],
        [-2.5724,  1.4596, -1.5684, -0.3723],
        [-0.9130,  0.3625,...87],
        [ 1.5599,  1.8954,  0.2003,  0.8250],
        [-1.8670,  1.3862, -0.1251,  0.5208]], grad_fn=<MmBackward>)
rhs_data = tensor([[-0.5586,  1.2995,  0.4686, -3.2376],
        [ 0.8194,  3.0440, -1.2207,  2.0622],
        [-0.4392,  4.5676,...44],
        [ 1.1046, -1.5310,  2.4979, -0.2353],
        [ 1.2784,  1.2943,  2.8227, -1.5528]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[4-1-1-g2-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'f_ni': Scheme(sh..._nj': Scheme(shape=(4,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int32, out_node_feats = 1, out_edge_feats = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0155B648>
op = 'add'
lhs_data = tensor([[ 0.2541, -0.7663,  1.9420,  0.0404],
        [-0.5131, -1.1529, -1.1021,  0.5091],
        [-0.1418, -1.5617,...32],
        [-0.9856,  1.2223, -1.0214,  2.2359],
        [ 0.3141,  1.1446, -0.0780,  0.0148]], grad_fn=<MmBackward>)
rhs_data = tensor([[ 1.4962,  0.3621, -0.1691, -0.1641],
        [ 3.4366, -0.4130,  0.2263, -2.2695],
        [-1.6926,  0.3727,...44],
        [-1.9140, -0.4491, -2.2426, -1.2523],
        [ 1.4527, -0.3189,  2.7364, -0.2511]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[4-1-1-g2-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'f_ni': Scheme(sh..._nj': Scheme(shape=(4,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int64, out_node_feats = 1, out_edge_feats = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF017EFF88>
op = 'add'
lhs_data = tensor([[-2.6586, -2.5621, -1.5969, -0.3606],
        [-4.0109, -5.7380, -0.9841, -0.5199],
        [-4.1142, -2.5933,...14],
        [ 0.3800,  1.0556, -0.4602,  0.7014],
        [ 1.2257,  1.0313,  1.9744,  1.8609]], grad_fn=<MmBackward>)
rhs_data = tensor([[-1.6076, -2.1986,  0.7022,  0.9287],
        [-3.0533,  1.2627, -0.6349,  0.3400],
        [ 0.6336, -3.4636,...36],
        [ 0.9242,  0.9622,  0.7106,  1.0568],
        [ 0.7987,  1.0452,  1.1953, -0.9553]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[4-1-1-g3-idtype0] _______________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'f_ni': Scheme(shape=(4,), dtype=torch.float32), 'f_nj': Scheme(shape=(4,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_node_feats = 1, out_edge_feats = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03CF3D88>
op = 'add'
lhs_data = tensor([[-0.1330, -4.3771, -3.4253, -1.9938],
        [-1.6464, -0.9109, -1.1013,  0.1960],
        [-0.1609, -0.7915,...71],
        [ 1.6249,  0.4494,  1.2005,  0.2167],
        [-0.3224,  0.5720, -2.3798,  1.1342]], grad_fn=<MmBackward>)
rhs_data = tensor([[ 3.3954, -2.6274, -1.4209, -1.3720],
        [ 0.4453, -1.5269, -3.4375,  1.8051],
        [-1.5277, -0.3332,...76],
        [-0.3833,  1.6089,  1.4920,  0.2292],
        [ 0.5944,  0.3316, -0.2435, -1.0885]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[4-1-1-g3-idtype1] _______________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'f_ni': Scheme(shape=(4,), dtype=torch.float32), 'f_nj': Scheme(shape=(4,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_node_feats = 1, out_edge_feats = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0184F8C8>
op = 'add'
lhs_data = tensor([[ 1.0806,  1.8655, -0.5368, -0.1033],
        [-2.4902, -0.5509, -1.6514,  1.1719],
        [-3.8806,  0.9989,...04],
        [ 2.7465,  0.0502, -0.6588, -0.1011],
        [ 1.5691,  0.1306, -0.5738, -0.9420]], grad_fn=<MmBackward>)
rhs_data = tensor([[ 0.3695,  1.4046, -0.9196, -1.6754],
        [ 3.1955,  0.2724,  1.1049,  0.5488],
        [-0.2718,  2.7186,...93],
        [ 0.2168, -1.1918, -0.9921, -1.2101],
        [ 0.6106,  0.1939,  0.1893, -1.2250]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[4-1-1-g4-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'f_ni': Scheme(shape=(4,), dtype=torch.float32), 'f_nj': Scheme(shape=(4,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_node_feats = 1, out_edge_feats = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03D27648>
op = 'add'
lhs_data = tensor([[ 0.9970,  1.5860, -0.9695,  1.7324],
        [ 1.3100,  0.6790,  0.1797,  2.1306],
        [-1.4844, -0.5437,...49],
        [-0.1265,  2.6317,  1.4199, -1.0537],
        [ 0.5047,  0.0384,  1.0749, -0.2805]], grad_fn=<MmBackward>)
rhs_data = tensor([[-3.1614e+00, -6.0420e+00, -2.4869e+00,  1.6781e+00],
        [-6.4173e-01, -8.4946e-01, -2.5066e-01, -1.7252e...  6.6390e-02, -2.4160e+00],
        [-1.7252e-01, -5.7819e-01, -1.6768e-01, -8.4636e-01]],
       grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[4-1-1-g4-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'f_ni': Scheme(shape=(4,), dtype=torch.float32), 'f_nj': Scheme(shape=(4,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_node_feats = 1, out_edge_feats = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0151CF88>
op = 'add'
lhs_data = tensor([[ 1.4317,  0.2051, -0.4386,  0.5630],
        [-2.1512,  0.0136,  1.2346, -1.2071],
        [-0.8090,  1.5302,...70],
        [-1.6954, -0.5430,  0.2314, -0.7219],
        [ 2.1417, -1.3347, -1.9166,  2.8811]], grad_fn=<MmBackward>)
rhs_data = tensor([[-1.9771, -0.9079, -2.0069, -0.5775],
        [-0.2323,  0.1495, -0.5343,  2.8827],
        [ 1.0017,  0.6641,...95],
        [-1.1003,  0.2216,  0.5521, -1.8042],
        [-1.0003,  1.2796, -0.3066,  0.0315]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[4-1-1-g5-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'f_ni': Scheme(shape=(4,), dtype=torch.float32), 'f_nj': Scheme(shape=(4,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_node_feats = 1, out_edge_feats = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF076A7E88>
op = 'add'
lhs_data = tensor([[ 1.2188,  2.1927,  2.2811, -0.4799],
        [-1.0089,  0.4566,  0.3029,  0.1925],
        [ 2.0592,  1.8421,...56],
        [ 0.0860,  3.8168,  2.2639,  4.0036],
        [-0.2505,  2.5956,  0.8886,  1.4244]], grad_fn=<MmBackward>)
rhs_data = tensor([[ 0.7030,  0.1426,  0.0691,  3.0107],
        [ 1.9122,  0.6484,  0.2837, -0.9625],
        [ 0.6895,  1.2586,...01],
        [ 0.3855, -1.0402,  0.5591,  1.9251],
        [ 1.6369,  2.1130, -0.3646, -0.3312]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[4-1-1-g5-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'f_ni': Scheme(shape=(4,), dtype=torch.float32), 'f_nj': Scheme(shape=(4,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_node_feats = 1, out_edge_feats = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14931248>
op = 'add'
lhs_data = tensor([[ 0.4427, -0.6951,  0.0676,  1.0345],
        [-2.9837,  0.5889, -0.0792,  0.0990],
        [-2.1386, -0.7936,...31],
        [ 0.3877, -0.9499, -0.7087,  1.9725],
        [ 1.2181,  1.5107,  0.1961, -1.7028]], grad_fn=<MmBackward>)
rhs_data = tensor([[-0.0164,  0.7725,  0.5298, -1.5721],
        [ 0.6917, -1.7708,  0.4378,  2.5899],
        [ 0.4570, -1.1392,...10],
        [ 0.1479, -0.0720, -1.2991, -0.2711],
        [-0.7566,  1.0521,  1.0432, -0.5895]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[4-1-5-g0-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'f_ni': Scheme(sh...), 'f_nj': Scheme(shape=(4,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
idtype = torch.int32, out_node_feats = 5, out_edge_feats = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03D407C8>
op = 'add'
lhs_data = tensor([[-3.1886,  1.3212, -1.8167, -2.0514],
        [ 1.3203, -0.6645, -0.1217, -0.2292],
        [-1.7978, -1.0207,...20],
        [-0.9057,  0.3921, -1.1115, -1.6404],
        [-0.1795,  0.6632, -0.4115,  0.7573]], grad_fn=<MmBackward>)
rhs_data = tensor([[ 0.7254,  0.1973, -0.0808, -0.5974],
        [ 0.2943,  1.9091, -2.0205, -0.4959],
        [-0.1741,  0.6722,...51],
        [ 0.8325, -0.9791, -0.4594,  1.4006],
        [-1.8936,  0.4258,  3.0047, -1.0940]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[4-1-5-g0-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'f_ni': Scheme(sh...), 'f_nj': Scheme(shape=(4,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
idtype = torch.int64, out_node_feats = 5, out_edge_feats = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03CD8488>
op = 'add'
lhs_data = tensor([[-0.2177, -1.2223, -0.8321,  0.4431],
        [-1.5519, -3.7386, -1.4848, -1.3837],
        [ 2.1457, -2.0602,...26],
        [ 2.9129, -3.4845, -0.0458,  2.3417],
        [-0.5298,  0.1458, -2.5092, -0.2638]], grad_fn=<MmBackward>)
rhs_data = tensor([[ 0.6079, -0.1443, -0.3838, -0.8270],
        [-1.0742, -1.3106,  0.1403,  2.6426],
        [ 0.6929,  0.9749,...66],
        [ 1.0807, -2.1269, -1.4642,  1.8988],
        [ 0.6468,  0.3654,  4.3858,  1.2625]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[4-1-5-g1-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'f_ni': Scheme(shape=(4,), dtype=torch.float32), 'f_nj': Scheme(shape=(4,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_node_feats = 5, out_edge_feats = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01731F08>
op = 'add'
lhs_data = tensor([[-2.3643, -2.3627, -1.5377,  2.2136],
        [ 1.4285,  2.6656,  0.8685,  1.0907],
        [ 3.4750, -1.9359,...11],
        [ 0.5168, -0.0379,  2.0932, -0.6070],
        [-1.7742, -0.7319,  1.0960,  0.1867]], grad_fn=<MmBackward>)
rhs_data = tensor([[ 4.5253, -1.6406,  0.9484,  0.4256],
        [ 1.2817,  0.3861,  1.9974,  2.8732],
        [-2.4887,  3.2689,...84],
        [-3.8410, -4.7006,  1.6751,  2.0618],
        [ 0.0761, -2.2310, -2.0439,  0.0618]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[4-1-5-g1-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'f_ni': Scheme(shape=(4,), dtype=torch.float32), 'f_nj': Scheme(shape=(4,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_node_feats = 5, out_edge_feats = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14ACAD08>
op = 'add'
lhs_data = tensor([[ 2.9831e-01,  3.2656e+00,  3.9567e+00, -4.1908e+00],
        [-2.0610e+00, -1.1100e+00, -3.8260e+00, -2.7758e... -1.1880e+00, -1.8185e+00],
        [ 6.8576e-01,  2.8559e+00, -9.5661e-01,  2.4070e-01]],
       grad_fn=<MmBackward>)
rhs_data = tensor([[ 3.1088, -1.8868,  0.2667, -1.1148],
        [-1.7955, -2.4547,  4.9673,  4.8564],
        [ 0.6461, -2.8934,...74],
        [ 0.4428, -3.0348,  1.8993,  1.2420],
        [-0.5464, -3.7737,  0.0888,  0.9691]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[4-1-5-g2-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'f_ni': Scheme(sh..._nj': Scheme(shape=(4,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int32, out_node_feats = 5, out_edge_feats = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF010B6488>
op = 'add'
lhs_data = tensor([[-1.9866e+00, -1.4745e+00, -2.2471e+00, -1.9539e+00],
        [ 2.5841e+00,  7.2719e-01, -1.4797e+00, -1.8424e... -3.9597e-03,  1.1277e+00],
        [ 4.3326e-01,  8.3875e-01, -3.4387e-01,  9.6885e-01]],
       grad_fn=<MmBackward>)
rhs_data = tensor([[ 0.6348, -0.0901, -1.7309,  1.3762],
        [ 0.7000,  2.3851, -0.5260, -0.1101],
        [-0.2585,  1.2528,...64],
        [ 0.0194,  0.3597, -0.3978, -1.9673],
        [-0.5673,  0.9465,  0.0478, -1.2157]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[4-1-5-g2-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'f_ni': Scheme(sh..._nj': Scheme(shape=(4,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int64, out_node_feats = 5, out_edge_feats = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14559148>
op = 'add'
lhs_data = tensor([[-1.2525, -1.5866,  2.0505,  1.8830],
        [ 0.5264,  3.4699, -1.9177, -1.7508],
        [ 0.9481,  1.3392,...71],
        [-0.0924, -1.6354,  1.8022, -0.8079],
        [-1.0340, -1.4842,  0.4683, -2.7473]], grad_fn=<MmBackward>)
rhs_data = tensor([[-0.1480,  0.6609,  1.2678,  0.3669],
        [-2.5142, -1.2031,  1.2148, -1.1793],
        [ 0.4836,  0.8027,...57],
        [-0.3987,  0.8789, -0.6952,  0.2705],
        [-3.0286, -0.4141,  0.7886, -0.1568]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[4-1-5-g3-idtype0] _______________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'f_ni': Scheme(shape=(4,), dtype=torch.float32), 'f_nj': Scheme(shape=(4,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_node_feats = 5, out_edge_feats = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1469BA88>
op = 'add'
lhs_data = tensor([[ 0.7737,  2.2102, -2.6027,  6.1020],
        [-0.5747, -0.7563,  4.2326, -3.1685],
        [ 0.5448, -3.2331,...53],
        [ 1.7322,  1.4712, -1.8540,  4.2923],
        [-0.1973,  3.3409, -1.5507, -0.7583]], grad_fn=<MmBackward>)
rhs_data = tensor([[ 0.4175,  1.1364, -4.1467, -1.5625],
        [ 0.3793,  0.8213,  3.7015, -1.6664],
        [-0.2226, -1.3215,...95],
        [-0.3825,  1.0516, -3.8249, -1.5104],
        [-1.0052,  0.1070,  1.3424,  1.2628]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[4-1-5-g3-idtype1] _______________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'f_ni': Scheme(shape=(4,), dtype=torch.float32), 'f_nj': Scheme(shape=(4,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_node_feats = 5, out_edge_feats = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14ACAB48>
op = 'add'
lhs_data = tensor([[-0.2843,  0.3852,  0.1904, -1.4449],
        [ 2.8285, -0.4158, -2.3034, -0.3183],
        [-1.7569, -0.3601,...41],
        [-0.8128, -1.7875,  1.3410,  2.2222],
        [-4.4570, -1.5183,  1.3367,  2.5890]], grad_fn=<MmBackward>)
rhs_data = tensor([[ 0.3204, -0.3112,  0.2016, -0.3800],
        [ 0.0988,  1.4551,  0.2763, -1.8006],
        [-2.3225, -2.9573,...44],
        [-2.2619, -1.0183, -0.7870,  2.0058],
        [-0.1445,  1.1549, -1.7810,  0.2983]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[4-1-5-g4-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'f_ni': Scheme(shape=(4,), dtype=torch.float32), 'f_nj': Scheme(shape=(4,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_node_feats = 5, out_edge_feats = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14AB0888>
op = 'add'
lhs_data = tensor([[ 0.7403, -2.6611,  2.0726, -0.1620],
        [ 0.1937, -3.9163,  2.7279, -0.3411],
        [ 0.0943,  0.1539,...84],
        [-0.2559, -2.0848,  1.9536, -0.4397],
        [ 0.9811, -2.8739, -2.6099, -1.8018]], grad_fn=<MmBackward>)
rhs_data = tensor([[-1.9081,  4.1517,  2.4007,  3.4005],
        [-0.2704,  1.5704, -0.4303, -1.6884],
        [ 2.2988,  1.9768,...25],
        [-1.3410,  0.5842, -1.2046, -0.3630],
        [ 2.1658,  2.4787,  1.4649, -0.0976]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[4-1-5-g4-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'f_ni': Scheme(shape=(4,), dtype=torch.float32), 'f_nj': Scheme(shape=(4,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_node_feats = 5, out_edge_feats = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF017C2BC8>
op = 'add'
lhs_data = tensor([[-0.1629,  0.0876,  0.9421, -1.8273],
        [ 1.1288,  0.0362, -0.6557,  1.4832],
        [-1.5502, -1.4937,...43],
        [ 3.5529,  2.4599,  0.6221, -1.9668],
        [ 0.3306, -0.3822, -0.6408,  0.7038]], grad_fn=<MmBackward>)
rhs_data = tensor([[ 1.9198, -1.0479, -0.1992,  0.8821],
        [-1.1449, -0.6629, -2.5579, -1.4012],
        [-1.4192, -0.3162,...03],
        [-0.5087,  2.1324,  0.7303,  4.4128],
        [ 0.3148, -1.7595, -1.7210, -1.4954]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[4-1-5-g5-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'f_ni': Scheme(shape=(4,), dtype=torch.float32), 'f_nj': Scheme(shape=(4,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_node_feats = 5, out_edge_feats = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03CD63C8>
op = 'add'
lhs_data = tensor([[ 1.0631,  0.8938,  1.1200,  1.9751],
        [ 1.8232,  1.4729,  2.2568, -0.0371],
        [ 0.1926,  0.4803,...69],
        [-1.2921, -4.3989, -2.3871,  1.9124],
        [-0.8024,  1.5462, -0.3333, -1.3744]], grad_fn=<MmBackward>)
rhs_data = tensor([[-2.9503,  1.3268, -0.3853, -2.6746],
        [-0.1582, -0.2348,  1.0995,  0.3570],
        [-2.4720,  1.2746,...31],
        [ 0.6830,  3.4650, -3.8762, -0.8625],
        [-2.1446,  0.5766, -0.1155, -0.7594]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[4-1-5-g5-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'f_ni': Scheme(shape=(4,), dtype=torch.float32), 'f_nj': Scheme(shape=(4,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_node_feats = 5, out_edge_feats = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF018B1A08>
op = 'add'
lhs_data = tensor([[-1.0509, -0.5268,  0.0349, -0.6368],
        [ 0.8334, -0.3870, -1.3914,  2.7050],
        [-0.0502,  0.1761,...63],
        [ 0.8315,  0.5500, -0.2115, -0.5934],
        [ 0.5705,  0.8046,  0.5534, -0.3663]], grad_fn=<MmBackward>)
rhs_data = tensor([[ 0.5355, -1.1845, -2.1908,  0.4937],
        [ 0.0783,  1.6918, -1.5063, -1.1174],
        [ 0.1490,  1.2118,...79],
        [-0.7770, -0.1401, -2.1277,  0.7843],
        [ 1.7185, -2.3144, -0.0133, -0.5802]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[4-5-1-g0-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'f_ni': Scheme(sh..., 'f_nj': Scheme(shape=(20,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
idtype = torch.int32, out_node_feats = 1, out_edge_feats = 5, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0167B308>
op = 'add'
lhs_data = tensor([[ 2.3418e+00, -2.5311e+00,  2.4129e+00, -9.1697e-01,  2.8740e+00,
          9.7840e-02,  8.7723e-01, -3.5627e-...,  1.4117e+00,
         -1.1561e-02,  1.6876e-01,  1.0952e+00, -1.6553e+00,  9.4926e-01]],
       grad_fn=<MmBackward>)
rhs_data = tensor([[ 1.9326e+00,  2.1042e-01, -4.2754e-01, -9.8492e-01, -2.8375e-01,
          1.7402e-01, -1.4331e-01,  1.1612e+..., -1.6642e+00,
          7.5962e-01, -8.4145e-01,  2.4269e-01,  3.2455e-01, -1.1221e+00]],
       grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[4-5-1-g0-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'f_ni': Scheme(sh..., 'f_nj': Scheme(shape=(20,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
idtype = torch.int64, out_node_feats = 1, out_edge_feats = 5, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE145C00C8>
op = 'add'
lhs_data = tensor([[-1.0858, -1.4119,  0.3398,  1.8595, -0.6600, -0.6716,  1.5300,  0.9475,
          0.0602, -0.9475,  0.9366, -...1314,  0.0681,  0.3897,  1.3603, -0.0389, -0.3280,
         -0.0632,  2.1367, -0.2880, -0.9100]], grad_fn=<MmBackward>)
rhs_data = tensor([[ 5.1795e-02, -1.2399e+00,  6.2994e-01, -2.7987e+00, -1.4160e+00,
          5.1281e-01, -4.0056e-01,  3.3425e-..., -2.7116e-01,
         -2.0198e+00, -2.5785e-01,  1.7359e+00, -2.9486e-01, -1.0544e+00]],
       grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[4-5-1-g1-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'f_ni': Scheme(shape=(20,), dtype=torch.float32), 'f_nj': Scheme(shape=(20,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_node_feats = 1, out_edge_feats = 5, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14948488>
op = 'add'
lhs_data = tensor([[ 0.0592, -2.9019, -0.4141,  1.4615,  2.4652, -0.5613,  0.1637, -0.5940,
          0.3767,  1.6885,  0.4685, -...8940, -0.2568, -2.3942, -1.1186, -0.8192,  2.5201,
         -0.5623, -0.2988,  0.6815,  0.5842]], grad_fn=<MmBackward>)
rhs_data = tensor([[ 0.8483, -0.9579,  0.1241, -0.6175, -0.8686,  1.9778, -3.0819,  0.9716,
          1.8992,  0.4874,  0.8220,  ...5765,  0.6131, -0.9150,  1.1984, -1.6667,  0.8757,
          0.6425,  0.6694, -2.4310,  0.9502]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[4-5-1-g1-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'f_ni': Scheme(shape=(20,), dtype=torch.float32), 'f_nj': Scheme(shape=(20,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_node_feats = 1, out_edge_feats = 5, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE149AB588>
op = 'add'
lhs_data = tensor([[ 0.1659,  1.3182, -1.1311,  0.7880,  0.1638, -1.6882,  1.0765, -0.2963,
         -1.1422,  0.0067,  0.3497, -...8238, -0.2561, -0.2937, -0.0645, -0.1885, -0.1951,
         -1.3943,  1.5823,  0.0948,  1.1425]], grad_fn=<MmBackward>)
rhs_data = tensor([[ 1.1902e+00,  1.2980e-01,  2.5972e-01,  2.5151e+00,  9.7190e-01,
          1.3649e-01, -1.1707e+00, -5.0471e-...,  1.1449e+00,
         -2.9666e-01, -5.9871e-01, -1.4830e+00,  2.4847e-01, -1.5507e+00]],
       grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[4-5-1-g2-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'f_ni': Scheme(sh...nj': Scheme(shape=(20,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int32, out_node_feats = 1, out_edge_feats = 5, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01690BC8>
op = 'add'
lhs_data = tensor([[ 2.3720, -1.8935, -0.4162,  0.5520, -1.4524,  0.1448, -0.4728,  0.4514,
          1.7599, -0.8936,  0.0332,  ...5884, -1.5337,  0.3644, -1.2351,  0.0875,  0.0422,
         -0.6128,  3.5685, -1.8449,  0.5394]], grad_fn=<MmBackward>)
rhs_data = tensor([[ 6.3970e-01,  1.6601e+00, -2.3311e+00, -2.0783e+00, -6.8123e-01,
         -4.8562e-01, -1.8884e+00, -2.5141e-...,  5.2333e-01,
          9.9181e-01, -1.7940e+00, -7.5259e-01,  1.7487e+00,  1.8527e+00]],
       grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[4-5-1-g2-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'f_ni': Scheme(sh...nj': Scheme(shape=(20,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int64, out_node_feats = 1, out_edge_feats = 5, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14AA2D88>
op = 'add'
lhs_data = tensor([[-4.4231e-01, -1.5895e+00,  1.0512e+00,  1.6224e-01, -1.7813e+00,
         -7.0237e-01,  6.1579e-01, -3.8878e-...,  1.1537e+00,
          5.5554e-01, -3.6554e-01, -1.3466e+00, -1.6178e-01, -6.8565e-01]],
       grad_fn=<MmBackward>)
rhs_data = tensor([[-0.5089,  1.7619, -1.4550,  2.9052, -1.6036, -2.6163,  1.5313, -0.4541,
          0.0209,  0.6515,  0.0631, -...5678,  0.6605, -0.4022, -0.2934,  0.0829,  1.2738,
          0.5716,  0.4643, -0.1542, -0.7089]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[4-5-1-g3-idtype0] _______________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'f_ni': Scheme(shape=(20,), dtype=torch.float32), 'f_nj': Scheme(shape=(20,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_node_feats = 1, out_edge_feats = 5, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF011D3E08>
op = 'add'
lhs_data = tensor([[ 1.4051e+00, -1.6456e+00,  8.6499e-01,  2.1046e+00, -3.2655e-01,
          6.8975e-01,  1.7024e+00,  3.4062e-...,  3.7050e-02,
          1.8741e+00, -5.0038e-01, -3.1436e-01,  7.5653e-01, -2.1477e-01]],
       grad_fn=<MmBackward>)
rhs_data = tensor([[ 0.6566,  1.5460, -0.9208,  0.5337, -1.3848,  0.9384,  0.3511, -0.3051,
          1.3190, -0.3054,  1.9916, -...4495, -1.5011,  0.7656,  1.2375,  0.8733,  1.2703,
         -0.1096, -0.3807, -2.6003, -0.3046]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[4-5-1-g3-idtype1] _______________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'f_ni': Scheme(shape=(20,), dtype=torch.float32), 'f_nj': Scheme(shape=(20,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_node_feats = 1, out_edge_feats = 5, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01756088>
op = 'add'
lhs_data = tensor([[ 1.5827, -0.7517,  2.2143,  0.1396,  1.5923, -0.2242,  1.5345, -0.5487,
          2.2328,  2.6643,  1.1048, -...1692,  0.9636, -0.8558, -0.9104,  0.5204,  0.7929,
          0.2878, -1.3475, -0.7272,  2.6748]], grad_fn=<MmBackward>)
rhs_data = tensor([[-0.6046,  0.8751,  1.5239,  0.3035,  2.6308,  0.1682, -0.1261, -1.8990,
          0.5793,  1.5350, -0.2932,  ...0225,  0.7329,  1.7413,  0.2510,  2.4960,  3.2739,
         -2.0400,  0.7201,  0.9248,  0.5579]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[4-5-1-g4-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'f_ni': Scheme(shape=(20,), dtype=torch.float32), 'f_nj': Scheme(shape=(20,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_node_feats = 1, out_edge_feats = 5, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0183BDC8>
op = 'add'
lhs_data = tensor([[ 2.1489e+00, -9.4475e-01, -6.9973e-01,  2.7737e+00,  4.2542e-01,
          6.1822e-01, -5.8451e-01, -1.7896e-...,  2.1459e+00,
         -2.0947e+00,  1.7494e+00, -1.2161e+00,  1.3367e-01, -2.2884e+00]],
       grad_fn=<MmBackward>)
rhs_data = tensor([[ 0.6186,  0.2262,  0.0786,  1.3415, -0.3437,  2.1063,  0.7720, -0.2089,
          2.4058, -2.0210,  1.1341,  ...1026, -0.8183, -0.5657,  0.5970, -0.8751, -2.0710,
         -0.1378,  0.3012,  1.9311, -0.5781]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[4-5-1-g4-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'f_ni': Scheme(shape=(20,), dtype=torch.float32), 'f_nj': Scheme(shape=(20,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_node_feats = 1, out_edge_feats = 5, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01545D08>
op = 'add'
lhs_data = tensor([[ 0.1110,  0.8908,  0.8170,  1.6353,  0.5071,  1.1498,  0.8443, -0.2035,
          0.3475,  0.1749, -1.2598, -...0959, -0.1554, -1.0228, -0.8037,  0.8342, -0.9882,
         -1.3433, -0.9329, -0.5659,  1.2995]], grad_fn=<MmBackward>)
rhs_data = tensor([[ 0.4166, -1.2834, -0.6489,  0.4947,  1.1288,  0.1873,  0.4988, -0.1183,
          1.0769, -1.0882, -0.4656,  ...1004,  0.5031,  0.1384,  1.4942,  0.2226, -0.0404,
         -0.4624, -0.7095,  1.2584,  0.6456]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[4-5-1-g5-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'f_ni': Scheme(shape=(20,), dtype=torch.float32), 'f_nj': Scheme(shape=(20,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_node_feats = 1, out_edge_feats = 5, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE145B07C8>
op = 'add'
lhs_data = tensor([[ 0.8046,  0.7871, -0.7802, -1.1572, -0.6657,  1.3246, -0.4318, -0.8336,
          0.0958, -0.8787,  1.0474, -...8857,  2.1098,  0.3639,  0.9637,  1.7194, -0.0812,
          4.0037,  0.3620, -2.5374, -1.3716]], grad_fn=<MmBackward>)
rhs_data = tensor([[-0.1920, -0.7119,  0.9303, -1.0677,  1.7210,  0.4289, -1.0316,  1.0797,
          0.6077,  0.3135, -0.7021, -...0760,  0.8812, -1.4685,  0.5581, -0.2039,  0.6168,
          0.9647, -0.1308, -0.8288, -1.9422]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[4-5-1-g5-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'f_ni': Scheme(shape=(20,), dtype=torch.float32), 'f_nj': Scheme(shape=(20,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_node_feats = 1, out_edge_feats = 5, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14CCFAC8>
op = 'add'
lhs_data = tensor([[-4.0623e-01, -8.9916e-01, -1.3417e+00,  6.8556e-01,  1.3740e-01,
         -4.4514e-01, -9.5944e-01,  9.3194e-..., -4.0303e-01,
         -2.3050e+00,  1.2788e+00,  1.2249e+00,  1.3205e+00,  3.2922e-01]],
       grad_fn=<MmBackward>)
rhs_data = tensor([[-1.5288e-01,  3.5973e-01,  7.5185e-01,  6.4194e-01, -6.3932e-01,
         -1.0915e+00,  1.3648e-01,  7.8767e-..., -3.1102e+00,
         -1.4462e+00, -1.7412e-01, -2.0356e+00,  5.6373e-01, -5.1349e-01]],
       grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[4-5-5-g0-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'f_ni': Scheme(sh..., 'f_nj': Scheme(shape=(20,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
idtype = torch.int32, out_node_feats = 5, out_edge_feats = 5, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0165AC08>
op = 'add'
lhs_data = tensor([[-6.8408e-01, -4.4241e-01,  9.9652e-01, -8.2476e-01,  2.4521e+00,
         -6.4895e-01, -3.7110e-01,  9.6443e-..., -2.9267e-02,
          2.3671e-01, -1.0841e-01,  1.1537e+00,  9.7124e-02,  5.2890e-01]],
       grad_fn=<MmBackward>)
rhs_data = tensor([[ 0.5928,  0.9564,  0.4271,  3.6479, -2.3172, -0.6711, -0.1610,  2.0091,
          1.4395, -1.1904,  0.9926,  ...1439,  1.3852, -1.7436, -1.4168,  0.4069, -2.2603,
         -0.1702, -1.2209,  1.0660,  0.5498]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[4-5-5-g0-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'f_ni': Scheme(sh..., 'f_nj': Scheme(shape=(20,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
idtype = torch.int64, out_node_feats = 5, out_edge_feats = 5, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF015453C8>
op = 'add'
lhs_data = tensor([[-1.5943e+00, -2.9214e-01, -1.8768e+00, -9.2384e-01, -2.1452e-01,
          2.1503e-01,  9.5467e-01,  1.1005e+..., -7.5066e-01,
         -3.2543e-01,  5.8007e-01, -1.2702e+00,  5.2915e-01, -7.8763e-02]],
       grad_fn=<MmBackward>)
rhs_data = tensor([[ 1.5485e+00,  3.0599e-01, -1.1678e-01, -9.8203e-01,  6.5697e-02,
          1.2137e-01, -6.3660e-01,  6.7201e-..., -1.4764e-01,
         -2.9830e-01, -1.0045e-01, -8.2377e-02, -6.7796e-02, -1.0676e-01]],
       grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[4-5-5-g1-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'f_ni': Scheme(shape=(20,), dtype=torch.float32), 'f_nj': Scheme(shape=(20,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_node_feats = 5, out_edge_feats = 5, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01539888>
op = 'add'
lhs_data = tensor([[ 0.8545, -3.3729,  0.5817,  0.6089, -1.8478, -1.7260, -1.2018, -0.8049,
         -1.2596, -2.1611,  1.2954,  ...3658,  1.0334, -1.2622,  0.6242,  1.5942,  0.7616,
         -2.2061, -0.1066, -1.7449, -1.9504]], grad_fn=<MmBackward>)
rhs_data = tensor([[-1.9390,  1.3609, -1.1060,  1.1583, -0.7940, -0.7214, -1.2001, -0.7164,
          1.8448, -0.3010,  1.0260, -...4955, -0.6047, -0.1624,  0.4444, -0.3747,  0.4804,
         -1.4386, -0.8819, -1.4224,  0.9308]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[4-5-5-g1-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'f_ni': Scheme(shape=(20,), dtype=torch.float32), 'f_nj': Scheme(shape=(20,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_node_feats = 5, out_edge_feats = 5, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03CBEB08>
op = 'add'
lhs_data = tensor([[ 1.4643,  0.3809,  0.6707,  1.5284,  1.6854, -0.4055, -1.1722,  2.2317,
          0.6627,  0.4365,  1.1575,  ...7317, -1.0370, -0.2921, -0.9101, -0.1197,  0.8746,
         -0.6404,  1.2474,  0.4315,  1.5950]], grad_fn=<MmBackward>)
rhs_data = tensor([[ 6.3378e-01, -1.2730e+00, -2.4526e+00, -1.1860e+00, -1.4681e+00,
          2.0937e+00, -1.0424e+00, -4.3698e-...,  9.4171e-01,
         -2.1466e-01, -4.3827e-01, -6.7339e-01,  8.0752e-01,  1.6748e-01]],
       grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[4-5-5-g2-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'f_ni': Scheme(sh...nj': Scheme(shape=(20,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int32, out_node_feats = 5, out_edge_feats = 5, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0189E708>
op = 'add'
lhs_data = tensor([[ 6.1441e-01, -1.1214e+00, -4.1723e-01,  5.8946e-01, -1.3968e+00,
          7.2983e-01,  2.9579e-01,  1.1614e+...,  9.1684e-01,
         -8.6391e-01,  1.7200e+00,  1.3682e+00, -1.9824e+00, -3.4557e-01]],
       grad_fn=<MmBackward>)
rhs_data = tensor([[-1.2485, -0.2628,  0.4073,  0.0113,  1.2264, -0.9985,  1.2004,  1.1076,
         -1.0174, -0.6624,  0.3301, -...5793, -1.7182, -1.2066, -2.0637,  0.4166,  0.4039,
          0.4172, -1.8494,  1.2681,  0.2591]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[4-5-5-g2-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'f_ni': Scheme(sh...nj': Scheme(shape=(20,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int64, out_node_feats = 5, out_edge_feats = 5, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1462EBC8>
op = 'add'
lhs_data = tensor([[-7.7342e-01, -1.3228e-02, -2.4738e+00, -2.7465e+00,  2.3475e-01,
          1.2156e+00, -2.1778e+00, -9.6917e-...,  8.9174e-02,
          2.0534e-01, -1.0169e+00, -5.6891e-01, -1.7862e+00, -1.7301e+00]],
       grad_fn=<MmBackward>)
rhs_data = tensor([[ 6.5833e-01,  2.2887e+00,  5.7301e-01, -7.5700e-02, -1.6388e+00,
          3.5509e+00, -7.7818e-01,  8.8199e-..., -1.6652e+00,
         -1.3639e+00,  6.2213e-01,  1.8194e+00, -5.9530e-01,  6.4701e-01]],
       grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[4-5-5-g3-idtype0] _______________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'f_ni': Scheme(shape=(20,), dtype=torch.float32), 'f_nj': Scheme(shape=(20,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_node_feats = 5, out_edge_feats = 5, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01888288>
op = 'add'
lhs_data = tensor([[ 1.8960,  0.6549, -1.6366,  0.1678, -0.6300,  2.4284, -2.3742, -0.9172,
          1.5790, -1.0337, -1.6099,  ...6177,  0.0981,  3.0807,  1.4327, -1.6914,  0.9035,
          2.6730,  0.0211,  0.2691,  0.8100]], grad_fn=<MmBackward>)
rhs_data = tensor([[ 1.0084e-01, -1.6607e+00, -2.4339e+00, -2.6906e-01, -2.5486e-01,
         -1.4958e+00, -5.3673e-01,  1.2756e-...,  7.5038e-01,
         -1.4541e+00,  3.1389e-01,  1.4213e+00,  1.3720e+00,  1.0100e+00]],
       grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[4-5-5-g3-idtype1] _______________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'f_ni': Scheme(shape=(20,), dtype=torch.float32), 'f_nj': Scheme(shape=(20,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_node_feats = 5, out_edge_feats = 5, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01741588>
op = 'add'
lhs_data = tensor([[ 2.0281e+00, -8.4660e-01,  3.2056e-01,  8.7663e-02, -2.1813e+00,
          4.2587e-01,  2.2207e-02,  2.9588e+..., -1.1805e+00,
          6.4614e-01, -6.1611e-01,  2.0874e-01, -2.8085e-01,  1.4730e+00]],
       grad_fn=<MmBackward>)
rhs_data = tensor([[ 4.3329e-03,  2.0706e-01, -3.4503e-01, -2.4975e+00,  1.0312e-01,
          3.8880e+00,  1.0967e+00, -9.5148e-...,  1.8279e-01,
          1.1232e+00,  1.0245e+00, -2.1430e+00, -1.1878e+00,  3.6890e-01]],
       grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[4-5-5-g4-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'f_ni': Scheme(shape=(20,), dtype=torch.float32), 'f_nj': Scheme(shape=(20,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_node_feats = 5, out_edge_feats = 5, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF019A1348>
op = 'add'
lhs_data = tensor([[-7.0024e-01, -2.1143e+00,  9.4483e-01, -1.0089e+00, -4.3028e-01,
          1.0309e+00,  9.6889e-01,  1.2470e+...,  2.6410e-01,
         -9.4642e-01, -1.0746e+00,  2.1012e+00, -7.6031e-01, -8.9569e-01]],
       grad_fn=<MmBackward>)
rhs_data = tensor([[-1.0757, -0.1064, -0.1964, -0.5468, -1.3376,  0.0035, -0.1566, -0.2380,
          0.1809, -0.2126, -0.5209, -...2021, -2.3571, -0.3253,  0.2685, -0.0733, -0.7760,
         -0.2067,  0.8493,  2.6851, -0.6573]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[4-5-5-g4-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'f_ni': Scheme(shape=(20,), dtype=torch.float32), 'f_nj': Scheme(shape=(20,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_node_feats = 5, out_edge_feats = 5, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03CC67C8>
op = 'add'
lhs_data = tensor([[ 0.0632, -0.0524,  0.9609, -0.7443,  1.4727,  1.0182, -0.3096, -1.4167,
         -2.1141,  0.7572, -0.9319,  ...6902,  0.9121,  1.1594, -0.0607,  0.2103, -0.7947,
          1.5204, -0.5483, -1.9902,  0.1371]], grad_fn=<MmBackward>)
rhs_data = tensor([[-0.7836,  0.1587, -0.4091, -1.9545, -0.4315, -0.3215,  1.9764,  1.7557,
          1.8379,  0.1764,  0.6760, -...0893,  0.2386, -0.5311,  1.3889, -0.8549, -0.1964,
         -0.5495, -0.8813, -1.7497,  1.2059]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[4-5-5-g5-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'f_ni': Scheme(shape=(20,), dtype=torch.float32), 'f_nj': Scheme(shape=(20,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_node_feats = 5, out_edge_feats = 5, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF011F05C8>
op = 'add'
lhs_data = tensor([[-0.7611, -1.6732,  1.6232, -0.2803, -0.2467,  0.1703,  0.3493, -0.2968,
          2.3206, -1.0701,  2.6926, -...6046, -1.1256, -0.8574, -0.1569,  0.6474,  1.0157,
         -0.8739, -0.1974,  0.2860, -0.7708]], grad_fn=<MmBackward>)
rhs_data = tensor([[-8.2038e-01, -1.2606e+00,  6.6014e-02,  8.6488e-01, -2.9163e+00,
         -7.5656e-01, -5.7392e-02,  4.2107e-...,  1.1068e-01,
          9.3745e-02,  8.5721e-01, -1.8557e-01,  1.0382e-01, -7.3513e-01]],
       grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_egat_conv[4-5-5-g5-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'f_ni': Scheme(shape=(20,), dtype=torch.float32), 'f_nj': Scheme(shape=(20,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_node_feats = 5, out_edge_feats = 5, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=10,
                           in_edge_feats=5,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = F.randn((g.number_of_nodes(), 10))
        efeat = F.randn((g.number_of_edges(), 5))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:564: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF016E2C08>
op = 'add'
lhs_data = tensor([[-2.0328e+00, -2.8940e+00,  1.0430e-01, -6.0096e-01, -3.2215e-01,
          9.1139e-01, -2.6772e+00, -8.2305e-...,  4.1666e-01,
          2.2483e+00, -6.4060e-01, -3.6514e-01, -1.1272e+00,  1.8958e-01]],
       grad_fn=<MmBackward>)
rhs_data = tensor([[-2.9879e-01, -3.1880e-01, -1.1936e+00, -9.3224e-01,  9.8483e-01,
         -8.1061e-01,  8.9513e-01,  3.1228e+..., -4.9517e-01,
          3.7901e-01,  1.8508e+00,  3.1310e-01, -2.7327e-01, -1.2691e-01]],
       grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_____________________ test_egat_conv_bi[1-1-1-g0-idtype0] _____________________

g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int32, out_node_feats = 1, out_edge_feats = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv_bi(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=(10,15),
                           in_edge_feats=7,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), 15)))
        efeat = F.randn((g.number_of_edges(), 7))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:589: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03D0C788>
op = 'add'
lhs_data = tensor([[ 2.8240],
        [-2.1112]], grad_fn=<MmBackward>)
rhs_data = tensor([[-0.7571],
        [ 0.5796],
        [-0.9385],
        [-2.8028]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_____________________ test_egat_conv_bi[1-1-1-g0-idtype1] _____________________

g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int64, out_node_feats = 1, out_edge_feats = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv_bi(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=(10,15),
                           in_edge_feats=7,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), 15)))
        efeat = F.randn((g.number_of_edges(), 7))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:589: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF011D0388>
op = 'add'
lhs_data = tensor([[-0.3210],
        [ 1.6125]], grad_fn=<MmBackward>)
rhs_data = tensor([[-1.1213],
        [ 2.6075],
        [ 1.9127],
        [ 0.4586]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_____________________ test_egat_conv_bi[1-1-1-g1-idtype0] _____________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int32
out_node_feats = 1, out_edge_feats = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv_bi(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=(10,15),
                           in_edge_feats=7,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), 15)))
        efeat = F.randn((g.number_of_edges(), 7))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:589: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF018CEA48>
op = 'add'
lhs_data = tensor([[-3.0569],
        [ 0.7531],
        [ 0.4107],
        [-0.8140],
        [ 5.4511],
        [-0.4450]], grad_fn=<MmBackward>)
rhs_data = tensor([[-0.5495],
        [ 1.7457],
        [ 3.0253]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_____________________ test_egat_conv_bi[1-1-1-g1-idtype1] _____________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int64
out_node_feats = 1, out_edge_feats = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv_bi(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=(10,15),
                           in_edge_feats=7,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), 15)))
        efeat = F.randn((g.number_of_edges(), 7))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:589: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF016794C8>
op = 'add'
lhs_data = tensor([[-1.1048],
        [-3.5164],
        [-1.3283],
        [-2.1665],
        [-4.5384],
        [ 2.3528]], grad_fn=<MmBackward>)
rhs_data = tensor([[2.3836],
        [0.4488],
        [3.1381]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_____________________ test_egat_conv_bi[1-1-5-g0-idtype0] _____________________

g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int32, out_node_feats = 5, out_edge_feats = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv_bi(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=(10,15),
                           in_edge_feats=7,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), 15)))
        efeat = F.randn((g.number_of_edges(), 7))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:589: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03CB8608>
op = 'add'
lhs_data = tensor([[ 0.9555],
        [-0.1825]], grad_fn=<MmBackward>)
rhs_data = tensor([[-1.1969],
        [-0.8480],
        [ 0.0436],
        [ 1.7280]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_____________________ test_egat_conv_bi[1-1-5-g0-idtype1] _____________________

g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int64, out_node_feats = 5, out_edge_feats = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv_bi(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=(10,15),
                           in_edge_feats=7,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), 15)))
        efeat = F.randn((g.number_of_edges(), 7))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:589: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01890888>
op = 'add'
lhs_data = tensor([[-1.2543],
        [ 0.5183]], grad_fn=<MmBackward>)
rhs_data = tensor([[-1.5706],
        [-1.0798],
        [ 0.0327],
        [-2.8392]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_____________________ test_egat_conv_bi[1-1-5-g1-idtype0] _____________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int32
out_node_feats = 5, out_edge_feats = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv_bi(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=(10,15),
                           in_edge_feats=7,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), 15)))
        efeat = F.randn((g.number_of_edges(), 7))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:589: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0148DBC8>
op = 'add'
lhs_data = tensor([[-0.4686],
        [ 1.9217],
        [-2.6619],
        [-1.0409],
        [ 0.2924],
        [-2.1099]], grad_fn=<MmBackward>)
rhs_data = tensor([[3.0384],
        [2.9239],
        [1.5217]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_____________________ test_egat_conv_bi[1-1-5-g1-idtype1] _____________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int64
out_node_feats = 5, out_edge_feats = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv_bi(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=(10,15),
                           in_edge_feats=7,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), 15)))
        efeat = F.randn((g.number_of_edges(), 7))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:589: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14CCFB08>
op = 'add'
lhs_data = tensor([[ 1.1670],
        [ 0.2266],
        [ 1.2627],
        [ 0.7198],
        [-0.1601],
        [-1.8073]], grad_fn=<MmBackward>)
rhs_data = tensor([[ 2.8252],
        [-1.1598],
        [-3.1703]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_____________________ test_egat_conv_bi[1-5-1-g0-idtype0] _____________________

g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int32, out_node_feats = 1, out_edge_feats = 5, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv_bi(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=(10,15),
                           in_edge_feats=7,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), 15)))
        efeat = F.randn((g.number_of_edges(), 7))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:589: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14C34088>
op = 'add'
lhs_data = tensor([[ 0.3146,  2.0922, -1.2694, -0.0680, -0.7776],
        [-1.0541, -0.4320, -1.4617, -2.2023,  0.1421]], grad_fn=<MmBackward>)
rhs_data = tensor([[-1.0656, -0.0955, -0.4764,  1.2581,  1.5510],
        [-1.8763, -1.9493,  0.3834, -0.4012,  1.2087],
        ...614,  0.2427,  0.1690,  0.7810,  0.6016],
        [-3.1626, -2.6896,  1.4366,  1.1721, -0.5961]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_____________________ test_egat_conv_bi[1-5-1-g0-idtype1] _____________________

g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int64, out_node_feats = 1, out_edge_feats = 5, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv_bi(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=(10,15),
                           in_edge_feats=7,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), 15)))
        efeat = F.randn((g.number_of_edges(), 7))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:589: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0109AF08>
op = 'add'
lhs_data = tensor([[-0.2371,  0.2344,  3.8942, -1.2320,  1.0402],
        [-1.5102, -0.1980, -2.0674,  1.0480, -1.8496]], grad_fn=<MmBackward>)
rhs_data = tensor([[ 2.2017e-01,  3.4319e-01,  2.4685e+00, -2.3145e+00,  3.5856e+00],
        [ 8.9920e-02,  1.2209e+00,  3.5252e... -2.1515e+00],
        [-1.2832e+00, -2.4450e+00, -1.0793e+00,  1.7236e+00, -1.9500e+00]],
       grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_____________________ test_egat_conv_bi[1-5-1-g1-idtype0] _____________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int32
out_node_feats = 1, out_edge_feats = 5, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv_bi(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=(10,15),
                           in_edge_feats=7,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), 15)))
        efeat = F.randn((g.number_of_edges(), 7))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:589: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03D2F408>
op = 'add'
lhs_data = tensor([[ 1.0916, -1.4950, -0.7666, -0.4751,  0.5945],
        [-0.3357,  0.3355,  0.2148,  2.9543,  0.0321],
        ...161,  2.2725, -2.6201, -0.4652, -1.8011],
        [-0.3900, -1.0558,  0.7978,  0.7933,  5.4721]], grad_fn=<MmBackward>)
rhs_data = tensor([[-0.9215,  0.8658, -2.3865,  1.3820, -0.6142],
        [ 0.4072,  1.6119,  4.1919, -1.5234,  0.4493],
        [-0.8402, -0.2129, -4.0229, -3.5910, -0.1222]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_____________________ test_egat_conv_bi[1-5-1-g1-idtype1] _____________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int64
out_node_feats = 1, out_edge_feats = 5, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv_bi(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=(10,15),
                           in_edge_feats=7,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), 15)))
        efeat = F.randn((g.number_of_edges(), 7))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:589: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01959948>
op = 'add'
lhs_data = tensor([[ 0.7888, -1.2142,  0.2092, -3.8654,  0.4653],
        [-0.5668,  0.8523, -1.1037, -0.9767,  2.9824],
        ...872, -0.8212,  0.1097, -1.7158,  1.7991],
        [-1.1656,  0.6755,  0.8520, -1.2901, -4.5112]], grad_fn=<MmBackward>)
rhs_data = tensor([[-0.2197, -1.6336, -3.5178, -0.2170,  1.6726],
        [ 0.2573,  4.4163, -1.4093,  1.1220,  1.7720],
        [-0.4822,  2.2817,  1.0312, -1.8688,  1.0071]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_____________________ test_egat_conv_bi[1-5-5-g0-idtype0] _____________________

g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int32, out_node_feats = 5, out_edge_feats = 5, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv_bi(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=(10,15),
                           in_edge_feats=7,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), 15)))
        efeat = F.randn((g.number_of_edges(), 7))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:589: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01723B48>
op = 'add'
lhs_data = tensor([[-2.2929,  1.3306,  1.2237,  2.6639,  3.3122],
        [-1.5867,  1.2147,  4.4359, -0.5106, -1.9450]], grad_fn=<MmBackward>)
rhs_data = tensor([[ 3.8562, -2.1296,  0.0457,  1.3962, -0.3416],
        [-3.0199,  0.2560,  0.9294, -3.2362, -0.7833],
        ...829, -0.0402, -2.3186,  1.4928, -0.4964],
        [-0.2262,  3.9483, -0.3972, -2.3051,  0.6604]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_____________________ test_egat_conv_bi[1-5-5-g0-idtype1] _____________________

g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int64, out_node_feats = 5, out_edge_feats = 5, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv_bi(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=(10,15),
                           in_edge_feats=7,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), 15)))
        efeat = F.randn((g.number_of_edges(), 7))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:589: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF015DE6C8>
op = 'add'
lhs_data = tensor([[-1.6821,  1.9085, -1.0450,  0.9774, -2.2073],
        [ 1.6780, -1.2726, -2.6771, -1.5571,  0.6557]], grad_fn=<MmBackward>)
rhs_data = tensor([[-2.1100, -1.8739, -1.0670,  1.2892, -2.3772],
        [-0.1286, -0.0339, -1.1950, -0.3152,  0.2632],
        ...022,  0.6798,  2.1355, -0.6326, -1.3051],
        [-1.6236, -2.0882, -0.6534, -2.7557, -0.2837]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_____________________ test_egat_conv_bi[1-5-5-g1-idtype0] _____________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int32
out_node_feats = 5, out_edge_feats = 5, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv_bi(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=(10,15),
                           in_edge_feats=7,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), 15)))
        efeat = F.randn((g.number_of_edges(), 7))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:589: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF015AAE08>
op = 'add'
lhs_data = tensor([[ 0.9805,  1.0242, -0.3608,  0.6343,  0.9700],
        [-0.9685, -0.9643, -2.3260, -0.6701,  0.8226],
        ...471, -0.4268,  0.4982,  3.3154, -0.1516],
        [-0.6576, -0.4776,  0.7304, -2.0443, -1.5253]], grad_fn=<MmBackward>)
rhs_data = tensor([[-0.8454, -0.9631, -0.3690, -0.6729, -0.2060],
        [-0.3451, -1.3987,  0.1778, -1.9388,  0.5186],
        [ 0.2851, -0.9529, -1.6814, -0.5284,  1.5315]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_____________________ test_egat_conv_bi[1-5-5-g1-idtype1] _____________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int64
out_node_feats = 5, out_edge_feats = 5, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv_bi(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=(10,15),
                           in_edge_feats=7,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), 15)))
        efeat = F.randn((g.number_of_edges(), 7))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:589: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03D15308>
op = 'add'
lhs_data = tensor([[-2.0645,  0.8260, -1.3832,  1.7390, -0.2179],
        [ 1.5741,  1.8801,  1.8280, -2.6369,  2.7830],
        ...763,  0.5599,  1.3475,  2.1505,  1.6288],
        [ 0.4243,  2.7415,  1.0676, -0.2745,  1.3554]], grad_fn=<MmBackward>)
rhs_data = tensor([[-3.1130, -1.0786, -1.7205,  0.5425, -1.2976],
        [-1.4849,  1.0256, -2.4215,  3.2899, -2.9724],
        [-1.0300,  0.2692, -0.4148,  0.7324, -0.6202]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_____________________ test_egat_conv_bi[4-1-1-g0-idtype0] _____________________

g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int32, out_node_feats = 1, out_edge_feats = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv_bi(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=(10,15),
                           in_edge_feats=7,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), 15)))
        efeat = F.randn((g.number_of_edges(), 7))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:589: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE146AB248>
op = 'add'
lhs_data = tensor([[-0.1962, -1.7739,  0.8247,  0.0619],
        [-0.4716, -1.6367, -1.1099, -1.4762]], grad_fn=<MmBackward>)
rhs_data = tensor([[ 1.3422, -0.9957,  0.1040, -0.7622],
        [ 0.3082, -0.4750,  0.8199,  0.7947],
        [ 0.8280, -4.0873, -0.0076, -2.0162],
        [ 1.0438,  0.4035, -0.0360,  1.9571]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_____________________ test_egat_conv_bi[4-1-1-g0-idtype1] _____________________

g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int64, out_node_feats = 1, out_edge_feats = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv_bi(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=(10,15),
                           in_edge_feats=7,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), 15)))
        efeat = F.randn((g.number_of_edges(), 7))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:589: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0180C208>
op = 'add'
lhs_data = tensor([[ 0.8192,  0.2947, -0.2068, -1.5144],
        [-4.5291,  3.0066, -0.7048, -0.8271]], grad_fn=<MmBackward>)
rhs_data = tensor([[-2.6707, -2.7875, -3.1676,  3.0346],
        [ 0.5352,  2.5069, -0.3904, -2.1632],
        [-2.3969,  0.5511, -3.9606,  0.0391],
        [-0.8857,  1.0355, -0.9382, -1.2227]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_____________________ test_egat_conv_bi[4-1-1-g1-idtype0] _____________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int32
out_node_feats = 1, out_edge_feats = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv_bi(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=(10,15),
                           in_edge_feats=7,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), 15)))
        efeat = F.randn((g.number_of_edges(), 7))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:589: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0148D088>
op = 'add'
lhs_data = tensor([[ 0.6757, -0.4633,  0.5540,  1.0501],
        [ 0.1781, -0.8037,  1.2925, -1.4955],
        [-1.4085,  0.3828,...84],
        [-0.1253, -0.5181,  1.7059, -2.7104],
        [-0.3109, -1.4386,  0.1302,  0.0192]], grad_fn=<MmBackward>)
rhs_data = tensor([[-0.2652,  0.7894,  0.1817,  0.4694],
        [-0.7414, -0.3457, -0.0176,  0.2824],
        [ 1.0822,  0.1103,  0.3402, -0.5809]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_____________________ test_egat_conv_bi[4-1-1-g1-idtype1] _____________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int64
out_node_feats = 1, out_edge_feats = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv_bi(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=(10,15),
                           in_edge_feats=7,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), 15)))
        efeat = F.randn((g.number_of_edges(), 7))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:589: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07703AC8>
op = 'add'
lhs_data = tensor([[-0.4065, -0.7620,  1.0641, -2.9007],
        [ 2.0265,  1.9998, -1.1969, -0.8955],
        [-1.4233, -1.3512,...41],
        [ 1.3946,  1.1776,  0.3965, -0.1368],
        [ 1.5669,  1.9249, -0.7914,  2.0394]], grad_fn=<MmBackward>)
rhs_data = tensor([[ 2.8247, -2.0020, -0.0164,  1.3055],
        [-1.7450,  0.5076, -0.8277,  2.4379],
        [ 2.0831,  0.5127, -2.3217, -0.0316]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_____________________ test_egat_conv_bi[4-1-5-g0-idtype0] _____________________

g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int32, out_node_feats = 5, out_edge_feats = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv_bi(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=(10,15),
                           in_edge_feats=7,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), 15)))
        efeat = F.randn((g.number_of_edges(), 7))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:589: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14A82908>
op = 'add'
lhs_data = tensor([[-0.2626, -1.4545,  1.2638, -0.8761],
        [ 2.0661, -1.7082,  1.0934, -1.4949]], grad_fn=<MmBackward>)
rhs_data = tensor([[-0.1110,  1.7318, -1.5468, -1.7342],
        [ 2.8438,  0.5388, -2.1282,  0.6741],
        [ 2.6437, -1.8629,  1.1669, -0.9728],
        [ 1.0125,  0.9813,  0.3377,  1.0628]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_____________________ test_egat_conv_bi[4-1-5-g0-idtype1] _____________________

g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int64, out_node_feats = 5, out_edge_feats = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv_bi(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=(10,15),
                           in_edge_feats=7,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), 15)))
        efeat = F.randn((g.number_of_edges(), 7))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:589: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03D2F108>
op = 'add'
lhs_data = tensor([[-0.6508,  3.3676, -1.8081,  2.1275],
        [ 0.0592,  0.3606,  0.5148, -0.1441]], grad_fn=<MmBackward>)
rhs_data = tensor([[-0.6177,  0.2453, -2.3402, -1.2833],
        [-2.9738, -0.0109, -1.9707,  0.7520],
        [-0.5154,  1.0631, -1.7359, -0.3899],
        [ 2.2389,  0.5466,  1.6230,  0.8275]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_____________________ test_egat_conv_bi[4-1-5-g1-idtype0] _____________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int32
out_node_feats = 5, out_edge_feats = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv_bi(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=(10,15),
                           in_edge_feats=7,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), 15)))
        efeat = F.randn((g.number_of_edges(), 7))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:589: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01809B48>
op = 'add'
lhs_data = tensor([[ 0.0990,  1.2184,  0.4267, -0.5217],
        [-2.5857, -0.6303, -1.6209, -0.2791],
        [ 0.4903,  3.1664,...10],
        [-0.0808,  0.9886, -0.2650,  1.4255],
        [-2.0096, -2.1572, -0.5341, -0.2613]], grad_fn=<MmBackward>)
rhs_data = tensor([[ 1.8884, -1.6914, -1.1863, -2.7802],
        [-0.4494,  1.7968,  0.2263,  2.5877],
        [ 1.3491, -0.7064, -2.3377,  0.4425]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_____________________ test_egat_conv_bi[4-1-5-g1-idtype1] _____________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int64
out_node_feats = 5, out_edge_feats = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv_bi(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=(10,15),
                           in_edge_feats=7,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), 15)))

        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:589: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE145F7F88>
op = 'add'
lhs_data = tensor([[ 1.2024, -0.1988,  0.8150, -3.4943],
        [ 1.4453,  0.7416, -1.7601,  1.6785],
        [-1.1386, -1.1007,...94],
        [-0.0228, -1.1398, -2.9624, -4.8255],
        [ 0.3311, -0.8166, -4.0241, -2.7917]], grad_fn=<MmBackward>)
rhs_data = tensor([[-1.3706,  0.2851, -2.7871,  0.5717],
        [-2.8678, -3.6779, -0.0455, -0.5709],
        [-1.8704, -1.1826, -2.3336,  1.5540]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_____________________ test_egat_conv_bi[4-5-1-g0-idtype0] _____________________

g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int32, out_node_feats = 1, out_edge_feats = 5, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv_bi(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=(10,15),
                           in_edge_feats=7,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), 15)))
        efeat = F.randn((g.number_of_edges(), 7))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:589: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01582FC8>
op = 'add'
lhs_data = tensor([[ 0.4097,  1.8984,  0.3551, -0.4480, -0.0687,  0.4001, -0.0705, -1.3342,
          0.1790, -0.2992, -0.6846,  ...1696, -0.0989,  0.0766,  1.0797,  0.3481,  0.4213,
         -0.5149,  0.2680, -1.7514, -0.5989]], grad_fn=<MmBackward>)
rhs_data = tensor([[ 0.4364,  1.4956, -0.8586,  0.2187,  0.8939,  0.8513,  1.2006, -0.6805,
         -0.7382, -0.0431, -0.0510,  ...4594,  2.6749,  0.4502,  0.0860, -1.8500, -1.4749,
         -0.3614,  0.0621, -1.3884,  1.5933]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_____________________ test_egat_conv_bi[4-5-1-g0-idtype1] _____________________

g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int64, out_node_feats = 1, out_edge_feats = 5, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv_bi(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=(10,15),
                           in_edge_feats=7,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), 15)))
        efeat = F.randn((g.number_of_edges(), 7))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:589: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03D172C8>
op = 'add'
lhs_data = tensor([[-0.3011, -0.5376,  0.3123, -0.4423, -0.0936,  0.1586, -0.1759, -0.1315,
         -0.3151,  0.2314,  0.0671,  ...9736,  0.5245,  0.7731, -1.3618, -0.5584,  0.5466,
         -0.9599, -2.0818, -0.2300,  0.5293]], grad_fn=<MmBackward>)
rhs_data = tensor([[ 3.2553,  2.3239, -0.1054,  0.6926, -0.1579, -2.2509, -0.1216,  0.8011,
          0.3188,  1.1837,  0.5312,  ...7275, -0.1071, -1.4916, -0.7246,  0.7349, -0.9231,
          0.1787,  0.0780, -0.7637,  1.1226]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_____________________ test_egat_conv_bi[4-5-1-g1-idtype0] _____________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int32
out_node_feats = 1, out_edge_feats = 5, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv_bi(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=(10,15),
                           in_edge_feats=7,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), 15)))
        efeat = F.randn((g.number_of_edges(), 7))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:589: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF030C1248>
op = 'add'
lhs_data = tensor([[ 2.2073e+00, -2.6027e-04,  3.4743e+00,  2.7274e+00, -8.1739e-01,
          1.3504e+00, -7.8289e-01,  6.1971e-..., -1.0887e+00,
         -9.9952e-01,  1.9203e-01, -2.6045e-01,  4.8717e-01, -6.8592e-01]],
       grad_fn=<MmBackward>)
rhs_data = tensor([[-0.5644,  0.0285,  0.2939, -0.3236, -0.6025, -2.1658,  0.2563, -1.6114,
         -0.0392,  1.7581, -1.0836, -...7246,  0.5448, -0.1695, -3.4277, -1.2917,  1.1475,
         -0.9609,  1.9271, -3.8454,  0.4098]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_____________________ test_egat_conv_bi[4-5-1-g1-idtype1] _____________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int64
out_node_feats = 1, out_edge_feats = 5, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv_bi(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=(10,15),
                           in_edge_feats=7,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), 15)))
        efeat = F.randn((g.number_of_edges(), 7))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:589: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF019B7A08>
op = 'add'
lhs_data = tensor([[-1.6426,  0.8838,  0.3667,  1.3293, -0.6857, -0.1028,  0.7211, -2.6559,
         -0.3012,  1.0810,  0.7554,  ...0170, -1.0157,  0.9363, -0.9162,  1.8780, -0.8444,
          1.3314, -0.1458, -0.4253, -0.5418]], grad_fn=<MmBackward>)
rhs_data = tensor([[-1.5545,  0.8220,  1.4938,  0.5802, -0.4761,  2.0218,  1.0873,  0.4447,
          1.3501,  1.5914, -0.2872, -...5345, -0.3573,  0.3685, -0.4908, -1.6937,  0.3961,
          2.6331,  0.3872,  0.1478,  2.9466]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_____________________ test_egat_conv_bi[4-5-5-g0-idtype0] _____________________

g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int32, out_node_feats = 5, out_edge_feats = 5, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv_bi(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=(10,15),
                           in_edge_feats=7,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), 15)))
        efeat = F.randn((g.number_of_edges(), 7))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:589: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01741EC8>
op = 'add'
lhs_data = tensor([[ 0.0046,  0.6153,  1.2086,  0.1239,  0.4620, -0.1714, -0.5258,  0.5707,
         -0.3771,  1.5341,  0.7290,  ...0234, -0.2427, -0.6409,  0.5663,  0.2867, -0.4876,
          0.4327, -0.9562, -0.8901,  1.4923]], grad_fn=<MmBackward>)
rhs_data = tensor([[-0.1430,  0.2631,  0.5120,  3.6576, -1.1857,  0.4625,  0.3265, -0.4424,
          1.4593, -2.7570,  0.5860,  ...0915,  0.6118,  0.0119, -3.1844,  0.8426, -0.5874,
         -0.2376,  1.3520,  1.4748,  0.6901]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_____________________ test_egat_conv_bi[4-5-5-g0-idtype1] _____________________

g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int64, out_node_feats = 5, out_edge_feats = 5, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv_bi(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=(10,15),
                           in_edge_feats=7,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), 15)))
        efeat = F.randn((g.number_of_edges(), 7))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:589: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0177F348>
op = 'add'
lhs_data = tensor([[-0.9737,  0.6102,  0.0785, -0.9298,  0.5088, -1.0626,  0.9676, -2.0778,
          1.1726,  0.1344, -0.3150, -...8315,  1.5974, -1.6376, -0.7398,  2.4316, -0.0376,
          0.8019, -1.0263,  0.3901,  0.4874]], grad_fn=<MmBackward>)
rhs_data = tensor([[ 0.8386, -0.3807, -1.1705,  2.1955,  0.8331, -1.2222,  0.4905,  2.6220,
         -0.6031, -0.1808,  0.3843,  ...2990,  1.7932,  2.6046,  0.7737,  0.1156, -0.7440,
          0.3782,  1.7154, -0.0047,  2.8158]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_____________________ test_egat_conv_bi[4-5-5-g1-idtype0] _____________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int32
out_node_feats = 5, out_edge_feats = 5, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv_bi(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=(10,15),
                           in_edge_feats=7,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), 15)))
        efeat = F.randn((g.number_of_edges(), 7))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:589: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE15216488>
op = 'add'
lhs_data = tensor([[-0.0281,  0.6139, -0.4069, -0.3891, -0.1169, -0.0984, -0.0637, -0.8928,
         -0.2248, -0.0232, -0.7644,  ...0111, -0.6119,  1.9497, -1.5755, -0.7964, -1.7147,
         -1.3416,  0.9161,  0.2530, -1.7544]], grad_fn=<MmBackward>)
rhs_data = tensor([[ 0.0109,  0.4432,  0.2941, -0.1281,  0.1075, -1.5179, -2.7440,  2.7387,
          0.7373,  0.0477,  2.1056, -...6200,  0.6960, -1.3931,  1.3815,  0.8730, -0.4332,
          0.3742, -0.0612, -2.0617, -0.7192]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_____________________ test_egat_conv_bi[4-5-5-g1-idtype1] _____________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int64
out_node_feats = 5, out_edge_feats = 5, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_node_feats', [1, 5])
    @pytest.mark.parametrize('out_edge_feats', [1, 5])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_egat_conv_bi(g, idtype, out_node_feats, out_edge_feats, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        egat = nn.EGATConv(in_node_feats=(10,15),
                           in_edge_feats=7,
                           out_node_feats=out_node_feats,
                           out_edge_feats=out_edge_feats,
                           num_heads=num_heads)
        nfeat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), 15)))
        efeat = F.randn((g.number_of_edges(), 7))
        egat = egat.to(ctx)
>       h, f = egat(g, nfeat, efeat)

tests\pytorch\test_nn.py:589: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egatconv.py:210: in forward
    graph.apply_edges(fn.u_add_v('f_ni', 'f_nj', 'f_tmp'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0179B3C8>
op = 'add'
lhs_data = tensor([[-1.0389e+00,  7.1546e-02, -3.4941e-01,  7.5242e-02,  3.9990e+00,
          1.3364e+00,  5.6844e-01,  9.8133e-...,  1.2899e+00,
         -2.0342e+00,  4.1273e-01, -1.7717e+00, -1.2472e+00, -1.5473e+00]],
       grad_fn=<MmBackward>)
rhs_data = tensor([[ 0.0752,  1.7975, -1.3407,  1.0803, -0.5635, -0.0950,  0.4054,  0.2015,
         -1.2371, -1.6147, -0.0560, -...0693,  0.2942,  0.6756, -1.7121,  0.9719, -1.2741,
         -0.2872,  0.6854,  1.4201, -2.9483]], grad_fn=<MmBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_______________________ test_sage_conv[mean-g0-idtype0] _______________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
aggre_type = 'mean'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    def test_sage_conv(idtype, g, aggre_type):
        g = g.astype(idtype).to(F.ctx())
        sage = nn.SAGEConv(5, 10, aggre_type)
        feat = F.randn((g.number_of_src_nodes(), 5))
        sage = sage.to(F.ctx())
        # test pickle
        th.save(sage, tmp_buffer)
>       h = sage(g, feat)

tests\pytorch\test_nn.py:608: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:235: in forward
    graph.update_all(msg_fn, fn.mean('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0198E188>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.5167, -0.5155, -0.3048,  0.3555, -0.6499],
        [ 0.2540, -1.6776,  0.3690, -1.3596, -1.4630],
        ...0.1157],
        [-1.2600,  0.6454,  1.8325,  0.6069, -0.0750],
        [ 1.4862, -0.2292, -0.5262,  0.9044,  1.9944]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_sage_conv[mean-g0-idtype1] _______________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
aggre_type = 'mean'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    def test_sage_conv(idtype, g, aggre_type):
        g = g.astype(idtype).to(F.ctx())
        sage = nn.SAGEConv(5, 10, aggre_type)
        feat = F.randn((g.number_of_src_nodes(), 5))
        sage = sage.to(F.ctx())
        # test pickle
        th.save(sage, tmp_buffer)
>       h = sage(g, feat)

tests\pytorch\test_nn.py:608: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:235: in forward
    graph.update_all(msg_fn, fn.mean('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF014B4DC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.8975,  1.1802, -1.5554, -1.4592,  0.2069],
        [-0.0925, -0.4041,  1.8117, -1.1794, -0.4688],
        ...3.2642],
        [-0.5249,  0.0258,  0.3991,  1.9249,  0.0129],
        [-0.5083,  1.5056, -1.9640, -0.4229,  1.1801]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_sage_conv[mean-g1-idtype0] _______________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
aggre_type = 'mean'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    def test_sage_conv(idtype, g, aggre_type):
        g = g.astype(idtype).to(F.ctx())
        sage = nn.SAGEConv(5, 10, aggre_type)
        feat = F.randn((g.number_of_src_nodes(), 5))
        sage = sage.to(F.ctx())
        # test pickle
        th.save(sage, tmp_buffer)
>       h = sage(g, feat)

tests\pytorch\test_nn.py:608: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:235: in forward
    graph.update_all(msg_fn, fn.mean('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14A16908>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.6935,  2.1613, -0.1898,  1.0169,  0.2102],
        [ 0.1352, -0.1715, -0.4494, -0.3250, -0.1710],
        ...0.9318],
        [-0.9497, -0.0270,  2.3339, -0.0300, -0.2625],
        [-0.8269, -1.9289,  0.1916,  0.0808, -1.6697]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_sage_conv[mean-g1-idtype1] _______________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
aggre_type = 'mean'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    def test_sage_conv(idtype, g, aggre_type):
        g = g.astype(idtype).to(F.ctx())
        sage = nn.SAGEConv(5, 10, aggre_type)
        feat = F.randn((g.number_of_src_nodes(), 5))
        sage = sage.to(F.ctx())
        # test pickle
        th.save(sage, tmp_buffer)
>       h = sage(g, feat)

tests\pytorch\test_nn.py:608: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:235: in forward
    graph.update_all(msg_fn, fn.mean('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03D3C908>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.4537, -0.1520,  1.4016,  0.8592,  2.5332],
        [ 0.7623, -0.2182,  0.5084, -0.0966, -0.0754],
        ...0.4105],
        [ 0.0675,  0.1108, -0.5819, -1.5778, -1.9720],
        [ 0.7386, -0.4546, -0.5872, -0.9520,  0.8888]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_sage_conv[mean-g2-idtype0] _______________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
aggre_type = 'mean'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    def test_sage_conv(idtype, g, aggre_type):
        g = g.astype(idtype).to(F.ctx())
        sage = nn.SAGEConv(5, 10, aggre_type)
        feat = F.randn((g.number_of_src_nodes(), 5))
        sage = sage.to(F.ctx())
        # test pickle
        th.save(sage, tmp_buffer)
>       h = sage(g, feat)

tests\pytorch\test_nn.py:608: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:235: in forward
    graph.update_all(msg_fn, fn.mean('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF016A81C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.1992, -1.9114,  1.7328, -0.7454,  0.1148],
        [ 0.8614,  0.4872,  0.6834,  0.1425,  1.2337],
        ...1.1650],
        [ 0.2462, -0.1469, -0.1363,  1.0008, -0.4970],
        [-0.5467, -1.6064, -0.4651,  0.1313, -0.3886]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_sage_conv[mean-g2-idtype1] _______________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
aggre_type = 'mean'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    def test_sage_conv(idtype, g, aggre_type):
        g = g.astype(idtype).to(F.ctx())
        sage = nn.SAGEConv(5, 10, aggre_type)
        feat = F.randn((g.number_of_src_nodes(), 5))
        sage = sage.to(F.ctx())
        # test pickle
        th.save(sage, tmp_buffer)
>       h = sage(g, feat)

tests\pytorch\test_nn.py:608: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:235: in forward
    graph.update_all(msg_fn, fn.mean('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01810F48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.6244, -2.6425, -0.5178,  0.3598,  0.0451],
        [ 3.6627,  0.0434, -0.4057, -0.4627, -0.5298],
        ...1.1128],
        [-0.6951, -0.1278,  0.2457,  0.1586, -0.2663],
        [ 0.4082,  0.1368,  0.5814,  0.9309, -1.6320]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_sage_conv[mean-g3-idtype0] _______________________

idtype = torch.int32
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
aggre_type = 'mean'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    def test_sage_conv(idtype, g, aggre_type):
        g = g.astype(idtype).to(F.ctx())
        sage = nn.SAGEConv(5, 10, aggre_type)
        feat = F.randn((g.number_of_src_nodes(), 5))
        sage = sage.to(F.ctx())
        # test pickle
        th.save(sage, tmp_buffer)
>       h = sage(g, feat)

tests\pytorch\test_nn.py:608: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:235: in forward
    graph.update_all(msg_fn, fn.mean('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03CF8AC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.6399, -0.3381,  0.0730, -0.5250, -0.5501],
        [-0.7266,  0.5355,  0.6045, -0.3308, -1.6098],
        ...1.0480],
        [ 0.7506, -0.2638, -0.7180,  0.4151, -0.5750],
        [-0.3106,  0.5783, -0.5130,  0.8730, -0.0179]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_sage_conv[mean-g3-idtype1] _______________________

idtype = torch.int64
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
aggre_type = 'mean'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    def test_sage_conv(idtype, g, aggre_type):
        g = g.astype(idtype).to(F.ctx())
        sage = nn.SAGEConv(5, 10, aggre_type)
        feat = F.randn((g.number_of_src_nodes(), 5))
        sage = sage.to(F.ctx())
        # test pickle
        th.save(sage, tmp_buffer)
>       h = sage(g, feat)

tests\pytorch\test_nn.py:608: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:235: in forward
    graph.update_all(msg_fn, fn.mean('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF018856C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.5509, -0.2188,  0.2724, -0.0069, -1.8051],
        [ 0.6083,  2.4782, -0.4405,  0.4490, -0.2045],
        ...0.1644],
        [ 0.0149,  0.1271, -0.7126,  0.0644,  0.0457],
        [ 1.4510,  1.8559, -0.4459, -1.2784, -2.8244]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_sage_conv[mean-g4-idtype0] _______________________

idtype = torch.int32
g = Graph(num_nodes=5, num_edges=6,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
aggre_type = 'mean'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    def test_sage_conv(idtype, g, aggre_type):
        g = g.astype(idtype).to(F.ctx())
        sage = nn.SAGEConv(5, 10, aggre_type)
        feat = F.randn((g.number_of_src_nodes(), 5))
        sage = sage.to(F.ctx())
        # test pickle
        th.save(sage, tmp_buffer)
>       h = sage(g, feat)

tests\pytorch\test_nn.py:608: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:235: in forward
    graph.update_all(msg_fn, fn.mean('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF018C05C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.4847, -0.1821, -0.3754, -1.2280, -1.0295],
        [-1.6525, -0.0031, -0.7570,  0.4880,  0.0231],
        ...1.0345],
        [-0.4572,  1.5477, -0.8569,  1.0986, -0.8180],
        [-0.4444, -1.6869,  0.7497,  0.8708, -0.1252]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_sage_conv[mean-g4-idtype1] _______________________

idtype = torch.int64
g = Graph(num_nodes=5, num_edges=6,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
aggre_type = 'mean'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    def test_sage_conv(idtype, g, aggre_type):
        g = g.astype(idtype).to(F.ctx())
        sage = nn.SAGEConv(5, 10, aggre_type)
        feat = F.randn((g.number_of_src_nodes(), 5))
        sage = sage.to(F.ctx())
        # test pickle
        th.save(sage, tmp_buffer)
>       h = sage(g, feat)

tests\pytorch\test_nn.py:608: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:235: in forward
    graph.update_all(msg_fn, fn.mean('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0185EF08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.0518,  0.4190,  0.0381, -0.7346, -0.8814],
        [ 0.2230,  0.3211, -0.1149,  2.4687, -1.0968],
        ...0.8669],
        [ 0.2742,  2.9805,  0.1265, -0.8976,  1.0096],
        [ 0.0063, -0.8539, -0.5710,  0.6938,  0.9824]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_sage_conv[mean-g5-idtype0] _______________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
aggre_type = 'mean'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    def test_sage_conv(idtype, g, aggre_type):
        g = g.astype(idtype).to(F.ctx())
        sage = nn.SAGEConv(5, 10, aggre_type)
        feat = F.randn((g.number_of_src_nodes(), 5))
        sage = sage.to(F.ctx())
        # test pickle
        th.save(sage, tmp_buffer)
>       h = sage(g, feat)

tests\pytorch\test_nn.py:608: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:235: in forward
    graph.update_all(msg_fn, fn.mean('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF016A8748>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.2012,  0.9740, -0.4616, -1.9719, -0.9165],
        [-0.6598,  1.1900, -1.2940, -0.1118, -0.4931],
        ...0.4581],
        [-0.5044, -0.2535,  0.8249,  0.9311, -0.5560],
        [ 0.7562,  1.7565,  0.2017,  0.4127, -0.4059]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_sage_conv[mean-g5-idtype1] _______________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
aggre_type = 'mean'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    def test_sage_conv(idtype, g, aggre_type):
        g = g.astype(idtype).to(F.ctx())
        sage = nn.SAGEConv(5, 10, aggre_type)
        feat = F.randn((g.number_of_src_nodes(), 5))
        sage = sage.to(F.ctx())
        # test pickle
        th.save(sage, tmp_buffer)
>       h = sage(g, feat)

tests\pytorch\test_nn.py:608: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:235: in forward
    graph.update_all(msg_fn, fn.mean('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF030D7648>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.3978,  1.6531, -0.0331,  1.1014,  0.3209],
        [-0.2445, -0.5199,  0.5895,  0.9006, -1.9214],
        ...0.8987],
        [ 0.2628, -1.0204,  1.7846, -0.3674,  0.4017],
        [ 0.2405, -0.4377,  1.4141,  1.7520,  0.2006]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_sage_conv[mean-g6-idtype0] _______________________

idtype = torch.int32, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
aggre_type = 'mean'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    def test_sage_conv(idtype, g, aggre_type):
        g = g.astype(idtype).to(F.ctx())
        sage = nn.SAGEConv(5, 10, aggre_type)
        feat = F.randn((g.number_of_src_nodes(), 5))
        sage = sage.to(F.ctx())
        # test pickle
        th.save(sage, tmp_buffer)
>       h = sage(g, feat)

tests\pytorch\test_nn.py:608: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:235: in forward
    graph.update_all(msg_fn, fn.mean('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE15BA2948>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 2.2969e-01,  5.8390e-01, -6.1020e-02, -1.4316e+00, -6.6701e-01],
        [ 9.9733e-01,  3.7042e-01, -9.6290e...01, -6.6794e-01,  1.6573e+00, -1.1488e+00],
        [-1.2569e+00, -7.2781e-01, -2.8847e-01,  1.5214e+00, -1.9161e+00]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_sage_conv[mean-g6-idtype1] _______________________

idtype = torch.int64, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
aggre_type = 'mean'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    def test_sage_conv(idtype, g, aggre_type):
        g = g.astype(idtype).to(F.ctx())
        sage = nn.SAGEConv(5, 10, aggre_type)
        feat = F.randn((g.number_of_src_nodes(), 5))
        sage = sage.to(F.ctx())
        # test pickle
        th.save(sage, tmp_buffer)
>       h = sage(g, feat)

tests\pytorch\test_nn.py:608: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:235: in forward
    graph.update_all(msg_fn, fn.mean('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14A5DA08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.4544, -0.1587, -1.6297, -3.1991, -1.0804],
        [-0.5030, -2.0614,  0.4025, -0.1783,  0.3222],
        ...3.5838],
        [-0.8967, -1.3463, -0.9606, -0.5578, -0.3875],
        [-2.1614,  0.8547, -0.7490, -0.8992,  0.1229]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_sage_conv[mean-g7-idtype0] _______________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
aggre_type = 'mean'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    def test_sage_conv(idtype, g, aggre_type):
        g = g.astype(idtype).to(F.ctx())
        sage = nn.SAGEConv(5, 10, aggre_type)
        feat = F.randn((g.number_of_src_nodes(), 5))
        sage = sage.to(F.ctx())
        # test pickle
        th.save(sage, tmp_buffer)
>       h = sage(g, feat)

tests\pytorch\test_nn.py:608: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:235: in forward
    graph.update_all(msg_fn, fn.mean('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF017AB408>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.1797, -0.1389,  0.3035,  0.2275,  1.2560],
        [ 0.4777,  1.0115, -0.6438,  0.4192,  0.3112],
        ...0.3222],
        [-0.9916, -0.0668,  0.5594, -0.5939, -0.8603],
        [ 0.3453,  0.4358, -1.0278, -0.4006,  1.7030]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_sage_conv[mean-g7-idtype1] _______________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
aggre_type = 'mean'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    def test_sage_conv(idtype, g, aggre_type):
        g = g.astype(idtype).to(F.ctx())
        sage = nn.SAGEConv(5, 10, aggre_type)
        feat = F.randn((g.number_of_src_nodes(), 5))
        sage = sage.to(F.ctx())
        # test pickle
        th.save(sage, tmp_buffer)
>       h = sage(g, feat)

tests\pytorch\test_nn.py:608: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:235: in forward
    graph.update_all(msg_fn, fn.mean('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14923588>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.8330, -0.1078, -0.2348,  0.3756,  1.1148],
        [-1.2159,  0.2977,  1.3930,  0.0479, -0.1370],
        ...0.8305],
        [-0.0466, -0.4962, -0.2039,  0.5966, -1.9703],
        [-0.5651, -0.1076,  0.4802,  0.7503,  0.2763]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_sage_conv[pool-g0-idtype0] _______________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
aggre_type = 'pool'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    def test_sage_conv(idtype, g, aggre_type):
        g = g.astype(idtype).to(F.ctx())
        sage = nn.SAGEConv(5, 10, aggre_type)
        feat = F.randn((g.number_of_src_nodes(), 5))
        sage = sage.to(F.ctx())
        # test pickle
        th.save(sage, tmp_buffer)
>       h = sage(g, feat)

tests\pytorch\test_nn.py:608: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:257: in forward
    graph.update_all(msg_fn, fn.max('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF010B67C8>
op = 'copy_lhs', reduce_op = 'max'
lhs_data = tensor([[0.1585, 0.0000, 0.0000, 0.0000, 0.5636],
        [0.6246, 0.0000, 0.0000, 0.0000, 1.0704],
        [0.0000, 0...  [0.0000, 0.5154, 0.0000, 0.0000, 0.8562],
        [0.0000, 0.0000, 0.0000, 0.3472, 1.0348]], grad_fn=<ReluBackward0>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_sage_conv[pool-g0-idtype1] _______________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
aggre_type = 'pool'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    def test_sage_conv(idtype, g, aggre_type):
        g = g.astype(idtype).to(F.ctx())
        sage = nn.SAGEConv(5, 10, aggre_type)
        feat = F.randn((g.number_of_src_nodes(), 5))
        sage = sage.to(F.ctx())
        # test pickle
        th.save(sage, tmp_buffer)
>       h = sage(g, feat)

tests\pytorch\test_nn.py:608: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:257: in forward
    graph.update_all(msg_fn, fn.max('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01196D08>
op = 'copy_lhs', reduce_op = 'max'
lhs_data = tensor([[0.0000, 2.7713, 0.0660, 1.2188, 0.4864],
        [0.0000, 0.2083, 0.0000, 1.1882, 0.2260],
        [0.0000, 1...  [0.3085, 2.5072, 0.0000, 0.0000, 0.5998],
        [0.0000, 1.1852, 1.0088, 0.0000, 1.4331]], grad_fn=<ReluBackward0>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_sage_conv[pool-g1-idtype0] _______________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
aggre_type = 'pool'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    def test_sage_conv(idtype, g, aggre_type):
        g = g.astype(idtype).to(F.ctx())
        sage = nn.SAGEConv(5, 10, aggre_type)
        feat = F.randn((g.number_of_src_nodes(), 5))
        sage = sage.to(F.ctx())
        # test pickle
        th.save(sage, tmp_buffer)
>       h = sage(g, feat)

tests\pytorch\test_nn.py:608: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:257: in forward
    graph.update_all(msg_fn, fn.max('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0705BA88>
op = 'copy_lhs', reduce_op = 'max'
lhs_data = tensor([[0.0000, 0.0000, 2.3390, 1.0991, 1.8052],
        [0.0000, 2.0598, 3.0507, 3.1248, 3.3416],
        [0.0000, 0...  [0.0000, 0.7962, 0.0000, 0.0635, 0.0000],
        [0.0000, 0.0000, 0.0000, 0.0000, 0.0000]], grad_fn=<ReluBackward0>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_sage_conv[pool-g1-idtype1] _______________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
aggre_type = 'pool'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    def test_sage_conv(idtype, g, aggre_type):
        g = g.astype(idtype).to(F.ctx())
        sage = nn.SAGEConv(5, 10, aggre_type)
        feat = F.randn((g.number_of_src_nodes(), 5))
        sage = sage.to(F.ctx())
        # test pickle
        th.save(sage, tmp_buffer)
>       h = sage(g, feat)

tests\pytorch\test_nn.py:608: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:257: in forward
    graph.update_all(msg_fn, fn.max('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF014B93C8>
op = 'copy_lhs', reduce_op = 'max'
lhs_data = tensor([[0.0000, 1.4257, 3.4639, 0.0000, 1.7592],
        [0.6720, 2.0973, 0.5037, 0.0000, 0.0127],
        [0.0000, 0...  [0.0000, 0.0000, 3.7724, 0.5622, 1.0016],
        [1.3971, 0.5749, 0.0000, 0.8470, 0.0000]], grad_fn=<ReluBackward0>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_sage_conv[pool-g2-idtype0] _______________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
aggre_type = 'pool'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    def test_sage_conv(idtype, g, aggre_type):
        g = g.astype(idtype).to(F.ctx())
        sage = nn.SAGEConv(5, 10, aggre_type)
        feat = F.randn((g.number_of_src_nodes(), 5))
        sage = sage.to(F.ctx())
        # test pickle
        th.save(sage, tmp_buffer)
>       h = sage(g, feat)

tests\pytorch\test_nn.py:608: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:257: in forward
    graph.update_all(msg_fn, fn.max('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03D2FFC8>
op = 'copy_lhs', reduce_op = 'max'
lhs_data = tensor([[0.4976, 1.0941, 1.6991, 0.8592, 0.1752],
        [1.0626, 0.0679, 1.2594, 0.5604, 0.0066],
        [0.1063, 0...  [0.0000, 0.5206, 1.2513, 1.6043, 0.3325],
        [0.0000, 0.4688, 0.1870, 1.2528, 0.1846]], grad_fn=<ReluBackward0>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_sage_conv[pool-g2-idtype1] _______________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
aggre_type = 'pool'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    def test_sage_conv(idtype, g, aggre_type):
        g = g.astype(idtype).to(F.ctx())
        sage = nn.SAGEConv(5, 10, aggre_type)
        feat = F.randn((g.number_of_src_nodes(), 5))
        sage = sage.to(F.ctx())
        # test pickle
        th.save(sage, tmp_buffer)
>       h = sage(g, feat)

tests\pytorch\test_nn.py:608: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:257: in forward
    graph.update_all(msg_fn, fn.max('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0198FAC8>
op = 'copy_lhs', reduce_op = 'max'
lhs_data = tensor([[1.4845, 0.0000, 0.0000, 0.0000, 2.8829],
        [0.0000, 0.0000, 0.0000, 0.0000, 0.0000],
        [0.0000, 2...  [2.5745, 0.0000, 0.0000, 0.0000, 1.1305],
        [0.2341, 0.0000, 0.0000, 0.0000, 0.0000]], grad_fn=<ReluBackward0>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_sage_conv[pool-g3-idtype0] _______________________

idtype = torch.int32
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
aggre_type = 'pool'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    def test_sage_conv(idtype, g, aggre_type):
        g = g.astype(idtype).to(F.ctx())
        sage = nn.SAGEConv(5, 10, aggre_type)
        feat = F.randn((g.number_of_src_nodes(), 5))
        sage = sage.to(F.ctx())
        # test pickle
        th.save(sage, tmp_buffer)
>       h = sage(g, feat)

tests\pytorch\test_nn.py:608: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:257: in forward
    graph.update_all(msg_fn, fn.max('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0151CE08>
op = 'copy_lhs', reduce_op = 'max'
lhs_data = tensor([[3.9357, 1.2618, 0.0000, 0.6383, 0.0000],
        [0.0000, 0.0000, 0.8488, 0.6486, 0.1943],
        [0.0000, 0...  [0.0000, 1.0634, 0.9607, 1.3736, 0.0000],
        [0.2360, 0.0000, 1.1149, 0.0000, 2.3872]], grad_fn=<ReluBackward0>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_sage_conv[pool-g3-idtype1] _______________________

idtype = torch.int64
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
aggre_type = 'pool'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    def test_sage_conv(idtype, g, aggre_type):
        g = g.astype(idtype).to(F.ctx())
        sage = nn.SAGEConv(5, 10, aggre_type)
        feat = F.randn((g.number_of_src_nodes(), 5))
        sage = sage.to(F.ctx())
        # test pickle
        th.save(sage, tmp_buffer)
>       h = sage(g, feat)

tests\pytorch\test_nn.py:608: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:257: in forward
    graph.update_all(msg_fn, fn.max('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0188B748>
op = 'copy_lhs', reduce_op = 'max'
lhs_data = tensor([[0.0000, 0.2410, 0.0415, 0.0000, 0.7263],
        [0.5832, 0.0000, 1.3551, 0.0000, 0.0000],
        [0.2466, 0...  [0.0000, 0.0000, 1.7840, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.4463, 0.0000, 1.5614]], grad_fn=<ReluBackward0>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_sage_conv[pool-g4-idtype0] _______________________

idtype = torch.int32
g = Graph(num_nodes=5, num_edges=6,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
aggre_type = 'pool'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    def test_sage_conv(idtype, g, aggre_type):
        g = g.astype(idtype).to(F.ctx())
        sage = nn.SAGEConv(5, 10, aggre_type)
        feat = F.randn((g.number_of_src_nodes(), 5))
        sage = sage.to(F.ctx())
        # test pickle
        th.save(sage, tmp_buffer)
>       h = sage(g, feat)

tests\pytorch\test_nn.py:608: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:257: in forward
    graph.update_all(msg_fn, fn.max('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF014A7348>
op = 'copy_lhs', reduce_op = 'max'
lhs_data = tensor([[0.0869, 0.9946, 0.8170, 0.3927, 0.8594],
        [3.0572, 0.0000, 1.1308, 0.0000, 1.7594],
        [0.0000, 0...  [1.1543, 0.0000, 0.5000, 0.0000, 0.0183],
        [0.0000, 1.0264, 0.0000, 0.0783, 0.0000]], grad_fn=<ReluBackward0>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_sage_conv[pool-g4-idtype1] _______________________

idtype = torch.int64
g = Graph(num_nodes=5, num_edges=6,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
aggre_type = 'pool'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    def test_sage_conv(idtype, g, aggre_type):
        g = g.astype(idtype).to(F.ctx())
        sage = nn.SAGEConv(5, 10, aggre_type)
        feat = F.randn((g.number_of_src_nodes(), 5))
        sage = sage.to(F.ctx())
        # test pickle
        th.save(sage, tmp_buffer)
>       h = sage(g, feat)

tests\pytorch\test_nn.py:608: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:257: in forward
    graph.update_all(msg_fn, fn.max('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01585F08>
op = 'copy_lhs', reduce_op = 'max'
lhs_data = tensor([[0.0000, 0.6308, 3.6605, 1.7524, 0.0000],
        [0.0000, 0.0000, 0.0000, 0.0000, 1.0386],
        [0.2584, 0...  [0.5457, 0.0000, 0.0000, 0.0000, 0.8775],
        [1.4006, 0.0000, 0.0000, 0.0000, 2.1852]], grad_fn=<ReluBackward0>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_sage_conv[pool-g5-idtype0] _______________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
aggre_type = 'pool'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    def test_sage_conv(idtype, g, aggre_type):
        g = g.astype(idtype).to(F.ctx())
        sage = nn.SAGEConv(5, 10, aggre_type)
        feat = F.randn((g.number_of_src_nodes(), 5))
        sage = sage.to(F.ctx())
        # test pickle
        th.save(sage, tmp_buffer)
>       h = sage(g, feat)

tests\pytorch\test_nn.py:608: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:257: in forward
    graph.update_all(msg_fn, fn.max('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0188BE08>
op = 'copy_lhs', reduce_op = 'max'
lhs_data = tensor([[0.0000e+00, 0.0000e+00, 0.0000e+00, 4.5259e-02, 0.0000e+00],
        [0.0000e+00, 6.2370e-04, 0.0000e+00, 4.6...00, 5.1441e-01],
        [9.9388e-01, 0.0000e+00, 8.2020e-01, 6.8841e-01, 1.5739e+00]],
       grad_fn=<ReluBackward0>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_sage_conv[pool-g5-idtype1] _______________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
aggre_type = 'pool'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    def test_sage_conv(idtype, g, aggre_type):
        g = g.astype(idtype).to(F.ctx())
        sage = nn.SAGEConv(5, 10, aggre_type)
        feat = F.randn((g.number_of_src_nodes(), 5))
        sage = sage.to(F.ctx())
        # test pickle
        th.save(sage, tmp_buffer)
>       h = sage(g, feat)

tests\pytorch\test_nn.py:608: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:257: in forward
    graph.update_all(msg_fn, fn.max('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03D11148>
op = 'copy_lhs', reduce_op = 'max'
lhs_data = tensor([[0.0000, 0.0000, 1.4082, 2.8034, 0.8778],
        [1.9190, 0.1259, 0.0248, 3.3276, 0.0000],
        [0.0000, 0...  [0.0000, 0.0000, 0.0000, 0.0000, 0.1052],
        [0.0000, 0.0000, 0.0000, 0.0000, 0.9237]], grad_fn=<ReluBackward0>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_sage_conv[pool-g6-idtype0] _______________________

idtype = torch.int32, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
aggre_type = 'pool'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    def test_sage_conv(idtype, g, aggre_type):
        g = g.astype(idtype).to(F.ctx())
        sage = nn.SAGEConv(5, 10, aggre_type)
        feat = F.randn((g.number_of_src_nodes(), 5))
        sage = sage.to(F.ctx())
        # test pickle
        th.save(sage, tmp_buffer)
>       h = sage(g, feat)

tests\pytorch\test_nn.py:608: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:257: in forward
    graph.update_all(msg_fn, fn.max('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03EC7648>
op = 'copy_lhs', reduce_op = 'max'
lhs_data = tensor([[0.0000, 1.8150, 0.3608, 0.0000, 1.3258],
        [0.0000, 1.0446, 0.4713, 0.4176, 2.7303],
        [0.0000, 1...  [0.0000, 1.1925, 1.1038, 0.0000, 0.0000],
        [0.4410, 2.2797, 1.2218, 0.0000, 0.0000]], grad_fn=<ReluBackward0>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_sage_conv[pool-g6-idtype1] _______________________

idtype = torch.int64, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
aggre_type = 'pool'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    def test_sage_conv(idtype, g, aggre_type):
        g = g.astype(idtype).to(F.ctx())
        sage = nn.SAGEConv(5, 10, aggre_type)
        feat = F.randn((g.number_of_src_nodes(), 5))
        sage = sage.to(F.ctx())
        # test pickle
        th.save(sage, tmp_buffer)
>       h = sage(g, feat)

tests\pytorch\test_nn.py:608: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:257: in forward
    graph.update_all(msg_fn, fn.max('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF014C5988>
op = 'copy_lhs', reduce_op = 'max'
lhs_data = tensor([[0.0000, 0.0000, 1.4520, 0.0000, 0.0000],
        [0.0858, 0.4558, 0.9923, 0.0000, 0.0000],
        [0.0000, 0...  [0.0000, 0.0000, 2.3225, 0.0000, 0.0000],
        [0.1389, 0.0000, 2.3790, 0.0000, 0.0000]], grad_fn=<ReluBackward0>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_sage_conv[pool-g7-idtype0] _______________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
aggre_type = 'pool'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    def test_sage_conv(idtype, g, aggre_type):
        g = g.astype(idtype).to(F.ctx())
        sage = nn.SAGEConv(5, 10, aggre_type)
        feat = F.randn((g.number_of_src_nodes(), 5))
        sage = sage.to(F.ctx())
        # test pickle
        th.save(sage, tmp_buffer)
>       h = sage(g, feat)

tests\pytorch\test_nn.py:608: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:257: in forward
    graph.update_all(msg_fn, fn.max('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE149F4708>
op = 'copy_lhs', reduce_op = 'max'
lhs_data = tensor([[0.0000, 0.9312, 0.0000, 0.0000, 0.0000],
        [1.4138, 0.0000, 0.0000, 1.1098, 0.0000],
        [0.6215, 0...  [0.3499, 0.0000, 0.0000, 0.4962, 0.0000],
        [1.6557, 0.0000, 3.7999, 0.0000, 1.4334]], grad_fn=<ReluBackward0>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_sage_conv[pool-g7-idtype1] _______________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
aggre_type = 'pool'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    def test_sage_conv(idtype, g, aggre_type):
        g = g.astype(idtype).to(F.ctx())
        sage = nn.SAGEConv(5, 10, aggre_type)
        feat = F.randn((g.number_of_src_nodes(), 5))
        sage = sage.to(F.ctx())
        # test pickle
        th.save(sage, tmp_buffer)
>       h = sage(g, feat)

tests\pytorch\test_nn.py:608: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:257: in forward
    graph.update_all(msg_fn, fn.max('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01762B08>
op = 'copy_lhs', reduce_op = 'max'
lhs_data = tensor([[0.0966, 0.0000, 0.0000, 1.7586, 4.4027],
        [0.0000, 0.0143, 0.7113, 1.0392, 0.0000],
        [0.0000, 0...  [0.0000, 1.8970, 1.5842, 0.0000, 0.0000],
        [0.0000, 2.1559, 0.0000, 0.5469, 0.0000]], grad_fn=<ReluBackward0>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_sage_conv[gcn-g0-idtype0] ________________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
aggre_type = 'gcn'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    def test_sage_conv(idtype, g, aggre_type):
        g = g.astype(idtype).to(F.ctx())
        sage = nn.SAGEConv(5, 10, aggre_type)
        feat = F.randn((g.number_of_src_nodes(), 5))
        sage = sage.to(F.ctx())
        # test pickle
        th.save(sage, tmp_buffer)
>       h = sage(g, feat)

tests\pytorch\test_nn.py:608: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:249: in forward
    graph.update_all(msg_fn, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF018700C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.1317,  0.8249, -1.9308, -0.4633,  1.3161],
        [-0.5204, -1.4748, -0.0529,  0.3877, -0.3051],
        ...0.8400],
        [-1.6038,  1.8152, -0.0849,  1.4837, -1.1731],
        [-1.0338,  0.2002,  0.6466, -0.7875, -3.2355]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_sage_conv[gcn-g0-idtype1] ________________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
aggre_type = 'gcn'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    def test_sage_conv(idtype, g, aggre_type):
        g = g.astype(idtype).to(F.ctx())
        sage = nn.SAGEConv(5, 10, aggre_type)
        feat = F.randn((g.number_of_src_nodes(), 5))
        sage = sage.to(F.ctx())
        # test pickle
        th.save(sage, tmp_buffer)
>       h = sage(g, feat)

tests\pytorch\test_nn.py:608: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:249: in forward
    graph.update_all(msg_fn, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF011A8E08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.2721, -0.6001, -1.2396, -0.8582, -1.6975],
        [-0.7693, -0.4166, -0.2530, -0.3657, -0.8675],
        ...0.4995],
        [ 1.9809,  1.0196, -0.2863,  2.1411, -1.2307],
        [ 0.0409, -0.2123,  0.0104, -0.0176, -1.3512]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_sage_conv[gcn-g1-idtype0] ________________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
aggre_type = 'gcn'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    def test_sage_conv(idtype, g, aggre_type):
        g = g.astype(idtype).to(F.ctx())
        sage = nn.SAGEConv(5, 10, aggre_type)
        feat = F.randn((g.number_of_src_nodes(), 5))
        sage = sage.to(F.ctx())
        # test pickle
        th.save(sage, tmp_buffer)
>       h = sage(g, feat)

tests\pytorch\test_nn.py:608: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:249: in forward
    graph.update_all(msg_fn, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE15BBA0C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.4227,  0.4643,  1.6042,  0.7281,  1.5001],
        [ 0.5219, -0.9932,  1.2095, -0.0889, -0.7277],
        ...0.1642],
        [-0.1649,  1.0272, -1.8370, -0.6352, -1.0884],
        [-0.8506,  1.8167, -0.1743, -0.4326, -1.5211]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_sage_conv[gcn-g1-idtype1] ________________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
aggre_type = 'gcn'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    def test_sage_conv(idtype, g, aggre_type):
        g = g.astype(idtype).to(F.ctx())
        sage = nn.SAGEConv(5, 10, aggre_type)
        feat = F.randn((g.number_of_src_nodes(), 5))
        sage = sage.to(F.ctx())
        # test pickle
        th.save(sage, tmp_buffer)
>       h = sage(g, feat)

tests\pytorch\test_nn.py:608: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:249: in forward
    graph.update_all(msg_fn, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01865AC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.4624,  0.4306, -0.7809,  1.3977,  1.5689],
        [ 0.0907,  1.0231, -0.5921, -0.6391,  1.6878],
        ...0.4652],
        [ 0.1927, -1.8232, -0.9590,  0.0984, -0.9140],
        [ 1.6839, -0.8043,  1.3981, -0.5844,  2.1040]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_sage_conv[gcn-g2-idtype0] ________________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
aggre_type = 'gcn'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    def test_sage_conv(idtype, g, aggre_type):
        g = g.astype(idtype).to(F.ctx())
        sage = nn.SAGEConv(5, 10, aggre_type)
        feat = F.randn((g.number_of_src_nodes(), 5))
        sage = sage.to(F.ctx())
        # test pickle
        th.save(sage, tmp_buffer)
>       h = sage(g, feat)

tests\pytorch\test_nn.py:608: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:249: in forward
    graph.update_all(msg_fn, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14572AC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.5539, -0.7815, -0.0902, -2.3258, -0.9464],
        [-0.5937, -0.4888, -0.2014,  0.6368,  0.1389],
        ...0.7342],
        [ 1.6375, -1.0547,  0.4803, -0.2956, -0.4565],
        [-0.3084, -0.5308,  0.0575,  0.2688, -0.5630]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_sage_conv[gcn-g2-idtype1] ________________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
aggre_type = 'gcn'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    def test_sage_conv(idtype, g, aggre_type):
        g = g.astype(idtype).to(F.ctx())
        sage = nn.SAGEConv(5, 10, aggre_type)
        feat = F.randn((g.number_of_src_nodes(), 5))
        sage = sage.to(F.ctx())
        # test pickle
        th.save(sage, tmp_buffer)
>       h = sage(g, feat)

tests\pytorch\test_nn.py:608: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:249: in forward
    graph.update_all(msg_fn, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14AC8F08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.8919, -0.9838, -0.0758,  0.1411, -0.2249],
        [ 1.2644, -1.5220, -0.4056,  0.4851, -0.5592],
        ...0.7041],
        [ 1.4738,  1.2664,  0.8684,  0.3717, -0.7432],
        [-0.6703,  0.9671, -1.5924,  0.3516, -0.0820]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_sage_conv[gcn-g3-idtype0] ________________________

idtype = torch.int32
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
aggre_type = 'gcn'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    def test_sage_conv(idtype, g, aggre_type):
        g = g.astype(idtype).to(F.ctx())
        sage = nn.SAGEConv(5, 10, aggre_type)
        feat = F.randn((g.number_of_src_nodes(), 5))
        sage = sage.to(F.ctx())
        # test pickle
        th.save(sage, tmp_buffer)
>       h = sage(g, feat)

tests\pytorch\test_nn.py:608: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:249: in forward
    graph.update_all(msg_fn, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0150E888>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-4.9485e-01,  6.3729e-01, -6.7693e-01,  2.7704e+00, -8.7528e-01],
        [-4.8373e-01,  7.8652e-01, -1.4407e...02, -1.2880e+00, -8.3344e-01,  6.3666e-01],
        [ 1.4071e+00, -1.9565e+00,  2.0912e+00, -7.8690e-01, -1.2858e+00]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_sage_conv[gcn-g3-idtype1] ________________________

idtype = torch.int64
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
aggre_type = 'gcn'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    def test_sage_conv(idtype, g, aggre_type):
        g = g.astype(idtype).to(F.ctx())
        sage = nn.SAGEConv(5, 10, aggre_type)
        feat = F.randn((g.number_of_src_nodes(), 5))
        sage = sage.to(F.ctx())
        # test pickle
        th.save(sage, tmp_buffer)
>       h = sage(g, feat)

tests\pytorch\test_nn.py:608: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:249: in forward
    graph.update_all(msg_fn, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14AA2588>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.7310,  0.9221,  0.3632,  1.9280,  0.4146],
        [-0.6335,  0.1429, -0.5864, -0.3491,  0.5359],
        ...0.4904],
        [ 0.4814,  0.9299, -1.0617, -0.0411,  0.7684],
        [-0.1227, -0.2402,  0.8567,  0.0870, -0.0233]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_sage_conv[gcn-g4-idtype0] ________________________

idtype = torch.int32
g = Graph(num_nodes=5, num_edges=6,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
aggre_type = 'gcn'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    def test_sage_conv(idtype, g, aggre_type):
        g = g.astype(idtype).to(F.ctx())
        sage = nn.SAGEConv(5, 10, aggre_type)
        feat = F.randn((g.number_of_src_nodes(), 5))
        sage = sage.to(F.ctx())
        # test pickle
        th.save(sage, tmp_buffer)
>       h = sage(g, feat)

tests\pytorch\test_nn.py:608: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:249: in forward
    graph.update_all(msg_fn, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF014EE2C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.3895,  1.0153, -1.8424,  1.2921,  0.3789],
        [-0.5200, -0.4616, -0.1287,  0.6863,  1.6166],
        ...1.6807],
        [ 0.1907,  0.2240, -1.9563,  1.3175, -0.0841],
        [-0.8794,  0.1521, -0.4919, -1.2872,  0.7056]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_sage_conv[gcn-g4-idtype1] ________________________

idtype = torch.int64
g = Graph(num_nodes=5, num_edges=6,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
aggre_type = 'gcn'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    def test_sage_conv(idtype, g, aggre_type):
        g = g.astype(idtype).to(F.ctx())
        sage = nn.SAGEConv(5, 10, aggre_type)
        feat = F.randn((g.number_of_src_nodes(), 5))
        sage = sage.to(F.ctx())
        # test pickle
        th.save(sage, tmp_buffer)
>       h = sage(g, feat)

tests\pytorch\test_nn.py:608: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:249: in forward
    graph.update_all(msg_fn, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF016FA2C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.6726, -0.4565, -0.1422, -0.1390,  0.2742],
        [ 1.3324, -1.2843, -0.5101,  0.5544, -0.4567],
        ...0.3577],
        [ 0.0221,  0.5931, -0.7505,  0.3465, -1.2055],
        [ 0.1464,  0.0984,  1.3545, -0.2890,  2.3422]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_sage_conv[gcn-g5-idtype0] ________________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
aggre_type = 'gcn'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    def test_sage_conv(idtype, g, aggre_type):
        g = g.astype(idtype).to(F.ctx())
        sage = nn.SAGEConv(5, 10, aggre_type)
        feat = F.randn((g.number_of_src_nodes(), 5))
        sage = sage.to(F.ctx())
        # test pickle
        th.save(sage, tmp_buffer)
>       h = sage(g, feat)

tests\pytorch\test_nn.py:608: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:249: in forward
    graph.update_all(msg_fn, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01513888>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.0646,  0.3022,  0.5060,  0.0165, -0.1680],
        [-0.1863, -0.5406,  0.3033, -0.2630, -1.2373],
        ...2.0721],
        [ 0.8471, -0.2928,  0.9058, -0.9212, -1.9147],
        [-0.0233, -1.2787, -0.6364, -1.3143, -0.5757]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_sage_conv[gcn-g5-idtype1] ________________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
aggre_type = 'gcn'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    def test_sage_conv(idtype, g, aggre_type):
        g = g.astype(idtype).to(F.ctx())
        sage = nn.SAGEConv(5, 10, aggre_type)
        feat = F.randn((g.number_of_src_nodes(), 5))
        sage = sage.to(F.ctx())
        # test pickle
        th.save(sage, tmp_buffer)
>       h = sage(g, feat)

tests\pytorch\test_nn.py:608: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:249: in forward
    graph.update_all(msg_fn, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0184F088>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.2002, -1.6793,  0.5727, -1.9986, -0.4839],
        [-1.4539,  1.2085, -0.2599, -2.9339, -0.1177],
        ...0.9278],
        [-0.9053,  0.0461,  0.2047, -0.1742,  0.8499],
        [-0.4104, -1.1370,  0.4894, -0.5011, -0.9028]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_sage_conv[gcn-g6-idtype0] ________________________

idtype = torch.int32, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
aggre_type = 'gcn'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    def test_sage_conv(idtype, g, aggre_type):
        g = g.astype(idtype).to(F.ctx())
        sage = nn.SAGEConv(5, 10, aggre_type)
        feat = F.randn((g.number_of_src_nodes(), 5))
        sage = sage.to(F.ctx())
        # test pickle
        th.save(sage, tmp_buffer)
>       h = sage(g, feat)

tests\pytorch\test_nn.py:608: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:249: in forward
    graph.update_all(msg_fn, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01196988>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-2.2947,  0.8801,  0.1600,  0.0176, -0.3397],
        [ 0.6011, -0.5153, -0.0909,  0.7892, -1.6008],
        ...0.3166],
        [ 0.0714, -0.8795, -1.0113, -0.9840, -2.0283],
        [-0.3023,  1.0605, -2.1490,  0.2935, -0.5128]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_sage_conv[gcn-g6-idtype1] ________________________

idtype = torch.int64, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
aggre_type = 'gcn'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    def test_sage_conv(idtype, g, aggre_type):
        g = g.astype(idtype).to(F.ctx())
        sage = nn.SAGEConv(5, 10, aggre_type)
        feat = F.randn((g.number_of_src_nodes(), 5))
        sage = sage.to(F.ctx())
        # test pickle
        th.save(sage, tmp_buffer)
>       h = sage(g, feat)

tests\pytorch\test_nn.py:608: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:249: in forward
    graph.update_all(msg_fn, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03D1B748>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.6671,  0.4295, -0.1758, -1.3110, -0.3408],
        [ 0.2610, -0.7964,  0.3317, -1.1380,  0.6887],
        ...0.4561],
        [ 1.3425,  1.2358, -0.8415, -0.9569,  0.5872],
        [ 0.3366, -1.4637,  0.8637, -0.9939, -1.0113]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_sage_conv[gcn-g7-idtype0] ________________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
aggre_type = 'gcn'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    def test_sage_conv(idtype, g, aggre_type):
        g = g.astype(idtype).to(F.ctx())
        sage = nn.SAGEConv(5, 10, aggre_type)
        feat = F.randn((g.number_of_src_nodes(), 5))
        sage = sage.to(F.ctx())
        # test pickle
        th.save(sage, tmp_buffer)
>       h = sage(g, feat)

tests\pytorch\test_nn.py:608: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:249: in forward
    graph.update_all(msg_fn, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01999788>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.2117, -0.6470,  0.7409, -0.2614,  1.5635],
        [ 0.6877, -1.8111,  1.5840,  1.4461, -0.8770],
        ...1.5089],
        [ 0.5453, -0.8471, -0.6357,  0.1641, -0.6634],
        [ 0.7166,  1.5923,  1.5609,  0.3108,  1.6669]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_sage_conv[gcn-g7-idtype1] ________________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
aggre_type = 'gcn'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    def test_sage_conv(idtype, g, aggre_type):
        g = g.astype(idtype).to(F.ctx())
        sage = nn.SAGEConv(5, 10, aggre_type)
        feat = F.randn((g.number_of_src_nodes(), 5))
        sage = sage.to(F.ctx())
        # test pickle
        th.save(sage, tmp_buffer)
>       h = sage(g, feat)

tests\pytorch\test_nn.py:608: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:249: in forward
    graph.update_all(msg_fn, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0171CB88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.3355,  0.2068, -0.1222,  0.7349, -0.9088],
        [-0.3700, -0.7060, -0.9482,  0.1384, -0.9776],
        ...0.3440],
        [ 0.7908, -0.4497, -1.0618,  1.0447, -0.3508],
        [ 0.3749, -0.1648,  0.2236, -0.8188,  0.3056]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_sage_conv[lstm-g0-idtype0] _______________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
aggre_type = 'lstm'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    def test_sage_conv(idtype, g, aggre_type):
        g = g.astype(idtype).to(F.ctx())
        sage = nn.SAGEConv(5, 10, aggre_type)
        feat = F.randn((g.number_of_src_nodes(), 5))
        sage = sage.to(F.ctx())
        # test pickle
        th.save(sage, tmp_buffer)
>       h = sage(g, feat)

tests\pytorch\test_nn.py:608: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:261: in forward
    graph.update_all(msg_fn, self._lstm_reducer)
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:362: in message_passing
    msgdata = invoke_gsddmm(g, mfunc)
python\dgl\core.py:276: in invoke_gsddmm
    z = op(graph, x)
python\dgl\ops\sddmm.py:164: in copy_u
    return gsddmm(g, 'copy_lhs', x, None)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0148D488>
op = 'copy_lhs'
lhs_data = tensor([[ 0.0544,  0.1762, -0.0123, -0.7046,  2.0941],
        [-0.8694,  0.4527,  0.5545, -0.6142, -0.4968],
        ...1.9667],
        [-0.0466,  0.7508,  0.6995, -0.3397, -0.5727],
        [ 1.8063,  0.8264,  0.2978,  0.0687, -0.2268]])
rhs_data = None, lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_______________________ test_sage_conv[lstm-g0-idtype1] _______________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
aggre_type = 'lstm'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    def test_sage_conv(idtype, g, aggre_type):
        g = g.astype(idtype).to(F.ctx())
        sage = nn.SAGEConv(5, 10, aggre_type)
        feat = F.randn((g.number_of_src_nodes(), 5))
        sage = sage.to(F.ctx())
        # test pickle
        th.save(sage, tmp_buffer)
>       h = sage(g, feat)

tests\pytorch\test_nn.py:608: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:261: in forward
    graph.update_all(msg_fn, self._lstm_reducer)
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:362: in message_passing
    msgdata = invoke_gsddmm(g, mfunc)
python\dgl\core.py:276: in invoke_gsddmm
    z = op(graph, x)
python\dgl\ops\sddmm.py:164: in copy_u
    return gsddmm(g, 'copy_lhs', x, None)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0146F108>
op = 'copy_lhs'
lhs_data = tensor([[-0.8851,  1.1868, -0.9101, -1.4355,  0.2530],
        [-1.2803, -0.0293,  0.4718,  0.5390, -0.3281],
        ...0.5542],
        [ 0.1706,  0.8044, -1.0459,  0.1604, -2.0331],
        [-0.8538,  1.6266, -1.3029,  1.0085,  0.9217]])
rhs_data = None, lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_______________________ test_sage_conv[lstm-g1-idtype0] _______________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
aggre_type = 'lstm'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    def test_sage_conv(idtype, g, aggre_type):
        g = g.astype(idtype).to(F.ctx())
        sage = nn.SAGEConv(5, 10, aggre_type)
        feat = F.randn((g.number_of_src_nodes(), 5))
        sage = sage.to(F.ctx())
        # test pickle
        th.save(sage, tmp_buffer)
>       h = sage(g, feat)

tests\pytorch\test_nn.py:608: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:261: in forward
    graph.update_all(msg_fn, self._lstm_reducer)
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:362: in message_passing
    msgdata = invoke_gsddmm(g, mfunc)
python\dgl\core.py:276: in invoke_gsddmm
    z = op(graph, x)
python\dgl\ops\sddmm.py:164: in copy_u
    return gsddmm(g, 'copy_lhs', x, None)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07249B88>
op = 'copy_lhs'
lhs_data = tensor([[-0.6543, -0.6353, -0.6893,  1.6505,  0.3748],
        [-0.0247,  0.0374,  1.8475, -0.5484, -1.5510],
        ...0.9698],
        [ 0.7876, -0.7301, -0.6944, -0.6687,  1.4542],
        [ 0.8526, -0.1140, -0.3254, -0.8657, -0.9688]])
rhs_data = None, lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_______________________ test_sage_conv[lstm-g1-idtype1] _______________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
aggre_type = 'lstm'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    def test_sage_conv(idtype, g, aggre_type):
        g = g.astype(idtype).to(F.ctx())
        sage = nn.SAGEConv(5, 10, aggre_type)
        feat = F.randn((g.number_of_src_nodes(), 5))
        sage = sage.to(F.ctx())
        # test pickle
        th.save(sage, tmp_buffer)
>       h = sage(g, feat)

tests\pytorch\test_nn.py:608: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:261: in forward
    graph.update_all(msg_fn, self._lstm_reducer)
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:362: in message_passing
    msgdata = invoke_gsddmm(g, mfunc)
python\dgl\core.py:276: in invoke_gsddmm
    z = op(graph, x)
python\dgl\ops\sddmm.py:164: in copy_u
    return gsddmm(g, 'copy_lhs', x, None)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0195D448>
op = 'copy_lhs'
lhs_data = tensor([[ 1.7050, -1.2768, -2.0223,  1.8231,  1.0498],
        [-0.4519, -0.8706,  0.7210, -0.3761, -1.0483],
        ...0.9027],
        [-0.9603,  0.5364,  0.7130, -0.0129,  1.1508],
        [-0.0767,  0.8617, -1.6927,  1.5943, -0.3463]])
rhs_data = None, lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_______________________ test_sage_conv[lstm-g2-idtype0] _______________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
aggre_type = 'lstm'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    def test_sage_conv(idtype, g, aggre_type):
        g = g.astype(idtype).to(F.ctx())
        sage = nn.SAGEConv(5, 10, aggre_type)
        feat = F.randn((g.number_of_src_nodes(), 5))
        sage = sage.to(F.ctx())
        # test pickle
        th.save(sage, tmp_buffer)
>       h = sage(g, feat)

tests\pytorch\test_nn.py:608: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:261: in forward
    graph.update_all(msg_fn, self._lstm_reducer)
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:362: in message_passing
    msgdata = invoke_gsddmm(g, mfunc)
python\dgl\core.py:276: in invoke_gsddmm
    z = op(graph, x)
python\dgl\ops\sddmm.py:164: in copy_u
    return gsddmm(g, 'copy_lhs', x, None)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14651A08>
op = 'copy_lhs'
lhs_data = tensor([[ 0.7758,  1.6994,  1.3097, -0.4352,  1.4428],
        [-1.8886,  1.5648, -0.8342, -0.1991, -0.3190],
        ...0.7745],
        [ 1.1931,  0.8746,  0.4170, -0.3326,  0.2879],
        [-1.6256, -0.1270,  0.3621, -1.2274,  0.5574]])
rhs_data = None, lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_______________________ test_sage_conv[lstm-g2-idtype1] _______________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
aggre_type = 'lstm'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    def test_sage_conv(idtype, g, aggre_type):
        g = g.astype(idtype).to(F.ctx())
        sage = nn.SAGEConv(5, 10, aggre_type)
        feat = F.randn((g.number_of_src_nodes(), 5))
        sage = sage.to(F.ctx())
        # test pickle
        th.save(sage, tmp_buffer)
>       h = sage(g, feat)

tests\pytorch\test_nn.py:608: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:261: in forward
    graph.update_all(msg_fn, self._lstm_reducer)
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:362: in message_passing
    msgdata = invoke_gsddmm(g, mfunc)
python\dgl\core.py:276: in invoke_gsddmm
    z = op(graph, x)
python\dgl\ops\sddmm.py:164: in copy_u
    return gsddmm(g, 'copy_lhs', x, None)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF014B9408>
op = 'copy_lhs'
lhs_data = tensor([[ 2.4081, -0.7503,  1.1796, -0.7593, -0.1321],
        [ 1.1219, -0.2574, -0.4210, -0.0353, -0.0903],
        ...0.7208],
        [-0.4674, -0.6958,  0.3037,  1.0330,  0.2633],
        [ 0.1138, -0.1878,  1.1796,  1.6160,  0.3083]])
rhs_data = None, lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_______________________ test_sage_conv[lstm-g3-idtype0] _______________________

idtype = torch.int32
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
aggre_type = 'lstm'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    def test_sage_conv(idtype, g, aggre_type):
        g = g.astype(idtype).to(F.ctx())
        sage = nn.SAGEConv(5, 10, aggre_type)
        feat = F.randn((g.number_of_src_nodes(), 5))
        sage = sage.to(F.ctx())
        # test pickle
        th.save(sage, tmp_buffer)
>       h = sage(g, feat)

tests\pytorch\test_nn.py:608: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:261: in forward
    graph.update_all(msg_fn, self._lstm_reducer)
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:362: in message_passing
    msgdata = invoke_gsddmm(g, mfunc)
python\dgl\core.py:276: in invoke_gsddmm
    z = op(graph, x)
python\dgl\ops\sddmm.py:164: in copy_u
    return gsddmm(g, 'copy_lhs', x, None)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0173C2C8>
op = 'copy_lhs'
lhs_data = tensor([[ 0.5247,  0.4713, -0.7374,  1.1520, -1.1929],
        [-1.1691, -0.3961,  0.8512, -0.0165,  0.0934],
        ...1.0899],
        [ 0.9810,  2.1578, -0.1592, -1.0056, -0.5694],
        [-0.0473, -0.0778, -0.1367, -0.7316, -0.4092]])
rhs_data = None, lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_______________________ test_sage_conv[lstm-g3-idtype1] _______________________

idtype = torch.int64
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
aggre_type = 'lstm'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    def test_sage_conv(idtype, g, aggre_type):
        g = g.astype(idtype).to(F.ctx())
        sage = nn.SAGEConv(5, 10, aggre_type)
        feat = F.randn((g.number_of_src_nodes(), 5))
        sage = sage.to(F.ctx())
        # test pickle
        th.save(sage, tmp_buffer)
>       h = sage(g, feat)

tests\pytorch\test_nn.py:608: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:261: in forward
    graph.update_all(msg_fn, self._lstm_reducer)
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:362: in message_passing
    msgdata = invoke_gsddmm(g, mfunc)
python\dgl\core.py:276: in invoke_gsddmm
    z = op(graph, x)
python\dgl\ops\sddmm.py:164: in copy_u
    return gsddmm(g, 'copy_lhs', x, None)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0168CD48>
op = 'copy_lhs'
lhs_data = tensor([[-0.0693, -0.6114, -0.3043, -0.7791, -0.1575],
        [ 0.1809, -0.2016,  1.0209, -1.1739,  2.1141],
        ...0.0052],
        [ 0.7676,  1.0426, -0.9768, -0.3668, -1.0214],
        [ 1.5826, -0.2150, -0.0765, -2.7617,  0.6783]])
rhs_data = None, lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_______________________ test_sage_conv[lstm-g4-idtype0] _______________________

idtype = torch.int32
g = Graph(num_nodes=5, num_edges=6,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
aggre_type = 'lstm'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    def test_sage_conv(idtype, g, aggre_type):
        g = g.astype(idtype).to(F.ctx())
        sage = nn.SAGEConv(5, 10, aggre_type)
        feat = F.randn((g.number_of_src_nodes(), 5))
        sage = sage.to(F.ctx())
        # test pickle
        th.save(sage, tmp_buffer)
>       h = sage(g, feat)

tests\pytorch\test_nn.py:608: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:261: in forward
    graph.update_all(msg_fn, self._lstm_reducer)
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:362: in message_passing
    msgdata = invoke_gsddmm(g, mfunc)
python\dgl\core.py:276: in invoke_gsddmm
    z = op(graph, x)
python\dgl\ops\sddmm.py:164: in copy_u
    return gsddmm(g, 'copy_lhs', x, None)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0199B3C8>
op = 'copy_lhs'
lhs_data = tensor([[-0.6559,  0.8538, -1.3506, -0.4642, -0.9858],
        [-0.8282,  0.1550, -1.8817, -1.9218,  1.4038],
        ...0.3345],
        [-0.5609,  0.6256,  1.3528, -0.4574,  0.4690],
        [-0.4918,  0.2804,  0.7062, -0.7548, -0.2902]])
rhs_data = None, lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_______________________ test_sage_conv[lstm-g4-idtype1] _______________________

idtype = torch.int64
g = Graph(num_nodes=5, num_edges=6,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
aggre_type = 'lstm'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    def test_sage_conv(idtype, g, aggre_type):
        g = g.astype(idtype).to(F.ctx())
        sage = nn.SAGEConv(5, 10, aggre_type)
        feat = F.randn((g.number_of_src_nodes(), 5))
        sage = sage.to(F.ctx())
        # test pickle
        th.save(sage, tmp_buffer)
>       h = sage(g, feat)

tests\pytorch\test_nn.py:608: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:261: in forward
    graph.update_all(msg_fn, self._lstm_reducer)
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:362: in message_passing
    msgdata = invoke_gsddmm(g, mfunc)
python\dgl\core.py:276: in invoke_gsddmm
    z = op(graph, x)
python\dgl\ops\sddmm.py:164: in copy_u
    return gsddmm(g, 'copy_lhs', x, None)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01846688>
op = 'copy_lhs'
lhs_data = tensor([[-1.4225,  0.0898,  0.2211,  2.1565, -0.3492],
        [ 2.1033, -1.4210,  0.0624, -0.4144, -0.3323],
        ...1.5091],
        [ 1.3346,  0.9257,  0.1212,  0.0998, -1.7716],
        [-1.2207,  0.3490, -0.9693,  1.5240, -1.7165]])
rhs_data = None, lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_______________________ test_sage_conv[lstm-g5-idtype0] _______________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
aggre_type = 'lstm'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    def test_sage_conv(idtype, g, aggre_type):
        g = g.astype(idtype).to(F.ctx())
        sage = nn.SAGEConv(5, 10, aggre_type)
        feat = F.randn((g.number_of_src_nodes(), 5))
        sage = sage.to(F.ctx())
        # test pickle
        th.save(sage, tmp_buffer)
>       h = sage(g, feat)

tests\pytorch\test_nn.py:608: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:261: in forward
    graph.update_all(msg_fn, self._lstm_reducer)
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:362: in message_passing
    msgdata = invoke_gsddmm(g, mfunc)
python\dgl\core.py:276: in invoke_gsddmm
    z = op(graph, x)
python\dgl\ops\sddmm.py:164: in copy_u
    return gsddmm(g, 'copy_lhs', x, None)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF030ED288>
op = 'copy_lhs'
lhs_data = tensor([[-0.6374, -0.9077, -0.9458, -0.0568,  0.8216],
        [ 0.0236, -0.4668, -0.1924,  1.3923,  0.4500],
        ...1.3198],
        [-0.5151, -0.2244, -0.1011, -0.3936, -1.6569],
        [-0.8406,  0.5607,  1.2703, -0.6100,  1.0739]])
rhs_data = None, lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_______________________ test_sage_conv[lstm-g5-idtype1] _______________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
aggre_type = 'lstm'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    def test_sage_conv(idtype, g, aggre_type):
        g = g.astype(idtype).to(F.ctx())
        sage = nn.SAGEConv(5, 10, aggre_type)
        feat = F.randn((g.number_of_src_nodes(), 5))
        sage = sage.to(F.ctx())
        # test pickle
        th.save(sage, tmp_buffer)
>       h = sage(g, feat)

tests\pytorch\test_nn.py:608: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:261: in forward
    graph.update_all(msg_fn, self._lstm_reducer)
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:362: in message_passing
    msgdata = invoke_gsddmm(g, mfunc)
python\dgl\core.py:276: in invoke_gsddmm
    z = op(graph, x)
python\dgl\ops\sddmm.py:164: in copy_u
    return gsddmm(g, 'copy_lhs', x, None)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01906C08>
op = 'copy_lhs'
lhs_data = tensor([[ 0.0962,  0.6598,  0.4890,  0.0233,  0.2376],
        [-1.5510, -0.7440, -0.3303,  0.8969,  1.4227],
        ...1.2613],
        [-0.2466,  1.2698, -0.8042,  1.0269, -1.2107],
        [ 1.6837, -0.3120,  0.1987, -0.6342, -0.1233]])
rhs_data = None, lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_______________________ test_sage_conv[lstm-g6-idtype0] _______________________

idtype = torch.int32, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
aggre_type = 'lstm'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    def test_sage_conv(idtype, g, aggre_type):
        g = g.astype(idtype).to(F.ctx())
        sage = nn.SAGEConv(5, 10, aggre_type)
        feat = F.randn((g.number_of_src_nodes(), 5))
        sage = sage.to(F.ctx())
        # test pickle
        th.save(sage, tmp_buffer)
>       h = sage(g, feat)

tests\pytorch\test_nn.py:608: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:261: in forward
    graph.update_all(msg_fn, self._lstm_reducer)
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:362: in message_passing
    msgdata = invoke_gsddmm(g, mfunc)
python\dgl\core.py:276: in invoke_gsddmm
    z = op(graph, x)
python\dgl\ops\sddmm.py:164: in copy_u
    return gsddmm(g, 'copy_lhs', x, None)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01100CC8>
op = 'copy_lhs'
lhs_data = tensor([[ 0.5407, -0.4451,  0.6384, -0.6729, -1.2293],
        [ 1.4116, -0.6518,  0.7805,  1.1927, -1.2893],
        ...0.8235],
        [ 0.9921, -0.3395,  0.8530,  1.2187,  1.7495],
        [-0.8063, -0.2532,  1.6409,  0.9818,  0.1885]])
rhs_data = None, lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_______________________ test_sage_conv[lstm-g6-idtype1] _______________________

idtype = torch.int64, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
aggre_type = 'lstm'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    def test_sage_conv(idtype, g, aggre_type):
        g = g.astype(idtype).to(F.ctx())
        sage = nn.SAGEConv(5, 10, aggre_type)
        feat = F.randn((g.number_of_src_nodes(), 5))
        sage = sage.to(F.ctx())
        # test pickle
        th.save(sage, tmp_buffer)
>       h = sage(g, feat)

tests\pytorch\test_nn.py:608: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:261: in forward
    graph.update_all(msg_fn, self._lstm_reducer)
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:362: in message_passing
    msgdata = invoke_gsddmm(g, mfunc)
python\dgl\core.py:276: in invoke_gsddmm
    z = op(graph, x)
python\dgl\ops\sddmm.py:164: in copy_u
    return gsddmm(g, 'copy_lhs', x, None)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03CC1E88>
op = 'copy_lhs'
lhs_data = tensor([[ 6.4376e-01, -2.1773e+00, -2.8430e-01, -8.4108e-01,  2.1034e-01],
        [ 1.9406e-03,  3.1938e-01,  1.2224e...01, -1.4588e+00,  7.5040e-01,  3.7610e-01],
        [-2.9940e-02, -5.8264e-01,  2.6539e-01, -2.4611e-01, -6.9269e-01]])
rhs_data = None, lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_______________________ test_sage_conv[lstm-g7-idtype0] _______________________

idtype = torch.int32
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
aggre_type = 'lstm'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    def test_sage_conv(idtype, g, aggre_type):
        g = g.astype(idtype).to(F.ctx())
        sage = nn.SAGEConv(5, 10, aggre_type)
        feat = F.randn((g.number_of_src_nodes(), 5))
        sage = sage.to(F.ctx())
        # test pickle
        th.save(sage, tmp_buffer)
>       h = sage(g, feat)

tests\pytorch\test_nn.py:608: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:261: in forward
    graph.update_all(msg_fn, self._lstm_reducer)
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:362: in message_passing
    msgdata = invoke_gsddmm(g, mfunc)
python\dgl\core.py:276: in invoke_gsddmm
    z = op(graph, x)
python\dgl\ops\sddmm.py:164: in copy_u
    return gsddmm(g, 'copy_lhs', x, None)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF014C52C8>
op = 'copy_lhs'
lhs_data = tensor([[-1.0313, -0.1786,  1.4341,  0.5708,  0.9777],
        [ 1.0823, -1.3628, -1.9957, -1.2569,  1.0563],
        ...0.3093],
        [-0.5011, -1.4219,  0.5885, -0.6005,  0.8242],
        [-0.5651,  0.7029,  0.2103, -0.4807,  0.3125]])
rhs_data = None, lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_______________________ test_sage_conv[lstm-g7-idtype1] _______________________

idtype = torch.int64
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
aggre_type = 'lstm'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    def test_sage_conv(idtype, g, aggre_type):
        g = g.astype(idtype).to(F.ctx())
        sage = nn.SAGEConv(5, 10, aggre_type)
        feat = F.randn((g.number_of_src_nodes(), 5))
        sage = sage.to(F.ctx())
        # test pickle
        th.save(sage, tmp_buffer)
>       h = sage(g, feat)

tests\pytorch\test_nn.py:608: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:261: in forward
    graph.update_all(msg_fn, self._lstm_reducer)
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:362: in message_passing
    msgdata = invoke_gsddmm(g, mfunc)
python\dgl\core.py:276: in invoke_gsddmm
    z = op(graph, x)
python\dgl\ops\sddmm.py:164: in copy_u
    return gsddmm(g, 'copy_lhs', x, None)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01925948>
op = 'copy_lhs'
lhs_data = tensor([[ 0.0700, -0.3559, -0.0451, -0.3154, -0.4079],
        [ 0.5923, -0.0578,  0.5590, -0.2135, -1.7571],
        ...0.4190],
        [ 0.2922,  1.2955,  0.6359, -1.0558,  0.9277],
        [ 0.9773, -0.5745, -0.2773,  1.2643,  1.4278]])
rhs_data = None, lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
____________________ test_sage_conv_bi[1-mean-g0-idtype0] _____________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 4, '_V': 5},
      num_edges={('_U', '_E', '_V'): 6},
      metagraph=[('_U', '_V', '_E')])
aggre_type = 'mean', out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_sage_conv_bi(idtype, g, aggre_type, out_dim):
        g = g.astype(idtype).to(F.ctx())
        dst_dim = 5 if aggre_type != 'gcn' else 10
        sage = nn.SAGEConv((10, dst_dim), out_dim, aggre_type)
        feat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), dst_dim)))
        sage = sage.to(F.ctx())
>       h = sage(g, feat)

tests\pytorch\test_nn.py:621: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:235: in forward
    graph.update_all(msg_fn, fn.mean('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07040F88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 2.1250],
        [ 2.2057],
        [-0.8784],
        [-0.9439]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
____________________ test_sage_conv_bi[1-mean-g0-idtype1] _____________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 4, '_V': 5},
      num_edges={('_U', '_E', '_V'): 6},
      metagraph=[('_U', '_V', '_E')])
aggre_type = 'mean', out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_sage_conv_bi(idtype, g, aggre_type, out_dim):
        g = g.astype(idtype).to(F.ctx())
        dst_dim = 5 if aggre_type != 'gcn' else 10
        sage = nn.SAGEConv((10, dst_dim), out_dim, aggre_type)
        feat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), dst_dim)))
        sage = sage.to(F.ctx())
>       h = sage(g, feat)

tests\pytorch\test_nn.py:621: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:235: in forward
    graph.update_all(msg_fn, fn.mean('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01731188>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 2.5425],
        [ 2.4914],
        [-1.5760],
        [-0.2270]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
____________________ test_sage_conv_bi[1-mean-g1-idtype0] _____________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
aggre_type = 'mean', out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_sage_conv_bi(idtype, g, aggre_type, out_dim):
        g = g.astype(idtype).to(F.ctx())
        dst_dim = 5 if aggre_type != 'gcn' else 10
        sage = nn.SAGEConv((10, dst_dim), out_dim, aggre_type)
        feat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), dst_dim)))
        sage = sage.to(F.ctx())
>       h = sage(g, feat)

tests\pytorch\test_nn.py:621: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:235: in forward
    graph.update_all(msg_fn, fn.mean('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE145DF808>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 3.6887],
        [-1.1183]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
____________________ test_sage_conv_bi[1-mean-g1-idtype1] _____________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
aggre_type = 'mean', out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_sage_conv_bi(idtype, g, aggre_type, out_dim):
        g = g.astype(idtype).to(F.ctx())
        dst_dim = 5 if aggre_type != 'gcn' else 10
        sage = nn.SAGEConv((10, dst_dim), out_dim, aggre_type)
        feat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), dst_dim)))
        sage = sage.to(F.ctx())
>       h = sage(g, feat)

tests\pytorch\test_nn.py:621: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:235: in forward
    graph.update_all(msg_fn, fn.mean('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF015B4148>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.1828],
        [-1.3385]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
____________________ test_sage_conv_bi[1-mean-g2-idtype0] _____________________

idtype = torch.int32, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
aggre_type = 'mean', out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_sage_conv_bi(idtype, g, aggre_type, out_dim):
        g = g.astype(idtype).to(F.ctx())
        dst_dim = 5 if aggre_type != 'gcn' else 10
        sage = nn.SAGEConv((10, dst_dim), out_dim, aggre_type)
        feat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), dst_dim)))
        sage = sage.to(F.ctx())
>       h = sage(g, feat)

tests\pytorch\test_nn.py:621: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:235: in forward
    graph.update_all(msg_fn, fn.mean('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07031808>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.4523],
        [ 1.2955],
        [ 0.8773],
        [-1.1241],
        [-0.8164],
        [-2.0507]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
____________________ test_sage_conv_bi[1-mean-g2-idtype1] _____________________

idtype = torch.int64, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
aggre_type = 'mean', out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_sage_conv_bi(idtype, g, aggre_type, out_dim):
        g = g.astype(idtype).to(F.ctx())
        dst_dim = 5 if aggre_type != 'gcn' else 10
        sage = nn.SAGEConv((10, dst_dim), out_dim, aggre_type)
        feat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), dst_dim)))
        sage = sage.to(F.ctx())
>       h = sage(g, feat)

tests\pytorch\test_nn.py:621: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:235: in forward
    graph.update_all(msg_fn, fn.mean('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF017C5DC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.8007],
        [ 0.7882],
        [-0.7265],
        [ 2.2861],
        [-2.0374],
        [ 1.6006]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
____________________ test_sage_conv_bi[1-pool-g0-idtype0] _____________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 4, '_V': 5},
      num_edges={('_U', '_E', '_V'): 6},
      metagraph=[('_U', '_V', '_E')])
aggre_type = 'pool', out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_sage_conv_bi(idtype, g, aggre_type, out_dim):
        g = g.astype(idtype).to(F.ctx())
        dst_dim = 5 if aggre_type != 'gcn' else 10
        sage = nn.SAGEConv((10, dst_dim), out_dim, aggre_type)
        feat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), dst_dim)))
        sage = sage.to(F.ctx())
>       h = sage(g, feat)

tests\pytorch\test_nn.py:621: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:257: in forward
    graph.update_all(msg_fn, fn.max('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF015399C8>
op = 'copy_lhs', reduce_op = 'max'
lhs_data = tensor([[0.0000, 0.0000, 1.7026, 0.0000, 0.0989, 1.8490, 0.0000, 0.0000, 2.6321,
         1.0383],
        [2.2170, 1....   [0.0000, 0.0482, 0.0000, 0.0000, 0.6505, 0.5561, 0.0000, 0.0000, 0.2682,
         0.6782]], grad_fn=<ReluBackward0>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
____________________ test_sage_conv_bi[1-pool-g0-idtype1] _____________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 4, '_V': 5},
      num_edges={('_U', '_E', '_V'): 6},
      metagraph=[('_U', '_V', '_E')])
aggre_type = 'pool', out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_sage_conv_bi(idtype, g, aggre_type, out_dim):
        g = g.astype(idtype).to(F.ctx())
        dst_dim = 5 if aggre_type != 'gcn' else 10
        sage = nn.SAGEConv((10, dst_dim), out_dim, aggre_type)
        feat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), dst_dim)))
        sage = sage.to(F.ctx())
>       h = sage(g, feat)

tests\pytorch\test_nn.py:621: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:257: in forward
    graph.update_all(msg_fn, fn.max('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01488048>
op = 'copy_lhs', reduce_op = 'max'
lhs_data = tensor([[0.8398, 0.0000, 0.0000, 0.9895, 0.0305, 0.0077, 0.0000, 1.1758, 0.0000,
         0.0000],
        [0.7710, 2....   [0.0000, 0.0000, 0.0000, 1.1543, 0.6486, 0.0000, 0.0000, 0.0000, 1.2503,
         0.0000]], grad_fn=<ReluBackward0>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
____________________ test_sage_conv_bi[1-pool-g1-idtype0] _____________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
aggre_type = 'pool', out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_sage_conv_bi(idtype, g, aggre_type, out_dim):
        g = g.astype(idtype).to(F.ctx())
        dst_dim = 5 if aggre_type != 'gcn' else 10
        sage = nn.SAGEConv((10, dst_dim), out_dim, aggre_type)
        feat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), dst_dim)))
        sage = sage.to(F.ctx())
>       h = sage(g, feat)

tests\pytorch\test_nn.py:621: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:257: in forward
    graph.update_all(msg_fn, fn.max('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14CCFA48>
op = 'copy_lhs', reduce_op = 'max'
lhs_data = tensor([[0.0000, 0.0000, 0.0000, 0.3445, 0.0000, 0.0000, 0.9821, 0.0000, 0.4790,
         0.0000],
        [0.0325, 0.0000, 0.2090, 0.1014, 0.0000, 0.3628, 0.0000, 0.0516, 0.0000,
         0.0000]], grad_fn=<ReluBackward0>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
____________________ test_sage_conv_bi[1-pool-g1-idtype1] _____________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
aggre_type = 'pool', out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_sage_conv_bi(idtype, g, aggre_type, out_dim):
        g = g.astype(idtype).to(F.ctx())
        dst_dim = 5 if aggre_type != 'gcn' else 10
        sage = nn.SAGEConv((10, dst_dim), out_dim, aggre_type)
        feat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), dst_dim)))
        sage = sage.to(F.ctx())
>       h = sage(g, feat)

tests\pytorch\test_nn.py:621: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:257: in forward
    graph.update_all(msg_fn, fn.max('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0167EAC8>
op = 'copy_lhs', reduce_op = 'max'
lhs_data = tensor([[1.0928, 1.2135, 0.0038, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000,
         0.0960],
        [0.0000, 2.6516, 0.1269, 0.0000, 0.0000, 0.0000, 0.1389, 0.0000, 0.0000,
         0.0000]], grad_fn=<ReluBackward0>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
____________________ test_sage_conv_bi[1-pool-g2-idtype0] _____________________

idtype = torch.int32, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
aggre_type = 'pool', out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_sage_conv_bi(idtype, g, aggre_type, out_dim):
        g = g.astype(idtype).to(F.ctx())
        dst_dim = 5 if aggre_type != 'gcn' else 10
        sage = nn.SAGEConv((10, dst_dim), out_dim, aggre_type)
        feat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), dst_dim)))
        sage = sage.to(F.ctx())
>       h = sage(g, feat)

tests\pytorch\test_nn.py:621: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:257: in forward
    graph.update_all(msg_fn, fn.max('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1452DFC8>
op = 'copy_lhs', reduce_op = 'max'
lhs_data = tensor([[0.0000, 0.6514, 0.0000, 0.0000, 0.8830, 4.0294, 0.2809, 0.0000, 0.0000,
         1.0419],
        [1.1047, 0....   [0.4095, 0.0000, 0.0000, 0.4766, 1.5695, 1.9555, 0.0000, 0.3870, 0.0000,
         1.3458]], grad_fn=<ReluBackward0>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
____________________ test_sage_conv_bi[1-pool-g2-idtype1] _____________________

idtype = torch.int64, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
aggre_type = 'pool', out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_sage_conv_bi(idtype, g, aggre_type, out_dim):
        g = g.astype(idtype).to(F.ctx())
        dst_dim = 5 if aggre_type != 'gcn' else 10
        sage = nn.SAGEConv((10, dst_dim), out_dim, aggre_type)
        feat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), dst_dim)))
        sage = sage.to(F.ctx())
>       h = sage(g, feat)

tests\pytorch\test_nn.py:621: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:257: in forward
    graph.update_all(msg_fn, fn.max('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01514808>
op = 'copy_lhs', reduce_op = 'max'
lhs_data = tensor([[0.0000, 0.8103, 0.0000, 0.0510, 2.6135, 0.6568, 2.1943, 0.0000, 0.1374,
         0.0000],
        [0.0035, 3....   [0.0000, 2.2380, 0.0000, 1.5116, 1.0177, 1.4157, 1.1628, 0.1126, 3.1787,
         0.3180]], grad_fn=<ReluBackward0>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_____________________ test_sage_conv_bi[1-gcn-g0-idtype0] _____________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 4, '_V': 5},
      num_edges={('_U', '_E', '_V'): 6},
      metagraph=[('_U', '_V', '_E')])
aggre_type = 'gcn', out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_sage_conv_bi(idtype, g, aggre_type, out_dim):
        g = g.astype(idtype).to(F.ctx())
        dst_dim = 5 if aggre_type != 'gcn' else 10
        sage = nn.SAGEConv((10, dst_dim), out_dim, aggre_type)
        feat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), dst_dim)))
        sage = sage.to(F.ctx())
>       h = sage(g, feat)

tests\pytorch\test_nn.py:621: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:249: in forward
    graph.update_all(msg_fn, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF030D7E08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.3361],
        [-0.1347],
        [ 3.9180],
        [ 2.4078]], grad_fn=<MmBackward>)


    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_____________________ test_sage_conv_bi[1-gcn-g0-idtype1] _____________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 4, '_V': 5},
      num_edges={('_U', '_E', '_V'): 6},
      metagraph=[('_U', '_V', '_E')])
aggre_type = 'gcn', out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_sage_conv_bi(idtype, g, aggre_type, out_dim):
        g = g.astype(idtype).to(F.ctx())
        dst_dim = 5 if aggre_type != 'gcn' else 10
        sage = nn.SAGEConv((10, dst_dim), out_dim, aggre_type)
        feat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), dst_dim)))
        sage = sage.to(F.ctx())
>       h = sage(g, feat)

tests\pytorch\test_nn.py:621: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:249: in forward
    graph.update_all(msg_fn, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF014C9388>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0.6484],
        [0.3302],
        [1.3775],
        [0.8802]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_____________________ test_sage_conv_bi[1-gcn-g1-idtype0] _____________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
aggre_type = 'gcn', out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_sage_conv_bi(idtype, g, aggre_type, out_dim):
        g = g.astype(idtype).to(F.ctx())
        dst_dim = 5 if aggre_type != 'gcn' else 10
        sage = nn.SAGEConv((10, dst_dim), out_dim, aggre_type)
        feat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), dst_dim)))
        sage = sage.to(F.ctx())
>       h = sage(g, feat)

tests\pytorch\test_nn.py:621: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:249: in forward
    graph.update_all(msg_fn, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF010F3048>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.8833],
        [-1.6978]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_____________________ test_sage_conv_bi[1-gcn-g1-idtype1] _____________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
aggre_type = 'gcn', out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_sage_conv_bi(idtype, g, aggre_type, out_dim):
        g = g.astype(idtype).to(F.ctx())
        dst_dim = 5 if aggre_type != 'gcn' else 10
        sage = nn.SAGEConv((10, dst_dim), out_dim, aggre_type)
        feat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), dst_dim)))
        sage = sage.to(F.ctx())
>       h = sage(g, feat)

tests\pytorch\test_nn.py:621: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:249: in forward
    graph.update_all(msg_fn, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF014B8E48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-2.1861],
        [ 1.3745]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_____________________ test_sage_conv_bi[1-gcn-g2-idtype0] _____________________

idtype = torch.int32, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
aggre_type = 'gcn', out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_sage_conv_bi(idtype, g, aggre_type, out_dim):
        g = g.astype(idtype).to(F.ctx())
        dst_dim = 5 if aggre_type != 'gcn' else 10
        sage = nn.SAGEConv((10, dst_dim), out_dim, aggre_type)
        feat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), dst_dim)))
        sage = sage.to(F.ctx())
>       h = sage(g, feat)

tests\pytorch\test_nn.py:621: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:249: in forward
    graph.update_all(msg_fn, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1226B988>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.0953],
        [-0.6711],
        [-2.9648],
        [ 1.3850],
        [-2.1642],
        [-0.4693]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_____________________ test_sage_conv_bi[1-gcn-g2-idtype1] _____________________

idtype = torch.int64, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
aggre_type = 'gcn', out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_sage_conv_bi(idtype, g, aggre_type, out_dim):
        g = g.astype(idtype).to(F.ctx())
        dst_dim = 5 if aggre_type != 'gcn' else 10
        sage = nn.SAGEConv((10, dst_dim), out_dim, aggre_type)
        feat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), dst_dim)))
        sage = sage.to(F.ctx())
>       h = sage(g, feat)

tests\pytorch\test_nn.py:621: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:249: in forward
    graph.update_all(msg_fn, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF014291C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.7613],
        [ 0.5151],
        [-1.8637],
        [-0.6371],
        [-0.3214],
        [ 0.7414]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
____________________ test_sage_conv_bi[1-lstm-g0-idtype0] _____________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 4, '_V': 5},
      num_edges={('_U', '_E', '_V'): 6},
      metagraph=[('_U', '_V', '_E')])
aggre_type = 'lstm', out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_sage_conv_bi(idtype, g, aggre_type, out_dim):
        g = g.astype(idtype).to(F.ctx())
        dst_dim = 5 if aggre_type != 'gcn' else 10
        sage = nn.SAGEConv((10, dst_dim), out_dim, aggre_type)
        feat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), dst_dim)))
        sage = sage.to(F.ctx())
>       h = sage(g, feat)

tests\pytorch\test_nn.py:621: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:261: in forward
    graph.update_all(msg_fn, self._lstm_reducer)
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:362: in message_passing
    msgdata = invoke_gsddmm(g, mfunc)
python\dgl\core.py:276: in invoke_gsddmm
    z = op(graph, x)
python\dgl\ops\sddmm.py:164: in copy_u
    return gsddmm(g, 'copy_lhs', x, None)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE145DCCC8>
op = 'copy_lhs'
lhs_data = tensor([[-1.7643, -2.0578,  1.7569,  1.0039,  1.9650,  0.7118,  0.9555, -0.6192,
         -0.3978, -1.2367],
        [... 0.2715],
        [ 0.2582,  2.0767, -0.2051,  1.0329,  0.7733, -0.1276,  1.2515, -0.2024,
         -0.6516,  0.4189]])
rhs_data = None, lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
____________________ test_sage_conv_bi[1-lstm-g0-idtype1] _____________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 4, '_V': 5},
      num_edges={('_U', '_E', '_V'): 6},
      metagraph=[('_U', '_V', '_E')])
aggre_type = 'lstm', out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_sage_conv_bi(idtype, g, aggre_type, out_dim):
        g = g.astype(idtype).to(F.ctx())
        dst_dim = 5 if aggre_type != 'gcn' else 10
        sage = nn.SAGEConv((10, dst_dim), out_dim, aggre_type)
        feat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), dst_dim)))
        sage = sage.to(F.ctx())
>       h = sage(g, feat)

tests\pytorch\test_nn.py:621: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:261: in forward
    graph.update_all(msg_fn, self._lstm_reducer)
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:362: in message_passing
    msgdata = invoke_gsddmm(g, mfunc)
python\dgl\core.py:276: in invoke_gsddmm
    z = op(graph, x)
python\dgl\ops\sddmm.py:164: in copy_u
    return gsddmm(g, 'copy_lhs', x, None)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF017A90C8>
op = 'copy_lhs'
lhs_data = tensor([[ 4.6498e-01,  1.3706e+00,  2.2719e-01,  3.5126e-01,  8.7833e-01,
         -5.2887e-01, -4.3218e-01, -1.1734e-...-01,  2.7957e-01,  6.6686e-01, -6.2588e-02,
          7.4312e-01,  8.7782e-01,  1.0510e+00,  3.7153e-01,  5.2126e-01]])
rhs_data = None, lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
____________________ test_sage_conv_bi[1-lstm-g1-idtype0] _____________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
aggre_type = 'lstm', out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_sage_conv_bi(idtype, g, aggre_type, out_dim):
        g = g.astype(idtype).to(F.ctx())
        dst_dim = 5 if aggre_type != 'gcn' else 10
        sage = nn.SAGEConv((10, dst_dim), out_dim, aggre_type)
        feat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), dst_dim)))
        sage = sage.to(F.ctx())
>       h = sage(g, feat)

tests\pytorch\test_nn.py:621: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:261: in forward
    graph.update_all(msg_fn, self._lstm_reducer)
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:362: in message_passing
    msgdata = invoke_gsddmm(g, mfunc)
python\dgl\core.py:276: in invoke_gsddmm
    z = op(graph, x)
python\dgl\ops\sddmm.py:164: in copy_u
    return gsddmm(g, 'copy_lhs', x, None)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0717C248>
op = 'copy_lhs'
lhs_data = tensor([[-0.1367,  0.5358,  2.4593,  0.8362, -1.9374, -0.2122, -0.0704, -0.1004,
         -1.3655, -0.0352],
        [ 0.1371,  0.2828, -0.2729, -0.1885, -0.8072,  0.4908, -0.4412, -0.5351,
         -0.4777, -1.2909]])
rhs_data = None, lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
____________________ test_sage_conv_bi[1-lstm-g1-idtype1] _____________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
aggre_type = 'lstm', out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_sage_conv_bi(idtype, g, aggre_type, out_dim):
        g = g.astype(idtype).to(F.ctx())
        dst_dim = 5 if aggre_type != 'gcn' else 10
        sage = nn.SAGEConv((10, dst_dim), out_dim, aggre_type)
        feat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), dst_dim)))
        sage = sage.to(F.ctx())
>       h = sage(g, feat)

tests\pytorch\test_nn.py:621: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:261: in forward
    graph.update_all(msg_fn, self._lstm_reducer)
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:362: in message_passing
    msgdata = invoke_gsddmm(g, mfunc)
python\dgl\core.py:276: in invoke_gsddmm
    z = op(graph, x)
python\dgl\ops\sddmm.py:164: in copy_u
    return gsddmm(g, 'copy_lhs', x, None)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF014C5308>
op = 'copy_lhs'
lhs_data = tensor([[ 1.4891, -0.1086, -0.2258, -0.9871,  0.2672, -0.2113,  0.0343, -0.5318,
         -0.5578,  0.4856],
        [-1.2217,  0.6697, -0.4637, -0.1946, -0.7720, -1.7658, -0.2480, -1.0679,
         -0.6720, -0.7057]])
rhs_data = None, lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
____________________ test_sage_conv_bi[1-lstm-g2-idtype0] _____________________

idtype = torch.int32, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
aggre_type = 'lstm', out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_sage_conv_bi(idtype, g, aggre_type, out_dim):
        g = g.astype(idtype).to(F.ctx())
        dst_dim = 5 if aggre_type != 'gcn' else 10
        sage = nn.SAGEConv((10, dst_dim), out_dim, aggre_type)
        feat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), dst_dim)))
        sage = sage.to(F.ctx())
>       h = sage(g, feat)

tests\pytorch\test_nn.py:621: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:261: in forward
    graph.update_all(msg_fn, self._lstm_reducer)
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:362: in message_passing
    msgdata = invoke_gsddmm(g, mfunc)
python\dgl\core.py:276: in invoke_gsddmm
    z = op(graph, x)
python\dgl\ops\sddmm.py:164: in copy_u
    return gsddmm(g, 'copy_lhs', x, None)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01892748>
op = 'copy_lhs'
lhs_data = tensor([[-1.1263, -1.0317,  2.3468,  0.5330, -0.2204,  0.2433, -1.3930,  0.2225,
         -0.0258, -1.0189],
        [...-0.3886],
        [ 0.9357, -0.5037, -1.3425,  0.3275, -0.5899, -0.1657,  0.0109,  0.7961,
         -0.2692,  0.9374]])
rhs_data = None, lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
____________________ test_sage_conv_bi[1-lstm-g2-idtype1] _____________________

idtype = torch.int64, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
aggre_type = 'lstm', out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_sage_conv_bi(idtype, g, aggre_type, out_dim):
        g = g.astype(idtype).to(F.ctx())
        dst_dim = 5 if aggre_type != 'gcn' else 10
        sage = nn.SAGEConv((10, dst_dim), out_dim, aggre_type)
        feat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), dst_dim)))
        sage = sage.to(F.ctx())
>       h = sage(g, feat)

tests\pytorch\test_nn.py:621: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:261: in forward
    graph.update_all(msg_fn, self._lstm_reducer)
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:362: in message_passing
    msgdata = invoke_gsddmm(g, mfunc)
python\dgl\core.py:276: in invoke_gsddmm
    z = op(graph, x)
python\dgl\ops\sddmm.py:164: in copy_u
    return gsddmm(g, 'copy_lhs', x, None)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01690DC8>
op = 'copy_lhs'
lhs_data = tensor([[-0.3295,  0.6882, -0.3799,  0.5197, -0.3487,  1.9932, -0.4821, -0.8118,
          1.5472,  1.6226],
        [...-0.6728],
        [ 0.2739,  1.2477,  0.9486, -0.0962,  0.2667, -0.5894,  0.7396,  0.3935,
          0.8567, -1.1778]])
rhs_data = None, lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
____________________ test_sage_conv_bi[2-mean-g0-idtype0] _____________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 4, '_V': 5},
      num_edges={('_U', '_E', '_V'): 6},
      metagraph=[('_U', '_V', '_E')])
aggre_type = 'mean', out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_sage_conv_bi(idtype, g, aggre_type, out_dim):
        g = g.astype(idtype).to(F.ctx())
        dst_dim = 5 if aggre_type != 'gcn' else 10
        sage = nn.SAGEConv((10, dst_dim), out_dim, aggre_type)
        feat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), dst_dim)))
        sage = sage.to(F.ctx())
>       h = sage(g, feat)

tests\pytorch\test_nn.py:621: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:235: in forward
    graph.update_all(msg_fn, fn.mean('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF017F7148>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.0489,  0.5645],
        [-0.3063, -2.2330],
        [-1.8266,  1.3116],
        [-3.9712, -2.3290]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
____________________ test_sage_conv_bi[2-mean-g0-idtype1] _____________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 4, '_V': 5},
      num_edges={('_U', '_E', '_V'): 6},
      metagraph=[('_U', '_V', '_E')])
aggre_type = 'mean', out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_sage_conv_bi(idtype, g, aggre_type, out_dim):
        g = g.astype(idtype).to(F.ctx())
        dst_dim = 5 if aggre_type != 'gcn' else 10
        sage = nn.SAGEConv((10, dst_dim), out_dim, aggre_type)
        feat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), dst_dim)))
        sage = sage.to(F.ctx())
>       h = sage(g, feat)

tests\pytorch\test_nn.py:621: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:235: in forward
    graph.update_all(msg_fn, fn.mean('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF016FA248>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.9249,  4.9890],
        [ 2.5473,  0.1085],
        [-3.9742,  1.1358],
        [-3.0672,  1.4615]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
____________________ test_sage_conv_bi[2-mean-g1-idtype0] _____________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
aggre_type = 'mean', out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_sage_conv_bi(idtype, g, aggre_type, out_dim):
        g = g.astype(idtype).to(F.ctx())
        dst_dim = 5 if aggre_type != 'gcn' else 10
        sage = nn.SAGEConv((10, dst_dim), out_dim, aggre_type)
        feat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), dst_dim)))
        sage = sage.to(F.ctx())
>       h = sage(g, feat)

tests\pytorch\test_nn.py:621: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:235: in forward
    graph.update_all(msg_fn, fn.mean('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14A828C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.8889,  0.2545],
        [-0.4874,  0.8715]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
____________________ test_sage_conv_bi[2-mean-g1-idtype1] _____________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
aggre_type = 'mean', out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_sage_conv_bi(idtype, g, aggre_type, out_dim):
        g = g.astype(idtype).to(F.ctx())
        dst_dim = 5 if aggre_type != 'gcn' else 10
        sage = nn.SAGEConv((10, dst_dim), out_dim, aggre_type)
        feat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), dst_dim)))
        sage = sage.to(F.ctx())
>       h = sage(g, feat)

tests\pytorch\test_nn.py:621: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:235: in forward
    graph.update_all(msg_fn, fn.mean('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF018CE488>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.5525, -1.4295],
        [-0.4075,  2.2417]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
____________________ test_sage_conv_bi[2-mean-g2-idtype0] _____________________

idtype = torch.int32, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
aggre_type = 'mean', out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_sage_conv_bi(idtype, g, aggre_type, out_dim):
        g = g.astype(idtype).to(F.ctx())
        dst_dim = 5 if aggre_type != 'gcn' else 10
        sage = nn.SAGEConv((10, dst_dim), out_dim, aggre_type)
        feat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), dst_dim)))
        sage = sage.to(F.ctx())
>       h = sage(g, feat)

tests\pytorch\test_nn.py:621: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:235: in forward
    graph.update_all(msg_fn, fn.mean('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE146ABC08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.8514, -0.5656],
        [ 1.8181, -2.6620],
        [-3.0471, -2.4134],
        [-1.3824, -0.6803],
        [ 1.6069, -0.8870],
        [-0.8269, -3.1130]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
____________________ test_sage_conv_bi[2-mean-g2-idtype1] _____________________

idtype = torch.int64, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
aggre_type = 'mean', out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_sage_conv_bi(idtype, g, aggre_type, out_dim):
        g = g.astype(idtype).to(F.ctx())
        dst_dim = 5 if aggre_type != 'gcn' else 10
        sage = nn.SAGEConv((10, dst_dim), out_dim, aggre_type)
        feat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), dst_dim)))
        sage = sage.to(F.ctx())
>       h = sage(g, feat)

tests\pytorch\test_nn.py:621: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:235: in forward
    graph.update_all(msg_fn, fn.mean('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01955388>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 2.1622, -0.7949],
        [-0.5807,  1.4902],
        [ 1.4357,  1.8260],
        [-1.1657, -0.6123],
        [-0.6318,  1.3783],
        [-1.6290, -0.4152]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
____________________ test_sage_conv_bi[2-pool-g0-idtype0] _____________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 4, '_V': 5},
      num_edges={('_U', '_E', '_V'): 6},
      metagraph=[('_U', '_V', '_E')])
aggre_type = 'pool', out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_sage_conv_bi(idtype, g, aggre_type, out_dim):
        g = g.astype(idtype).to(F.ctx())
        dst_dim = 5 if aggre_type != 'gcn' else 10
        sage = nn.SAGEConv((10, dst_dim), out_dim, aggre_type)
        feat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), dst_dim)))
        sage = sage.to(F.ctx())
>       h = sage(g, feat)

tests\pytorch\test_nn.py:621: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:257: in forward
    graph.update_all(msg_fn, fn.max('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01865108>
op = 'copy_lhs', reduce_op = 'max'
lhs_data = tensor([[0.0000, 0.0000, 0.0000, 0.0000, 0.2835, 2.0616, 0.0000, 0.9364, 0.0000,
         0.0000],
        [0.0137, 1....   [0.0000, 0.5000, 0.0000, 0.0000, 0.0000, 0.0000, 3.8947, 0.0000, 1.3107,
         0.0000]], grad_fn=<ReluBackward0>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
____________________ test_sage_conv_bi[2-pool-g0-idtype1] _____________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 4, '_V': 5},
      num_edges={('_U', '_E', '_V'): 6},
      metagraph=[('_U', '_V', '_E')])
aggre_type = 'pool', out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_sage_conv_bi(idtype, g, aggre_type, out_dim):
        g = g.astype(idtype).to(F.ctx())
        dst_dim = 5 if aggre_type != 'gcn' else 10
        sage = nn.SAGEConv((10, dst_dim), out_dim, aggre_type)
        feat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), dst_dim)))
        sage = sage.to(F.ctx())
>       h = sage(g, feat)

tests\pytorch\test_nn.py:621: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:257: in forward
    graph.update_all(msg_fn, fn.max('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14969E08>
op = 'copy_lhs', reduce_op = 'max'
lhs_data = tensor([[0.0000, 0.0000, 0.1842, 0.0000, 0.0000, 0.0000, 0.0000, 3.4215, 2.3464,
         0.0000],
        [0.0000, 0....   [0.0000, 0.0000, 0.1347, 0.0000, 0.0000, 0.0000, 0.0000, 3.2485, 3.9125,
         0.0000]], grad_fn=<ReluBackward0>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
____________________ test_sage_conv_bi[2-pool-g1-idtype0] _____________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
aggre_type = 'pool', out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_sage_conv_bi(idtype, g, aggre_type, out_dim):
        g = g.astype(idtype).to(F.ctx())
        dst_dim = 5 if aggre_type != 'gcn' else 10
        sage = nn.SAGEConv((10, dst_dim), out_dim, aggre_type)
        feat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), dst_dim)))
        sage = sage.to(F.ctx())
>       h = sage(g, feat)

tests\pytorch\test_nn.py:621: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:257: in forward
    graph.update_all(msg_fn, fn.max('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF010E7308>
op = 'copy_lhs', reduce_op = 'max'
lhs_data = tensor([[1.2699, 1.1667, 2.4329, 2.2943, 0.0000, 0.0000, 0.0000, 0.7549, 0.2093,
         0.0000],
        [1.8079, 0.0000, 0.0000, 1.2063, 1.3089, 0.0000, 0.0000, 0.7123, 0.5319,
         0.9047]], grad_fn=<ReluBackward0>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
____________________ test_sage_conv_bi[2-pool-g1-idtype1] _____________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
aggre_type = 'pool', out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_sage_conv_bi(idtype, g, aggre_type, out_dim):
        g = g.astype(idtype).to(F.ctx())
        dst_dim = 5 if aggre_type != 'gcn' else 10
        sage = nn.SAGEConv((10, dst_dim), out_dim, aggre_type)
        feat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), dst_dim)))
        sage = sage.to(F.ctx())
>       h = sage(g, feat)

tests\pytorch\test_nn.py:621: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:257: in forward
    graph.update_all(msg_fn, fn.max('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE149F4488>
op = 'copy_lhs', reduce_op = 'max'
lhs_data = tensor([[0.0000, 0.2261, 0.0000, 0.0000, 0.0000, 1.9870, 0.0000, 0.0000, 0.5471,
         0.0000],
        [0.0000, 0.0000, 0.0000, 0.0000, 3.4257, 0.1079, 0.0000, 0.0000, 3.2329,
         0.0000]], grad_fn=<ReluBackward0>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
____________________ test_sage_conv_bi[2-pool-g2-idtype0] _____________________

idtype = torch.int32, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
aggre_type = 'pool', out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_sage_conv_bi(idtype, g, aggre_type, out_dim):
        g = g.astype(idtype).to(F.ctx())
        dst_dim = 5 if aggre_type != 'gcn' else 10
        sage = nn.SAGEConv((10, dst_dim), out_dim, aggre_type)
        feat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), dst_dim)))
        sage = sage.to(F.ctx())
>       h = sage(g, feat)

tests\pytorch\test_nn.py:621: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:257: in forward
    graph.update_all(msg_fn, fn.max('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0175E648>
op = 'copy_lhs', reduce_op = 'max'
lhs_data = tensor([[0.0000, 0.0000, 2.3856, 0.0000, 3.5579, 0.0747, 0.0000, 0.2106, 1.3117,
         0.4992],
        [0.0000, 0....   [0.1936, 1.3845, 0.0000, 0.1639, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000,
         0.0000]], grad_fn=<ReluBackward0>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
____________________ test_sage_conv_bi[2-pool-g2-idtype1] _____________________

idtype = torch.int64, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
aggre_type = 'pool', out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_sage_conv_bi(idtype, g, aggre_type, out_dim):
        g = g.astype(idtype).to(F.ctx())
        dst_dim = 5 if aggre_type != 'gcn' else 10
        sage = nn.SAGEConv((10, dst_dim), out_dim, aggre_type)
        feat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), dst_dim)))
        sage = sage.to(F.ctx())
>       h = sage(g, feat)

tests\pytorch\test_nn.py:621: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:257: in forward
    graph.update_all(msg_fn, fn.max('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01738548>
op = 'copy_lhs', reduce_op = 'max'
lhs_data = tensor([[0.0000, 0.0000, 1.6486, 0.8147, 0.0000, 0.0000, 0.8223, 0.0685, 1.3873,
         0.2088],
        [0.0000, 2....   [0.7101, 0.0000, 1.3674, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000,
         0.0000]], grad_fn=<ReluBackward0>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_____________________ test_sage_conv_bi[2-gcn-g0-idtype0] _____________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 4, '_V': 5},
      num_edges={('_U', '_E', '_V'): 6},
      metagraph=[('_U', '_V', '_E')])
aggre_type = 'gcn', out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_sage_conv_bi(idtype, g, aggre_type, out_dim):
        g = g.astype(idtype).to(F.ctx())
        dst_dim = 5 if aggre_type != 'gcn' else 10
        sage = nn.SAGEConv((10, dst_dim), out_dim, aggre_type)
        feat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), dst_dim)))
        sage = sage.to(F.ctx())
>       h = sage(g, feat)

tests\pytorch\test_nn.py:621: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:249: in forward
    graph.update_all(msg_fn, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF014F4F88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.0646,  0.8017],
        [ 0.3282,  0.2623],
        [-1.7307,  1.4024],
        [ 1.3608, -1.4565]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_____________________ test_sage_conv_bi[2-gcn-g0-idtype1] _____________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 4, '_V': 5},
      num_edges={('_U', '_E', '_V'): 6},
      metagraph=[('_U', '_V', '_E')])
aggre_type = 'gcn', out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_sage_conv_bi(idtype, g, aggre_type, out_dim):
        g = g.astype(idtype).to(F.ctx())
        dst_dim = 5 if aggre_type != 'gcn' else 10
        sage = nn.SAGEConv((10, dst_dim), out_dim, aggre_type)
        feat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), dst_dim)))
        sage = sage.to(F.ctx())
>       h = sage(g, feat)

tests\pytorch\test_nn.py:621: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:249: in forward
    graph.update_all(msg_fn, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03D5EFC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 2.4131, -1.0959],
        [ 0.2039, -2.2723],
        [-3.7041,  2.3179],
        [ 1.1539,  0.8174]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_____________________ test_sage_conv_bi[2-gcn-g1-idtype0] _____________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
aggre_type = 'gcn', out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_sage_conv_bi(idtype, g, aggre_type, out_dim):
        g = g.astype(idtype).to(F.ctx())
        dst_dim = 5 if aggre_type != 'gcn' else 10
        sage = nn.SAGEConv((10, dst_dim), out_dim, aggre_type)
        feat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), dst_dim)))
        sage = sage.to(F.ctx())
>       h = sage(g, feat)

tests\pytorch\test_nn.py:621: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:249: in forward
    graph.update_all(msg_fn, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0720FD88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[4.0023, 2.3261],
        [1.6924, 1.8011]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_____________________ test_sage_conv_bi[2-gcn-g1-idtype1] _____________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
aggre_type = 'gcn', out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_sage_conv_bi(idtype, g, aggre_type, out_dim):
        g = g.astype(idtype).to(F.ctx())
        dst_dim = 5 if aggre_type != 'gcn' else 10
        sage = nn.SAGEConv((10, dst_dim), out_dim, aggre_type)
        feat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), dst_dim)))
        sage = sage.to(F.ctx())
>       h = sage(g, feat)

tests\pytorch\test_nn.py:621: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:249: in forward
    graph.update_all(msg_fn, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF011D0A48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.9533, -0.7720],
        [-2.8326, -1.3237]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_____________________ test_sage_conv_bi[2-gcn-g2-idtype0] _____________________

idtype = torch.int32, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
aggre_type = 'gcn', out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_sage_conv_bi(idtype, g, aggre_type, out_dim):
        g = g.astype(idtype).to(F.ctx())
        dst_dim = 5 if aggre_type != 'gcn' else 10
        sage = nn.SAGEConv((10, dst_dim), out_dim, aggre_type)
        feat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), dst_dim)))
        sage = sage.to(F.ctx())
>       h = sage(g, feat)

tests\pytorch\test_nn.py:621: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:249: in forward
    graph.update_all(msg_fn, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14617708>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.7798,  2.0964],
        [-3.1119,  2.5519],
        [-1.0109, -1.0529],
        [-3.0597,  0.5489],
        [-1.7212, -2.1765],
        [ 0.3687,  0.0886]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_____________________ test_sage_conv_bi[2-gcn-g2-idtype1] _____________________

idtype = torch.int64, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
aggre_type = 'gcn', out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_sage_conv_bi(idtype, g, aggre_type, out_dim):
        g = g.astype(idtype).to(F.ctx())
        dst_dim = 5 if aggre_type != 'gcn' else 10
        sage = nn.SAGEConv((10, dst_dim), out_dim, aggre_type)
        feat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), dst_dim)))
        sage = sage.to(F.ctx())
>       h = sage(g, feat)

tests\pytorch\test_nn.py:621: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:249: in forward
    graph.update_all(msg_fn, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF018E8188>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-3.9102,  0.3674],
        [ 0.0179, -0.5913],
        [ 4.9297, -0.6748],
        [-0.8237,  2.1899],
        [ 0.2711,  0.5452],
        [-4.4050,  2.0423]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
____________________ test_sage_conv_bi[2-lstm-g0-idtype0] _____________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 4, '_V': 5},
      num_edges={('_U', '_E', '_V'): 6},
      metagraph=[('_U', '_V', '_E')])
aggre_type = 'lstm', out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_sage_conv_bi(idtype, g, aggre_type, out_dim):
        g = g.astype(idtype).to(F.ctx())
        dst_dim = 5 if aggre_type != 'gcn' else 10
        sage = nn.SAGEConv((10, dst_dim), out_dim, aggre_type)
        feat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), dst_dim)))
        sage = sage.to(F.ctx())
>       h = sage(g, feat)

tests\pytorch\test_nn.py:621: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:261: in forward
    graph.update_all(msg_fn, self._lstm_reducer)
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:362: in message_passing
    msgdata = invoke_gsddmm(g, mfunc)
python\dgl\core.py:276: in invoke_gsddmm
    z = op(graph, x)
python\dgl\ops\sddmm.py:164: in copy_u
    return gsddmm(g, 'copy_lhs', x, None)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01898B48>
op = 'copy_lhs'
lhs_data = tensor([[-0.8607,  0.3098,  0.2808, -0.1653, -0.1592,  0.0485, -0.5745, -0.5345,
          1.8666, -1.4028],
        [... 0.1401],
        [ 0.2440,  0.5697,  1.3561,  1.3104, -0.9470, -1.6598,  0.3752,  0.2317,
          0.1118,  0.7133]])
rhs_data = None, lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
____________________ test_sage_conv_bi[2-lstm-g0-idtype1] _____________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 4, '_V': 5},
      num_edges={('_U', '_E', '_V'): 6},
      metagraph=[('_U', '_V', '_E')])
aggre_type = 'lstm', out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_sage_conv_bi(idtype, g, aggre_type, out_dim):
        g = g.astype(idtype).to(F.ctx())
        dst_dim = 5 if aggre_type != 'gcn' else 10
        sage = nn.SAGEConv((10, dst_dim), out_dim, aggre_type)
        feat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), dst_dim)))
        sage = sage.to(F.ctx())
>       h = sage(g, feat)

tests\pytorch\test_nn.py:621: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:261: in forward
    graph.update_all(msg_fn, self._lstm_reducer)
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:362: in message_passing
    msgdata = invoke_gsddmm(g, mfunc)
python\dgl\core.py:276: in invoke_gsddmm
    z = op(graph, x)
python\dgl\ops\sddmm.py:164: in copy_u
    return gsddmm(g, 'copy_lhs', x, None)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF011E18C8>
op = 'copy_lhs'
lhs_data = tensor([[ 0.3738,  0.6188,  1.1808, -0.2378,  0.4823,  0.5211, -1.4715,  0.7935,
          0.7130,  0.1403],
        [...-1.3522],
        [ 0.3115,  0.4445, -0.1987, -1.0467, -2.6066, -1.7669,  0.4564,  1.2884,
         -0.3591, -1.5631]])
rhs_data = None, lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
____________________ test_sage_conv_bi[2-lstm-g1-idtype0] _____________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
aggre_type = 'lstm', out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_sage_conv_bi(idtype, g, aggre_type, out_dim):
        g = g.astype(idtype).to(F.ctx())
        dst_dim = 5 if aggre_type != 'gcn' else 10
        sage = nn.SAGEConv((10, dst_dim), out_dim, aggre_type)
        feat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), dst_dim)))
        sage = sage.to(F.ctx())
>       h = sage(g, feat)

tests\pytorch\test_nn.py:621: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:261: in forward
    graph.update_all(msg_fn, self._lstm_reducer)
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:362: in message_passing
    msgdata = invoke_gsddmm(g, mfunc)
python\dgl\core.py:276: in invoke_gsddmm
    z = op(graph, x)
python\dgl\ops\sddmm.py:164: in copy_u
    return gsddmm(g, 'copy_lhs', x, None)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0198EA08>
op = 'copy_lhs'
lhs_data = tensor([[ 1.3395, -0.0973,  1.1509,  0.4748, -0.2030,  0.3357, -0.1795, -0.7663,
         -0.7051, -0.6345],
        [ 0.3183,  2.5592,  0.5120,  0.2009,  1.2238,  1.0704,  0.7836,  2.6911,
          1.4359, -0.3083]])
rhs_data = None, lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
____________________ test_sage_conv_bi[2-lstm-g1-idtype1] _____________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
aggre_type = 'lstm', out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_sage_conv_bi(idtype, g, aggre_type, out_dim):
        g = g.astype(idtype).to(F.ctx())
        dst_dim = 5 if aggre_type != 'gcn' else 10
        sage = nn.SAGEConv((10, dst_dim), out_dim, aggre_type)
        feat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), dst_dim)))
        sage = sage.to(F.ctx())
>       h = sage(g, feat)

tests\pytorch\test_nn.py:621: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:261: in forward
    graph.update_all(msg_fn, self._lstm_reducer)
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:362: in message_passing
    msgdata = invoke_gsddmm(g, mfunc)
python\dgl\core.py:276: in invoke_gsddmm
    z = op(graph, x)
python\dgl\ops\sddmm.py:164: in copy_u
    return gsddmm(g, 'copy_lhs', x, None)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF015C08C8>
op = 'copy_lhs'
lhs_data = tensor([[-0.4320,  1.2556,  0.8403,  0.0075, -0.4551, -1.0147, -1.4030, -0.1795,
         -0.6305,  1.2127],
        [ 1.3403, -0.9065, -0.0187, -1.0294, -1.0403,  1.7941,  0.2398,  1.3540,
          0.0667,  1.2675]])
rhs_data = None, lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
____________________ test_sage_conv_bi[2-lstm-g2-idtype0] _____________________

idtype = torch.int32, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
aggre_type = 'lstm', out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_sage_conv_bi(idtype, g, aggre_type, out_dim):
        g = g.astype(idtype).to(F.ctx())
        dst_dim = 5 if aggre_type != 'gcn' else 10
        sage = nn.SAGEConv((10, dst_dim), out_dim, aggre_type)
        feat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), dst_dim)))
        sage = sage.to(F.ctx())
>       h = sage(g, feat)

tests\pytorch\test_nn.py:621: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:261: in forward
    graph.update_all(msg_fn, self._lstm_reducer)
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:362: in message_passing
    msgdata = invoke_gsddmm(g, mfunc)
python\dgl\core.py:276: in invoke_gsddmm
    z = op(graph, x)
python\dgl\ops\sddmm.py:164: in copy_u
    return gsddmm(g, 'copy_lhs', x, None)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01709B88>
op = 'copy_lhs'
lhs_data = tensor([[ 0.4768,  0.8830, -0.5621,  1.6969,  0.0225, -1.3958, -1.3762, -0.7441,
         -0.7116,  0.4947],
        [... 0.2954],
        [ 0.2342,  0.3976, -0.3529,  0.1594, -0.2597, -0.3330, -0.6725, -0.9375,
          0.8374, -1.2583]])
rhs_data = None, lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
____________________ test_sage_conv_bi[2-lstm-g2-idtype1] _____________________

idtype = torch.int64, g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
aggre_type = 'lstm', out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite']))
    @pytest.mark.parametrize('aggre_type', ['mean', 'pool', 'gcn', 'lstm'])
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_sage_conv_bi(idtype, g, aggre_type, out_dim):
        g = g.astype(idtype).to(F.ctx())
        dst_dim = 5 if aggre_type != 'gcn' else 10
        sage = nn.SAGEConv((10, dst_dim), out_dim, aggre_type)
        feat = (F.randn((g.number_of_src_nodes(), 10)), F.randn((g.number_of_dst_nodes(), dst_dim)))
        sage = sage.to(F.ctx())
>       h = sage(g, feat)

tests\pytorch\test_nn.py:621: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:261: in forward
    graph.update_all(msg_fn, self._lstm_reducer)
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:362: in message_passing
    msgdata = invoke_gsddmm(g, mfunc)
python\dgl\core.py:276: in invoke_gsddmm
    z = op(graph, x)
python\dgl\ops\sddmm.py:164: in copy_u
    return gsddmm(g, 'copy_lhs', x, None)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14A7D488>
op = 'copy_lhs'
lhs_data = tensor([[ 1.0471e+00, -1.2653e+00, -5.2789e-01,  5.7426e-02, -1.0818e+00,
          4.4435e-01,  2.9689e-01, -7.2202e-...+00, -4.0678e-01, -1.2995e+00, -7.9521e-01,
          7.2025e-01, -1.5633e+00, -1.6645e+00, -1.9601e+00, -1.4754e+00]])
rhs_data = None, lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_________________________ test_sage_conv2[1-idtype0] __________________________

idtype = torch.int32, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_sage_conv2(idtype, out_dim):
        # TODO: add test for blocks
        # Test the case for graphs without edges
        g = dgl.heterograph({('_U', '_E', '_V'): ([], [])}, {'_U': 5, '_V': 3})
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        sage = nn.SAGEConv((3, 3), out_dim, 'gcn')
        feat = (F.randn((5, 3)), F.randn((3, 3)))
        sage = sage.to(ctx)
>       h = sage(g, (F.copy_to(feat[0], F.ctx()), F.copy_to(feat[1], F.ctx())))

tests\pytorch\test_nn.py:636: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:249: in forward
    graph.update_all(msg_fn, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14A1E248>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.7503],
        [-0.3877],
        [ 1.7361],
        [-0.7325],
        [-0.4650]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_sage_conv2[1-idtype1] __________________________

idtype = torch.int64, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_sage_conv2(idtype, out_dim):
        # TODO: add test for blocks
        # Test the case for graphs without edges
        g = dgl.heterograph({('_U', '_E', '_V'): ([], [])}, {'_U': 5, '_V': 3})
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        sage = nn.SAGEConv((3, 3), out_dim, 'gcn')
        feat = (F.randn((5, 3)), F.randn((3, 3)))
        sage = sage.to(ctx)
>       h = sage(g, (F.copy_to(feat[0], F.ctx()), F.copy_to(feat[1], F.ctx())))

tests\pytorch\test_nn.py:636: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:249: in forward
    graph.update_all(msg_fn, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF018B15C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.8477],
        [ 0.3394],
        [ 2.8783],
        [-1.5177],
        [-1.7766]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_sage_conv2[2-idtype0] __________________________

idtype = torch.int32, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_sage_conv2(idtype, out_dim):
        # TODO: add test for blocks
        # Test the case for graphs without edges
        g = dgl.heterograph({('_U', '_E', '_V'): ([], [])}, {'_U': 5, '_V': 3})
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        sage = nn.SAGEConv((3, 3), out_dim, 'gcn')
        feat = (F.randn((5, 3)), F.randn((3, 3)))
        sage = sage.to(ctx)
>       h = sage(g, (F.copy_to(feat[0], F.ctx()), F.copy_to(feat[1], F.ctx())))

tests\pytorch\test_nn.py:636: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:249: in forward
    graph.update_all(msg_fn, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0189DA08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.4078, -1.2594],
        [-0.0893, -0.0618],
        [ 0.4615, -0.6097],
        [ 0.9645,  0.6322],
        [ 0.0364, -1.0941]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_sage_conv2[2-idtype1] __________________________

idtype = torch.int64, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_sage_conv2(idtype, out_dim):
        # TODO: add test for blocks
        # Test the case for graphs without edges
        g = dgl.heterograph({('_U', '_E', '_V'): ([], [])}, {'_U': 5, '_V': 3})
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        sage = nn.SAGEConv((3, 3), out_dim, 'gcn')
        feat = (F.randn((5, 3)), F.randn((3, 3)))
        sage = sage.to(ctx)
>       h = sage(g, (F.copy_to(feat[0], F.ctx()), F.copy_to(feat[1], F.ctx())))

tests\pytorch\test_nn.py:636: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:249: in forward
    graph.update_all(msg_fn, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0712E148>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 4.0099, -4.2069],
        [ 3.6503, -0.0156],
        [-0.8242, -0.3775],
        [-0.6331, -3.0260],
        [-0.0086, -1.8557]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_sgc_conv[1-g0-idtype0] _________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
idtype = torch.int32, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_sgc_conv(g, idtype, out_dim):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        # not cached
        sgc = nn.SGConv(5, out_dim, 3)
    
        # test pickle
        th.save(sgc, tmp_buffer)
    
        feat = F.randn((g.number_of_nodes(), 5))
        sgc = sgc.to(ctx)
    
>       h = sgc(g, feat)

tests\pytorch\test_nn.py:662: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sgconv.py:203: in forward
    fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0712C688>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.1153e+00,  6.2224e-01,  7.5383e-01, -7.6151e-01, -1.6290e+00],
        [ 2.9963e-01, -9.1867e-01, -2.6921e...00,  1.2709e-01,  3.8796e-01,  2.5487e-01],
        [-1.1475e+00, -1.1530e+00, -8.7164e-01, -1.9866e-01,  1.6638e-01]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_sgc_conv[1-g0-idtype1] _________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
idtype = torch.int64, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_sgc_conv(g, idtype, out_dim):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        # not cached
        sgc = nn.SGConv(5, out_dim, 3)
    
        # test pickle
        th.save(sgc, tmp_buffer)
    
        feat = F.randn((g.number_of_nodes(), 5))
        sgc = sgc.to(ctx)
    
>       h = sgc(g, feat)

tests\pytorch\test_nn.py:662: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sgconv.py:203: in forward
    fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF017B1D48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.0928, -0.3243, -1.0257, -0.0284,  0.2070],
        [ 1.2326, -0.3536, -0.2209, -0.4852,  0.1602],
        ...0.2397],
        [-0.8215, -1.0241,  0.1723, -0.0351,  0.6604],
        [-0.7579,  0.2230, -0.0809, -0.6123,  0.8807]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_sgc_conv[1-g1-idtype0] _________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_sgc_conv(g, idtype, out_dim):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        # not cached
        sgc = nn.SGConv(5, out_dim, 3)
    
        # test pickle
        th.save(sgc, tmp_buffer)
    
        feat = F.randn((g.number_of_nodes(), 5))
        sgc = sgc.to(ctx)
    
>       h = sgc(g, feat)

tests\pytorch\test_nn.py:662: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sgconv.py:203: in forward
    fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01860748>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.2158,  0.4157, -0.2189, -0.5853, -1.0790],
        [-0.5006,  0.2078, -0.0058,  0.0240,  1.2029],
        ...0.2636],
        [-1.0410, -0.9868, -0.9353,  0.5991,  0.6517],
        [-0.7061,  0.8314, -0.0315,  0.7336,  0.0050]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_sgc_conv[1-g1-idtype1] _________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_sgc_conv(g, idtype, out_dim):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        # not cached
        sgc = nn.SGConv(5, out_dim, 3)
    
        # test pickle
        th.save(sgc, tmp_buffer)
    
        feat = F.randn((g.number_of_nodes(), 5))
        sgc = sgc.to(ctx)
    
>       h = sgc(g, feat)

tests\pytorch\test_nn.py:662: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sgconv.py:203: in forward
    fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01650088>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.7625, -1.1514, -0.9397,  0.2973, -0.9160],
        [-0.2303, -0.2627,  0.3862,  0.4743,  0.5665],
        ...0.9913],
        [-0.1590,  0.6508,  0.4640,  0.3804, -0.4416],
        [ 1.1729,  1.4288,  0.4643,  0.3029,  0.1413]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_sgc_conv[1-g2-idtype0] _________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int32, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_sgc_conv(g, idtype, out_dim):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        # not cached
        sgc = nn.SGConv(5, out_dim, 3)
    
        # test pickle
        th.save(sgc, tmp_buffer)
    
        feat = F.randn((g.number_of_nodes(), 5))
        sgc = sgc.to(ctx)
    
>       h = sgc(g, feat)

tests\pytorch\test_nn.py:662: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sgconv.py:203: in forward
    fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1228ACC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.1421, -0.0257, -0.3185, -0.6716, -0.2598],
        [ 0.6564, -0.3693, -0.7975,  0.0162,  0.1493],
        ...0.2139],
        [-0.0872, -0.6665,  0.5276, -1.5325,  0.7190],
        [ 0.0667, -1.0048, -0.2295, -0.6781, -1.7817]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_sgc_conv[1-g2-idtype1] _________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int64, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_sgc_conv(g, idtype, out_dim):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        # not cached
        sgc = nn.SGConv(5, out_dim, 3)
    
        # test pickle
        th.save(sgc, tmp_buffer)
    
        feat = F.randn((g.number_of_nodes(), 5))
        sgc = sgc.to(ctx)
    
>       h = sgc(g, feat)

tests\pytorch\test_nn.py:662: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sgconv.py:203: in forward
    fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01571988>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.1237,  1.1311,  0.7043,  1.1315,  0.3190],
        [ 0.0840, -0.8494,  0.6983, -0.8182, -0.4864],
        ...1.6545],
        [-1.3567,  0.4707,  0.8853, -1.4294,  0.1069],
        [-0.6326, -0.0563,  0.7955,  0.3189, -1.7650]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_sgc_conv[1-g3-idtype0] _________________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_sgc_conv(g, idtype, out_dim):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        # not cached
        sgc = nn.SGConv(5, out_dim, 3)
    
        # test pickle
        th.save(sgc, tmp_buffer)
    
        feat = F.randn((g.number_of_nodes(), 5))
        sgc = sgc.to(ctx)
    
>       h = sgc(g, feat)

tests\pytorch\test_nn.py:662: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sgconv.py:203: in forward
    fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01620208>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.5359, -0.7373, -0.1986,  0.3809, -1.0554],
        [-0.4912,  0.2131,  0.1595, -1.0143,  0.6709],
        ...0.6299],
        [-1.2511, -1.0104,  0.0441,  0.0149,  1.2652],
        [ 0.5878,  0.4335, -0.0434, -0.5360, -0.6394]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_sgc_conv[1-g3-idtype1] _________________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_sgc_conv(g, idtype, out_dim):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        # not cached
        sgc = nn.SGConv(5, out_dim, 3)
    
        # test pickle
        th.save(sgc, tmp_buffer)
    
        feat = F.randn((g.number_of_nodes(), 5))
        sgc = sgc.to(ctx)
    
>       h = sgc(g, feat)

tests\pytorch\test_nn.py:662: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sgconv.py:203: in forward
    fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07037A08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-6.9358e-01, -2.4879e-01, -2.4465e+00,  8.9066e-01,  4.7676e-01],
        [-1.7994e-01,  2.2614e-01, -7.1317e...02,  8.3151e-01, -4.9775e-01,  1.3894e-01],
        [ 1.6549e-02,  9.4562e-01, -1.2167e-01, -8.4982e-01,  1.4768e+00]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_sgc_conv[1-g4-idtype0] _________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_sgc_conv(g, idtype, out_dim):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        # not cached
        sgc = nn.SGConv(5, out_dim, 3)
    
        # test pickle
        th.save(sgc, tmp_buffer)
    
        feat = F.randn((g.number_of_nodes(), 5))
        sgc = sgc.to(ctx)
    
>       h = sgc(g, feat)

tests\pytorch\test_nn.py:662: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sgconv.py:203: in forward
    fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0720FC08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.4010,  0.1829,  0.1227, -0.0140, -0.6890],
        [ 0.3091, -0.2111,  0.0620, -0.2190,  0.8252],
        ...1.1525],
        [-1.5799,  0.3866,  0.0354, -0.9569, -0.4465],
        [-0.2135, -1.0207,  0.3849, -0.7626, -0.2160]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_sgc_conv[1-g4-idtype1] _________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_sgc_conv(g, idtype, out_dim):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        # not cached
        sgc = nn.SGConv(5, out_dim, 3)
    
        # test pickle
        th.save(sgc, tmp_buffer)
    
        feat = F.randn((g.number_of_nodes(), 5))
        sgc = sgc.to(ctx)
    
>       h = sgc(g, feat)

tests\pytorch\test_nn.py:662: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sgconv.py:203: in forward
    fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07113848>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.4565, -0.1102,  0.7633, -0.4977, -0.7755],
        [-0.6357,  0.5309,  1.2306,  0.2407,  0.2489],
        ...1.3964],
        [-0.2957,  0.8520, -1.6005,  0.1575,  0.1486],
        [ 2.0756,  0.0430, -0.1943, -3.5298, -0.8414]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_sgc_conv[1-g5-idtype0] _________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_sgc_conv(g, idtype, out_dim):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        # not cached
        sgc = nn.SGConv(5, out_dim, 3)
    
        # test pickle
        th.save(sgc, tmp_buffer)
    
        feat = F.randn((g.number_of_nodes(), 5))
        sgc = sgc.to(ctx)
    
>       h = sgc(g, feat)

tests\pytorch\test_nn.py:662: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sgconv.py:203: in forward
    fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF018E8448>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.9835,  1.1724,  0.4163,  0.7120, -0.6683],
        [ 0.7105,  0.3790, -0.0564, -1.0632,  0.6019],
        ...0.2274],
        [-1.0582, -0.6426, -1.3442,  1.1950,  0.0438],
        [ 1.1084, -0.7226, -2.8490,  0.4675,  0.7968]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_sgc_conv[1-g5-idtype1] _________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_sgc_conv(g, idtype, out_dim):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        # not cached
        sgc = nn.SGConv(5, out_dim, 3)
    
        # test pickle
        th.save(sgc, tmp_buffer)
    
        feat = F.randn((g.number_of_nodes(), 5))
        sgc = sgc.to(ctx)
    
>       h = sgc(g, feat)

tests\pytorch\test_nn.py:662: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sgconv.py:203: in forward
    fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0181DC88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.2725, -0.6251, -0.7310,  0.1805, -0.7205],
        [ 0.5960,  0.1132,  0.3324, -0.2837,  0.2707],
        ...0.0091],
        [-0.4190,  0.7394, -0.0046, -1.3882, -1.1518],
        [ 0.5905, -0.0313,  0.5909,  0.7899,  0.7637]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_sgc_conv[2-g0-idtype0] _________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
idtype = torch.int32, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_sgc_conv(g, idtype, out_dim):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        # not cached
        sgc = nn.SGConv(5, out_dim, 3)
    
        # test pickle
        th.save(sgc, tmp_buffer)
    
        feat = F.randn((g.number_of_nodes(), 5))
        sgc = sgc.to(ctx)
    
>       h = sgc(g, feat)

tests\pytorch\test_nn.py:662: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sgconv.py:203: in forward
    fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF019409C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.4751,  0.5176,  1.0550,  0.7969, -0.4435],
        [ 0.0225,  0.5541,  0.1618,  0.5451,  0.2055],
        ...0.8868],
        [ 0.6143,  1.6895, -0.1782, -0.1688, -0.1521],
        [-0.2698,  1.2282, -0.2693, -1.5887, -0.0491]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_sgc_conv[2-g0-idtype1] _________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
idtype = torch.int64, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_sgc_conv(g, idtype, out_dim):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        # not cached
        sgc = nn.SGConv(5, out_dim, 3)
    
        # test pickle
        th.save(sgc, tmp_buffer)
    
        feat = F.randn((g.number_of_nodes(), 5))
        sgc = sgc.to(ctx)
    
>       h = sgc(g, feat)

tests\pytorch\test_nn.py:662: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sgconv.py:203: in forward
    fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF011D0588>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.0581, -0.8305, -1.3066,  0.5630, -0.4314],
        [-0.0966,  0.1328, -0.1938,  0.4745,  0.1688],
        ...0.6270],
        [-1.2290, -0.2114, -0.4526,  1.0747, -1.3044],
        [ 2.3246,  0.0307,  1.4267, -1.9825,  1.0128]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_sgc_conv[2-g1-idtype0] _________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_sgc_conv(g, idtype, out_dim):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        # not cached
        sgc = nn.SGConv(5, out_dim, 3)
    
        # test pickle
        th.save(sgc, tmp_buffer)
    
        feat = F.randn((g.number_of_nodes(), 5))
        sgc = sgc.to(ctx)
    
>       h = sgc(g, feat)

tests\pytorch\test_nn.py:662: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sgconv.py:203: in forward
    fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0107EA88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.9076,  0.2035,  0.6050,  1.0509, -0.2774],
        [-0.1183,  0.5136,  1.0129, -0.0960,  0.4883],
        ...0.7544],
        [-0.7401, -0.3792, -0.4190, -1.4912,  0.8041],
        [ 0.6352,  1.4845, -0.1968,  0.3076, -0.8869]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_sgc_conv[2-g1-idtype1] _________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_sgc_conv(g, idtype, out_dim):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        # not cached
        sgc = nn.SGConv(5, out_dim, 3)
    
        # test pickle
        th.save(sgc, tmp_buffer)
    
        feat = F.randn((g.number_of_nodes(), 5))
        sgc = sgc.to(ctx)
    
>       h = sgc(g, feat)

tests\pytorch\test_nn.py:662: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sgconv.py:203: in forward
    fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01472308>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.9261, -0.8832, -0.6830, -0.6839,  0.1409],
        [ 0.4890, -0.1235, -0.2927, -0.0656, -1.3658],
        ...0.5201],
        [ 1.8555,  3.2268, -1.0963,  0.0656, -1.2751],
        [-0.3706, -0.4129,  0.1674, -0.4970,  1.3016]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_sgc_conv[2-g2-idtype0] _________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int32, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_sgc_conv(g, idtype, out_dim):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        # not cached
        sgc = nn.SGConv(5, out_dim, 3)
    
        # test pickle
        th.save(sgc, tmp_buffer)
    
        feat = F.randn((g.number_of_nodes(), 5))
        sgc = sgc.to(ctx)
    
>       h = sgc(g, feat)

tests\pytorch\test_nn.py:662: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sgconv.py:203: in forward
    fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1463CD08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.3531,  0.0983, -0.2150,  0.2558, -0.0818],
        [-0.1776,  0.1130, -0.1080, -0.0919, -0.0484],
        ...1.6309],
        [-1.1397,  1.6878,  0.4944, -0.4484,  0.8537],
        [-0.1935,  0.1629, -2.5101,  1.1200,  1.3222]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_sgc_conv[2-g2-idtype1] _________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int64, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_sgc_conv(g, idtype, out_dim):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        # not cached
        sgc = nn.SGConv(5, out_dim, 3)
    
        # test pickle
        th.save(sgc, tmp_buffer)
    
        feat = F.randn((g.number_of_nodes(), 5))
        sgc = sgc.to(ctx)
    
>       h = sgc(g, feat)

tests\pytorch\test_nn.py:662: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sgconv.py:203: in forward
    fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1228ECC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.3318, -0.9306, -1.0793, -0.3981, -0.8571],
        [-0.2308, -0.5392, -0.7182,  0.3514, -0.4415],
        ...0.6080],
        [ 0.0759, -0.1025,  0.1057,  0.2420, -0.1594],
        [-1.2331, -0.4894, -0.0777,  0.9134,  0.0118]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_sgc_conv[2-g3-idtype0] _________________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_sgc_conv(g, idtype, out_dim):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        # not cached
        sgc = nn.SGConv(5, out_dim, 3)
    
        # test pickle
        th.save(sgc, tmp_buffer)
    
        feat = F.randn((g.number_of_nodes(), 5))
        sgc = sgc.to(ctx)
    
>       h = sgc(g, feat)

tests\pytorch\test_nn.py:662: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sgconv.py:203: in forward
    fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03CC6CC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.3044e+00, -1.8580e+00, -3.3866e-01, -7.0749e-01, -1.9628e+00],
        [ 1.4719e+00,  9.3374e-01, -1.2402e...00,  5.2632e-01, -1.0365e+00,  5.8922e-01],
        [-6.0602e-01,  2.2662e-01, -5.4622e-01,  2.2425e-01, -5.6242e-01]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_sgc_conv[2-g3-idtype1] _________________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_sgc_conv(g, idtype, out_dim):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        # not cached
        sgc = nn.SGConv(5, out_dim, 3)
    
        # test pickle
        th.save(sgc, tmp_buffer)
    
        feat = F.randn((g.number_of_nodes(), 5))
        sgc = sgc.to(ctx)
    
>       h = sgc(g, feat)

tests\pytorch\test_nn.py:662: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sgconv.py:203: in forward
    fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE12242B08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.5670,  1.3418,  0.0205,  0.0554,  1.2156],
        [-0.5563,  0.0522, -1.0762,  0.0051, -0.2892],
        ...1.1610],
        [ 0.1171,  0.5560,  0.6516,  0.3490, -1.0903],
        [-0.5307, -1.2562,  0.6655, -0.0182,  0.9046]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_sgc_conv[2-g4-idtype0] _________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_sgc_conv(g, idtype, out_dim):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        # not cached
        sgc = nn.SGConv(5, out_dim, 3)
    
        # test pickle
        th.save(sgc, tmp_buffer)
    
        feat = F.randn((g.number_of_nodes(), 5))
        sgc = sgc.to(ctx)
    
>       h = sgc(g, feat)

tests\pytorch\test_nn.py:662: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sgconv.py:203: in forward
    fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF018A50C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.8036, -0.6416,  0.9105,  0.2474,  0.8440],
        [ 0.5363, -0.1496, -0.5685,  0.2297, -0.2812],
        ...0.5207],
        [-0.5266,  0.2349,  1.3866,  0.1286, -0.1605],
        [-1.4979, -0.2791, -1.1554,  0.5632,  0.6840]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_sgc_conv[2-g4-idtype1] _________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_sgc_conv(g, idtype, out_dim):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        # not cached
        sgc = nn.SGConv(5, out_dim, 3)
    
        # test pickle
        th.save(sgc, tmp_buffer)
    
        feat = F.randn((g.number_of_nodes(), 5))
        sgc = sgc.to(ctx)
    
>       h = sgc(g, feat)

tests\pytorch\test_nn.py:662: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sgconv.py:203: in forward
    fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF070A6AC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.4208, -0.6790, -0.1098,  0.6655,  0.1232],
        [-0.3847, -0.6806,  0.1624, -0.9245, -0.5730],
        ...0.0143],
        [-1.4645, -0.8081, -2.3578, -1.1136,  0.6103],
        [-1.8674,  1.3260,  0.1762,  0.1219, -0.6558]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_sgc_conv[2-g5-idtype0] _________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_sgc_conv(g, idtype, out_dim):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        # not cached
        sgc = nn.SGConv(5, out_dim, 3)
    
        # test pickle
        th.save(sgc, tmp_buffer)
    
        feat = F.randn((g.number_of_nodes(), 5))
        sgc = sgc.to(ctx)
    
>       h = sgc(g, feat)

tests\pytorch\test_nn.py:662: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sgconv.py:203: in forward
    fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF016F59C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.2771,  0.8985,  1.1932, -0.1613, -0.3753],
        [-1.0308, -1.7162, -0.6974, -0.0261,  0.2392],
        ...0.6472],
        [ 0.1996,  0.3721,  0.4608, -1.5719,  2.0905],
        [ 0.5852, -1.3563,  0.8762,  1.9243, -0.0025]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_sgc_conv[2-g5-idtype1] _________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_sgc_conv(g, idtype, out_dim):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        # not cached
        sgc = nn.SGConv(5, out_dim, 3)
    
        # test pickle
        th.save(sgc, tmp_buffer)
    
        feat = F.randn((g.number_of_nodes(), 5))
        sgc = sgc.to(ctx)
    
>       h = sgc(g, feat)

tests\pytorch\test_nn.py:662: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sgconv.py:203: in forward
    fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF017106C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.2411, -0.4401, -0.0199,  0.3107,  0.1241],
        [ 0.3906,  0.5180,  0.1621, -1.2393, -0.5666],
        ...0.4666],
        [-0.6296,  1.0790, -1.0514, -0.7966, -0.8324],
        [-0.4257, -0.0182, -1.7608,  1.9983,  0.4763]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_appnp_conv[g0-idtype0] _________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float32)})
idtype = torch.int32

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_appnp_conv(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        appnp = nn.APPNPConv(10, 0.1)
        feat = F.randn((g.number_of_nodes(), 5))
        appnp = appnp.to(ctx)
    
        # test pickle
        th.save(appnp, tmp_buffer)
    
>       h = appnp(g, feat)

tests\pytorch\test_nn.py:685: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\appnpconv.py:117: in forward
    fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF071E90C8>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[ 0.3022,  0.6082,  0.5632, -0.6902,  0.7791],
        [-0.4253,  0.2931, -0.1326, -0.7940, -0.2531],
        ...0.3433],
        [ 0.8229,  0.6415, -0.6227, -1.2144,  0.1676],
        [-0.6429, -0.3646,  0.4035,  0.0946, -0.7868]])
rhs_data = tensor([[1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_appnp_conv[g0-idtype1] _________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float32)})
idtype = torch.int64

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_appnp_conv(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        appnp = nn.APPNPConv(10, 0.1)
        feat = F.randn((g.number_of_nodes(), 5))
        appnp = appnp.to(ctx)
    
        # test pickle
        th.save(appnp, tmp_buffer)
    
>       h = appnp(g, feat)

tests\pytorch\test_nn.py:685: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\appnpconv.py:117: in forward
    fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0718A5C8>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[-1.3359, -0.1955, -0.5525,  0.7520,  0.0383],
        [-0.3386, -0.2548, -1.0032,  1.5583,  0.3742],
        ...0.0520],
        [-0.1452, -0.5809,  0.3677, -1.6490,  0.6598],
        [-1.0950,  0.5411,  1.0554, -0.4085, -0.1354]])
rhs_data = tensor([[1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_appnp_conv[g1-idtype0] _________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float32)})
idtype = torch.int32

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_appnp_conv(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        appnp = nn.APPNPConv(10, 0.1)
        feat = F.randn((g.number_of_nodes(), 5))
        appnp = appnp.to(ctx)
    
        # test pickle
        th.save(appnp, tmp_buffer)
    
>       h = appnp(g, feat)

tests\pytorch\test_nn.py:685: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\appnpconv.py:117: in forward
    fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01570A48>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[ 0.2455,  0.3232,  0.3403,  0.5423, -0.2726],
        [-0.0069,  0.4237, -0.2699, -1.1848, -0.5432],
        ...0.4339],
        [-0.3943, -2.1058,  0.8110,  0.6022, -1.0830],
        [ 1.3945, -0.4357, -0.1637, -1.0308,  0.4060]])
rhs_data = tensor([[1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_appnp_conv[g1-idtype1] _________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float32)})
idtype = torch.int64

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_appnp_conv(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        appnp = nn.APPNPConv(10, 0.1)
        feat = F.randn((g.number_of_nodes(), 5))
        appnp = appnp.to(ctx)
    
        # test pickle
        th.save(appnp, tmp_buffer)
    
>       h = appnp(g, feat)

tests\pytorch\test_nn.py:685: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\appnpconv.py:117: in forward
    fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14923B08>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[ 0.7470,  0.5226, -0.2514,  0.3301,  0.3837],
        [-0.5294,  0.9001,  1.2214,  0.2357, -0.6622],
        ...1.1281],
        [ 0.6578, -0.2730,  0.4807, -2.0040,  1.6251],
        [ 1.9827, -0.4308, -0.4573, -0.6509,  0.2080]])
rhs_data = tensor([[1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_appnp_conv[g2-idtype0] _________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), 'w': Scheme(shape=(1,), dtype=torch.float32)})
idtype = torch.int32

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_appnp_conv(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        appnp = nn.APPNPConv(10, 0.1)
        feat = F.randn((g.number_of_nodes(), 5))
        appnp = appnp.to(ctx)
    
        # test pickle
        th.save(appnp, tmp_buffer)
    
>       h = appnp(g, feat)

tests\pytorch\test_nn.py:685: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\appnpconv.py:117: in forward
    fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF016F5B88>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[-0.7430, -0.0635, -0.1236, -0.9118, -0.1409],
        [-0.2851, -0.4702, -0.3732,  0.3772,  0.1108],
        ...1.2701],
        [-0.7047,  0.6995,  0.7522,  0.9739, -0.0363],
        [ 0.6585,  0.2290, -1.5503, -0.5822,  0.0714]])
rhs_data = tensor([[1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_appnp_conv[g2-idtype1] _________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), 'w': Scheme(shape=(1,), dtype=torch.float32)})
idtype = torch.int64

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_appnp_conv(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        appnp = nn.APPNPConv(10, 0.1)
        feat = F.randn((g.number_of_nodes(), 5))
        appnp = appnp.to(ctx)
    
        # test pickle
        th.save(appnp, tmp_buffer)
    
>       h = appnp(g, feat)

tests\pytorch\test_nn.py:685: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\appnpconv.py:117: in forward
    fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01100E48>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[ 0.5062,  0.0161, -0.4136,  0.6292, -0.1412],
        [-0.2489,  0.4808, -0.6800,  0.6940, -0.0739],
        ...1.3135],
        [-0.4537, -0.9981, -0.2241,  0.9888,  1.2246],
        [-0.2838, -0.4010,  2.1741, -0.4732, -0.3427]])
rhs_data = tensor([[1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_appnp_conv[g3-idtype0] _________________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float32)})
idtype = torch.int32

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_appnp_conv(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        appnp = nn.APPNPConv(10, 0.1)
        feat = F.randn((g.number_of_nodes(), 5))
        appnp = appnp.to(ctx)
    
        # test pickle
        th.save(appnp, tmp_buffer)
    
>       h = appnp(g, feat)

tests\pytorch\test_nn.py:685: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\appnpconv.py:117: in forward
    fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE12365988>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[-1.0299,  0.2926, -0.0097, -0.4929,  0.4381],
        [-0.1529,  0.2010,  0.7588, -1.0679,  0.1471],
        ...0.0425],
        [-0.7706, -1.2581,  0.5618, -1.7124,  0.4557],
        [-0.6841,  1.0255,  0.0610, -2.3333, -0.1727]])
rhs_data = tensor([[1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_appnp_conv[g3-idtype1] _________________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float32)})
idtype = torch.int64

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_appnp_conv(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        appnp = nn.APPNPConv(10, 0.1)
        feat = F.randn((g.number_of_nodes(), 5))
        appnp = appnp.to(ctx)
    
        # test pickle
        th.save(appnp, tmp_buffer)
    
>       h = appnp(g, feat)

tests\pytorch\test_nn.py:685: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\appnpconv.py:117: in forward
    fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE12351588>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[ 0.7933,  0.4328, -0.3607,  0.4800,  1.2979],
        [-0.7450, -0.7528, -0.6951, -0.0517, -1.4109],
        ...0.5642],
        [ 0.0882,  0.4170,  0.4050, -0.8288, -1.3514],
        [ 0.5588,  0.0540, -0.7828, -0.5787,  0.8467]])
rhs_data = tensor([[1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_appnp_conv[g4-idtype0] _________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float32)})
idtype = torch.int32

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_appnp_conv(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        appnp = nn.APPNPConv(10, 0.1)
        feat = F.randn((g.number_of_nodes(), 5))
        appnp = appnp.to(ctx)
    
        # test pickle
        th.save(appnp, tmp_buffer)
    
>       h = appnp(g, feat)

tests\pytorch\test_nn.py:685: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\appnpconv.py:117: in forward
    fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01955C48>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[-4.9494e-01,  1.6106e-01,  6.0797e-01,  9.2947e-01,  2.6904e-02],
        [-6.8278e-01,  1.8360e-01, -3.9227e...01,  1.6273e+00, -7.2576e-01,  9.2706e-02],
        [-5.6638e-01,  1.3135e+00, -2.0288e-02, -4.4796e-01,  2.8753e+00]])
rhs_data = tensor([[1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_appnp_conv[g4-idtype1] _________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float32)})
idtype = torch.int64

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_appnp_conv(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        appnp = nn.APPNPConv(10, 0.1)
        feat = F.randn((g.number_of_nodes(), 5))
        appnp = appnp.to(ctx)
    
        # test pickle
        th.save(appnp, tmp_buffer)
    
>       h = appnp(g, feat)

tests\pytorch\test_nn.py:685: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\appnpconv.py:117: in forward
    fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0198F9C8>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[ 0.3172,  0.3583, -0.8368, -0.2492, -0.0303],
        [ 0.1314, -0.7922, -0.7403,  0.7373,  0.0136],
        ...0.7224],
        [-1.4246,  0.6161, -0.0501,  1.2807,  0.5537],
        [ 1.3446, -1.5174, -1.2142,  0.9258,  2.1938]])
rhs_data = tensor([[1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_appnp_conv[g5-idtype0] _________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float32)})
idtype = torch.int32

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_appnp_conv(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        appnp = nn.APPNPConv(10, 0.1)
        feat = F.randn((g.number_of_nodes(), 5))
        appnp = appnp.to(ctx)
    
        # test pickle
        th.save(appnp, tmp_buffer)
    
>       h = appnp(g, feat)

tests\pytorch\test_nn.py:685: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\appnpconv.py:117: in forward
    fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07246AC8>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[-0.4701, -0.2182, -0.0842,  0.1956,  0.6631],
        [-0.7034,  0.7063,  0.8337, -0.8359,  0.6053],
        ...0.0968],
        [-0.2572,  0.8915,  0.8062, -1.6188,  1.6586],
        [-2.3579,  0.6590,  0.2064,  2.4589, -0.6835]])
rhs_data = tensor([[1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_appnp_conv[g5-idtype1] _________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float32)})
idtype = torch.int64

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_appnp_conv(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        appnp = nn.APPNPConv(10, 0.1)
        feat = F.randn((g.number_of_nodes(), 5))
        appnp = appnp.to(ctx)
    
        # test pickle
        th.save(appnp, tmp_buffer)
    
>       h = appnp(g, feat)

tests\pytorch\test_nn.py:685: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\appnpconv.py:117: in forward
    fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07218248>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[ 0.1347,  0.5045,  0.3906,  1.3026, -0.3070],
        [ 1.1250,  1.4391, -0.2138,  0.4424,  0.1221],
        ...1.1022],
        [ 0.5674, -0.8320,  1.0354,  1.3648, -1.5096],
        [ 0.9106, -0.6724, -0.5468,  0.7959,  0.1443]])
rhs_data = tensor([[1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
____________________ test_appnp_conv_e_weight[g0-idtype0] _____________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), '_src_out_w': Sch...}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32), '_edge_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int32

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_appnp_conv_e_weight(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        appnp = nn.APPNPConv(10, 0.1)
        feat = F.randn((g.number_of_nodes(), 5))
        eweight = F.ones((g.num_edges(), ))
        appnp = appnp.to(ctx)
    
>       h = appnp(g, feat, edge_weight=eweight)

tests\pytorch\test_nn.py:699: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\appnpconv.py:106: in forward
    'both')(graph, edge_weight)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:117: in forward
    reversed_g.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'out_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01568F88>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1.])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
____________________ test_appnp_conv_e_weight[g0-idtype1] _____________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), '_src_out_w': Sch...}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32), '_edge_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int64

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_appnp_conv_e_weight(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        appnp = nn.APPNPConv(10, 0.1)
        feat = F.randn((g.number_of_nodes(), 5))
        eweight = F.ones((g.num_edges(), ))
        appnp = appnp.to(ctx)
    
>       h = appnp(g, feat, edge_weight=eweight)

tests\pytorch\test_nn.py:699: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\appnpconv.py:106: in forward
    'both')(graph, edge_weight)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:117: in forward
    reversed_g.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'out_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF016502C8>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1.])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
____________________ test_appnp_conv_e_weight[g1-idtype0] _____________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'_src_out_w': Scheme(shape=(), dtype=torch.float32), '_dst_in_w': Scheme(shape=(), dtype=torch.float32)}
      edata_schemes={'_edge_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int32

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_appnp_conv_e_weight(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        appnp = nn.APPNPConv(10, 0.1)
        feat = F.randn((g.number_of_nodes(), 5))
        eweight = F.ones((g.num_edges(), ))
        appnp = appnp.to(ctx)
    
>       h = appnp(g, feat, edge_weight=eweight)

tests\pytorch\test_nn.py:699: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\appnpconv.py:106: in forward
    'both')(graph, edge_weight)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:117: in forward
    reversed_g.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'out_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0195C5C8>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1.])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
____________________ test_appnp_conv_e_weight[g1-idtype1] _____________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'_src_out_w': Scheme(shape=(), dtype=torch.float32), '_dst_in_w': Scheme(shape=(), dtype=torch.float32)}
      edata_schemes={'_edge_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int64

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_appnp_conv_e_weight(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        appnp = nn.APPNPConv(10, 0.1)
        feat = F.randn((g.number_of_nodes(), 5))
        eweight = F.ones((g.num_edges(), ))
        appnp = appnp.to(ctx)
    
>       h = appnp(g, feat, edge_weight=eweight)

tests\pytorch\test_nn.py:699: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\appnpconv.py:106: in forward
    'both')(graph, edge_weight)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:117: in forward
    reversed_g.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'out_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF011E01C8>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1.])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
____________________ test_appnp_conv_e_weight[g2-idtype0] _____________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), '_src_out_w': Sch...   edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), '_edge_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int32

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_appnp_conv_e_weight(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        appnp = nn.APPNPConv(10, 0.1)
        feat = F.randn((g.number_of_nodes(), 5))
        eweight = F.ones((g.num_edges(), ))
        appnp = appnp.to(ctx)
    
>       h = appnp(g, feat, edge_weight=eweight)

tests\pytorch\test_nn.py:699: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\appnpconv.py:106: in forward
    'both')(graph, edge_weight)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:117: in forward
    reversed_g.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'out_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF016FA5C8>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1.])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
____________________ test_appnp_conv_e_weight[g2-idtype1] _____________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), '_src_out_w': Sch...   edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), '_edge_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int64

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_appnp_conv_e_weight(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        appnp = nn.APPNPConv(10, 0.1)
        feat = F.randn((g.number_of_nodes(), 5))
        eweight = F.ones((g.num_edges(), ))
        appnp = appnp.to(ctx)
    
>       h = appnp(g, feat, edge_weight=eweight)

tests\pytorch\test_nn.py:699: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\appnpconv.py:106: in forward
    'both')(graph, edge_weight)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:117: in forward
    reversed_g.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'out_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF017A9588>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1.])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
____________________ test_appnp_conv_e_weight[g3-idtype0] _____________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'_src_out_w': Scheme(shape=(), dtype=torch.float32), '_dst_in_w': Scheme(shape=(), dtype=torch.float32)}
      edata_schemes={'_edge_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int32

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_appnp_conv_e_weight(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)

        feat = F.randn((g.number_of_nodes(), 5))
        eweight = F.ones((g.num_edges(), ))
        appnp = appnp.to(ctx)
    
>       h = appnp(g, feat, edge_weight=eweight)

tests\pytorch\test_nn.py:699: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\appnpconv.py:106: in forward
    'both')(graph, edge_weight)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:117: in forward
    reversed_g.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'out_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07188208>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1.])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
____________________ test_appnp_conv_e_weight[g3-idtype1] _____________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'_src_out_w': Scheme(shape=(), dtype=torch.float32), '_dst_in_w': Scheme(shape=(), dtype=torch.float32)}
      edata_schemes={'_edge_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int64

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_appnp_conv_e_weight(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        appnp = nn.APPNPConv(10, 0.1)
        feat = F.randn((g.number_of_nodes(), 5))
        eweight = F.ones((g.num_edges(), ))
        appnp = appnp.to(ctx)
    
>       h = appnp(g, feat, edge_weight=eweight)

tests\pytorch\test_nn.py:699: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\appnpconv.py:106: in forward
    'both')(graph, edge_weight)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:117: in forward
    reversed_g.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'out_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE149CFA08>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1.])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
____________________ test_appnp_conv_e_weight[g4-idtype0] _____________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'_src_out_w': Scheme(shape=(), dtype=torch.float32), '_dst_in_w': Scheme(shape=(), dtype=torch.float32)}
      edata_schemes={'_edge_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int32

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_appnp_conv_e_weight(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        appnp = nn.APPNPConv(10, 0.1)
        feat = F.randn((g.number_of_nodes(), 5))
        eweight = F.ones((g.num_edges(), ))
        appnp = appnp.to(ctx)
    
>       h = appnp(g, feat, edge_weight=eweight)

tests\pytorch\test_nn.py:699: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\appnpconv.py:106: in forward
    'both')(graph, edge_weight)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:117: in forward
    reversed_g.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'out_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01860988>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1.])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
____________________ test_appnp_conv_e_weight[g4-idtype1] _____________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'_src_out_w': Scheme(shape=(), dtype=torch.float32), '_dst_in_w': Scheme(shape=(), dtype=torch.float32)}
      edata_schemes={'_edge_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int64

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_appnp_conv_e_weight(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        appnp = nn.APPNPConv(10, 0.1)
        feat = F.randn((g.number_of_nodes(), 5))
        eweight = F.ones((g.num_edges(), ))
        appnp = appnp.to(ctx)
    
>       h = appnp(g, feat, edge_weight=eweight)

tests\pytorch\test_nn.py:699: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\appnpconv.py:106: in forward
    'both')(graph, edge_weight)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:117: in forward
    reversed_g.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'out_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03D150C8>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1.])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
____________________ test_appnp_conv_e_weight[g5-idtype0] _____________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'_src_out_w': Scheme(shape=(), dtype=torch.float32), '_dst_in_w': Scheme(shape=(), dtype=torch.float32)}
      edata_schemes={'_edge_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int32

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_appnp_conv_e_weight(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        appnp = nn.APPNPConv(10, 0.1)
        feat = F.randn((g.number_of_nodes(), 5))
        eweight = F.ones((g.num_edges(), ))
        appnp = appnp.to(ctx)
    
>       h = appnp(g, feat, edge_weight=eweight)

tests\pytorch\test_nn.py:699: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\appnpconv.py:106: in forward
    'both')(graph, edge_weight)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:117: in forward
    reversed_g.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'out_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF071A2608>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1.])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
____________________ test_appnp_conv_e_weight[g5-idtype1] _____________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'_src_out_w': Scheme(shape=(), dtype=torch.float32), '_dst_in_w': Scheme(shape=(), dtype=torch.float32)}
      edata_schemes={'_edge_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int64

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_appnp_conv_e_weight(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        appnp = nn.APPNPConv(10, 0.1)
        feat = F.randn((g.number_of_nodes(), 5))
        eweight = F.ones((g.num_edges(), ))
        appnp = appnp.to(ctx)
    
>       h = appnp(g, feat, edge_weight=eweight)

tests\pytorch\test_nn.py:699: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\appnpconv.py:106: in forward
    'both')(graph, edge_weight)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:117: in forward
    reversed_g.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'out_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07182F88>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1.])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
___________________ test_gcn2conv_e_weight[True-g0-idtype0] ___________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), '_src_out_w': Sch...}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32), '_edge_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int32, bias = True

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize("bias", [True, False])
    def test_gcn2conv_e_weight(g, idtype, bias):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        gcn2conv = nn.GCN2Conv(5, layer=2, alpha=0.5, bias=bias,
                               project_initial_features=True)
        feat = F.randn((g.number_of_nodes(), 5))
        eweight = F.ones((g.num_edges(), ))
        gcn2conv = gcn2conv.to(ctx)
        res = feat
>       h = gcn2conv(g, res, feat, edge_weight=eweight)

tests\pytorch\test_nn.py:714: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gcn2conv.py:236: in forward
    edge_weight = EdgeWeightNorm('both')(graph, edge_weight)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:117: in forward
    reversed_g.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'out_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0718E948>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1.])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
___________________ test_gcn2conv_e_weight[True-g0-idtype1] ___________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), '_src_out_w': Sch...}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32), '_edge_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int64, bias = True

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize("bias", [True, False])
    def test_gcn2conv_e_weight(g, idtype, bias):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        gcn2conv = nn.GCN2Conv(5, layer=2, alpha=0.5, bias=bias,
                               project_initial_features=True)
        feat = F.randn((g.number_of_nodes(), 5))
        eweight = F.ones((g.num_edges(), ))
        gcn2conv = gcn2conv.to(ctx)
        res = feat
>       h = gcn2conv(g, res, feat, edge_weight=eweight)

tests\pytorch\test_nn.py:714: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gcn2conv.py:236: in forward
    edge_weight = EdgeWeightNorm('both')(graph, edge_weight)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:117: in forward
    reversed_g.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'out_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0195CB48>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1.])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
___________________ test_gcn2conv_e_weight[True-g1-idtype0] ___________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'_src_out_w': Scheme(shape=(), dtype=torch.float32), '_dst_in_w': Scheme(shape=(), dtype=torch.float32)}
      edata_schemes={'_edge_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int32, bias = True

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize("bias", [True, False])
    def test_gcn2conv_e_weight(g, idtype, bias):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        gcn2conv = nn.GCN2Conv(5, layer=2, alpha=0.5, bias=bias,
                               project_initial_features=True)
        feat = F.randn((g.number_of_nodes(), 5))
        eweight = F.ones((g.num_edges(), ))
        gcn2conv = gcn2conv.to(ctx)
        res = feat
>       h = gcn2conv(g, res, feat, edge_weight=eweight)

tests\pytorch\test_nn.py:714: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gcn2conv.py:236: in forward
    edge_weight = EdgeWeightNorm('both')(graph, edge_weight)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:117: in forward
    reversed_g.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'out_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07248648>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1.])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
___________________ test_gcn2conv_e_weight[True-g1-idtype1] ___________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'_src_out_w': Scheme(shape=(), dtype=torch.float32), '_dst_in_w': Scheme(shape=(), dtype=torch.float32)}
      edata_schemes={'_edge_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int64, bias = True

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize("bias", [True, False])
    def test_gcn2conv_e_weight(g, idtype, bias):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        gcn2conv = nn.GCN2Conv(5, layer=2, alpha=0.5, bias=bias,
                               project_initial_features=True)
        feat = F.randn((g.number_of_nodes(), 5))
        eweight = F.ones((g.num_edges(), ))
        gcn2conv = gcn2conv.to(ctx)
        res = feat
>       h = gcn2conv(g, res, feat, edge_weight=eweight)

tests\pytorch\test_nn.py:714: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gcn2conv.py:236: in forward
    edge_weight = EdgeWeightNorm('both')(graph, edge_weight)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:117: in forward
    reversed_g.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'out_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF018C0F48>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1.])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
___________________ test_gcn2conv_e_weight[True-g2-idtype0] ___________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), '_src_out_w': Sch...   edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), '_edge_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int32, bias = True

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize("bias", [True, False])
    def test_gcn2conv_e_weight(g, idtype, bias):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        gcn2conv = nn.GCN2Conv(5, layer=2, alpha=0.5, bias=bias,
                               project_initial_features=True)
        feat = F.randn((g.number_of_nodes(), 5))
        eweight = F.ones((g.num_edges(), ))
        gcn2conv = gcn2conv.to(ctx)
        res = feat
>       h = gcn2conv(g, res, feat, edge_weight=eweight)

tests\pytorch\test_nn.py:714: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gcn2conv.py:236: in forward
    edge_weight = EdgeWeightNorm('both')(graph, edge_weight)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:117: in forward
    reversed_g.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'out_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF071BEB08>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1.])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
___________________ test_gcn2conv_e_weight[True-g2-idtype1] ___________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), '_src_out_w': Sch...   edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), '_edge_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int64, bias = True

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize("bias", [True, False])
    def test_gcn2conv_e_weight(g, idtype, bias):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        gcn2conv = nn.GCN2Conv(5, layer=2, alpha=0.5, bias=bias,
                               project_initial_features=True)
        feat = F.randn((g.number_of_nodes(), 5))
        eweight = F.ones((g.num_edges(), ))
        gcn2conv = gcn2conv.to(ctx)
        res = feat
>       h = gcn2conv(g, res, feat, edge_weight=eweight)

tests\pytorch\test_nn.py:714: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gcn2conv.py:236: in forward
    edge_weight = EdgeWeightNorm('both')(graph, edge_weight)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:117: in forward
    reversed_g.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'out_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0185E4C8>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1.])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
___________________ test_gcn2conv_e_weight[True-g3-idtype0] ___________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'_src_out_w': Scheme(shape=(), dtype=torch.float32), '_dst_in_w': Scheme(shape=(), dtype=torch.float32)}
      edata_schemes={'_edge_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int32, bias = True

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize("bias", [True, False])
    def test_gcn2conv_e_weight(g, idtype, bias):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        gcn2conv = nn.GCN2Conv(5, layer=2, alpha=0.5, bias=bias,
                               project_initial_features=True)
        feat = F.randn((g.number_of_nodes(), 5))
        eweight = F.ones((g.num_edges(), ))
        gcn2conv = gcn2conv.to(ctx)
        res = feat
>       h = gcn2conv(g, res, feat, edge_weight=eweight)

tests\pytorch\test_nn.py:714: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gcn2conv.py:236: in forward
    edge_weight = EdgeWeightNorm('both')(graph, edge_weight)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:117: in forward
    reversed_g.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'out_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01738448>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1.])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
___________________ test_gcn2conv_e_weight[True-g3-idtype1] ___________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'_src_out_w': Scheme(shape=(), dtype=torch.float32), '_dst_in_w': Scheme(shape=(), dtype=torch.float32)}
      edata_schemes={'_edge_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int64, bias = True

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize("bias", [True, False])
    def test_gcn2conv_e_weight(g, idtype, bias):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        gcn2conv = nn.GCN2Conv(5, layer=2, alpha=0.5, bias=bias,
                               project_initial_features=True)
        feat = F.randn((g.number_of_nodes(), 5))
        eweight = F.ones((g.num_edges(), ))
        gcn2conv = gcn2conv.to(ctx)
        res = feat
>       h = gcn2conv(g, res, feat, edge_weight=eweight)

tests\pytorch\test_nn.py:714: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gcn2conv.py:236: in forward
    edge_weight = EdgeWeightNorm('both')(graph, edge_weight)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:117: in forward
    reversed_g.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'out_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03D1B108>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1.])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
___________________ test_gcn2conv_e_weight[True-g4-idtype0] ___________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'_src_out_w': Scheme(shape=(), dtype=torch.float32), '_dst_in_w': Scheme(shape=(), dtype=torch.float32)}
      edata_schemes={'_edge_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int32, bias = True

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize("bias", [True, False])
    def test_gcn2conv_e_weight(g, idtype, bias):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        gcn2conv = nn.GCN2Conv(5, layer=2, alpha=0.5, bias=bias,
                               project_initial_features=True)
        feat = F.randn((g.number_of_nodes(), 5))
        eweight = F.ones((g.num_edges(), ))
        gcn2conv = gcn2conv.to(ctx)
        res = feat
>       h = gcn2conv(g, res, feat, edge_weight=eweight)

tests\pytorch\test_nn.py:714: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gcn2conv.py:236: in forward
    edge_weight = EdgeWeightNorm('both')(graph, edge_weight)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:117: in forward
    reversed_g.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'out_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF071BE188>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1.])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
___________________ test_gcn2conv_e_weight[True-g4-idtype1] ___________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'_src_out_w': Scheme(shape=(), dtype=torch.float32), '_dst_in_w': Scheme(shape=(), dtype=torch.float32)}
      edata_schemes={'_edge_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int64, bias = True

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize("bias", [True, False])
    def test_gcn2conv_e_weight(g, idtype, bias):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        gcn2conv = nn.GCN2Conv(5, layer=2, alpha=0.5, bias=bias,
                               project_initial_features=True)
        feat = F.randn((g.number_of_nodes(), 5))
        eweight = F.ones((g.num_edges(), ))
        gcn2conv = gcn2conv.to(ctx)
        res = feat
>       h = gcn2conv(g, res, feat, edge_weight=eweight)

tests\pytorch\test_nn.py:714: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gcn2conv.py:236: in forward
    edge_weight = EdgeWeightNorm('both')(graph, edge_weight)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:117: in forward
    reversed_g.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'out_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0172C608>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1.])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
___________________ test_gcn2conv_e_weight[True-g5-idtype0] ___________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'_src_out_w': Scheme(shape=(), dtype=torch.float32), '_dst_in_w': Scheme(shape=(), dtype=torch.float32)}
      edata_schemes={'_edge_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int32, bias = True

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize("bias", [True, False])
    def test_gcn2conv_e_weight(g, idtype, bias):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        gcn2conv = nn.GCN2Conv(5, layer=2, alpha=0.5, bias=bias,
                               project_initial_features=True)
        feat = F.randn((g.number_of_nodes(), 5))
        eweight = F.ones((g.num_edges(), ))
        gcn2conv = gcn2conv.to(ctx)
        res = feat
>       h = gcn2conv(g, res, feat, edge_weight=eweight)

tests\pytorch\test_nn.py:714: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gcn2conv.py:236: in forward
    edge_weight = EdgeWeightNorm('both')(graph, edge_weight)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:117: in forward
    reversed_g.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'out_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07090548>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1.])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
___________________ test_gcn2conv_e_weight[True-g5-idtype1] ___________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'_src_out_w': Scheme(shape=(), dtype=torch.float32), '_dst_in_w': Scheme(shape=(), dtype=torch.float32)}
      edata_schemes={'_edge_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int64, bias = True

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize("bias", [True, False])
    def test_gcn2conv_e_weight(g, idtype, bias):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        gcn2conv = nn.GCN2Conv(5, layer=2, alpha=0.5, bias=bias,
                               project_initial_features=True)
        feat = F.randn((g.number_of_nodes(), 5))
        eweight = F.ones((g.num_edges(), ))
        gcn2conv = gcn2conv.to(ctx)
        res = feat
>       h = gcn2conv(g, res, feat, edge_weight=eweight)

tests\pytorch\test_nn.py:714: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gcn2conv.py:236: in forward
    edge_weight = EdgeWeightNorm('both')(graph, edge_weight)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:117: in forward
    reversed_g.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'out_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF071B0488>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1.])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_gcn2conv_e_weight[False-g0-idtype0] ___________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), '_src_out_w': Sch...}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32), '_edge_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int32, bias = False

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize("bias", [True, False])
    def test_gcn2conv_e_weight(g, idtype, bias):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        gcn2conv = nn.GCN2Conv(5, layer=2, alpha=0.5, bias=bias,
                               project_initial_features=True)
        feat = F.randn((g.number_of_nodes(), 5))
        eweight = F.ones((g.num_edges(), ))
        gcn2conv = gcn2conv.to(ctx)
        res = feat
>       h = gcn2conv(g, res, feat, edge_weight=eweight)

tests\pytorch\test_nn.py:714: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gcn2conv.py:236: in forward
    edge_weight = EdgeWeightNorm('both')(graph, edge_weight)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:117: in forward
    reversed_g.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'out_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07204B88>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1.])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_gcn2conv_e_weight[False-g0-idtype1] ___________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), '_src_out_w': Sch...}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32), '_edge_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int64, bias = False

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize("bias", [True, False])
    def test_gcn2conv_e_weight(g, idtype, bias):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        gcn2conv = nn.GCN2Conv(5, layer=2, alpha=0.5, bias=bias,
                               project_initial_features=True)
        feat = F.randn((g.number_of_nodes(), 5))
        eweight = F.ones((g.num_edges(), ))
        gcn2conv = gcn2conv.to(ctx)
        res = feat
>       h = gcn2conv(g, res, feat, edge_weight=eweight)

tests\pytorch\test_nn.py:714: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gcn2conv.py:236: in forward
    edge_weight = EdgeWeightNorm('both')(graph, edge_weight)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:117: in forward
    reversed_g.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'out_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0147B508>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1.])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_gcn2conv_e_weight[False-g1-idtype0] ___________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'_src_out_w': Scheme(shape=(), dtype=torch.float32), '_dst_in_w': Scheme(shape=(), dtype=torch.float32)}
      edata_schemes={'_edge_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int32, bias = False

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize("bias", [True, False])
    def test_gcn2conv_e_weight(g, idtype, bias):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        gcn2conv = nn.GCN2Conv(5, layer=2, alpha=0.5, bias=bias,
                               project_initial_features=True)
        feat = F.randn((g.number_of_nodes(), 5))
        eweight = F.ones((g.num_edges(), ))
        gcn2conv = gcn2conv.to(ctx)
        res = feat
>       h = gcn2conv(g, res, feat, edge_weight=eweight)

tests\pytorch\test_nn.py:714: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gcn2conv.py:236: in forward
    edge_weight = EdgeWeightNorm('both')(graph, edge_weight)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:117: in forward
    reversed_g.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'out_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF071D58C8>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1.])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_gcn2conv_e_weight[False-g1-idtype1] ___________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'_src_out_w': Scheme(shape=(), dtype=torch.float32), '_dst_in_w': Scheme(shape=(), dtype=torch.float32)}
      edata_schemes={'_edge_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int64, bias = False

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize("bias", [True, False])
    def test_gcn2conv_e_weight(g, idtype, bias):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        gcn2conv = nn.GCN2Conv(5, layer=2, alpha=0.5, bias=bias,
                               project_initial_features=True)
        feat = F.randn((g.number_of_nodes(), 5))
        eweight = F.ones((g.num_edges(), ))
        gcn2conv = gcn2conv.to(ctx)
        res = feat
>       h = gcn2conv(g, res, feat, edge_weight=eweight)

tests\pytorch\test_nn.py:714: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gcn2conv.py:236: in forward
    edge_weight = EdgeWeightNorm('both')(graph, edge_weight)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:117: in forward
    reversed_g.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'out_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF011D0588>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1.])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_gcn2conv_e_weight[False-g2-idtype0] ___________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), '_src_out_w': Sch...   edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), '_edge_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int32, bias = False

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize("bias", [True, False])
    def test_gcn2conv_e_weight(g, idtype, bias):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        gcn2conv = nn.GCN2Conv(5, layer=2, alpha=0.5, bias=bias,
                               project_initial_features=True)
        feat = F.randn((g.number_of_nodes(), 5))
        eweight = F.ones((g.num_edges(), ))
        gcn2conv = gcn2conv.to(ctx)
        res = feat
>       h = gcn2conv(g, res, feat, edge_weight=eweight)

tests\pytorch\test_nn.py:714: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gcn2conv.py:236: in forward
    edge_weight = EdgeWeightNorm('both')(graph, edge_weight)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:117: in forward
    reversed_g.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'out_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF070F67C8>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1.])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_gcn2conv_e_weight[False-g2-idtype1] ___________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), '_src_out_w': Sch...   edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), '_edge_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int64, bias = False

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize("bias", [True, False])
    def test_gcn2conv_e_weight(g, idtype, bias):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        gcn2conv = nn.GCN2Conv(5, layer=2, alpha=0.5, bias=bias,
                               project_initial_features=True)
        feat = F.randn((g.number_of_nodes(), 5))
        eweight = F.ones((g.num_edges(), ))
        gcn2conv = gcn2conv.to(ctx)
        res = feat
>       h = gcn2conv(g, res, feat, edge_weight=eweight)

tests\pytorch\test_nn.py:714: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gcn2conv.py:236: in forward
    edge_weight = EdgeWeightNorm('both')(graph, edge_weight)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:117: in forward
    reversed_g.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'out_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07281648>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1.])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_gcn2conv_e_weight[False-g3-idtype0] ___________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'_src_out_w': Scheme(shape=(), dtype=torch.float32), '_dst_in_w': Scheme(shape=(), dtype=torch.float32)}
      edata_schemes={'_edge_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int32, bias = False

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize("bias", [True, False])
    def test_gcn2conv_e_weight(g, idtype, bias):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        gcn2conv = nn.GCN2Conv(5, layer=2, alpha=0.5, bias=bias,
                               project_initial_features=True)
        feat = F.randn((g.number_of_nodes(), 5))
        eweight = F.ones((g.num_edges(), ))
        gcn2conv = gcn2conv.to(ctx)
        res = feat
>       h = gcn2conv(g, res, feat, edge_weight=eweight)

tests\pytorch\test_nn.py:714: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gcn2conv.py:236: in forward
    edge_weight = EdgeWeightNorm('both')(graph, edge_weight)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:117: in forward
    reversed_g.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'out_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF019D6BC8>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1.])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_gcn2conv_e_weight[False-g3-idtype1] ___________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'_src_out_w': Scheme(shape=(), dtype=torch.float32), '_dst_in_w': Scheme(shape=(), dtype=torch.float32)}
      edata_schemes={'_edge_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int64, bias = False

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize("bias", [True, False])
    def test_gcn2conv_e_weight(g, idtype, bias):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        gcn2conv = nn.GCN2Conv(5, layer=2, alpha=0.5, bias=bias,
                               project_initial_features=True)
        feat = F.randn((g.number_of_nodes(), 5))
        eweight = F.ones((g.num_edges(), ))
        gcn2conv = gcn2conv.to(ctx)
        res = feat
>       h = gcn2conv(g, res, feat, edge_weight=eweight)

tests\pytorch\test_nn.py:714: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gcn2conv.py:236: in forward
    edge_weight = EdgeWeightNorm('both')(graph, edge_weight)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:117: in forward
    reversed_g.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'out_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF017BF648>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1.])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_gcn2conv_e_weight[False-g4-idtype0] ___________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'_src_out_w': Scheme(shape=(), dtype=torch.float32), '_dst_in_w': Scheme(shape=(), dtype=torch.float32)}
      edata_schemes={'_edge_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int32, bias = False

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize("bias", [True, False])
    def test_gcn2conv_e_weight(g, idtype, bias):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        gcn2conv = nn.GCN2Conv(5, layer=2, alpha=0.5, bias=bias,
                               project_initial_features=True)
        feat = F.randn((g.number_of_nodes(), 5))
        eweight = F.ones((g.num_edges(), ))
        gcn2conv = gcn2conv.to(ctx)
        res = feat
>       h = gcn2conv(g, res, feat, edge_weight=eweight)

tests\pytorch\test_nn.py:714: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gcn2conv.py:236: in forward
    edge_weight = EdgeWeightNorm('both')(graph, edge_weight)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:117: in forward
    reversed_g.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'out_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01570B08>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1.])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_gcn2conv_e_weight[False-g4-idtype1] ___________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'_src_out_w': Scheme(shape=(), dtype=torch.float32), '_dst_in_w': Scheme(shape=(), dtype=torch.float32)}
      edata_schemes={'_edge_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int64, bias = False

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize("bias", [True, False])
    def test_gcn2conv_e_weight(g, idtype, bias):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        gcn2conv = nn.GCN2Conv(5, layer=2, alpha=0.5, bias=bias,
                               project_initial_features=True)
        feat = F.randn((g.number_of_nodes(), 5))
        eweight = F.ones((g.num_edges(), ))
        gcn2conv = gcn2conv.to(ctx)
        res = feat
>       h = gcn2conv(g, res, feat, edge_weight=eweight)

tests\pytorch\test_nn.py:714: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gcn2conv.py:236: in forward
    edge_weight = EdgeWeightNorm('both')(graph, edge_weight)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:117: in forward
    reversed_g.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'out_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF018998C8>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1.])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_gcn2conv_e_weight[False-g5-idtype0] ___________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'_src_out_w': Scheme(shape=(), dtype=torch.float32), '_dst_in_w': Scheme(shape=(), dtype=torch.float32)}
      edata_schemes={'_edge_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int32, bias = False

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize("bias", [True, False])
    def test_gcn2conv_e_weight(g, idtype, bias):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        gcn2conv = nn.GCN2Conv(5, layer=2, alpha=0.5, bias=bias,
                               project_initial_features=True)
        feat = F.randn((g.number_of_nodes(), 5))
        eweight = F.ones((g.num_edges(), ))
        gcn2conv = gcn2conv.to(ctx)
        res = feat
>       h = gcn2conv(g, res, feat, edge_weight=eweight)

tests\pytorch\test_nn.py:714: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gcn2conv.py:236: in forward
    edge_weight = EdgeWeightNorm('both')(graph, edge_weight)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:117: in forward
    reversed_g.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'out_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07287808>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1.])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_gcn2conv_e_weight[False-g5-idtype1] ___________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'_src_out_w': Scheme(shape=(), dtype=torch.float32), '_dst_in_w': Scheme(shape=(), dtype=torch.float32)}
      edata_schemes={'_edge_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int64, bias = False

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize("bias", [True, False])
    def test_gcn2conv_e_weight(g, idtype, bias):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        gcn2conv = nn.GCN2Conv(5, layer=2, alpha=0.5, bias=bias,
                               project_initial_features=True)
        feat = F.randn((g.number_of_nodes(), 5))
        eweight = F.ones((g.num_edges(), ))
        gcn2conv = gcn2conv.to(ctx)
        res = feat
>       h = gcn2conv(g, res, feat, edge_weight=eweight)

tests\pytorch\test_nn.py:714: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gcn2conv.py:236: in forward
    edge_weight = EdgeWeightNorm('both')(graph, edge_weight)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:117: in forward
    reversed_g.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'out_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14CCF5C8>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1.])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________________ test_sgconv_e_weight[g0-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), '_src_out_w': Sch...}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32), '_edge_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int32

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_sgconv_e_weight(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        sgconv = nn.SGConv(5, 5, 3)
        feat = F.randn((g.number_of_nodes(), 5))
        eweight = F.ones((g.num_edges(), ))
        sgconv = sgconv.to(ctx)
>       h = sgconv(g, feat, edge_weight=eweight)

tests\pytorch\test_nn.py:727: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sgconv.py:186: in forward
    'both')(graph, edge_weight)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:117: in forward
    reversed_g.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'out_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE12313548>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1.])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________________ test_sgconv_e_weight[g0-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), '_src_out_w': Sch...}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32), '_edge_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int64

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_sgconv_e_weight(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        sgconv = nn.SGConv(5, 5, 3)
        feat = F.randn((g.number_of_nodes(), 5))
        eweight = F.ones((g.num_edges(), ))
        sgconv = sgconv.to(ctx)
>       h = sgconv(g, feat, edge_weight=eweight)

tests\pytorch\test_nn.py:727: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sgconv.py:186: in forward
    'both')(graph, edge_weight)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:117: in forward
    reversed_g.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'out_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0192BE08>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1.])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________________ test_sgconv_e_weight[g1-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'_src_out_w': Scheme(shape=(), dtype=torch.float32), '_dst_in_w': Scheme(shape=(), dtype=torch.float32)}
      edata_schemes={'_edge_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int32

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_sgconv_e_weight(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        sgconv = nn.SGConv(5, 5, 3)
        feat = F.randn((g.number_of_nodes(), 5))
        eweight = F.ones((g.num_edges(), ))
        sgconv = sgconv.to(ctx)
>       h = sgconv(g, feat, edge_weight=eweight)

tests\pytorch\test_nn.py:727: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sgconv.py:186: in forward
    'both')(graph, edge_weight)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:117: in forward
    reversed_g.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'out_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0720C148>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1.])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/xbyak'...
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________________ test_sgconv_e_weight[g1-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'_src_out_w': Scheme(shape=(), dtype=torch.float32), '_dst_in_w': Scheme(shape=(), dtype=torch.float32)}
      edata_schemes={'_edge_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int64

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_sgconv_e_weight(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        sgconv = nn.SGConv(5, 5, 3)
        feat = F.randn((g.number_of_nodes(), 5))
        eweight = F.ones((g.num_edges(), ))
        sgconv = sgconv.to(ctx)
>       h = sgconv(g, feat, edge_weight=eweight)

tests\pytorch\test_nn.py:727: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sgconv.py:186: in forward
    'both')(graph, edge_weight)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:117: in forward
    reversed_g.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'out_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF071F6848>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1.])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________________ test_sgconv_e_weight[g2-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), '_src_out_w': Sch...   edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), '_edge_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int32

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_sgconv_e_weight(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        sgconv = nn.SGConv(5, 5, 3)
        feat = F.randn((g.number_of_nodes(), 5))
        eweight = F.ones((g.num_edges(), ))
        sgconv = sgconv.to(ctx)
>       h = sgconv(g, feat, edge_weight=eweight)

tests\pytorch\test_nn.py:727: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sgconv.py:186: in forward
    'both')(graph, edge_weight)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:117: in forward
    reversed_g.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'out_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07078688>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1.])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________________ test_sgconv_e_weight[g2-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), '_src_out_w': Sch...   edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), '_edge_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int64

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_sgconv_e_weight(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        sgconv = nn.SGConv(5, 5, 3)
        feat = F.randn((g.number_of_nodes(), 5))
        eweight = F.ones((g.num_edges(), ))
        sgconv = sgconv.to(ctx)
>       h = sgconv(g, feat, edge_weight=eweight)

tests\pytorch\test_nn.py:727: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sgconv.py:186: in forward
    'both')(graph, edge_weight)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:117: in forward
    reversed_g.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'out_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1454FBC8>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1.])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________________ test_sgconv_e_weight[g3-idtype0] _______________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'_src_out_w': Scheme(shape=(), dtype=torch.float32), '_dst_in_w': Scheme(shape=(), dtype=torch.float32)}
      edata_schemes={'_edge_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int32

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_sgconv_e_weight(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        sgconv = nn.SGConv(5, 5, 3)
        feat = F.randn((g.number_of_nodes(), 5))
        eweight = F.ones((g.num_edges(), ))
        sgconv = sgconv.to(ctx)
>       h = sgconv(g, feat, edge_weight=eweight)

tests\pytorch\test_nn.py:727: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sgconv.py:186: in forward
    'both')(graph, edge_weight)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:117: in forward
    reversed_g.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'out_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01899448>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1.])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________________ test_sgconv_e_weight[g3-idtype1] _______________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'_src_out_w': Scheme(shape=(), dtype=torch.float32), '_dst_in_w': Scheme(shape=(), dtype=torch.float32)}
      edata_schemes={'_edge_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int64

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_sgconv_e_weight(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        sgconv = nn.SGConv(5, 5, 3)
        feat = F.randn((g.number_of_nodes(), 5))
        eweight = F.ones((g.num_edges(), ))
        sgconv = sgconv.to(ctx)
>       h = sgconv(g, feat, edge_weight=eweight)

tests\pytorch\test_nn.py:727: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sgconv.py:186: in forward
    'both')(graph, edge_weight)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:117: in forward
    reversed_g.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'out_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF017BF408>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1.])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________________ test_sgconv_e_weight[g4-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'_src_out_w': Scheme(shape=(), dtype=torch.float32), '_dst_in_w': Scheme(shape=(), dtype=torch.float32)}
      edata_schemes={'_edge_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int32

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_sgconv_e_weight(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        sgconv = nn.SGConv(5, 5, 3)
        feat = F.randn((g.number_of_nodes(), 5))
        eweight = F.ones((g.num_edges(), ))
        sgconv = sgconv.to(ctx)
>       h = sgconv(g, feat, edge_weight=eweight)

tests\pytorch\test_nn.py:727: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sgconv.py:186: in forward
    'both')(graph, edge_weight)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:117: in forward
    reversed_g.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'out_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07078BC8>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1.])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________________ test_sgconv_e_weight[g4-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'_src_out_w': Scheme(shape=(), dtype=torch.float32), '_dst_in_w': Scheme(shape=(), dtype=torch.float32)}
      edata_schemes={'_edge_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int64

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_sgconv_e_weight(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        sgconv = nn.SGConv(5, 5, 3)
        feat = F.randn((g.number_of_nodes(), 5))
        eweight = F.ones((g.num_edges(), ))
        sgconv = sgconv.to(ctx)
>       h = sgconv(g, feat, edge_weight=eweight)

tests\pytorch\test_nn.py:727: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sgconv.py:186: in forward
    'both')(graph, edge_weight)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:117: in forward
    reversed_g.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'out_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE122E6408>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1.])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________________ test_sgconv_e_weight[g5-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'_src_out_w': Scheme(shape=(), dtype=torch.float32), '_dst_in_w': Scheme(shape=(), dtype=torch.float32)}
      edata_schemes={'_edge_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int32

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_sgconv_e_weight(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        sgconv = nn.SGConv(5, 5, 3)
        feat = F.randn((g.number_of_nodes(), 5))
        eweight = F.ones((g.num_edges(), ))
        sgconv = sgconv.to(ctx)
>       h = sgconv(g, feat, edge_weight=eweight)

tests\pytorch\test_nn.py:727: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sgconv.py:186: in forward
    'both')(graph, edge_weight)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:117: in forward
    reversed_g.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'out_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01679E48>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1.])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________________ test_sgconv_e_weight[g5-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'_src_out_w': Scheme(shape=(), dtype=torch.float32), '_dst_in_w': Scheme(shape=(), dtype=torch.float32)}
      edata_schemes={'_edge_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int64

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_sgconv_e_weight(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        sgconv = nn.SGConv(5, 5, 3)
        feat = F.randn((g.number_of_nodes(), 5))
        eweight = F.ones((g.num_edges(), ))
        sgconv = sgconv.to(ctx)
>       h = sgconv(g, feat, edge_weight=eweight)

tests\pytorch\test_nn.py:727: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sgconv.py:186: in forward
    'both')(graph, edge_weight)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:117: in forward
    reversed_g.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'out_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03D27E48>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1.])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________________ test_tagconv_e_weight[g0-idtype0] ______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), '_src_out_w': Sch...}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32), '_edge_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int32

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_tagconv_e_weight(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        conv = nn.TAGConv(5, 5, bias=True)
        conv = conv.to(ctx)
        feat = F.randn((g.number_of_nodes(), 5))
        eweight = F.ones((g.num_edges(), ))
        conv = conv.to(ctx)
>       h = conv(g, feat, edge_weight=eweight)

tests\pytorch\test_nn.py:740: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\tagconv.py:126: in forward
    'both')(graph, edge_weight)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:117: in forward
    reversed_g.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'out_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03D23C08>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1.])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________________ test_tagconv_e_weight[g0-idtype1] ______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), '_src_out_w': Sch...}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32), '_edge_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int64

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_tagconv_e_weight(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        conv = nn.TAGConv(5, 5, bias=True)
        conv = conv.to(ctx)
        feat = F.randn((g.number_of_nodes(), 5))
        eweight = F.ones((g.num_edges(), ))
        conv = conv.to(ctx)
>       h = conv(g, feat, edge_weight=eweight)

tests\pytorch\test_nn.py:740: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\tagconv.py:126: in forward
    'both')(graph, edge_weight)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:117: in forward
    reversed_g.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'out_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF016FA048>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1.])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________________ test_tagconv_e_weight[g1-idtype0] ______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'_src_out_w': Scheme(shape=(), dtype=torch.float32), '_dst_in_w': Scheme(shape=(), dtype=torch.float32)}
      edata_schemes={'_edge_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int32

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_tagconv_e_weight(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        conv = nn.TAGConv(5, 5, bias=True)
        conv = conv.to(ctx)
        feat = F.randn((g.number_of_nodes(), 5))
        eweight = F.ones((g.num_edges(), ))
        conv = conv.to(ctx)
>       h = conv(g, feat, edge_weight=eweight)

tests\pytorch\test_nn.py:740: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\tagconv.py:126: in forward
    'both')(graph, edge_weight)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:117: in forward
    reversed_g.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'out_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1237AC88>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1.])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________________ test_tagconv_e_weight[g1-idtype1] ______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'_src_out_w': Scheme(shape=(), dtype=torch.float32), '_dst_in_w': Scheme(shape=(), dtype=torch.float32)}
      edata_schemes={'_edge_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int64

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_tagconv_e_weight(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        conv = nn.TAGConv(5, 5, bias=True)
        conv = conv.to(ctx)
        feat = F.randn((g.number_of_nodes(), 5))
        eweight = F.ones((g.num_edges(), ))
        conv = conv.to(ctx)
>       h = conv(g, feat, edge_weight=eweight)

tests\pytorch\test_nn.py:740: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\tagconv.py:126: in forward
    'both')(graph, edge_weight)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:117: in forward
    reversed_g.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'out_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03D1B108>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1.])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________________ test_tagconv_e_weight[g2-idtype0] ______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), '_src_out_w': Sch...   edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), '_edge_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int32

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_tagconv_e_weight(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        conv = nn.TAGConv(5, 5, bias=True)
        conv = conv.to(ctx)
        feat = F.randn((g.number_of_nodes(), 5))
        eweight = F.ones((g.num_edges(), ))
        conv = conv.to(ctx)
>       h = conv(g, feat, edge_weight=eweight)

tests\pytorch\test_nn.py:740: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\tagconv.py:126: in forward
    'both')(graph, edge_weight)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:117: in forward
    reversed_g.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'out_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0712C5C8>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1.])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________________ test_tagconv_e_weight[g2-idtype1] ______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), '_src_out_w': Sch...   edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), '_edge_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int64

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_tagconv_e_weight(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        conv = nn.TAGConv(5, 5, bias=True)
        conv = conv.to(ctx)
        feat = F.randn((g.number_of_nodes(), 5))
        eweight = F.ones((g.num_edges(), ))
        conv = conv.to(ctx)
>       h = conv(g, feat, edge_weight=eweight)

tests\pytorch\test_nn.py:740: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\tagconv.py:126: in forward
    'both')(graph, edge_weight)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:117: in forward
    reversed_g.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'out_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07275648>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1.])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________________ test_tagconv_e_weight[g3-idtype0] ______________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'_src_out_w': Scheme(shape=(), dtype=torch.float32), '_dst_in_w': Scheme(shape=(), dtype=torch.float32)}
      edata_schemes={'_edge_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int32

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_tagconv_e_weight(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        conv = nn.TAGConv(5, 5, bias=True)
        conv = conv.to(ctx)
        feat = F.randn((g.number_of_nodes(), 5))
        eweight = F.ones((g.num_edges(), ))
        conv = conv.to(ctx)
>       h = conv(g, feat, edge_weight=eweight)

tests\pytorch\test_nn.py:740: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\tagconv.py:126: in forward
    'both')(graph, edge_weight)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:117: in forward
    reversed_g.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'out_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14AA2D88>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1.])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________________ test_tagconv_e_weight[g3-idtype1] ______________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'_src_out_w': Scheme(shape=(), dtype=torch.float32), '_dst_in_w': Scheme(shape=(), dtype=torch.float32)}
      edata_schemes={'_edge_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int64

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_tagconv_e_weight(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        conv = nn.TAGConv(5, 5, bias=True)
        conv = conv.to(ctx)
        feat = F.randn((g.number_of_nodes(), 5))
        eweight = F.ones((g.num_edges(), ))
        conv = conv.to(ctx)
>       h = conv(g, feat, edge_weight=eweight)

tests\pytorch\test_nn.py:740: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\tagconv.py:126: in forward
    'both')(graph, edge_weight)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:117: in forward
    reversed_g.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'out_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE122FA3C8>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1.])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________________ test_tagconv_e_weight[g4-idtype0] ______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'_src_out_w': Scheme(shape=(), dtype=torch.float32), '_dst_in_w': Scheme(shape=(), dtype=torch.float32)}
      edata_schemes={'_edge_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int32

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_tagconv_e_weight(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        conv = nn.TAGConv(5, 5, bias=True)
        conv = conv.to(ctx)
        feat = F.randn((g.number_of_nodes(), 5))
        eweight = F.ones((g.num_edges(), ))
        conv = conv.to(ctx)
>       h = conv(g, feat, edge_weight=eweight)

tests\pytorch\test_nn.py:740: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\tagconv.py:126: in forward
    'both')(graph, edge_weight)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:117: in forward
    reversed_g.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'out_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01487108>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1.])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________________ test_tagconv_e_weight[g4-idtype1] ______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'_src_out_w': Scheme(shape=(), dtype=torch.float32), '_dst_in_w': Scheme(shape=(), dtype=torch.float32)}
      edata_schemes={'_edge_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int64

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_tagconv_e_weight(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        conv = nn.TAGConv(5, 5, bias=True)
        conv = conv.to(ctx)
        feat = F.randn((g.number_of_nodes(), 5))
        eweight = F.ones((g.num_edges(), ))
        conv = conv.to(ctx)
>       h = conv(g, feat, edge_weight=eweight)

tests\pytorch\test_nn.py:740: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\tagconv.py:126: in forward
    'both')(graph, edge_weight)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:117: in forward
    reversed_g.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'out_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE12369D48>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1.])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________________ test_tagconv_e_weight[g5-idtype0] ______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'_src_out_w': Scheme(shape=(), dtype=torch.float32), '_dst_in_w': Scheme(shape=(), dtype=torch.float32)}
      edata_schemes={'_edge_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int32

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_tagconv_e_weight(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        conv = nn.TAGConv(5, 5, bias=True)
        conv = conv.to(ctx)
        feat = F.randn((g.number_of_nodes(), 5))
        eweight = F.ones((g.num_edges(), ))
        conv = conv.to(ctx)
>       h = conv(g, feat, edge_weight=eweight)

tests\pytorch\test_nn.py:740: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\tagconv.py:126: in forward
    'both')(graph, edge_weight)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:117: in forward
    reversed_g.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'out_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF06FFF1C8>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1.])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________________ test_tagconv_e_weight[g5-idtype1] ______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'_src_out_w': Scheme(shape=(), dtype=torch.float32), '_dst_in_w': Scheme(shape=(), dtype=torch.float32)}
      edata_schemes={'_edge_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int64

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_tagconv_e_weight(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        conv = nn.TAGConv(5, 5, bias=True)
        conv = conv.to(ctx)
        feat = F.randn((g.number_of_nodes(), 5))
        eweight = F.ones((g.num_edges(), ))
        conv = conv.to(ctx)
>       h = conv(g, feat, edge_weight=eweight)

tests\pytorch\test_nn.py:740: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\tagconv.py:126: in forward
    'both')(graph, edge_weight)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:117: in forward
    reversed_g.update_all(fn.copy_edge('_edge_w', 'm'), fn.sum('m', 'out_weight'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF06FF8B08>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1.])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_gin_conv[mean-g0-idtype0] ________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
idtype = torch.int32, aggregator_type = 'mean'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('aggregator_type', ['mean', 'max', 'sum'])
    def test_gin_conv(g, idtype, aggregator_type):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gin = nn.GINConv(
            th.nn.Linear(5, 12),
            aggregator_type
        )
        th.save(gin, tmp_buffer)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gin = gin.to(ctx)
>       h = gin(g, feat)

tests\pytorch\test_nn.py:756: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\ginconv.py:147: in forward
    graph.update_all(aggregate_fn, _reducer('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07204D48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.1578,  0.9274, -0.3338, -1.0208, -1.7261],
        [-0.9936, -1.2473,  0.3943,  0.6899, -0.1839],
        ...0.1618],
        [-0.0325,  0.3879,  0.8579,  0.0226,  0.6298],
        [-0.9232, -0.7955, -0.4868, -0.0830,  0.7374]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_gin_conv[mean-g0-idtype1] ________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
idtype = torch.int64, aggregator_type = 'mean'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('aggregator_type', ['mean', 'max', 'sum'])
    def test_gin_conv(g, idtype, aggregator_type):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gin = nn.GINConv(
            th.nn.Linear(5, 12),
            aggregator_type
        )
        th.save(gin, tmp_buffer)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gin = gin.to(ctx)
>       h = gin(g, feat)

tests\pytorch\test_nn.py:756: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\ginconv.py:147: in forward
    graph.update_all(aggregate_fn, _reducer('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF017E5348>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.0631,  1.7769, -1.5049,  0.4322, -0.6622],
        [ 0.4781,  1.2847,  1.2636,  0.9980,  1.5270],
        ...1.0043],
        [ 1.1717,  0.3404, -0.2702,  1.1591,  0.4501],
        [-1.4900, -0.3216,  0.1101,  0.3037, -0.7344]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_gin_conv[mean-g1-idtype0] ________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, aggregator_type = 'mean'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('aggregator_type', ['mean', 'max', 'sum'])
    def test_gin_conv(g, idtype, aggregator_type):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gin = nn.GINConv(
            th.nn.Linear(5, 12),
            aggregator_type
        )
        th.save(gin, tmp_buffer)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gin = gin.to(ctx)
>       h = gin(g, feat)

tests\pytorch\test_nn.py:756: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\ginconv.py:147: in forward
    graph.update_all(aggregate_fn, _reducer('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF016F5AC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.1399,  1.2955,  0.1397,  0.2387, -0.4162],
        [ 1.7430, -0.1689,  0.9269,  0.0209,  0.4076],
        ...0.6856],
        [-1.1189,  0.7819,  1.8479,  2.0297, -0.6487],
        [ 0.9764, -0.0551, -0.6369, -2.4758,  0.6960]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_gin_conv[mean-g1-idtype1] ________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, aggregator_type = 'mean'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('aggregator_type', ['mean', 'max', 'sum'])
    def test_gin_conv(g, idtype, aggregator_type):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gin = nn.GINConv(
            th.nn.Linear(5, 12),
            aggregator_type
        )
        th.save(gin, tmp_buffer)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gin = gin.to(ctx)
>       h = gin(g, feat)

tests\pytorch\test_nn.py:756: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\ginconv.py:147: in forward
    graph.update_all(aggregate_fn, _reducer('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01723948>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.1854,  0.6338, -2.5521,  0.6106, -0.2610],
        [-0.7510, -0.6518,  2.0678,  1.6954, -0.2636],
        ...0.0880],
        [ 0.4399,  0.0286, -0.3754,  0.9592, -1.1065],
        [ 0.8646, -0.7142,  0.4081,  1.4733, -0.2494]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_gin_conv[mean-g2-idtype0] ________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int32, aggregator_type = 'mean'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('aggregator_type', ['mean', 'max', 'sum'])
    def test_gin_conv(g, idtype, aggregator_type):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gin = nn.GINConv(
            th.nn.Linear(5, 12),
            aggregator_type
        )
        th.save(gin, tmp_buffer)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gin = gin.to(ctx)
>       h = gin(g, feat)

tests\pytorch\test_nn.py:756: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\ginconv.py:147: in forward
    graph.update_all(aggregate_fn, _reducer('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07281448>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.0171, -0.8855,  1.4332, -1.4715,  0.6713],
        [-0.2114,  0.7143,  0.7356, -0.5021,  0.3700],
        ...0.2789],
        [ 0.6205,  0.8589,  0.8029, -0.9879,  1.8120],
        [ 1.5201,  0.4331,  0.5148, -1.2150, -1.2043]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_gin_conv[mean-g2-idtype1] ________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int64, aggregator_type = 'mean'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('aggregator_type', ['mean', 'max', 'sum'])
    def test_gin_conv(g, idtype, aggregator_type):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gin = nn.GINConv(
            th.nn.Linear(5, 12),
            aggregator_type
        )
        th.save(gin, tmp_buffer)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gin = gin.to(ctx)
>       h = gin(g, feat)

tests\pytorch\test_nn.py:756: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\ginconv.py:147: in forward
    graph.update_all(aggregate_fn, _reducer('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE122E6A88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.4136, -0.0842, -0.0607,  2.7614,  0.9834],
        [-0.4795,  1.7599, -1.2918,  0.4193,  1.5680],
        ...0.2102],
        [ 0.4479,  2.2428,  0.7148, -0.2024,  1.0989],
        [ 2.2271,  0.1086,  0.6063, -0.8233, -1.8636]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_gin_conv[mean-g3-idtype0] ________________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, aggregator_type = 'mean'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('aggregator_type', ['mean', 'max', 'sum'])
    def test_gin_conv(g, idtype, aggregator_type):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gin = nn.GINConv(
            th.nn.Linear(5, 12),
            aggregator_type
        )
        th.save(gin, tmp_buffer)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gin = gin.to(ctx)
>       h = gin(g, feat)

tests\pytorch\test_nn.py:756: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\ginconv.py:147: in forward
    graph.update_all(aggregate_fn, _reducer('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07079B08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.0597, -0.3504, -0.1914,  0.0507,  0.5362],
        [ 2.2427, -1.5097, -1.5408,  0.3203,  0.7216],
        ...2.1791],
        [ 1.0410, -0.5246, -1.3653, -0.7680, -0.9917],
        [-0.6230,  1.2993, -0.3299,  0.9430,  0.8012]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_gin_conv[mean-g3-idtype1] ________________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, aggregator_type = 'mean'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('aggregator_type', ['mean', 'max', 'sum'])
    def test_gin_conv(g, idtype, aggregator_type):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gin = nn.GINConv(
            th.nn.Linear(5, 12),
            aggregator_type
        )
        th.save(gin, tmp_buffer)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gin = gin.to(ctx)
>       h = gin(g, feat)

tests\pytorch\test_nn.py:756: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\ginconv.py:147: in forward
    graph.update_all(aggregate_fn, _reducer('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1227FD48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.5972,  0.0490,  0.0438, -1.3045, -0.4567],
        [-2.1196, -0.0708, -2.3856,  0.4803,  0.1487],
        ...1.8917],
        [ 2.1346,  0.1067,  0.2688,  0.3583,  0.2093],
        [-1.2903,  0.5724, -1.0920, -0.2209,  0.9032]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_gin_conv[mean-g4-idtype0] ________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, aggregator_type = 'mean'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('aggregator_type', ['mean', 'max', 'sum'])
    def test_gin_conv(g, idtype, aggregator_type):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gin = nn.GINConv(
            th.nn.Linear(5, 12),
            aggregator_type
        )
        th.save(gin, tmp_buffer)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gin = gin.to(ctx)
>       h = gin(g, feat)

tests\pytorch\test_nn.py:756: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\ginconv.py:147: in forward
    graph.update_all(aggregate_fn, _reducer('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF017C5848>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.4177, -0.5618,  0.4309, -0.9013, -1.3876],
        [ 0.7270,  0.1227,  1.0965, -0.4204,  0.6108],
        ...1.7923],
        [ 1.0580, -0.3592,  0.5705,  1.2009,  1.9194],
        [-0.4993,  0.3096, -0.6783, -0.6336,  0.6429]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_gin_conv[mean-g4-idtype1] ________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, aggregator_type = 'mean'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('aggregator_type', ['mean', 'max', 'sum'])
    def test_gin_conv(g, idtype, aggregator_type):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gin = nn.GINConv(
            th.nn.Linear(5, 12),
            aggregator_type
        )
        th.save(gin, tmp_buffer)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gin = gin.to(ctx)
>       h = gin(g, feat)

tests\pytorch\test_nn.py:756: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\ginconv.py:147: in forward
    graph.update_all(aggregate_fn, _reducer('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF071EE208>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.3501, -1.4167, -1.0396, -0.3281,  0.5279],
        [ 0.4625, -0.6377, -1.2637,  1.2002, -0.5915],
        ...1.6001],
        [-0.1340,  0.3802,  0.7768, -0.2667, -0.5524],
        [ 0.6077,  0.0641, -0.5888, -1.6979, -1.3596]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_gin_conv[mean-g5-idtype0] ________________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int32
aggregator_type = 'mean'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('aggregator_type', ['mean', 'max', 'sum'])
    def test_gin_conv(g, idtype, aggregator_type):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gin = nn.GINConv(
            th.nn.Linear(5, 12),
            aggregator_type
        )
        th.save(gin, tmp_buffer)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gin = gin.to(ctx)
>       h = gin(g, feat)

tests\pytorch\test_nn.py:756: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\ginconv.py:147: in forward
    graph.update_all(aggregate_fn, _reducer('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF019AD8C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.1331, -1.1189, -0.6516,  1.0083, -0.3977],
        [ 1.2300,  0.0524,  0.1886, -0.6744,  0.0113],
        ...0.2066],
        [ 0.7235, -1.1525,  0.0567, -0.6923,  0.4674],
        [ 0.4098,  1.3261, -1.3476,  2.8055, -1.6596]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_gin_conv[mean-g5-idtype1] ________________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int64
aggregator_type = 'mean'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('aggregator_type', ['mean', 'max', 'sum'])
    def test_gin_conv(g, idtype, aggregator_type):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gin = nn.GINConv(
            th.nn.Linear(5, 12),
            aggregator_type
        )
        th.save(gin, tmp_buffer)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gin = gin.to(ctx)
>       h = gin(g, feat)

tests\pytorch\test_nn.py:756: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\ginconv.py:147: in forward
    graph.update_all(aggregate_fn, _reducer('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07049CC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-2.4072, -0.9553, -2.5469, -1.1501,  0.2915],
        [-1.3309, -2.0913,  0.3251,  0.1511, -0.5105],
        ...2.0513],
        [ 0.0969, -0.6941,  0.5711,  1.2488,  0.6080],
        [ 0.0606,  0.9393,  0.7249,  0.3028,  0.0175]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_gin_conv[mean-g6-idtype0] ________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, aggregator_type = 'mean'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('aggregator_type', ['mean', 'max', 'sum'])
    def test_gin_conv(g, idtype, aggregator_type):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gin = nn.GINConv(
            th.nn.Linear(5, 12),
            aggregator_type
        )
        th.save(gin, tmp_buffer)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gin = gin.to(ctx)
>       h = gin(g, feat)

tests\pytorch\test_nn.py:756: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\ginconv.py:147: in forward
    graph.update_all(aggregate_fn, _reducer('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0712C908>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.2209,  0.7108,  0.5997, -0.0613,  0.3337],
        [-0.9931,  0.8302, -0.5921,  0.2965, -0.2925],
        ...0.2222],
        [ 1.4555, -0.3253, -0.5436,  0.0366,  0.4359],
        [-1.6682, -1.2368, -0.5446,  0.9565,  0.7209]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_gin_conv[mean-g6-idtype1] ________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, aggregator_type = 'mean'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('aggregator_type', ['mean', 'max', 'sum'])
    def test_gin_conv(g, idtype, aggregator_type):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gin = nn.GINConv(
            th.nn.Linear(5, 12),
            aggregator_type
        )
        th.save(gin, tmp_buffer)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gin = gin.to(ctx)
>       h = gin(g, feat)

tests\pytorch\test_nn.py:756: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\ginconv.py:147: in forward
    graph.update_all(aggregate_fn, _reducer('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF06FFF908>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.2938e+00, -1.2575e+00,  5.1674e-01,  2.1682e+00, -4.2185e-01],
        [ 2.2190e-03, -3.9464e-01,  7.4524e...01, -6.6164e-01, -4.9276e-01, -1.2803e+00],
        [-2.6160e-01, -9.0249e-01, -5.9452e-01, -2.2122e+00,  4.5144e-01]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________________ test_gin_conv[max-g0-idtype0] ________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
idtype = torch.int32, aggregator_type = 'max'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('aggregator_type', ['mean', 'max', 'sum'])
    def test_gin_conv(g, idtype, aggregator_type):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gin = nn.GINConv(
            th.nn.Linear(5, 12),
            aggregator_type
        )
        th.save(gin, tmp_buffer)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gin = gin.to(ctx)
>       h = gin(g, feat)

tests\pytorch\test_nn.py:756: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\ginconv.py:147: in forward
    graph.update_all(aggregate_fn, _reducer('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE12287908>
op = 'copy_lhs', reduce_op = 'max'
lhs_data = tensor([[ 0.4521, -0.0279, -0.5417,  0.4252,  0.4747],
        [-1.1361,  1.6613,  0.2875,  1.0265, -0.6965],
        ...0.7936],
        [-1.3712,  0.8832,  0.0483, -0.4799, -1.0016],
        [ 0.4824,  1.4265, -0.9215,  0.5297,  0.1564]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________________ test_gin_conv[max-g0-idtype1] ________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
idtype = torch.int64, aggregator_type = 'max'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('aggregator_type', ['mean', 'max', 'sum'])
    def test_gin_conv(g, idtype, aggregator_type):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gin = nn.GINConv(
            th.nn.Linear(5, 12),
            aggregator_type
        )
        th.save(gin, tmp_buffer)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gin = gin.to(ctx)
>       h = gin(g, feat)

tests\pytorch\test_nn.py:756: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\ginconv.py:147: in forward
    graph.update_all(aggregate_fn, _reducer('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07150608>
op = 'copy_lhs', reduce_op = 'max'
lhs_data = tensor([[-0.0891,  0.0099,  1.2986, -0.2674,  0.8560],
        [ 0.9391, -0.2227,  0.0610, -1.5881,  1.7273],
        ...0.5334],
        [-0.3030,  0.4995, -1.1955,  0.8334,  0.0156],
        [-0.2388,  1.5657,  0.4686,  0.7966, -0.8837]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________________ test_gin_conv[max-g1-idtype0] ________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, aggregator_type = 'max'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('aggregator_type', ['mean', 'max', 'sum'])
    def test_gin_conv(g, idtype, aggregator_type):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gin = nn.GINConv(
            th.nn.Linear(5, 12),
            aggregator_type
        )
        th.save(gin, tmp_buffer)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gin = gin.to(ctx)
>       h = gin(g, feat)

tests\pytorch\test_nn.py:756: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\ginconv.py:147: in forward
    graph.update_all(aggregate_fn, _reducer('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF072AA9C8>
op = 'copy_lhs', reduce_op = 'max'
lhs_data = tensor([[ 1.0191, -0.8505,  0.1193, -0.9345, -1.5733],
        [-0.8477,  2.5242,  0.8921,  0.4702,  1.6213],
        ...0.5669],
        [ 0.1155, -0.9951, -1.2259,  0.2155,  0.2644],
        [-0.3859, -0.2998, -0.2218,  0.3657,  0.3656]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________________ test_gin_conv[max-g1-idtype1] ________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, aggregator_type = 'max'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('aggregator_type', ['mean', 'max', 'sum'])
    def test_gin_conv(g, idtype, aggregator_type):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gin = nn.GINConv(
            th.nn.Linear(5, 12),
            aggregator_type
        )
        th.save(gin, tmp_buffer)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gin = gin.to(ctx)
>       h = gin(g, feat)

tests\pytorch\test_nn.py:756: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\ginconv.py:147: in forward
    graph.update_all(aggregate_fn, _reducer('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF070A5948>
op = 'copy_lhs', reduce_op = 'max'
lhs_data = tensor([[ 0.1430,  0.4813, -3.2906,  0.2665,  1.8577],
        [-0.0933,  1.5538, -0.7399,  1.0764,  0.2280],
        ...0.1311],
        [ 0.8158, -1.0273,  0.5022,  2.4275,  0.3870],
        [ 0.6560,  0.6469,  1.0354, -0.3586,  1.8573]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________________ test_gin_conv[max-g2-idtype0] ________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int32, aggregator_type = 'max'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('aggregator_type', ['mean', 'max', 'sum'])
    def test_gin_conv(g, idtype, aggregator_type):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gin = nn.GINConv(
            th.nn.Linear(5, 12),
            aggregator_type
        )
        th.save(gin, tmp_buffer)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gin = gin.to(ctx)
>       h = gin(g, feat)

tests\pytorch\test_nn.py:756: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\ginconv.py:147: in forward
    graph.update_all(aggregate_fn, _reducer('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01518F08>
op = 'copy_lhs', reduce_op = 'max'
lhs_data = tensor([[ 0.0590, -1.1471, -0.1841,  1.3831,  0.1204],
        [-0.6261,  0.0087,  0.0347, -0.3270,  0.9536],
        ...0.6791],
        [-0.6236,  0.5164,  2.5420, -1.5327,  0.9765],
        [ 1.1259, -0.8221,  0.3412, -0.4780, -1.5303]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________________ test_gin_conv[max-g2-idtype1] ________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int64, aggregator_type = 'max'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('aggregator_type', ['mean', 'max', 'sum'])
    def test_gin_conv(g, idtype, aggregator_type):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gin = nn.GINConv(
            th.nn.Linear(5, 12),
            aggregator_type
        )
        th.save(gin, tmp_buffer)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gin = gin.to(ctx)
>       h = gin(g, feat)

tests\pytorch\test_nn.py:756: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\ginconv.py:147: in forward
    graph.update_all(aggregate_fn, _reducer('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF014C8708>
op = 'copy_lhs', reduce_op = 'max'
lhs_data = tensor([[ 1.5038e-01, -1.0726e+00,  2.3176e+00,  1.6168e+00,  6.8585e-01],
        [-1.2723e+00, -4.8183e-01, -1.8165e...01, -9.9712e-01,  5.2204e-01,  1.1504e+00],
        [ 6.2358e-01,  4.0150e-02, -1.1157e+00, -4.8686e-02,  1.0958e+00]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________________ test_gin_conv[max-g3-idtype0] ________________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, aggregator_type = 'max'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('aggregator_type', ['mean', 'max', 'sum'])
    def test_gin_conv(g, idtype, aggregator_type):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gin = nn.GINConv(
            th.nn.Linear(5, 12),
            aggregator_type
        )
        th.save(gin, tmp_buffer)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gin = gin.to(ctx)
>       h = gin(g, feat)

tests\pytorch\test_nn.py:756: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\ginconv.py:147: in forward
    graph.update_all(aggregate_fn, _reducer('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07090D88>
op = 'copy_lhs', reduce_op = 'max'
lhs_data = tensor([[-0.5151,  0.0405, -2.1773, -0.2679, -0.3728],
        [-0.2094, -0.1573,  0.8179,  1.0720,  0.4018],
        ...0.2388],
        [-2.9170, -0.5191,  0.4771, -1.5897, -0.0954],
        [ 1.4005, -0.5708,  0.0108, -1.6201, -0.1515]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________________ test_gin_conv[max-g3-idtype1] ________________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, aggregator_type = 'max'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('aggregator_type', ['mean', 'max', 'sum'])
    def test_gin_conv(g, idtype, aggregator_type):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gin = nn.GINConv(
            th.nn.Linear(5, 12),
            aggregator_type
        )
        th.save(gin, tmp_buffer)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gin = gin.to(ctx)
>       h = gin(g, feat)

tests\pytorch\test_nn.py:756: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\ginconv.py:147: in forward
    graph.update_all(aggregate_fn, _reducer('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03D1B908>
op = 'copy_lhs', reduce_op = 'max'
lhs_data = tensor([[ 0.0801,  1.3213,  0.5106, -0.2097, -1.1012],
        [-0.4281, -0.1443, -0.5178, -1.2243,  0.8792],
        ...0.2926],
        [-0.1945, -0.1867,  0.2236, -0.8844,  0.1374],
        [ 2.4159,  2.0315,  1.0913, -0.0300,  0.3863]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________________ test_gin_conv[max-g4-idtype0] ________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, aggregator_type = 'max'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('aggregator_type', ['mean', 'max', 'sum'])
    def test_gin_conv(g, idtype, aggregator_type):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gin = nn.GINConv(
            th.nn.Linear(5, 12),
            aggregator_type
        )
        th.save(gin, tmp_buffer)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gin = gin.to(ctx)
>       h = gin(g, feat)

tests\pytorch\test_nn.py:756: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\ginconv.py:147: in forward
    graph.update_all(aggregate_fn, _reducer('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01680FC8>
op = 'copy_lhs', reduce_op = 'max'
lhs_data = tensor([[-1.9890,  1.5519, -0.2495, -0.8028, -0.2431],
        [ 0.1178, -0.6053, -0.3829,  1.6680,  1.0775],
        ...0.0449],
        [-0.9177,  0.9627,  2.9129,  1.7519,  1.4250],
        [ 0.5581,  0.7741, -0.3476,  0.2286, -0.9125]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________________ test_gin_conv[max-g4-idtype1] ________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, aggregator_type = 'max'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('aggregator_type', ['mean', 'max', 'sum'])
    def test_gin_conv(g, idtype, aggregator_type):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gin = nn.GINConv(
            th.nn.Linear(5, 12),
            aggregator_type
        )
        th.save(gin, tmp_buffer)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gin = gin.to(ctx)
>       h = gin(g, feat)

tests\pytorch\test_nn.py:756: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\ginconv.py:147: in forward
    graph.update_all(aggregate_fn, _reducer('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07330E48>
op = 'copy_lhs', reduce_op = 'max'
lhs_data = tensor([[ 0.1476,  0.7466, -0.7712,  0.9318,  1.1660],
        [-0.8856, -0.3422,  1.2024, -0.6597,  0.4257],
        ...0.8630],
        [ 1.5186, -0.8865, -1.2251,  0.5028, -0.3721],
        [ 1.0502,  2.1292, -0.8927,  0.9118, -0.1986]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________________ test_gin_conv[max-g5-idtype0] ________________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int32
aggregator_type = 'max'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('aggregator_type', ['mean', 'max', 'sum'])
    def test_gin_conv(g, idtype, aggregator_type):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gin = nn.GINConv(
            th.nn.Linear(5, 12),
            aggregator_type
        )
        th.save(gin, tmp_buffer)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gin = gin.to(ctx)
>       h = gin(g, feat)

tests\pytorch\test_nn.py:756: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\ginconv.py:147: in forward
    graph.update_all(aggregate_fn, _reducer('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE122E9E48>
op = 'copy_lhs', reduce_op = 'max'
lhs_data = tensor([[-0.7037, -0.3709, -1.1668, -0.3380, -1.5027],
        [-0.2211,  0.5467,  0.6273, -0.1265,  0.1151],
        ...0.1411],
        [ 0.9007, -0.9212, -0.2278,  0.7187, -0.2668],
        [ 1.0508, -0.0484, -0.2209,  1.3864,  2.8027]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________________ test_gin_conv[max-g5-idtype1] ________________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int64
aggregator_type = 'max'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('aggregator_type', ['mean', 'max', 'sum'])
    def test_gin_conv(g, idtype, aggregator_type):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gin = nn.GINConv(
            th.nn.Linear(5, 12),
            aggregator_type
        )
        th.save(gin, tmp_buffer)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gin = gin.to(ctx)
>       h = gin(g, feat)

tests\pytorch\test_nn.py:756: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\ginconv.py:147: in forward
    graph.update_all(aggregate_fn, _reducer('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01518EC8>
op = 'copy_lhs', reduce_op = 'max'
lhs_data = tensor([[ 0.5347, -1.3889,  1.0142,  0.9753,  1.1949],
        [-0.0308, -0.8895,  0.6418,  0.1913, -0.4456],
        ...2.7313],
        [ 0.9385,  0.0660,  0.8051, -0.7371,  0.7952],
        [-0.2538,  1.3538,  0.2255, -0.2889, -0.6752]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________________ test_gin_conv[max-g6-idtype0] ________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, aggregator_type = 'max'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('aggregator_type', ['mean', 'max', 'sum'])
    def test_gin_conv(g, idtype, aggregator_type):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gin = nn.GINConv(
            th.nn.Linear(5, 12),
            aggregator_type
        )
        th.save(gin, tmp_buffer)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gin = gin.to(ctx)
>       h = gin(g, feat)

tests\pytorch\test_nn.py:756: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\ginconv.py:147: in forward
    graph.update_all(aggregate_fn, _reducer('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF011F0888>
op = 'copy_lhs', reduce_op = 'max'
lhs_data = tensor([[-1.3043, -1.4004, -0.8137, -0.4508, -1.3827],
        [-1.1136,  1.1781,  0.0123, -0.2525, -0.1335],
        ...1.4103],
        [-0.9252,  1.4059, -1.1902, -0.8723, -0.8557],
        [ 1.7429, -2.2370,  0.4029,  1.3454,  1.0214]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________________ test_gin_conv[max-g6-idtype1] ________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, aggregator_type = 'max'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('aggregator_type', ['mean', 'max', 'sum'])
    def test_gin_conv(g, idtype, aggregator_type):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gin = nn.GINConv(
            th.nn.Linear(5, 12),
            aggregator_type
        )
        th.save(gin, tmp_buffer)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gin = gin.to(ctx)
>       h = gin(g, feat)

tests\pytorch\test_nn.py:756: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\ginconv.py:147: in forward
    graph.update_all(aggregate_fn, _reducer('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE12294D08>
op = 'copy_lhs', reduce_op = 'max'
lhs_data = tensor([[ 1.1308, -1.1358,  0.6479, -2.0350,  0.4969],
        [ 0.4511, -1.2825,  0.0696,  0.6036,  1.2983],
        ...0.9945],
        [ 0.9168,  0.0463, -0.5607,  1.3766,  1.4715],
        [ 0.9571,  2.1828,  0.2147, -0.7896,  0.7244]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________________ test_gin_conv[sum-g0-idtype0] ________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
idtype = torch.int32, aggregator_type = 'sum'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('aggregator_type', ['mean', 'max', 'sum'])
    def test_gin_conv(g, idtype, aggregator_type):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gin = nn.GINConv(
            th.nn.Linear(5, 12),
            aggregator_type
        )
        th.save(gin, tmp_buffer)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gin = gin.to(ctx)
>       h = gin(g, feat)

tests\pytorch\test_nn.py:756: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\ginconv.py:147: in forward
    graph.update_all(aggregate_fn, _reducer('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0199EFC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.5834e+00,  1.5742e+00, -1.0513e+00,  2.3277e+00,  3.8425e-01],
        [ 5.8044e-01, -5.5099e-04,  1.4138e...00, -2.0909e+00,  8.4466e-01, -4.4795e-01],
        [ 7.9799e-01,  8.3325e-02, -2.5030e-01,  1.4930e-01,  1.7076e+00]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________________ test_gin_conv[sum-g0-idtype1] ________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
idtype = torch.int64, aggregator_type = 'sum'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('aggregator_type', ['mean', 'max', 'sum'])
    def test_gin_conv(g, idtype, aggregator_type):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gin = nn.GINConv(
            th.nn.Linear(5, 12),
            aggregator_type
        )
        th.save(gin, tmp_buffer)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gin = gin.to(ctx)
>       h = gin(g, feat)

tests\pytorch\test_nn.py:756: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\ginconv.py:147: in forward
    graph.update_all(aggregate_fn, _reducer('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF030C1888>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.6558, -0.9535, -0.9455, -1.4452,  0.6496],
        [ 1.4147, -1.5048,  0.7882, -0.5024,  0.3199],
        ...1.8232],
        [ 0.6205, -1.6389, -0.0143, -1.6953, -0.0999],
        [ 0.5037, -2.0145,  1.4395,  1.2640, -1.7950]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________________ test_gin_conv[sum-g1-idtype0] ________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, aggregator_type = 'sum'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('aggregator_type', ['mean', 'max', 'sum'])
    def test_gin_conv(g, idtype, aggregator_type):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gin = nn.GINConv(
            th.nn.Linear(5, 12),
            aggregator_type
        )
        th.save(gin, tmp_buffer)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gin = gin.to(ctx)
>       h = gin(g, feat)

tests\pytorch\test_nn.py:756: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\ginconv.py:147: in forward
    graph.update_all(aggregate_fn, _reducer('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01756548>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-5.8432e-01, -8.6030e-02, -5.8102e-01, -1.8006e+00, -1.3131e-02],
        [-1.2956e+00,  1.8444e-01,  2.0077e...01,  6.4482e-01, -7.4000e-02,  1.5370e-01],
        [ 1.8526e+00, -2.7075e-01,  6.1798e-01, -1.9956e-01,  2.0947e-01]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________________ test_gin_conv[sum-g1-idtype1] ________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, aggregator_type = 'sum'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('aggregator_type', ['mean', 'max', 'sum'])
    def test_gin_conv(g, idtype, aggregator_type):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gin = nn.GINConv(
            th.nn.Linear(5, 12),
            aggregator_type
        )
        th.save(gin, tmp_buffer)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gin = gin.to(ctx)
>       h = gin(g, feat)

tests\pytorch\test_nn.py:756: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\ginconv.py:147: in forward
    graph.update_all(aggregate_fn, _reducer('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01680C48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.0516, -1.1410,  1.5335, -0.0352,  0.5650],
        [ 0.4788,  0.5885,  0.8207,  1.3696, -0.2886],
        ...0.8250],
        [ 0.7430,  0.0500,  1.3693,  0.2391, -0.6103],
        [-0.1752, -0.0939, -0.2892, -0.5817, -1.4244]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________________ test_gin_conv[sum-g2-idtype0] ________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int32, aggregator_type = 'sum'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('aggregator_type', ['mean', 'max', 'sum'])
    def test_gin_conv(g, idtype, aggregator_type):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gin = nn.GINConv(
            th.nn.Linear(5, 12),
            aggregator_type
        )
        th.save(gin, tmp_buffer)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gin = gin.to(ctx)
>       h = gin(g, feat)

tests\pytorch\test_nn.py:756: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\ginconv.py:147: in forward
    graph.update_all(aggregate_fn, _reducer('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0702CB88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 8.5057e-02,  1.3194e+00, -4.5185e-01,  1.9764e+00,  1.1466e-03],
        [ 2.1355e+00, -6.3669e-01,  7.4861e...01, -1.2152e+00,  3.2271e+00, -1.2409e-01],
        [-2.0427e-01,  5.1231e-01, -7.7207e-01, -2.0429e+00,  6.8140e-01]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________________ test_gin_conv[sum-g2-idtype1] ________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int64, aggregator_type = 'sum'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('aggregator_type', ['mean', 'max', 'sum'])
    def test_gin_conv(g, idtype, aggregator_type):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gin = nn.GINConv(
            th.nn.Linear(5, 12),
            aggregator_type
        )
        th.save(gin, tmp_buffer)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gin = gin.to(ctx)
>       h = gin(g, feat)

tests\pytorch\test_nn.py:756: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\ginconv.py:147: in forward
    graph.update_all(aggregate_fn, _reducer('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01756A48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.3646, -1.2615,  0.2554, -0.5286,  0.7959],
        [-0.9727,  0.2038,  0.1229, -0.6403, -1.9579],
        ...1.0620],
        [ 0.3966, -0.8025, -0.4366, -0.4278,  1.4910],
        [ 1.5326, -0.3118, -0.5521,  0.7089,  0.1363]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________________ test_gin_conv[sum-g3-idtype0] ________________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, aggregator_type = 'sum'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('aggregator_type', ['mean', 'max', 'sum'])
    def test_gin_conv(g, idtype, aggregator_type):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gin = nn.GINConv(
            th.nn.Linear(5, 12),
            aggregator_type
        )
        th.save(gin, tmp_buffer)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gin = gin.to(ctx)
>       h = gin(g, feat)

tests\pytorch\test_nn.py:756: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\ginconv.py:147: in forward
    graph.update_all(aggregate_fn, _reducer('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07197388>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.2012, -0.4383,  0.4052,  0.6541,  0.1973],
        [ 0.7956,  1.3828, -0.1019, -0.1491, -2.6134],
        ...0.1178],
        [ 0.2065,  0.6472,  0.0783,  0.5443,  0.6133],
        [-1.7336, -0.3745, -0.4009,  0.1746,  0.0723]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________________ test_gin_conv[sum-g3-idtype1] ________________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, aggregator_type = 'sum'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('aggregator_type', ['mean', 'max', 'sum'])
    def test_gin_conv(g, idtype, aggregator_type):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gin = nn.GINConv(
            th.nn.Linear(5, 12),
            aggregator_type
        )
        th.save(gin, tmp_buffer)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gin = gin.to(ctx)
>       h = gin(g, feat)

tests\pytorch\test_nn.py:756: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\ginconv.py:147: in forward
    graph.update_all(aggregate_fn, _reducer('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03D492C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.3851e+00, -1.6531e+00,  3.1130e-01, -1.0454e+00,  5.5015e-01],
        [-9.2134e-01,  3.6062e-01, -1.5127e...01, -1.1124e+00,  1.1652e+00,  2.0797e+00],
        [-1.2967e+00,  3.8297e-01, -1.3795e+00, -5.6451e-01,  1.6123e-01]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________________ test_gin_conv[sum-g4-idtype0] ________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, aggregator_type = 'sum'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('aggregator_type', ['mean', 'max', 'sum'])
    def test_gin_conv(g, idtype, aggregator_type):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gin = nn.GINConv(
            th.nn.Linear(5, 12),
            aggregator_type
        )
        th.save(gin, tmp_buffer)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gin = gin.to(ctx)
>       h = gin(g, feat)

tests\pytorch\test_nn.py:756: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\ginconv.py:147: in forward
    graph.update_all(aggregate_fn, _reducer('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF030F01C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.0411,  0.6564, -0.3562, -2.7572,  2.1971],
        [ 1.6199,  1.8120, -1.4224,  0.5631,  1.5604],
        ...0.2546],
        [ 0.1667,  0.4026,  0.5303,  0.1654,  0.3171],
        [-0.2986, -0.9095, -0.8957,  0.4363,  2.2527]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________________ test_gin_conv[sum-g4-idtype1] ________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, aggregator_type = 'sum'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('aggregator_type', ['mean', 'max', 'sum'])
    def test_gin_conv(g, idtype, aggregator_type):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gin = nn.GINConv(
            th.nn.Linear(5, 12),
            aggregator_type
        )
        th.save(gin, tmp_buffer)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gin = gin.to(ctx)
>       h = gin(g, feat)

tests\pytorch\test_nn.py:756: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\ginconv.py:147: in forward
    graph.update_all(aggregate_fn, _reducer('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF070D08C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.7913,  1.2642,  0.2892,  1.0980,  0.1619],
        [-0.7544, -1.3295, -3.0947,  0.9309, -2.8356],
        ...0.7324],
        [ 0.0822, -0.2266,  0.0478,  3.0692,  0.3972],
        [ 0.6488,  0.7146,  0.3584, -0.8233,  0.3279]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________________ test_gin_conv[sum-g5-idtype0] ________________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int32
aggregator_type = 'sum'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('aggregator_type', ['mean', 'max', 'sum'])
    def test_gin_conv(g, idtype, aggregator_type):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gin = nn.GINConv(
            th.nn.Linear(5, 12),
            aggregator_type
        )
        th.save(gin, tmp_buffer)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gin = gin.to(ctx)
>       h = gin(g, feat)

tests\pytorch\test_nn.py:756: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\ginconv.py:147: in forward
    graph.update_all(aggregate_fn, _reducer('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01908B08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.6447,  0.8006,  1.0161, -0.1675, -0.3446],
        [ 1.0337, -0.6133,  0.9268, -0.1681,  0.1925],
        ...1.2575],
        [-0.6836, -0.0145, -0.7349, -1.3590,  0.5426],
        [ 1.0773, -0.6362,  0.1814, -0.4191, -1.3380]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________________ test_gin_conv[sum-g5-idtype1] ________________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int64
aggregator_type = 'sum'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('aggregator_type', ['mean', 'max', 'sum'])
    def test_gin_conv(g, idtype, aggregator_type):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gin = nn.GINConv(
            th.nn.Linear(5, 12),
            aggregator_type
        )
        th.save(gin, tmp_buffer)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gin = gin.to(ctx)
>       h = gin(g, feat)

tests\pytorch\test_nn.py:756: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\ginconv.py:147: in forward
    graph.update_all(aggregate_fn, _reducer('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01999988>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-7.8052e-01,  8.0543e-03, -7.0697e-01, -4.1293e-01,  1.8542e+00],
        [ 1.0028e-01,  1.7926e+00,  9.3330e...01, -1.8927e+00,  8.7882e-01,  2.3596e-01],
        [-1.5516e+00, -9.8148e-01,  4.5159e-01,  5.4020e-02, -1.0345e+00]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________________ test_gin_conv[sum-g6-idtype0] ________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, aggregator_type = 'sum'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('aggregator_type', ['mean', 'max', 'sum'])
    def test_gin_conv(g, idtype, aggregator_type):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gin = nn.GINConv(
            th.nn.Linear(5, 12),
            aggregator_type
        )
        th.save(gin, tmp_buffer)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gin = gin.to(ctx)
>       h = gin(g, feat)

tests\pytorch\test_nn.py:756: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\ginconv.py:147: in forward
    graph.update_all(aggregate_fn, _reducer('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE122E2648>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.5355,  0.0127, -1.2381, -0.4347, -0.4400],
        [-0.1713, -1.3838,  0.1933, -0.0511, -0.5979],
        ...0.1377],
        [ 0.8662,  1.3950, -1.0709,  0.0998, -0.0266],
        [ 0.4959, -0.6125, -1.3473, -0.0432,  0.5150]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________________ test_gin_conv[sum-g6-idtype1] ________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, aggregator_type = 'sum'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('aggregator_type', ['mean', 'max', 'sum'])
    def test_gin_conv(g, idtype, aggregator_type):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gin = nn.GINConv(
            th.nn.Linear(5, 12),
            aggregator_type
        )
        th.save(gin, tmp_buffer)
        feat = F.randn((g.number_of_src_nodes(), 5))
        gin = gin.to(ctx)
>       h = gin(g, feat)

tests\pytorch\test_nn.py:756: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\ginconv.py:147: in forward
    graph.update_all(aggregate_fn, _reducer('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03D55748>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-4.1125e-01, -1.3159e+00,  4.4266e-01, -1.9466e+00, -1.0117e+00],
        [-2.4040e-01, -2.3057e-01, -5.3927e...01,  2.3986e-01,  9.3775e-01, -1.6689e+00],
        [-4.9664e-01,  4.8946e-01, -5.9527e-01, -4.8277e-02,  1.8125e+00]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_gine_conv[g0-idtype0] __________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'hn': Scheme(shap...32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32), 'he': Scheme(shape=(5,), dtype=torch.float32)})
idtype = torch.int32

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    def test_gine_conv(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        gine = nn.GINEConv(
            th.nn.Linear(5, 12)
        )
        th.save(gine, tmp_buffer)
        nfeat = F.randn((g.number_of_src_nodes(), 5))
        efeat = F.randn((g.num_edges(), 5))
        gine = gine.to(ctx)
>       h = gine(g, nfeat, efeat)

tests\pytorch\test_nn.py:780: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gineconv.py:94: in forward
    graph.update_all(self.message, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:369: in message_passing
    ndata = invoke_gspmm(g, fn.copy_e(msg, msg), rfunc, edata=msgdata)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE12287148>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([[0.0000, 0.0000, 0.0000, 0.8256, 0.0299],
        [0.0000, 0.0000, 0.4868, 0.0000, 0.0000],
        [0.0000, 0..., 1.2652, 1.1702],
        [0.0000, 0.0000, 1.1649, 0.0000, 0.6106],
        [2.7970, 0.0000, 0.0000, 0.8303, 0.0000]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_gine_conv[g0-idtype1] __________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'hn': Scheme(shap...32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32), 'he': Scheme(shape=(5,), dtype=torch.float32)})
idtype = torch.int64

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    def test_gine_conv(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        gine = nn.GINEConv(
            th.nn.Linear(5, 12)
        )
        th.save(gine, tmp_buffer)
        nfeat = F.randn((g.number_of_src_nodes(), 5))
        efeat = F.randn((g.num_edges(), 5))
        gine = gine.to(ctx)
>       h = gine(g, nfeat, efeat)

tests\pytorch\test_nn.py:780: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gineconv.py:94: in forward
    graph.update_all(self.message, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:369: in message_passing
    ndata = invoke_gspmm(g, fn.copy_e(msg, msg), rfunc, edata=msgdata)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01545FC8>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([[1.7903, 2.6116, 0.0000, 0.0000, 1.1927],
        [2.0723, 2.7633, 0.0000, 0.0000, 0.6454],
        [1.6419, 0..., 0.0000, 0.0000],
        [0.0000, 0.8300, 0.6411, 0.0000, 0.4664],
        [1.0613, 0.0000, 1.4560, 1.9865, 1.3356]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_gine_conv[g1-idtype0] __________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'hn': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={'he': Scheme(shape=(5,), dtype=torch.float32)})
idtype = torch.int32

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    def test_gine_conv(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        gine = nn.GINEConv(
            th.nn.Linear(5, 12)
        )
        th.save(gine, tmp_buffer)
        nfeat = F.randn((g.number_of_src_nodes(), 5))
        efeat = F.randn((g.num_edges(), 5))
        gine = gine.to(ctx)
>       h = gine(g, nfeat, efeat)

tests\pytorch\test_nn.py:780: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gineconv.py:94: in forward
    graph.update_all(self.message, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:369: in message_passing
    ndata = invoke_gspmm(g, fn.copy_e(msg, msg), rfunc, edata=msgdata)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF070E70C8>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([[1.6155, 1.9008, 0.0000, 2.6108, 0.7356],
        [0.0000, 0.0000, 0.0000, 1.3317, 0.5397],
        [0.1237, 0..., 0.0000, 0.1963],
        [0.0000, 0.0000, 1.7064, 0.7801, 0.0000],
        [0.0000, 0.0000, 1.8982, 0.0120, 1.6840]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_gine_conv[g1-idtype1] __________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'hn': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={'he': Scheme(shape=(5,), dtype=torch.float32)})
idtype = torch.int64

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    def test_gine_conv(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        gine = nn.GINEConv(
            th.nn.Linear(5, 12)
        )
        th.save(gine, tmp_buffer)
        nfeat = F.randn((g.number_of_src_nodes(), 5))
        efeat = F.randn((g.num_edges(), 5))
        gine = gine.to(ctx)
>       h = gine(g, nfeat, efeat)

tests\pytorch\test_nn.py:780: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gineconv.py:94: in forward
    graph.update_all(self.message, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:369: in message_passing
    ndata = invoke_gspmm(g, fn.copy_e(msg, msg), rfunc, edata=msgdata)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01056C88>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([[0.0000, 0.0000, 0.0000, 0.3305, 0.0000],
        [1.3401, 3.5389, 0.0000, 1.2314, 0.0568],
        [0.5184, 0..., 2.3965, 0.7931],
        [0.0000, 1.9356, 0.0000, 0.0000, 0.0000],
        [0.0000, 1.4872, 0.0000, 0.7977, 0.0692]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_gine_conv[g2-idtype0] __________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'hn': Scheme(shap...      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), 'he': Scheme(shape=(5,), dtype=torch.float32)})
idtype = torch.int32

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    def test_gine_conv(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        gine = nn.GINEConv(
            th.nn.Linear(5, 12)
        )
        th.save(gine, tmp_buffer)
        nfeat = F.randn((g.number_of_src_nodes(), 5))
        efeat = F.randn((g.num_edges(), 5))
        gine = gine.to(ctx)
>       h = gine(g, nfeat, efeat)

tests\pytorch\test_nn.py:780: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gineconv.py:94: in forward
    graph.update_all(self.message, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:369: in message_passing
    ndata = invoke_gspmm(g, fn.copy_e(msg, msg), rfunc, edata=msgdata)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07246608>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([[1.5884, 0.0000, 0.4125, 0.1621, 0.0000],
        [0.0000, 0.0000, 1.1134, 1.9789, 0.0000],
        [0.6874, 0..., 1.0716, 0.0000],
        [0.0000, 0.0000, 0.0000, 0.0000, 3.2395],
        [0.0000, 1.4643, 0.0000, 0.0000, 0.0000]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_gine_conv[g2-idtype1] __________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'hn': Scheme(shap...      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), 'he': Scheme(shape=(5,), dtype=torch.float32)})
idtype = torch.int64

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    def test_gine_conv(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        gine = nn.GINEConv(
            th.nn.Linear(5, 12)
        )
        th.save(gine, tmp_buffer)
        nfeat = F.randn((g.number_of_src_nodes(), 5))
        efeat = F.randn((g.num_edges(), 5))
        gine = gine.to(ctx)
>       h = gine(g, nfeat, efeat)

tests\pytorch\test_nn.py:780: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gineconv.py:94: in forward
    graph.update_all(self.message, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:369: in message_passing
    ndata = invoke_gspmm(g, fn.copy_e(msg, msg), rfunc, edata=msgdata)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE12377708>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([[1.3293, 0.0000, 0.0000, 1.2173, 0.0000],
        [0.0000, 0.0000, 0.0000, 1.9565, 0.0000],
        [1.1898, 1..., 0.0000, 0.0000],
        [0.0000, 0.0000, 1.0648, 0.0000, 0.0000],
        [0.6323, 0.0000, 0.0000, 0.0000, 1.6663]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_gine_conv[g3-idtype0] __________________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'hn': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={'he': Scheme(shape=(5,), dtype=torch.float32)})
idtype = torch.int32

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    def test_gine_conv(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        gine = nn.GINEConv(
            th.nn.Linear(5, 12)
        )
        th.save(gine, tmp_buffer)
        nfeat = F.randn((g.number_of_src_nodes(), 5))
        efeat = F.randn((g.num_edges(), 5))
        gine = gine.to(ctx)
>       h = gine(g, nfeat, efeat)

tests\pytorch\test_nn.py:780: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gineconv.py:94: in forward
    graph.update_all(self.message, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:369: in message_passing
    ndata = invoke_gspmm(g, fn.copy_e(msg, msg), rfunc, edata=msgdata)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF073C2D88>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([[0.3667, 1.9691, 0.4603, 0.0000, 0.0000],
        [1.7481, 0.9016, 0.0000, 0.0000, 0.2810],
        [0.0000, 0..., 1.8820, 0.0000],
        [0.5558, 0.3186, 2.0039, 1.2265, 0.0000],
        [0.0000, 0.0000, 0.0000, 0.0000, 0.3794]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_gine_conv[g3-idtype1] __________________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'hn': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={'he': Scheme(shape=(5,), dtype=torch.float32)})
idtype = torch.int64

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    def test_gine_conv(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        gine = nn.GINEConv(
            th.nn.Linear(5, 12)
        )
        th.save(gine, tmp_buffer)
        nfeat = F.randn((g.number_of_src_nodes(), 5))
        efeat = F.randn((g.num_edges(), 5))
        gine = gine.to(ctx)
>       h = gine(g, nfeat, efeat)

tests\pytorch\test_nn.py:780: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gineconv.py:94: in forward
    graph.update_all(self.message, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:369: in message_passing
    ndata = invoke_gspmm(g, fn.copy_e(msg, msg), rfunc, edata=msgdata)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE122FA688>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([[1.0244, 0.0000, 0.0000, 1.4446, 0.0000],
        [2.8809, 0.0000, 2.4141, 0.0000, 0.0000],
        [2.1454, 0..., 0.0000, 0.0000],
        [2.4529, 0.3300, 0.7640, 0.0000, 0.5678],
        [3.2922, 0.0000, 0.0000, 0.0842, 0.9365]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_gine_conv[g4-idtype0] __________________________

g = Graph(num_nodes=5, num_edges=6,
      ndata_schemes={'hn': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={'he': Scheme(shape=(5,), dtype=torch.float32)})
idtype = torch.int32

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    def test_gine_conv(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        gine = nn.GINEConv(
            th.nn.Linear(5, 12)
        )
        th.save(gine, tmp_buffer)
        nfeat = F.randn((g.number_of_src_nodes(), 5))
        efeat = F.randn((g.num_edges(), 5))
        gine = gine.to(ctx)
>       h = gine(g, nfeat, efeat)

tests\pytorch\test_nn.py:780: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gineconv.py:94: in forward
    graph.update_all(self.message, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:369: in message_passing
    ndata = invoke_gspmm(g, fn.copy_e(msg, msg), rfunc, edata=msgdata)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1225B388>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([[0.0000, 2.6615, 0.0000, 0.0000, 0.0000],
        [0.0000, 0.1395, 0.0000, 0.9603, 0.0000],
        [0.0503, 2..., 0.3030, 0.2011],
        [2.0162, 0.8744, 0.8356, 0.0000, 0.0000],
        [0.0000, 0.7378, 0.0000, 0.0000, 0.0000]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_gine_conv[g4-idtype1] __________________________

g = Graph(num_nodes=5, num_edges=6,
      ndata_schemes={'hn': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={'he': Scheme(shape=(5,), dtype=torch.float32)})
idtype = torch.int64

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    def test_gine_conv(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        gine = nn.GINEConv(
            th.nn.Linear(5, 12)
        )
        th.save(gine, tmp_buffer)
        nfeat = F.randn((g.number_of_src_nodes(), 5))
        efeat = F.randn((g.num_edges(), 5))
        gine = gine.to(ctx)
>       h = gine(g, nfeat, efeat)

tests\pytorch\test_nn.py:780: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gineconv.py:94: in forward
    graph.update_all(self.message, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:369: in message_passing
    ndata = invoke_gspmm(g, fn.copy_e(msg, msg), rfunc, edata=msgdata)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0188C688>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([[0.5706, 0.0000, 0.0000, 0.6787, 0.0000],
        [0.2065, 0.0000, 0.0000, 0.0000, 0.0000],
        [0.0000, 0..., 1.2663, 0.0000],
        [0.0000, 0.7969, 0.8487, 0.7118, 1.7750],
        [0.0000, 0.0000, 0.9077, 0.0000, 0.0000]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_gine_conv[g5-idtype0] __________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'hn': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={'he': Scheme(shape=(5,), dtype=torch.float32)})
idtype = torch.int32

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    def test_gine_conv(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        gine = nn.GINEConv(
            th.nn.Linear(5, 12)
        )
        th.save(gine, tmp_buffer)
        nfeat = F.randn((g.number_of_src_nodes(), 5))
        efeat = F.randn((g.num_edges(), 5))
        gine = gine.to(ctx)
>       h = gine(g, nfeat, efeat)

tests\pytorch\test_nn.py:780: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gineconv.py:94: in forward
    graph.update_all(self.message, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:369: in message_passing
    ndata = invoke_gspmm(g, fn.copy_e(msg, msg), rfunc, edata=msgdata)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1231CFC8>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([[0.0477, 0.0000, 1.1184, 1.8947, 0.8979],
        [1.0122, 0.0000, 0.3587, 0.0000, 2.0935],
        [1.9103, 0..., 0.0000, 0.0000],
        [0.3241, 0.0000, 1.2479, 1.6853, 1.9686],
        [0.1791, 0.6173, 0.0000, 0.3224, 0.0966]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_gine_conv[g5-idtype1] __________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'hn': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={'he': Scheme(shape=(5,), dtype=torch.float32)})
idtype = torch.int64

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    def test_gine_conv(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        gine = nn.GINEConv(
            th.nn.Linear(5, 12)
        )
        th.save(gine, tmp_buffer)
        nfeat = F.randn((g.number_of_src_nodes(), 5))
        efeat = F.randn((g.num_edges(), 5))
        gine = gine.to(ctx)
>       h = gine(g, nfeat, efeat)

tests\pytorch\test_nn.py:780: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gineconv.py:94: in forward
    graph.update_all(self.message, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:369: in message_passing
    ndata = invoke_gspmm(g, fn.copy_e(msg, msg), rfunc, edata=msgdata)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE123608C8>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([[2.0392, 0.9983, 0.0000, 0.0000, 0.0000],
        [1.9022, 0.0000, 0.0000, 1.0003, 0.4638],
        [0.0000, 0..., 0.0000, 0.0000],
        [0.0000, 0.2917, 0.0000, 1.4858, 0.2600],
        [0.5705, 0.0000, 0.0000, 0.0102, 1.2529]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_gine_conv[g6-idtype0] __________________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int32

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    def test_gine_conv(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        gine = nn.GINEConv(
            th.nn.Linear(5, 12)
        )
        th.save(gine, tmp_buffer)
        nfeat = F.randn((g.number_of_src_nodes(), 5))
        efeat = F.randn((g.num_edges(), 5))
        gine = gine.to(ctx)
>       h = gine(g, nfeat, efeat)

tests\pytorch\test_nn.py:780: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gineconv.py:94: in forward
    graph.update_all(self.message, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:369: in message_passing
    ndata = invoke_gspmm(g, fn.copy_e(msg, msg), rfunc, edata=msgdata)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF06FF1B48>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([[0.0000, 0.0000, 0.9681, 0.1453, 0.0000],
        [0.6030, 0.0000, 0.0000, 0.7270, 0.0000],
        [0.0000, 0.0000, 0.7993, 0.4565, 0.1917]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_gine_conv[g6-idtype1] __________________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int64

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    def test_gine_conv(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        gine = nn.GINEConv(
            th.nn.Linear(5, 12)
        )
        th.save(gine, tmp_buffer)
        nfeat = F.randn((g.number_of_src_nodes(), 5))
        efeat = F.randn((g.num_edges(), 5))
        gine = gine.to(ctx)
>       h = gine(g, nfeat, efeat)

tests\pytorch\test_nn.py:780: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gineconv.py:94: in forward
    graph.update_all(self.message, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:369: in message_passing
    ndata = invoke_gspmm(g, fn.copy_e(msg, msg), rfunc, edata=msgdata)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07281A88>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([[0.7894, 1.5370, 1.3254, 0.0000, 0.0000],
        [2.0225, 0.0000, 2.1215, 0.6305, 0.0000],
        [0.0000, 0.0000, 0.0000, 0.1327, 2.2635]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_gine_conv[g7-idtype0] __________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'hn': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={'he': Scheme(shape=(5,), dtype=torch.float32)})
idtype = torch.int32

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    def test_gine_conv(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        gine = nn.GINEConv(
            th.nn.Linear(5, 12)
        )
        th.save(gine, tmp_buffer)
        nfeat = F.randn((g.number_of_src_nodes(), 5))
        efeat = F.randn((g.num_edges(), 5))
        gine = gine.to(ctx)
>       h = gine(g, nfeat, efeat)

tests\pytorch\test_nn.py:780: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gineconv.py:94: in forward
    graph.update_all(self.message, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:369: in message_passing
    ndata = invoke_gspmm(g, fn.copy_e(msg, msg), rfunc, edata=msgdata)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF014C4548>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([[0.0000, 0.0000, 0.2371, 0.3413, 0.0000],
        [0.0000, 0.1846, 0.1482, 0.0000, 0.0000],
        [0.0000, 0..., 0.0000, 0.0000],
        [0.6473, 0.0000, 0.6255, 0.0000, 0.0000],
        [0.0000, 0.0635, 0.0000, 0.0000, 0.0000]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_gine_conv[g7-idtype1] __________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'hn': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={'he': Scheme(shape=(5,), dtype=torch.float32)})
idtype = torch.int64

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite']))
    def test_gine_conv(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        gine = nn.GINEConv(
            th.nn.Linear(5, 12)
        )
        th.save(gine, tmp_buffer)
        nfeat = F.randn((g.number_of_src_nodes(), 5))
        efeat = F.randn((g.num_edges(), 5))
        gine = gine.to(ctx)
>       h = gine(g, nfeat, efeat)

tests\pytorch\test_nn.py:780: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gineconv.py:94: in forward
    graph.update_all(self.message, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:369: in message_passing
    ndata = invoke_gspmm(g, fn.copy_e(msg, msg), rfunc, edata=msgdata)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF016BBB88>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([[1.0612, 0.2476, 0.0000, 0.2346, 0.0000],
        [0.0000, 1.4963, 0.7653, 0.0000, 0.1703],
        [0.0000, 0..., 0.7863, 2.5533],
        [0.7673, 0.1044, 0.0000, 0.0000, 0.0000],
        [3.3851, 0.0000, 2.6849, 0.5741, 0.0000]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________________ test_gin_conv_bi[mean-g0-idtype0] ______________________

g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int32, aggregator_type = 'mean'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('aggregator_type', ['mean', 'max', 'sum'])
    def test_gin_conv_bi(g, idtype, aggregator_type):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gin = nn.GINConv(
            th.nn.Linear(5, 12),
            aggregator_type
        )
        feat = (F.randn((g.number_of_src_nodes(), 5)), F.randn((g.number_of_dst_nodes(), 5)))
        gin = gin.to(ctx)
>       h = gin(g, feat)

tests\pytorch\test_nn.py:803: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\ginconv.py:147: in forward
    graph.update_all(aggregate_fn, _reducer('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF017C2148>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.3758,  0.9582, -0.1891, -1.0195, -1.0334],
        [-1.0342,  0.4768,  1.0738,  0.3700,  0.2138]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________________ test_gin_conv_bi[mean-g0-idtype1] ______________________

g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int64, aggregator_type = 'mean'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('aggregator_type', ['mean', 'max', 'sum'])
    def test_gin_conv_bi(g, idtype, aggregator_type):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gin = nn.GINConv(
            th.nn.Linear(5, 12),
            aggregator_type
        )
        feat = (F.randn((g.number_of_src_nodes(), 5)), F.randn((g.number_of_dst_nodes(), 5)))
        gin = gin.to(ctx)
>       h = gin(g, feat)

tests\pytorch\test_nn.py:803: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\ginconv.py:147: in forward
    graph.update_all(aggregate_fn, _reducer('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF014F7588>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.5494, -0.9025,  0.0296,  1.7940, -1.3260],
        [-0.7833, -0.3094,  0.1658,  0.7158, -0.4555]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________________ test_gin_conv_bi[mean-g1-idtype0] ______________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int32
aggregator_type = 'mean'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('aggregator_type', ['mean', 'max', 'sum'])
    def test_gin_conv_bi(g, idtype, aggregator_type):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gin = nn.GINConv(
            th.nn.Linear(5, 12),
            aggregator_type
        )
        feat = (F.randn((g.number_of_src_nodes(), 5)), F.randn((g.number_of_dst_nodes(), 5)))
        gin = gin.to(ctx)
>       h = gin(g, feat)

tests\pytorch\test_nn.py:803: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\ginconv.py:147: in forward
    graph.update_all(aggregate_fn, _reducer('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE123080C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.3486,  1.7885,  1.6667, -1.0430,  0.2748],
        [-0.1044,  1.4744, -0.1749,  0.7598, -0.7891],
        ...0.7025],
        [-1.3281,  0.3173, -0.2705,  0.0836, -1.8137],
        [ 1.1399, -0.1268,  2.2136, -0.2804,  1.3412]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________________ test_gin_conv_bi[mean-g1-idtype1] ______________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int64
aggregator_type = 'mean'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('aggregator_type', ['mean', 'max', 'sum'])
    def test_gin_conv_bi(g, idtype, aggregator_type):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gin = nn.GINConv(
            th.nn.Linear(5, 12),
            aggregator_type
        )
        feat = (F.randn((g.number_of_src_nodes(), 5)), F.randn((g.number_of_dst_nodes(), 5)))
        gin = gin.to(ctx)
>       h = gin(g, feat)

tests\pytorch\test_nn.py:803: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\ginconv.py:147: in forward
    graph.update_all(aggregate_fn, _reducer('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03CBEB08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.2667,  0.6800, -0.3013, -1.3854, -1.0632],
        [-0.6010,  0.5442,  0.1422,  0.9451, -0.1827],
        ...0.7827],
        [ 0.7350, -0.3517, -0.9609, -0.6293, -0.1873],
        [ 1.5129,  0.0341, -1.7065, -0.0139, -0.1646]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________________ test_gin_conv_bi[max-g0-idtype0] _______________________

g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int32, aggregator_type = 'max'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('aggregator_type', ['mean', 'max', 'sum'])
    def test_gin_conv_bi(g, idtype, aggregator_type):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gin = nn.GINConv(
            th.nn.Linear(5, 12),
            aggregator_type
        )
        feat = (F.randn((g.number_of_src_nodes(), 5)), F.randn((g.number_of_dst_nodes(), 5)))
        gin = gin.to(ctx)
>       h = gin(g, feat)

tests\pytorch\test_nn.py:803: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\ginconv.py:147: in forward
    graph.update_all(aggregate_fn, _reducer('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF014F7E88>
op = 'copy_lhs', reduce_op = 'max'
lhs_data = tensor([[-0.9783,  0.7490,  1.7305,  0.7558,  0.9908],
        [-0.8629, -1.5892, -0.5262,  0.3558,  0.6287]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________________ test_gin_conv_bi[max-g0-idtype1] _______________________

g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int64, aggregator_type = 'max'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('aggregator_type', ['mean', 'max', 'sum'])
    def test_gin_conv_bi(g, idtype, aggregator_type):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gin = nn.GINConv(
            th.nn.Linear(5, 12),
            aggregator_type
        )
        feat = (F.randn((g.number_of_src_nodes(), 5)), F.randn((g.number_of_dst_nodes(), 5)))
        gin = gin.to(ctx)
>       h = gin(g, feat)

tests\pytorch\test_nn.py:803: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\ginconv.py:147: in forward
    graph.update_all(aggregate_fn, _reducer('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF070A68C8>
op = 'copy_lhs', reduce_op = 'max'
lhs_data = tensor([[ 0.1971,  0.1732, -1.0025,  0.5148,  0.1992],
        [ 0.8224,  0.9831, -0.5843,  0.2126,  1.2017]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________________ test_gin_conv_bi[max-g1-idtype0] _______________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int32
aggregator_type = 'max'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('aggregator_type', ['mean', 'max', 'sum'])
    def test_gin_conv_bi(g, idtype, aggregator_type):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gin = nn.GINConv(
            th.nn.Linear(5, 12),
            aggregator_type
        )
        feat = (F.randn((g.number_of_src_nodes(), 5)), F.randn((g.number_of_dst_nodes(), 5)))
        gin = gin.to(ctx)
>       h = gin(g, feat)

tests\pytorch\test_nn.py:803: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\ginconv.py:147: in forward
    graph.update_all(aggregate_fn, _reducer('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE12308F48>
op = 'copy_lhs', reduce_op = 'max'
lhs_data = tensor([[-1.1819, -0.0746, -0.6652, -0.5348, -1.4933],
        [-0.7484, -0.2941,  0.6376,  0.5217, -0.8410],
        ...0.0198],
        [-0.8757, -1.9040, -0.4350, -1.6887, -1.0629],
        [ 0.0620,  0.0948, -0.5812,  0.6093, -0.2238]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________________ test_gin_conv_bi[max-g1-idtype1] _______________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int64
aggregator_type = 'max'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('aggregator_type', ['mean', 'max', 'sum'])
    def test_gin_conv_bi(g, idtype, aggregator_type):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gin = nn.GINConv(
            th.nn.Linear(5, 12),
            aggregator_type
        )
        feat = (F.randn((g.number_of_src_nodes(), 5)), F.randn((g.number_of_dst_nodes(), 5)))
        gin = gin.to(ctx)
>       h = gin(g, feat)

tests\pytorch\test_nn.py:803: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\ginconv.py:147: in forward
    graph.update_all(aggregate_fn, _reducer('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07204F48>
op = 'copy_lhs', reduce_op = 'max'
lhs_data = tensor([[ 1.0092,  0.0657, -1.2470,  2.0574,  0.7177],
        [-1.4195, -1.1926, -0.3136,  0.0221,  0.0828],
        ...0.0683],
        [-1.3504,  1.8403, -0.0311,  0.3088, -0.2242],
        [ 0.2382, -1.4836, -0.2210, -0.3738, -0.9012]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________________ test_gin_conv_bi[sum-g0-idtype0] _______________________

g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int32, aggregator_type = 'sum'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('aggregator_type', ['mean', 'max', 'sum'])
    def test_gin_conv_bi(g, idtype, aggregator_type):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gin = nn.GINConv(
            th.nn.Linear(5, 12),
            aggregator_type
        )
        feat = (F.randn((g.number_of_src_nodes(), 5)), F.randn((g.number_of_dst_nodes(), 5)))
        gin = gin.to(ctx)
>       h = gin(g, feat)

tests\pytorch\test_nn.py:803: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\ginconv.py:147: in forward
    graph.update_all(aggregate_fn, _reducer('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01895508>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.0292, -0.4030,  0.0877, -0.4273,  0.8367],
        [ 0.3599,  0.1665,  0.5510, -0.6425,  0.1151]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________________ test_gin_conv_bi[sum-g0-idtype1] _______________________

g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int64, aggregator_type = 'sum'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('aggregator_type', ['mean', 'max', 'sum'])
    def test_gin_conv_bi(g, idtype, aggregator_type):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gin = nn.GINConv(
            th.nn.Linear(5, 12),
            aggregator_type
        )
        feat = (F.randn((g.number_of_src_nodes(), 5)), F.randn((g.number_of_dst_nodes(), 5)))
        gin = gin.to(ctx)
>       h = gin(g, feat)

tests\pytorch\test_nn.py:803: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\ginconv.py:147: in forward
    graph.update_all(aggregate_fn, _reducer('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF018CE0C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.8157, -0.4870,  0.9038,  0.0176,  1.4160],
        [-0.6496,  0.2211, -1.6208, -1.5805, -0.3934]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________________ test_gin_conv_bi[sum-g1-idtype0] _______________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int32
aggregator_type = 'sum'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('aggregator_type', ['mean', 'max', 'sum'])
    def test_gin_conv_bi(g, idtype, aggregator_type):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gin = nn.GINConv(
            th.nn.Linear(5, 12),
            aggregator_type
        )
        feat = (F.randn((g.number_of_src_nodes(), 5)), F.randn((g.number_of_dst_nodes(), 5)))
        gin = gin.to(ctx)
>       h = gin(g, feat)

tests\pytorch\test_nn.py:803: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\ginconv.py:147: in forward
    graph.update_all(aggregate_fn, _reducer('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF072927C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.1510,  0.0308,  1.0674, -0.4953,  0.0634],
        [-1.1121,  1.2815,  1.5259,  1.1717,  0.3433],
        ...0.2117],
        [ 0.5163, -0.6465, -0.4302,  0.2935, -1.4722],
        [-0.1389, -0.4731,  0.7472, -0.6184, -0.5542]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________________ test_gin_conv_bi[sum-g1-idtype1] _______________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int64
aggregator_type = 'sum'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('aggregator_type', ['mean', 'max', 'sum'])
    def test_gin_conv_bi(g, idtype, aggregator_type):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gin = nn.GINConv(
            th.nn.Linear(5, 12),
            aggregator_type
        )
        feat = (F.randn((g.number_of_src_nodes(), 5)), F.randn((g.number_of_dst_nodes(), 5)))
        gin = gin.to(ctx)
>       h = gin(g, feat)

tests\pytorch\test_nn.py:803: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\ginconv.py:147: in forward
    graph.update_all(aggregate_fn, _reducer('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03CBD248>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.9048,  1.1661,  0.7406, -1.3596, -0.7848],
        [ 1.1724,  1.0570, -0.9294,  0.4843, -0.8379],
        ...0.9104],
        [-0.3894, -0.3575, -0.9595, -1.4570,  1.5026],
        [-0.0308, -0.7701, -0.8868, -1.0777,  0.3063]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_agnn_conv[g0-idtype0] __________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32), 'norm_h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
idtype = torch.int32

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    def test_agnn_conv(g, idtype):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        agnn = nn.AGNNConv(1)
        feat = F.randn((g.number_of_src_nodes(), 5))
        agnn = agnn.to(ctx)
>       h = agnn(g, feat)

tests\pytorch\test_nn.py:814: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\agnnconv.py:153: in forward
    graph.apply_edges(fn.u_dot_v('norm_h', 'norm_h', 'cos'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE12351348>
op = 'dot'
lhs_data = tensor([[ 0.4286, -0.5061,  0.4009, -0.6207, -0.1192],
        [ 0.5157,  0.2645, -0.3096, -0.4872,  0.5752],
        ...0.8080],
        [-0.7636,  0.0049, -0.3576, -0.4648, -0.2701],
        [ 0.1654,  0.5234,  0.0197, -0.7705,  0.3234]])
rhs_data = tensor([[ 0.4286, -0.5061,  0.4009, -0.6207, -0.1192],
        [ 0.5157,  0.2645, -0.3096, -0.4872,  0.5752],
        ...0.8080],
        [-0.7636,  0.0049, -0.3576, -0.4648, -0.2701],
        [ 0.1654,  0.5234,  0.0197, -0.7705,  0.3234]])
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_________________________ test_agnn_conv[g0-idtype1] __________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32), 'norm_h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
idtype = torch.int64

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    def test_agnn_conv(g, idtype):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        agnn = nn.AGNNConv(1)
        feat = F.randn((g.number_of_src_nodes(), 5))
        agnn = agnn.to(ctx)
>       h = agnn(g, feat)

tests\pytorch\test_nn.py:814: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\agnnconv.py:153: in forward
    graph.apply_edges(fn.u_dot_v('norm_h', 'norm_h', 'cos'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF072A8048>
op = 'dot'
lhs_data = tensor([[-0.6243, -0.0730, -0.2763, -0.1868,  0.7026],
        [ 0.2801,  0.4135, -0.4730,  0.1745,  0.7045],
        ...0.0982],
        [-0.5889, -0.2168,  0.3906, -0.2736,  0.6154],
        [-0.1575,  0.1891,  0.9489,  0.0608, -0.1881]])
rhs_data = tensor([[-0.6243, -0.0730, -0.2763, -0.1868,  0.7026],
        [ 0.2801,  0.4135, -0.4730,  0.1745,  0.7045],
        ...0.0982],
        [-0.5889, -0.2168,  0.3906, -0.2736,  0.6154],
        [-0.1575,  0.1891,  0.9489,  0.0608, -0.1881]])
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_________________________ test_agnn_conv[g1-idtype0] __________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32), 'norm_h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    def test_agnn_conv(g, idtype):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        agnn = nn.AGNNConv(1)
        feat = F.randn((g.number_of_src_nodes(), 5))
        agnn = agnn.to(ctx)
>       h = agnn(g, feat)

tests\pytorch\test_nn.py:814: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\agnnconv.py:153: in forward
    graph.apply_edges(fn.u_dot_v('norm_h', 'norm_h', 'cos'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0727AEC8>
op = 'dot'
lhs_data = tensor([[ 0.0038,  0.5131,  0.7743,  0.1452, -0.3407],
        [-0.8095,  0.5654, -0.1402,  0.0566,  0.0479],
        ...0.0758],
        [-0.0254, -0.2235,  0.6427, -0.6270, -0.3785],
        [ 0.2479,  0.8567,  0.1856,  0.3713,  0.1796]])
rhs_data = tensor([[ 0.0038,  0.5131,  0.7743,  0.1452, -0.3407],
        [-0.8095,  0.5654, -0.1402,  0.0566,  0.0479],
        ...0.0758],
        [-0.0254, -0.2235,  0.6427, -0.6270, -0.3785],
        [ 0.2479,  0.8567,  0.1856,  0.3713,  0.1796]])
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_________________________ test_agnn_conv[g1-idtype1] __________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32), 'norm_h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    def test_agnn_conv(g, idtype):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        agnn = nn.AGNNConv(1)
        feat = F.randn((g.number_of_src_nodes(), 5))
        agnn = agnn.to(ctx)
>       h = agnn(g, feat)

tests\pytorch\test_nn.py:814: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\agnnconv.py:153: in forward
    graph.apply_edges(fn.u_dot_v('norm_h', 'norm_h', 'cos'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0180C848>
op = 'dot'
lhs_data = tensor([[-0.5450, -0.1808,  0.5547,  0.1318, -0.5876],
        [ 0.5856, -0.5868,  0.4113, -0.3571,  0.1268],
        ...0.3070],
        [ 0.3371,  0.5547, -0.7204, -0.2011,  0.1392],
        [ 0.1642,  0.0729, -0.2250,  0.8531, -0.4351]])
rhs_data = tensor([[-0.5450, -0.1808,  0.5547,  0.1318, -0.5876],
        [ 0.5856, -0.5868,  0.4113, -0.3571,  0.1268],
        ...0.3070],
        [ 0.3371,  0.5547, -0.7204, -0.2011,  0.1392],
        [ 0.1642,  0.0729, -0.2250,  0.8531, -0.4351]])
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_________________________ test_agnn_conv[g2-idtype0] __________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32), 'norm_h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int32

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    def test_agnn_conv(g, idtype):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        agnn = nn.AGNNConv(1)
        feat = F.randn((g.number_of_src_nodes(), 5))
        agnn = agnn.to(ctx)
>       h = agnn(g, feat)

tests\pytorch\test_nn.py:814: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\agnnconv.py:153: in forward
    graph.apply_edges(fn.u_dot_v('norm_h', 'norm_h', 'cos'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07284B08>
op = 'dot'
lhs_data = tensor([[-0.6645,  0.5695, -0.3784, -0.1809, -0.2414],
        [ 0.2985, -0.7870,  0.2828,  0.2643, -0.3764],
        ...0.5883],
        [ 0.2286,  0.1554,  0.0372,  0.9457, -0.1668],
        [-0.0631,  0.0087, -0.8602,  0.3612,  0.3543]])
rhs_data = tensor([[-0.6645,  0.5695, -0.3784, -0.1809, -0.2414],
        [ 0.2985, -0.7870,  0.2828,  0.2643, -0.3764],
        ...0.5883],
        [ 0.2286,  0.1554,  0.0372,  0.9457, -0.1668],
        [-0.0631,  0.0087, -0.8602,  0.3612,  0.3543]])
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_________________________ test_agnn_conv[g2-idtype1] __________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32), 'norm_h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int64

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    def test_agnn_conv(g, idtype):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        agnn = nn.AGNNConv(1)
        feat = F.randn((g.number_of_src_nodes(), 5))
        agnn = agnn.to(ctx)
>       h = agnn(g, feat)

tests\pytorch\test_nn.py:814: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\agnnconv.py:153: in forward
    graph.apply_edges(fn.u_dot_v('norm_h', 'norm_h', 'cos'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF017E5E48>
op = 'dot'
lhs_data = tensor([[ 0.3846,  0.6352, -0.6198,  0.2259,  0.1157],
        [ 0.1377,  0.3615, -0.5599,  0.7189,  0.1415],
        ...0.8033],
        [ 0.3070,  0.0908,  0.8182,  0.4413, -0.1826],
        [-0.6200, -0.1613, -0.4019, -0.4914,  0.4320]])
rhs_data = tensor([[ 0.3846,  0.6352, -0.6198,  0.2259,  0.1157],
        [ 0.1377,  0.3615, -0.5599,  0.7189,  0.1415],
        ...0.8033],
        [ 0.3070,  0.0908,  0.8182,  0.4413, -0.1826],
        [-0.6200, -0.1613, -0.4019, -0.4914,  0.4320]])
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_________________________ test_agnn_conv[g3-idtype0] __________________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32), 'norm_h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    def test_agnn_conv(g, idtype):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        agnn = nn.AGNNConv(1)
        feat = F.randn((g.number_of_src_nodes(), 5))
        agnn = agnn.to(ctx)
>       h = agnn(g, feat)

tests\pytorch\test_nn.py:814: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\agnnconv.py:153: in forward
    graph.apply_edges(fn.u_dot_v('norm_h', 'norm_h', 'cos'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF070CA6C8>
op = 'dot'
lhs_data = tensor([[ 0.1998,  0.6931, -0.4464,  0.5090, -0.1461],
        [ 0.3102, -0.0704, -0.1842,  0.2706,  0.8898],
        ...0.1013],
        [ 0.3624,  0.3080,  0.0491,  0.2483,  0.8425],
        [-0.7690,  0.3942, -0.2130,  0.4483,  0.0835]])
rhs_data = tensor([[ 0.1998,  0.6931, -0.4464,  0.5090, -0.1461],
        [ 0.3102, -0.0704, -0.1842,  0.2706,  0.8898],
        ...0.1013],
        [ 0.3624,  0.3080,  0.0491,  0.2483,  0.8425],
        [-0.7690,  0.3942, -0.2130,  0.4483,  0.0835]])
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_________________________ test_agnn_conv[g3-idtype1] __________________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32), 'norm_h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    def test_agnn_conv(g, idtype):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        agnn = nn.AGNNConv(1)
        feat = F.randn((g.number_of_src_nodes(), 5))
        agnn = agnn.to(ctx)
>       h = agnn(g, feat)

tests\pytorch\test_nn.py:814: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\agnnconv.py:153: in forward
    graph.apply_edges(fn.u_dot_v('norm_h', 'norm_h', 'cos'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE122EB8C8>
op = 'dot'
lhs_data = tensor([[-0.1265, -0.7315,  0.5625, -0.3529, -0.0896],
        [-0.5474,  0.1471,  0.7415,  0.0966,  0.3458],
        ...0.0119],
        [-0.2535, -0.1942,  0.7298,  0.1781,  0.5776],
        [ 0.7240,  0.2574,  0.0281, -0.0540, -0.6371]])
rhs_data = tensor([[-0.1265, -0.7315,  0.5625, -0.3529, -0.0896],
        [-0.5474,  0.1471,  0.7415,  0.0966,  0.3458],
        ...0.0119],
        [-0.2535, -0.1942,  0.7298,  0.1781,  0.5776],
        [ 0.7240,  0.2574,  0.0281, -0.0540, -0.6371]])
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_________________________ test_agnn_conv[g4-idtype0] __________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32), 'norm_h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    def test_agnn_conv(g, idtype):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        agnn = nn.AGNNConv(1)
        feat = F.randn((g.number_of_src_nodes(), 5))
        agnn = agnn.to(ctx)
>       h = agnn(g, feat)

tests\pytorch\test_nn.py:814: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\agnnconv.py:153: in forward
    graph.apply_edges(fn.u_dot_v('norm_h', 'norm_h', 'cos'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07292488>
op = 'dot'
lhs_data = tensor([[-0.2267,  0.8288, -0.5061, -0.0075, -0.0735],
        [ 0.2321, -0.5285,  0.3473, -0.2443,  0.6975],
        ...0.6429],
        [-0.2003,  0.3505,  0.8049,  0.1947,  0.3890],
        [-0.4860,  0.1049,  0.7086,  0.2431,  0.4377]])
rhs_data = tensor([[-0.2267,  0.8288, -0.5061, -0.0075, -0.0735],
        [ 0.2321, -0.5285,  0.3473, -0.2443,  0.6975],
        ...0.6429],
        [-0.2003,  0.3505,  0.8049,  0.1947,  0.3890],
        [-0.4860,  0.1049,  0.7086,  0.2431,  0.4377]])
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_________________________ test_agnn_conv[g4-idtype1] __________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32), 'norm_h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    def test_agnn_conv(g, idtype):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        agnn = nn.AGNNConv(1)
        feat = F.randn((g.number_of_src_nodes(), 5))
        agnn = agnn.to(ctx)
>       h = agnn(g, feat)

tests\pytorch\test_nn.py:814: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\agnnconv.py:153: in forward
    graph.apply_edges(fn.u_dot_v('norm_h', 'norm_h', 'cos'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1239F648>
op = 'dot'
lhs_data = tensor([[-0.6677,  0.2939,  0.0287, -0.0568, -0.6810],
        [ 0.2680, -0.8927,  0.2729, -0.2278,  0.0696],
        ...0.1314],
        [ 0.3086,  0.0765, -0.1365,  0.6202, -0.7040],
        [ 0.2117,  0.0861, -0.6551, -0.7159, -0.0782]])
rhs_data = tensor([[-0.6677,  0.2939,  0.0287, -0.0568, -0.6810],
        [ 0.2680, -0.8927,  0.2729, -0.2278,  0.0696],
        ...0.1314],
        [ 0.3086,  0.0765, -0.1365,  0.6202, -0.7040],
        [ 0.2117,  0.0861, -0.6551, -0.7159, -0.0782]])
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_________________________ test_agnn_conv[g5-idtype0] __________________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int32

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    def test_agnn_conv(g, idtype):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        agnn = nn.AGNNConv(1)
        feat = F.randn((g.number_of_src_nodes(), 5))
        agnn = agnn.to(ctx)
>       h = agnn(g, feat)

tests\pytorch\test_nn.py:814: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\agnnconv.py:153: in forward
    graph.apply_edges(fn.u_dot_v('norm_h', 'norm_h', 'cos'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0714BB88>
op = 'dot'
lhs_data = tensor([[-0.3977,  0.2251, -0.8785, -0.0774,  0.1164],
        [-0.4034,  0.6895, -0.2428,  0.1087, -0.5395],
        ...0.0890],
        [-0.1768, -0.8609,  0.2600, -0.3795,  0.1266],
        [ 0.3800,  0.6862, -0.2795,  0.0117,  0.5535]])
rhs_data = tensor([[-0.3977,  0.2251, -0.8785, -0.0774,  0.1164],
        [-0.4034,  0.6895, -0.2428,  0.1087, -0.5395],
        [ 0.8165, -0.1868,  0.4614,  0.2877,  0.0517]])
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_________________________ test_agnn_conv[g5-idtype1] __________________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int64

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    def test_agnn_conv(g, idtype):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        agnn = nn.AGNNConv(1)
        feat = F.randn((g.number_of_src_nodes(), 5))
        agnn = agnn.to(ctx)
>       h = agnn(g, feat)

tests\pytorch\test_nn.py:814: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\agnnconv.py:153: in forward
    graph.apply_edges(fn.u_dot_v('norm_h', 'norm_h', 'cos'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01723A48>
op = 'dot'
lhs_data = tensor([[ 0.6973,  0.0855, -0.1773, -0.5932, -0.3510],
        [-0.4243,  0.5767, -0.0593, -0.6103,  0.3336],
        ...0.3158],
        [-0.7679, -0.3975, -0.0985,  0.3186,  0.3757],
        [ 0.7443, -0.5032,  0.3005,  0.2271,  0.2257]])
rhs_data = tensor([[ 0.6973,  0.0855, -0.1773, -0.5932, -0.3510],
        [-0.4243,  0.5767, -0.0593, -0.6103,  0.3336],
        [-0.1072,  0.0329, -0.5464, -0.5862, -0.5876]])
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_________________________ test_agnn_conv[g6-idtype0] __________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32), 'norm_h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    def test_agnn_conv(g, idtype):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        agnn = nn.AGNNConv(1)
        feat = F.randn((g.number_of_src_nodes(), 5))
        agnn = agnn.to(ctx)
>       h = agnn(g, feat)

tests\pytorch\test_nn.py:814: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\agnnconv.py:153: in forward
    graph.apply_edges(fn.u_dot_v('norm_h', 'norm_h', 'cos'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF018850C8>
op = 'dot'
lhs_data = tensor([[-0.9683, -0.1114, -0.0521,  0.0175, -0.2166],
        [ 0.3572, -0.7885,  0.2879, -0.3184,  0.2576],
        ...0.3020],
        [-0.3313, -0.0581, -0.4103, -0.2573, -0.8077],
        [ 0.7230,  0.4639,  0.0691,  0.3755,  0.3411]])
rhs_data = tensor([[-0.9683, -0.1114, -0.0521,  0.0175, -0.2166],
        [ 0.3572, -0.7885,  0.2879, -0.3184,  0.2576],
        ...0.3020],
        [-0.3313, -0.0581, -0.4103, -0.2573, -0.8077],
        [ 0.7230,  0.4639,  0.0691,  0.3755,  0.3411]])
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_________________________ test_agnn_conv[g6-idtype1] __________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5,), dtype=torch.float32), 'norm_h': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    def test_agnn_conv(g, idtype):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        agnn = nn.AGNNConv(1)
        feat = F.randn((g.number_of_src_nodes(), 5))
        agnn = agnn.to(ctx)
>       h = agnn(g, feat)

tests\pytorch\test_nn.py:814: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\agnnconv.py:153: in forward
    graph.apply_edges(fn.u_dot_v('norm_h', 'norm_h', 'cos'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0195C488>
op = 'dot'
lhs_data = tensor([[-0.4028,  0.1882, -0.0703,  0.6326, -0.6303],
        [ 0.7948,  0.1143,  0.3665, -0.1892,  0.4302],
        ...0.3724],
        [-0.0222, -0.3507,  0.2915, -0.8669, -0.2002],
        [-0.2782, -0.5298, -0.3185, -0.6043, -0.4187]])
rhs_data = tensor([[-0.4028,  0.1882, -0.0703,  0.6326, -0.6303],
        [ 0.7948,  0.1143,  0.3665, -0.1892,  0.4302],
        ...0.3724],
        [-0.0222, -0.3507,  0.2915, -0.8669, -0.2002],
        [-0.2782, -0.5298, -0.3185, -0.6043, -0.4187]])
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
________________________ test_agnn_conv_bi[g0-idtype0] ________________________

g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int32

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    def test_agnn_conv_bi(g, idtype):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        agnn = nn.AGNNConv(1)
        feat = (F.randn((g.number_of_src_nodes(), 5)), F.randn((g.number_of_dst_nodes(), 5)))
        agnn = agnn.to(ctx)
>       h = agnn(g, feat)

tests\pytorch\test_nn.py:825: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\agnnconv.py:153: in forward
    graph.apply_edges(fn.u_dot_v('norm_h', 'norm_h', 'cos'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE12300F88>
op = 'dot'
lhs_data = tensor([[ 0.0225,  0.9611,  0.0094,  0.2695,  0.0559],
        [-0.0736, -0.6987,  0.5495, -0.2351, -0.3863]])
rhs_data = tensor([[ 0.0939,  0.5455, -0.2364, -0.4320, -0.6717],
        [ 0.2838,  0.1115, -0.1927,  0.1573, -0.9193],
        [ 0.4359, -0.4086,  0.6233,  0.1291, -0.4877],
        [-0.3743,  0.7116,  0.3030, -0.4810,  0.1743]])
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
________________________ test_agnn_conv_bi[g0-idtype1] ________________________

g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int64

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    def test_agnn_conv_bi(g, idtype):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        agnn = nn.AGNNConv(1)
        feat = (F.randn((g.number_of_src_nodes(), 5)), F.randn((g.number_of_dst_nodes(), 5)))
        agnn = agnn.to(ctx)
>       h = agnn(g, feat)

tests\pytorch\test_nn.py:825: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\agnnconv.py:153: in forward
    graph.apply_edges(fn.u_dot_v('norm_h', 'norm_h', 'cos'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0718ED08>
op = 'dot'
lhs_data = tensor([[ 0.7318, -0.2515,  0.5226, -0.0065, -0.3578],
        [-0.3914, -0.0922, -0.2712,  0.7205,  0.4957]])
rhs_data = tensor([[-0.0176,  0.5481,  0.2368,  0.6908,  0.4074],
        [-0.1797, -0.6486,  0.1299,  0.6318,  0.3619],
        [ 0.4815,  0.2730, -0.0349, -0.1926,  0.8095],
        [-0.3582,  0.6670,  0.4827, -0.0349,  0.4387]])
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
________________________ test_agnn_conv_bi[g1-idtype0] ________________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int32

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    def test_agnn_conv_bi(g, idtype):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        agnn = nn.AGNNConv(1)
        feat = (F.randn((g.number_of_src_nodes(), 5)), F.randn((g.number_of_dst_nodes(), 5)))
        agnn = agnn.to(ctx)
>       h = agnn(g, feat)

tests\pytorch\test_nn.py:825: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\agnnconv.py:153: in forward
    graph.apply_edges(fn.u_dot_v('norm_h', 'norm_h', 'cos'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF014728C8>
op = 'dot'
lhs_data = tensor([[-0.2893, -0.2946,  0.3116, -0.6415,  0.5665],
        [ 0.1616,  0.1738,  0.1269,  0.0087, -0.9631],
        ...0.4032],
        [-0.4777, -0.5367, -0.4411,  0.3914, -0.3688],
        [-0.1720, -0.7269, -0.3300,  0.3665,  0.4460]])
rhs_data = tensor([[-0.3404, -0.5240, -0.1260, -0.7705,  0.0011],
        [ 0.0942, -0.3938,  0.2178, -0.8228,  0.3340],
        [ 0.0639, -0.1988,  0.0763,  0.6668, -0.7113]])
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
________________________ test_agnn_conv_bi[g1-idtype1] ________________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int64

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    def test_agnn_conv_bi(g, idtype):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        agnn = nn.AGNNConv(1)
        feat = (F.randn((g.number_of_src_nodes(), 5)), F.randn((g.number_of_dst_nodes(), 5)))
        agnn = agnn.to(ctx)
>       h = agnn(g, feat)

tests\pytorch\test_nn.py:825: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\agnnconv.py:153: in forward
    graph.apply_edges(fn.u_dot_v('norm_h', 'norm_h', 'cos'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF030D7D48>
op = 'dot'
lhs_data = tensor([[ 0.0763,  0.6179,  0.7062,  0.2729,  0.1978],
        [-0.0653, -0.9371, -0.2681,  0.2031,  0.0670],
        ...0.2984],
        [-0.9346, -0.1018,  0.0289,  0.2653,  0.2121],
        [ 0.3888, -0.3734, -0.4271,  0.4352,  0.5810]])
rhs_data = tensor([[-3.1542e-01, -1.5139e-01,  2.1705e-01,  1.4768e-02, -9.1119e-01],
        [-2.5411e-01,  8.1217e-01, -1.1445e-01, -2.8103e-01,  4.2864e-01],
        [ 2.6396e-04,  2.4084e-01, -6.0702e-01, -5.4679e-01,  5.2397e-01]])
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_gated_graph_conv[g0-idtype0] ______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(10,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32), 'W_e*h': Scheme(shape=(10,), dtype=torch.float32)})
idtype = torch.int32

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_gated_graph_conv(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        ggconv = nn.GatedGraphConv(5, 10, 5, 3)
        etypes = th.arange(g.number_of_edges()) % 3
        feat = F.randn((g.number_of_nodes(), 5))
        ggconv = ggconv.to(ctx)
        etypes = etypes.to(ctx)
    
>       h = ggconv(g, feat, etypes)

tests\pytorch\test_nn.py:839: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatedgraphconv.py:166: in forward
    graph.update_all(fn.copy_e('W_e*h', 'm'), fn.sum('m', 'a'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0173C1C8>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([[ 1.9701, -1.5809, -0.9174,  1.1354, -2.0402, -2.6653,  0.3838, -0.1625,
          0.7666,  0.8643],
        [...0.1865, -2.9586, -1.7290, -0.2510,  3.2699,  0.5544, -0.9090,
          1.6071,  0.8992]], grad_fn=<IndexCopyBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________________ test_gated_graph_conv[g0-idtype1] ______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(10,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32), 'W_e*h': Scheme(shape=(10,), dtype=torch.float32)})
idtype = torch.int64

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_gated_graph_conv(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        ggconv = nn.GatedGraphConv(5, 10, 5, 3)
        etypes = th.arange(g.number_of_edges()) % 3
        feat = F.randn((g.number_of_nodes(), 5))
        ggconv = ggconv.to(ctx)
        etypes = etypes.to(ctx)
    
>       h = ggconv(g, feat, etypes)

tests\pytorch\test_nn.py:839: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatedgraphconv.py:166: in forward
    graph.update_all(fn.copy_e('W_e*h', 'm'), fn.sum('m', 'a'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07113708>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([[ 1.0021, -1.2659,  0.7103, -0.0434,  1.8240,  0.2811, -0.3031, -0.7373,
         -0.1379, -1.0032],
        [...0.3836, -0.4799,  1.2181, -1.2509,  0.0240, -1.1274, -0.3704,
         -0.3235, -0.2490]], grad_fn=<IndexCopyBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________________ test_gated_graph_conv[g1-idtype0] ______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(10,), dtype=torch.float32)}
      edata_schemes={'W_e*h': Scheme(shape=(10,), dtype=torch.float32)})
idtype = torch.int32

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_gated_graph_conv(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        ggconv = nn.GatedGraphConv(5, 10, 5, 3)
        etypes = th.arange(g.number_of_edges()) % 3
        feat = F.randn((g.number_of_nodes(), 5))
        ggconv = ggconv.to(ctx)
        etypes = etypes.to(ctx)
    
>       h = ggconv(g, feat, etypes)

tests\pytorch\test_nn.py:839: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatedgraphconv.py:166: in forward
    graph.update_all(fn.copy_e('W_e*h', 'm'), fn.sum('m', 'a'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0723C1C8>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([[-2.5566, -0.6239, -1.9764,  1.5942,  0.0807, -3.3504,  0.9141, -1.7515,
          0.8498, -0.7732],
        [...1.0434,  0.0997, -0.0060, -0.0236,  0.1211,  0.2912, -1.6525,
         -1.1316, -0.4106]], grad_fn=<IndexCopyBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________________ test_gated_graph_conv[g1-idtype1] ______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(10,), dtype=torch.float32)}
      edata_schemes={'W_e*h': Scheme(shape=(10,), dtype=torch.float32)})
idtype = torch.int64

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_gated_graph_conv(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        ggconv = nn.GatedGraphConv(5, 10, 5, 3)
        etypes = th.arange(g.number_of_edges()) % 3
        feat = F.randn((g.number_of_nodes(), 5))
        ggconv = ggconv.to(ctx)
        etypes = etypes.to(ctx)
    
>       h = ggconv(g, feat, etypes)

tests\pytorch\test_nn.py:839: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatedgraphconv.py:166: in forward
    graph.update_all(fn.copy_e('W_e*h', 'm'), fn.sum('m', 'a'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1228A848>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([[ 3.3253e-02, -2.2311e-01, -9.8281e-01,  6.7130e-01, -2.9924e-02,
         -1.1228e+00, -1.9009e-03,  7.5394e-...42e-01,
         -2.5462e-01, -3.9647e-01,  4.9793e-01,  4.5027e-02, -9.9199e-01]],
       grad_fn=<IndexCopyBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________________ test_gated_graph_conv[g2-idtype0] ______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(10,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), 'W_e*h': Scheme(shape=(10,), dtype=torch.float32)})
idtype = torch.int32

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_gated_graph_conv(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        ggconv = nn.GatedGraphConv(5, 10, 5, 3)
        etypes = th.arange(g.number_of_edges()) % 3
        feat = F.randn((g.number_of_nodes(), 5))
        ggconv = ggconv.to(ctx)
        etypes = etypes.to(ctx)
    
>       h = ggconv(g, feat, etypes)

tests\pytorch\test_nn.py:839: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatedgraphconv.py:166: in forward
    graph.update_all(fn.copy_e('W_e*h', 'm'), fn.sum('m', 'a'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07036BC8>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([[ 5.5361e-01, -1.5942e-01, -1.3038e+00, -4.9414e-02, -4.2178e-01,
         -1.4795e+00, -5.2905e-01, -1.7544e-...04e+00,
          1.1562e+00, -1.9064e+00,  1.1469e+00,  1.1027e+00,  2.0168e-02]],
       grad_fn=<IndexCopyBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________________ test_gated_graph_conv[g2-idtype1] ______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(10,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), 'W_e*h': Scheme(shape=(10,), dtype=torch.float32)})
idtype = torch.int64

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_gated_graph_conv(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        ggconv = nn.GatedGraphConv(5, 10, 5, 3)
        etypes = th.arange(g.number_of_edges()) % 3
        feat = F.randn((g.number_of_nodes(), 5))
        ggconv = ggconv.to(ctx)
        etypes = etypes.to(ctx)
    
>       h = ggconv(g, feat, etypes)

tests\pytorch\test_nn.py:839: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatedgraphconv.py:166: in forward
    graph.update_all(fn.copy_e('W_e*h', 'm'), fn.sum('m', 'a'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01756888>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([[ 0.5290, -3.4911,  0.6114, -0.0053, -1.8625, -1.4583,  2.6483,  0.3852,
         -1.3438, -1.0683],
        [...0.0120, -1.2531, -0.1098, -0.3702, -0.9222, -1.0065, -0.5864,
         -1.6965, -0.5853]], grad_fn=<IndexCopyBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________________ test_gated_graph_conv[g3-idtype0] ______________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(10,), dtype=torch.float32)}
      edata_schemes={'W_e*h': Scheme(shape=(10,), dtype=torch.float32)})
idtype = torch.int32

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_gated_graph_conv(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        ggconv = nn.GatedGraphConv(5, 10, 5, 3)
        etypes = th.arange(g.number_of_edges()) % 3
        feat = F.randn((g.number_of_nodes(), 5))
        ggconv = ggconv.to(ctx)
        etypes = etypes.to(ctx)
    
>       h = ggconv(g, feat, etypes)

tests\pytorch\test_nn.py:839: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatedgraphconv.py:166: in forward
    graph.update_all(fn.copy_e('W_e*h', 'm'), fn.sum('m', 'a'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF018B4808>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([[-0.3450,  0.4842,  0.1922,  0.0640, -0.3673,  0.7398, -1.1455, -1.0007,
          0.4845,  2.0563],
        [...0.3085, -0.0098, -1.4459, -0.1898, -0.5901, -0.6204,  0.6072,
          1.3917,  1.1367]], grad_fn=<IndexCopyBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________________ test_gated_graph_conv[g3-idtype1] ______________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(10,), dtype=torch.float32)}
      edata_schemes={'W_e*h': Scheme(shape=(10,), dtype=torch.float32)})
idtype = torch.int64

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_gated_graph_conv(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        ggconv = nn.GatedGraphConv(5, 10, 5, 3)
        etypes = th.arange(g.number_of_edges()) % 3
        feat = F.randn((g.number_of_nodes(), 5))
        ggconv = ggconv.to(ctx)
        etypes = etypes.to(ctx)
    
>       h = ggconv(g, feat, etypes)

tests\pytorch\test_nn.py:839: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatedgraphconv.py:166: in forward
    graph.update_all(fn.copy_e('W_e*h', 'm'), fn.sum('m', 'a'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0705DFC8>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([[-1.2025e+00, -1.6061e+00, -9.0034e-01, -1.0437e+00,  6.4090e-01,
         -9.1958e-01,  6.8953e-02, -1.1184e+...76e-01,
          5.7717e-01,  1.5296e+00,  7.3805e-01, -1.4248e-01, -2.4844e+00]],
       grad_fn=<IndexCopyBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________________ test_gated_graph_conv[g4-idtype0] ______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(10,), dtype=torch.float32)}
      edata_schemes={'W_e*h': Scheme(shape=(10,), dtype=torch.float32)})
idtype = torch.int32

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_gated_graph_conv(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        ggconv = nn.GatedGraphConv(5, 10, 5, 3)
        etypes = th.arange(g.number_of_edges()) % 3
        feat = F.randn((g.number_of_nodes(), 5))
        ggconv = ggconv.to(ctx)
        etypes = etypes.to(ctx)
    
>       h = ggconv(g, feat, etypes)

tests\pytorch\test_nn.py:839: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatedgraphconv.py:166: in forward
    graph.update_all(fn.copy_e('W_e*h', 'm'), fn.sum('m', 'a'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03CBE8C8>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([[ 1.0152e+00, -2.4924e-01,  1.5077e+00,  5.9481e-01,  1.4205e+00,
         -1.7157e+00,  1.6703e+00,  9.4764e-...93e-01,
         -3.7687e-01, -3.5135e-01, -1.1963e+00,  3.3918e-01, -2.8068e-01]],
       grad_fn=<IndexCopyBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________________ test_gated_graph_conv[g4-idtype1] ______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(10,), dtype=torch.float32)}
      edata_schemes={'W_e*h': Scheme(shape=(10,), dtype=torch.float32)})
idtype = torch.int64

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_gated_graph_conv(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        ggconv = nn.GatedGraphConv(5, 10, 5, 3)
        etypes = th.arange(g.number_of_edges()) % 3
        feat = F.randn((g.number_of_nodes(), 5))
        ggconv = ggconv.to(ctx)
        etypes = etypes.to(ctx)
    
>       h = ggconv(g, feat, etypes)

tests\pytorch\test_nn.py:839: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatedgraphconv.py:166: in forward
    graph.update_all(fn.copy_e('W_e*h', 'm'), fn.sum('m', 'a'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0726E948>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([[-0.7958,  0.0785, -0.4371, -1.3536, -1.0355,  0.8085,  0.2263,  1.0418,
         -0.3464,  0.9260],
        [...0.1908, -0.1929, -1.6428, -1.6145,  0.4732,  1.4833, -0.3811,
          0.2378,  0.5424]], grad_fn=<IndexCopyBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________________ test_gated_graph_conv[g5-idtype0] ______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(10,), dtype=torch.float32)}
      edata_schemes={'W_e*h': Scheme(shape=(10,), dtype=torch.float32)})
idtype = torch.int32

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_gated_graph_conv(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        ggconv = nn.GatedGraphConv(5, 10, 5, 3)
        etypes = th.arange(g.number_of_edges()) % 3
        feat = F.randn((g.number_of_nodes(), 5))
        ggconv = ggconv.to(ctx)
        etypes = etypes.to(ctx)
    
>       h = ggconv(g, feat, etypes)

tests\pytorch\test_nn.py:839: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatedgraphconv.py:166: in forward
    graph.update_all(fn.copy_e('W_e*h', 'm'), fn.sum('m', 'a'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03D3C548>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([[-6.5399e-01,  2.3313e-01,  1.1944e+00, -1.1540e+00, -4.4744e-01,
         -9.4751e-01,  3.2691e-01,  8.1827e-...70e+00,
          6.6351e-01,  5.5303e-01,  1.0211e+00, -1.5889e+00,  6.1944e-02]],
       grad_fn=<IndexCopyBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
______________________ test_gated_graph_conv[g5-idtype1] ______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(10,), dtype=torch.float32)}
      edata_schemes={'W_e*h': Scheme(shape=(10,), dtype=torch.float32)})
idtype = torch.int64

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_gated_graph_conv(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        ggconv = nn.GatedGraphConv(5, 10, 5, 3)
        etypes = th.arange(g.number_of_edges()) % 3
        feat = F.randn((g.number_of_nodes(), 5))
        ggconv = ggconv.to(ctx)
        etypes = etypes.to(ctx)
    
>       h = ggconv(g, feat, etypes)

tests\pytorch\test_nn.py:839: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatedgraphconv.py:166: in forward
    graph.update_all(fn.copy_e('W_e*h', 'm'), fn.sum('m', 'a'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01635E88>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([[-1.5005,  0.7843, -0.2795, -0.1400, -0.8133,  0.1751, -0.2120, -0.6873,
         -1.0987,  0.4442],
        [...0.3670, -0.3157,  0.5572,  0.6258, -1.5046,  1.2371, -0.6116,
         -1.0062, -0.3811]], grad_fn=<IndexCopyBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_gated_graph_conv_one_etype[g0-idtype0] _________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(10,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32), 'W_e*h': Scheme(shape=(10,), dtype=torch.float32)})
idtype = torch.int32

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_gated_graph_conv_one_etype(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        ggconv = nn.GatedGraphConv(5, 10, 5, 1)
        etypes = th.zeros(g.number_of_edges())
        feat = F.randn((g.number_of_nodes(), 5))
        ggconv = ggconv.to(ctx)
        etypes = etypes.to(ctx)
    
>       h = ggconv(g, feat, etypes)

tests\pytorch\test_nn.py:854: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatedgraphconv.py:166: in forward
    graph.update_all(fn.copy_e('W_e*h', 'm'), fn.sum('m', 'a'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07090E88>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([[-0.6897, -1.1861,  1.6512, -0.7090, -0.0858, -0.9571,  0.5387,  1.1565,
         -1.6776,  0.1337],
        [...0.8486,  1.9717, -0.7752,  1.3342,  0.0961, -0.5752, -0.5053,
          0.6519, -0.4360]], grad_fn=<IndexCopyBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_gated_graph_conv_one_etype[g0-idtype1] _________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(10,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32), 'W_e*h': Scheme(shape=(10,), dtype=torch.float32)})
idtype = torch.int64

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_gated_graph_conv_one_etype(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        ggconv = nn.GatedGraphConv(5, 10, 5, 1)
        etypes = th.zeros(g.number_of_edges())
        feat = F.randn((g.number_of_nodes(), 5))
        ggconv = ggconv.to(ctx)
        etypes = etypes.to(ctx)
    
>       h = ggconv(g, feat, etypes)

tests\pytorch\test_nn.py:854: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatedgraphconv.py:166: in forward
    graph.update_all(fn.copy_e('W_e*h', 'm'), fn.sum('m', 'a'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01756948>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([[-0.0134,  0.3718,  0.3940,  0.4956,  0.8686,  0.9926, -0.4224,  0.5083,
          0.2102,  2.1514],
        [...0.4049,  0.2793,  0.0671, -1.3619, -0.9489,  0.4725,  0.1181,
         -0.0904, -3.2025]], grad_fn=<IndexCopyBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_gated_graph_conv_one_etype[g1-idtype0] _________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(10,), dtype=torch.float32)}
      edata_schemes={'W_e*h': Scheme(shape=(10,), dtype=torch.float32)})
idtype = torch.int32

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_gated_graph_conv_one_etype(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        ggconv = nn.GatedGraphConv(5, 10, 5, 1)
        etypes = th.zeros(g.number_of_edges())
        feat = F.randn((g.number_of_nodes(), 5))
        ggconv = ggconv.to(ctx)
        etypes = etypes.to(ctx)
    
>       h = ggconv(g, feat, etypes)

tests\pytorch\test_nn.py:854: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatedgraphconv.py:166: in forward
    graph.update_all(fn.copy_e('W_e*h', 'm'), fn.sum('m', 'a'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF016BDC88>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([[-0.3178, -0.8938, -1.6870,  0.9792, -1.5222,  0.0930, -0.6055, -0.6056,
         -0.3570,  0.8798],
        [...0.8081, -3.2159,  1.3163, -0.5642, -1.9705,  0.1376, -1.6199,
         -2.5948,  1.2127]], grad_fn=<IndexCopyBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_gated_graph_conv_one_etype[g1-idtype1] _________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(10,), dtype=torch.float32)}
      edata_schemes={'W_e*h': Scheme(shape=(10,), dtype=torch.float32)})
idtype = torch.int64

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_gated_graph_conv_one_etype(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        ggconv = nn.GatedGraphConv(5, 10, 5, 1)
        etypes = th.zeros(g.number_of_edges())
        feat = F.randn((g.number_of_nodes(), 5))
        ggconv = ggconv.to(ctx)
        etypes = etypes.to(ctx)
    
>       h = ggconv(g, feat, etypes)

tests\pytorch\test_nn.py:854: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatedgraphconv.py:166: in forward
    graph.update_all(fn.copy_e('W_e*h', 'm'), fn.sum('m', 'a'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0182E348>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([[ 0.8741, -0.6940,  1.2184,  0.9291, -1.1192,  0.5512,  0.5371,  0.4918,
          1.0928, -0.6343],
        [...0.7946, -0.1105,  0.4282, -0.0574,  0.4263, -0.8628,  0.6048,
          1.9482, -0.7608]], grad_fn=<IndexCopyBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_gated_graph_conv_one_etype[g2-idtype0] _________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(10,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), 'W_e*h': Scheme(shape=(10,), dtype=torch.float32)})
idtype = torch.int32

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_gated_graph_conv_one_etype(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        ggconv = nn.GatedGraphConv(5, 10, 5, 1)
        etypes = th.zeros(g.number_of_edges())
        feat = F.randn((g.number_of_nodes(), 5))
        ggconv = ggconv.to(ctx)
        etypes = etypes.to(ctx)
    
>       h = ggconv(g, feat, etypes)

tests\pytorch\test_nn.py:854: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatedgraphconv.py:166: in forward
    graph.update_all(fn.copy_e('W_e*h', 'm'), fn.sum('m', 'a'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03D3A188>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([[-0.0136,  0.7415, -0.5897, -0.2060, -1.4802, -0.3587, -0.7062,  0.1650,
          0.8711,  0.2058],
        [...1.1643, -2.7958, -1.1792, -3.6901, -1.6968, -0.1925,  0.0420,
          0.9647,  1.8737]], grad_fn=<IndexCopyBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_gated_graph_conv_one_etype[g2-idtype1] _________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(10,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), 'W_e*h': Scheme(shape=(10,), dtype=torch.float32)})
idtype = torch.int64

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_gated_graph_conv_one_etype(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        ggconv = nn.GatedGraphConv(5, 10, 5, 1)
        etypes = th.zeros(g.number_of_edges())
        feat = F.randn((g.number_of_nodes(), 5))
        ggconv = ggconv.to(ctx)
        etypes = etypes.to(ctx)
    
>       h = ggconv(g, feat, etypes)

tests\pytorch\test_nn.py:854: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatedgraphconv.py:166: in forward
    graph.update_all(fn.copy_e('W_e*h', 'm'), fn.sum('m', 'a'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF073EB9C8>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([[-0.5424, -1.6713, -0.2660,  0.5454,  0.5041,  0.7466,  1.0852,  2.6828,
         -1.1898, -0.3039],
        [...0.8101, -1.2362, -0.3035, -0.5102,  0.0238,  0.9539,  0.6480,
          0.3275, -1.3877]], grad_fn=<IndexCopyBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_gated_graph_conv_one_etype[g3-idtype0] _________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(10,), dtype=torch.float32)}
      edata_schemes={'W_e*h': Scheme(shape=(10,), dtype=torch.float32)})
idtype = torch.int32

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_gated_graph_conv_one_etype(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        ggconv = nn.GatedGraphConv(5, 10, 5, 1)
        etypes = th.zeros(g.number_of_edges())
        feat = F.randn((g.number_of_nodes(), 5))
        ggconv = ggconv.to(ctx)
        etypes = etypes.to(ctx)
    
>       h = ggconv(g, feat, etypes)

tests\pytorch\test_nn.py:854: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatedgraphconv.py:166: in forward
    graph.update_all(fn.copy_e('W_e*h', 'm'), fn.sum('m', 'a'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF016C5E88>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([[-1.2203e+00,  4.9314e-01,  7.1594e-02, -1.1298e+00, -1.5235e+00,
          1.0951e+00,  1.2548e+00, -7.3806e-...61e+00,
          8.3022e-01,  7.0108e-01,  1.1366e+00,  1.0659e+00,  1.8892e+00]],
       grad_fn=<IndexCopyBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_gated_graph_conv_one_etype[g3-idtype1] _________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(10,), dtype=torch.float32)}
      edata_schemes={'W_e*h': Scheme(shape=(10,), dtype=torch.float32)})
idtype = torch.int64

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_gated_graph_conv_one_etype(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        ggconv = nn.GatedGraphConv(5, 10, 5, 1)
        etypes = th.zeros(g.number_of_edges())
        feat = F.randn((g.number_of_nodes(), 5))
        ggconv = ggconv.to(ctx)
        etypes = etypes.to(ctx)
    
>       h = ggconv(g, feat, etypes)

tests\pytorch\test_nn.py:854: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatedgraphconv.py:166: in forward
    graph.update_all(fn.copy_e('W_e*h', 'm'), fn.sum('m', 'a'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF073E7608>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([[-0.0729, -0.1870,  0.1397,  0.5697, -0.0602,  0.2616, -0.7796,  1.1400,
         -0.4169, -0.5673],
        [...0.2804, -0.2232,  0.6668, -0.6743,  0.2663, -0.1057,  0.4649,
          0.3657, -0.4250]], grad_fn=<IndexCopyBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_gated_graph_conv_one_etype[g4-idtype0] _________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(10,), dtype=torch.float32)}
      edata_schemes={'W_e*h': Scheme(shape=(10,), dtype=torch.float32)})
idtype = torch.int32

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_gated_graph_conv_one_etype(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        ggconv = nn.GatedGraphConv(5, 10, 5, 1)
        etypes = th.zeros(g.number_of_edges())
        feat = F.randn((g.number_of_nodes(), 5))
        ggconv = ggconv.to(ctx)
        etypes = etypes.to(ctx)
    
>       h = ggconv(g, feat, etypes)

tests\pytorch\test_nn.py:854: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatedgraphconv.py:166: in forward
    graph.update_all(fn.copy_e('W_e*h', 'm'), fn.sum('m', 'a'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE123C4A48>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([[-2.2347e-02, -4.0493e-01, -1.7914e-01,  4.0358e-01,  1.2097e-01,
          1.1547e+00,  3.5966e-01,  1.2723e+...95e+00,
          1.0301e+00, -8.5566e-01, -1.0406e-01, -1.5340e+00, -1.2449e+00]],
       grad_fn=<IndexCopyBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_gated_graph_conv_one_etype[g4-idtype1] _________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(10,), dtype=torch.float32)}
      edata_schemes={'W_e*h': Scheme(shape=(10,), dtype=torch.float32)})
idtype = torch.int64

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_gated_graph_conv_one_etype(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        ggconv = nn.GatedGraphConv(5, 10, 5, 1)
        etypes = th.zeros(g.number_of_edges())
        feat = F.randn((g.number_of_nodes(), 5))
        ggconv = ggconv.to(ctx)
        etypes = etypes.to(ctx)
    
>       h = ggconv(g, feat, etypes)

tests\pytorch\test_nn.py:854: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatedgraphconv.py:166: in forward
    graph.update_all(fn.copy_e('W_e*h', 'm'), fn.sum('m', 'a'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01568748>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([[-0.4417, -1.9879, -1.2170,  1.7168, -0.5137, -0.3298, -0.5861, -1.4225,
         -2.3076, -0.4068],
        [...2.6698, -3.0065, -0.3171,  0.1134,  2.5033, -2.0765, -1.7769,
         -1.3350,  0.9011]], grad_fn=<IndexCopyBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_gated_graph_conv_one_etype[g5-idtype0] _________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(10,), dtype=torch.float32)}
      edata_schemes={'W_e*h': Scheme(shape=(10,), dtype=torch.float32)})
idtype = torch.int32

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_gated_graph_conv_one_etype(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        ggconv = nn.GatedGraphConv(5, 10, 5, 1)
        etypes = th.zeros(g.number_of_edges())
        feat = F.randn((g.number_of_nodes(), 5))
        ggconv = ggconv.to(ctx)
        etypes = etypes.to(ctx)
    
>       h = ggconv(g, feat, etypes)

tests\pytorch\test_nn.py:854: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatedgraphconv.py:166: in forward
    graph.update_all(fn.copy_e('W_e*h', 'm'), fn.sum('m', 'a'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01056AC8>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([[ 0.9016, -3.2595, -0.7682,  1.6626, -1.0176, -1.0124,  0.1687,  0.1683,
          2.0540,  1.1160],
        [...1.6954, -0.5533,  0.1701,  0.3437,  0.9253,  1.1491,  0.1864,
         -0.1712, -0.0799]], grad_fn=<IndexCopyBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_gated_graph_conv_one_etype[g5-idtype1] _________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(10,), dtype=torch.float32)}
      edata_schemes={'W_e*h': Scheme(shape=(10,), dtype=torch.float32)})
idtype = torch.int64

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_gated_graph_conv_one_etype(g, idtype):
        ctx = F.ctx()
        g = g.astype(idtype).to(ctx)
        ggconv = nn.GatedGraphConv(5, 10, 5, 1)
        etypes = th.zeros(g.number_of_edges())
        feat = F.randn((g.number_of_nodes(), 5))
        ggconv = ggconv.to(ctx)
        etypes = etypes.to(ctx)
    
>       h = ggconv(g, feat, etypes)

tests\pytorch\test_nn.py:854: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gatedgraphconv.py:166: in forward
    graph.update_all(fn.copy_e('W_e*h', 'm'), fn.sum('m', 'a'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE123000C8>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([[-1.1769e+00, -3.1107e-02, -6.1376e-01, -3.3871e-01, -6.5688e-01,
          2.9310e-01, -4.7821e-01,  6.7372e-...63e+00,
          1.3877e+00,  3.2194e-01,  6.0567e-01, -7.0362e-01,  3.1825e-01]],
       grad_fn=<IndexCopyBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________________ test_nn_conv[g0-idtype0] ___________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5, 1), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(5, 10), dtype=torch.float32)})
idtype = torch.int32

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    def test_nn_conv(g, idtype):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        edge_func = th.nn.Linear(4, 5 * 10)
        nnconv = nn.NNConv(5, 10, edge_func, 'mean')
        feat = F.randn((g.number_of_src_nodes(), 5))
        efeat = F.randn((g.number_of_edges(), 4))
        nnconv = nnconv.to(ctx)
>       h = nnconv(g, feat, efeat)

tests\pytorch\test_nn.py:870: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\nnconv.py:168: in forward
    graph.update_all(fn.u_mul_e('h', 'w', 'm'), self.reducer('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF072719C8>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[[ 1.6105],
         [ 1.1448],
         [ 1.4999],
         [-0.5556],
         [-0.4088]],

        [[ 0.094...     [ 0.1699]],

        [[-0.2861],
         [ 0.7597],
         [-2.4939],
         [-1.0669],
         [ 0.1899]]])
rhs_data = tensor([[[ 8.2576e-01, -4.7050e-01,  7.7563e-01,  8.2536e-01,  2.4392e-01,
          -4.5868e-01, -6.0147e-01,  1.8605....2142e-01,
          -4.8486e-01, -6.3138e-02, -4.2237e-01,  3.2299e-01, -2.8527e-01]]],
       grad_fn=<ViewBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________________ test_nn_conv[g0-idtype1] ___________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5, 1), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(5, 10), dtype=torch.float32)})
idtype = torch.int64

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    def test_nn_conv(g, idtype):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        edge_func = th.nn.Linear(4, 5 * 10)
        nnconv = nn.NNConv(5, 10, edge_func, 'mean')
        feat = F.randn((g.number_of_src_nodes(), 5))
        efeat = F.randn((g.number_of_edges(), 4))
        nnconv = nnconv.to(ctx)
>       h = nnconv(g, feat, efeat)

tests\pytorch\test_nn.py:870: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\nnconv.py:168: in forward
    graph.update_all(fn.u_mul_e('h', 'w', 'm'), self.reducer('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF072E4348>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[[-0.7352],
         [ 0.2378],
         [-0.0286],
         [ 0.2340],
         [-1.9917]],

        [[-0.078...     [ 0.5483]],

        [[-2.4911],
         [-0.8546],
         [ 1.1509],
         [-0.1186],
         [-0.0410]]])
rhs_data = tensor([[[ 8.8120e-01,  1.3773e-02, -5.2749e-01, -3.9969e-01,  3.7879e-01,
           2.7060e-01,  3.0852e-01, -2.6666....8076e-01,
           2.8510e-01,  2.8224e-01,  1.8351e+00, -2.6469e-01, -5.4991e-02]]],
       grad_fn=<ViewBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________________ test_nn_conv[g1-idtype0] ___________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5, 1), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(5, 10), dtype=torch.float32)})
idtype = torch.int32

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    def test_nn_conv(g, idtype):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        edge_func = th.nn.Linear(4, 5 * 10)
        nnconv = nn.NNConv(5, 10, edge_func, 'mean')
        feat = F.randn((g.number_of_src_nodes(), 5))
        efeat = F.randn((g.number_of_edges(), 4))
        nnconv = nnconv.to(ctx)
>       h = nnconv(g, feat, efeat)

tests\pytorch\test_nn.py:870: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\nnconv.py:168: in forward
    graph.update_all(fn.u_mul_e('h', 'w', 'm'), self.reducer('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF018D24C8>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[[-0.6594],
         [-1.3087],
         [-1.2918],
         [-1.1973],
         [ 0.3114]],

        [[ 0.131...     [ 1.2619]],

        [[-1.7862],
         [-1.4680],
         [-0.0675],
         [ 0.0438],
         [ 2.2645]]])
rhs_data = tensor([[[-3.7857e-01, -1.0376e-01, -5.6935e-01,  3.1129e-02,  6.5841e-01,
          -3.5279e-02, -5.4120e-01, -1.7347....5553e-02,
           1.4305e+00, -8.9141e-01, -2.9831e-01, -8.3306e-01, -3.5016e-01]]],
       grad_fn=<ViewBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________________ test_nn_conv[g1-idtype1] ___________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5, 1), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(5, 10), dtype=torch.float32)})
idtype = torch.int64

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    def test_nn_conv(g, idtype):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        edge_func = th.nn.Linear(4, 5 * 10)
        nnconv = nn.NNConv(5, 10, edge_func, 'mean')
        feat = F.randn((g.number_of_src_nodes(), 5))
        efeat = F.randn((g.number_of_edges(), 4))
        nnconv = nnconv.to(ctx)
>       h = nnconv(g, feat, efeat)

tests\pytorch\test_nn.py:870: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\nnconv.py:168: in forward
    graph.update_all(fn.u_mul_e('h', 'w', 'm'), self.reducer('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1232B608>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[[-1.4509e+00],
         [ 8.9792e-01],
         [-1.1763e+00],
         [-6.7349e-01],
         [ 2.3233e+00]...      [[ 6.9189e-01],
         [ 2.8482e-01],
         [-4.7072e-01],
         [ 1.3066e+00],
         [ 7.4275e-01]]])
rhs_data = tensor([[[-4.3428e-01,  1.0300e+00,  1.5574e-01, -4.3855e-02, -1.2201e+00,
           2.5047e+00,  6.4188e-01, -4.0361....6567e-01,
           1.1190e+00, -2.1165e-01,  2.9749e-01, -1.3590e+00,  4.3597e-01]]],
       grad_fn=<ViewBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________________ test_nn_conv[g2-idtype0] ___________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5, 1), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), 'w': Scheme(shape=(5, 10), dtype=torch.float32)})
idtype = torch.int32

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    def test_nn_conv(g, idtype):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        edge_func = th.nn.Linear(4, 5 * 10)
        nnconv = nn.NNConv(5, 10, edge_func, 'mean')
        feat = F.randn((g.number_of_src_nodes(), 5))
        efeat = F.randn((g.number_of_edges(), 4))
        nnconv = nnconv.to(ctx)
>       h = nnconv(g, feat, efeat)

tests\pytorch\test_nn.py:870: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\nnconv.py:168: in forward
    graph.update_all(fn.u_mul_e('h', 'w', 'm'), self.reducer('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01892FC8>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[[ 1.8373],
         [ 0.2732],
         [-0.3591],
         [ 0.4750],
         [-0.6398]],

        [[-0.053...     [-0.0438]],

        [[-0.2916],
         [-1.6749],
         [ 0.6482],
         [ 0.6713],
         [-1.7338]]])
rhs_data = tensor([[[-2.9097e-01,  5.5693e-01, -1.9549e-01, -6.0476e-01, -2.7407e-01,
           6.0159e-01,  2.1911e-04,  1.0940....9840e-01,
           1.8744e-02,  4.7853e-01,  1.0043e-01, -6.6755e-02,  6.8396e-01]]],
       grad_fn=<ViewBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________________ test_nn_conv[g2-idtype1] ___________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5, 1), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), 'w': Scheme(shape=(5, 10), dtype=torch.float32)})
idtype = torch.int64

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    def test_nn_conv(g, idtype):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        edge_func = th.nn.Linear(4, 5 * 10)
        nnconv = nn.NNConv(5, 10, edge_func, 'mean')
        feat = F.randn((g.number_of_src_nodes(), 5))
        efeat = F.randn((g.number_of_edges(), 4))
        nnconv = nnconv.to(ctx)
>       h = nnconv(g, feat, efeat)

tests\pytorch\test_nn.py:870: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\nnconv.py:168: in forward
    graph.update_all(fn.u_mul_e('h', 'w', 'm'), self.reducer('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF072E6548>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[[ 0.7839],
         [-0.5126],
         [-1.2214],
         [-0.8958],
         [-0.4937]],

        [[ 1.390...     [-1.4965]],

        [[ 0.5894],
         [ 0.7096],
         [ 1.2569],
         [ 0.6219],
         [-0.9380]]])
rhs_data = tensor([[[-4.2609e-01, -2.0803e-01,  1.8228e-01,  4.0750e-01, -5.4358e-01,
           5.1417e-01,  4.4881e-01,  1.1557....4265e+00,
           2.6584e-01, -4.9678e-01,  6.7047e-01, -7.9749e-01, -4.5425e-01]]],
       grad_fn=<ViewBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________________ test_nn_conv[g3-idtype0] ___________________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(5, 1), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(5, 10), dtype=torch.float32)})
idtype = torch.int32

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    def test_nn_conv(g, idtype):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        edge_func = th.nn.Linear(4, 5 * 10)
        nnconv = nn.NNConv(5, 10, edge_func, 'mean')
        feat = F.randn((g.number_of_src_nodes(), 5))
        efeat = F.randn((g.number_of_edges(), 4))
        nnconv = nnconv.to(ctx)
>       h = nnconv(g, feat, efeat)

tests\pytorch\test_nn.py:870: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\nnconv.py:168: in forward
    graph.update_all(fn.u_mul_e('h', 'w', 'm'), self.reducer('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0725E208>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[[-0.0903],
         [-0.3125],
         [ 0.0568],
         [-1.8826],
         [ 0.2567]],

        [[-0.364...     [ 0.7269]],

        [[ 0.7012],
         [ 0.6696],
         [ 0.1366],
         [ 1.5190],
         [-1.6184]]])
rhs_data = tensor([[[-5.6525e-01,  1.2339e-02, -2.5646e-01, -1.3117e+00, -5.1914e-01,
           1.7384e-01,  1.0961e-01, -6.7540....4561e-01,
           6.0565e-01,  3.5897e-01,  8.4287e-01, -1.9219e-01, -2.2849e-01]]],
       grad_fn=<ViewBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________________ test_nn_conv[g3-idtype1] ___________________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(5, 1), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(5, 10), dtype=torch.float32)})
idtype = torch.int64

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    def test_nn_conv(g, idtype):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        edge_func = th.nn.Linear(4, 5 * 10)
        nnconv = nn.NNConv(5, 10, edge_func, 'mean')
        feat = F.randn((g.number_of_src_nodes(), 5))
        efeat = F.randn((g.number_of_edges(), 4))
        nnconv = nnconv.to(ctx)
>       h = nnconv(g, feat, efeat)

tests\pytorch\test_nn.py:870: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\nnconv.py:168: in forward
    graph.update_all(fn.u_mul_e('h', 'w', 'm'), self.reducer('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01890F88>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[[ 0.5097],
         [-0.3208],
         [ 0.1462],
         [ 0.5108],
         [ 0.1423]],

        [[-1.235...     [ 1.6017]],

        [[ 0.9733],
         [-0.5189],
         [-1.4980],
         [-1.3272],
         [-1.5542]]])
rhs_data = tensor([[[-6.3054e-01,  1.9095e-01,  1.3562e+00, -8.8070e-01, -2.8085e-01,
           9.3081e-01,  1.1122e+00, -1.1670....6591e-01,
          -1.6500e-02,  5.7519e-01,  7.3849e-02, -2.4296e-01, -3.7532e-01]]],
       grad_fn=<ViewBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________________ test_nn_conv[g4-idtype0] ___________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5, 1), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(5, 10), dtype=torch.float32)})
idtype = torch.int32

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    def test_nn_conv(g, idtype):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        edge_func = th.nn.Linear(4, 5 * 10)
        nnconv = nn.NNConv(5, 10, edge_func, 'mean')
        feat = F.randn((g.number_of_src_nodes(), 5))
        efeat = F.randn((g.number_of_edges(), 4))
        nnconv = nnconv.to(ctx)
>       h = nnconv(g, feat, efeat)

tests\pytorch\test_nn.py:870: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\nnconv.py:168: in forward
    graph.update_all(fn.u_mul_e('h', 'w', 'm'), self.reducer('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03D8A7C8>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[[-9.0886e-01],
         [-5.5997e-01],
         [ 1.8524e-01],
         [-1.7846e+00],
         [ 2.8472e-01]...      [[-8.1769e-01],
         [-1.2141e+00],
         [ 1.9991e+00],
         [ 2.4733e+00],
         [-1.9136e-02]]])
rhs_data = tensor([[[ 4.2237e-01,  2.6490e-02, -4.1391e-01, -2.8815e-01, -1.4226e-01,
          -5.9593e-01, -3.9922e-01, -1.8630....8466e-01,
           8.7111e-01,  2.7804e-02,  1.6239e-01, -2.7963e-01,  5.6791e-02]]],
       grad_fn=<ViewBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________________ test_nn_conv[g4-idtype1] ___________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5, 1), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(5, 10), dtype=torch.float32)})
idtype = torch.int64

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    def test_nn_conv(g, idtype):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        edge_func = th.nn.Linear(4, 5 * 10)
        nnconv = nn.NNConv(5, 10, edge_func, 'mean')
        feat = F.randn((g.number_of_src_nodes(), 5))
        efeat = F.randn((g.number_of_edges(), 4))
        nnconv = nnconv.to(ctx)
>       h = nnconv(g, feat, efeat)

tests\pytorch\test_nn.py:870: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\nnconv.py:168: in forward
    graph.update_all(fn.u_mul_e('h', 'w', 'm'), self.reducer('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF072BF408>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[[ 0.7062],
         [ 0.4698],
         [-1.2409],
         [-0.9853],
         [ 2.2252]],

        [[ 0.131...     [-2.6137]],

        [[ 0.9596],
         [-0.6359],
         [ 0.1255],
         [ 0.3129],
         [ 1.5072]]])
rhs_data = tensor([[[ 5.5353e-01,  1.4861e-01,  1.8772e-01, -5.8536e-01,  8.9622e-01,
          -6.2239e-01, -5.7919e-01, -3.5456....0265e-01,
          -8.3651e-01, -5.6061e-01,  1.7238e-01,  1.8418e-01, -8.0785e-01]]],
       grad_fn=<ViewBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________________ test_nn_conv[g5-idtype0] ___________________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int32

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    def test_nn_conv(g, idtype):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        edge_func = th.nn.Linear(4, 5 * 10)
        nnconv = nn.NNConv(5, 10, edge_func, 'mean')
        feat = F.randn((g.number_of_src_nodes(), 5))
        efeat = F.randn((g.number_of_edges(), 4))
        nnconv = nnconv.to(ctx)
>       h = nnconv(g, feat, efeat)

tests\pytorch\test_nn.py:870: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\nnconv.py:168: in forward
    graph.update_all(fn.u_mul_e('h', 'w', 'm'), self.reducer('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1469EE08>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[[-2.5708e+00],
         [-4.9630e-02],
         [-1.0974e+00],
         [ 6.9190e-01],
         [-1.3132e-01]...      [[ 2.1599e-01],
         [ 1.3091e-01],
         [ 1.9167e+00],
         [ 9.7499e-01],
         [-1.3350e-01]]])
rhs_data = tensor([[[-0.1897,  0.6466, -0.5478, -0.2022,  0.9300,  0.7670,  0.6228,
          -0.5767, -0.2971, -0.1333],
       ..., -0.0234, -0.3985,  0.0166,  0.5951,  0.2847, -0.6190,
           0.4502,  0.1349, -0.1918]]], grad_fn=<ViewBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________________ test_nn_conv[g5-idtype1] ___________________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int64

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    def test_nn_conv(g, idtype):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        edge_func = th.nn.Linear(4, 5 * 10)
        nnconv = nn.NNConv(5, 10, edge_func, 'mean')
        feat = F.randn((g.number_of_src_nodes(), 5))
        efeat = F.randn((g.number_of_edges(), 4))
        nnconv = nnconv.to(ctx)
>       h = nnconv(g, feat, efeat)

tests\pytorch\test_nn.py:870: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\nnconv.py:168: in forward
    graph.update_all(fn.u_mul_e('h', 'w', 'm'), self.reducer('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03D66A48>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[[-0.8700],
         [ 0.4072],
         [ 0.6403],
         [-2.1423],
         [-0.1622]],

        [[ 0.828...     [-1.0972]],

        [[-1.2113],
         [ 0.2733],
         [-0.7043],
         [ 0.9543],
         [-1.6574]]])
rhs_data = tensor([[[-0.3872,  0.2797,  0.1140, -0.2533,  0.0966,  0.0362, -0.5189,
          -0.5645,  0.2884, -0.1082],
       ...,  0.1587, -0.5384,  0.0141, -0.4856, -0.6456,  0.4106,
          -0.2256, -0.2534,  0.3953]]], grad_fn=<ViewBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________________ test_nn_conv[g6-idtype0] ___________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5, 1), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(5, 10), dtype=torch.float32)})
idtype = torch.int32

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    def test_nn_conv(g, idtype):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        edge_func = th.nn.Linear(4, 5 * 10)
        nnconv = nn.NNConv(5, 10, edge_func, 'mean')
        feat = F.randn((g.number_of_src_nodes(), 5))
        efeat = F.randn((g.number_of_edges(), 4))
        nnconv = nnconv.to(ctx)
>       h = nnconv(g, feat, efeat)

tests\pytorch\test_nn.py:870: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\nnconv.py:168: in forward
    graph.update_all(fn.u_mul_e('h', 'w', 'm'), self.reducer('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03D0CD88>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[[-0.4216],
         [ 1.7510],
         [-0.6387],
         [ 0.6388],
         [ 1.2004]],

        [[ 1.158...     [ 0.4181]],

        [[ 0.3234],
         [-0.6070],
         [ 0.4604],
         [ 0.2601],
         [ 2.0716]]])
rhs_data = tensor([[[-7.1898e-01, -4.0984e-01, -4.5876e-01,  5.2486e-01,  9.4229e-01,
           1.0168e+00, -2.4729e-03, -9.7308....4241e-01,
          -7.1283e-01, -1.8350e-01, -1.0275e+00, -3.6497e-02, -6.1681e-01]]],
       grad_fn=<ViewBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________________ test_nn_conv[g6-idtype1] ___________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(5, 1), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(5, 10), dtype=torch.float32)})
idtype = torch.int64

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    def test_nn_conv(g, idtype):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        edge_func = th.nn.Linear(4, 5 * 10)
        nnconv = nn.NNConv(5, 10, edge_func, 'mean')
        feat = F.randn((g.number_of_src_nodes(), 5))
        efeat = F.randn((g.number_of_edges(), 4))
        nnconv = nnconv.to(ctx)
>       h = nnconv(g, feat, efeat)

tests\pytorch\test_nn.py:870: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\nnconv.py:168: in forward
    graph.update_all(fn.u_mul_e('h', 'w', 'm'), self.reducer('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF070C2A08>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[[ 0.8483],
         [ 0.6652],
         [-0.8181],
         [ 0.1647],
         [-0.8461]],

        [[-3.496...     [-0.7719]],

        [[ 1.3900],
         [-0.0283],
         [-0.5718],
         [ 0.2895],
         [ 0.0654]]])
rhs_data = tensor([[[ 3.1462e-01, -6.8390e-01,  4.3771e-01, -9.8854e-01, -6.3900e-02,
          -3.8192e-01, -6.0243e-01,  4.4032....5639e-01,
          -6.9813e-01,  2.1549e-01,  7.4389e-01, -3.0876e-01, -5.9535e-01]]],
       grad_fn=<ViewBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_nn_conv_bi[g0-idtype0] _________________________

g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int32

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    def test_nn_conv_bi(g, idtype):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        edge_func = th.nn.Linear(4, 5 * 10)
        nnconv = nn.NNConv((5, 2), 10, edge_func, 'mean')
        feat = F.randn((g.number_of_src_nodes(), 5))
        feat_dst = F.randn((g.number_of_dst_nodes(), 2))
        efeat = F.randn((g.number_of_edges(), 4))
        nnconv = nnconv.to(ctx)
>       h = nnconv(g, (feat, feat_dst), efeat)

tests\pytorch\test_nn.py:885: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\nnconv.py:168: in forward
    graph.update_all(fn.u_mul_e('h', 'w', 'm'), self.reducer('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0146E508>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[[-1.8129],
         [ 0.0198],
         [-1.0557],
         [-1.8359],
         [ 0.0123]],

        [[-0.6516],
         [-0.4092],
         [-0.1734],
         [-1.2135],
         [-0.7169]]])
rhs_data = tensor([[[-2.4746e-01,  5.3657e-01, -1.0729e-01,  1.2967e-01,  7.4316e-01,
          -5.5548e-01, -1.3860e-01, -1.6261....9881e-02,
           4.5117e-01,  2.6996e-01,  3.5154e-01, -7.7012e-01, -1.4374e-01]]],
       grad_fn=<ViewBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_nn_conv_bi[g0-idtype1] _________________________

g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int64

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    def test_nn_conv_bi(g, idtype):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        edge_func = th.nn.Linear(4, 5 * 10)
        nnconv = nn.NNConv((5, 2), 10, edge_func, 'mean')
        feat = F.randn((g.number_of_src_nodes(), 5))
        feat_dst = F.randn((g.number_of_dst_nodes(), 2))
        efeat = F.randn((g.number_of_edges(), 4))
        nnconv = nnconv.to(ctx)
>       h = nnconv(g, (feat, feat_dst), efeat)

tests\pytorch\test_nn.py:885: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\nnconv.py:168: in forward
    graph.update_all(fn.u_mul_e('h', 'w', 'm'), self.reducer('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1231C748>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[[-2.4119],
         [-0.9670],
         [-1.7840],
         [-0.6586],
         [ 1.4624]],

        [[ 1.9595],
         [-0.1084],
         [-0.3641],
         [-0.9541],
         [ 0.8652]]])
rhs_data = tensor([[[-0.3702,  0.3196, -0.5047,  0.1652, -0.8513,  0.5248, -1.1570,
          -0.9323,  0.4260,  0.6290],
       ...,  0.7234,  0.5843,  0.8130, -1.1965, -0.3629,  1.5739,
          -1.0989, -1.4331, -0.4681]]], grad_fn=<ViewBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_nn_conv_bi[g1-idtype0] _________________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int32

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    def test_nn_conv_bi(g, idtype):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        edge_func = th.nn.Linear(4, 5 * 10)
        nnconv = nn.NNConv((5, 2), 10, edge_func, 'mean')
        feat = F.randn((g.number_of_src_nodes(), 5))
        feat_dst = F.randn((g.number_of_dst_nodes(), 2))
        efeat = F.randn((g.number_of_edges(), 4))
        nnconv = nnconv.to(ctx)
>       h = nnconv(g, (feat, feat_dst), efeat)

tests\pytorch\test_nn.py:885: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\nnconv.py:168: in forward
    graph.update_all(fn.u_mul_e('h', 'w', 'm'), self.reducer('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03DA0888>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[[ 0.0668],
         [-1.0075],
         [-0.3981],
         [-1.2391],
         [-0.5691]],

        [[ 0.973...     [ 0.3488]],

        [[-0.4802],
         [-0.2115],
         [ 0.7891],
         [-0.0660],
         [-0.6463]]])
rhs_data = tensor([[[ 2.8435e-01,  1.1743e+00,  1.5007e+00,  4.5370e-01,  7.7482e-01,
          -7.2070e-01,  4.7197e-01, -4.4477....4213e-01,
          -8.0057e-01, -2.5315e-01,  6.6837e-01,  1.0907e-01, -6.1113e-01]]],
       grad_fn=<ViewBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_nn_conv_bi[g1-idtype1] _________________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int64

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    def test_nn_conv_bi(g, idtype):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        edge_func = th.nn.Linear(4, 5 * 10)
        nnconv = nn.NNConv((5, 2), 10, edge_func, 'mean')
        feat = F.randn((g.number_of_src_nodes(), 5))
        feat_dst = F.randn((g.number_of_dst_nodes(), 2))
        efeat = F.randn((g.number_of_edges(), 4))
        nnconv = nnconv.to(ctx)
>       h = nnconv(g, (feat, feat_dst), efeat)

tests\pytorch\test_nn.py:885: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\nnconv.py:168: in forward
    graph.update_all(fn.u_mul_e('h', 'w', 'm'), self.reducer('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF011E18C8>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[[-0.7108],
         [ 0.5723],
         [ 1.5076],
         [ 0.0324],
         [-0.8884]],

        [[ 0.489...     [-0.0056]],

        [[ 1.2583],
         [-0.9806],
         [ 0.1414],
         [ 1.1724],
         [-1.6388]]])
rhs_data = tensor([[[-0.0077,  0.6619,  0.2604, -0.3285,  0.4877, -0.3253, -0.7351,
          -0.2655, -1.2512, -0.4471],
       ...,  0.0377,  0.6960,  0.1912, -0.9976, -0.7429,  0.0805,
           0.7918,  0.9663, -0.3223]]], grad_fn=<ViewBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________________ test_gmm_conv[g0-idtype0] __________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(4, 10), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(4, 1), dtype=torch.float32)})
idtype = torch.int32

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_gmm_conv(g, idtype):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gmmconv = nn.GMMConv(5, 10, 3, 4, 'mean')
        feat = F.randn((g.number_of_nodes(), 5))
        pseudo = F.randn((g.number_of_edges(), 3))
        gmmconv = gmmconv.to(ctx)
>       h = gmmconv(g, feat, pseudo)

tests\pytorch\test_nn.py:898: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gmmconv.py:240: in forward
    graph.update_all(fn.u_mul_e('h', 'w', 'm'), self._reducer('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0704B308>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[[-1.0394e-01,  1.5901e-01, -8.2195e-01, -5.5929e-01, -1.8217e+00,
           9.4200e-01, -8.4482e-01,  8.1142....2907e-01,
          -2.8834e-01,  1.7658e-01,  1.8434e-02, -2.3371e-02,  3.7818e-01]]],
       grad_fn=<ViewBackward>)
rhs_data = tensor([[[0.3099],
         [0.2443],
         [0.2831],
         [0.1945]],

        [[0.4584],
         [0.5196],
  ...       [0.4078]],

        [[0.7378],
         [0.6698],
         [0.7158],
         [0.6567]]], grad_fn=<ExpBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________________ test_gmm_conv[g0-idtype1] __________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(4, 10), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(4, 1), dtype=torch.float32)})
idtype = torch.int64

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_gmm_conv(g, idtype):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gmmconv = nn.GMMConv(5, 10, 3, 4, 'mean')
        feat = F.randn((g.number_of_nodes(), 5))
        pseudo = F.randn((g.number_of_edges(), 3))
        gmmconv = gmmconv.to(ctx)
>       h = gmmconv(g, feat, pseudo)

tests\pytorch\test_nn.py:898: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gmmconv.py:240: in forward
    graph.update_all(fn.u_mul_e('h', 'w', 'm'), self._reducer('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1226B748>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[[-1.0087, -0.5836,  0.8994,  0.4047,  0.5372,  1.1467, -0.4768,
          -0.1266, -1.0438,  0.5160],
       ...,  0.3299, -0.0093,  0.6166, -1.2858, -0.4726,  0.5087,
          -1.7000, -0.8218, -0.0943]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[4.9052e-01],
         [4.3691e-01],
         [5.2787e-01],
         [5.0224e-01]],

        [[5.8906e-01],
 ...,

        [[9.0779e-01],
         [9.2133e-01],
         [9.4710e-01],
         [9.4793e-01]]], grad_fn=<ExpBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________________ test_gmm_conv[g1-idtype0] __________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(4, 10), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(4, 1), dtype=torch.float32)})
idtype = torch.int32

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_gmm_conv(g, idtype):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gmmconv = nn.GMMConv(5, 10, 3, 4, 'mean')
        feat = F.randn((g.number_of_nodes(), 5))
        pseudo = F.randn((g.number_of_edges(), 3))
        gmmconv = gmmconv.to(ctx)
>       h = gmmconv(g, feat, pseudo)

tests\pytorch\test_nn.py:898: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gmmconv.py:240: in forward
    graph.update_all(fn.u_mul_e('h', 'w', 'm'), self._reducer('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01714848>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[[-5.2860e-02,  9.9621e-01,  5.3027e-01, -3.3924e-01,  2.1454e+00,
           6.8411e-01, -1.7225e+00, -7.6966....8524e-01,
           1.0855e+00,  3.9795e-01, -1.4060e+00, -3.7768e-01, -1.0134e+00]]],
       grad_fn=<ViewBackward>)
rhs_data = tensor([[[1.1542e-01],
         [1.2960e-01],
         [1.3857e-01],
         [1.7276e-01]],

        [[1.9985e-01],
 ...,

        [[6.8592e-01],
         [6.7675e-01],
         [7.3623e-01],
         [6.6932e-01]]], grad_fn=<ExpBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________________ test_gmm_conv[g1-idtype1] __________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(4, 10), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(4, 1), dtype=torch.float32)})
idtype = torch.int64

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_gmm_conv(g, idtype):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gmmconv = nn.GMMConv(5, 10, 3, 4, 'mean')
        feat = F.randn((g.number_of_nodes(), 5))
        pseudo = F.randn((g.number_of_edges(), 3))
        gmmconv = gmmconv.to(ctx)
>       h = gmmconv(g, feat, pseudo)

tests\pytorch\test_nn.py:898: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gmmconv.py:240: in forward
    graph.update_all(fn.u_mul_e('h', 'w', 'm'), self._reducer('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03E09A08>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[[ 1.0668e+00,  7.9595e-01,  3.7557e-01, -1.8368e-01,  8.4075e-01,
           8.7313e-02,  3.1826e-01, -4.3885....6818e-01,
          -1.0887e-01, -4.2135e-02, -5.4503e-04,  3.7206e-01, -7.2532e-02]]],
       grad_fn=<ViewBackward>)
rhs_data = tensor([[[0.4501],
         [0.3756],
         [0.3836],
         [0.4343]],

        [[0.5577],
         [0.5648],
  ...       [0.7938]],

        [[0.0612],
         [0.0565],
         [0.0564],
         [0.0629]]], grad_fn=<ExpBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________________ test_gmm_conv[g2-idtype0] __________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(4, 10), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), 'w': Scheme(shape=(4, 1), dtype=torch.float32)})
idtype = torch.int32

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_gmm_conv(g, idtype):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gmmconv = nn.GMMConv(5, 10, 3, 4, 'mean')
        feat = F.randn((g.number_of_nodes(), 5))
        pseudo = F.randn((g.number_of_edges(), 3))
        gmmconv = gmmconv.to(ctx)
>       h = gmmconv(g, feat, pseudo)

tests\pytorch\test_nn.py:898: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gmmconv.py:240: in forward
    graph.update_all(fn.u_mul_e('h', 'w', 'm'), self._reducer('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0189E6C8>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[[-1.1562e-01,  2.8064e-01, -3.8650e-01,  6.6696e-02, -2.6928e-02,
          -2.8173e-01,  7.1332e-02,  1.0294....8192e-01,
          -1.6281e-02, -4.2456e-01, -1.9981e-01, -6.9853e-01,  9.9441e-02]]],
       grad_fn=<ViewBackward>)
rhs_data = tensor([[[0.1800],
         [0.2041],
         [0.3098],
         [0.2772]],

        [[0.2622],
         [0.2848],
  ...       [0.3757]],

        [[0.0098],
         [0.0123],
         [0.0246],
         [0.0220]]], grad_fn=<ExpBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________________ test_gmm_conv[g2-idtype1] __________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(4, 10), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), 'w': Scheme(shape=(4, 1), dtype=torch.float32)})
idtype = torch.int64

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_gmm_conv(g, idtype):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gmmconv = nn.GMMConv(5, 10, 3, 4, 'mean')
        feat = F.randn((g.number_of_nodes(), 5))
        pseudo = F.randn((g.number_of_edges(), 3))
        gmmconv = gmmconv.to(ctx)
>       h = gmmconv(g, feat, pseudo)

tests\pytorch\test_nn.py:898: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gmmconv.py:240: in forward
    graph.update_all(fn.u_mul_e('h', 'w', 'm'), self._reducer('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01809548>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[[ 0.3245,  0.3329, -1.0036, -0.3040,  1.4024, -0.3393,  1.6174,
           1.0631, -0.6760, -0.7969],
       ..., -0.5083, -0.6526,  0.0906, -0.0647, -0.5312,  0.3471,
          -0.1384,  0.3147, -0.2978]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[0.6044],
         [0.6429],
         [0.6066],
         [0.5427]],

        [[0.0683],
         [0.0644],
  ...       [0.5994]],

        [[0.1766],
         [0.1585],
         [0.1949],
         [0.1432]]], grad_fn=<ExpBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________________ test_gmm_conv[g3-idtype0] __________________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(4, 10), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(4, 1), dtype=torch.float32)})
idtype = torch.int32

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_gmm_conv(g, idtype):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gmmconv = nn.GMMConv(5, 10, 3, 4, 'mean')
        feat = F.randn((g.number_of_nodes(), 5))
        pseudo = F.randn((g.number_of_edges(), 3))
        gmmconv = gmmconv.to(ctx)
>       h = gmmconv(g, feat, pseudo)

tests\pytorch\test_nn.py:898: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gmmconv.py:240: in forward
    graph.update_all(fn.u_mul_e('h', 'w', 'm'), self._reducer('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03D661C8>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[[ 6.9404e-01, -6.0983e-01,  7.1854e-01, -6.1359e-01,  1.6982e+00,
          -9.7232e-02, -8.8549e-01,  5.9038....5970e-01,
           4.2481e-01, -7.8436e-02, -1.9853e-01,  1.0576e+00,  7.9524e-01]]],
       grad_fn=<ViewBackward>)
rhs_data = tensor([[[0.0410],
         [0.1067],
         [0.0470],
         [0.0842]],

        [[0.1473],
         [0.2526],
  ...       [0.0728]],

        [[0.2209],
         [0.1347],
         [0.2877],
         [0.1334]]], grad_fn=<ExpBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________________ test_gmm_conv[g3-idtype1] __________________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(4, 10), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(4, 1), dtype=torch.float32)})
idtype = torch.int64

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_gmm_conv(g, idtype):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gmmconv = nn.GMMConv(5, 10, 3, 4, 'mean')
        feat = F.randn((g.number_of_nodes(), 5))
        pseudo = F.randn((g.number_of_edges(), 3))
        gmmconv = gmmconv.to(ctx)
>       h = gmmconv(g, feat, pseudo)

tests\pytorch\test_nn.py:898: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gmmconv.py:240: in forward
    graph.update_all(fn.u_mul_e('h', 'w', 'm'), self._reducer('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF019737C8>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[[-1.5042e-01, -3.8891e-01,  1.7365e+00,  1.1596e+00, -7.9376e-01,
           7.2831e-01, -5.9966e-01, -2.6602....5965e-01,
           7.5041e-01,  1.7507e+00,  2.4121e-01, -6.4155e-01,  7.0471e-01]]],
       grad_fn=<ViewBackward>)
rhs_data = tensor([[[0.0362],
         [0.0387],
         [0.0530],
         [0.0311]],

        [[0.1989],
         [0.2230],
  ...       [0.0465]],

        [[0.0350],
         [0.0223],
         [0.0215],
         [0.0462]]], grad_fn=<ExpBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________________ test_gmm_conv[g4-idtype0] __________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(4, 10), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(4, 1), dtype=torch.float32)})
idtype = torch.int32

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_gmm_conv(g, idtype):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gmmconv = nn.GMMConv(5, 10, 3, 4, 'mean')
        feat = F.randn((g.number_of_nodes(), 5))
        pseudo = F.randn((g.number_of_edges(), 3))
        gmmconv = gmmconv.to(ctx)
>       h = gmmconv(g, feat, pseudo)

tests\pytorch\test_nn.py:898: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gmmconv.py:240: in forward
    graph.update_all(fn.u_mul_e('h', 'w', 'm'), self._reducer('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0184F348>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[[-7.2804e-01,  9.3219e-01,  2.3345e-01, -7.8081e-01, -1.6124e+00,
          -1.3781e+00,  7.7109e-01,  7.1306....7545e-01,
           3.5383e-01,  2.4587e-01,  1.6901e+00,  3.8637e-01, -1.4900e+00]]],
       grad_fn=<ViewBackward>)
rhs_data = tensor([[[0.8726],
         [0.7758],
         [0.8965],
         [0.8581]],

        [[0.2396],
         [0.2013],
  ...       [0.7634]],

        [[0.5795],
         [0.4999],
         [0.5012],
         [0.6198]]], grad_fn=<ExpBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________________ test_gmm_conv[g4-idtype1] __________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(4, 10), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(4, 1), dtype=torch.float32)})
idtype = torch.int64

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_gmm_conv(g, idtype):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gmmconv = nn.GMMConv(5, 10, 3, 4, 'mean')
        feat = F.randn((g.number_of_nodes(), 5))
        pseudo = F.randn((g.number_of_edges(), 3))
        gmmconv = gmmconv.to(ctx)
>       h = gmmconv(g, feat, pseudo)

tests\pytorch\test_nn.py:898: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gmmconv.py:240: in forward
    graph.update_all(fn.u_mul_e('h', 'w', 'm'), self._reducer('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF071133C8>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[[ 1.0576e+00, -6.4621e-01, -4.1629e-01, -9.0518e-02, -1.2911e+00,
           1.1726e+00,  2.3198e-01, -4.5388....1093e-02,
           3.4171e-01, -4.2320e-01, -5.1958e-01, -5.2254e-02,  1.5779e+00]]],
       grad_fn=<ViewBackward>)
rhs_data = tensor([[[0.1811],
         [0.2556],
         [0.2531],
         [0.1786]],

        [[0.2156],
         [0.2901],
  ...       [0.0832]],

        [[0.1494],
         [0.1385],
         [0.1432],
         [0.1771]]], grad_fn=<ExpBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________________ test_gmm_conv[g5-idtype0] __________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(4, 10), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(4, 1), dtype=torch.float32)})
idtype = torch.int32

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_gmm_conv(g, idtype):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gmmconv = nn.GMMConv(5, 10, 3, 4, 'mean')
        feat = F.randn((g.number_of_nodes(), 5))
        pseudo = F.randn((g.number_of_edges(), 3))
        gmmconv = gmmconv.to(ctx)
>       h = gmmconv(g, feat, pseudo)

tests\pytorch\test_nn.py:898: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gmmconv.py:240: in forward
    graph.update_all(fn.u_mul_e('h', 'w', 'm'), self._reducer('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03DC6348>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[[-8.5381e-01, -1.9209e+00,  5.5114e-03, -6.2952e-01, -1.4025e+00,
          -7.9472e-01, -5.0047e-01, -2.9629....2686e-01,
          -5.5830e-01, -6.8988e-01, -1.3553e+00,  1.4776e+00,  1.1147e+00]]],
       grad_fn=<ViewBackward>)
rhs_data = tensor([[[5.3326e-01],
         [5.7278e-01],
         [5.3645e-01],
         [6.2758e-01]],

        [[8.5498e-01],
 ...,

        [[4.9567e-01],
         [5.5898e-01],
         [5.3344e-01],
         [5.7535e-01]]], grad_fn=<ExpBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________________ test_gmm_conv[g5-idtype1] __________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(4, 10), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(4, 1), dtype=torch.float32)})
idtype = torch.int64

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    def test_gmm_conv(g, idtype):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gmmconv = nn.GMMConv(5, 10, 3, 4, 'mean')
        feat = F.randn((g.number_of_nodes(), 5))
        pseudo = F.randn((g.number_of_edges(), 3))
        gmmconv = gmmconv.to(ctx)
>       h = gmmconv(g, feat, pseudo)

tests\pytorch\test_nn.py:898: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gmmconv.py:240: in forward
    graph.update_all(fn.u_mul_e('h', 'w', 'm'), self._reducer('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03DD9088>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[[-3.9792e-01, -5.6246e-01, -1.0099e+00, -6.0917e-01, -2.2818e+00,
           1.6990e+00, -3.4209e-01,  3.2864....9608e-01,
          -5.8952e-01, -3.5950e-01,  8.0589e-01, -2.1694e-01,  1.2902e-01]]],
       grad_fn=<ViewBackward>)
rhs_data = tensor([[[0.8668],
         [0.8903],
         [0.8428],
         [0.9275]],

        [[0.0793],
         [0.0645],
  ...       [0.0894]],

        [[0.9182],
         [0.8660],
         [0.8941],
         [0.7923]]], grad_fn=<ExpBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________________ test_gmm_conv_bi[g0-idtype0] _________________________

g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int32

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite', 'block-bipartite'], exclude=['zero-degree']))
    def test_gmm_conv_bi(g, idtype):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gmmconv = nn.GMMConv((5, 2), 10, 3, 4, 'mean')
        feat = F.randn((g.number_of_src_nodes(), 5))
        feat_dst = F.randn((g.number_of_dst_nodes(), 2))
        pseudo = F.randn((g.number_of_edges(), 3))
        gmmconv = gmmconv.to(ctx)
>       h = gmmconv(g, (feat, feat_dst), pseudo)

tests\pytorch\test_nn.py:912: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gmmconv.py:240: in forward
    graph.update_all(fn.u_mul_e('h', 'w', 'm'), self._reducer('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03DB3A88>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[[ 0.0667,  0.0939, -0.5682, -0.4721, -0.5217, -0.1882,  0.2747,
           0.1830, -0.2771,  0.1894],
       ..., -0.0132,  0.1187, -0.4078,  0.7262,  0.0545, -0.3632,
          -0.2224,  0.4490,  0.2743]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[0.0676],
         [0.0520],
         [0.0625],
         [0.0742]],

        [[0.7224],
         [0.7285],
  ...       [0.0230]],

        [[0.0024],
         [0.0023],
         [0.0030],
         [0.0034]]], grad_fn=<ExpBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________________ test_gmm_conv_bi[g0-idtype1] _________________________

g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int64

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite', 'block-bipartite'], exclude=['zero-degree']))
    def test_gmm_conv_bi(g, idtype):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gmmconv = nn.GMMConv((5, 2), 10, 3, 4, 'mean')
        feat = F.randn((g.number_of_src_nodes(), 5))
        feat_dst = F.randn((g.number_of_dst_nodes(), 2))
        pseudo = F.randn((g.number_of_edges(), 3))
        gmmconv = gmmconv.to(ctx)
>       h = gmmconv(g, (feat, feat_dst), pseudo)

tests\pytorch\test_nn.py:912: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gmmconv.py:240: in forward
    graph.update_all(fn.u_mul_e('h', 'w', 'm'), self._reducer('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0168E608>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[[-0.2430, -0.1104,  0.0136, -0.7433, -0.3674,  0.2859,  0.4140,
           0.2443, -0.4271,  0.2678],
       ..., -0.2325,  0.5585,  0.2391,  0.1514, -0.1036,  0.1283,
           0.2434,  0.1330, -0.1417]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[0.8505],
         [0.7463],
         [0.8756],
         [0.8577]],

        [[0.2809],
         [0.2789],
  ...       [0.3135]],

        [[0.0750],
         [0.0566],
         [0.0774],
         [0.0823]]], grad_fn=<ExpBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________________ test_gmm_conv_bi[g1-idtype0] _________________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int32

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite', 'block-bipartite'], exclude=['zero-degree']))
    def test_gmm_conv_bi(g, idtype):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gmmconv = nn.GMMConv((5, 2), 10, 3, 4, 'mean')
        feat = F.randn((g.number_of_src_nodes(), 5))
        feat_dst = F.randn((g.number_of_dst_nodes(), 2))
        pseudo = F.randn((g.number_of_edges(), 3))
        gmmconv = gmmconv.to(ctx)
>       h = gmmconv(g, (feat, feat_dst), pseudo)

tests\pytorch\test_nn.py:912: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gmmconv.py:240: in forward
    graph.update_all(fn.u_mul_e('h', 'w', 'm'), self._reducer('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03E17108>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[[-3.8567e-01,  1.8279e-02, -4.1301e-01,  1.0180e+00,  6.4230e-01,
           3.2729e-01, -2.3017e-01,  1.2685....0580e-01,
          -6.9238e-01,  2.1038e-01, -7.4467e-01,  2.1979e-01, -6.8019e-01]]],
       grad_fn=<ViewBackward>)
rhs_data = tensor([[[0.2555],
         [0.4781],
         [0.4722],
         [0.3243]],

        [[0.5087],
         [0.6912],
  ...       [0.4522]],

        [[0.5181],
         [0.6382],
         [0.5075],
         [0.4310]]], grad_fn=<ExpBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________________ test_gmm_conv_bi[g1-idtype1] _________________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int64

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite', 'block-bipartite'], exclude=['zero-degree']))
    def test_gmm_conv_bi(g, idtype):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        gmmconv = nn.GMMConv((5, 2), 10, 3, 4, 'mean')
        feat = F.randn((g.number_of_src_nodes(), 5))
        feat_dst = F.randn((g.number_of_dst_nodes(), 2))
        pseudo = F.randn((g.number_of_edges(), 3))
        gmmconv = gmmconv.to(ctx)
>       h = gmmconv(g, (feat, feat_dst), pseudo)

tests\pytorch\test_nn.py:912: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\gmmconv.py:240: in forward
    graph.update_all(fn.u_mul_e('h', 'w', 'm'), self._reducer('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03D550C8>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[[ 1.4966e-01, -7.5348e-01,  4.9827e-01, -3.5936e-01,  7.9210e-01,
          -6.2673e-01,  2.8023e-01, -9.7601....0406e-01,
           2.1381e-01,  6.1623e-01,  5.2469e-02,  2.0718e-02, -3.5698e-02]]],
       grad_fn=<ViewBackward>)
rhs_data = tensor([[[0.4400],
         [0.6046],
         [0.4899],
         [0.5654]],

        [[0.0035],
         [0.0070],
  ...       [0.0050]],

        [[0.4921],
         [0.6128],
         [0.5089],
         [0.6100]]], grad_fn=<ExpBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[1-g0-both-idtype0] ___________________

norm_type = 'both'
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int32, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03D5EFC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.6894],
        [-0.2656]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[1-g0-both-idtype1] ___________________

norm_type = 'both'
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int64, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1469E188>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.7086],
        [-0.7572]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[1-g0-right-idtype0] __________________

norm_type = 'right'
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int32, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF015680C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0.4984],
        [0.7740]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[1-g0-right-idtype1] __________________

norm_type = 'right'
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int64, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01993648>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0.2060],
        [2.0173]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[1-g0-none-idtype0] ___________________

norm_type = 'none'
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int32, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03E00C48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.5704],
        [-0.0931]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[1-g0-none-idtype1] ___________________

norm_type = 'none'
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int64, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07193548>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[4.3463],
        [0.3246]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[1-g1-both-idtype0] ___________________

norm_type = 'both'
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
idtype = torch.int32, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03E1E5C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.3120],
        [-0.2244],
        [ 0.6342],
        [-0.3165],
        [-0.3298],
        [ 0.5249],
        [-0.3563],
        [-0.2584],
        [-0.3699],
        [-0.1353]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[1-g1-both-idtype1] ___________________

norm_type = 'both'
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
idtype = torch.int64, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF071892C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.5592],
        [-0.7154],
        [-0.9441],
        [-0.2006],
        [-1.0904],
        [ 2.1187],
        [ 0.6767],
        [-0.3431],
        [ 1.2696],
        [-2.1132]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[1-g1-right-idtype0] __________________

norm_type = 'right'
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
idtype = torch.int32, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1231C288>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.9529],
        [ 1.8418],
        [-1.1630],
        [ 1.0026],
        [-0.7052],
        [-1.7842],
        [ 1.4343],
        [ 0.6845],
        [ 3.1847],
        [-0.5136]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[1-g1-right-idtype1] __________________

norm_type = 'right'
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
idtype = torch.int64, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE12242D08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.4136],
        [-0.6471],
        [-0.6668],
        [-0.8303],
        [-1.8566],
        [ 0.8562],
        [-0.3528],
        [-0.4386],
        [ 0.4342],
        [ 0.7688]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[1-g1-none-idtype0] ___________________

norm_type = 'none'
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
idtype = torch.int32, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0718E5C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.2274],
        [-0.2330],
        [ 1.2957],
        [ 1.4184],
        [-1.4058],
        [ 1.8791],
        [ 1.5203],
        [-0.4152],
        [ 0.5592],
        [ 0.4297]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[1-g1-none-idtype1] ___________________

norm_type = 'none'
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
idtype = torch.int64, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0188B288>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.9057],
        [ 1.3922],
        [ 0.5030],
        [-0.1469],
        [-2.7244],
        [ 0.0816],
        [ 0.4461],
        [-0.6698],
        [ 1.0289],
        [-1.2757]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[1-g2-both-idtype0] ___________________

norm_type = 'both'
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03DF8AC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.3752],
        [ 0.4392],
        [ 0.0116],
        [-0.0381],
        [-0.7210],
        [-2.1855],
        [ 0.2577],
        [ 0.8623],
        [-1.7276],
        [-2.3790]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[1-g2-both-idtype1] ___________________

norm_type = 'both'
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03DD0A88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.1585],
        [ 0.2463],
        [-0.6888],
        [-0.5436],
        [-1.0343],
        [-0.5276],
        [-0.7115],
        [ 0.1568],
        [-2.7433],
        [-0.0563]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[1-g2-right-idtype0] __________________

norm_type = 'right'
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07247348>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.2162],
        [-0.6256],
        [ 0.9568],
        [ 0.5510],
        [ 0.1020],
        [-0.0593],
        [-0.0745],
        [ 1.5067],
        [ 0.5401],
        [ 0.8200]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[1-g2-right-idtype1] __________________

norm_type = 'right'
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07040F08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.1530],
        [-0.7762],
        [-0.1017],
        [ 0.3125],
        [ 0.0215],
        [ 0.7056],
        [-4.0200],
        [-0.8108],
        [ 2.0894],
        [ 0.3449]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[1-g2-none-idtype0] ___________________

norm_type = 'none'
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03D06548>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.0819],
        [-1.6520],
        [-0.6316],
        [ 1.1696],
        [-1.4296],
        [ 1.4547],
        [ 1.0833],
        [-0.9426],
        [ 0.6800],
        [-0.3857]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[1-g2-none-idtype1] ___________________

norm_type = 'none'
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03D83048>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.4034],
        [ 1.0044],
        [-1.3102],
        [ 0.0622],
        [-1.7712],
        [ 0.2954],
        [ 1.6652],
        [ 0.6685],
        [-0.3154],
        [-0.3576]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[1-g3-both-idtype0] ___________________

norm_type = 'both'
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int32, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0736EAC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.1242],
        [ 0.2950],
        [-0.3126],
        [ 0.2107],
        [ 0.1367],
        [ 1.6240],
        [-1.2324],
        [-0.0837],
        [ 0.6581],
        [ 0.3335]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[1-g3-both-idtype1] ___________________

norm_type = 'both'
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int64, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF016C4CC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.0131],
        [-0.5568],
        [-1.4773],
        [-0.3668],
        [-0.1241],
        [-0.4529],
        [ 1.4737],
        [ 0.5293],
        [ 0.5611],
        [-0.3041]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[1-g3-right-idtype0] __________________

norm_type = 'right'
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int32, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF071A2288>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.3975],
        [-1.4615],
        [ 0.2483],
        [ 1.1704],
        [-0.9989],
        [ 1.9454],
        [ 1.0105],
        [-3.3315],
        [ 1.2930],
        [-0.0354]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[1-g3-right-idtype1] __________________

norm_type = 'right'
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int64, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03D14588>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.2123],
        [ 0.8149],
        [-2.5953],
        [ 0.1614],
        [ 2.0813],
        [ 0.8167],
        [-0.4310],
        [ 2.6282],
        [-2.9950],
        [-0.7267]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[1-g3-none-idtype0] ___________________

norm_type = 'none'
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int32, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03D6DE08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-3.9283],
        [ 0.5980],
        [-1.5557],
        [ 0.6387],
        [-1.8197],
        [-0.2934],
        [-0.2531],
        [ 0.7039],
        [ 2.2292],
        [ 0.8129]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[1-g3-none-idtype1] ___________________

norm_type = 'none'
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int64, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03D78848>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.1846],
        [-0.3051],
        [ 0.9604],
        [-0.1530],
        [-0.2766],
        [-0.4155],
        [-1.3227],
        [-0.1277],
        [ 1.8083],
        [ 1.3529]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[1-g4-both-idtype0] ___________________

norm_type = 'both'
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF016ED688>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.7385],
        [-0.0813],
        [-0.5557],
        [-1.0997],
        [-0.0441],
        [-0.3412],
        [ 0.9611],
        [-0.8789],
        [-0.4156]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[1-g4-both-idtype1] ___________________

norm_type = 'both'
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03D75048>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.2416],
        [ 0.4034],
        [ 1.2644],
        [-1.2527],
        [ 2.0439],
        [ 0.6820],
        [ 0.0128],
        [-0.1895],
        [-0.0925]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[1-g4-right-idtype0] __________________

norm_type = 'right'
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0181C788>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.8641],
        [ 1.3333],
        [-0.7130],
        [ 2.6710],
        [-0.4644],
        [-1.3119],
        [-1.0645],
        [-0.0212],
        [ 0.7361]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[1-g4-right-idtype1] __________________

norm_type = 'right'
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF073BA248>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 2.4108],
        [-0.0280],
        [ 0.6262],
        [ 0.1617],
        [-2.1480],
        [ 1.3224],
        [-0.1878],
        [-0.4613],
        [-0.7966]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[1-g4-none-idtype0] ___________________

norm_type = 'none'
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF017FF788>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.6051],
        [ 1.9308],
        [ 1.9788],
        [ 1.3928],
        [ 0.3757],
        [-0.6361],
        [ 2.0586],
        [-0.5461],
        [ 2.9449]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[1-g4-none-idtype1] ___________________

norm_type = 'none'
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03DF3EC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.2084],
        [-1.7593],
        [-1.2945],
        [ 0.5207],
        [-0.3053],
        [ 0.5228],
        [ 1.4860],
        [ 0.1124],
        [ 1.8734]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[1-g5-both-idtype0] ___________________

norm_type = 'both'
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01082448>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.4974],
        [-0.0746],
        [ 1.0799],
        [-1.0802],
        [ 0.9766],
        [ 0.7196],
        [-0.5129],
        [ 1.4496],
        [-0.1521],
        [-1.2853]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[1-g5-both-idtype1] ___________________

norm_type = 'both'
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE122E92C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.2248],
        [ 0.1445],
        [-0.4187],
        [ 0.1567],
        [ 0.5457],
        [-0.0232],
        [-0.2720],
        [ 0.1133],
        [-0.5212],
        [-0.5260]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[1-g5-right-idtype0] __________________

norm_type = 'right'
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07078788>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.7863],
        [ 0.2211],
        [ 1.0069],
        [-0.6882],
        [ 0.3965],
        [ 0.7213],
        [-1.1229],
        [ 0.6901],
        [-0.4128],
        [-0.3472]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[1-g5-right-idtype1] __________________

norm_type = 'right'
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0192BCC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 2.8666],
        [-1.3170],
        [-1.7080],
        [-1.4374],
        [-1.3648],
        [-0.3498],
        [-0.3272],
        [ 1.2613],
        [ 1.0615],
        [-1.7627]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[1-g5-none-idtype0] ___________________

norm_type = 'none'
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE12336B48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.1201],
        [ 1.8022],
        [-2.1805],
        [-1.3612],
        [ 0.1846],
        [ 0.3148],
        [-1.0696],
        [-1.1247],
        [-0.5497],
        [ 1.3347]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[1-g5-none-idtype1] ___________________

norm_type = 'none'
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01518C88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.7504],
        [ 0.6316],
        [-0.7456],
        [-1.2564],
        [-1.4312],
        [-0.2252],
        [ 3.1701],
        [-0.3720],
        [ 0.3011],
        [-0.8195]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[1-g6-both-idtype0] ___________________

norm_type = 'both', g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
idtype = torch.int32, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1234BC48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.5740],
        [-0.6218],
        [-1.0874],
        [ 3.8511],
        [ 0.3914],
        [-0.5728]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[1-g6-both-idtype1] ___________________

norm_type = 'both', g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
idtype = torch.int64, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07066C88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.0634],
        [-0.4338],
        [ 0.2534],
        [-0.1352],
        [-0.6974],
        [-1.7371]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[1-g6-right-idtype0] __________________

norm_type = 'right', g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
idtype = torch.int32, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03D3AF88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.3309],
        [-1.8538],
        [ 0.8958],
        [-1.0850],
        [-3.3681],
        [-3.0249]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[1-g6-right-idtype1] __________________

norm_type = 'right', g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
idtype = torch.int64, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE12352BC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-2.1006],
        [ 1.4585],
        [-0.9599],
        [ 2.0109],
        [ 0.7593],
        [-0.7545]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[1-g6-none-idtype0] ___________________

norm_type = 'none', g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
idtype = torch.int32, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE121EBBC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.8062],
        [-0.5509],
        [ 0.4601],
        [ 0.7325],
        [ 0.0354],
        [-0.2172]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[1-g6-none-idtype1] ___________________

norm_type = 'none', g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
idtype = torch.int64, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF072E4888>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.2768],
        [-0.1300],
        [ 0.3197],
        [-0.4221],
        [-0.3408],
        [ 0.4516]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[1-g7-both-idtype0] ___________________

norm_type = 'both'
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07079708>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.1307],
        [-1.1302],
        [ 0.1583],
        [ 1.2993],
        [-0.8542],
        [ 1.4163],
        [ 0.0415],
        [-0.8618],
        [ 0.6743],
        [ 1.1464]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[1-g7-both-idtype1] ___________________

norm_type = 'both'
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03E03FC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.0031],
        [ 1.4958],
        [-0.9031],
        [-0.5207],
        [ 0.6594],
        [ 0.9867],
        [ 0.1700],
        [-2.2626],
        [ 0.8617],
        [ 1.6734]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[1-g7-right-idtype0] __________________

norm_type = 'right'
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03D27C08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.7116],
        [ 1.6214],
        [ 0.7151],
        [ 0.8486],
        [ 0.5530],
        [ 2.7667],
        [ 0.9301],
        [-1.6352],
        [-1.1589],
        [-0.5660]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[1-g7-right-idtype1] __________________

norm_type = 'right'
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07163108>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.7224],
        [ 0.1494],
        [-1.1859],
        [-1.0782],
        [-1.0921],
        [-0.7552],
        [-1.2500],
        [-0.6734],
        [-0.0864],
        [ 0.1035]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[1-g7-none-idtype0] ___________________

norm_type = 'none'
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03E1B6C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.7045],
        [-2.4822],
        [ 1.0656],
        [-0.9557],
        [ 0.7185],
        [ 0.9086],
        [-0.3823],
        [ 0.0951],
        [-0.6454],
        [-2.5429]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[1-g7-none-idtype1] ___________________

norm_type = 'none'
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE12326108>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.3929],
        [ 1.7412],
        [-2.3667],
        [ 0.3414],
        [-0.2181],
        [-1.9539],
        [-0.1957],
        [-0.7050],
        [-2.2120],
        [ 2.3472]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[2-g0-both-idtype0] ___________________

norm_type = 'both'
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int32, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03CC6848>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.0535,  0.2615],
        [-0.1835, -0.3259]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[2-g0-both-idtype1] ___________________

norm_type = 'both'
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int64, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07324848>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.6879, -0.6169],
        [-0.3183,  0.1158]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[2-g0-right-idtype0] __________________

norm_type = 'right'
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int32, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03D15748>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.4571,  0.3367],
        [ 1.2559, -1.3124]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[2-g0-right-idtype1] __________________

norm_type = 'right'
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int64, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07246E08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.8919,  0.5680],
        [ 0.9213,  0.5819]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[2-g0-none-idtype0] ___________________

norm_type = 'none'
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int32, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF017BF9C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.7677, -0.6142],
        [ 0.4679,  0.8218]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[2-g0-none-idtype1] ___________________

norm_type = 'none'
g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int64, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03D78AC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.5966, -0.7014],
        [ 0.4207, -1.0862]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[2-g1-both-idtype0] ___________________

norm_type = 'both'
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
idtype = torch.int32, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03DFE408>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 6.3572e-02, -2.3987e-01],
        [ 1.4900e+00,  5.2945e-01],
        [ 9.0326e-02, -6.0087e-01],
        [ ....4811e-01,  6.1425e-01],
        [-2.1378e+00, -5.2975e-02],
        [-3.0054e-01, -5.1798e-02]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[2-g1-both-idtype1] ___________________

norm_type = 'both'
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
idtype = torch.int64, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE121EA948>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.3212, -0.2879],
        [-0.0113, -0.1146],
        [ 0.2062, -0.1645],
        [ 0.0081,  0.0204],
      ...9, -1.1613],
        [-3.7358, -0.8968],
        [-0.9263,  1.6934],
        [ 0.9882, -1.1352]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[2-g1-right-idtype0] __________________

norm_type = 'right'
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
idtype = torch.int32, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07336E88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.2161, -3.3258],
        [ 0.3899, -0.4552],
        [ 1.3248, -1.3175],
        [-2.8084,  1.9503],
      ...3, -2.0930],
        [-0.5776, -0.7462],
        [-2.0418,  0.7217],
        [ 2.4984,  0.0508]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[2-g1-right-idtype1] __________________

norm_type = 'right'
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
idtype = torch.int64, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01700888>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.0740, -0.7849],
        [ 1.7928,  0.2345],
        [ 0.5456, -0.7928],
        [ 0.8768, -1.2575],
      ...2, -1.0160],
        [-0.2619, -1.5572],
        [-1.1251,  0.3327],
        [ 1.1032, -0.0503]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[2-g1-none-idtype0] ___________________

norm_type = 'none'
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
idtype = torch.int32, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03DFAA88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.4197,  0.8359],
        [ 1.3335,  0.4224],
        [ 1.5052, -0.8819],
        [-0.6998,  0.7681],
      ...9, -1.4529],
        [-2.2650,  0.3032],
        [ 0.6548, -0.0101],
        [ 0.7254,  0.6259]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[2-g1-none-idtype1] ___________________

norm_type = 'none'
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
idtype = torch.int64, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01802548>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.7308,  1.8016],
        [-0.6890,  0.5785],
        [ 1.1303, -1.7625],
        [ 2.9705, -1.3467],
      ...8, -3.2297],
        [ 0.7221,  1.0870],
        [-0.4338, -0.9184],
        [-0.7838, -0.7252]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[2-g2-both-idtype0] ___________________

norm_type = 'both'
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0735C508>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.2431,  1.6273],
        [ 0.1684,  0.2813],
        [-0.9821, -1.1271],
        [ 0.5550,  0.6853],
      ...1, -1.7264],
        [-1.3279, -2.6044],
        [ 0.0895, -1.2487],
        [ 1.7991,  0.3724]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[2-g2-both-idtype1] ___________________

norm_type = 'both'
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF073C1348>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.5246, -0.4132],
        [ 0.5062, -0.0038],
        [-0.5340,  0.3477],
        [-0.2291,  0.0294],
      ...1,  1.1515],
        [ 0.9314, -0.5209],
        [ 0.2423, -1.3572],
        [ 1.0417, -1.1523]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[2-g2-right-idtype0] __________________

norm_type = 'right'
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF071BE848>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.9216, -1.6763],
        [ 1.7834, -2.0112],
        [-0.1023,  0.4966],
        [ 0.5120,  0.4627],
      ...5, -0.3781],
        [-1.0866,  0.9257],
        [ 1.3443, -0.2144],
        [-0.9199, -0.9447]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[2-g2-right-idtype1] __________________

norm_type = 'right'
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03CD6108>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.4034,  0.1835],
        [-0.3035, -1.2629],
        [-0.2233, -1.7079],
        [ 1.8583,  1.5782],
      ...4,  0.9317],
        [ 2.1824,  1.6918],
        [ 2.3706,  1.0933],
        [ 1.6049,  1.4703]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[2-g2-none-idtype0] ___________________

norm_type = 'none'
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03E01808>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.8105, -0.7356],
        [-0.6959,  0.2979],
        [-1.3562,  0.1273],
        [-1.1062,  0.5512],
      ...1, -0.0236],
        [-0.7779,  1.6138],
        [-0.9351,  0.6571],
        [-0.9040, -0.3278]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[2-g2-none-idtype1] ___________________

norm_type = 'none'
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03DA6AC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.5904, -0.2524],
        [ 1.4974,  0.6287],
        [-1.7777,  0.2143],
        [-0.5061,  0.6147],
      ...8, -1.0503],
        [ 1.1370,  1.0193],
        [-3.5506,  1.2077],
        [ 0.2752,  0.4062]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[2-g3-both-idtype0] ___________________

norm_type = 'both'
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int32, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1225AB48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.3493, -0.9388],
        [-0.2284,  0.9643],
        [ 0.3206,  0.4100],
        [-0.0895, -1.4870],
      ...9, -0.1860],
        [ 1.2083, -0.5335],
        [ 2.5894,  1.3947],
        [ 0.4751, -0.6062]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[2-g3-both-idtype1] ___________________

norm_type = 'both'
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int64, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03D1BF48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.2779,  0.5695],
        [-0.2369, -0.9207],
        [ 0.1799,  0.7140],
        [-0.3010, -0.6443],
      ...4,  0.2508],
        [-0.7480,  0.4197],
        [-0.8351,  2.5504],
        [ 1.7923,  0.7883]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[2-g3-right-idtype0] __________________

norm_type = 'right'
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int32, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE145C01C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.0159, -0.3147],
        [ 0.3722, -0.1600],
        [-1.4089, -1.2514],
        [ 0.5023, -0.6940],
      ...4,  0.3467],
        [-0.6323, -0.3194],
        [ 0.0221,  0.3430],
        [ 1.2794,  2.6821]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[2-g3-right-idtype1] __________________

norm_type = 'right'
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int64, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE12223F88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.7702, -1.0171],
        [-0.3998,  0.0453],
        [-2.9881, -0.3333],
        [-0.5108,  0.9458],
      ...5,  0.7667],
        [-0.2178,  0.1502],
        [-0.5126, -0.6607],
        [ 0.6344, -0.1215]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[2-g3-none-idtype0] ___________________

norm_type = 'none'
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int32, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0198EAC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.6036,  1.1359],
        [ 0.4259,  1.2713],
        [-0.4820,  1.5359],
        [ 0.5081, -3.0142],
      ...6, -2.0203],
        [ 0.7775,  1.7450],
        [-0.4477, -1.1044],
        [ 0.0479, -1.2119]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[2-g3-none-idtype1] ___________________

norm_type = 'none'
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int64, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF073DE5C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.3831, -0.1623],
        [-0.2736, -1.5970],
        [ 0.0663,  0.4000],
        [-0.2874,  0.9317],
      ...0,  0.8166],
        [-1.4072,  0.5512],
        [ 0.4142,  0.0910],
        [-2.1473,  0.1844]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[2-g4-both-idtype0] ___________________

norm_type = 'both'
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE122E6FC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.1397, -0.2550],
        [ 0.5466,  0.7480],
        [ 0.6702,  1.0812],
        [ 0.7443,  0.4168],
      ...9, -1.0333],
        [-2.1606, -2.4702],
        [ 0.7381,  1.0124],
        [ 0.2424, -0.2181]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[2-g4-both-idtype1] ___________________

norm_type = 'both'
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE122584C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.3001, -0.2268],
        [ 0.2175, -0.9202],
        [ 0.5277,  1.1526],
        [ 1.2172,  3.2172],
      ...0, -0.7017],
        [-0.3155, -1.4023],
        [ 0.2472,  1.4226],
        [ 0.4283, -2.9465]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[2-g4-right-idtype0] __________________

norm_type = 'right'
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE12371B48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.0733,  1.1845],
        [ 0.6701, -1.0438],
        [-1.0266,  1.9384],
        [ 0.3960, -0.9826],
      ...9,  0.2036],
        [ 0.4507, -1.2627],
        [ 0.5897, -1.5299],
        [-0.8876,  1.9282]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[2-g4-right-idtype1] __________________

norm_type = 'right'
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE122F8648>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 2.6527, -1.6535],
        [-1.2821, -2.0337],
        [ 0.9056,  0.6876],
        [-0.8874,  0.8351],
      ...8, -3.1637],
        [ 0.2045,  0.0760],
        [ 2.8796,  1.1828],
        [-3.8797, -0.9333]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[2-g4-none-idtype0] ___________________

norm_type = 'none'
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07295888>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.4946,  2.6052],
        [-2.3344,  0.6533],
        [ 0.8924, -0.6174],
        [-1.4227, -0.4393],
      ...4,  0.0735],
        [ 0.8834,  2.7240],
        [ 2.4323, -0.0342],
        [-0.1249, -0.3558]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[2-g4-none-idtype1] ___________________

norm_type = 'none'
g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03D6C208>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.9265,  0.3782],
        [-1.2697,  0.3816],
        [-0.8689, -1.7110],
        [-0.1680, -0.8853],
      ...9,  1.8692],
        [ 0.8399, -0.8960],
        [-0.3965, -0.8522],
        [ 2.4670,  1.6708]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[2-g5-both-idtype0] ___________________

norm_type = 'both'
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03D43148>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.1287,  0.8881],
        [ 0.7294, -0.4425],
        [-0.9015,  0.6449],
        [ 0.6682, -0.3910],
      ...5,  0.0996],
        [ 1.3359,  1.1110],
        [ 0.1402,  0.2153],
        [ 0.9476, -0.3550]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[2-g5-both-idtype1] ___________________

norm_type = 'both'
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01948808>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.1358, -0.5234],
        [ 0.0036, -0.4743],
        [ 0.1652,  1.9433],
        [-0.7140,  0.5022],
      ...3, -0.2514],
        [ 1.2993, -0.1255],
        [-0.9504, -0.8335],
        [-0.3711,  0.3592]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[2-g5-right-idtype0] __________________

norm_type = 'right'
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1237C088>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.7265,  0.9246],
        [-1.2256, -1.9229],
        [ 0.2901, -0.4275],
        [-0.1803, -1.2172],
      ...3,  1.2344],
        [-0.1240,  0.6153],
        [ 0.4432, -0.4997],
        [-1.4621, -2.1706]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[2-g5-right-idtype1] __________________

norm_type = 'right'
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03DA3408>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.2621, -0.5773],
        [-1.1340, -0.9531],
        [-1.9158, -0.6562],
        [ 0.0861, -0.2835],
      ...8, -0.3894],
        [ 0.7520,  0.8223],
        [-0.8265,  2.0287],
        [ 0.0596,  0.5591]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[2-g5-none-idtype0] ___________________

norm_type = 'none'
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1225F548>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.9798, -2.1137],
        [-0.6788, -0.6855],
        [-1.9849, -1.5732],
        [ 0.2896,  0.4231],
      ...6, -1.7439],
        [ 0.2564,  0.3711],
        [-0.5102,  0.7334],
        [-0.4523, -1.2525]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[2-g5-none-idtype1] ___________________

norm_type = 'none'
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE123A4508>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.2330, -0.8414],
        [-0.6792,  0.3635],
        [-0.0952, -0.7079],
        [ 1.1991, -0.1358],
      ...7, -0.0605],
        [-0.0884, -0.8879],
        [-0.3777,  0.4511],
        [ 0.1352, -0.5070]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[2-g6-both-idtype0] ___________________

norm_type = 'both', g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
idtype = torch.int32, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07049448>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.2521,  1.2476],
        [-0.2045, -0.0489],
        [ 0.7095, -1.5968],
        [ 0.4035,  1.7978],
        [ 0.4832,  3.3781],
        [ 0.3848,  1.4405]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[2-g6-both-idtype1] ___________________

norm_type = 'both', g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
idtype = torch.int64, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE121EA1C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.0866,  3.4502],
        [-0.4442,  0.4656],
        [ 0.8157,  0.7665],
        [-0.1684,  0.3274],
        [-0.1779, -0.3314],
        [-0.2699,  0.6767]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[2-g6-right-idtype0] __________________

norm_type = 'right', g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
idtype = torch.int32, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03E4ACC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.8771, -1.3386],
        [ 1.0482,  1.1702],
        [ 2.1215,  3.8488],
        [-0.8253, -2.5760],
        [ 0.8233, -0.1761],
        [-0.1109, -0.4918]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[2-g6-right-idtype1] __________________

norm_type = 'right', g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
idtype = torch.int64, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01700488>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.7307, -0.5785],
        [-0.2896, -1.2098],
        [-0.2114,  1.6255],
        [ 0.3616, -0.6996],
        [-0.4073,  0.6564],
        [ 0.5113,  0.7748]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[2-g6-none-idtype0] ___________________

norm_type = 'none', g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
idtype = torch.int32, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03DB9EC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.3509,  0.8456],
        [ 1.7724,  0.1835],
        [-4.2332, -1.7631],
        [ 0.0956, -1.1362],
        [ 0.0673,  0.3801],
        [-2.3810,  0.6680]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[2-g6-none-idtype1] ___________________

norm_type = 'none', g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)
idtype = torch.int64, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF072944C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.1505, -1.0392],
        [-0.5703, -0.1622],
        [ 1.1073, -0.7158],
        [ 1.9212, -1.2566],
        [ 0.7413,  0.6405],
        [-0.3005, -0.2180]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[2-g7-both-idtype0] ___________________

norm_type = 'both'
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03D06088>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.1269, -0.1796],
        [-0.4185, -0.3096],
        [ 0.8369,  0.2194],
        [ 0.7925,  0.5175],
      ...1, -0.9139],
        [ 2.9453,  0.1506],
        [-1.6357, -0.0334],
        [ 3.2946,  0.7874]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[2-g7-both-idtype1] ___________________

norm_type = 'both'
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0146E5C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.2975, -0.4201],
        [ 0.6562, -0.4391],
        [-1.3458,  0.1775],
        [-0.3952,  0.2400],
      ...0, -1.4653],
        [ 0.1183, -0.0391],
        [-0.1510,  0.1255],
        [ 1.0389, -0.7123]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[2-g7-right-idtype0] __________________

norm_type = 'right'
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1228E648>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.5662,  0.9773],
        [-0.1745, -0.1897],
        [-0.2755, -1.0459],
        [-0.8420,  1.1128],
      ...7, -0.7359],
        [-0.0192,  0.4320],
        [-2.0010,  0.3151],
        [-0.6168, -0.4310]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[2-g7-right-idtype1] __________________

norm_type = 'right'
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1235D588>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.3473, -1.4332],
        [-0.0938,  2.1125],
        [ 0.2526,  0.1867],
        [-1.5942,  1.0249],
      ...3, -0.1221],
        [-0.2958, -1.8970],
        [ 0.6376,  0.2688],
        [ 0.2782,  1.5352]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[2-g7-none-idtype0] ___________________

norm_type = 'none'
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0714B388>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.6513,  1.9487],
        [-0.7282,  1.3980],
        [-0.4052,  0.9704],
        [-0.0414, -0.1981],
      ...5, -1.9093],
        [-1.5928,  0.6559],
        [-0.4361, -0.7507],
        [-0.0270,  0.2611]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_dense_graph_conv[2-g7-none-idtype1] ___________________

norm_type = 'none'
g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('norm_type', ['both', 'right', 'none'])
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_graph_conv(norm_type, g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        # TODO(minjie): enable the following option after #1385
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        conv = nn.GraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv = nn.DenseGraphConv(5, out_dim, norm=norm_type, bias=True)
        dense_conv.weight.data = conv.weight.data
        dense_conv.bias.data = conv.bias.data
        feat = F.randn((g.number_of_src_nodes(), 5))
        conv = conv.to(ctx)
        dense_conv = dense_conv.to(ctx)
>       out_conv = conv(g, feat)

tests\pytorch\test_nn.py:932: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03E324C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.9374,  1.1745],
        [ 1.1024,  0.4289],
        [-0.2881, -0.2376],
        [ 2.5788, -1.1294],
      ...6,  1.1821],
        [-0.4488, -0.5485],
        [-0.2000, -1.7327],
        [-2.4283, -1.7645]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_____________________ test_dense_sage_conv[1-g0-idtype0] ______________________

g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int32, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_sage_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        sage = nn.SAGEConv(5, out_dim, 'gcn')
        dense_sage = nn.DenseSAGEConv(5, out_dim)
        dense_sage.fc.weight.data = sage.fc_neigh.weight.data
        dense_sage.fc.bias.data = sage.bias.data
        if len(g.ntypes) == 2:
            feat = (
                F.randn((g.number_of_src_nodes(), 5)),
                F.randn((g.number_of_dst_nodes(), 5))
            )
        else:
            feat = F.randn((g.number_of_nodes(), 5))
        sage = sage.to(ctx)
        dense_sage = dense_sage.to(ctx)
>       out_sage = sage(g, feat)

tests\pytorch\test_nn.py:956: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:249: in forward
    graph.update_all(msg_fn, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF070E6BC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.4381],
        [ 3.1397]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_____________________ test_dense_sage_conv[1-g0-idtype1] ______________________

g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int64, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_sage_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        sage = nn.SAGEConv(5, out_dim, 'gcn')
        dense_sage = nn.DenseSAGEConv(5, out_dim)
        dense_sage.fc.weight.data = sage.fc_neigh.weight.data
        dense_sage.fc.bias.data = sage.bias.data
        if len(g.ntypes) == 2:
            feat = (
                F.randn((g.number_of_src_nodes(), 5)),
                F.randn((g.number_of_dst_nodes(), 5))
            )
        else:
            feat = F.randn((g.number_of_nodes(), 5))
        sage = sage.to(ctx)
        dense_sage = dense_sage.to(ctx)
>       out_sage = sage(g, feat)

tests\pytorch\test_nn.py:956: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:249: in forward
    graph.update_all(msg_fn, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03D277C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[1.7132],
        [0.4321]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_____________________ test_dense_sage_conv[1-g1-idtype0] ______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
idtype = torch.int32, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_sage_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        sage = nn.SAGEConv(5, out_dim, 'gcn')
        dense_sage = nn.DenseSAGEConv(5, out_dim)
        dense_sage.fc.weight.data = sage.fc_neigh.weight.data
        dense_sage.fc.bias.data = sage.bias.data
        if len(g.ntypes) == 2:
            feat = (
                F.randn((g.number_of_src_nodes(), 5)),
                F.randn((g.number_of_dst_nodes(), 5))
            )
        else:
            feat = F.randn((g.number_of_nodes(), 5))
        sage = sage.to(ctx)
        dense_sage = dense_sage.to(ctx)
>       out_sage = sage(g, feat)

tests\pytorch\test_nn.py:956: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:249: in forward
    graph.update_all(msg_fn, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0740C448>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.0178],
        [-4.2534],
        [ 8.1334],
        [-0.2057],
        [ 2.5950],
        [-1.2583],
        [ 0.5171],
        [ 0.0298],
        [-1.0775],
        [-1.0520]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_____________________ test_dense_sage_conv[1-g1-idtype1] ______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
idtype = torch.int64, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_sage_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        sage = nn.SAGEConv(5, out_dim, 'gcn')
        dense_sage = nn.DenseSAGEConv(5, out_dim)
        dense_sage.fc.weight.data = sage.fc_neigh.weight.data
        dense_sage.fc.bias.data = sage.bias.data
        if len(g.ntypes) == 2:
            feat = (
                F.randn((g.number_of_src_nodes(), 5)),
                F.randn((g.number_of_dst_nodes(), 5))
            )
        else:
            feat = F.randn((g.number_of_nodes(), 5))
        sage = sage.to(ctx)
        dense_sage = dense_sage.to(ctx)
>       out_sage = sage(g, feat)

tests\pytorch\test_nn.py:956: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:249: in forward
    graph.update_all(msg_fn, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03D8AF48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.2534],
        [-1.4586],
        [ 0.2344],
        [ 1.5878],
        [-0.3027],
        [-0.5381],
        [-1.3500],
        [-0.6228],
        [ 0.9664],
        [-0.1548]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_____________________ test_dense_sage_conv[1-g2-idtype0] ______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_sage_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        sage = nn.SAGEConv(5, out_dim, 'gcn')
        dense_sage = nn.DenseSAGEConv(5, out_dim)
        dense_sage.fc.weight.data = sage.fc_neigh.weight.data
        dense_sage.fc.bias.data = sage.bias.data
        if len(g.ntypes) == 2:
            feat = (
                F.randn((g.number_of_src_nodes(), 5)),
                F.randn((g.number_of_dst_nodes(), 5))
            )
        else:
            feat = F.randn((g.number_of_nodes(), 5))
        sage = sage.to(ctx)
        dense_sage = dense_sage.to(ctx)
>       out_sage = sage(g, feat)

tests\pytorch\test_nn.py:956: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:249: in forward
    graph.update_all(msg_fn, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01505C48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.3687],
        [-2.4167],
        [ 0.2740],
        [-0.9262],
        [ 1.2199],
        [ 0.3057],
        [ 1.0072],
        [-1.2542],
        [-0.0223],
        [-0.8523]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_____________________ test_dense_sage_conv[1-g2-idtype1] ______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_sage_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        sage = nn.SAGEConv(5, out_dim, 'gcn')
        dense_sage = nn.DenseSAGEConv(5, out_dim)
        dense_sage.fc.weight.data = sage.fc_neigh.weight.data
        dense_sage.fc.bias.data = sage.bias.data
        if len(g.ntypes) == 2:
            feat = (
                F.randn((g.number_of_src_nodes(), 5)),
                F.randn((g.number_of_dst_nodes(), 5))
            )
        else:
            feat = F.randn((g.number_of_nodes(), 5))
        sage = sage.to(ctx)
        dense_sage = dense_sage.to(ctx)
>       out_sage = sage(g, feat)

tests\pytorch\test_nn.py:956: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:249: in forward
    graph.update_all(msg_fn, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03DC6F88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.1472],
        [ 2.9466],
        [-2.1785],
        [ 0.3249],
        [-0.1859],
        [-0.4245],
        [-1.1086],
        [ 0.5914],
        [-0.4578],
        [ 1.2486]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_____________________ test_dense_sage_conv[1-g3-idtype0] ______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int32, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_sage_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        sage = nn.SAGEConv(5, out_dim, 'gcn')
        dense_sage = nn.DenseSAGEConv(5, out_dim)
        dense_sage.fc.weight.data = sage.fc_neigh.weight.data
        dense_sage.fc.bias.data = sage.bias.data
        if len(g.ntypes) == 2:
            feat = (
                F.randn((g.number_of_src_nodes(), 5)),
                F.randn((g.number_of_dst_nodes(), 5))
            )
        else:
            feat = F.randn((g.number_of_nodes(), 5))
        sage = sage.to(ctx)
        dense_sage = dense_sage.to(ctx)
>       out_sage = sage(g, feat)

tests\pytorch\test_nn.py:956: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:249: in forward
    graph.update_all(msg_fn, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03D0FA08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.0970],
        [-3.1663],
        [ 3.3163],
        [ 0.2478],
        [ 1.7958],
        [-0.4478],
        [-0.1523],
        [ 4.3373],
        [ 2.5558],
        [-1.8230]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_____________________ test_dense_sage_conv[1-g3-idtype1] ______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int64, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_sage_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        sage = nn.SAGEConv(5, out_dim, 'gcn')
        dense_sage = nn.DenseSAGEConv(5, out_dim)
        dense_sage.fc.weight.data = sage.fc_neigh.weight.data
        dense_sage.fc.bias.data = sage.bias.data
        if len(g.ntypes) == 2:
            feat = (
                F.randn((g.number_of_src_nodes(), 5)),
                F.randn((g.number_of_dst_nodes(), 5))
            )
        else:
            feat = F.randn((g.number_of_nodes(), 5))
        sage = sage.to(ctx)
        dense_sage = dense_sage.to(ctx)
>       out_sage = sage(g, feat)

tests\pytorch\test_nn.py:956: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:249: in forward
    graph.update_all(msg_fn, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE14A92308>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.4889],
        [ 0.1853],
        [-3.6972],
        [-0.7642],
        [-3.8500],
        [-0.1650],
        [ 1.4379],
        [-1.4547],
        [ 1.9583],
        [ 2.4875]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_____________________ test_dense_sage_conv[1-g4-idtype0] ______________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_sage_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        sage = nn.SAGEConv(5, out_dim, 'gcn')
        dense_sage = nn.DenseSAGEConv(5, out_dim)
        dense_sage.fc.weight.data = sage.fc_neigh.weight.data
        dense_sage.fc.bias.data = sage.bias.data
        if len(g.ntypes) == 2:
            feat = (
                F.randn((g.number_of_src_nodes(), 5)),
                F.randn((g.number_of_dst_nodes(), 5))
            )
        else:
            feat = F.randn((g.number_of_nodes(), 5))
        sage = sage.to(ctx)
        dense_sage = dense_sage.to(ctx)
>       out_sage = sage(g, feat)

tests\pytorch\test_nn.py:956: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:249: in forward
    graph.update_all(msg_fn, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF014E9508>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-3.2967],
        [-2.0195],
        [-1.9533],
        [ 0.6717],
        [ 1.0001],
        [-2.0296],
        [-2.1718],
        [ 5.2240],
        [ 1.6120]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_____________________ test_dense_sage_conv[1-g4-idtype1] ______________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_sage_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        sage = nn.SAGEConv(5, out_dim, 'gcn')
        dense_sage = nn.DenseSAGEConv(5, out_dim)
        dense_sage.fc.weight.data = sage.fc_neigh.weight.data
        dense_sage.fc.bias.data = sage.bias.data
        if len(g.ntypes) == 2:
            feat = (
                F.randn((g.number_of_src_nodes(), 5)),
                F.randn((g.number_of_dst_nodes(), 5))
            )
        else:
            feat = F.randn((g.number_of_nodes(), 5))
        sage = sage.to(ctx)
        dense_sage = dense_sage.to(ctx)
>       out_sage = sage(g, feat)

tests\pytorch\test_nn.py:956: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:249: in forward
    graph.update_all(msg_fn, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07309348>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.7903],
        [-1.7348],
        [ 3.2055],
        [-0.0260],
        [ 1.7800],
        [ 1.0049],
        [ 5.5662],
        [-0.7228],
        [-0.4006]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_____________________ test_dense_sage_conv[1-g5-idtype0] ______________________

g = Graph(num_nodes=5, num_edges=6,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_sage_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        sage = nn.SAGEConv(5, out_dim, 'gcn')
        dense_sage = nn.DenseSAGEConv(5, out_dim)
        dense_sage.fc.weight.data = sage.fc_neigh.weight.data
        dense_sage.fc.bias.data = sage.bias.data
        if len(g.ntypes) == 2:
            feat = (
                F.randn((g.number_of_src_nodes(), 5)),
                F.randn((g.number_of_dst_nodes(), 5))
            )
        else:
            feat = F.randn((g.number_of_nodes(), 5))
        sage = sage.to(ctx)
        dense_sage = dense_sage.to(ctx)
>       out_sage = sage(g, feat)

tests\pytorch\test_nn.py:956: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:249: in forward
    graph.update_all(msg_fn, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF071B3488>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.6542],
        [ 0.3013],
        [ 1.9912],
        [ 1.6111],
        [ 1.0939]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_____________________ test_dense_sage_conv[1-g5-idtype1] ______________________

g = Graph(num_nodes=5, num_edges=6,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_sage_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        sage = nn.SAGEConv(5, out_dim, 'gcn')
        dense_sage = nn.DenseSAGEConv(5, out_dim)
        dense_sage.fc.weight.data = sage.fc_neigh.weight.data
        dense_sage.fc.bias.data = sage.bias.data
        if len(g.ntypes) == 2:
            feat = (
                F.randn((g.number_of_src_nodes(), 5)),
                F.randn((g.number_of_dst_nodes(), 5))
            )
        else:
            feat = F.randn((g.number_of_nodes(), 5))
        sage = sage.to(ctx)
        dense_sage = dense_sage.to(ctx)
>       out_sage = sage(g, feat)

tests\pytorch\test_nn.py:956: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:249: in forward
    graph.update_all(msg_fn, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE12308FC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.5626],
        [ 1.5025],
        [-0.7117],
        [ 0.4166],
        [ 1.7778]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_____________________ test_dense_sage_conv[1-g6-idtype0] ______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_sage_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        sage = nn.SAGEConv(5, out_dim, 'gcn')
        dense_sage = nn.DenseSAGEConv(5, out_dim)
        dense_sage.fc.weight.data = sage.fc_neigh.weight.data
        dense_sage.fc.bias.data = sage.bias.data
        if len(g.ntypes) == 2:
            feat = (
                F.randn((g.number_of_src_nodes(), 5)),
                F.randn((g.number_of_dst_nodes(), 5))
            )
        else:
            feat = F.randn((g.number_of_nodes(), 5))
        sage = sage.to(ctx)
        dense_sage = dense_sage.to(ctx)
>       out_sage = sage(g, feat)

tests\pytorch\test_nn.py:956: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:249: in forward
    graph.update_all(msg_fn, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE12215F48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 4.7120],
        [-3.4943],
        [-4.2556],
        [ 1.2847],
        [ 2.0320],
        [ 0.7146],
        [ 3.4068],
        [ 0.3882],
        [-1.5304],
        [ 0.5456]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_____________________ test_dense_sage_conv[1-g6-idtype1] ______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_sage_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        sage = nn.SAGEConv(5, out_dim, 'gcn')
        dense_sage = nn.DenseSAGEConv(5, out_dim)
        dense_sage.fc.weight.data = sage.fc_neigh.weight.data
        dense_sage.fc.bias.data = sage.bias.data
        if len(g.ntypes) == 2:
            feat = (
                F.randn((g.number_of_src_nodes(), 5)),
                F.randn((g.number_of_dst_nodes(), 5))
            )
        else:
            feat = F.randn((g.number_of_nodes(), 5))
        sage = sage.to(ctx)
        dense_sage = dense_sage.to(ctx)
>       out_sage = sage(g, feat)

tests\pytorch\test_nn.py:956: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:249: in forward
    graph.update_all(msg_fn, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF017C5808>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.8245],
        [ 1.3567],
        [-1.1342],
        [-1.8312],
        [-0.0789],
        [-0.7032],
        [-1.6584],
        [-0.4862],
        [-2.5145],
        [ 1.5917]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_____________________ test_dense_sage_conv[1-g7-idtype0] ______________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int32
out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_sage_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        sage = nn.SAGEConv(5, out_dim, 'gcn')
        dense_sage = nn.DenseSAGEConv(5, out_dim)
        dense_sage.fc.weight.data = sage.fc_neigh.weight.data
        dense_sage.fc.bias.data = sage.bias.data
        if len(g.ntypes) == 2:
            feat = (
                F.randn((g.number_of_src_nodes(), 5)),
                F.randn((g.number_of_dst_nodes(), 5))
            )
        else:
            feat = F.randn((g.number_of_nodes(), 5))
        sage = sage.to(ctx)
        dense_sage = dense_sage.to(ctx)
>       out_sage = sage(g, feat)

tests\pytorch\test_nn.py:956: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:249: in forward
    graph.update_all(msg_fn, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07287048>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.3729],
        [ 0.5553],
        [ 1.4915],
        [ 2.5368],
        [-1.1188],
        [-1.9325]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_____________________ test_dense_sage_conv[1-g7-idtype1] ______________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int64
out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_sage_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        sage = nn.SAGEConv(5, out_dim, 'gcn')
        dense_sage = nn.DenseSAGEConv(5, out_dim)
        dense_sage.fc.weight.data = sage.fc_neigh.weight.data
        dense_sage.fc.bias.data = sage.bias.data
        if len(g.ntypes) == 2:
            feat = (
                F.randn((g.number_of_src_nodes(), 5)),
                F.randn((g.number_of_dst_nodes(), 5))
            )
        else:
            feat = F.randn((g.number_of_nodes(), 5))
        sage = sage.to(ctx)
        dense_sage = dense_sage.to(ctx)
>       out_sage = sage(g, feat)

tests\pytorch\test_nn.py:956: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:249: in forward
    graph.update_all(msg_fn, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07256EC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-2.3041],
        [-0.4078],
        [ 1.7749],
        [ 0.7627],
        [-0.0916],
        [ 0.5965]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_____________________ test_dense_sage_conv[1-g8-idtype0] ______________________

g = Graph(num_nodes={'_U': 4, '_V': 5},
      num_edges={('_U', '_E', '_V'): 6},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int32, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_sage_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        sage = nn.SAGEConv(5, out_dim, 'gcn')
        dense_sage = nn.DenseSAGEConv(5, out_dim)
        dense_sage.fc.weight.data = sage.fc_neigh.weight.data
        dense_sage.fc.bias.data = sage.bias.data
        if len(g.ntypes) == 2:
            feat = (
                F.randn((g.number_of_src_nodes(), 5)),
                F.randn((g.number_of_dst_nodes(), 5))
            )
        else:
            feat = F.randn((g.number_of_nodes(), 5))
        sage = sage.to(ctx)
        dense_sage = dense_sage.to(ctx)
>       out_sage = sage(g, feat)

tests\pytorch\test_nn.py:956: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:249: in forward
    graph.update_all(msg_fn, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0746FE48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.7658],
        [-1.2219],
        [ 0.1969],
        [-1.3367]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_____________________ test_dense_sage_conv[1-g8-idtype1] ______________________

g = Graph(num_nodes={'_U': 4, '_V': 5},
      num_edges={('_U', '_E', '_V'): 6},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int64, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_sage_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        sage = nn.SAGEConv(5, out_dim, 'gcn')
        dense_sage = nn.DenseSAGEConv(5, out_dim)
        dense_sage.fc.weight.data = sage.fc_neigh.weight.data
        dense_sage.fc.bias.data = sage.bias.data
        if len(g.ntypes) == 2:
            feat = (
                F.randn((g.number_of_src_nodes(), 5)),
                F.randn((g.number_of_dst_nodes(), 5))
            )
        else:
            feat = F.randn((g.number_of_nodes(), 5))
        sage = sage.to(ctx)
        dense_sage = dense_sage.to(ctx)
>       out_sage = sage(g, feat)

tests\pytorch\test_nn.py:956: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:249: in forward
    graph.update_all(msg_fn, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE123A4F88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.4340],
        [-1.4126],
        [-0.0205],
        [ 2.3952]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_____________________ test_dense_sage_conv[1-g9-idtype0] ______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_sage_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        sage = nn.SAGEConv(5, out_dim, 'gcn')
        dense_sage = nn.DenseSAGEConv(5, out_dim)
        dense_sage.fc.weight.data = sage.fc_neigh.weight.data
        dense_sage.fc.bias.data = sage.bias.data
        if len(g.ntypes) == 2:
            feat = (
                F.randn((g.number_of_src_nodes(), 5)),
                F.randn((g.number_of_dst_nodes(), 5))
            )
        else:
            feat = F.randn((g.number_of_nodes(), 5))
        sage = sage.to(ctx)
        dense_sage = dense_sage.to(ctx)
>       out_sage = sage(g, feat)

tests\pytorch\test_nn.py:956: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:249: in forward
    graph.update_all(msg_fn, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07393BC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 2.3321],
        [-1.5974],
        [-2.3200],
        [-0.5495],
        [ 2.2735],
        [-1.2528],
        [-1.0960],
        [-1.9671],
        [-3.1997],
        [ 1.0793]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_____________________ test_dense_sage_conv[1-g9-idtype1] ______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(1,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_sage_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        sage = nn.SAGEConv(5, out_dim, 'gcn')
        dense_sage = nn.DenseSAGEConv(5, out_dim)
        dense_sage.fc.weight.data = sage.fc_neigh.weight.data
        dense_sage.fc.bias.data = sage.bias.data
        if len(g.ntypes) == 2:
            feat = (
                F.randn((g.number_of_src_nodes(), 5)),
                F.randn((g.number_of_dst_nodes(), 5))
            )
        else:
            feat = F.randn((g.number_of_nodes(), 5))
        sage = sage.to(ctx)
        dense_sage = dense_sage.to(ctx)
>       out_sage = sage(g, feat)

tests\pytorch\test_nn.py:956: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:249: in forward
    graph.update_all(msg_fn, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF073363C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.2176],
        [ 0.7364],
        [-0.8227],
        [-2.8005],
        [-1.6167],
        [ 0.3160],
        [ 0.0540],
        [-0.6274],
        [-0.3670],
        [-1.2996]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_____________________ test_dense_sage_conv[2-g0-idtype0] ______________________

g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int32, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_sage_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        sage = nn.SAGEConv(5, out_dim, 'gcn')
        dense_sage = nn.DenseSAGEConv(5, out_dim)
        dense_sage.fc.weight.data = sage.fc_neigh.weight.data
        dense_sage.fc.bias.data = sage.bias.data
        if len(g.ntypes) == 2:
            feat = (
                F.randn((g.number_of_src_nodes(), 5)),
                F.randn((g.number_of_dst_nodes(), 5))
            )
        else:
            feat = F.randn((g.number_of_nodes(), 5))
        sage = sage.to(ctx)
        dense_sage = dense_sage.to(ctx)
>       out_sage = sage(g, feat)

tests\pytorch\test_nn.py:956: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:249: in forward
    graph.update_all(msg_fn, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE121EA508>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.9490, -1.0154],
        [ 1.6560, -0.6989]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_____________________ test_dense_sage_conv[2-g0-idtype1] ______________________

g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int64, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_sage_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        sage = nn.SAGEConv(5, out_dim, 'gcn')
        dense_sage = nn.DenseSAGEConv(5, out_dim)
        dense_sage.fc.weight.data = sage.fc_neigh.weight.data
        dense_sage.fc.bias.data = sage.bias.data
        if len(g.ntypes) == 2:
            feat = (
                F.randn((g.number_of_src_nodes(), 5)),
                F.randn((g.number_of_dst_nodes(), 5))
            )
        else:
            feat = F.randn((g.number_of_nodes(), 5))
        sage = sage.to(ctx)
        dense_sage = dense_sage.to(ctx)
>       out_sage = sage(g, feat)

tests\pytorch\test_nn.py:956: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:249: in forward
    graph.update_all(msg_fn, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF073C2E08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.5184, -1.3156],
        [-0.6252,  2.6336]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_____________________ test_dense_sage_conv[2-g1-idtype0] ______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
idtype = torch.int32, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_sage_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        sage = nn.SAGEConv(5, out_dim, 'gcn')
        dense_sage = nn.DenseSAGEConv(5, out_dim)
        dense_sage.fc.weight.data = sage.fc_neigh.weight.data
        dense_sage.fc.bias.data = sage.bias.data
        if len(g.ntypes) == 2:
            feat = (
                F.randn((g.number_of_src_nodes(), 5)),
                F.randn((g.number_of_dst_nodes(), 5))
            )
        else:
            feat = F.randn((g.number_of_nodes(), 5))
        sage = sage.to(ctx)
        dense_sage = dense_sage.to(ctx)
>       out_sage = sage(g, feat)

tests\pytorch\test_nn.py:956: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:249: in forward
    graph.update_all(msg_fn, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF071BEE08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 2.7403,  2.1781],
        [ 0.9852, -0.1202],
        [-4.5816, -1.0097],
        [ 1.0206, -0.3369],
      ...2,  0.3126],
        [-3.5051, -5.3337],
        [ 4.4905,  1.1021],
        [ 1.8589, -1.3745]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_____________________ test_dense_sage_conv[2-g1-idtype1] ______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
idtype = torch.int64, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_sage_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        sage = nn.SAGEConv(5, out_dim, 'gcn')
        dense_sage = nn.DenseSAGEConv(5, out_dim)
        dense_sage.fc.weight.data = sage.fc_neigh.weight.data
        dense_sage.fc.bias.data = sage.bias.data
        if len(g.ntypes) == 2:
            feat = (
                F.randn((g.number_of_src_nodes(), 5)),
                F.randn((g.number_of_dst_nodes(), 5))
            )
        else:
            feat = F.randn((g.number_of_nodes(), 5))
        sage = sage.to(ctx)
        dense_sage = dense_sage.to(ctx)
>       out_sage = sage(g, feat)

tests\pytorch\test_nn.py:956: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:249: in forward
    graph.update_all(msg_fn, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0712C488>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.4170,  0.4901],
        [ 1.9065,  1.2294],
        [-0.8915, -0.7057],
        [-1.4940,  2.0305],
      ...9,  2.4506],
        [-3.2785, -0.2264],
        [ 2.1410,  0.7881],
        [-1.4397, -1.1673]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_____________________ test_dense_sage_conv[2-g2-idtype0] ______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_sage_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        sage = nn.SAGEConv(5, out_dim, 'gcn')
        dense_sage = nn.DenseSAGEConv(5, out_dim)
        dense_sage.fc.weight.data = sage.fc_neigh.weight.data
        dense_sage.fc.bias.data = sage.bias.data
        if len(g.ntypes) == 2:
            feat = (
                F.randn((g.number_of_src_nodes(), 5)),
                F.randn((g.number_of_dst_nodes(), 5))
            )
        else:
            feat = F.randn((g.number_of_nodes(), 5))
        sage = sage.to(ctx)
        dense_sage = dense_sage.to(ctx)
>       out_sage = sage(g, feat)

tests\pytorch\test_nn.py:956: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:249: in forward
    graph.update_all(msg_fn, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE12255208>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.7075,  1.5021],
        [ 0.3436, -1.5150],
        [ 0.8802, -0.0318],
        [ 0.3353, -1.5854],
      ...3,  0.2115],
        [-1.4992,  1.5468],
        [-0.1891,  0.5314],
        [ 0.7700, -1.1053]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_____________________ test_dense_sage_conv[2-g2-idtype1] ______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_sage_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        sage = nn.SAGEConv(5, out_dim, 'gcn')
        dense_sage = nn.DenseSAGEConv(5, out_dim)
        dense_sage.fc.weight.data = sage.fc_neigh.weight.data
        dense_sage.fc.bias.data = sage.bias.data
        if len(g.ntypes) == 2:
            feat = (
                F.randn((g.number_of_src_nodes(), 5)),
                F.randn((g.number_of_dst_nodes(), 5))
            )
        else:
            feat = F.randn((g.number_of_nodes(), 5))
        sage = sage.to(ctx)
        dense_sage = dense_sage.to(ctx)
>       out_sage = sage(g, feat)

tests\pytorch\test_nn.py:956: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:249: in forward
    graph.update_all(msg_fn, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF073AAAC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.0237, -3.9334],
        [ 2.0856,  0.9338],
        [-0.2846,  0.1558],
        [-0.9458,  2.4010],
      ...2,  1.7329],
        [-0.1940, -1.0803],
        [-0.8622, -0.0419],
        [ 2.7706,  0.0070]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_____________________ test_dense_sage_conv[2-g3-idtype0] ______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int32, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_sage_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        sage = nn.SAGEConv(5, out_dim, 'gcn')
        dense_sage = nn.DenseSAGEConv(5, out_dim)
        dense_sage.fc.weight.data = sage.fc_neigh.weight.data
        dense_sage.fc.bias.data = sage.bias.data
        if len(g.ntypes) == 2:
            feat = (
                F.randn((g.number_of_src_nodes(), 5)),
                F.randn((g.number_of_dst_nodes(), 5))
            )
        else:
            feat = F.randn((g.number_of_nodes(), 5))
        sage = sage.to(ctx)
        dense_sage = dense_sage.to(ctx)
>       out_sage = sage(g, feat)

tests\pytorch\test_nn.py:956: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:249: in forward
    graph.update_all(msg_fn, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF070C2C88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.0378, -2.9151],
        [ 0.6657,  1.2088],
        [ 1.1628,  1.5059],
        [-1.4201, -1.1553],
      ...6, -0.4502],
        [-0.8209,  0.7523],
        [-2.6323, -3.9257],
        [ 1.0499, -0.3561]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_____________________ test_dense_sage_conv[2-g3-idtype1] ______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int64, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_sage_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        sage = nn.SAGEConv(5, out_dim, 'gcn')
        dense_sage = nn.DenseSAGEConv(5, out_dim)
        dense_sage.fc.weight.data = sage.fc_neigh.weight.data
        dense_sage.fc.bias.data = sage.bias.data
        if len(g.ntypes) == 2:
            feat = (
                F.randn((g.number_of_src_nodes(), 5)),
                F.randn((g.number_of_dst_nodes(), 5))
            )
        else:
            feat = F.randn((g.number_of_nodes(), 5))
        sage = sage.to(ctx)
        dense_sage = dense_sage.to(ctx)
>       out_sage = sage(g, feat)

tests\pytorch\test_nn.py:956: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:249: in forward
    graph.update_all(msg_fn, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0198E308>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-2.2368, -1.0537],
        [ 1.7805, -1.4698],
        [-3.3394,  2.0873],
        [-1.8355,  2.3716],
      ...2,  0.3513],
        [-1.5665,  1.1912],
        [-2.5450,  0.5936],
        [ 0.2189, -0.2958]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_____________________ test_dense_sage_conv[2-g4-idtype0] ______________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_sage_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        sage = nn.SAGEConv(5, out_dim, 'gcn')
        dense_sage = nn.DenseSAGEConv(5, out_dim)
        dense_sage.fc.weight.data = sage.fc_neigh.weight.data
        dense_sage.fc.bias.data = sage.bias.data
        if len(g.ntypes) == 2:
            feat = (
                F.randn((g.number_of_src_nodes(), 5)),
                F.randn((g.number_of_dst_nodes(), 5))
            )
        else:
            feat = F.randn((g.number_of_nodes(), 5))
        sage = sage.to(ctx)
        dense_sage = dense_sage.to(ctx)
>       out_sage = sage(g, feat)

tests\pytorch\test_nn.py:956: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:249: in forward
    graph.update_all(msg_fn, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1225AF88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.5831, -0.0939],
        [ 0.5439, -0.5663],
        [ 1.2872,  1.1971],
        [ 0.8992, -0.0141],
      ...2,  0.8591],
        [-0.2988,  1.7972],
        [ 0.1536, -0.2179],
        [ 1.5247, -0.1977]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_____________________ test_dense_sage_conv[2-g4-idtype1] ______________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_sage_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        sage = nn.SAGEConv(5, out_dim, 'gcn')
        dense_sage = nn.DenseSAGEConv(5, out_dim)
        dense_sage.fc.weight.data = sage.fc_neigh.weight.data
        dense_sage.fc.bias.data = sage.bias.data
        if len(g.ntypes) == 2:
            feat = (
                F.randn((g.number_of_src_nodes(), 5)),
                F.randn((g.number_of_dst_nodes(), 5))
            )
        else:
            feat = F.randn((g.number_of_nodes(), 5))
        sage = sage.to(ctx)
        dense_sage = dense_sage.to(ctx)
>       out_sage = sage(g, feat)

tests\pytorch\test_nn.py:956: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:249: in forward
    graph.update_all(msg_fn, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE123DE608>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.2655,  1.4906],
        [ 2.9605, -1.6507],
        [ 0.2846, -0.0444],
        [ 0.3875,  0.5456],
      ...0, -1.8850],
        [ 0.6776, -1.0694],
        [ 1.2769, -0.4919],
        [-0.7781,  1.5584]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_____________________ test_dense_sage_conv[2-g5-idtype0] ______________________

g = Graph(num_nodes=5, num_edges=6,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_sage_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        sage = nn.SAGEConv(5, out_dim, 'gcn')
        dense_sage = nn.DenseSAGEConv(5, out_dim)
        dense_sage.fc.weight.data = sage.fc_neigh.weight.data
        dense_sage.fc.bias.data = sage.bias.data
        if len(g.ntypes) == 2:
            feat = (
                F.randn((g.number_of_src_nodes(), 5)),
                F.randn((g.number_of_dst_nodes(), 5))
            )
        else:
            feat = F.randn((g.number_of_nodes(), 5))
        sage = sage.to(ctx)
        dense_sage = dense_sage.to(ctx)
>       out_sage = sage(g, feat)

tests\pytorch\test_nn.py:956: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:249: in forward
    graph.update_all(msg_fn, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF017A93C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-2.3531, -0.0311],
        [ 1.4617,  0.0862],
        [-2.0810,  0.7310],
        [-0.0582,  0.1861],
        [ 0.3801,  1.2621]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_____________________ test_dense_sage_conv[2-g5-idtype1] ______________________

g = Graph(num_nodes=5, num_edges=6,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_sage_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        sage = nn.SAGEConv(5, out_dim, 'gcn')
        dense_sage = nn.DenseSAGEConv(5, out_dim)
        dense_sage.fc.weight.data = sage.fc_neigh.weight.data
        dense_sage.fc.bias.data = sage.bias.data
        if len(g.ntypes) == 2:
            feat = (
                F.randn((g.number_of_src_nodes(), 5)),
                F.randn((g.number_of_dst_nodes(), 5))
            )
        else:
            feat = F.randn((g.number_of_nodes(), 5))
        sage = sage.to(ctx)
        dense_sage = dense_sage.to(ctx)
>       out_sage = sage(g, feat)

tests\pytorch\test_nn.py:956: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:249: in forward
    graph.update_all(msg_fn, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF018A5888>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 2.4476,  2.4650],
        [-1.1160, -1.3316],
        [ 1.6811,  1.7069],
        [ 2.8628,  0.4882],
        [ 1.4515,  1.0017]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_____________________ test_dense_sage_conv[2-g6-idtype0] ______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_sage_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        sage = nn.SAGEConv(5, out_dim, 'gcn')
        dense_sage = nn.DenseSAGEConv(5, out_dim)
        dense_sage.fc.weight.data = sage.fc_neigh.weight.data
        dense_sage.fc.bias.data = sage.bias.data
        if len(g.ntypes) == 2:
            feat = (
                F.randn((g.number_of_src_nodes(), 5)),
                F.randn((g.number_of_dst_nodes(), 5))
            )
        else:
            feat = F.randn((g.number_of_nodes(), 5))
        sage = sage.to(ctx)
        dense_sage = dense_sage.to(ctx)
>       out_sage = sage(g, feat)

tests\pytorch\test_nn.py:956: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:249: in forward
    graph.update_all(msg_fn, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF073E7148>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-3.3322,  2.8596],
        [ 0.8547, -0.0254],
        [-0.5779, -0.3182],
        [-2.1621,  1.2550],
      ...3, -0.2722],
        [-1.3559,  1.5236],
        [-1.5852,  0.3707],
        [ 0.3736, -0.2405]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_____________________ test_dense_sage_conv[2-g6-idtype1] ______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_sage_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        sage = nn.SAGEConv(5, out_dim, 'gcn')
        dense_sage = nn.DenseSAGEConv(5, out_dim)
        dense_sage.fc.weight.data = sage.fc_neigh.weight.data
        dense_sage.fc.bias.data = sage.bias.data
        if len(g.ntypes) == 2:
            feat = (
                F.randn((g.number_of_src_nodes(), 5)),
                F.randn((g.number_of_dst_nodes(), 5))
            )
        else:
            feat = F.randn((g.number_of_nodes(), 5))
        sage = sage.to(ctx)
        dense_sage = dense_sage.to(ctx)
>       out_sage = sage(g, feat)

tests\pytorch\test_nn.py:956: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:249: in forward
    graph.update_all(msg_fn, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0195C8C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.2600,  1.2842],
        [ 0.6575,  1.4200],
        [-1.6534, -0.3604],
        [-1.7044,  0.8796],
      ...4, -0.7638],
        [-0.9467,  0.2027],
        [ 0.7426, -1.1356],
        [-0.2186, -0.7314]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_____________________ test_dense_sage_conv[2-g7-idtype0] ______________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int32
out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_sage_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        sage = nn.SAGEConv(5, out_dim, 'gcn')
        dense_sage = nn.DenseSAGEConv(5, out_dim)
        dense_sage.fc.weight.data = sage.fc_neigh.weight.data
        dense_sage.fc.bias.data = sage.bias.data
        if len(g.ntypes) == 2:
            feat = (
                F.randn((g.number_of_src_nodes(), 5)),
                F.randn((g.number_of_dst_nodes(), 5))
            )
        else:
            feat = F.randn((g.number_of_nodes(), 5))
        sage = sage.to(ctx)
        dense_sage = dense_sage.to(ctx)
>       out_sage = sage(g, feat)

tests\pytorch\test_nn.py:956: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:249: in forward
    graph.update_all(msg_fn, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF015559C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.0483,  1.0428],
        [-2.1856, -2.6527],
        [-0.4817,  0.2964],
        [-0.6922, -1.1735],
        [ 0.4996, -0.8630],
        [ 0.5500,  0.5962]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_____________________ test_dense_sage_conv[2-g7-idtype1] ______________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int64
out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_sage_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        sage = nn.SAGEConv(5, out_dim, 'gcn')
        dense_sage = nn.DenseSAGEConv(5, out_dim)
        dense_sage.fc.weight.data = sage.fc_neigh.weight.data
        dense_sage.fc.bias.data = sage.bias.data
        if len(g.ntypes) == 2:
            feat = (
                F.randn((g.number_of_src_nodes(), 5)),
                F.randn((g.number_of_dst_nodes(), 5))
            )
        else:
            feat = F.randn((g.number_of_nodes(), 5))
        sage = sage.to(ctx)
        dense_sage = dense_sage.to(ctx)
>       out_sage = sage(g, feat)

tests\pytorch\test_nn.py:956: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:249: in forward
    graph.update_all(msg_fn, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03E59648>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.8900, -0.1851],
        [-0.4296, -0.3760],
        [ 0.3017, -0.4540],
        [ 0.5240, -0.4295],
        [-0.0224, -0.3005],
        [-0.4279,  0.3025]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_____________________ test_dense_sage_conv[2-g8-idtype0] ______________________

g = Graph(num_nodes={'_U': 4, '_V': 5},
      num_edges={('_U', '_E', '_V'): 6},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int32, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_sage_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        sage = nn.SAGEConv(5, out_dim, 'gcn')
        dense_sage = nn.DenseSAGEConv(5, out_dim)
        dense_sage.fc.weight.data = sage.fc_neigh.weight.data
        dense_sage.fc.bias.data = sage.bias.data
        if len(g.ntypes) == 2:
            feat = (
                F.randn((g.number_of_src_nodes(), 5)),
                F.randn((g.number_of_dst_nodes(), 5))
            )
        else:
            feat = F.randn((g.number_of_nodes(), 5))
        sage = sage.to(ctx)
        dense_sage = dense_sage.to(ctx)
>       out_sage = sage(g, feat)

tests\pytorch\test_nn.py:956: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:249: in forward
    graph.update_all(msg_fn, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07295888>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.8889, -0.2677],
        [ 0.0987, -0.0481],
        [-0.2068,  0.5036],
        [ 0.2338, -0.4842]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_____________________ test_dense_sage_conv[2-g8-idtype1] ______________________

g = Graph(num_nodes={'_U': 4, '_V': 5},
      num_edges={('_U', '_E', '_V'): 6},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int64, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_sage_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        sage = nn.SAGEConv(5, out_dim, 'gcn')
        dense_sage = nn.DenseSAGEConv(5, out_dim)
        dense_sage.fc.weight.data = sage.fc_neigh.weight.data
        dense_sage.fc.bias.data = sage.bias.data
        if len(g.ntypes) == 2:
            feat = (
                F.randn((g.number_of_src_nodes(), 5)),
                F.randn((g.number_of_dst_nodes(), 5))
            )
        else:
            feat = F.randn((g.number_of_nodes(), 5))
        sage = sage.to(ctx)
        dense_sage = dense_sage.to(ctx)
>       out_sage = sage(g, feat)

tests\pytorch\test_nn.py:956: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:249: in forward
    graph.update_all(msg_fn, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1239F208>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.4788,  0.2700],
        [-0.3567, -0.6105],
        [-0.8950, -0.3237],
        [ 1.5183,  1.0003]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_____________________ test_dense_sage_conv[2-g9-idtype0] ______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_sage_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        sage = nn.SAGEConv(5, out_dim, 'gcn')
        dense_sage = nn.DenseSAGEConv(5, out_dim)
        dense_sage.fc.weight.data = sage.fc_neigh.weight.data
        dense_sage.fc.bias.data = sage.bias.data
        if len(g.ntypes) == 2:
            feat = (
                F.randn((g.number_of_src_nodes(), 5)),
                F.randn((g.number_of_dst_nodes(), 5))
            )
        else:
            feat = F.randn((g.number_of_nodes(), 5))
        sage = sage.to(ctx)
        dense_sage = dense_sage.to(ctx)
>       out_sage = sage(g, feat)

tests\pytorch\test_nn.py:956: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:249: in forward
    graph.update_all(msg_fn, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0181DA08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-2.8393, -0.6429],
        [ 2.8410,  1.2053],
        [ 1.8389,  1.3328],
        [ 1.8312, -0.6356],
      ...0,  0.3235],
        [ 0.7834,  0.0416],
        [-0.9999,  0.0309],
        [ 0.5188, -0.0297]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_____________________ test_dense_sage_conv[2-g9-idtype1] ______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_sage_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
        sage = nn.SAGEConv(5, out_dim, 'gcn')
        dense_sage = nn.DenseSAGEConv(5, out_dim)
        dense_sage.fc.weight.data = sage.fc_neigh.weight.data
        dense_sage.fc.bias.data = sage.bias.data
        if len(g.ntypes) == 2:
            feat = (
                F.randn((g.number_of_src_nodes(), 5)),
                F.randn((g.number_of_dst_nodes(), 5))
            )
        else:
            feat = F.randn((g.number_of_nodes(), 5))
        sage = sage.to(ctx)
        dense_sage = dense_sage.to(ctx)
>       out_sage = sage(g, feat)

tests\pytorch\test_nn.py:956: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\sageconv.py:249: in forward
    graph.update_all(msg_fn, fn.sum('m', 'neigh'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF073561C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.8633,  0.4435],
        [-2.5902,  0.2579],
        [ 1.0817,  0.9767],
        [-1.2130, -0.7126],
      ...3, -0.2996],
        [ 0.3737,  0.1230],
        [ 4.3576, -0.8524],
        [ 0.5457,  1.3922]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________________ test_edge_conv[1-g0-idtype0] _________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'x': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
idtype = torch.int32, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_edge_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        edge_conv = nn.EdgeConv(5, out_dim).to(ctx)
        print(edge_conv)
    
        # test pickle
        th.save(edge_conv, tmp_buffer)
    
        h0 = F.randn((g.number_of_src_nodes(), 5))
>       h1 = edge_conv(g, h0)

tests\pytorch\test_nn.py:973: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\edgeconv.py:171: in forward
    g.apply_edges(fn.v_sub_u('x', 'x', 'theta'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03E3AB48>
op = 'add'
lhs_data = tensor([[ 0.0486,  0.3352,  0.4841,  0.1585, -1.2404],
        [-0.2769, -0.2891,  0.2224, -0.4155, -1.2190],
        ...2.1852],
        [-0.0289,  1.6883, -0.8614,  0.4308, -0.3867],
        [-0.1819, -0.4257,  0.6875,  1.6385, -0.6501]])
rhs_data = tensor([[-0.0486, -0.3352, -0.4841, -0.1585,  1.2404],
        [ 0.2769,  0.2891, -0.2224,  0.4155,  1.2190],
        ...2.1852],
        [ 0.0289, -1.6883,  0.8614, -0.4308,  0.3867],
        [ 0.1819,  0.4257, -0.6875, -1.6385,  0.6501]])
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
---------------------------- Captured stdout call -----------------------------
EdgeConv(
  (theta): Linear(in_features=5, out_features=1, bias=True)
  (phi): Linear(in_features=5, out_features=1, bias=True)
)
________________________ test_edge_conv[1-g0-idtype1] _________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'x': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
idtype = torch.int64, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_edge_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        edge_conv = nn.EdgeConv(5, out_dim).to(ctx)
        print(edge_conv)
    
        # test pickle
        th.save(edge_conv, tmp_buffer)
    
        h0 = F.randn((g.number_of_src_nodes(), 5))
>       h1 = edge_conv(g, h0)

tests\pytorch\test_nn.py:973: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\edgeconv.py:171: in forward
    g.apply_edges(fn.v_sub_u('x', 'x', 'theta'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07456748>
op = 'add'
lhs_data = tensor([[-0.5136, -1.4639,  1.5337,  1.4827, -0.6921],
        [-0.7769, -0.3269,  0.3802,  0.8433,  0.2765],
        ...1.5659],
        [ 0.3988,  0.5652,  1.1266, -0.1006,  0.0758],
        [ 1.7179,  0.0114,  1.8013,  1.3129,  1.3405]])
rhs_data = tensor([[ 0.5136,  1.4639, -1.5337, -1.4827,  0.6921],
        [ 0.7769,  0.3269, -0.3802, -0.8433, -0.2765],
        ...1.5659],
        [-0.3988, -0.5652, -1.1266,  0.1006, -0.0758],
        [-1.7179, -0.0114, -1.8013, -1.3129, -1.3405]])
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
---------------------------- Captured stdout call -----------------------------
EdgeConv(
  (theta): Linear(in_features=5, out_features=1, bias=True)
  (phi): Linear(in_features=5, out_features=1, bias=True)
)
________________________ test_edge_conv[1-g1-idtype0] _________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'x': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_edge_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        edge_conv = nn.EdgeConv(5, out_dim).to(ctx)
        print(edge_conv)
    
        # test pickle
        th.save(edge_conv, tmp_buffer)
    
        h0 = F.randn((g.number_of_src_nodes(), 5))
>       h1 = edge_conv(g, h0)

tests\pytorch\test_nn.py:973: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\edgeconv.py:171: in forward
    g.apply_edges(fn.v_sub_u('x', 'x', 'theta'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1237A148>
op = 'add'
lhs_data = tensor([[ 1.2480, -1.8522, -0.1975, -0.6914,  1.2823],
        [-1.0842,  0.3590, -1.1475,  0.0960, -0.9253],
        ...1.6976],
        [ 1.0603,  0.2190, -0.4218, -1.0665,  0.2321],
        [ 0.2228,  0.3022, -0.6097, -0.4752, -1.2877]])
rhs_data = tensor([[-1.2480,  1.8522,  0.1975,  0.6914, -1.2823],
        [ 1.0842, -0.3590,  1.1475, -0.0960,  0.9253],
        ...1.6976],
        [-1.0603, -0.2190,  0.4218,  1.0665, -0.2321],
        [-0.2228, -0.3022,  0.6097,  0.4752,  1.2877]])
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
---------------------------- Captured stdout call -----------------------------
EdgeConv(
  (theta): Linear(in_features=5, out_features=1, bias=True)
  (phi): Linear(in_features=5, out_features=1, bias=True)
)
________________________ test_edge_conv[1-g1-idtype1] _________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'x': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_edge_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        edge_conv = nn.EdgeConv(5, out_dim).to(ctx)
        print(edge_conv)
    
        # test pickle
        th.save(edge_conv, tmp_buffer)
    
        h0 = F.randn((g.number_of_src_nodes(), 5))
>       h1 = edge_conv(g, h0)

tests\pytorch\test_nn.py:973: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\edgeconv.py:171: in forward
    g.apply_edges(fn.v_sub_u('x', 'x', 'theta'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07411FC8>
op = 'add'
lhs_data = tensor([[ 0.7542,  1.2048, -1.3002,  0.5082,  0.0886],
        [-0.9296,  0.4856, -0.5248,  0.0413, -0.2500],
        ...0.0868],
        [ 0.9264,  0.9928, -0.7169, -1.0721, -1.5658],
        [ 0.5802,  0.7989, -0.0451,  0.0844,  0.3129]])
rhs_data = tensor([[-0.7542, -1.2048,  1.3002, -0.5082, -0.0886],
        [ 0.9296, -0.4856,  0.5248, -0.0413,  0.2500],
        ...0.0868],
        [-0.9264, -0.9928,  0.7169,  1.0721,  1.5658],
        [-0.5802, -0.7989,  0.0451, -0.0844, -0.3129]])
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
---------------------------- Captured stdout call -----------------------------
EdgeConv(
  (theta): Linear(in_features=5, out_features=1, bias=True)
  (phi): Linear(in_features=5, out_features=1, bias=True)
)
________________________ test_edge_conv[1-g2-idtype0] _________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'x': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int32, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_edge_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        edge_conv = nn.EdgeConv(5, out_dim).to(ctx)
        print(edge_conv)
    
        # test pickle
        th.save(edge_conv, tmp_buffer)
    
        h0 = F.randn((g.number_of_src_nodes(), 5))
>       h1 = edge_conv(g, h0)

tests\pytorch\test_nn.py:973: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\edgeconv.py:171: in forward
    g.apply_edges(fn.v_sub_u('x', 'x', 'theta'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE123D1808>
op = 'add'
lhs_data = tensor([[-0.9559,  1.6076,  1.2898,  1.0605, -0.2799],
        [-0.7241,  0.0056, -1.3432, -0.3699,  0.5676],
        ...1.8477],
        [ 1.6227,  0.1986,  0.9330, -0.5169,  0.7565],
        [ 0.1157,  0.5162, -0.4353, -1.0177, -1.0691]])
rhs_data = tensor([[ 0.9559, -1.6076, -1.2898, -1.0605,  0.2799],
        [ 0.7241, -0.0056,  1.3432,  0.3699, -0.5676],
        ...1.8477],
        [-1.6227, -0.1986, -0.9330,  0.5169, -0.7565],
        [-0.1157, -0.5162,  0.4353,  1.0177,  1.0691]])
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
---------------------------- Captured stdout call -----------------------------
EdgeConv(
  (theta): Linear(in_features=5, out_features=1, bias=True)
  (phi): Linear(in_features=5, out_features=1, bias=True)
)
________________________ test_edge_conv[1-g2-idtype1] _________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'x': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int64, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_edge_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        edge_conv = nn.EdgeConv(5, out_dim).to(ctx)
        print(edge_conv)
    
        # test pickle
        th.save(edge_conv, tmp_buffer)
    
        h0 = F.randn((g.number_of_src_nodes(), 5))
>       h1 = edge_conv(g, h0)

tests\pytorch\test_nn.py:973: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\edgeconv.py:171: in forward
    g.apply_edges(fn.v_sub_u('x', 'x', 'theta'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03DD0248>
op = 'add'
lhs_data = tensor([[-0.1355, -0.8171, -1.8280,  2.6747,  0.1938],
        [ 0.2894, -0.5278, -2.1278, -0.6207,  0.2268],
        ...0.1360],
        [ 0.4286,  1.5195, -0.8199,  0.5392, -1.9381],
        [ 1.0424,  1.7486, -0.4171, -0.1380, -1.6583]])
rhs_data = tensor([[ 0.1355,  0.8171,  1.8280, -2.6747, -0.1938],
        [-0.2894,  0.5278,  2.1278,  0.6207, -0.2268],
        ...0.1360],
        [-0.4286, -1.5195,  0.8199, -0.5392,  1.9381],
        [-1.0424, -1.7486,  0.4171,  0.1380,  1.6583]])
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
---------------------------- Captured stdout call -----------------------------
EdgeConv(
  (theta): Linear(in_features=5, out_features=1, bias=True)
  (phi): Linear(in_features=5, out_features=1, bias=True)
)
________________________ test_edge_conv[1-g3-idtype0] _________________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'x': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_edge_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        edge_conv = nn.EdgeConv(5, out_dim).to(ctx)
        print(edge_conv)
    
        # test pickle
        th.save(edge_conv, tmp_buffer)
    
        h0 = F.randn((g.number_of_src_nodes(), 5))
>       h1 = edge_conv(g, h0)

tests\pytorch\test_nn.py:973: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\edgeconv.py:171: in forward
    g.apply_edges(fn.v_sub_u('x', 'x', 'theta'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03E3ABC8>
op = 'add'
lhs_data = tensor([[-7.0816e-01,  4.1643e-02, -1.0095e-01,  9.8841e-01,  7.6864e-01],
        [ 1.8465e-01,  6.2129e-01,  3.3920e...00, -7.4651e-01, -2.8184e-02,  6.6963e-01],
        [ 5.6946e-01,  1.5591e+00, -1.2900e+00, -5.4384e-01, -7.9795e-01]])
rhs_data = tensor([[ 7.0816e-01, -4.1643e-02,  1.0095e-01, -9.8841e-01, -7.6864e-01],
        [-1.8465e-01, -6.2129e-01, -3.3920e...00,  7.4651e-01,  2.8184e-02, -6.6963e-01],
        [-5.6946e-01, -1.5591e+00,  1.2900e+00,  5.4384e-01,  7.9795e-01]])
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
---------------------------- Captured stdout call -----------------------------
EdgeConv(
  (theta): Linear(in_features=5, out_features=1, bias=True)
  (phi): Linear(in_features=5, out_features=1, bias=True)
)
________________________ test_edge_conv[1-g3-idtype1] _________________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'x': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_edge_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        edge_conv = nn.EdgeConv(5, out_dim).to(ctx)
        print(edge_conv)
    
        # test pickle
        th.save(edge_conv, tmp_buffer)
    
        h0 = F.randn((g.number_of_src_nodes(), 5))
>       h1 = edge_conv(g, h0)

tests\pytorch\test_nn.py:973: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\edgeconv.py:171: in forward
    g.apply_edges(fn.v_sub_u('x', 'x', 'theta'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03E26B88>
op = 'add'
lhs_data = tensor([[-0.5552,  1.1992,  0.9214, -1.3064, -0.3379],
        [ 0.8146, -0.6228, -1.6886, -0.4540, -1.0247],
        ...1.0450],
        [-0.8386,  0.8527,  1.8547,  0.5756, -0.4845],
        [-2.2683, -0.7701,  2.9316, -3.1989, -1.3097]])
rhs_data = tensor([[ 0.5552, -1.1992, -0.9214,  1.3064,  0.3379],
        [-0.8146,  0.6228,  1.6886,  0.4540,  1.0247],
        ...1.0450],
        [ 0.8386, -0.8527, -1.8547, -0.5756,  0.4845],
        [ 2.2683,  0.7701, -2.9316,  3.1989,  1.3097]])
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
---------------------------- Captured stdout call -----------------------------
EdgeConv(
  (theta): Linear(in_features=5, out_features=1, bias=True)
  (phi): Linear(in_features=5, out_features=1, bias=True)
)
________________________ test_edge_conv[1-g4-idtype0] _________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'x': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_edge_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        edge_conv = nn.EdgeConv(5, out_dim).to(ctx)
        print(edge_conv)
    
        # test pickle
        th.save(edge_conv, tmp_buffer)
    
        h0 = F.randn((g.number_of_src_nodes(), 5))
>       h1 = edge_conv(g, h0)

tests\pytorch\test_nn.py:973: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\edgeconv.py:171: in forward
    g.apply_edges(fn.v_sub_u('x', 'x', 'theta'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE122DD108>
op = 'add'
lhs_data = tensor([[-0.3758,  0.6346,  0.3090,  0.6035, -0.8740],
        [ 0.4706,  0.6435, -1.1937,  0.1546, -0.1859],
        ...0.7161],
        [ 0.0214, -1.0867,  0.4669, -1.1128, -0.8790],
        [-1.3629, -1.0645,  0.6803,  0.7043, -1.8198]])
rhs_data = tensor([[ 0.3758, -0.6346, -0.3090, -0.6035,  0.8740],
        [-0.4706, -0.6435,  1.1937, -0.1546,  0.1859],
        ...0.7161],
        [-0.0214,  1.0867, -0.4669,  1.1128,  0.8790],
        [ 1.3629,  1.0645, -0.6803, -0.7043,  1.8198]])
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
---------------------------- Captured stdout call -----------------------------
EdgeConv(
  (theta): Linear(in_features=5, out_features=1, bias=True)
  (phi): Linear(in_features=5, out_features=1, bias=True)
)
________________________ test_edge_conv[1-g4-idtype1] _________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'x': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_edge_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        edge_conv = nn.EdgeConv(5, out_dim).to(ctx)
        print(edge_conv)
    
        # test pickle
        th.save(edge_conv, tmp_buffer)
    
        h0 = F.randn((g.number_of_src_nodes(), 5))
>       h1 = edge_conv(g, h0)

tests\pytorch\test_nn.py:973: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\edgeconv.py:171: in forward
    g.apply_edges(fn.v_sub_u('x', 'x', 'theta'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07036A08>
op = 'add'
lhs_data = tensor([[ 1.3937,  0.6486,  0.4478, -1.1440,  0.4023],
        [-1.0325, -0.1026, -0.1505, -0.6028,  0.1455],
        ...0.4223],
        [ 0.6643, -1.6333, -0.2763, -0.9425,  1.3614],
        [-1.3606, -0.9159,  0.6320,  2.0441, -1.0964]])
rhs_data = tensor([[-1.3937, -0.6486, -0.4478,  1.1440, -0.4023],
        [ 1.0325,  0.1026,  0.1505,  0.6028, -0.1455],
        ...0.4223],
        [-0.6643,  1.6333,  0.2763,  0.9425, -1.3614],
        [ 1.3606,  0.9159, -0.6320, -2.0441,  1.0964]])
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
---------------------------- Captured stdout call -----------------------------
EdgeConv(
  (theta): Linear(in_features=5, out_features=1, bias=True)
  (phi): Linear(in_features=5, out_features=1, bias=True)
)
________________________ test_edge_conv[1-g5-idtype0] _________________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int32
out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_edge_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        edge_conv = nn.EdgeConv(5, out_dim).to(ctx)
        print(edge_conv)
    
        # test pickle
        th.save(edge_conv, tmp_buffer)
    
        h0 = F.randn((g.number_of_src_nodes(), 5))
>       h1 = edge_conv(g, h0)

tests\pytorch\test_nn.py:973: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\edgeconv.py:171: in forward
    g.apply_edges(fn.v_sub_u('x', 'x', 'theta'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF073C5E48>
op = 'add'
lhs_data = tensor([[-0.1372,  0.3909, -0.2051, -0.1314, -0.6487],
        [-1.0132, -0.4128, -0.8828,  1.2638,  0.2821],
        [ 0.1515,  1.5877,  0.8232,  0.9604,  0.5296]])
rhs_data = tensor([[ 0.1372, -0.3909,  0.2051,  0.1314,  0.6487],
        [ 1.0132,  0.4128,  0.8828, -1.2638, -0.2821],
        ...0.4779],
        [-0.1748,  1.0415,  0.2310,  0.8693, -1.6413],
        [-0.1419, -0.6125,  1.5626,  0.8603, -0.7813]])
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
---------------------------- Captured stdout call -----------------------------
EdgeConv(
  (theta): Linear(in_features=5, out_features=1, bias=True)
  (phi): Linear(in_features=5, out_features=1, bias=True)
)
________________________ test_edge_conv[1-g5-idtype1] _________________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int64
out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_edge_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        edge_conv = nn.EdgeConv(5, out_dim).to(ctx)
        print(edge_conv)
    
        # test pickle
        th.save(edge_conv, tmp_buffer)
    
        h0 = F.randn((g.number_of_src_nodes(), 5))
>       h1 = edge_conv(g, h0)

tests\pytorch\test_nn.py:973: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\edgeconv.py:171: in forward
    g.apply_edges(fn.v_sub_u('x', 'x', 'theta'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF06FFFB48>
op = 'add'
lhs_data = tensor([[ 1.4553, -0.9854, -0.0216, -1.9448, -1.2369],
        [ 1.7016,  0.0099, -0.8279, -0.0410, -0.0900],
        [ 1.1371,  0.7308, -0.3376, -0.0487,  0.3502]])
rhs_data = tensor([[-1.4553,  0.9854,  0.0216,  1.9448,  1.2369],
        [-1.7016, -0.0099,  0.8279,  0.0410,  0.0900],
        ...0.1876],
        [ 0.1043,  0.2221,  1.3706,  1.6082,  1.4585],
        [ 1.5756, -1.2806, -1.6523, -1.9453,  0.4100]])
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
---------------------------- Captured stdout call -----------------------------
EdgeConv(
  (theta): Linear(in_features=5, out_features=1, bias=True)
  (phi): Linear(in_features=5, out_features=1, bias=True)
)
________________________ test_edge_conv[1-g6-idtype0] _________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'x': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_edge_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        edge_conv = nn.EdgeConv(5, out_dim).to(ctx)
        print(edge_conv)
    
        # test pickle
        th.save(edge_conv, tmp_buffer)
    
        h0 = F.randn((g.number_of_src_nodes(), 5))
>       h1 = edge_conv(g, h0)

tests\pytorch\test_nn.py:973: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\edgeconv.py:171: in forward
    g.apply_edges(fn.v_sub_u('x', 'x', 'theta'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0759EC88>
op = 'add'
lhs_data = tensor([[ 1.2712, -0.6177, -1.8023,  0.1760,  0.1517],
        [-2.1062, -0.1434,  0.1742, -0.4430, -0.1527],
        ...0.5282],
        [-1.8916, -1.2768,  0.6588,  0.9638, -0.6020],
        [-0.1970, -0.4955,  1.9567, -1.8647,  0.7032]])
rhs_data = tensor([[-1.2712,  0.6177,  1.8023, -0.1760, -0.1517],
        [ 2.1062,  0.1434, -0.1742,  0.4430,  0.1527],
        ...0.5282],
        [ 1.8916,  1.2768, -0.6588, -0.9638,  0.6020],
        [ 0.1970,  0.4955, -1.9567,  1.8647, -0.7032]])
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
---------------------------- Captured stdout call -----------------------------
EdgeConv(
  (theta): Linear(in_features=5, out_features=1, bias=True)
  (phi): Linear(in_features=5, out_features=1, bias=True)
)
________________________ test_edge_conv[1-g6-idtype1] _________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'x': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_edge_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        edge_conv = nn.EdgeConv(5, out_dim).to(ctx)
        print(edge_conv)
    
        # test pickle
        th.save(edge_conv, tmp_buffer)
    
        h0 = F.randn((g.number_of_src_nodes(), 5))
>       h1 = edge_conv(g, h0)

tests\pytorch\test_nn.py:973: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\edgeconv.py:171: in forward
    g.apply_edges(fn.v_sub_u('x', 'x', 'theta'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1225B108>
op = 'add'
lhs_data = tensor([[ 0.5704, -0.1709,  1.6287, -0.3893, -1.8803],
        [-0.3791,  0.8389, -0.9468,  0.2351, -0.5759],
        ...0.0056],
        [ 0.8449,  0.4759, -0.3591, -1.3274, -0.4474],
        [-0.0172, -0.8281,  0.3922, -0.3871, -1.1245]])
rhs_data = tensor([[-0.5704,  0.1709, -1.6287,  0.3893,  1.8803],
        [ 0.3791, -0.8389,  0.9468, -0.2351,  0.5759],
        ...0.0056],
        [-0.8449, -0.4759,  0.3591,  1.3274,  0.4474],
        [ 0.0172,  0.8281, -0.3922,  0.3871,  1.1245]])
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
---------------------------- Captured stdout call -----------------------------
EdgeConv(
  (theta): Linear(in_features=5, out_features=1, bias=True)
  (phi): Linear(in_features=5, out_features=1, bias=True)
)
________________________ test_edge_conv[2-g0-idtype0] _________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'x': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
idtype = torch.int32, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_edge_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        edge_conv = nn.EdgeConv(5, out_dim).to(ctx)
        print(edge_conv)
    
        # test pickle
        th.save(edge_conv, tmp_buffer)
    
        h0 = F.randn((g.number_of_src_nodes(), 5))
>       h1 = edge_conv(g, h0)

tests\pytorch\test_nn.py:973: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\edgeconv.py:171: in forward
    g.apply_edges(fn.v_sub_u('x', 'x', 'theta'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03DB97C8>
op = 'add'
lhs_data = tensor([[-1.0467,  0.6822, -2.0434, -0.6843,  1.1932],
        [ 0.3530,  0.3166,  2.8570,  0.5172,  0.3971],
        ...0.3209],
        [ 0.1236, -0.0431, -1.2726,  0.9989, -0.0455],
        [ 0.9604,  1.7321, -2.0296, -0.6182, -0.6574]])
rhs_data = tensor([[ 1.0467, -0.6822,  2.0434,  0.6843, -1.1932],
        [-0.3530, -0.3166, -2.8570, -0.5172, -0.3971],
        ...0.3209],
        [-0.1236,  0.0431,  1.2726, -0.9989,  0.0455],
        [-0.9604, -1.7321,  2.0296,  0.6182,  0.6574]])
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
---------------------------- Captured stdout call -----------------------------
EdgeConv(
  (theta): Linear(in_features=5, out_features=2, bias=True)
  (phi): Linear(in_features=5, out_features=2, bias=True)
)
________________________ test_edge_conv[2-g0-idtype1] _________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'x': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
idtype = torch.int64, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_edge_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        edge_conv = nn.EdgeConv(5, out_dim).to(ctx)
        print(edge_conv)
    
        # test pickle
        th.save(edge_conv, tmp_buffer)
    
        h0 = F.randn((g.number_of_src_nodes(), 5))
>       h1 = edge_conv(g, h0)

tests\pytorch\test_nn.py:973: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\edgeconv.py:171: in forward
    g.apply_edges(fn.v_sub_u('x', 'x', 'theta'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03E1CE88>
op = 'add'
lhs_data = tensor([[ 4.6883e-02, -1.3889e+00,  7.9162e-01, -1.6002e+00, -5.4146e-01],
        [ 9.1432e-02,  8.7884e-01,  1.0614e...00,  4.5172e-01, -6.7632e-01, -5.2416e-02],
        [-3.3287e-01, -4.9955e-01,  6.9548e-01, -1.7118e+00,  4.6134e-01]])
rhs_data = tensor([[-4.6883e-02,  1.3889e+00, -7.9162e-01,  1.6002e+00,  5.4146e-01],
        [-9.1432e-02, -8.7884e-01, -1.0614e...00, -4.5172e-01,  6.7632e-01,  5.2416e-02],
        [ 3.3287e-01,  4.9955e-01, -6.9548e-01,  1.7118e+00, -4.6134e-01]])
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
---------------------------- Captured stdout call -----------------------------
EdgeConv(
  (theta): Linear(in_features=5, out_features=2, bias=True)
  (phi): Linear(in_features=5, out_features=2, bias=True)
)
________________________ test_edge_conv[2-g1-idtype0] _________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'x': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_edge_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        edge_conv = nn.EdgeConv(5, out_dim).to(ctx)
        print(edge_conv)
    
        # test pickle
        th.save(edge_conv, tmp_buffer)
    
        h0 = F.randn((g.number_of_src_nodes(), 5))
>       h1 = edge_conv(g, h0)

tests\pytorch\test_nn.py:973: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\edgeconv.py:171: in forward
    g.apply_edges(fn.v_sub_u('x', 'x', 'theta'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE123B8288>
op = 'add'
lhs_data = tensor([[ 2.6122,  0.6351, -0.6148, -0.3339,  0.5510],
        [-1.1324, -0.8523,  0.7315, -0.7203,  1.4622],
        ...0.4787],
        [-0.0921, -1.3091,  0.8404,  1.8181, -0.7372],
        [ 0.3272,  1.2601,  0.0613,  0.7541,  1.1493]])
rhs_data = tensor([[-2.6122, -0.6351,  0.6148,  0.3339, -0.5510],
        [ 1.1324,  0.8523, -0.7315,  0.7203, -1.4622],
        ...0.4787],
        [ 0.0921,  1.3091, -0.8404, -1.8181,  0.7372],
        [-0.3272, -1.2601, -0.0613, -0.7541, -1.1493]])
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
---------------------------- Captured stdout call -----------------------------
EdgeConv(
  (theta): Linear(in_features=5, out_features=2, bias=True)
  (phi): Linear(in_features=5, out_features=2, bias=True)
)
________________________ test_edge_conv[2-g1-idtype1] _________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'x': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_edge_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        edge_conv = nn.EdgeConv(5, out_dim).to(ctx)
        print(edge_conv)
    
        # test pickle
        th.save(edge_conv, tmp_buffer)
    
        h0 = F.randn((g.number_of_src_nodes(), 5))
>       h1 = edge_conv(g, h0)

tests\pytorch\test_nn.py:973: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\edgeconv.py:171: in forward
    g.apply_edges(fn.v_sub_u('x', 'x', 'theta'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0727AE08>
op = 'add'
lhs_data = tensor([[-0.9536, -0.5266,  1.3723, -1.6593,  1.2216],
        [ 0.2664, -0.2312,  0.7005, -1.5385, -0.7313],
        ...1.1688],
        [ 0.4359,  1.1523, -0.8423,  0.2683, -0.5432],
        [-0.4415,  1.5636, -0.4892, -0.7642,  2.7993]])
rhs_data = tensor([[ 0.9536,  0.5266, -1.3723,  1.6593, -1.2216],
        [-0.2664,  0.2312, -0.7005,  1.5385,  0.7313],
        ...1.1688],
        [-0.4359, -1.1523,  0.8423, -0.2683,  0.5432],
        [ 0.4415, -1.5636,  0.4892,  0.7642, -2.7993]])
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
---------------------------- Captured stdout call -----------------------------
EdgeConv(
  (theta): Linear(in_features=5, out_features=2, bias=True)
  (phi): Linear(in_features=5, out_features=2, bias=True)
)
________________________ test_edge_conv[2-g2-idtype0] _________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'x': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int32, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_edge_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        edge_conv = nn.EdgeConv(5, out_dim).to(ctx)
        print(edge_conv)
    
        # test pickle
        th.save(edge_conv, tmp_buffer)
    
        h0 = F.randn((g.number_of_src_nodes(), 5))
>       h1 = edge_conv(g, h0)

tests\pytorch\test_nn.py:973: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\edgeconv.py:171: in forward
    g.apply_edges(fn.v_sub_u('x', 'x', 'theta'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1223BF48>
op = 'add'
lhs_data = tensor([[-0.6287, -1.1842, -0.4102,  0.5920,  1.0011],
        [-0.1276,  0.5879,  0.3422,  1.3506, -0.5213],
        ...0.8062],
        [-0.6234, -0.4602,  0.4185,  0.7299, -0.7100],
        [ 1.1459,  0.3243,  1.2374, -0.6434, -0.7038]])
rhs_data = tensor([[ 0.6287,  1.1842,  0.4102, -0.5920, -1.0011],
        [ 0.1276, -0.5879, -0.3422, -1.3506,  0.5213],
        ...0.8062],
        [ 0.6234,  0.4602, -0.4185, -0.7299,  0.7100],
        [-1.1459, -0.3243, -1.2374,  0.6434,  0.7038]])
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
---------------------------- Captured stdout call -----------------------------
EdgeConv(
  (theta): Linear(in_features=5, out_features=2, bias=True)
  (phi): Linear(in_features=5, out_features=2, bias=True)
)
________________________ test_edge_conv[2-g2-idtype1] _________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'x': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int64, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_edge_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        edge_conv = nn.EdgeConv(5, out_dim).to(ctx)
        print(edge_conv)
    
        # test pickle
        th.save(edge_conv, tmp_buffer)
    
        h0 = F.randn((g.number_of_src_nodes(), 5))
>       h1 = edge_conv(g, h0)

tests\pytorch\test_nn.py:973: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\edgeconv.py:171: in forward
    g.apply_edges(fn.v_sub_u('x', 'x', 'theta'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF070CA208>
op = 'add'
lhs_data = tensor([[ 1.7821,  0.8230, -0.6369, -1.4772,  0.9777],
        [ 0.5996,  1.3512,  0.2797, -1.1523,  1.2788],
        ...0.5372],
        [-0.1322,  1.6087, -0.7401,  0.5466,  0.9431],
        [ 0.0514, -2.7608, -0.1321, -1.4494,  0.1496]])
rhs_data = tensor([[-1.7821, -0.8230,  0.6369,  1.4772, -0.9777],
        [-0.5996, -1.3512, -0.2797,  1.1523, -1.2788],
        ...0.5372],
        [ 0.1322, -1.6087,  0.7401, -0.5466, -0.9431],
        [-0.0514,  2.7608,  0.1321,  1.4494, -0.1496]])
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
---------------------------- Captured stdout call -----------------------------
EdgeConv(
  (theta): Linear(in_features=5, out_features=2, bias=True)
  (phi): Linear(in_features=5, out_features=2, bias=True)
)
________________________ test_edge_conv[2-g3-idtype0] _________________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'x': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_edge_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        edge_conv = nn.EdgeConv(5, out_dim).to(ctx)
        print(edge_conv)
    
        # test pickle
        th.save(edge_conv, tmp_buffer)
    
        h0 = F.randn((g.number_of_src_nodes(), 5))
>       h1 = edge_conv(g, h0)

tests\pytorch\test_nn.py:973: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\edgeconv.py:171: in forward
    g.apply_edges(fn.v_sub_u('x', 'x', 'theta'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0731EA08>
op = 'add'
lhs_data = tensor([[ 0.6476, -0.1591,  0.2446, -1.8975,  1.0186],
        [ 0.2461, -0.0846, -0.5721, -0.8421, -0.7333],
        ...0.7965],
        [-0.6208,  0.1667, -0.2236,  0.1140,  1.6281],
        [ 0.3169, -0.4587, -0.2594,  0.8978, -1.5922]])
rhs_data = tensor([[-0.6476,  0.1591, -0.2446,  1.8975, -1.0186],
        [-0.2461,  0.0846,  0.5721,  0.8421,  0.7333],
        ...0.7965],
        [ 0.6208, -0.1667,  0.2236, -0.1140, -1.6281],
        [-0.3169,  0.4587,  0.2594, -0.8978,  1.5922]])
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
---------------------------- Captured stdout call -----------------------------
EdgeConv(
  (theta): Linear(in_features=5, out_features=2, bias=True)
  (phi): Linear(in_features=5, out_features=2, bias=True)
)
________________________ test_edge_conv[2-g3-idtype1] _________________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'x': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_edge_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        edge_conv = nn.EdgeConv(5, out_dim).to(ctx)
        print(edge_conv)
    
        # test pickle
        th.save(edge_conv, tmp_buffer)
    
        h0 = F.randn((g.number_of_src_nodes(), 5))
>       h1 = edge_conv(g, h0)

tests\pytorch\test_nn.py:973: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\edgeconv.py:171: in forward
    g.apply_edges(fn.v_sub_u('x', 'x', 'theta'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03D8A548>
op = 'add'
lhs_data = tensor([[-0.4314,  0.1708, -0.0978,  0.5323,  0.3766],
        [-0.6014,  0.2944,  0.2122,  2.1762, -1.1645],
        ...0.6536],
        [-0.3113,  1.5779, -1.2486,  0.2441,  0.6326],
        [ 0.4450,  0.7251,  0.6062, -2.2069,  0.2868]])
rhs_data = tensor([[ 0.4314, -0.1708,  0.0978, -0.5323, -0.3766],
        [ 0.6014, -0.2944, -0.2122, -2.1762,  1.1645],
        ...0.6536],
        [ 0.3113, -1.5779,  1.2486, -0.2441, -0.6326],
        [-0.4450, -0.7251, -0.6062,  2.2069, -0.2868]])
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
---------------------------- Captured stdout call -----------------------------
EdgeConv(
  (theta): Linear(in_features=5, out_features=2, bias=True)
  (phi): Linear(in_features=5, out_features=2, bias=True)
)
________________________ test_edge_conv[2-g4-idtype0] _________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'x': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_edge_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        edge_conv = nn.EdgeConv(5, out_dim).to(ctx)
        print(edge_conv)
    
        # test pickle
        th.save(edge_conv, tmp_buffer)
    
        h0 = F.randn((g.number_of_src_nodes(), 5))
>       h1 = edge_conv(g, h0)

tests\pytorch\test_nn.py:973: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\edgeconv.py:171: in forward
    g.apply_edges(fn.v_sub_u('x', 'x', 'theta'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF073E9C88>
op = 'add'
lhs_data = tensor([[ 0.2836,  0.7614, -1.5352, -1.2427, -2.0061],
        [ 0.8611, -2.7123,  0.1213, -0.8119,  0.0168],
        ...0.0256],
        [ 0.0291, -0.3433,  1.2727, -0.4527, -0.3901],
        [ 0.3806,  1.8876, -0.2478, -0.9003,  0.4304]])
rhs_data = tensor([[-0.2836, -0.7614,  1.5352,  1.2427,  2.0061],
        [-0.8611,  2.7123, -0.1213,  0.8119, -0.0168],
        ...0.0256],
        [-0.0291,  0.3433, -1.2727,  0.4527,  0.3901],
        [-0.3806, -1.8876,  0.2478,  0.9003, -0.4304]])
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
---------------------------- Captured stdout call -----------------------------
EdgeConv(
  (theta): Linear(in_features=5, out_features=2, bias=True)
  (phi): Linear(in_features=5, out_features=2, bias=True)
)
________________________ test_edge_conv[2-g4-idtype1] _________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'x': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_edge_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        edge_conv = nn.EdgeConv(5, out_dim).to(ctx)
        print(edge_conv)
    
        # test pickle
        th.save(edge_conv, tmp_buffer)
    
        h0 = F.randn((g.number_of_src_nodes(), 5))
>       h1 = edge_conv(g, h0)

tests\pytorch\test_nn.py:973: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\edgeconv.py:171: in forward
    g.apply_edges(fn.v_sub_u('x', 'x', 'theta'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE123A3688>
op = 'add'
lhs_data = tensor([[-1.9751e-01,  2.1486e+00, -3.5153e-01, -1.0911e+00, -8.3551e-01],
        [ 1.1672e+00, -7.0716e-01,  1.3496e...01, -1.0408e+00,  6.1995e-01,  4.8403e-01],
        [ 1.0102e+00, -9.4458e-01,  3.0140e-01,  1.3150e-03,  3.1131e-01]])
rhs_data = tensor([[ 1.9751e-01, -2.1486e+00,  3.5153e-01,  1.0911e+00,  8.3551e-01],
        [-1.1672e+00,  7.0716e-01, -1.3496e...01,  1.0408e+00, -6.1995e-01, -4.8403e-01],
        [-1.0102e+00,  9.4458e-01, -3.0140e-01, -1.3150e-03, -3.1131e-01]])
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
---------------------------- Captured stdout call -----------------------------
EdgeConv(
  (theta): Linear(in_features=5, out_features=2, bias=True)
  (phi): Linear(in_features=5, out_features=2, bias=True)
)
________________________ test_edge_conv[2-g5-idtype0] _________________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int32
out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_edge_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        edge_conv = nn.EdgeConv(5, out_dim).to(ctx)
        print(edge_conv)
    
        # test pickle
        th.save(edge_conv, tmp_buffer)
    
        h0 = F.randn((g.number_of_src_nodes(), 5))
>       h1 = edge_conv(g, h0)

tests\pytorch\test_nn.py:973: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\edgeconv.py:171: in forward
    g.apply_edges(fn.v_sub_u('x', 'x', 'theta'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0714B3C8>
op = 'add'
lhs_data = tensor([[ 0.3245,  0.9317,  0.9036, -0.6901, -1.0062],
        [ 0.4788,  0.7190, -0.9677,  0.8543,  0.3720],
        [ 0.2968,  0.2801, -0.8357,  1.2710, -1.7688]])
rhs_data = tensor([[-0.3245, -0.9317, -0.9036,  0.6901,  1.0062],
        [-0.4788, -0.7190,  0.9677, -0.8543, -0.3720],
        ...1.2238],
        [-1.6393,  1.1642,  1.5238, -0.2397,  0.2302],
        [-0.8122, -0.1605,  2.1358,  0.5376, -1.1153]])
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
---------------------------- Captured stdout call -----------------------------
EdgeConv(
  (theta): Linear(in_features=5, out_features=2, bias=True)
  (phi): Linear(in_features=5, out_features=2, bias=True)
)
________________________ test_edge_conv[2-g5-idtype1] _________________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int64
out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_edge_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        edge_conv = nn.EdgeConv(5, out_dim).to(ctx)
        print(edge_conv)
    
        # test pickle
        th.save(edge_conv, tmp_buffer)
    
        h0 = F.randn((g.number_of_src_nodes(), 5))
>       h1 = edge_conv(g, h0)

tests\pytorch\test_nn.py:973: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\edgeconv.py:171: in forward
    g.apply_edges(fn.v_sub_u('x', 'x', 'theta'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0192B388>
op = 'add'
lhs_data = tensor([[ 0.4703,  1.3877,  0.3466, -1.1047, -0.0428],
        [-2.0723,  1.2491, -0.8536,  2.2255, -1.0305],
        [ 0.7438,  0.8698,  1.1144,  2.3243, -1.1786]])
rhs_data = tensor([[-0.4703, -1.3877, -0.3466,  1.1047,  0.0428],
        [ 2.0723, -1.2491,  0.8536, -2.2255,  1.0305],
        ...0.1292],
        [ 0.1026, -0.6543, -0.7291, -0.8119,  1.4093],
        [-0.2768, -0.9552,  0.5928, -0.2453, -1.2137]])
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
---------------------------- Captured stdout call -----------------------------
EdgeConv(
  (theta): Linear(in_features=5, out_features=2, bias=True)
  (phi): Linear(in_features=5, out_features=2, bias=True)
)
________________________ test_edge_conv[2-g6-idtype0] _________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'x': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_edge_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        edge_conv = nn.EdgeConv(5, out_dim).to(ctx)
        print(edge_conv)
    
        # test pickle
        th.save(edge_conv, tmp_buffer)
    
        h0 = F.randn((g.number_of_src_nodes(), 5))
>       h1 = edge_conv(g, h0)

tests\pytorch\test_nn.py:973: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\edgeconv.py:171: in forward
    g.apply_edges(fn.v_sub_u('x', 'x', 'theta'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07290A48>
op = 'add'
lhs_data = tensor([[ 0.7281,  1.3727, -1.5217,  0.8874, -1.5569],
        [ 1.2091,  0.4248,  0.1020,  1.1541,  0.8557],
        ...0.8428],
        [ 3.4457, -0.8480, -0.9980, -1.2621,  0.1704],
        [-1.0108,  0.6933, -1.0230, -1.2971,  1.0327]])
rhs_data = tensor([[-0.7281, -1.3727,  1.5217, -0.8874,  1.5569],
        [-1.2091, -0.4248, -0.1020, -1.1541, -0.8557],
        ...0.8428],
        [-3.4457,  0.8480,  0.9980,  1.2621, -0.1704],
        [ 1.0108, -0.6933,  1.0230,  1.2971, -1.0327]])
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
---------------------------- Captured stdout call -----------------------------
EdgeConv(
  (theta): Linear(in_features=5, out_features=2, bias=True)
  (phi): Linear(in_features=5, out_features=2, bias=True)
)
________________________ test_edge_conv[2-g6-idtype1] _________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'x': Scheme(shape=(5,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_edge_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        edge_conv = nn.EdgeConv(5, out_dim).to(ctx)
        print(edge_conv)
    
        # test pickle
        th.save(edge_conv, tmp_buffer)
    
        h0 = F.randn((g.number_of_src_nodes(), 5))
>       h1 = edge_conv(g, h0)

tests\pytorch\test_nn.py:973: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\edgeconv.py:171: in forward
    g.apply_edges(fn.v_sub_u('x', 'x', 'theta'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF073FBEC8>
op = 'add'
lhs_data = tensor([[-0.7974,  0.7729,  1.0826,  0.3204, -1.1722],
        [ 0.4221,  2.0494, -1.5402,  0.7005, -0.0130],
        ...0.6546],
        [-0.9132, -1.2859,  0.9105,  1.4827,  0.7854],
        [-0.7712,  0.7687, -0.5193,  0.1445,  0.2761]])
rhs_data = tensor([[ 0.7974, -0.7729, -1.0826, -0.3204,  1.1722],
        [-0.4221, -2.0494,  1.5402, -0.7005,  0.0130],
        ...0.6546],
        [ 0.9132,  1.2859, -0.9105, -1.4827, -0.7854],
        [ 0.7712, -0.7687,  0.5193, -0.1445, -0.2761]])
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
---------------------------- Captured stdout call -----------------------------
EdgeConv(
  (theta): Linear(in_features=5, out_features=2, bias=True)
  (phi): Linear(in_features=5, out_features=2, bias=True)
)
_______________________ test_edge_conv_bi[1-g0-idtype0] _______________________

g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int32, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_edge_conv_bi(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        edge_conv = nn.EdgeConv(5, out_dim).to(ctx)
        print(edge_conv)
        h0 = F.randn((g.number_of_src_nodes(), 5))
        x0 = F.randn((g.number_of_dst_nodes(), 5))
>       h1 = edge_conv(g, (h0, x0))

tests\pytorch\test_nn.py:986: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\edgeconv.py:171: in forward
    g.apply_edges(fn.v_sub_u('x', 'x', 'theta'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07427108>
op = 'add'
lhs_data = tensor([[-0.4482, -1.0390,  2.3526, -0.3421, -0.8654],
        [ 1.7475, -1.2230,  0.4174,  1.0663, -0.0574],
        [-0.6349, -0.3127, -0.6098,  2.8471,  0.0724],
        [-1.8267,  0.7960, -0.1545,  2.3544,  1.3370]])
rhs_data = tensor([[-0.1172, -1.4900,  0.4269,  0.6653, -0.7449],
        [ 0.2760,  0.4600, -0.4274,  0.2719,  1.2312]])
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
---------------------------- Captured stdout call -----------------------------
EdgeConv(
  (theta): Linear(in_features=5, out_features=1, bias=True)
  (phi): Linear(in_features=5, out_features=1, bias=True)
)
_______________________ test_edge_conv_bi[1-g0-idtype1] _______________________

g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int64, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_edge_conv_bi(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        edge_conv = nn.EdgeConv(5, out_dim).to(ctx)
        print(edge_conv)
        h0 = F.randn((g.number_of_src_nodes(), 5))
        x0 = F.randn((g.number_of_dst_nodes(), 5))
>       h1 = edge_conv(g, (h0, x0))

tests\pytorch\test_nn.py:986: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\edgeconv.py:171: in forward
    g.apply_edges(fn.v_sub_u('x', 'x', 'theta'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0717ABC8>
op = 'add'
lhs_data = tensor([[-0.4233,  1.2786, -0.2211,  0.4992, -1.1761],
        [-1.3803,  1.0027, -0.4542,  0.7570,  0.0147],
        [ 0.8226,  0.2086, -0.9733,  1.6012,  1.1022],
        [ 1.2063, -0.9509, -2.0677, -0.8150,  0.7429]])
rhs_data = tensor([[-0.2381,  0.4569, -0.5258,  0.6573,  1.5449],
        [-0.3189, -0.5085,  0.2481, -1.9428, -0.4919]])
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
---------------------------- Captured stdout call -----------------------------
EdgeConv(
  (theta): Linear(in_features=5, out_features=1, bias=True)
  (phi): Linear(in_features=5, out_features=1, bias=True)
)
_______________________ test_edge_conv_bi[1-g1-idtype0] _______________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int32
out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_edge_conv_bi(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        edge_conv = nn.EdgeConv(5, out_dim).to(ctx)
        print(edge_conv)
        h0 = F.randn((g.number_of_src_nodes(), 5))
        x0 = F.randn((g.number_of_dst_nodes(), 5))
>       h1 = edge_conv(g, (h0, x0))

tests\pytorch\test_nn.py:986: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\edgeconv.py:171: in forward
    g.apply_edges(fn.v_sub_u('x', 'x', 'theta'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07407608>
op = 'add'
lhs_data = tensor([[-0.8523, -1.5086,  0.5729, -1.4259, -1.1680],
        [-0.5033, -0.7529,  0.1385, -1.1288, -0.8033],
        [-0.4452,  0.4813, -0.9189, -0.9522, -0.0080]])
rhs_data = tensor([[ 1.0079, -1.6104, -1.1301, -1.4617,  0.2413],
        [ 1.6081,  2.0187, -1.2759,  1.7802,  0.0058],
        ...0.1624],
        [ 0.2959,  0.7917, -0.2627, -1.4337, -0.6803],
        [-0.2641, -1.2889,  0.2008,  1.6873,  0.4732]])
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
---------------------------- Captured stdout call -----------------------------
EdgeConv(
  (theta): Linear(in_features=5, out_features=1, bias=True)
  (phi): Linear(in_features=5, out_features=1, bias=True)
)
_______________________ test_edge_conv_bi[1-g1-idtype1] _______________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int64
out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_edge_conv_bi(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        edge_conv = nn.EdgeConv(5, out_dim).to(ctx)
        print(edge_conv)
        h0 = F.randn((g.number_of_src_nodes(), 5))
        x0 = F.randn((g.number_of_dst_nodes(), 5))
>       h1 = edge_conv(g, (h0, x0))

tests\pytorch\test_nn.py:986: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\edgeconv.py:171: in forward
    g.apply_edges(fn.v_sub_u('x', 'x', 'theta'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF018513C8>
op = 'add'
lhs_data = tensor([[-0.0807,  2.2112,  1.4780, -0.5147,  0.3133],
        [-1.7687,  0.9437, -0.8441, -1.3617, -0.2863],
        [ 1.6733, -1.2069,  0.0911,  0.2173, -0.4771]])
rhs_data = tensor([[-0.8197, -0.8039,  0.8894,  0.2984, -0.7267],
        [ 0.0263,  0.7333,  0.5352,  1.2911,  0.4380],
        ...0.0802],
        [-0.4048,  0.6796,  0.0595, -0.2775, -0.3414],
        [ 0.2645, -1.3595,  1.3935, -0.3734,  1.0322]])
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
---------------------------- Captured stdout call -----------------------------
EdgeConv(
  (theta): Linear(in_features=5, out_features=1, bias=True)
  (phi): Linear(in_features=5, out_features=1, bias=True)
)
_______________________ test_edge_conv_bi[2-g0-idtype0] _______________________

g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int32, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_edge_conv_bi(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        edge_conv = nn.EdgeConv(5, out_dim).to(ctx)
        print(edge_conv)
        h0 = F.randn((g.number_of_src_nodes(), 5))
        x0 = F.randn((g.number_of_dst_nodes(), 5))
>       h1 = edge_conv(g, (h0, x0))

tests\pytorch\test_nn.py:986: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\edgeconv.py:171: in forward
    g.apply_edges(fn.v_sub_u('x', 'x', 'theta'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0741E788>
op = 'add'
lhs_data = tensor([[-0.1190, -1.5178,  1.2033, -0.7864, -1.0407],
        [-0.5289, -0.3738, -0.0890,  0.4506, -0.4597],
        [ 1.9414,  2.4532,  0.1634, -1.3393, -2.2887],
        [ 0.0984,  1.2950,  0.3104, -0.5171,  0.2836]])
rhs_data = tensor([[ 0.1230, -0.6114,  0.2976,  0.8652,  2.2191],
        [ 1.0884, -0.8765,  0.2848, -0.6950, -1.5342]])
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
---------------------------- Captured stdout call -----------------------------
EdgeConv(
  (theta): Linear(in_features=5, out_features=2, bias=True)
  (phi): Linear(in_features=5, out_features=2, bias=True)
)
_______________________ test_edge_conv_bi[2-g0-idtype1] _______________________

g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int64, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_edge_conv_bi(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        edge_conv = nn.EdgeConv(5, out_dim).to(ctx)
        print(edge_conv)
        h0 = F.randn((g.number_of_src_nodes(), 5))
        x0 = F.randn((g.number_of_dst_nodes(), 5))
>       h1 = edge_conv(g, (h0, x0))

tests\pytorch\test_nn.py:986: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\edgeconv.py:171: in forward
    g.apply_edges(fn.v_sub_u('x', 'x', 'theta'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0717AF08>
op = 'add'
lhs_data = tensor([[-0.0321, -0.3868, -0.7811,  0.7588, -0.6651],
        [ 0.6558,  1.0836,  1.4311,  0.2528,  1.1681],
        [-0.9874,  1.3505,  0.5802, -0.9096, -0.3701],
        [ 1.0805,  0.6331,  1.9989, -0.1749, -0.4437]])
rhs_data = tensor([[ 1.4245,  0.0220, -1.5458, -2.1044,  2.2117],
        [ 2.4226, -0.0622, -0.0212, -0.0769,  0.1360]])
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
---------------------------- Captured stdout call -----------------------------
EdgeConv(
  (theta): Linear(in_features=5, out_features=2, bias=True)
  (phi): Linear(in_features=5, out_features=2, bias=True)
)
_______________________ test_edge_conv_bi[2-g1-idtype0] _______________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int32
out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_edge_conv_bi(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        edge_conv = nn.EdgeConv(5, out_dim).to(ctx)
        print(edge_conv)
        h0 = F.randn((g.number_of_src_nodes(), 5))
        x0 = F.randn((g.number_of_dst_nodes(), 5))
>       h1 = edge_conv(g, (h0, x0))

tests\pytorch\test_nn.py:986: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\edgeconv.py:171: in forward
    g.apply_edges(fn.v_sub_u('x', 'x', 'theta'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF018A0688>
op = 'add'
lhs_data = tensor([[-0.0464, -1.4496,  2.5057,  0.5077,  0.7538],
        [-1.6744,  0.1616,  1.1393,  0.1523, -0.3104],
        [-0.1068, -0.0039,  0.1814,  0.3441, -0.6041]])
rhs_data = tensor([[ 0.2958, -0.1471, -0.5130, -0.2249, -0.3845],
        [-0.3507, -1.4648,  1.1246,  0.7618,  1.5231],
        ...0.7443],
        [ 0.5420,  0.0401,  1.1182, -0.0514,  0.5618],
        [ 0.8810,  0.3216, -0.0967,  0.7743,  0.2607]])
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
---------------------------- Captured stdout call -----------------------------
EdgeConv(
  (theta): Linear(in_features=5, out_features=2, bias=True)
  (phi): Linear(in_features=5, out_features=2, bias=True)
)
_______________________ test_edge_conv_bi[2-g1-idtype1] _______________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int64
out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_edge_conv_bi(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        edge_conv = nn.EdgeConv(5, out_dim).to(ctx)
        print(edge_conv)
        h0 = F.randn((g.number_of_src_nodes(), 5))
        x0 = F.randn((g.number_of_dst_nodes(), 5))
>       h1 = edge_conv(g, (h0, x0))

tests\pytorch\test_nn.py:986: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\edgeconv.py:171: in forward
    g.apply_edges(fn.v_sub_u('x', 'x', 'theta'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE12360708>
op = 'add'
lhs_data = tensor([[ 0.4952, -0.5542,  0.0810,  1.6316, -0.8154],
        [ 1.2885, -0.0511, -0.5976,  0.2904,  0.5324],
        [-0.5083,  0.8421, -1.8440,  0.1749,  0.1064]])
rhs_data = tensor([[ 1.0526,  0.0246, -0.4953,  1.4015,  0.0063],
        [-0.9872, -1.0135,  0.9635, -0.6248, -0.9060],
        ...0.2247],
        [ 0.2773, -0.5805, -0.4164, -1.7540, -1.0479],
        [-0.6316,  0.9465, -0.1150, -0.8930, -0.4859]])
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
---------------------------- Captured stdout call -----------------------------
EdgeConv(
  (theta): Linear(in_features=5, out_features=2, bias=True)
  (phi): Linear(in_features=5, out_features=2, bias=True)
)
______________________ test_dotgat_conv[1-1-g0-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
idtype = torch.int32, out_dim = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_dotgat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        dotgat = nn.DotGatConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        dotgat = dotgat.to(ctx)
    
        # test pickle
        th.save(dotgat, tmp_buffer)
    
>       h = dotgat(g, feat)

tests\pytorch\test_nn.py:1003: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\dotgatconv.py:205: in forward
    graph.apply_edges(fn.u_dot_v('ft', 'ft', 'a'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0718E2C8>
op = 'dot'
lhs_data = tensor([[[-0.4171]],

        [[ 0.1313]],

        [[-0.0037]],

        [[ 0.3055]],

        [[-0.8149]],

        ...054]],

        [[-0.4450]],

        [[ 0.1160]],

        [[ 0.1449]],

        [[ 0.5042]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[-0.4171]],

        [[ 0.1313]],

        [[-0.0037]],

        [[ 0.3055]],

        [[-0.8149]],

        ...054]],

        [[-0.4450]],

        [[ 0.1160]],

        [[ 0.1449]],

        [[ 0.5042]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_dotgat_conv[1-1-g0-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
idtype = torch.int64, out_dim = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_dotgat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        dotgat = nn.DotGatConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        dotgat = dotgat.to(ctx)
    
        # test pickle
        th.save(dotgat, tmp_buffer)
    
>       h = dotgat(g, feat)

tests\pytorch\test_nn.py:1003: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\dotgatconv.py:205: in forward
    graph.apply_edges(fn.u_dot_v('ft', 'ft', 'a'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF017E7E48>
op = 'dot'
lhs_data = tensor([[[ 1.1489]],

        [[ 0.5041]],

        [[ 0.0415]],

        [[-0.1781]],

        [[-0.3042]],

        ...948]],

        [[ 0.0383]],

        [[-0.6830]],

        [[ 0.4236]],

        [[-0.2631]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[ 1.1489]],

        [[ 0.5041]],

        [[ 0.0415]],

        [[-0.1781]],

        [[-0.3042]],

        ...948]],

        [[ 0.0383]],

        [[-0.6830]],

        [[ 0.4236]],

        [[-0.2631]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_dotgat_conv[1-1-g1-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={}
      edata_schemes={})
idtype = torch.int32, out_dim = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_dotgat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        dotgat = nn.DotGatConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        dotgat = dotgat.to(ctx)
    
        # test pickle
        th.save(dotgat, tmp_buffer)
    
>       h = dotgat(g, feat)

tests\pytorch\test_nn.py:1003: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\dotgatconv.py:205: in forward
    graph.apply_edges(fn.u_dot_v('ft', 'ft', 'a'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07425708>
op = 'dot'
lhs_data = tensor([[[ 0.0316]],

        [[-0.2107]],

        [[-0.3332]],

        [[ 0.7550]],

        [[ 0.8863]],

        ...091]],

        [[-0.0951]],

        [[ 0.3141]],

        [[ 1.2935]],

        [[-0.8214]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[ 0.0316]],

        [[-0.2107]],

        [[-0.3332]],

        [[ 0.7550]],

        [[ 0.8863]],

        ...091]],

        [[-0.0951]],

        [[ 0.3141]],

        [[ 1.2935]],

        [[-0.8214]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_dotgat_conv[1-1-g1-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={}
      edata_schemes={})
idtype = torch.int64, out_dim = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_dotgat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        dotgat = nn.DotGatConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        dotgat = dotgat.to(ctx)
    
        # test pickle
        th.save(dotgat, tmp_buffer)
    
>       h = dotgat(g, feat)

tests\pytorch\test_nn.py:1003: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\dotgatconv.py:205: in forward
    graph.apply_edges(fn.u_dot_v('ft', 'ft', 'a'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0732E048>
op = 'dot'
lhs_data = tensor([[[-0.0455]],

        [[ 1.3344]],

        [[ 1.0264]],

        [[ 0.2271]],

        [[ 0.4164]],

        ...385]],

        [[-0.6401]],

        [[ 0.2951]],

        [[ 0.0372]],

        [[-0.8960]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[-0.0455]],

        [[ 1.3344]],

        [[ 1.0264]],

        [[ 0.2271]],

        [[ 0.4164]],

        ...385]],

        [[-0.6401]],

        [[ 0.2951]],

        [[ 0.0372]],

        [[-0.8960]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_dotgat_conv[1-1-g2-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int32, out_dim = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_dotgat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        dotgat = nn.DotGatConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        dotgat = dotgat.to(ctx)
    
        # test pickle
        th.save(dotgat, tmp_buffer)
    
>       h = dotgat(g, feat)

tests\pytorch\test_nn.py:1003: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\dotgatconv.py:205: in forward
    graph.apply_edges(fn.u_dot_v('ft', 'ft', 'a'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03DFAEC8>
op = 'dot'
lhs_data = tensor([[[-0.4019]],

        [[-0.1592]],

        [[ 0.6658]],

        [[-0.5920]],

        [[-0.9961]],

        ...380]],

        [[-1.0740]],

        [[ 1.1565]],

        [[-0.2682]],

        [[-0.8298]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[-0.4019]],

        [[-0.1592]],

        [[ 0.6658]],

        [[-0.5920]],

        [[-0.9961]],

        ...380]],

        [[-1.0740]],

        [[ 1.1565]],

        [[-0.2682]],

        [[-0.8298]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_dotgat_conv[1-1-g2-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int64, out_dim = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_dotgat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        dotgat = nn.DotGatConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        dotgat = dotgat.to(ctx)
    
        # test pickle
        th.save(dotgat, tmp_buffer)
    
>       h = dotgat(g, feat)

tests\pytorch\test_nn.py:1003: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\dotgatconv.py:205: in forward
    graph.apply_edges(fn.u_dot_v('ft', 'ft', 'a'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03D17D48>
op = 'dot'
lhs_data = tensor([[[ 0.1679]],

        [[ 0.0703]],

        [[-0.9326]],

        [[-0.5227]],

        [[ 0.5781]],

        ...232]],

        [[ 0.2986]],

        [[-0.5231]],

        [[ 1.1090]],

        [[ 0.5377]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[ 0.1679]],

        [[ 0.0703]],

        [[-0.9326]],

        [[-0.5227]],

        [[ 0.5781]],

        ...232]],

        [[ 0.2986]],

        [[-0.5231]],

        [[ 1.1090]],

        [[ 0.5377]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_dotgat_conv[1-1-g3-idtype0] _______________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={}
      edata_schemes={})
idtype = torch.int32, out_dim = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_dotgat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        dotgat = nn.DotGatConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        dotgat = dotgat.to(ctx)
    
        # test pickle
        th.save(dotgat, tmp_buffer)
    
>       h = dotgat(g, feat)

tests\pytorch\test_nn.py:1003: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\dotgatconv.py:205: in forward
    graph.apply_edges(fn.u_dot_v('ft', 'ft', 'a'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0188CF88>
op = 'dot'
lhs_data = tensor([[[ 0.1912]],

        [[ 0.3232]],

        [[ 0.3688]],

        [[-0.2081]],

        [[-0.6082]],

        [[-0.7385]],

        [[-0.0534]],

        [[ 0.2735]],

        [[-0.3544]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[ 0.1912]],

        [[ 0.3232]],

        [[ 0.3688]],

        [[-0.2081]],

        [[-0.6082]],

        [[-0.7385]],

        [[-0.0534]],

        [[ 0.2735]],

        [[-0.3544]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_dotgat_conv[1-1-g3-idtype1] _______________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={}
      edata_schemes={})
idtype = torch.int64, out_dim = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_dotgat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        dotgat = nn.DotGatConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        dotgat = dotgat.to(ctx)
    
        # test pickle
        th.save(dotgat, tmp_buffer)
    
>       h = dotgat(g, feat)

tests\pytorch\test_nn.py:1003: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\dotgatconv.py:205: in forward
    graph.apply_edges(fn.u_dot_v('ft', 'ft', 'a'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE122CDA48>
op = 'dot'
lhs_data = tensor([[[ 0.3769]],

        [[-0.1504]],

        [[ 0.5771]],

        [[ 0.9457]],

        [[-0.1547]],

        [[-0.5143]],

        [[-0.1559]],

        [[ 0.9152]],

        [[ 0.6502]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[ 0.3769]],

        [[-0.1504]],

        [[ 0.5771]],

        [[ 0.9457]],

        [[-0.1547]],

        [[-0.5143]],

        [[-0.1559]],

        [[ 0.9152]],

        [[ 0.6502]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_dotgat_conv[1-1-g4-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={}
      edata_schemes={})
idtype = torch.int32, out_dim = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_dotgat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        dotgat = nn.DotGatConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        dotgat = dotgat.to(ctx)
    
        # test pickle
        th.save(dotgat, tmp_buffer)
    
>       h = dotgat(g, feat)

tests\pytorch\test_nn.py:1003: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\dotgatconv.py:205: in forward
    graph.apply_edges(fn.u_dot_v('ft', 'ft', 'a'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07365448>
op = 'dot'
lhs_data = tensor([[[-0.0502]],

        [[ 1.2972]],

        [[-0.6443]],

        [[ 0.2333]],

        [[-0.8309]],

        ...263]],

        [[-0.6932]],

        [[-0.4154]],

        [[-0.1857]],

        [[ 0.9152]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[-0.0502]],

        [[ 1.2972]],

        [[-0.6443]],

        [[ 0.2333]],

        [[-0.8309]],

        ...263]],

        [[-0.6932]],

        [[-0.4154]],

        [[-0.1857]],

        [[ 0.9152]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_dotgat_conv[1-1-g4-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={}
      edata_schemes={})
idtype = torch.int64, out_dim = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_dotgat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        dotgat = nn.DotGatConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        dotgat = dotgat.to(ctx)
    
        # test pickle
        th.save(dotgat, tmp_buffer)
    
>       h = dotgat(g, feat)

tests\pytorch\test_nn.py:1003: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\dotgatconv.py:205: in forward
    graph.apply_edges(fn.u_dot_v('ft', 'ft', 'a'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF017C5E48>
op = 'dot'
lhs_data = tensor([[[ 0.1828]],

        [[-0.4214]],

        [[-0.1871]],

        [[ 0.6381]],

        [[ 0.2076]],

        ...946]],

        [[ 0.0509]],

        [[ 0.4553]],

        [[ 0.2720]],

        [[ 0.2266]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[ 0.1828]],

        [[-0.4214]],

        [[-0.1871]],

        [[ 0.6381]],

        [[ 0.2076]],

        ...946]],

        [[ 0.0509]],

        [[ 0.4553]],

        [[ 0.2720]],

        [[ 0.2266]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_dotgat_conv[1-1-g5-idtype0] _______________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int32
out_dim = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_dotgat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        dotgat = nn.DotGatConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        dotgat = dotgat.to(ctx)
    
        # test pickle
        th.save(dotgat, tmp_buffer)
    
>       h = dotgat(g, feat)

tests\pytorch\test_nn.py:1003: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\dotgatconv.py:205: in forward
    graph.apply_edges(fn.u_dot_v('ft', 'ft', 'a'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0743C708>
op = 'dot'
lhs_data = tensor([[[ 1.1225]],

        [[ 0.5864]],

        [[ 0.0394]],

        [[-0.2209]],

        [[-0.4053]],

        [[ 1.6284]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[1.1225]],

        [[0.5864]],

        [[0.0394]]], grad_fn=<SliceBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_dotgat_conv[1-1-g5-idtype1] _______________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int64
out_dim = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_dotgat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        dotgat = nn.DotGatConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        dotgat = dotgat.to(ctx)
    
        # test pickle
        th.save(dotgat, tmp_buffer)
    
>       h = dotgat(g, feat)

tests\pytorch\test_nn.py:1003: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\dotgatconv.py:205: in forward
    graph.apply_edges(fn.u_dot_v('ft', 'ft', 'a'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE122CDD48>
op = 'dot'
lhs_data = tensor([[[-0.1615]],

        [[ 0.2548]],

        [[ 0.6277]],

        [[-0.2624]],

        [[ 0.3890]],

        [[ 0.3343]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[-0.1615]],

        [[ 0.2548]],

        [[ 0.6277]]], grad_fn=<SliceBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_dotgat_conv[1-1-g6-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={}
      edata_schemes={})
idtype = torch.int32, out_dim = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_dotgat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        dotgat = nn.DotGatConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        dotgat = dotgat.to(ctx)
    
        # test pickle
        th.save(dotgat, tmp_buffer)
    
>       h = dotgat(g, feat)

tests\pytorch\test_nn.py:1003: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\dotgatconv.py:205: in forward
    graph.apply_edges(fn.u_dot_v('ft', 'ft', 'a'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1223B148>
op = 'dot'
lhs_data = tensor([[[-0.8170]],

        [[-0.1959]],

        [[ 0.4661]],

        [[ 0.8583]],

        [[ 0.5681]],

        ...143]],

        [[ 0.4732]],

        [[-0.5134]],

        [[-0.4693]],

        [[-0.4494]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[-0.8170]],

        [[-0.1959]],

        [[ 0.4661]],

        [[ 0.8583]],

        [[ 0.5681]],

        ...143]],

        [[ 0.4732]],

        [[-0.5134]],

        [[-0.4693]],

        [[-0.4494]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_dotgat_conv[1-1-g6-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={}
      edata_schemes={})
idtype = torch.int64, out_dim = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_dotgat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        dotgat = nn.DotGatConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        dotgat = dotgat.to(ctx)
    
        # test pickle
        th.save(dotgat, tmp_buffer)
    
>       h = dotgat(g, feat)

tests\pytorch\test_nn.py:1003: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\dotgatconv.py:205: in forward
    graph.apply_edges(fn.u_dot_v('ft', 'ft', 'a'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF073F5188>
op = 'dot'
lhs_data = tensor([[[-0.1568]],

        [[ 0.4040]],

        [[-0.4166]],

        [[-0.4490]],

        [[ 0.1679]],

        ...186]],

        [[ 0.0533]],

        [[ 0.2031]],

        [[ 0.1075]],

        [[-0.4217]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[-0.1568]],

        [[ 0.4040]],

        [[-0.4166]],

        [[-0.4490]],

        [[ 0.1679]],

        ...186]],

        [[ 0.0533]],

        [[ 0.2031]],

        [[ 0.1075]],

        [[-0.4217]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_dotgat_conv[1-2-g0-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
idtype = torch.int32, out_dim = 2, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_dotgat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        dotgat = nn.DotGatConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        dotgat = dotgat.to(ctx)
    
        # test pickle
        th.save(dotgat, tmp_buffer)
    
>       h = dotgat(g, feat)

tests\pytorch\test_nn.py:1003: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\dotgatconv.py:205: in forward
    graph.apply_edges(fn.u_dot_v('ft', 'ft', 'a'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07443F08>
op = 'dot'
lhs_data = tensor([[[-0.1434,  0.3204]],

        [[-0.8403, -0.3625]],

        [[ 0.8121,  0.2371]],

        [[ 0.2334, -0.221...,

        [[ 0.3428,  0.4673]],

        [[ 0.0996,  0.1325]],

        [[ 0.8194,  0.6052]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[-0.1434,  0.3204]],

        [[-0.8403, -0.3625]],

        [[ 0.8121,  0.2371]],

        [[ 0.2334, -0.221...,

        [[ 0.3428,  0.4673]],

        [[ 0.0996,  0.1325]],

        [[ 0.8194,  0.6052]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_dotgat_conv[1-2-g0-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
idtype = torch.int64, out_dim = 2, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_dotgat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        dotgat = nn.DotGatConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        dotgat = dotgat.to(ctx)
    
        # test pickle
        th.save(dotgat, tmp_buffer)
    
>       h = dotgat(g, feat)

tests\pytorch\test_nn.py:1003: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\dotgatconv.py:205: in forward
    graph.apply_edges(fn.u_dot_v('ft', 'ft', 'a'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03DD0248>
op = 'dot'
lhs_data = tensor([[[ 0.8656,  0.0412]],

        [[-0.2546, -0.6253]],

        [[ 0.7924,  0.8917]],

        [[ 0.0384,  1.233...,

        [[ 0.0440, -0.6401]],

        [[ 0.0309, -1.5200]],

        [[ 0.2442, -0.6399]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[ 0.8656,  0.0412]],

        [[-0.2546, -0.6253]],

        [[ 0.7924,  0.8917]],

        [[ 0.0384,  1.233...,

        [[ 0.0440, -0.6401]],

        [[ 0.0309, -1.5200]],

        [[ 0.2442, -0.6399]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_dotgat_conv[1-2-g1-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={}
      edata_schemes={})
idtype = torch.int32, out_dim = 2, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_dotgat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        dotgat = nn.DotGatConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        dotgat = dotgat.to(ctx)
    
        # test pickle
        th.save(dotgat, tmp_buffer)
    
>       h = dotgat(g, feat)

tests\pytorch\test_nn.py:1003: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\dotgatconv.py:205: in forward
    graph.apply_edges(fn.u_dot_v('ft', 'ft', 'a'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07396708>
op = 'dot'
lhs_data = tensor([[[-6.1478e-01, -2.5440e-01]],

        [[ 6.9270e-01,  2.9113e-01]],

        [[-1.3624e+00, -8.7509e-01]],

 ...,  1.8246e-01]],

        [[-7.5135e-01, -7.3620e-01]],

        [[ 3.1665e-01,  6.3831e-02]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[-6.1478e-01, -2.5440e-01]],

        [[ 6.9270e-01,  2.9113e-01]],

        [[-1.3624e+00, -8.7509e-01]],

 ...,  1.8246e-01]],

        [[-7.5135e-01, -7.3620e-01]],

        [[ 3.1665e-01,  6.3831e-02]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_dotgat_conv[1-2-g1-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={}
      edata_schemes={})
idtype = torch.int64, out_dim = 2, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_dotgat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        dotgat = nn.DotGatConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        dotgat = dotgat.to(ctx)
    
        # test pickle
        th.save(dotgat, tmp_buffer)
    
>       h = dotgat(g, feat)

tests\pytorch\test_nn.py:1003: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\dotgatconv.py:205: in forward
    graph.apply_edges(fn.u_dot_v('ft', 'ft', 'a'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0742C988>
op = 'dot'
lhs_data = tensor([[[ 0.0764,  0.0497]],

        [[ 0.1910,  0.5693]],

        [[ 0.9781,  1.2985]],

        [[-0.0146, -0.807...,

        [[-0.9079,  1.0759]],

        [[ 0.6067,  0.3260]],

        [[ 0.0073,  0.8853]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[ 0.0764,  0.0497]],

        [[ 0.1910,  0.5693]],

        [[ 0.9781,  1.2985]],

        [[-0.0146, -0.807...,

        [[-0.9079,  1.0759]],

        [[ 0.6067,  0.3260]],

        [[ 0.0073,  0.8853]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_dotgat_conv[1-2-g2-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int32, out_dim = 2, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_dotgat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        dotgat = nn.DotGatConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        dotgat = dotgat.to(ctx)
    
        # test pickle
        th.save(dotgat, tmp_buffer)
    
>       h = dotgat(g, feat)

tests\pytorch\test_nn.py:1003: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\dotgatconv.py:205: in forward
    graph.apply_edges(fn.u_dot_v('ft', 'ft', 'a'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07411288>
op = 'dot'
lhs_data = tensor([[[ 2.2440, -0.8786]],

        [[-0.1177, -0.0369]],

        [[-1.1287,  1.2949]],

        [[-0.5776, -0.003...,

        [[-0.8601,  0.3388]],

        [[-0.2230, -0.5133]],

        [[ 1.3118, -1.2113]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[ 2.2440, -0.8786]],

        [[-0.1177, -0.0369]],

        [[-1.1287,  1.2949]],

        [[-0.5776, -0.003...,

        [[-0.8601,  0.3388]],

        [[-0.2230, -0.5133]],

        [[ 1.3118, -1.2113]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_dotgat_conv[1-2-g2-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int64, out_dim = 2, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_dotgat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        dotgat = nn.DotGatConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        dotgat = dotgat.to(ctx)
    
        # test pickle
        th.save(dotgat, tmp_buffer)
    
>       h = dotgat(g, feat)

tests\pytorch\test_nn.py:1003: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\dotgatconv.py:205: in forward
    graph.apply_edges(fn.u_dot_v('ft', 'ft', 'a'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07472708>
op = 'dot'
lhs_data = tensor([[[-1.3013, -0.8579]],

        [[-0.3501, -0.1292]],

        [[-0.4198, -0.7377]],

        [[-0.7654, -0.321...,

        [[ 0.5930, -0.2829]],

        [[ 1.1805,  0.4619]],

        [[-1.4865, -1.5151]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[-1.3013, -0.8579]],

        [[-0.3501, -0.1292]],

        [[-0.4198, -0.7377]],

        [[-0.7654, -0.321...,

        [[ 0.5930, -0.2829]],

        [[ 1.1805,  0.4619]],

        [[-1.4865, -1.5151]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_dotgat_conv[1-2-g3-idtype0] _______________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={}
      edata_schemes={})
idtype = torch.int32, out_dim = 2, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_dotgat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        dotgat = nn.DotGatConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        dotgat = dotgat.to(ctx)
    
        # test pickle
        th.save(dotgat, tmp_buffer)
    
>       h = dotgat(g, feat)

tests\pytorch\test_nn.py:1003: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\dotgatconv.py:205: in forward
    graph.apply_edges(fn.u_dot_v('ft', 'ft', 'a'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF018EAB08>
op = 'dot'
lhs_data = tensor([[[ 0.6153, -1.6369]],

        [[ 0.7997,  0.1144]],

        [[-0.5679, -0.8437]],

        [[-0.6164,  0.159...,

        [[-0.4480,  0.4896]],

        [[ 1.2006, -0.5437]],

        [[ 0.5249, -0.6020]]], grad_fn=<ViewBackward>)Submodule path 'third_party/METIS': checked out '10603482cc055626ff99d2ac3ab7e73d1119439d'
Submodule 'GKlib' (https://github.com/KarypisLab/GKlib.git) registered for path 'third_party/METIS/GKlib'
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/METIS/GKlib'...
Submodule path 'third_party/METIS/GKlib': checked out '62de20c7f97c719abdc96d7c3e266a1f7cb52bc5'
Submodule path 'third_party/dlpack': checked out 'e2bdd3bee8cb6501558042633fa59144cc8b7f5f'
Submodule path 'third_party/dmlc-core': checked out 'bfad207b448480783a1f428ae3d93d87032d8349'

rhs_data = tensor([[[ 0.6153, -1.6369]],

        [[ 0.7997,  0.1144]],

        [[-0.5679, -0.8437]],

        [[-0.6164,  0.159...,

        [[-0.4480,  0.4896]],

        [[ 1.2006, -0.5437]],

        [[ 0.5249, -0.6020]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_dotgat_conv[1-2-g3-idtype1] _______________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={}
      edata_schemes={})
idtype = torch.int64, out_dim = 2, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_dotgat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        dotgat = nn.DotGatConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        dotgat = dotgat.to(ctx)
    
        # test pickle
        th.save(dotgat, tmp_buffer)
    
>       h = dotgat(g, feat)

tests\pytorch\test_nn.py:1003: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\dotgatconv.py:205: in forward
    graph.apply_edges(fn.u_dot_v('ft', 'ft', 'a'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE123A48C8>
op = 'dot'
lhs_data = tensor([[[-1.1685, -0.0319]],

        [[ 0.3663, -0.3267]],

        [[ 0.4925, -0.2051]],

        [[-0.4170, -1.306...,

        [[ 0.2416,  0.6640]],

        [[-0.0258,  0.5421]],

        [[-0.3742,  0.0142]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[-1.1685, -0.0319]],

        [[ 0.3663, -0.3267]],

        [[ 0.4925, -0.2051]],

        [[-0.4170, -1.306...,

        [[ 0.2416,  0.6640]],

        [[-0.0258,  0.5421]],

        [[-0.3742,  0.0142]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_dotgat_conv[1-2-g4-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={}
      edata_schemes={})
idtype = torch.int32, out_dim = 2, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_dotgat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        dotgat = nn.DotGatConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        dotgat = dotgat.to(ctx)
    
        # test pickle
        th.save(dotgat, tmp_buffer)
    
>       h = dotgat(g, feat)

tests\pytorch\test_nn.py:1003: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\dotgatconv.py:205: in forward
    graph.apply_edges(fn.u_dot_v('ft', 'ft', 'a'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE123DCC08>
op = 'dot'
lhs_data = tensor([[[-0.6726, -0.8222]],

        [[ 0.2960, -0.3414]],

        [[ 0.6060,  0.4467]],

        [[ 0.4379,  0.035...,

        [[-0.9098, -0.6567]],

        [[-0.8609, -1.7556]],

        [[ 0.9726,  0.8830]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[-0.6726, -0.8222]],

        [[ 0.2960, -0.3414]],

        [[ 0.6060,  0.4467]],

        [[ 0.4379,  0.035...,

        [[-0.9098, -0.6567]],

        [[-0.8609, -1.7556]],

        [[ 0.9726,  0.8830]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_dotgat_conv[1-2-g4-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={}
      edata_schemes={})
idtype = torch.int64, out_dim = 2, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_dotgat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        dotgat = nn.DotGatConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        dotgat = dotgat.to(ctx)
    
        # test pickle
        th.save(dotgat, tmp_buffer)
    
>       h = dotgat(g, feat)

tests\pytorch\test_nn.py:1003: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\dotgatconv.py:205: in forward
    graph.apply_edges(fn.u_dot_v('ft', 'ft', 'a'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07472C88>
op = 'dot'
lhs_data = tensor([[[-0.3572,  0.5924]],

        [[ 0.3099, -0.3014]],

        [[ 0.5014, -0.3931]],

        [[ 0.2107, -0.216...,

        [[-0.7109,  0.0086]],

        [[-0.0064, -0.0604]],

        [[-0.0854, -1.9289]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[-0.3572,  0.5924]],

        [[ 0.3099, -0.3014]],

        [[ 0.5014, -0.3931]],

        [[ 0.2107, -0.216...,

        [[-0.7109,  0.0086]],

        [[-0.0064, -0.0604]],

        [[-0.0854, -1.9289]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_dotgat_conv[1-2-g5-idtype0] _______________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int32
out_dim = 2, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_dotgat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        dotgat = nn.DotGatConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        dotgat = dotgat.to(ctx)
    
        # test pickle
        th.save(dotgat, tmp_buffer)
    
>       h = dotgat(g, feat)

tests\pytorch\test_nn.py:1003: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\dotgatconv.py:205: in forward
    graph.apply_edges(fn.u_dot_v('ft', 'ft', 'a'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03DA8148>
op = 'dot'
lhs_data = tensor([[[ 0.9792, -0.1193]],

        [[-0.1085,  0.1791]],

        [[ 0.0034,  0.4012]],

        [[ 0.7534, -0.2448]],

        [[ 0.2681, -0.4127]],

        [[ 0.0963, -0.4894]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[ 0.9792, -0.1193]],

        [[-0.1085,  0.1791]],

        [[ 0.0034,  0.4012]]], grad_fn=<SliceBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_dotgat_conv[1-2-g5-idtype1] _______________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int64
out_dim = 2, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_dotgat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        dotgat = nn.DotGatConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        dotgat = dotgat.to(ctx)
    
        # test pickle
        th.save(dotgat, tmp_buffer)
    
>       h = dotgat(g, feat)

tests\pytorch\test_nn.py:1003: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\dotgatconv.py:205: in forward
    graph.apply_edges(fn.u_dot_v('ft', 'ft', 'a'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE122C5C48>
op = 'dot'
lhs_data = tensor([[[ 0.9397, -0.3649]],

        [[-0.2243, -0.4058]],

        [[-1.0624,  0.1599]],

        [[ 0.3047, -0.5614]],

        [[-0.5164, -0.0374]],

        [[-0.4025,  0.2307]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[ 0.9397, -0.3649]],

        [[-0.2243, -0.4058]],

        [[-1.0624,  0.1599]]], grad_fn=<SliceBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_dotgat_conv[1-2-g6-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={}
      edata_schemes={})
idtype = torch.int32, out_dim = 2, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_dotgat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        dotgat = nn.DotGatConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        dotgat = dotgat.to(ctx)
    
        # test pickle
        th.save(dotgat, tmp_buffer)
    
>       h = dotgat(g, feat)

tests\pytorch\test_nn.py:1003: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\dotgatconv.py:205: in forward
    graph.apply_edges(fn.u_dot_v('ft', 'ft', 'a'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE122BCA08>
op = 'dot'
lhs_data = tensor([[[ 0.2864,  0.2502]],

        [[ 0.2364,  0.2355]],

        [[-0.1403, -0.3415]],

        [[-0.0506, -0.166...,

        [[ 1.0967,  0.4553]],

        [[ 0.2856,  0.1267]],

        [[ 0.2275,  0.0083]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[ 0.2864,  0.2502]],

        [[ 0.2364,  0.2355]],

        [[-0.1403, -0.3415]],

        [[-0.0506, -0.166...,

        [[ 1.0967,  0.4553]],

        [[ 0.2856,  0.1267]],

        [[ 0.2275,  0.0083]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_dotgat_conv[1-2-g6-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={}
      edata_schemes={})
idtype = torch.int64, out_dim = 2, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_dotgat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        dotgat = nn.DotGatConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        dotgat = dotgat.to(ctx)
    
        # test pickle
        th.save(dotgat, tmp_buffer)
    
>       h = dotgat(g, feat)

tests\pytorch\test_nn.py:1003: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\dotgatconv.py:205: in forward
    graph.apply_edges(fn.u_dot_v('ft', 'ft', 'a'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03DA8888>
op = 'dot'
lhs_data = tensor([[[-0.3321, -0.0618]],

        [[ 1.1738,  0.3312]],

        [[-0.4129, -0.3142]],

        [[ 0.6808,  0.482...,

        [[ 0.2299, -0.2887]],

        [[-0.0339, -0.3812]],

        [[-1.2863, -0.6661]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[-0.3321, -0.0618]],

        [[ 1.1738,  0.3312]],

        [[-0.4129, -0.3142]],

        [[ 0.6808,  0.482...,

        [[ 0.2299, -0.2887]],

        [[-0.0339, -0.3812]],

        [[-1.2863, -0.6661]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_dotgat_conv[4-1-g0-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
idtype = torch.int32, out_dim = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_dotgat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        dotgat = nn.DotGatConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        dotgat = dotgat.to(ctx)
    
        # test pickle
        th.save(dotgat, tmp_buffer)
    
>       h = dotgat(g, feat)

tests\pytorch\test_nn.py:1003: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\dotgatconv.py:205: in forward
    graph.apply_edges(fn.u_dot_v('ft', 'ft', 'a'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE123D1688>
op = 'dot'
lhs_data = tensor([[[ 0.2628],
         [ 0.4857],
         [-0.4749],
         [-0.5497]],

        [[ 0.1419],
         [ 0.040... [ 0.4177]],

        [[-0.2613],
         [-0.5927],
         [-0.1104],
         [ 0.7921]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[ 0.2628],
         [ 0.4857],
         [-0.4749],
         [-0.5497]],

        [[ 0.1419],
         [ 0.040... [ 0.4177]],

        [[-0.2613],
         [-0.5927],
         [-0.1104],
         [ 0.7921]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_dotgat_conv[4-1-g0-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
idtype = torch.int64, out_dim = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_dotgat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        dotgat = nn.DotGatConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        dotgat = dotgat.to(ctx)
    
        # test pickle
        th.save(dotgat, tmp_buffer)
    
>       h = dotgat(g, feat)

tests\pytorch\test_nn.py:1003: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\dotgatconv.py:205: in forward
    graph.apply_edges(fn.u_dot_v('ft', 'ft', 'a'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE122DF688>
op = 'dot'
lhs_data = tensor([[[ 0.9615],
         [-0.5329],
         [-0.6953],
         [-0.0452]],

        [[-0.0382],
         [-0.020... [ 0.2506]],

        [[ 0.7174],
         [-0.3499],
         [-0.1276],
         [ 0.5352]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[ 0.9615],
         [-0.5329],
         [-0.6953],
         [-0.0452]],

        [[-0.0382],
         [-0.020... [ 0.2506]],

        [[ 0.7174],
         [-0.3499],
         [-0.1276],
         [ 0.5352]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_dotgat_conv[4-1-g1-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={}
      edata_schemes={})
idtype = torch.int32, out_dim = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_dotgat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        dotgat = nn.DotGatConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        dotgat = dotgat.to(ctx)
    
        # test pickle
        th.save(dotgat, tmp_buffer)
    
>       h = dotgat(g, feat)

tests\pytorch\test_nn.py:1003: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\dotgatconv.py:205: in forward
    graph.apply_edges(fn.u_dot_v('ft', 'ft', 'a'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF073A8BC8>
op = 'dot'
lhs_data = tensor([[[ 2.5271e-01],
         [ 7.7926e-02],
         [-1.6030e-01],
         [ 2.4801e-01]],

        [[ 4.8758e-0...      [[ 6.9294e-01],
         [ 1.3202e+00],
         [-1.4746e+00],
         [ 2.4003e-02]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[ 2.5271e-01],
         [ 7.7926e-02],
         [-1.6030e-01],
         [ 2.4801e-01]],

        [[ 4.8758e-0...      [[ 6.9294e-01],
         [ 1.3202e+00],
         [-1.4746e+00],
         [ 2.4003e-02]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_dotgat_conv[4-1-g1-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={}
      edata_schemes={})
idtype = torch.int64, out_dim = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_dotgat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        dotgat = nn.DotGatConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        dotgat = dotgat.to(ctx)
    
        # test pickle
        th.save(dotgat, tmp_buffer)
    
>       h = dotgat(g, feat)

tests\pytorch\test_nn.py:1003: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\dotgatconv.py:205: in forward
    graph.apply_edges(fn.u_dot_v('ft', 'ft', 'a'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF018F7748>
op = 'dot'
lhs_data = tensor([[[ 0.3748],
         [ 0.7783],
         [ 0.1260],
         [ 0.2537]],

        [[-0.3613],
         [-0.277... [ 0.1384]],

        [[ 0.1680],
         [ 0.2012],
         [-0.1219],
         [-0.4177]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[ 0.3748],
         [ 0.7783],
         [ 0.1260],
         [ 0.2537]],

        [[-0.3613],
         [-0.277... [ 0.1384]],

        [[ 0.1680],
         [ 0.2012],
         [-0.1219],
         [-0.4177]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_dotgat_conv[4-1-g2-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int32, out_dim = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_dotgat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        dotgat = nn.DotGatConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        dotgat = dotgat.to(ctx)
    
        # test pickle
        th.save(dotgat, tmp_buffer)
    
>       h = dotgat(g, feat)

tests\pytorch\test_nn.py:1003: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\dotgatconv.py:205: in forward
    graph.apply_edges(fn.u_dot_v('ft', 'ft', 'a'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0731E4C8>
op = 'dot'
lhs_data = tensor([[[ 0.0588],
         [-0.9932],
         [ 0.7851],
         [ 0.4306]],

        [[-0.5548],
         [ 0.921... [-0.1073]],

        [[ 0.1017],
         [-0.2846],
         [-0.1900],
         [-0.1226]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[ 0.0588],
         [-0.9932],
         [ 0.7851],
         [ 0.4306]],

        [[-0.5548],
         [ 0.921... [-0.1073]],

        [[ 0.1017],
         [-0.2846],
         [-0.1900],
         [-0.1226]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_dotgat_conv[4-1-g2-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int64, out_dim = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_dotgat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        dotgat = nn.DotGatConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        dotgat = dotgat.to(ctx)
    
        # test pickle
        th.save(dotgat, tmp_buffer)
    
>       h = dotgat(g, feat)

tests\pytorch\test_nn.py:1003: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\dotgatconv.py:205: in forward
    graph.apply_edges(fn.u_dot_v('ft', 'ft', 'a'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE122C0A08>
op = 'dot'
lhs_data = tensor([[[-0.4551],
         [-0.8097],
         [ 0.7426],
         [-0.2365]],

        [[-0.7178],
         [ 0.066... [ 0.2029]],

        [[ 0.0706],
         [ 0.4135],
         [-0.6640],
         [ 0.7454]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[-0.4551],
         [-0.8097],
         [ 0.7426],
         [-0.2365]],

        [[-0.7178],
         [ 0.066... [ 0.2029]],

        [[ 0.0706],
         [ 0.4135],
         [-0.6640],
         [ 0.7454]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_dotgat_conv[4-1-g3-idtype0] _______________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={}
      edata_schemes={})
idtype = torch.int32, out_dim = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_dotgat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        dotgat = nn.DotGatConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        dotgat = dotgat.to(ctx)
    
        # test pickle
        th.save(dotgat, tmp_buffer)
    
>       h = dotgat(g, feat)

tests\pytorch\test_nn.py:1003: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\dotgatconv.py:205: in forward
    graph.apply_edges(fn.u_dot_v('ft', 'ft', 'a'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF077C60C8>
op = 'dot'
lhs_data = tensor([[[ 0.5471],
         [-0.0331],
         [-0.1496],
         [-0.2486]],

        [[-0.2693],
         [-0.449... [ 0.2819]],

        [[ 0.1578],
         [ 0.1815],
         [-0.0084],
         [-0.0032]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[ 0.5471],
         [-0.0331],
         [-0.1496],
         [-0.2486]],

        [[-0.2693],
         [-0.449... [ 0.2819]],

        [[ 0.1578],
         [ 0.1815],
         [-0.0084],
         [-0.0032]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_dotgat_conv[4-1-g3-idtype1] _______________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={}
      edata_schemes={})
idtype = torch.int64, out_dim = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_dotgat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        dotgat = nn.DotGatConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        dotgat = dotgat.to(ctx)
    
        # test pickle
        th.save(dotgat, tmp_buffer)
    
>       h = dotgat(g, feat)

tests\pytorch\test_nn.py:1003: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\dotgatconv.py:205: in forward
    graph.apply_edges(fn.u_dot_v('ft', 'ft', 'a'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0736D108>
op = 'dot'
lhs_data = tensor([[[-8.0979e-01],
         [-4.1980e-01],
         [ 1.5915e+00],
         [ 8.0311e-01]],

        [[-1.2550e-0...      [[ 5.2418e-01],
         [ 2.4671e-01],
         [-4.2148e-01],
         [-3.8784e-01]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[-8.0979e-01],
         [-4.1980e-01],
         [ 1.5915e+00],
         [ 8.0311e-01]],

        [[-1.2550e-0...      [[ 5.2418e-01],
         [ 2.4671e-01],
         [-4.2148e-01],
         [-3.8784e-01]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_dotgat_conv[4-1-g4-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={}
      edata_schemes={})
idtype = torch.int32, out_dim = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_dotgat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        dotgat = nn.DotGatConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        dotgat = dotgat.to(ctx)
    
        # test pickle
        th.save(dotgat, tmp_buffer)
    
>       h = dotgat(g, feat)

tests\pytorch\test_nn.py:1003: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\dotgatconv.py:205: in forward
    graph.apply_edges(fn.u_dot_v('ft', 'ft', 'a'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE122BF788>
op = 'dot'
lhs_data = tensor([[[ 2.3130],
         [ 0.1837],
         [ 0.8431],
         [-1.8198]],

        [[-0.4535],
         [ 0.331... [-0.7223]],

        [[ 0.1381],
         [ 0.7241],
         [ 0.5618],
         [ 0.4803]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[ 2.3130],
         [ 0.1837],
         [ 0.8431],
         [-1.8198]],

        [[-0.4535],
         [ 0.331... [-0.7223]],

        [[ 0.1381],
         [ 0.7241],
         [ 0.5618],
         [ 0.4803]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_dotgat_conv[4-1-g4-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={}
      edata_schemes={})
idtype = torch.int64, out_dim = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_dotgat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        dotgat = nn.DotGatConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        dotgat = dotgat.to(ctx)
    
        # test pickle
        th.save(dotgat, tmp_buffer)
    
>       h = dotgat(g, feat)

tests\pytorch\test_nn.py:1003: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\dotgatconv.py:205: in forward
    graph.apply_edges(fn.u_dot_v('ft', 'ft', 'a'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03DF8888>
op = 'dot'
lhs_data = tensor([[[ 0.5456],
         [ 0.0891],
         [ 0.8867],
         [-0.5136]],

        [[ 0.1496],
         [-0.124... [-0.0805]],

        [[-0.5656],
         [-0.2194],
         [-0.8757],
         [ 0.3788]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[ 0.5456],
         [ 0.0891],
         [ 0.8867],
         [-0.5136]],

        [[ 0.1496],
         [-0.124... [-0.0805]],

        [[-0.5656],
         [-0.2194],
         [-0.8757],
         [ 0.3788]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_dotgat_conv[4-1-g5-idtype0] _______________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int32
out_dim = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_dotgat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        dotgat = nn.DotGatConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        dotgat = dotgat.to(ctx)
    
        # test pickle
        th.save(dotgat, tmp_buffer)
    
>       h = dotgat(g, feat)

tests\pytorch\test_nn.py:1003: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\dotgatconv.py:205: in forward
    graph.apply_edges(fn.u_dot_v('ft', 'ft', 'a'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07995CC8>
op = 'dot'
lhs_data = tensor([[[ 0.3592],
         [-0.3527],
         [-0.2985],
         [ 0.0539]],

        [[-1.6622],
         [ 0.424... [ 0.3472]],

        [[ 0.2734],
         [ 0.8207],
         [ 0.4990],
         [-0.0491]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[ 0.3592],
         [-0.3527],
         [-0.2985],
         [ 0.0539]],

        [[-1.6622],
         [ 0.424...[ 0.7632]],

        [[-0.0163],
         [ 0.0141],
         [ 0.6946],
         [ 0.1270]]], grad_fn=<SliceBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_dotgat_conv[4-1-g5-idtype1] _______________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int64
out_dim = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_dotgat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        dotgat = nn.DotGatConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        dotgat = dotgat.to(ctx)
    
        # test pickle
        th.save(dotgat, tmp_buffer)
    
>       h = dotgat(g, feat)

tests\pytorch\test_nn.py:1003: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\dotgatconv.py:205: in forward
    graph.apply_edges(fn.u_dot_v('ft', 'ft', 'a'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0736D3C8>
op = 'dot'
lhs_data = tensor([[[ 0.0602],
         [ 0.1747],
         [-0.2630],
         [-0.1540]],

        [[ 0.7330],
         [-1.072... [-0.4085]],

        [[-0.2021],
         [-0.8088],
         [ 0.1552],
         [-0.3202]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[ 0.0602],
         [ 0.1747],
         [-0.2630],
         [-0.1540]],

        [[ 0.7330],
         [-1.072...[ 0.1855]],

        [[-0.6858],
         [-0.0074],
         [-0.1308],
         [-0.3257]]], grad_fn=<SliceBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_dotgat_conv[4-1-g6-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={}
      edata_schemes={})
idtype = torch.int32, out_dim = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_dotgat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        dotgat = nn.DotGatConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        dotgat = dotgat.to(ctx)
    
        # test pickle
        th.save(dotgat, tmp_buffer)
    
>       h = dotgat(g, feat)

tests\pytorch\test_nn.py:1003: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\dotgatconv.py:205: in forward
    graph.apply_edges(fn.u_dot_v('ft', 'ft', 'a'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07427CC8>
op = 'dot'
lhs_data = tensor([[[ 0.0898],
         [-0.0545],
         [-0.3276],
         [ 0.4689]],

        [[-0.0208],
         [ 0.596... [-0.0067]],

        [[ 0.4353],
         [-0.4514],
         [-0.5834],
         [ 0.7401]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[ 0.0898],
         [-0.0545],
         [-0.3276],
         [ 0.4689]],

        [[-0.0208],
         [ 0.596... [-0.0067]],

        [[ 0.4353],
         [-0.4514],
         [-0.5834],
         [ 0.7401]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_dotgat_conv[4-1-g6-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={}
      edata_schemes={})
idtype = torch.int64, out_dim = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_dotgat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        dotgat = nn.DotGatConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        dotgat = dotgat.to(ctx)
    
        # test pickle
        th.save(dotgat, tmp_buffer)
    
>       h = dotgat(g, feat)

tests\pytorch\test_nn.py:1003: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\dotgatconv.py:205: in forward
    graph.apply_edges(fn.u_dot_v('ft', 'ft', 'a'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE122232C8>
op = 'dot'
lhs_data = tensor([[[ 0.2853],
         [-0.0968],
         [-0.1306],
         [-0.1212]],

        [[-0.4239],
         [ 0.325... [-0.1908]],

        [[ 0.9640],
         [-0.2568],
         [ 0.3576],
         [-0.0646]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[ 0.2853],
         [-0.0968],
         [-0.1306],
         [-0.1212]],

        [[-0.4239],
         [ 0.325... [-0.1908]],

        [[ 0.9640],
         [-0.2568],
         [ 0.3576],
         [-0.0646]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_dotgat_conv[4-2-g0-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
idtype = torch.int32, out_dim = 2, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_dotgat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        dotgat = nn.DotGatConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        dotgat = dotgat.to(ctx)
    
        # test pickle
        th.save(dotgat, tmp_buffer)
    
>       h = dotgat(g, feat)

tests\pytorch\test_nn.py:1003: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\dotgatconv.py:205: in forward
    graph.apply_edges(fn.u_dot_v('ft', 'ft', 'a'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF078B0AC8>
op = 'dot'
lhs_data = tensor([[[ 0.4931, -0.3066],
         [-0.7155, -1.3766],
         [-0.0161, -0.2475],
         [-0.3122, -0.5728]],

...2334],
         [ 0.0299, -0.1628],
         [-0.0382, -0.4620],
         [-0.0407,  0.5088]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[ 0.4931, -0.3066],
         [-0.7155, -1.3766],
         [-0.0161, -0.2475],
         [-0.3122, -0.5728]],

...2334],
         [ 0.0299, -0.1628],
         [-0.0382, -0.4620],
         [-0.0407,  0.5088]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_dotgat_conv[4-2-g0-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
idtype = torch.int64, out_dim = 2, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_dotgat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        dotgat = nn.DotGatConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        dotgat = dotgat.to(ctx)
    
        # test pickle
        th.save(dotgat, tmp_buffer)
    
>       h = dotgat(g, feat)

tests\pytorch\test_nn.py:1003: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\dotgatconv.py:205: in forward
    graph.apply_edges(fn.u_dot_v('ft', 'ft', 'a'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF070E6A48>
op = 'dot'
lhs_data = tensor([[[-1.1386e+00, -4.4391e-01],
         [ 3.7796e-01,  8.4842e-01],
         [ 3.2564e-01, -2.6513e-01],
       ...e-01,  9.8109e-02],
         [-1.7835e-02, -7.3542e-01],
         [ 6.7590e-01, -1.6140e-01]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[-1.1386e+00, -4.4391e-01],
         [ 3.7796e-01,  8.4842e-01],
         [ 3.2564e-01, -2.6513e-01],
       ...e-01,  9.8109e-02],
         [-1.7835e-02, -7.3542e-01],
         [ 6.7590e-01, -1.6140e-01]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_dotgat_conv[4-2-g1-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={}
      edata_schemes={})
idtype = torch.int32, out_dim = 2, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_dotgat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        dotgat = nn.DotGatConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        dotgat = dotgat.to(ctx)
    
        # test pickle
        th.save(dotgat, tmp_buffer)
    
>       h = dotgat(g, feat)

tests\pytorch\test_nn.py:1003: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\dotgatconv.py:205: in forward
    graph.apply_edges(fn.u_dot_v('ft', 'ft', 'a'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF030E8248>
op = 'dot'
lhs_data = tensor([[[-0.1413,  0.0599],
         [ 0.4125, -0.1218],
         [ 0.4898, -0.3687],
         [-0.0029, -0.3604]],

...6279],
         [ 0.3096, -0.2107],
         [ 0.4226, -0.1336],
         [ 0.1092, -0.0199]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[-0.1413,  0.0599],
         [ 0.4125, -0.1218],
         [ 0.4898, -0.3687],
         [-0.0029, -0.3604]],

...6279],
         [ 0.3096, -0.2107],
         [ 0.4226, -0.1336],
         [ 0.1092, -0.0199]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_dotgat_conv[4-2-g1-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={}
      edata_schemes={})
idtype = torch.int64, out_dim = 2, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_dotgat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        dotgat = nn.DotGatConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        dotgat = dotgat.to(ctx)
    
        # test pickle
        th.save(dotgat, tmp_buffer)
    
>       h = dotgat(g, feat)

tests\pytorch\test_nn.py:1003: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\dotgatconv.py:205: in forward
    graph.apply_edges(fn.u_dot_v('ft', 'ft', 'a'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07779048>
op = 'dot'
lhs_data = tensor([[[-0.4017, -0.6442],
         [ 0.2983, -0.1432],
         [-1.0643, -0.5747],
         [-0.3928, -0.8110]],

...5875],
         [ 0.2696,  0.8920],
         [ 0.4336, -0.6929],
         [ 0.0164, -0.7280]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[-0.4017, -0.6442],
         [ 0.2983, -0.1432],
         [-1.0643, -0.5747],
         [-0.3928, -0.8110]],

...5875],
         [ 0.2696,  0.8920],
         [ 0.4336, -0.6929],
         [ 0.0164, -0.7280]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_dotgat_conv[4-2-g2-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int32, out_dim = 2, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_dotgat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        dotgat = nn.DotGatConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        dotgat = dotgat.to(ctx)
    
        # test pickle
        th.save(dotgat, tmp_buffer)
    
>       h = dotgat(g, feat)

tests\pytorch\test_nn.py:1003: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\dotgatconv.py:205: in forward
    graph.apply_edges(fn.u_dot_v('ft', 'ft', 'a'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE122BE9C8>
op = 'dot'
lhs_data = tensor([[[ 0.3967, -1.1596],
         [-0.2107, -0.0525],
         [-0.3646,  0.6873],
         [ 1.3665,  0.2432]],

...4121],
         [ 0.6462, -0.2959],
         [-0.8253,  0.0852],
         [ 0.5070,  0.5118]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[ 0.3967, -1.1596],
         [-0.2107, -0.0525],
         [-0.3646,  0.6873],
         [ 1.3665,  0.2432]],

...4121],
         [ 0.6462, -0.2959],
         [-0.8253,  0.0852],
         [ 0.5070,  0.5118]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_dotgat_conv[4-2-g2-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int64, out_dim = 2, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_dotgat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        dotgat = nn.DotGatConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        dotgat = dotgat.to(ctx)
    
        # test pickle
        th.save(dotgat, tmp_buffer)
    
>       h = dotgat(g, feat)

tests\pytorch\test_nn.py:1003: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\dotgatconv.py:205: in forward
    graph.apply_edges(fn.u_dot_v('ft', 'ft', 'a'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03E265C8>
op = 'dot'
lhs_data = tensor([[[ 4.6382e-01,  8.1709e-01],
         [ 4.9443e-01,  3.6309e-01],
         [-8.6739e-01, -5.9274e-01],
       ...e-01, -6.8233e-01],
         [ 8.4753e-01,  7.3353e-01],
         [-6.8092e-01, -2.1011e-01]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[ 4.6382e-01,  8.1709e-01],
         [ 4.9443e-01,  3.6309e-01],
         [-8.6739e-01, -5.9274e-01],
       ...e-01, -6.8233e-01],
         [ 8.4753e-01,  7.3353e-01],
         [-6.8092e-01, -2.1011e-01]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_dotgat_conv[4-2-g3-idtype0] _______________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={}
      edata_schemes={})
idtype = torch.int32, out_dim = 2, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_dotgat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        dotgat = nn.DotGatConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        dotgat = dotgat.to(ctx)
    
        # test pickle
        th.save(dotgat, tmp_buffer)
    
>       h = dotgat(g, feat)

tests\pytorch\test_nn.py:1003: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\dotgatconv.py:205: in forward
    graph.apply_edges(fn.u_dot_v('ft', 'ft', 'a'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07800B08>
op = 'dot'
lhs_data = tensor([[[-0.4989, -0.9164],
         [-0.5876, -0.9108],
         [-1.1572,  0.7104],
         [-0.5259, -0.4017]],

...1412],
         [ 0.0400, -0.5273],
         [-0.9923,  0.7896],
         [-0.0921, -1.1406]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[-0.4989, -0.9164],
         [-0.5876, -0.9108],
         [-1.1572,  0.7104],
         [-0.5259, -0.4017]],

...1412],
         [ 0.0400, -0.5273],
         [-0.9923,  0.7896],
         [-0.0921, -1.1406]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_dotgat_conv[4-2-g3-idtype1] _______________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={}
      edata_schemes={})
idtype = torch.int64, out_dim = 2, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_dotgat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        dotgat = nn.DotGatConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        dotgat = dotgat.to(ctx)
    
        # test pickle
        th.save(dotgat, tmp_buffer)
    
>       h = dotgat(g, feat)

tests\pytorch\test_nn.py:1003: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\dotgatconv.py:205: in forward
    graph.apply_edges(fn.u_dot_v('ft', 'ft', 'a'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07396A48>
op = 'dot'
lhs_data = tensor([[[-0.1732,  0.1927],
         [ 0.0238,  0.0952],
         [ 0.0835,  0.4412],
         [-0.4363, -0.3036]],

...2506],
         [-0.2434,  0.4468],
         [-0.0999,  0.1084],
         [ 0.5013,  0.6683]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[-0.1732,  0.1927],
         [ 0.0238,  0.0952],
         [ 0.0835,  0.4412],
         [-0.4363, -0.3036]],

...2506],
         [-0.2434,  0.4468],
         [-0.0999,  0.1084],
         [ 0.5013,  0.6683]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_dotgat_conv[4-2-g4-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={}
      edata_schemes={})
idtype = torch.int32, out_dim = 2, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_dotgat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        dotgat = nn.DotGatConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        dotgat = dotgat.to(ctx)
    
        # test pickle
        th.save(dotgat, tmp_buffer)
    
>       h = dotgat(g, feat)

tests\pytorch\test_nn.py:1003: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\dotgatconv.py:205: in forward
    graph.apply_edges(fn.u_dot_v('ft', 'ft', 'a'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE122ACF48>
op = 'dot'
lhs_data = tensor([[[ 0.8579, -0.0380],
         [ 0.6015, -0.5070],
         [-0.0785,  0.5414],
         [-1.0351,  0.0215]],

...0254],
         [-0.0309,  0.5453],
         [-0.2483, -0.4041],
         [ 0.3795, -0.6567]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[ 0.8579, -0.0380],
         [ 0.6015, -0.5070],
         [-0.0785,  0.5414],
         [-1.0351,  0.0215]],

...0254],
         [-0.0309,  0.5453],
         [-0.2483, -0.4041],
         [ 0.3795, -0.6567]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_dotgat_conv[4-2-g4-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={}
      edata_schemes={})
idtype = torch.int64, out_dim = 2, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_dotgat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        dotgat = nn.DotGatConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        dotgat = dotgat.to(ctx)
    
        # test pickle
        th.save(dotgat, tmp_buffer)
    
>       h = dotgat(g, feat)

tests\pytorch\test_nn.py:1003: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\dotgatconv.py:205: in forward
    graph.apply_edges(fn.u_dot_v('ft', 'ft', 'a'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF077AEE88>
op = 'dot'
lhs_data = tensor([[[ 0.0511,  0.0641],
         [-0.3525,  0.5721],
         [-0.1648, -1.1612],
         [-0.0997, -0.0737]],

...6743],
         [ 0.2523, -0.1218],
         [-0.1366, -0.3037],
         [-0.1647,  0.2105]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[ 0.0511,  0.0641],
         [-0.3525,  0.5721],
         [-0.1648, -1.1612],
         [-0.0997, -0.0737]],

...6743],
         [ 0.2523, -0.1218],
         [-0.1366, -0.3037],
         [-0.1647,  0.2105]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_dotgat_conv[4-2-g5-idtype0] _______________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int32
out_dim = 2, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_dotgat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        dotgat = nn.DotGatConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        dotgat = dotgat.to(ctx)
    
        # test pickle
        th.save(dotgat, tmp_buffer)
    
>       h = dotgat(g, feat)

tests\pytorch\test_nn.py:1003: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\dotgatconv.py:205: in forward
    graph.apply_edges(fn.u_dot_v('ft', 'ft', 'a'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0739C248>
op = 'dot'
lhs_data = tensor([[[ 0.8223,  0.1636],
         [-0.1698,  0.9940],
         [ 0.7054,  0.1483],
         [ 0.2376,  0.9035]],

...1099],
         [-0.0374, -0.2017],
         [ 0.0092,  0.2701],
         [ 0.2782,  0.1475]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[ 0.8223,  0.1636],
         [-0.1698,  0.9940],
         [ 0.7054,  0.1483],
         [ 0.2376,  0.9035]],

...781],
         [-0.9772,  1.0206],
         [ 0.4823,  1.7351],
         [-0.4486,  0.4658]]], grad_fn=<SliceBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_dotgat_conv[4-2-g5-idtype1] _______________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int64
out_dim = 2, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_dotgat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        dotgat = nn.DotGatConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        dotgat = dotgat.to(ctx)
    
        # test pickle
        th.save(dotgat, tmp_buffer)
    
>       h = dotgat(g, feat)

tests\pytorch\test_nn.py:1003: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\dotgatconv.py:205: in forward
    graph.apply_edges(fn.u_dot_v('ft', 'ft', 'a'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF073DED08>
op = 'dot'
lhs_data = tensor([[[-0.0454, -0.6071],
         [-0.0934, -0.2434],
         [ 0.7391,  1.0351],
         [-0.3213, -1.2373]],

...0815],
         [ 0.2930, -0.0952],
         [-0.4833, -0.0196],
         [ 0.0548,  0.3162]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[-0.0454, -0.6071],
         [-0.0934, -0.2434],
         [ 0.7391,  1.0351],
         [-0.3213, -1.2373]],

...286],
         [ 0.2600, -0.2596],
         [ 0.3131,  0.3927],
         [ 0.4486, -0.2778]]], grad_fn=<SliceBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_dotgat_conv[4-2-g6-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={}
      edata_schemes={})
idtype = torch.int32, out_dim = 2, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_dotgat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        dotgat = nn.DotGatConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        dotgat = dotgat.to(ctx)
    
        # test pickle
        th.save(dotgat, tmp_buffer)
    
>       h = dotgat(g, feat)

tests\pytorch\test_nn.py:1003: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\dotgatconv.py:205: in forward
    graph.apply_edges(fn.u_dot_v('ft', 'ft', 'a'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03E113C8>
op = 'dot'
lhs_data = tensor([[[-0.1223, -0.0246],
         [-0.2674,  0.6928],
         [-0.0086,  0.2498],
         [-0.3322,  0.2452]],

...2817],
         [-0.0403, -0.7375],
         [ 0.3018,  0.1425],
         [ 0.7630,  0.1669]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[-0.1223, -0.0246],
         [-0.2674,  0.6928],
         [-0.0086,  0.2498],
         [-0.3322,  0.2452]],

...2817],
         [-0.0403, -0.7375],
         [ 0.3018,  0.1425],
         [ 0.7630,  0.1669]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
______________________ test_dotgat_conv[4-2-g6-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={}
      edata_schemes={})
idtype = torch.int64, out_dim = 2, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'block-bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_dotgat_conv(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        dotgat = nn.DotGatConv(5, out_dim, num_heads)
        feat = F.randn((g.number_of_src_nodes(), 5))
        dotgat = dotgat.to(ctx)
    
        # test pickle
        th.save(dotgat, tmp_buffer)
    
>       h = dotgat(g, feat)

tests\pytorch\test_nn.py:1003: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\dotgatconv.py:205: in forward
    graph.apply_edges(fn.u_dot_v('ft', 'ft', 'a'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF077B12C8>
op = 'dot'
lhs_data = tensor([[[ 0.0693,  1.3967],
         [-0.3664,  0.3257],
         [-0.1604, -0.1506],
         [-1.0128,  0.5001]],

...2718],
         [ 0.1667,  0.1350],
         [-0.5472, -0.0505],
         [-0.0309,  0.5017]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[ 0.0693,  1.3967],
         [-0.3664,  0.3257],
         [-0.1604, -0.1506],
         [-1.0128,  0.5001]],

...2718],
         [ 0.1667,  0.1350],
         [-0.5472, -0.0505],
         [-0.0309,  0.5017]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_____________________ test_dotgat_conv_bi[1-1-g0-idtype0] _____________________

g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int32, out_dim = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_dotgat_conv_bi(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        dotgat = nn.DotGatConv((5, 5), out_dim, num_heads)
        feat = (F.randn((g.number_of_src_nodes(), 5)), F.randn((g.number_of_dst_nodes(), 5)))
        dotgat = dotgat.to(ctx)
>       h = dotgat(g, feat)

tests\pytorch\test_nn.py:1018: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\dotgatconv.py:205: in forward
    graph.apply_edges(fn.u_dot_v('ft', 'ft', 'a'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07218848>
op = 'dot'
lhs_data = tensor([[[-0.9694]],

        [[ 0.5359]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[-0.5836]],

        [[ 0.7600]],

        [[-0.4577]],

        [[-0.2557]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_____________________ test_dotgat_conv_bi[1-1-g0-idtype1] _____________________

g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int64, out_dim = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_dotgat_conv_bi(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        dotgat = nn.DotGatConv((5, 5), out_dim, num_heads)
        feat = (F.randn((g.number_of_src_nodes(), 5)), F.randn((g.number_of_dst_nodes(), 5)))
        dotgat = dotgat.to(ctx)
>       h = dotgat(g, feat)

tests\pytorch\test_nn.py:1018: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\dotgatconv.py:205: in forward
    graph.apply_edges(fn.u_dot_v('ft', 'ft', 'a'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF077AE448>
op = 'dot'
lhs_data = tensor([[[-0.8335]],

        [[ 0.1975]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[-0.3493]],

        [[-0.3559]],

        [[ 0.2142]],

        [[-0.5639]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_____________________ test_dotgat_conv_bi[1-1-g1-idtype0] _____________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int32
out_dim = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_dotgat_conv_bi(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        dotgat = nn.DotGatConv((5, 5), out_dim, num_heads)
        feat = (F.randn((g.number_of_src_nodes(), 5)), F.randn((g.number_of_dst_nodes(), 5)))
        dotgat = dotgat.to(ctx)
>       h = dotgat(g, feat)

tests\pytorch\test_nn.py:1018: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\dotgatconv.py:205: in forward
    graph.apply_edges(fn.u_dot_v('ft', 'ft', 'a'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1228E108>
op = 'dot'
lhs_data = tensor([[[-0.5240]],

        [[ 1.3672]],

        [[ 0.8047]],

        [[-0.0724]],

        [[-0.4126]],

        [[-0.6942]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[0.2445]],

        [[0.4472]],

        [[0.9522]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_____________________ test_dotgat_conv_bi[1-1-g1-idtype1] _____________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int64
out_dim = 1, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_dotgat_conv_bi(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        dotgat = nn.DotGatConv((5, 5), out_dim, num_heads)
        feat = (F.randn((g.number_of_src_nodes(), 5)), F.randn((g.number_of_dst_nodes(), 5)))
        dotgat = dotgat.to(ctx)
>       h = dotgat(g, feat)

tests\pytorch\test_nn.py:1018: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\dotgatconv.py:205: in forward
    graph.apply_edges(fn.u_dot_v('ft', 'ft', 'a'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF077A82C8>
op = 'dot'
lhs_data = tensor([[[ 0.5805]],

        [[-0.4286]],

        [[ 0.6778]],

        [[ 0.1553]],

        [[ 0.0926]],

        [[ 0.3704]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[ 0.8152]],

        [[-0.3201]],

        [[-0.6047]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_____________________ test_dotgat_conv_bi[1-2-g0-idtype0] _____________________

g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int32, out_dim = 2, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_dotgat_conv_bi(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        dotgat = nn.DotGatConv((5, 5), out_dim, num_heads)
        feat = (F.randn((g.number_of_src_nodes(), 5)), F.randn((g.number_of_dst_nodes(), 5)))
        dotgat = dotgat.to(ctx)
>       h = dotgat(g, feat)

tests\pytorch\test_nn.py:1018: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\dotgatconv.py:205: in forward
    graph.apply_edges(fn.u_dot_v('ft', 'ft', 'a'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF077D1E08>
op = 'dot'
lhs_data = tensor([[[-0.9265, -0.1147]],

        [[-0.7990,  0.3471]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[-0.6096,  0.4431]],

        [[ 0.4772, -0.1395]],

        [[-0.7384,  0.5005]],

        [[ 0.1076, -0.6642]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_____________________ test_dotgat_conv_bi[1-2-g0-idtype1] _____________________

g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int64, out_dim = 2, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_dotgat_conv_bi(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        dotgat = nn.DotGatConv((5, 5), out_dim, num_heads)
        feat = (F.randn((g.number_of_src_nodes(), 5)), F.randn((g.number_of_dst_nodes(), 5)))
        dotgat = dotgat.to(ctx)
>       h = dotgat(g, feat)

tests\pytorch\test_nn.py:1018: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\dotgatconv.py:205: in forward
    graph.apply_edges(fn.u_dot_v('ft', 'ft', 'a'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE12254488>
op = 'dot'
lhs_data = tensor([[[-0.1724, -0.4161]],

        [[ 0.5570, -0.0535]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[-0.2410,  1.0052]],

        [[ 0.1189,  0.0172]],

        [[ 0.0714, -0.1548]],

        [[-0.7236, -0.0960]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_____________________ test_dotgat_conv_bi[1-2-g1-idtype0] _____________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int32
out_dim = 2, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_dotgat_conv_bi(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        dotgat = nn.DotGatConv((5, 5), out_dim, num_heads)
        feat = (F.randn((g.number_of_src_nodes(), 5)), F.randn((g.number_of_dst_nodes(), 5)))
        dotgat = dotgat.to(ctx)
>       h = dotgat(g, feat)

tests\pytorch\test_nn.py:1018: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\dotgatconv.py:205: in forward
    graph.apply_edges(fn.u_dot_v('ft', 'ft', 'a'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07018088>
op = 'dot'
lhs_data = tensor([[[-0.2222,  0.4907]],

        [[-0.6510,  0.5893]],

        [[-0.1243, -0.0870]],

        [[-1.3795,  1.1179]],

        [[ 1.1190, -0.7951]],

        [[-0.9992,  0.6057]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[-0.0155,  0.0023]],

        [[ 0.2366, -0.0762]],

        [[ 0.1407, -0.3928]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_____________________ test_dotgat_conv_bi[1-2-g1-idtype1] _____________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int64
out_dim = 2, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_dotgat_conv_bi(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        dotgat = nn.DotGatConv((5, 5), out_dim, num_heads)
        feat = (F.randn((g.number_of_src_nodes(), 5)), F.randn((g.number_of_dst_nodes(), 5)))
        dotgat = dotgat.to(ctx)
>       h = dotgat(g, feat)

tests\pytorch\test_nn.py:1018: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\dotgatconv.py:205: in forward
    graph.apply_edges(fn.u_dot_v('ft', 'ft', 'a'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF077C5B08>
op = 'dot'
lhs_data = tensor([[[ 0.4264,  0.1563]],

        [[-0.4641,  1.5523]],

        [[-0.5701, -0.0172]],

        [[-0.6126,  0.7367]],

        [[ 0.1619,  0.2153]],

        [[ 1.2304, -0.9524]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[ 0.0657,  0.1810]],

        [[-0.2508,  0.4212]],

        [[-0.0944,  0.0794]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_____________________ test_dotgat_conv_bi[4-1-g0-idtype0] _____________________

g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int32, out_dim = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_dotgat_conv_bi(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        dotgat = nn.DotGatConv((5, 5), out_dim, num_heads)
        feat = (F.randn((g.number_of_src_nodes(), 5)), F.randn((g.number_of_dst_nodes(), 5)))
        dotgat = dotgat.to(ctx)
>       h = dotgat(g, feat)

tests\pytorch\test_nn.py:1018: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\dotgatconv.py:205: in forward
    graph.apply_edges(fn.u_dot_v('ft', 'ft', 'a'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF073E7808>
op = 'dot'
lhs_data = tensor([[[ 0.4830],
         [-0.1613],
         [ 0.1324],
         [ 0.3634]],

        [[ 0.5974],
         [ 0.1193],
         [-0.0037],
         [-0.1589]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[-0.4929],
         [ 1.2497],
         [ 1.1664],
         [ 0.9875]],

        [[-0.6167],
         [ 0.577... [ 0.5114]],

        [[-1.1181],
         [ 0.7334],
         [ 0.7711],
         [ 0.9342]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_____________________ test_dotgat_conv_bi[4-1-g0-idtype1] _____________________

g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int64, out_dim = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_dotgat_conv_bi(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        dotgat = nn.DotGatConv((5, 5), out_dim, num_heads)
        feat = (F.randn((g.number_of_src_nodes(), 5)), F.randn((g.number_of_dst_nodes(), 5)))
        dotgat = dotgat.to(ctx)
>       h = dotgat(g, feat)

tests\pytorch\test_nn.py:1018: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\dotgatconv.py:205: in forward
    graph.apply_edges(fn.u_dot_v('ft', 'ft', 'a'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE122B8148>
op = 'dot'
lhs_data = tensor([[[0.6500],
         [0.1414],
         [0.1042],
         [0.8920]],

        [[0.3820],
         [0.7157],
         [0.7959],
         [0.3495]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[-0.1814],
         [ 0.2764],
         [-0.3175],
         [-0.2684]],

        [[-0.3241],
         [ 0.849... [-0.2943]],

        [[-0.0895],
         [-0.3085],
         [ 0.3221],
         [ 0.0168]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_____________________ test_dotgat_conv_bi[4-1-g1-idtype0] _____________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int32
out_dim = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_dotgat_conv_bi(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        dotgat = nn.DotGatConv((5, 5), out_dim, num_heads)
        feat = (F.randn((g.number_of_src_nodes(), 5)), F.randn((g.number_of_dst_nodes(), 5)))
        dotgat = dotgat.to(ctx)
>       h = dotgat(g, feat)

tests\pytorch\test_nn.py:1018: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\dotgatconv.py:205: in forward
    graph.apply_edges(fn.u_dot_v('ft', 'ft', 'a'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF073E7C88>
op = 'dot'
lhs_data = tensor([[[-0.4656],
         [-0.6085],
         [-0.0667],
         [-0.5637]],

        [[-0.0830],
         [-0.151... [ 0.6698]],

        [[-0.3442],
         [ 0.2146],
         [ 0.7411],
         [-0.5289]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[ 0.4106],
         [-0.9586],
         [-1.4979],
         [-0.7013]],

        [[ 0.7881],
         [-0.826... [-0.1119]],

        [[-0.8860],
         [ 0.8015],
         [-0.0748],
         [ 0.3975]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_____________________ test_dotgat_conv_bi[4-1-g1-idtype1] _____________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int64
out_dim = 1, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_dotgat_conv_bi(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        dotgat = nn.DotGatConv((5, 5), out_dim, num_heads)
        feat = (F.randn((g.number_of_src_nodes(), 5)), F.randn((g.number_of_dst_nodes(), 5)))
        dotgat = dotgat.to(ctx)
>       h = dotgat(g, feat)

tests\pytorch\test_nn.py:1018: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\dotgatconv.py:205: in forward
    graph.apply_edges(fn.u_dot_v('ft', 'ft', 'a'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF077B8B08>
op = 'dot'
lhs_data = tensor([[[ 0.0849],
         [ 0.3540],
         [ 0.0689],
         [ 0.0529]],

        [[-0.3363],
         [-0.627... [ 0.4526]],

        [[-0.6140],
         [ 0.5132],
         [ 1.0290],
         [ 0.6017]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[ 0.8828],
         [-0.1587],
         [ 0.7704],
         [-0.5350]],

        [[ 0.1671],
         [ 0.111... [-0.5320]],

        [[ 0.1875],
         [ 0.0708],
         [-0.0523],
         [ 0.1025]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_____________________ test_dotgat_conv_bi[4-2-g0-idtype0] _____________________

g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int32, out_dim = 2, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_dotgat_conv_bi(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        dotgat = nn.DotGatConv((5, 5), out_dim, num_heads)
        feat = (F.randn((g.number_of_src_nodes(), 5)), F.randn((g.number_of_dst_nodes(), 5)))
        dotgat = dotgat.to(ctx)
>       h = dotgat(g, feat)

tests\pytorch\test_nn.py:1018: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\dotgatconv.py:205: in forward
    graph.apply_edges(fn.u_dot_v('ft', 'ft', 'a'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF019ADEC8>
op = 'dot'
lhs_data = tensor([[[ 0.0498,  0.5550],
         [ 0.2347,  0.6748],
         [ 0.2003,  0.4885],
         [ 0.0236,  0.2013]],

...2404],
         [-0.6822, -0.3418],
         [-0.4372,  0.5612],
         [ 0.5821, -0.4192]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[-0.3013,  0.6716],
         [-0.8220,  0.9957],
         [-1.1052, -0.6140],
         [-0.6755, -0.5250]],

...3560],
         [-0.2628, -0.8790],
         [ 0.3457,  0.1349],
         [-0.1346,  0.4277]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_____________________ test_dotgat_conv_bi[4-2-g0-idtype1] _____________________

g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int64, out_dim = 2, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_dotgat_conv_bi(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        dotgat = nn.DotGatConv((5, 5), out_dim, num_heads)
        feat = (F.randn((g.number_of_src_nodes(), 5)), F.randn((g.number_of_dst_nodes(), 5)))
        dotgat = dotgat.to(ctx)
>       h = dotgat(g, feat)

tests\pytorch\test_nn.py:1018: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\dotgatconv.py:205: in forward
    graph.apply_edges(fn.u_dot_v('ft', 'ft', 'a'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07309B88>
op = 'dot'
lhs_data = tensor([[[ 0.2981,  0.2188],
         [-0.2406, -0.8796],
         [ 0.2949, -0.0015],
         [-0.3675, -0.2658]],

...4028],
         [ 0.3603,  1.1409],
         [-1.1468, -0.0301],
         [ 0.8810, -0.4608]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[ 0.2881, -0.2651],
         [ 0.2372, -0.0791],
         [ 0.1272,  0.1410],
         [-0.1721, -0.0077]],

...1961],
         [-0.1210,  0.2333],
         [-0.4226,  0.1699],
         [-0.7182,  0.5729]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_____________________ test_dotgat_conv_bi[4-2-g1-idtype0] _____________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int32
out_dim = 2, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_dotgat_conv_bi(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        dotgat = nn.DotGatConv((5, 5), out_dim, num_heads)
        feat = (F.randn((g.number_of_src_nodes(), 5)), F.randn((g.number_of_dst_nodes(), 5)))
        dotgat = dotgat.to(ctx)
>       h = dotgat(g, feat)

tests\pytorch\test_nn.py:1018: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\dotgatconv.py:205: in forward
    graph.apply_edges(fn.u_dot_v('ft', 'ft', 'a'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF074271C8>
op = 'dot'
lhs_data = tensor([[[-0.8512, -0.3414],
         [ 1.5003, -0.6760],
         [-0.4036,  0.1045],
         [ 0.3344, -1.4628]],

...1348],
         [ 1.0049, -0.2356],
         [-0.5077,  0.2977],
         [ 0.4937, -1.0519]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[ 0.6201, -0.2883],
         [-0.3675, -0.0990],
         [ 0.1445, -0.1637],
         [-0.1185,  0.0800]],

...3691],
         [-0.5112, -0.2334],
         [ 0.2504, -0.2135],
         [-0.2540,  0.1269]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_____________________ test_dotgat_conv_bi[4-2-g1-idtype1] _____________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int64
out_dim = 2, num_heads = 4

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    @pytest.mark.parametrize('num_heads', [1, 4])
    def test_dotgat_conv_bi(g, idtype, out_dim, num_heads):
        g = g.astype(idtype).to(F.ctx())
        ctx = F.ctx()
        dotgat = nn.DotGatConv((5, 5), out_dim, num_heads)
        feat = (F.randn((g.number_of_src_nodes(), 5)), F.randn((g.number_of_dst_nodes(), 5)))
        dotgat = dotgat.to(ctx)
>       h = dotgat(g, feat)

tests\pytorch\test_nn.py:1018: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\dotgatconv.py:205: in forward
    graph.apply_edges(fn.u_dot_v('ft', 'ft', 'a'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF077D0988>
op = 'dot'
lhs_data = tensor([[[-0.3785,  0.4268],
         [ 0.1302,  0.3379],
         [ 0.1440, -0.0128],
         [-0.1614, -0.0936]],

...9188],
         [-0.4152, -0.5345],
         [ 0.0572,  0.5560],
         [-0.8934,  0.4026]]], grad_fn=<ViewBackward>)
rhs_data = tensor([[[-0.6983,  0.6638],
         [-1.0231, -0.5278],
         [-0.7298,  0.1425],
         [-0.4834, -0.4650]],

...0780],
         [-0.1142, -0.2039],
         [ 0.0024,  0.4723],
         [ 0.2392, -0.2105]]], grad_fn=<ViewBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
___________________________ test_dense_cheb_conv[1] ___________________________

out_dim = 1

    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_cheb_conv(out_dim):
        for k in range(1, 4):
            ctx = F.ctx()
            g = dgl.DGLGraph(sp.sparse.random(100, 100, density=0.1), readonly=True)
            g = g.to(F.ctx())
            adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
            cheb = nn.ChebConv(5, out_dim, k, None)
            dense_cheb = nn.DenseChebConv(5, out_dim, k)
            #for i in range(len(cheb.fc)):
            #    dense_cheb.W.data[i] = cheb.fc[i].weight.data.t()
            dense_cheb.W.data = cheb.linear.weight.data.transpose(-1, -2).view(k, 5, out_dim)
            if cheb.linear.bias is not None:
                dense_cheb.bias.data = cheb.linear.bias.data
            feat = F.randn((100, 5))
            cheb = cheb.to(ctx)
            dense_cheb = dense_cheb.to(ctx)
>           out_cheb = cheb(g, feat, [2.0])

tests\pytorch\test_nn.py:1040: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\chebconv.py:130: in forward
    h = unnLaplacian(X_0, D_invsqrt, graph)
python\dgl\nn\pytorch\conv\chebconv.py:103: in unnLaplacian
    graph.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF077D21C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.5630,  0.4032,  0.3565, -0.0388, -0.0166],
        [ 0.0446, -0.0604,  0.1367, -0.0020, -0.0208],
        ...0.3573],
        [-0.1222, -0.3434, -0.2911,  0.1425, -0.6669],
        [-0.0159, -0.0764, -0.2173, -0.0154,  0.0497]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
---------------------------- Captured stdout call -----------------------------
1 tensor([[-0.1483],
        [ 0.5646],
        [-0.1799],
        [ 0.4515],
        [ 0.4828],
        [ 0.1645],
        [-0.2216],
        [ 0.2288],
        [-0.0157],
        [-0.3422],
        [ 0.2072],
        [ 0.3209],
        [ 0.3846],
        [ 0.2681],
        [-0.0924],
        [-0.2921],
        [ 0.4173],
        [ 0.4343],
        [ 0.1131],
        [-0.0686],
        [ 0.4608],
        [ 0.3049],
        [ 0.2280],
        [ 0.5373],
        [ 0.2868],
        [ 0.3676],
        [ 0.3435],
        [ 0.1384],
        [ 0.3128],
        [-0.0893],
        [ 0.3390],
        [-0.1982],
        [ 0.6095],
        [-0.4951],
        [ 0.3317],
        [-0.1628],
        [-0.0403],
        [ 0.4532],
        [-0.1243],
        [ 0.0797],
        [ 0.4808],
        [-0.0392],
        [ 0.0935],
        [ 0.1351],
        [ 0.0286],
        [ 0.0858],
        [ 0.2192],
        [ 0.4419],
        [ 0.3408],
        [ 0.5475],
        [ 0.0143],
        [ 0.1049],
        [ 0.5359],
        [ 0.1154],
        [ 0.2162],
        [-0.0771],
        [ 0.5497],
        [ 0.1472],
        [ 0.4032],
        [ 0.6115],
        [ 0.1225],
        [-0.2362],
        [-0.1780],
        [-0.0076],
        [ 0.0823],
        [-0.1912],
        [ 0.1414],
        [-0.1802],
        [ 0.2309],
        [ 0.5352],
        [ 0.3177],
        [-0.1316],
        [ 0.3189],
        [ 0.1216],
        [ 0.2709],
        [-0.1656],
        [ 0.0413],
        [ 0.0174],
        [ 0.3980],
        [ 0.5580],
        [ 0.3642],
        [ 0.1090],
        [ 0.2172],
        [ 0.2753],
        [ 0.6316],
        [-0.3075],
        [ 0.4241],
        [ 0.4387],
        [ 0.2734],
        [ 0.4583],
        [-0.0487],
        [ 0.5076],
        [ 0.3659],
        [ 0.5654],
        [-0.3401],
        [ 0.3943],
        [-0.0047],
        [ 0.4762],
        [-0.0104],
        [-0.0068]], grad_fn=<AddmmBackward>) tensor([[-0.1483],
        [ 0.5646],
        [-0.1799],
        [ 0.4515],
        [ 0.4828],
        [ 0.1645],
        [-0.2216],
        [ 0.2288],
        [-0.0157],
        [-0.3422],
        [ 0.2072],
        [ 0.3209],
        [ 0.3846],
        [ 0.2681],
        [-0.0924],
        [-0.2921],
        [ 0.4173],
        [ 0.4343],
        [ 0.1131],
        [-0.0686],
        [ 0.4608],
        [ 0.3049],
        [ 0.2280],
        [ 0.5373],
        [ 0.2868],
        [ 0.3676],
        [ 0.3435],
        [ 0.1384],
        [ 0.3128],
        [-0.0893],
        [ 0.3390],
        [-0.1982],
        [ 0.6095],
        [-0.4951],
        [ 0.3317],
        [-0.1628],
        [-0.0403],
        [ 0.4532],
        [-0.1243],
        [ 0.0797],
        [ 0.4808],
        [-0.0392],
        [ 0.0935],
        [ 0.1351],
        [ 0.0286],
        [ 0.0858],
        [ 0.2192],
        [ 0.4419],
        [ 0.3408],
        [ 0.5475],
        [ 0.0143],
        [ 0.1049],
        [ 0.5359],
        [ 0.1154],
        [ 0.2162],
        [-0.0771],
        [ 0.5497],
        [ 0.1472],
        [ 0.4032],
        [ 0.6115],
        [ 0.1225],
        [-0.2362],
        [-0.1780],
        [-0.0076],
        [ 0.0823],
        [-0.1912],
        [ 0.1414],
        [-0.1802],
        [ 0.2309],
        [ 0.5352],
        [ 0.3177],
        [-0.1316],
        [ 0.3189],
        [ 0.1216],
        [ 0.2709],
        [-0.1656],
        [ 0.0413],
        [ 0.0174],
        [ 0.3980],
        [ 0.5580],
        [ 0.3642],
        [ 0.1090],
        [ 0.2172],
        [ 0.2753],
        [ 0.6316],
        [-0.3075],
        [ 0.4241],
        [ 0.4387],
        [ 0.2734],
        [ 0.4583],
        [-0.0487],
        [ 0.5076],
        [ 0.3659],
        [ 0.5654],
        [-0.3401],
        [ 0.3943],
        [-0.0047],
        [ 0.4762],
        [-0.0104],
        [-0.0068]], grad_fn=<AddBackward0>)
___________________________ test_dense_cheb_conv[2] ___________________________

out_dim = 2

    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_dense_cheb_conv(out_dim):
        for k in range(1, 4):
            ctx = F.ctx()
            g = dgl.DGLGraph(sp.sparse.random(100, 100, density=0.1), readonly=True)
            g = g.to(F.ctx())
            adj = g.adjacency_matrix(transpose=True, ctx=ctx).to_dense()
            cheb = nn.ChebConv(5, out_dim, k, None)
            dense_cheb = nn.DenseChebConv(5, out_dim, k)
            #for i in range(len(cheb.fc)):
            #    dense_cheb.W.data[i] = cheb.fc[i].weight.data.t()
            dense_cheb.W.data = cheb.linear.weight.data.transpose(-1, -2).view(k, 5, out_dim)
            if cheb.linear.bias is not None:
                dense_cheb.bias.data = cheb.linear.bias.data
            feat = F.randn((100, 5))
            cheb = cheb.to(ctx)
            dense_cheb = dense_cheb.to(ctx)
>           out_cheb = cheb(g, feat, [2.0])

tests\pytorch\test_nn.py:1040: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\chebconv.py:130: in forward
    h = unnLaplacian(X_0, D_invsqrt, graph)
python\dgl\nn\pytorch\conv\chebconv.py:103: in unnLaplacian
    graph.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE123C5CC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-2.2905e-01,  6.4358e-02,  2.5134e-01, -1.3856e-01, -1.7534e-01],
        [-3.5620e-02,  3.4000e-01, -2.6566e...01, -2.8050e-02,  4.4176e-02,  2.9358e-01],
        [-2.9128e-01, -3.2529e-01, -3.4963e-01,  7.1537e-01,  7.6390e-01]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
---------------------------- Captured stdout call -----------------------------
1 tensor([[-0.2465, -0.6337],
        [ 0.7033,  0.0666],
        [ 0.1377, -0.7721],
        [-0.0125, -0.6798],
        [-0.0136, -0.8350],
        [ 0.5796,  0.6321],
        [ 0.6563, -0.1389],
        [ 0.8157,  0.5572],
        [ 0.5124, -0.7402],
        [ 0.0373, -0.4913],
        [ 0.0628, -0.2320],
        [-0.1307, -0.1633],
        [ 0.3715, -0.1237],
        [-0.2811, -0.9184],
        [ 0.5474, -0.2947],
        [ 0.1764,  0.4816],
        [ 0.9328,  0.7598],
        [ 0.8769,  0.6227],
        [ 0.1643, -0.1913],
        [-0.1127,  0.0972],
        [ 0.4026, -0.2674],
        [ 0.6276,  0.1105],
        [ 0.5037, -0.2416],
        [ 0.4262,  0.6545],
        [ 0.8236,  0.3003],
        [ 0.1003, -0.4666],
        [ 0.2677, -0.3673],
        [ 0.5369,  0.2568],
        [ 0.6343,  0.3808],
        [ 0.3723, -0.5850],
        [ 0.7706,  0.4047],
        [ 0.8998,  1.4726],
        [ 0.3942,  0.2480],
        [ 0.7651,  0.2438],
        [-0.1149, -0.6694],
        [ 0.1618, -0.1905],
        [ 0.0557, -0.6954],
        [ 0.1738,  0.3303],
        [ 0.4794,  1.0118],
        [ 0.4832,  0.3378],
        [ 0.3867,  0.3986],
        [-0.1022, -1.6046],
        [ 0.4078,  0.2784],
        [ 0.0392, -0.8187],
        [ 0.1138, -0.7104],
        [-0.7036, -1.4081],
        [ 0.6625,  1.5324],
        [ 0.8624,  1.3113],
        [ 0.3953, -0.5044],
        [ 0.6495,  0.1119],
        [ 0.6778,  0.5930],
        [ 0.6782,  0.2112],
        [ 0.3920,  1.0177],
        [ 0.0276, -0.6887],
        [ 0.6563,  0.9138],
        [ 0.4774,  0.0135],
        [ 0.0406, -0.3739],
        [ 0.3774,  0.9451],
        [ 1.2535,  1.4316],
        [ 0.8844,  0.0901],
        [ 0.3512,  0.3770],
        [ 0.3120, -0.2171],
        [-0.3557, -0.8810],
        [ 0.4348, -0.9008],
        [ 0.3494,  0.7059],
        [ 1.3247,  1.3000],
        [ 0.5807, -0.3567],
        [ 0.6728,  0.5278],
        [ 0.5968,  0.4298],
        [ 0.5961, -0.0569],
        [ 0.8899,  1.0007],
        [-0.1936, -0.6426],
        [-0.0083,  0.0156],
        [ 1.1627,  0.5378],
        [ 0.8846,  1.5960],
        [ 0.4053, -0.2602],
        [-0.0170,  0.3704],
        [ 0.3956,  0.8193],
        [ 0.6795,  1.0290],
        [ 1.1320,  0.9030],
        [ 0.4816,  0.0789],
        [ 0.3430,  0.3855],
        [ 0.5907,  1.1944],
        [ 0.8586,  0.5325],
        [ 0.1185,  0.7967],
        [ 0.5796, -0.4615],
        [ 0.5472,  0.4449],
        [ 0.6968,  0.3300],
        [ 0.2125,  0.5799],
        [ 1.3935,  1.1025],
        [-0.2531, -0.2949],
        [ 1.4429,  2.4493],
        [-0.0854, -0.4195],
        [ 1.0228,  0.9628],
        [ 0.2092, -0.4127],
        [ 0.0841,  0.3869],
        [ 0.7690,  0.5580],
        [ 0.4555, -0.5889],
        [ 0.4778,  0.1505],
        [ 0.1846, -0.1337]], grad_fn=<AddmmBackward>) tensor([[-0.2465, -0.6337],
        [ 0.7033,  0.0666],
        [ 0.1377, -0.7721],
        [-0.0125, -0.6798],
        [-0.0136, -0.8350],
        [ 0.5796,  0.6321],
        [ 0.6563, -0.1389],
        [ 0.8157,  0.5572],
        [ 0.5124, -0.7402],
        [ 0.0373, -0.4913],
        [ 0.0628, -0.2320],
        [-0.1307, -0.1633],
        [ 0.3715, -0.1237],
        [-0.2811, -0.9184],
        [ 0.5474, -0.2947],
        [ 0.1764,  0.4816],
        [ 0.9328,  0.7598],
        [ 0.8769,  0.6227],
        [ 0.1643, -0.1913],
        [-0.1127,  0.0972],
        [ 0.4026, -0.2674],
        [ 0.6276,  0.1105],
        [ 0.5037, -0.2416],
        [ 0.4262,  0.6545],
        [ 0.8236,  0.3003],
        [ 0.1003, -0.4666],
        [ 0.2677, -0.3673],
        [ 0.5369,  0.2568],
        [ 0.6343,  0.3808],
        [ 0.3723, -0.5850],
        [ 0.7706,  0.4047],
        [ 0.8998,  1.4726],
        [ 0.3942,  0.2480],
        [ 0.7651,  0.2438],
        [-0.1149, -0.6694],
        [ 0.1618, -0.1905],
        [ 0.0557, -0.6954],
        [ 0.1738,  0.3303],
        [ 0.4794,  1.0118],
        [ 0.4832,  0.3378],
        [ 0.3867,  0.3986],
        [-0.1022, -1.6046],
        [ 0.4078,  0.2784],
        [ 0.0392, -0.8187],
        [ 0.1138, -0.7104],
        [-0.7036, -1.4081],
        [ 0.6625,  1.5324],
        [ 0.8624,  1.3113],
        [ 0.3953, -0.5044],
        [ 0.6495,  0.1119],
        [ 0.6778,  0.5930],
        [ 0.6782,  0.2112],
        [ 0.3920,  1.0177],
        [ 0.0276, -0.6887],
        [ 0.6563,  0.9138],
        [ 0.4774,  0.0135],
        [ 0.0406, -0.3739],
        [ 0.3774,  0.9451],
        [ 1.2535,  1.4316],
        [ 0.8844,  0.0901],
        [ 0.3512,  0.3770],
        [ 0.3120, -0.2171],
        [-0.3557, -0.8810],
        [ 0.4348, -0.9008],
        [ 0.3494,  0.7059],
        [ 1.3247,  1.3000],
        [ 0.5807, -0.3567],
        [ 0.6728,  0.5278],
        [ 0.5968,  0.4298],
        [ 0.5961, -0.0569],
        [ 0.8899,  1.0007],
        [-0.1936, -0.6426],
        [-0.0083,  0.0156],
        [ 1.1627,  0.5378],
        [ 0.8846,  1.5960],
        [ 0.4053, -0.2602],
        [-0.0170,  0.3704],
        [ 0.3956,  0.8193],
        [ 0.6795,  1.0290],
        [ 1.1320,  0.9030],
        [ 0.4816,  0.0789],
        [ 0.3430,  0.3855],
        [ 0.5907,  1.1944],
        [ 0.8586,  0.5325],
        [ 0.1185,  0.7967],
        [ 0.5796, -0.4615],
        [ 0.5472,  0.4449],
        [ 0.6968,  0.3300],
        [ 0.2125,  0.5799],
        [ 1.3935,  1.1025],
        [-0.2531, -0.2949],
        [ 1.4429,  2.4493],
        [-0.0854, -0.4195],
        [ 1.0228,  0.9628],
        [ 0.2092, -0.4127],
        [ 0.0841,  0.3869],
        [ 0.7690,  0.5580],
        [ 0.4555, -0.5889],
        [ 0.4778,  0.1505],
        [ 0.1846, -0.1337]], grad_fn=<AddBackward0>)
_______________________________ test_sequential _______________________________

    def test_sequential():
        ctx = F.ctx()
        # Test single graph
        class ExampleLayer(th.nn.Module):
            def __init__(self):
                super().__init__()
    
            def forward(self, graph, n_feat, e_feat):
                graph = graph.local_var()
                graph.ndata['h'] = n_feat
                graph.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
                n_feat += graph.ndata['h']
                graph.apply_edges(fn.u_add_v('h', 'h', 'e'))
                e_feat += graph.edata['e']
                return n_feat, e_feat
    
        g = dgl.DGLGraph()
        g.add_nodes(3)
        g.add_edges([0, 1, 2, 0, 1, 2, 0, 1, 2], [0, 0, 0, 1, 1, 1, 2, 2, 2])
        g = g.to(F.ctx())
        net = nn.Sequential(ExampleLayer(), ExampleLayer(), ExampleLayer())
        n_feat = F.randn((3, 4))
        e_feat = F.randn((9, 4))
        net = net.to(ctx)
>       n_feat, e_feat = net(g, n_feat, e_feat)

tests\pytorch\test_nn.py:1069: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:221: in forward
    feats = module(graph, *feats)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
tests\pytorch\test_nn.py:1055: in forward
    graph.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF072A8108>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.7821,  1.0798,  0.4059,  0.4878],
        [ 0.2633, -0.6495,  0.9429,  0.3172],
        [ 1.2196, -1.3653,  0.2878,  0.1060]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_cf_conv[1-g0-idtype0] __________________________

g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int32, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 3])
    def test_cf_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        cfconv = nn.CFConv(node_in_feats=2,
                           edge_in_feats=3,
                           hidden_feats=2,
                           out_feats=out_dim)
    
        ctx = F.ctx()
        if F.gpu_ctx():
            cfconv = cfconv.to(ctx)
    
        src_feats = F.randn((g.number_of_src_nodes(), 2))
        edge_feats = F.randn((g.number_of_edges(), 3))
>       h = cfconv(g, src_feats, edge_feats)

tests\pytorch\test_nn.py:1131: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\cfconv.py:141: in forward
    g.update_all(fn.u_mul_e('hv', 'he', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01993CC8>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[-2.1406,  0.1408],
        [-0.3480, -0.1269]], grad_fn=<AddmmBackward>)
rhs_data = tensor([[-0.0792,  0.0550],
        [-0.0116,  0.3700],
        [-0.1034, -0.0047],
        [ 0.0517,  0.4648],
      ...  0.0475],
        [-0.0846,  0.1412],
        [-0.1397,  0.0826],
        [-0.0326,  0.2453]], grad_fn=<SubBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_cf_conv[1-g0-idtype1] __________________________

g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int64, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 3])
    def test_cf_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        cfconv = nn.CFConv(node_in_feats=2,
                           edge_in_feats=3,
                           hidden_feats=2,
                           out_feats=out_dim)
    
        ctx = F.ctx()
        if F.gpu_ctx():
            cfconv = cfconv.to(ctx)
    
        src_feats = F.randn((g.number_of_src_nodes(), 2))
        edge_feats = F.randn((g.number_of_edges(), 3))
>       h = cfconv(g, src_feats, edge_feats)

tests\pytorch\test_nn.py:1131: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\cfconv.py:141: in forward
    g.update_all(fn.u_mul_e('hv', 'he', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE12352F08>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[ 1.1229, -0.9106],
        [ 0.3747, -0.9606]], grad_fn=<AddmmBackward>)
rhs_data = tensor([[ 0.0743, -0.0270],
        [ 0.1241,  0.0067],
        [-0.1070, -0.4083],
        [ 0.2114,  0.0599],
      ... -0.1385],
        [ 0.0028, -0.1408],
        [-0.1319, -0.2701],
        [-0.1204, -0.4106]], grad_fn=<SubBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_cf_conv[1-g1-idtype0] __________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'hv': Scheme(shap...32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32), 'he': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int32, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 3])
    def test_cf_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        cfconv = nn.CFConv(node_in_feats=2,
                           edge_in_feats=3,
                           hidden_feats=2,
                           out_feats=out_dim)
    
        ctx = F.ctx()
        if F.gpu_ctx():
            cfconv = cfconv.to(ctx)
    
        src_feats = F.randn((g.number_of_src_nodes(), 2))
        edge_feats = F.randn((g.number_of_edges(), 3))
>       h = cfconv(g, src_feats, edge_feats)

tests\pytorch\test_nn.py:1131: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\cfconv.py:141: in forward
    g.update_all(fn.u_mul_e('hv', 'he', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF070E62C8>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[-1.3516,  0.1996],
        [-0.6192,  0.4839],
        [-0.1795,  0.6388],
        [-0.3852,  0.5561],
      ... 0.4915],
        [-0.2954,  0.6594],
        [-0.4244,  0.5406],
        [-0.5093,  0.6077]], grad_fn=<AddmmBackward>)
rhs_data = tensor([[0.1160, 0.5277],
        [0.2252, 0.5788],
        [0.0765, 0.2678],
        [0.1610, 0.5306],
        [0.122...0.0812, 0.2122],
        [0.1763, 0.4388],
        [0.2475, 0.6353],
        [0.1036, 0.3019]], grad_fn=<SubBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_cf_conv[1-g1-idtype1] __________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'hv': Scheme(shap...32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32), 'he': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int64, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 3])
    def test_cf_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        cfconv = nn.CFConv(node_in_feats=2,
                           edge_in_feats=3,
                           hidden_feats=2,
                           out_feats=out_dim)
    
        ctx = F.ctx()
        if F.gpu_ctx():
            cfconv = cfconv.to(ctx)
    
        src_feats = F.randn((g.number_of_src_nodes(), 2))
        edge_feats = F.randn((g.number_of_edges(), 3))
>       h = cfconv(g, src_feats, edge_feats)

tests\pytorch\test_nn.py:1131: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\cfconv.py:141: in forward
    g.update_all(fn.u_mul_e('hv', 'he', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE12225348>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[ 0.4853, -0.4872],
        [-0.5734, -0.5029],
        [-0.7382, -0.4998],
        [-0.8256, -0.4660],
      ...-0.4933],
        [ 1.7662, -0.5391],
        [-0.9096, -0.4787],
        [-0.0372, -0.5028]], grad_fn=<AddmmBackward>)
rhs_data = tensor([[ 0.1204,  0.1568],
        [ 0.0907, -0.1325],
        [ 0.1201,  0.1643],
        [ 0.0878, -0.1580],
      ... -0.0334],
        [ 0.1355,  0.3316],
        [ 0.1002, -0.0463],
        [ 0.0871, -0.1614]], grad_fn=<SubBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_cf_conv[1-g2-idtype0] __________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'hv': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'he': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int32, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 3])
    def test_cf_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        cfconv = nn.CFConv(node_in_feats=2,
                           edge_in_feats=3,
                           hidden_feats=2,
                           out_feats=out_dim)
    
        ctx = F.ctx()
        if F.gpu_ctx():
            cfconv = cfconv.to(ctx)
    
        src_feats = F.randn((g.number_of_src_nodes(), 2))
        edge_feats = F.randn((g.number_of_edges(), 3))
>       h = cfconv(g, src_feats, edge_feats)

tests\pytorch\test_nn.py:1131: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\cfconv.py:141: in forward
    g.update_all(fn.u_mul_e('hv', 'he', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF073AA948>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[-0.4129,  0.2770],
        [-0.3130,  1.1279],
        [-0.3422,  1.3653],
        [-0.1728,  0.5021],
      ... 1.7256],
        [-0.2666,  2.1003],
        [ 0.1972, -0.0280],
        [-0.2829, -0.8954]], grad_fn=<AddmmBackward>)
rhs_data = tensor([[-0.2630, -0.2328],
        [-0.3101, -0.2241],
        [-0.3978, -0.2103],
        [-0.4193, -0.2163],
      ... -0.2444],
        [-0.4194, -0.2164],
        [-0.2891, -0.2358],
        [-0.3747, -0.2184]], grad_fn=<SubBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):Submodule path 'third_party/googletest': checked out 'f71fb4f9a912ec945401cc49a287a759b6131026'

E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_cf_conv[1-g2-idtype1] __________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'hv': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'he': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int64, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 3])
    def test_cf_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        cfconv = nn.CFConv(node_in_feats=2,
                           edge_in_feats=3,
                           hidden_feats=2,
                           out_feats=out_dim)
    
        ctx = F.ctx()
        if F.gpu_ctx():
            cfconv = cfconv.to(ctx)
    
        src_feats = F.randn((g.number_of_src_nodes(), 2))
        edge_feats = F.randn((g.number_of_edges(), 3))
>       h = cfconv(g, src_feats, edge_feats)

tests\pytorch\test_nn.py:1131: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\cfconv.py:141: in forward
    g.update_all(fn.u_mul_e('hv', 'he', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF077F42C8>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[ 0.5189,  0.3301],
        [-0.3752, -1.0341],
        [ 0.1102, -1.3705],
        [ 1.4589,  0.3245],
      ...-0.2583],
        [ 0.4323,  0.0608],
        [ 1.1519,  0.1163],
        [ 1.0316, -0.9775]], grad_fn=<AddmmBackward>)
rhs_data = tensor([[-0.1110, -0.1610],
        [ 0.0553, -0.2688],
        [-0.1121, -0.1618],
        [-0.2924, -0.0733],
      ... -0.2423],
        [-0.2189, -0.1163],
        [-0.0722, -0.1944],
        [-0.2043, -0.1133]], grad_fn=<SubBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_cf_conv[1-g3-idtype0] __________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'hv': Scheme(shap...      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), 'he': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int32, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 3])
    def test_cf_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        cfconv = nn.CFConv(node_in_feats=2,
                           edge_in_feats=3,
                           hidden_feats=2,
                           out_feats=out_dim)
    
        ctx = F.ctx()
        if F.gpu_ctx():
            cfconv = cfconv.to(ctx)
    
        src_feats = F.randn((g.number_of_src_nodes(), 2))
        edge_feats = F.randn((g.number_of_edges(), 3))
>       h = cfconv(g, src_feats, edge_feats)

tests\pytorch\test_nn.py:1131: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\cfconv.py:141: in forward
    g.update_all(fn.u_mul_e('hv', 'he', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF070FEC88>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[ 0.1345, -0.1930],
        [ 0.4006, -0.3208],
        [-0.5630,  0.1903],
        [-0.3892,  0.1667],
      ... 0.2174],
        [ 0.3060, -0.8140],
        [ 0.0836, -0.1974],
        [-0.0428,  0.1000]], grad_fn=<AddmmBackward>)
rhs_data = tensor([[ 0.3497, -0.0621],
        [ 0.3217, -0.0380],
        [ 0.2832, -0.1647],
        [ 0.2790, -0.0421],
      ... -0.1822],
        [ 0.2879, -0.0920],
        [ 0.2708,  0.0134],
        [ 0.3551, -0.1914]], grad_fn=<SubBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_cf_conv[1-g3-idtype1] __________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'hv': Scheme(shap...      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), 'he': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int64, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 3])
    def test_cf_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        cfconv = nn.CFConv(node_in_feats=2,
                           edge_in_feats=3,
                           hidden_feats=2,
                           out_feats=out_dim)
    
        ctx = F.ctx()
        if F.gpu_ctx():
            cfconv = cfconv.to(ctx)
    
        src_feats = F.randn((g.number_of_src_nodes(), 2))
        edge_feats = F.randn((g.number_of_edges(), 3))
>       h = cfconv(g, src_feats, edge_feats)

tests\pytorch\test_nn.py:1131: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\cfconv.py:141: in forward
    g.update_all(fn.u_mul_e('hv', 'he', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07172548>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[ 0.1889, -0.2599],
        [ 0.0762, -0.3780],
        [-0.0091, -1.2315],
        [ 0.2648, -0.9686],
      ...-0.9417],
        [-0.2435, -0.5967],
        [-0.3035, -0.1244],
        [-0.3837, -0.3527]], grad_fn=<AddmmBackward>)
rhs_data = tensor([[-0.1891, -0.2623],
        [-0.2137, -0.2967],
        [-0.1221, -0.2329],
        [-0.0936, -0.1789],
      ... -0.2179],
        [-0.0493, -0.1008],
        [-0.0831, -0.1617],
        [-0.1045, -0.1982]], grad_fn=<SubBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_cf_conv[1-g4-idtype0] __________________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'hv': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'he': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int32, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 3])
    def test_cf_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        cfconv = nn.CFConv(node_in_feats=2,
                           edge_in_feats=3,
                           hidden_feats=2,
                           out_feats=out_dim)
    
        ctx = F.ctx()
        if F.gpu_ctx():
            cfconv = cfconv.to(ctx)
    
        src_feats = F.randn((g.number_of_src_nodes(), 2))
        edge_feats = F.randn((g.number_of_edges(), 3))
>       h = cfconv(g, src_feats, edge_feats)

tests\pytorch\test_nn.py:1131: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\cfconv.py:141: in forward
    g.update_all(fn.u_mul_e('hv', 'he', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE123005C8>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[ 6.4941e-01,  2.6193e-01],
        [ 1.7130e-01, -1.8927e-01],
        [ 4.3637e-01, -4.7804e-01],
        [ ...42e-01, -1.3411e-03],
        [ 5.1762e-03, -8.2245e-01],
        [-8.4013e-02, -1.2175e+00]], grad_fn=<AddmmBackward>)
rhs_data = tensor([[ 0.1037, -0.1837],
        [-0.0162, -0.0123],
        [ 0.0193, -0.0827],
        [ 0.0224, -0.0707],
      ... -0.1169],
        [ 0.0114, -0.0611],
        [ 0.0378, -0.1025],
        [ 0.0290, -0.0979]], grad_fn=<SubBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_cf_conv[1-g4-idtype1] __________________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'hv': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'he': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int64, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 3])
    def test_cf_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        cfconv = nn.CFConv(node_in_feats=2,
                           edge_in_feats=3,
                           hidden_feats=2,
                           out_feats=out_dim)
    
        ctx = F.ctx()
        if F.gpu_ctx():
            cfconv = cfconv.to(ctx)
    
        src_feats = F.randn((g.number_of_src_nodes(), 2))
        edge_feats = F.randn((g.number_of_edges(), 3))
>       h = cfconv(g, src_feats, edge_feats)

tests\pytorch\test_nn.py:1131: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\cfconv.py:141: in forward
    g.update_all(fn.u_mul_e('hv', 'he', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF070D06C8>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[ 0.3384, -0.0174],
        [-1.5532, -0.8653],
        [ 0.4789,  0.4795],
        [-2.1891,  1.3919],
      ...-1.3279],
        [ 1.0438, -1.4701],
        [-0.6542,  1.1595],
        [-2.2503,  0.0839]], grad_fn=<AddmmBackward>)
rhs_data = tensor([[ 0.2910, -0.0387],
        [ 0.2101, -0.0203],
        [ 0.2307, -0.0703],
        [ 0.1284, -0.0570],
      ... -0.0212],
        [ 0.1311, -0.0133],
        [ 0.1838, -0.0766],
        [ 0.4051,  0.0837]], grad_fn=<SubBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_cf_conv[1-g5-idtype0] __________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'hv': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'he': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int32, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 3])
    def test_cf_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        cfconv = nn.CFConv(node_in_feats=2,
                           edge_in_feats=3,
                           hidden_feats=2,
                           out_feats=out_dim)
    
        ctx = F.ctx()
        if F.gpu_ctx():
            cfconv = cfconv.to(ctx)
    
        src_feats = F.randn((g.number_of_src_nodes(), 2))
        edge_feats = F.randn((g.number_of_edges(), 3))
>       h = cfconv(g, src_feats, edge_feats)

tests\pytorch\test_nn.py:1131: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\cfconv.py:141: in forward
    g.update_all(fn.u_mul_e('hv', 'he', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01709748>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[ 1.1695,  0.9294],
        [ 0.0270,  0.1165],
        [ 0.5317,  0.6153],
        [ 1.1138,  0.9625],
      ...-0.2283],
        [ 1.2296,  1.2729],
        [ 0.6608,  0.9211],
        [ 0.8738,  0.5206]], grad_fn=<AddmmBackward>)
rhs_data = tensor([[-0.0205, -0.0561],
        [ 0.0805,  0.0736],
        [ 0.1955,  0.1686],
        [-0.1237, -0.0482],
      ...  0.1129],
        [ 0.0796,  0.0208],
        [-0.0708, -0.0053],
        [ 0.0764,  0.0372]], grad_fn=<SubBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_cf_conv[1-g5-idtype1] __________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'hv': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'he': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int64, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 3])
    def test_cf_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        cfconv = nn.CFConv(node_in_feats=2,
                           edge_in_feats=3,
                           hidden_feats=2,
                           out_feats=out_dim)
    
        ctx = F.ctx()
        if F.gpu_ctx():
            cfconv = cfconv.to(ctx)
    
        src_feats = F.randn((g.number_of_src_nodes(), 2))
        edge_feats = F.randn((g.number_of_edges(), 3))
>       h = cfconv(g, src_feats, edge_feats)

tests\pytorch\test_nn.py:1131: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\cfconv.py:141: in forward
    g.update_all(fn.u_mul_e('hv', 'he', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0743CB88>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[ 0.5434, -0.5901],
        [ 0.5241, -0.5419],
        [ 0.3922, -0.3747],
        [ 0.2414, -0.2652],
      ...-0.6070],
        [ 0.7640, -0.8502],
        [ 0.5522, -0.5358],
        [-0.0091,  0.0380]], grad_fn=<AddmmBackward>)
rhs_data = tensor([[-0.1240, -0.0303],
        [-0.1103, -0.1254],
        [-0.1058, -0.1557],
        [-0.0941, -0.1944],
      ... -0.2634],
        [-0.0852, -0.2465],
        [-0.1101, -0.1272],
        [-0.0804, -0.2778]], grad_fn=<SubBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_cf_conv[1-g6-idtype0] __________________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int32
out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 3])
    def test_cf_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        cfconv = nn.CFConv(node_in_feats=2,
                           edge_in_feats=3,
                           hidden_feats=2,
                           out_feats=out_dim)
    
        ctx = F.ctx()
        if F.gpu_ctx():
            cfconv = cfconv.to(ctx)
    
        src_feats = F.randn((g.number_of_src_nodes(), 2))
        edge_feats = F.randn((g.number_of_edges(), 3))
>       h = cfconv(g, src_feats, edge_feats)

tests\pytorch\test_nn.py:1131: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\cfconv.py:141: in forward
    g.update_all(fn.u_mul_e('hv', 'he', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF077B1548>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[ 0.1677,  0.4548],
        [-0.4094, -0.4531],
        [ 0.2504,  0.6026],
        [ 0.0457,  0.2744],
        [-0.0430,  0.1149],
        [ 0.0645,  0.3395]], grad_fn=<AddmmBackward>)
rhs_data = tensor([[ 0.4264, -0.1747],
        [ 0.4726, -0.2277],
        [ 0.4874, -0.2363]], grad_fn=<SubBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_cf_conv[1-g6-idtype1] __________________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int64
out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 3])
    def test_cf_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        cfconv = nn.CFConv(node_in_feats=2,
                           edge_in_feats=3,
                           hidden_feats=2,
                           out_feats=out_dim)
    
        ctx = F.ctx()
        if F.gpu_ctx():
            cfconv = cfconv.to(ctx)
    
        src_feats = F.randn((g.number_of_src_nodes(), 2))
        edge_feats = F.randn((g.number_of_edges(), 3))
>       h = cfconv(g, src_feats, edge_feats)

tests\pytorch\test_nn.py:1131: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\cfconv.py:141: in forward
    g.update_all(fn.u_mul_e('hv', 'he', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03E01608>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[ 0.5529, -1.0355],
        [-1.3675,  0.4194],
        [ 0.1287, -0.8793],
        [-0.1844, -0.4283],
        [ 0.8775, -1.2841],
        [ 0.3139, -0.8368]], grad_fn=<AddmmBackward>)
rhs_data = tensor([[ 0.1153, -0.3116],
        [ 0.2156, -0.2842],
        [ 0.2108, -0.2907]], grad_fn=<SubBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_cf_conv[1-g7-idtype0] __________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'hv': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'he': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int32, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 3])
    def test_cf_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        cfconv = nn.CFConv(node_in_feats=2,
                           edge_in_feats=3,
                           hidden_feats=2,
                           out_feats=out_dim)
    
        ctx = F.ctx()
        if F.gpu_ctx():
            cfconv = cfconv.to(ctx)
    
        src_feats = F.randn((g.number_of_src_nodes(), 2))
        edge_feats = F.randn((g.number_of_edges(), 3))
>       h = cfconv(g, src_feats, edge_feats)

tests\pytorch\test_nn.py:1131: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\cfconv.py:141: in forward
    g.update_all(fn.u_mul_e('hv', 'he', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03E51D08>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[ 0.7240, -1.0462],
        [ 0.3820, -0.6722],
        [ 0.6814, -1.0605],
        [ 0.5390, -0.8234],
      ...-0.4344],
        [ 0.2305, -0.4538],
        [ 0.1931, -0.5229],
        [ 0.1417, -0.4344]], grad_fn=<AddmmBackward>)
rhs_data = tensor([[-0.0488, -0.0816],
        [-0.4168, -0.0973],
        [-0.2214, -0.0750],
        [-0.1562, -0.0219],
      ... -0.0837],
        [-0.2122, -0.0864],
        [-0.1754, -0.1021],
        [-0.0857, -0.0861]], grad_fn=<SubBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_cf_conv[1-g7-idtype1] __________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'hv': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'he': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int64, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 3])
    def test_cf_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        cfconv = nn.CFConv(node_in_feats=2,
                           edge_in_feats=3,
                           hidden_feats=2,
                           out_feats=out_dim)
    
        ctx = F.ctx()
        if F.gpu_ctx():
            cfconv = cfconv.to(ctx)
    
        src_feats = F.randn((g.number_of_src_nodes(), 2))
        edge_feats = F.randn((g.number_of_edges(), 3))
>       h = cfconv(g, src_feats, edge_feats)

tests\pytorch\test_nn.py:1131: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\cfconv.py:141: in forward
    g.update_all(fn.u_mul_e('hv', 'he', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01973588>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[ 0.1321, -1.3864],
        [ 0.4641, -0.7318],
        [ 0.5015, -0.5702],
        [-0.1072, -0.3683],
      ...-0.7871],
        [ 0.1309,  0.7834],
        [ 0.0371, -1.2877],
        [ 0.7487, -1.1976]], grad_fn=<AddmmBackward>)
rhs_data = tensor([[ 0.1121, -0.1911],
        [ 0.0757, -0.2146],
        [ 0.1208, -0.2143],
        [ 0.1944, -0.1423],
      ... -0.2241],
        [ 0.1900, -0.1446],
        [ 0.2349, -0.1105],
        [ 0.0994, -0.2069]], grad_fn=<SubBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_cf_conv[3-g0-idtype0] __________________________

g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int32, out_dim = 3

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 3])
    def test_cf_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        cfconv = nn.CFConv(node_in_feats=2,
                           edge_in_feats=3,
                           hidden_feats=2,
                           out_feats=out_dim)
    
        ctx = F.ctx()
        if F.gpu_ctx():
            cfconv = cfconv.to(ctx)
    
        src_feats = F.randn((g.number_of_src_nodes(), 2))
        edge_feats = F.randn((g.number_of_edges(), 3))
>       h = cfconv(g, src_feats, edge_feats)

tests\pytorch\test_nn.py:1131: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\cfconv.py:141: in forward
    g.update_all(fn.u_mul_e('hv', 'he', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0793D748>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[-1.3601,  1.3875],
        [ 0.2260,  1.2157]], grad_fn=<AddmmBackward>)
rhs_data = tensor([[ 0.0706,  0.2858],
        [-0.0163,  0.3519],
        [ 0.2913,  0.1362],
        [ 0.2742,  0.1458],
      ... -0.0076],
        [ 0.3882,  0.0806],
        [-0.0010,  0.3395],
        [ 0.1611,  0.2151]], grad_fn=<SubBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_cf_conv[3-g0-idtype1] __________________________

g = Graph(num_nodes={'_U': 2, '_V': 4},
      num_edges={('_U', '_E', '_V'): 8},
      metagraph=[('_U', '_V', '_E')])
idtype = torch.int64, out_dim = 3

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 3])
    def test_cf_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        cfconv = nn.CFConv(node_in_feats=2,
                           edge_in_feats=3,
                           hidden_feats=2,
                           out_feats=out_dim)
    
        ctx = F.ctx()
        if F.gpu_ctx():
            cfconv = cfconv.to(ctx)
    
        src_feats = F.randn((g.number_of_src_nodes(), 2))
        edge_feats = F.randn((g.number_of_edges(), 3))
>       h = cfconv(g, src_feats, edge_feats)

tests\pytorch\test_nn.py:1131: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\cfconv.py:141: in forward
    g.update_all(fn.u_mul_e('hv', 'he', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07309648>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[1.1212, 1.0092],
        [0.3330, 0.2156]], grad_fn=<AddmmBackward>)
rhs_data = tensor([[-0.1770,  0.1776],
        [-0.2192,  0.2054],
        [-0.1689,  0.2583],
        [-0.1929,  0.1956],
      ...  0.2251],
        [-0.1899,  0.1061],
        [-0.2263,  0.3646],
        [-0.1952,  0.1148]], grad_fn=<SubBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_cf_conv[3-g1-idtype0] __________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'hv': Scheme(shap...32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32), 'he': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int32, out_dim = 3

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 3])
    def test_cf_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        cfconv = nn.CFConv(node_in_feats=2,
                           edge_in_feats=3,
                           hidden_feats=2,
                           out_feats=out_dim)
    
        ctx = F.ctx()
        if F.gpu_ctx():
            cfconv = cfconv.to(ctx)
    
        src_feats = F.randn((g.number_of_src_nodes(), 2))
        edge_feats = F.randn((g.number_of_edges(), 3))
>       h = cfconv(g, src_feats, edge_feats)

tests\pytorch\test_nn.py:1131: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\cfconv.py:141: in forward
    g.update_all(fn.u_mul_e('hv', 'he', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1237C188>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[-0.9417, -0.9901],
        [ 0.0051,  0.0779],
        [ 1.8375,  0.3163],
        [ 0.6293, -0.8953],
      ...-1.6287],
        [-0.1944,  0.4275],
        [ 0.3914, -1.1927],
        [ 0.1526,  0.2306]], grad_fn=<AddmmBackward>)
rhs_data = tensor([[-0.1140,  0.4099],
        [-0.1283,  0.2489],
        [-0.1456,  0.2014],
        [-0.1062,  0.3335],
      ...  0.2252],
        [-0.0057,  0.5714],
        [-0.0824,  0.2928],
        [-0.1410,  0.3006]], grad_fn=<SubBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_cf_conv[3-g1-idtype1] __________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'hv': Scheme(shap...32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32), 'he': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int64, out_dim = 3

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 3])
    def test_cf_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        cfconv = nn.CFConv(node_in_feats=2,
                           edge_in_feats=3,
                           hidden_feats=2,
                           out_feats=out_dim)
    
        ctx = F.ctx()
        if F.gpu_ctx():
            cfconv = cfconv.to(ctx)
    
        src_feats = F.randn((g.number_of_src_nodes(), 2))
        edge_feats = F.randn((g.number_of_edges(), 3))
>       h = cfconv(g, src_feats, edge_feats)

tests\pytorch\test_nn.py:1131: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\cfconv.py:141: in forward
    g.update_all(fn.u_mul_e('hv', 'he', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1225FF88>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[-0.6605, -0.9031],
        [ 0.7032, -0.6744],
        [ 0.9075, -1.0427],
        [ 1.3722, -0.7992],
      ...-0.2457],
        [ 0.7234,  0.6957],
        [-0.5191,  0.4515],
        [ 1.1244,  0.1045]], grad_fn=<AddmmBackward>)
rhs_data = tensor([[ 0.4939,  0.1642],
        [ 0.7238, -0.0041],
        [ 1.1079, -0.2024],
        [ 0.9937, -0.1238],
      ...  0.1780],
        [ 0.3609,  0.2623],
        [ 0.2757,  0.3295],
        [ 0.5029,  0.1205]], grad_fn=<SubBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_cf_conv[3-g2-idtype0] __________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'hv': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'he': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int32, out_dim = 3

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 3])
    def test_cf_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        cfconv = nn.CFConv(node_in_feats=2,
                           edge_in_feats=3,
                           hidden_feats=2,
                           out_feats=out_dim)
    
        ctx = F.ctx()
        if F.gpu_ctx():
            cfconv = cfconv.to(ctx)
    
        src_feats = F.randn((g.number_of_src_nodes(), 2))
        edge_feats = F.randn((g.number_of_edges(), 3))
>       h = cfconv(g, src_feats, edge_feats)

tests\pytorch\test_nn.py:1131: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\cfconv.py:141: in forward
    g.update_all(fn.u_mul_e('hv', 'he', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01890188>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[-0.6089, -0.5040],
        [-0.6448, -0.2721],
        [-0.6322, -0.8929],
        [-0.6227, -0.5643],
      ...-0.4440],
        [-0.6397, -0.1292],
        [-0.7402, -1.1689],
        [-0.6188, -0.1321]], grad_fn=<AddmmBackward>)
rhs_data = tensor([[ 0.3053, -0.2366],
        [ 0.3231, -0.2165],
        [ 0.3293, -0.2146],
        [ 0.3204, -0.2247],
      ... -0.1773],
        [ 0.3335, -0.1930],
        [ 0.3335, -0.2011],
        [ 0.2740, -0.2644]], grad_fn=<SubBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_cf_conv[3-g2-idtype1] __________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'hv': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'he': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int64, out_dim = 3

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 3])
    def test_cf_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        cfconv = nn.CFConv(node_in_feats=2,
                           edge_in_feats=3,
                           hidden_feats=2,
                           out_feats=out_dim)
    
        ctx = F.ctx()
        if F.gpu_ctx():
            cfconv = cfconv.to(ctx)
    
        src_feats = F.randn((g.number_of_src_nodes(), 2))
        edge_feats = F.randn((g.number_of_edges(), 3))
>       h = cfconv(g, src_feats, edge_feats)

tests\pytorch\test_nn.py:1131: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\cfconv.py:141: in forward
    g.update_all(fn.u_mul_e('hv', 'he', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0742CF88>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[ 0.4588,  0.0313],
        [-0.5335, -0.0400],
        [-0.3967,  0.5772],
        [-0.1241,  0.5410],
      ...-0.3400],
        [ 0.4190,  0.3312],
        [ 1.2772, -1.4966],
        [ 0.7217, -0.4742]], grad_fn=<AddmmBackward>)
rhs_data = tensor([[ 0.0317,  0.5005],
        [ 0.0516,  0.5524],
        [ 0.0077,  0.4559],
        [-0.0283,  0.4155],
      ...  0.3351],
        [-0.0642,  0.3377],
        [-0.0470,  0.3476],
        [-0.0184,  0.4145]], grad_fn=<SubBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_cf_conv[3-g3-idtype0] __________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'hv': Scheme(shap...      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), 'he': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int32, out_dim = 3

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 3])
    def test_cf_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        cfconv = nn.CFConv(node_in_feats=2,
                           edge_in_feats=3,
                           hidden_feats=2,
                           out_feats=out_dim)
    
        ctx = F.ctx()
        if F.gpu_ctx():
            cfconv = cfconv.to(ctx)
    
        src_feats = F.randn((g.number_of_src_nodes(), 2))
        edge_feats = F.randn((g.number_of_edges(), 3))
>       h = cfconv(g, src_feats, edge_feats)

tests\pytorch\test_nn.py:1131: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\cfconv.py:141: in forward
    g.update_all(fn.u_mul_e('hv', 'he', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0732E888>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[-0.1415,  0.3143],
        [-0.5523, -0.1056],
        [-0.6425,  1.7113],
        [ 0.0089, -0.0386],
      ...-0.1144],
        [-0.4237,  0.2457],
        [-0.6992,  0.7787],
        [-0.6499,  0.9044]], grad_fn=<AddmmBackward>)
rhs_data = tensor([[-0.0592, -0.2300],
        [-0.0439, -0.2201],
        [-0.0511, -0.2210],
        [-0.1340, -0.3341],
      ... -0.2747],
        [-0.0295, -0.1982],
        [-0.0683, -0.2408],
        [-0.0857, -0.2714]], grad_fn=<SubBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_cf_conv[3-g3-idtype1] __________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'hv': Scheme(shap...      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), 'he': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int64, out_dim = 3

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 3])
    def test_cf_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        cfconv = nn.CFConv(node_in_feats=2,
                           edge_in_feats=3,
                           hidden_feats=2,
                           out_feats=out_dim)
    
        ctx = F.ctx()
        if F.gpu_ctx():
            cfconv = cfconv.to(ctx)
    
        src_feats = F.randn((g.number_of_src_nodes(), 2))
        edge_feats = F.randn((g.number_of_edges(), 3))
>       h = cfconv(g, src_feats, edge_feats)

tests\pytorch\test_nn.py:1131: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\cfconv.py:141: in forward
    g.update_all(fn.u_mul_e('hv', 'he', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF077D2D08>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[-0.4380, -0.4481],
        [ 0.3606, -1.0650],
        [ 0.1241, -0.3717],
        [ 1.2115, -3.1293],
      ...-1.1628],
        [ 0.1798, -1.0541],
        [-0.2272, -1.0274],
        [-0.0821,  0.1897]], grad_fn=<AddmmBackward>)
rhs_data = tensor([[-0.2521,  0.0691],
        [-0.2956, -0.0417],
        [-0.3016, -0.0219],
        [-0.2817,  0.0363],
      ...  0.0504],
        [-0.2361,  0.0184],
        [-0.1943,  0.1093],
        [-0.2175,  0.0664]], grad_fn=<SubBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_cf_conv[3-g4-idtype0] __________________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'hv': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'he': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int32, out_dim = 3

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 3])
    def test_cf_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        cfconv = nn.CFConv(node_in_feats=2,
                           edge_in_feats=3,
                           hidden_feats=2,
                           out_feats=out_dim)
    
        ctx = F.ctx()
        if F.gpu_ctx():
            cfconv = cfconv.to(ctx)
    
        src_feats = F.randn((g.number_of_src_nodes(), 2))
        edge_feats = F.randn((g.number_of_edges(), 3))
>       h = cfconv(g, src_feats, edge_feats)

tests\pytorch\test_nn.py:1131: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\cfconv.py:141: in forward
    g.update_all(fn.u_mul_e('hv', 'he', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF070F61C8>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[ 0.2628,  1.0869],
        [ 0.6771,  0.6082],
        [ 0.5164,  0.7393],
        [-0.6578,  1.3990],
      ... 1.0551],
        [ 1.1883,  0.5425],
        [ 1.1203,  0.0504],
        [ 0.7922,  0.9932]], grad_fn=<AddmmBackward>)
rhs_data = tensor([[-0.3165,  0.2238],
        [-0.2727,  0.2404],
        [-0.2912,  0.2499],
        [-0.2670,  0.2527],
      ...  0.2024],
        [-0.3545,  0.1823],
        [-0.3255,  0.2262],
        [-0.2844,  0.2451]], grad_fn=<SubBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_cf_conv[3-g4-idtype1] __________________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'hv': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'he': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int64, out_dim = 3

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 3])
    def test_cf_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        cfconv = nn.CFConv(node_in_feats=2,
                           edge_in_feats=3,
                           hidden_feats=2,
                           out_feats=out_dim)
    
        ctx = F.ctx()
        if F.gpu_ctx():
            cfconv = cfconv.to(ctx)
    
        src_feats = F.randn((g.number_of_src_nodes(), 2))
        edge_feats = F.randn((g.number_of_edges(), 3))
>       h = cfconv(g, src_feats, edge_feats)

tests\pytorch\test_nn.py:1131: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\cfconv.py:141: in forward
    g.update_all(fn.u_mul_e('hv', 'he', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07779808>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[ 0.4941,  0.4901],
        [ 0.4641,  0.4366],
        [ 0.5565,  0.7978],
        [ 0.4979,  0.4381],
      ... 0.8441],
        [ 0.5789,  0.7933],
        [ 0.5945,  0.9087],
        [ 0.7685,  1.5726]], grad_fn=<AddmmBackward>)
rhs_data = tensor([[ 0.0670,  0.0383],
        [ 0.1258,  0.1065],
        [-0.0134, -0.0965],
        [ 0.1072,  0.0710],
      ...  0.1001],
        [ 0.1355,  0.1150],
        [ 0.1127,  0.0859],
        [ 0.0716,  0.0386]], grad_fn=<SubBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_cf_conv[3-g5-idtype0] __________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'hv': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'he': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int32, out_dim = 3

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 3])
    def test_cf_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        cfconv = nn.CFConv(node_in_feats=2,
                           edge_in_feats=3,
                           hidden_feats=2,
                           out_feats=out_dim)
    
        ctx = F.ctx()
        if F.gpu_ctx():
            cfconv = cfconv.to(ctx)
    
        src_feats = F.randn((g.number_of_src_nodes(), 2))
        edge_feats = F.randn((g.number_of_edges(), 3))
>       h = cfconv(g, src_feats, edge_feats)

tests\pytorch\test_nn.py:1131: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\cfconv.py:141: in forward
    g.update_all(fn.u_mul_e('hv', 'he', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1232B448>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[-0.0736,  0.4641],
        [ 1.9047, -1.5977],
        [ 0.3181,  0.2111],
        [ 0.1287,  0.3254],
      ...-0.8896],
        [ 0.0654,  0.4333],
        [ 0.2686,  0.3084],
        [ 0.2286,  0.2036]], grad_fn=<AddmmBackward>)
rhs_data = tensor([[ 0.0152, -0.2629],
        [-0.0362, -0.2077],
        [-0.0189, -0.2596],
        [-0.1453, -0.1980],
      ... -0.2746],
        [ 0.1936, -0.3303],
        [ 0.0974, -0.2755],
        [ 0.1182, -0.2668]], grad_fn=<SubBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_cf_conv[3-g5-idtype1] __________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'hv': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'he': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int64, out_dim = 3

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 3])
    def test_cf_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        cfconv = nn.CFConv(node_in_feats=2,
                           edge_in_feats=3,
                           hidden_feats=2,
                           out_feats=out_dim)
    
        ctx = F.ctx()
        if F.gpu_ctx():
            cfconv = cfconv.to(ctx)
    
        src_feats = F.randn((g.number_of_src_nodes(), 2))
        edge_feats = F.randn((g.number_of_edges(), 3))
>       h = cfconv(g, src_feats, edge_feats)

tests\pytorch\test_nn.py:1131: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\cfconv.py:141: in forward
    g.update_all(fn.u_mul_e('hv', 'he', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF018C0C48>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[ 0.9200, -0.2535],
        [-0.7085, -0.2813],
        [ 0.3873, -0.0866],
        [-0.5302, -0.8738],
      ...-1.1078],
        [-0.0254, -0.8007],
        [-0.0522, -0.1383],
        [ 0.3218, -0.3205]], grad_fn=<AddmmBackward>)
rhs_data = tensor([[ 0.1152, -0.2811],
        [ 0.1181, -0.1706],
        [ 0.1433, -0.2596],
        [ 0.1307, -0.1279],
      ... -0.1837],
        [ 0.1467, -0.1241],
        [ 0.0818, -0.2227],
        [ 0.0809, -0.1871]], grad_fn=<SubBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_cf_conv[3-g6-idtype0] __________________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int32
out_dim = 3

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 3])
    def test_cf_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        cfconv = nn.CFConv(node_in_feats=2,
                           edge_in_feats=3,
                           hidden_feats=2,
                           out_feats=out_dim)
    
        ctx = F.ctx()
        if F.gpu_ctx():
            cfconv = cfconv.to(ctx)
    
        src_feats = F.randn((g.number_of_src_nodes(), 2))
        edge_feats = F.randn((g.number_of_edges(), 3))
>       h = cfconv(g, src_feats, edge_feats)

tests\pytorch\test_nn.py:1131: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\cfconv.py:141: in forward
    g.update_all(fn.u_mul_e('hv', 'he', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03DD9648>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[-0.1436,  0.4064],
        [ 0.6517,  1.6997],
        [-0.2646,  0.2199],
        [-1.1011, -1.2607],
        [-1.2178, -1.4861],
        [-0.4590, -0.1612]], grad_fn=<AddmmBackward>)
rhs_data = tensor([[0.0357, 0.2791],
        [0.1345, 0.2601],
        [0.0985, 0.2724]], grad_fn=<SubBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_cf_conv[3-g6-idtype1] __________________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3), idtype = torch.int64
out_dim = 3

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 3])
    def test_cf_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        cfconv = nn.CFConv(node_in_feats=2,
                           edge_in_feats=3,
                           hidden_feats=2,
                           out_feats=out_dim)
    
        ctx = F.ctx()
        if F.gpu_ctx():
            cfconv = cfconv.to(ctx)
    
        src_feats = F.randn((g.number_of_src_nodes(), 2))
        edge_feats = F.randn((g.number_of_edges(), 3))
>       h = cfconv(g, src_feats, edge_feats)

tests\pytorch\test_nn.py:1131: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\cfconv.py:141: in forward
    g.update_all(fn.u_mul_e('hv', 'he', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF077D1548>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[ 0.2804,  0.1498],
        [-0.5499, -0.6719],
        [ 0.2159, -0.0975],
        [ 0.0350, -0.2297],
        [ 0.7875,  0.5588],
        [ 0.6448,  0.5562]], grad_fn=<AddmmBackward>)
rhs_data = tensor([[ 0.0125,  0.2376],
        [ 0.2731, -0.0329],
        [ 0.0458,  0.2169]], grad_fn=<SubBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_cf_conv[3-g7-idtype0] __________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'hv': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'he': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int32, out_dim = 3

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 3])
    def test_cf_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        cfconv = nn.CFConv(node_in_feats=2,
                           edge_in_feats=3,
                           hidden_feats=2,
                           out_feats=out_dim)
    
        ctx = F.ctx()
        if F.gpu_ctx():
            cfconv = cfconv.to(ctx)
    
        src_feats = F.randn((g.number_of_src_nodes(), 2))
        edge_feats = F.randn((g.number_of_edges(), 3))
>       h = cfconv(g, src_feats, edge_feats)

tests\pytorch\test_nn.py:1131: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\cfconv.py:141: in forward
    g.update_all(fn.u_mul_e('hv', 'he', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0192C148>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[ 0.3197, -1.0854],
        [ 0.2007, -0.8984],
        [-0.0698, -0.5147],
        [ 0.1016, -0.4987],
      ...-0.5964],
        [-0.0471, -0.5245],
        [ 0.1367, -0.7330],
        [-0.3348, -0.7299]], grad_fn=<AddmmBackward>)
rhs_data = tensor([[ 0.3022, -0.2503],
        [ 0.3084, -0.2160],
        [ 0.3207, -0.1390],
        [ 0.3032, -0.2470],
      ... -0.2428],
        [ 0.2802, -0.3631],
        [ 0.2809, -0.3598],
        [ 0.3133, -0.1875]], grad_fn=<SubBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_cf_conv[3-g7-idtype1] __________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'hv': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'he': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int64, out_dim = 3

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo', 'bipartite'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 3])
    def test_cf_conv(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        cfconv = nn.CFConv(node_in_feats=2,
                           edge_in_feats=3,
                           hidden_feats=2,
                           out_feats=out_dim)
    
        ctx = F.ctx()
        if F.gpu_ctx():
            cfconv = cfconv.to(ctx)
    
        src_feats = F.randn((g.number_of_src_nodes(), 2))
        edge_feats = F.randn((g.number_of_edges(), 3))
>       h = cfconv(g, src_feats, edge_feats)

tests\pytorch\test_nn.py:1131: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\cfconv.py:141: in forward
    g.update_all(fn.u_mul_e('hv', 'he', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0796ABC8>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[0.4647, 1.6154],
        [0.5008, 0.4884],
        [0.4233, 0.0997],
        [0.4943, 0.2448],
        [0.501....5051, 0.7847],
        [0.5804, 1.2393],
        [0.4814, 0.6121],
        [0.4941, 0.0775]], grad_fn=<AddmmBackward>)
rhs_data = tensor([[ 0.5990, -0.0943],
        [ 0.2034, -0.1971],
        [ 0.0647, -0.2374],
        [ 0.6960, -0.0682],
      ... -0.1513],
        [ 0.4471, -0.1353],
        [ 0.3549, -0.1566],
        [ 0.2290, -0.1916]], grad_fn=<SubBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_____________________ test_hetero_conv[False-sum-idtype0] _____________________

agg = 'sum', idtype = torch.int32, canonical_keys = False

    @parametrize_idtype
    @pytest.mark.parametrize('agg', ['sum', 'max', 'min', 'mean', 'stack', myagg])
    @pytest.mark.parametrize('canonical_keys', [False, True])
    def test_hetero_conv(agg, idtype, canonical_keys):
        g = dgl.heterograph({
            ('user', 'follows', 'user'): ([0, 0, 2, 1], [1, 2, 1, 3]),
            ('user', 'plays', 'game'): ([0, 0, 0, 1, 2], [0, 2, 3, 0, 2]),
            ('store', 'sells', 'game'): ([0, 0, 1, 1], [0, 3, 1, 2])},
            idtype=idtype, device=F.ctx())
        if not canonical_keys:
            conv = nn.HeteroGraphConv({
                'follows': nn.GraphConv(2, 3, allow_zero_in_degree=True),
                'plays': nn.GraphConv(2, 4, allow_zero_in_degree=True),
                'sells': nn.GraphConv(3, 4, allow_zero_in_degree=True)},
                agg)
        else:
            conv = nn.HeteroGraphConv({
                ('user', 'follows', 'user'): nn.GraphConv(2, 3, allow_zero_in_degree=True),
                ('user', 'plays', 'game'): nn.GraphConv(2, 4, allow_zero_in_degree=True),
                ('store', 'sells', 'game'): nn.GraphConv(3, 4, allow_zero_in_degree=True)},
                agg)
    
        conv = conv.to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        uf = F.randn((4, 2))
        gf = F.randn((4, 4))
        sf = F.randn((2, 3))
    
>       h = conv(g, {'user': uf, 'game': gf, 'store': sf})

tests\pytorch\test_nn.py:1178: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\hetero.py:205: in forward
    **mod_kwargs.get(etype, {}))
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:428: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07018188>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.1480, -0.2703, -1.3044],
        [ 1.4912, -0.3858, -0.5647]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_____________________ test_hetero_conv[False-sum-idtype1] _____________________

agg = 'sum', idtype = torch.int64, canonical_keys = False

    @parametrize_idtype
    @pytest.mark.parametrize('agg', ['sum', 'max', 'min', 'mean', 'stack', myagg])
    @pytest.mark.parametrize('canonical_keys', [False, True])
    def test_hetero_conv(agg, idtype, canonical_keys):
        g = dgl.heterograph({
            ('user', 'follows', 'user'): ([0, 0, 2, 1], [1, 2, 1, 3]),
            ('user', 'plays', 'game'): ([0, 0, 0, 1, 2], [0, 2, 3, 0, 2]),
            ('store', 'sells', 'game'): ([0, 0, 1, 1], [0, 3, 1, 2])},
            idtype=idtype, device=F.ctx())
        if not canonical_keys:
            conv = nn.HeteroGraphConv({
                'follows': nn.GraphConv(2, 3, allow_zero_in_degree=True),
                'plays': nn.GraphConv(2, 4, allow_zero_in_degree=True),
                'sells': nn.GraphConv(3, 4, allow_zero_in_degree=True)},
                agg)
        else:
            conv = nn.HeteroGraphConv({
                ('user', 'follows', 'user'): nn.GraphConv(2, 3, allow_zero_in_degree=True),
                ('user', 'plays', 'game'): nn.GraphConv(2, 4, allow_zero_in_degree=True),
                ('store', 'sells', 'game'): nn.GraphConv(3, 4, allow_zero_in_degree=True)},
                agg)
    
        conv = conv.to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        uf = F.randn((4, 2))
        gf = F.randn((4, 4))
        sf = F.randn((2, 3))
    
>       h = conv(g, {'user': uf, 'game': gf, 'store': sf})

tests\pytorch\test_nn.py:1178: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\hetero.py:205: in forward
    **mod_kwargs.get(etype, {}))
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:428: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07999A88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.1559,  1.1797, -0.2209],
        [ 0.1136, -0.5057, -0.6093]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_____________________ test_hetero_conv[False-max-idtype0] _____________________

agg = 'max', idtype = torch.int32, canonical_keys = False

    @parametrize_idtype
    @pytest.mark.parametrize('agg', ['sum', 'max', 'min', 'mean', 'stack', myagg])
    @pytest.mark.parametrize('canonical_keys', [False, True])
    def test_hetero_conv(agg, idtype, canonical_keys):
        g = dgl.heterograph({
            ('user', 'follows', 'user'): ([0, 0, 2, 1], [1, 2, 1, 3]),
            ('user', 'plays', 'game'): ([0, 0, 0, 1, 2], [0, 2, 3, 0, 2]),
            ('store', 'sells', 'game'): ([0, 0, 1, 1], [0, 3, 1, 2])},
            idtype=idtype, device=F.ctx())
        if not canonical_keys:
            conv = nn.HeteroGraphConv({
                'follows': nn.GraphConv(2, 3, allow_zero_in_degree=True),
                'plays': nn.GraphConv(2, 4, allow_zero_in_degree=True),
                'sells': nn.GraphConv(3, 4, allow_zero_in_degree=True)},
                agg)
        else:
            conv = nn.HeteroGraphConv({
                ('user', 'follows', 'user'): nn.GraphConv(2, 3, allow_zero_in_degree=True),
                ('user', 'plays', 'game'): nn.GraphConv(2, 4, allow_zero_in_degree=True),
                ('store', 'sells', 'game'): nn.GraphConv(3, 4, allow_zero_in_degree=True)},
                agg)
    
        conv = conv.to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        uf = F.randn((4, 2))
        gf = F.randn((4, 4))
        sf = F.randn((2, 3))
    
>       h = conv(g, {'user': uf, 'game': gf, 'store': sf})

tests\pytorch\test_nn.py:1178: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\hetero.py:205: in forward
    **mod_kwargs.get(etype, {}))
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:428: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF070A5808>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.6523, -1.2169, -0.0322],
        [ 0.3153,  1.4738, -0.4358]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_____________________ test_hetero_conv[False-max-idtype1] _____________________

agg = 'max', idtype = torch.int64, canonical_keys = False

    @parametrize_idtype
    @pytest.mark.parametrize('agg', ['sum', 'max', 'min', 'mean', 'stack', myagg])
    @pytest.mark.parametrize('canonical_keys', [False, True])
    def test_hetero_conv(agg, idtype, canonical_keys):
        g = dgl.heterograph({
            ('user', 'follows', 'user'): ([0, 0, 2, 1], [1, 2, 1, 3]),
            ('user', 'plays', 'game'): ([0, 0, 0, 1, 2], [0, 2, 3, 0, 2]),
            ('store', 'sells', 'game'): ([0, 0, 1, 1], [0, 3, 1, 2])},
            idtype=idtype, device=F.ctx())
        if not canonical_keys:
            conv = nn.HeteroGraphConv({
                'follows': nn.GraphConv(2, 3, allow_zero_in_degree=True),
                'plays': nn.GraphConv(2, 4, allow_zero_in_degree=True),
                'sells': nn.GraphConv(3, 4, allow_zero_in_degree=True)},
                agg)
        else:
            conv = nn.HeteroGraphConv({
                ('user', 'follows', 'user'): nn.GraphConv(2, 3, allow_zero_in_degree=True),
                ('user', 'plays', 'game'): nn.GraphConv(2, 4, allow_zero_in_degree=True),
                ('store', 'sells', 'game'): nn.GraphConv(3, 4, allow_zero_in_degree=True)},
                agg)
    
        conv = conv.to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        uf = F.randn((4, 2))
        gf = F.randn((4, 4))
        sf = F.randn((2, 3))
    
>       h = conv(g, {'user': uf, 'game': gf, 'store': sf})

tests\pytorch\test_nn.py:1178: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\hetero.py:205: in forward
    **mod_kwargs.get(etype, {}))
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:428: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03E5C888>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.1729, -0.1719,  0.7675],
        [ 0.4450,  0.1347, -0.5818]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_____________________ test_hetero_conv[False-min-idtype0] _____________________

agg = 'min', idtype = torch.int32, canonical_keys = False

    @parametrize_idtype
    @pytest.mark.parametrize('agg', ['sum', 'max', 'min', 'mean', 'stack', myagg])
    @pytest.mark.parametrize('canonical_keys', [False, True])
    def test_hetero_conv(agg, idtype, canonical_keys):
        g = dgl.heterograph({
            ('user', 'follows', 'user'): ([0, 0, 2, 1], [1, 2, 1, 3]),
            ('user', 'plays', 'game'): ([0, 0, 0, 1, 2], [0, 2, 3, 0, 2]),
            ('store', 'sells', 'game'): ([0, 0, 1, 1], [0, 3, 1, 2])},
            idtype=idtype, device=F.ctx())
        if not canonical_keys:
            conv = nn.HeteroGraphConv({
                'follows': nn.GraphConv(2, 3, allow_zero_in_degree=True),
                'plays': nn.GraphConv(2, 4, allow_zero_in_degree=True),
                'sells': nn.GraphConv(3, 4, allow_zero_in_degree=True)},
                agg)
        else:
            conv = nn.HeteroGraphConv({
                ('user', 'follows', 'user'): nn.GraphConv(2, 3, allow_zero_in_degree=True),
                ('user', 'plays', 'game'): nn.GraphConv(2, 4, allow_zero_in_degree=True),
                ('store', 'sells', 'game'): nn.GraphConv(3, 4, allow_zero_in_degree=True)},
                agg)
    
        conv = conv.to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        uf = F.randn((4, 2))
        gf = F.randn((4, 4))
        sf = F.randn((2, 3))
    
>       h = conv(g, {'user': uf, 'game': gf, 'store': sf})

tests\pytorch\test_nn.py:1178: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\hetero.py:205: in forward
    **mod_kwargs.get(etype, {}))
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:428: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE12314D88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.0833,  0.9648, -0.7251],
        [-0.2482, -1.1087, -0.2997]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_____________________ test_hetero_conv[False-min-idtype1] _____________________

agg = 'min', idtype = torch.int64, canonical_keys = False

    @parametrize_idtype
    @pytest.mark.parametrize('agg', ['sum', 'max', 'min', 'mean', 'stack', myagg])
    @pytest.mark.parametrize('canonical_keys', [False, True])
    def test_hetero_conv(agg, idtype, canonical_keys):
        g = dgl.heterograph({
            ('user', 'follows', 'user'): ([0, 0, 2, 1], [1, 2, 1, 3]),
            ('user', 'plays', 'game'): ([0, 0, 0, 1, 2], [0, 2, 3, 0, 2]),
            ('store', 'sells', 'game'): ([0, 0, 1, 1], [0, 3, 1, 2])},
            idtype=idtype, device=F.ctx())
        if not canonical_keys:
            conv = nn.HeteroGraphConv({
                'follows': nn.GraphConv(2, 3, allow_zero_in_degree=True),
                'plays': nn.GraphConv(2, 4, allow_zero_in_degree=True),
                'sells': nn.GraphConv(3, 4, allow_zero_in_degree=True)},
                agg)
        else:
            conv = nn.HeteroGraphConv({
                ('user', 'follows', 'user'): nn.GraphConv(2, 3, allow_zero_in_degree=True),
                ('user', 'plays', 'game'): nn.GraphConv(2, 4, allow_zero_in_degree=True),
                ('store', 'sells', 'game'): nn.GraphConv(3, 4, allow_zero_in_degree=True)},
                agg)
    
        conv = conv.to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        uf = F.randn((4, 2))
        gf = F.randn((4, 4))
        sf = F.randn((2, 3))
    
>       h = conv(g, {'user': uf, 'game': gf, 'store': sf})

tests\pytorch\test_nn.py:1178: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\hetero.py:205: in forward
    **mod_kwargs.get(etype, {}))
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:428: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0171EE88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.0219, -0.1872, -1.8041],
        [-0.2160, -0.6605, -0.5506]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
____________________ test_hetero_conv[False-mean-idtype0] _____________________

agg = 'mean', idtype = torch.int32, canonical_keys = False

    @parametrize_idtype
    @pytest.mark.parametrize('agg', ['sum', 'max', 'min', 'mean', 'stack', myagg])
    @pytest.mark.parametrize('canonical_keys', [False, True])
    def test_hetero_conv(agg, idtype, canonical_keys):
        g = dgl.heterograph({
            ('user', 'follows', 'user'): ([0, 0, 2, 1], [1, 2, 1, 3]),
            ('user', 'plays', 'game'): ([0, 0, 0, 1, 2], [0, 2, 3, 0, 2]),
            ('store', 'sells', 'game'): ([0, 0, 1, 1], [0, 3, 1, 2])},
            idtype=idtype, device=F.ctx())
        if not canonical_keys:
            conv = nn.HeteroGraphConv({
                'follows': nn.GraphConv(2, 3, allow_zero_in_degree=True),
                'plays': nn.GraphConv(2, 4, allow_zero_in_degree=True),
                'sells': nn.GraphConv(3, 4, allow_zero_in_degree=True)},
                agg)
        else:
            conv = nn.HeteroGraphConv({
                ('user', 'follows', 'user'): nn.GraphConv(2, 3, allow_zero_in_degree=True),
                ('user', 'plays', 'game'): nn.GraphConv(2, 4, allow_zero_in_degree=True),
                ('store', 'sells', 'game'): nn.GraphConv(3, 4, allow_zero_in_degree=True)},
                agg)
    
        conv = conv.to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        uf = F.randn((4, 2))
        gf = F.randn((4, 4))
        sf = F.randn((2, 3))
    
>       h = conv(g, {'user': uf, 'game': gf, 'store': sf})

tests\pytorch\test_nn.py:1178: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\hetero.py:205: in forward
    **mod_kwargs.get(etype, {}))
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:428: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03CC15C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.2757, -0.7083, -0.6424],
        [-0.8338, -0.6223,  0.0424]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
____________________ test_hetero_conv[False-mean-idtype1] _____________________

agg = 'mean', idtype = torch.int64, canonical_keys = False

    @parametrize_idtype
    @pytest.mark.parametrize('agg', ['sum', 'max', 'min', 'mean', 'stack', myagg])
    @pytest.mark.parametrize('canonical_keys', [False, True])
    def test_hetero_conv(agg, idtype, canonical_keys):
        g = dgl.heterograph({
            ('user', 'follows', 'user'): ([0, 0, 2, 1], [1, 2, 1, 3]),
            ('user', 'plays', 'game'): ([0, 0, 0, 1, 2], [0, 2, 3, 0, 2]),
            ('store', 'sells', 'game'): ([0, 0, 1, 1], [0, 3, 1, 2])},
            idtype=idtype, device=F.ctx())
        if not canonical_keys:
            conv = nn.HeteroGraphConv({
                'follows': nn.GraphConv(2, 3, allow_zero_in_degree=True),
                'plays': nn.GraphConv(2, 4, allow_zero_in_degree=True),
                'sells': nn.GraphConv(3, 4, allow_zero_in_degree=True)},
                agg)
        else:
            conv = nn.HeteroGraphConv({
                ('user', 'follows', 'user'): nn.GraphConv(2, 3, allow_zero_in_degree=True),
                ('user', 'plays', 'game'): nn.GraphConv(2, 4, allow_zero_in_degree=True),
                ('store', 'sells', 'game'): nn.GraphConv(3, 4, allow_zero_in_degree=True)},
                agg)
    
        conv = conv.to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        uf = F.randn((4, 2))
        gf = F.randn((4, 4))
        sf = F.randn((2, 3))
    
>       h = conv(g, {'user': uf, 'game': gf, 'store': sf})

tests\pytorch\test_nn.py:1178: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\hetero.py:205: in forward
    **mod_kwargs.get(etype, {}))
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:428: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07381F88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.8001,  0.6591, -1.7937],
        [ 0.4277,  1.2046,  0.8283]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
____________________ test_hetero_conv[False-stack-idtype0] ____________________

agg = 'stack', idtype = torch.int32, canonical_keys = False

    @parametrize_idtype
    @pytest.mark.parametrize('agg', ['sum', 'max', 'min', 'mean', 'stack', myagg])
    @pytest.mark.parametrize('canonical_keys', [False, True])
    def test_hetero_conv(agg, idtype, canonical_keys):
        g = dgl.heterograph({
            ('user', 'follows', 'user'): ([0, 0, 2, 1], [1, 2, 1, 3]),
            ('user', 'plays', 'game'): ([0, 0, 0, 1, 2], [0, 2, 3, 0, 2]),
            ('store', 'sells', 'game'): ([0, 0, 1, 1], [0, 3, 1, 2])},
            idtype=idtype, device=F.ctx())
        if not canonical_keys:
            conv = nn.HeteroGraphConv({
                'follows': nn.GraphConv(2, 3, allow_zero_in_degree=True),
                'plays': nn.GraphConv(2, 4, allow_zero_in_degree=True),
                'sells': nn.GraphConv(3, 4, allow_zero_in_degree=True)},
                agg)
        else:
            conv = nn.HeteroGraphConv({
                ('user', 'follows', 'user'): nn.GraphConv(2, 3, allow_zero_in_degree=True),
                ('user', 'plays', 'game'): nn.GraphConv(2, 4, allow_zero_in_degree=True),
                ('store', 'sells', 'game'): nn.GraphConv(3, 4, allow_zero_in_degree=True)},
                agg)
    
        conv = conv.to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        uf = F.randn((4, 2))
        gf = F.randn((4, 4))
        sf = F.randn((2, 3))
    
>       h = conv(g, {'user': uf, 'game': gf, 'store': sf})

tests\pytorch\test_nn.py:1178: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\hetero.py:205: in forward
    **mod_kwargs.get(etype, {}))
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:428: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0181F0C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.2573, -0.3374,  0.5159],
        [-0.5254,  0.7733, -0.4114]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
____________________ test_hetero_conv[False-stack-idtype1] ____________________

agg = 'stack', idtype = torch.int64, canonical_keys = False

    @parametrize_idtype
    @pytest.mark.parametrize('agg', ['sum', 'max', 'min', 'mean', 'stack', myagg])
    @pytest.mark.parametrize('canonical_keys', [False, True])
    def test_hetero_conv(agg, idtype, canonical_keys):
        g = dgl.heterograph({
            ('user', 'follows', 'user'): ([0, 0, 2, 1], [1, 2, 1, 3]),
            ('user', 'plays', 'game'): ([0, 0, 0, 1, 2], [0, 2, 3, 0, 2]),
            ('store', 'sells', 'game'): ([0, 0, 1, 1], [0, 3, 1, 2])},
            idtype=idtype, device=F.ctx())
        if not canonical_keys:
            conv = nn.HeteroGraphConv({
                'follows': nn.GraphConv(2, 3, allow_zero_in_degree=True),
                'plays': nn.GraphConv(2, 4, allow_zero_in_degree=True),
                'sells': nn.GraphConv(3, 4, allow_zero_in_degree=True)},
                agg)
        else:
            conv = nn.HeteroGraphConv({
                ('user', 'follows', 'user'): nn.GraphConv(2, 3, allow_zero_in_degree=True),
                ('user', 'plays', 'game'): nn.GraphConv(2, 4, allow_zero_in_degree=True),
                ('store', 'sells', 'game'): nn.GraphConv(3, 4, allow_zero_in_degree=True)},
                agg)
    
        conv = conv.to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        uf = F.randn((4, 2))
        gf = F.randn((4, 4))
        sf = F.randn((2, 3))
    
>       h = conv(g, {'user': uf, 'game': gf, 'store': sf})

tests\pytorch\test_nn.py:1178: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\hetero.py:205: in forward
    **mod_kwargs.get(etype, {}))
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:428: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE12215AC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.8973, -1.4752, -0.3907],
        [-0.4775,  0.2110, -0.7121]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
____________________ test_hetero_conv[False-myagg-idtype0] ____________________

agg = <function myagg at 0x000000AE126FB620>, idtype = torch.int32
canonical_keys = False

    @parametrize_idtype
    @pytest.mark.parametrize('agg', ['sum', 'max', 'min', 'mean', 'stack', myagg])
    @pytest.mark.parametrize('canonical_keys', [False, True])
    def test_hetero_conv(agg, idtype, canonical_keys):
        g = dgl.heterograph({
            ('user', 'follows', 'user'): ([0, 0, 2, 1], [1, 2, 1, 3]),
            ('user', 'plays', 'game'): ([0, 0, 0, 1, 2], [0, 2, 3, 0, 2]),
            ('store', 'sells', 'game'): ([0, 0, 1, 1], [0, 3, 1, 2])},
            idtype=idtype, device=F.ctx())
        if not canonical_keys:
            conv = nn.HeteroGraphConv({
                'follows': nn.GraphConv(2, 3, allow_zero_in_degree=True),
                'plays': nn.GraphConv(2, 4, allow_zero_in_degree=True),
                'sells': nn.GraphConv(3, 4, allow_zero_in_degree=True)},
                agg)
        else:
            conv = nn.HeteroGraphConv({
                ('user', 'follows', 'user'): nn.GraphConv(2, 3, allow_zero_in_degree=True),
                ('user', 'plays', 'game'): nn.GraphConv(2, 4, allow_zero_in_degree=True),
                ('store', 'sells', 'game'): nn.GraphConv(3, 4, allow_zero_in_degree=True)},
                agg)
    
        conv = conv.to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        uf = F.randn((4, 2))
        gf = F.randn((4, 4))
        sf = F.randn((2, 3))
    
>       h = conv(g, {'user': uf, 'game': gf, 'store': sf})

tests\pytorch\test_nn.py:1178: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\hetero.py:205: in forward
    **mod_kwargs.get(etype, {}))
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:428: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE122F8648>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.6001,  0.6006, -0.1172],
        [ 0.5491, -0.4640,  0.7142]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
____________________ test_hetero_conv[False-myagg-idtype1] ____________________

agg = <function myagg at 0x000000AE126FB620>, idtype = torch.int64
canonical_keys = False

    @parametrize_idtype
    @pytest.mark.parametrize('agg', ['sum', 'max', 'min', 'mean', 'stack', myagg])
    @pytest.mark.parametrize('canonical_keys', [False, True])
    def test_hetero_conv(agg, idtype, canonical_keys):
        g = dgl.heterograph({
            ('user', 'follows', 'user'): ([0, 0, 2, 1], [1, 2, 1, 3]),
            ('user', 'plays', 'game'): ([0, 0, 0, 1, 2], [0, 2, 3, 0, 2]),
            ('store', 'sells', 'game'): ([0, 0, 1, 1], [0, 3, 1, 2])},
            idtype=idtype, device=F.ctx())
        if not canonical_keys:
            conv = nn.HeteroGraphConv({
                'follows': nn.GraphConv(2, 3, allow_zero_in_degree=True),
                'plays': nn.GraphConv(2, 4, allow_zero_in_degree=True),
                'sells': nn.GraphConv(3, 4, allow_zero_in_degree=True)},
                agg)
        else:
            conv = nn.HeteroGraphConv({
                ('user', 'follows', 'user'): nn.GraphConv(2, 3, allow_zero_in_degree=True),
                ('user', 'plays', 'game'): nn.GraphConv(2, 4, allow_zero_in_degree=True),
                ('store', 'sells', 'game'): nn.GraphConv(3, 4, allow_zero_in_degree=True)},
                agg)
    
        conv = conv.to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        uf = F.randn((4, 2))
        gf = F.randn((4, 4))
        sf = F.randn((2, 3))
    
>       h = conv(g, {'user': uf, 'game': gf, 'store': sf})

tests\pytorch\test_nn.py:1178: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\hetero.py:205: in forward
    **mod_kwargs.get(etype, {}))
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:428: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07974988>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.1008, -0.2457, -0.4536],
        [-0.7911, -0.0238,  0.3008]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_____________________ test_hetero_conv[True-sum-idtype0] ______________________

agg = 'sum', idtype = torch.int32, canonical_keys = True

    @parametrize_idtype
    @pytest.mark.parametrize('agg', ['sum', 'max', 'min', 'mean', 'stack', myagg])
    @pytest.mark.parametrize('canonical_keys', [False, True])
    def test_hetero_conv(agg, idtype, canonical_keys):
        g = dgl.heterograph({
            ('user', 'follows', 'user'): ([0, 0, 2, 1], [1, 2, 1, 3]),
            ('user', 'plays', 'game'): ([0, 0, 0, 1, 2], [0, 2, 3, 0, 2]),
            ('store', 'sells', 'game'): ([0, 0, 1, 1], [0, 3, 1, 2])},
            idtype=idtype, device=F.ctx())
        if not canonical_keys:
            conv = nn.HeteroGraphConv({
                'follows': nn.GraphConv(2, 3, allow_zero_in_degree=True),
                'plays': nn.GraphConv(2, 4, allow_zero_in_degree=True),
                'sells': nn.GraphConv(3, 4, allow_zero_in_degree=True)},
                agg)
        else:
            conv = nn.HeteroGraphConv({
                ('user', 'follows', 'user'): nn.GraphConv(2, 3, allow_zero_in_degree=True),
                ('user', 'plays', 'game'): nn.GraphConv(2, 4, allow_zero_in_degree=True),
                ('store', 'sells', 'game'): nn.GraphConv(3, 4, allow_zero_in_degree=True)},
                agg)
    
        conv = conv.to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        uf = F.randn((4, 2))
        gf = F.randn((4, 4))
        sf = F.randn((2, 3))
    
>       h = conv(g, {'user': uf, 'game': gf, 'store': sf})

tests\pytorch\test_nn.py:1178: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\hetero.py:205: in forward
    **mod_kwargs.get(etype, {}))
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:428: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07411208>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.1678, -0.5335,  0.0572],
        [ 0.6498, -0.7910, -1.5448]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_____________________ test_hetero_conv[True-sum-idtype1] ______________________

agg = 'sum', idtype = torch.int64, canonical_keys = True

    @parametrize_idtype
    @pytest.mark.parametrize('agg', ['sum', 'max', 'min', 'mean', 'stack', myagg])
    @pytest.mark.parametrize('canonical_keys', [False, True])
    def test_hetero_conv(agg, idtype, canonical_keys):
        g = dgl.heterograph({
            ('user', 'follows', 'user'): ([0, 0, 2, 1], [1, 2, 1, 3]),
            ('user', 'plays', 'game'): ([0, 0, 0, 1, 2], [0, 2, 3, 0, 2]),
            ('store', 'sells', 'game'): ([0, 0, 1, 1], [0, 3, 1, 2])},
            idtype=idtype, device=F.ctx())
        if not canonical_keys:
            conv = nn.HeteroGraphConv({
                'follows': nn.GraphConv(2, 3, allow_zero_in_degree=True),
                'plays': nn.GraphConv(2, 4, allow_zero_in_degree=True),
                'sells': nn.GraphConv(3, 4, allow_zero_in_degree=True)},
                agg)
        else:
            conv = nn.HeteroGraphConv({
                ('user', 'follows', 'user'): nn.GraphConv(2, 3, allow_zero_in_degree=True),
                ('user', 'plays', 'game'): nn.GraphConv(2, 4, allow_zero_in_degree=True),
                ('store', 'sells', 'game'): nn.GraphConv(3, 4, allow_zero_in_degree=True)},
                agg)
    
        conv = conv.to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        uf = F.randn((4, 2))
        gf = F.randn((4, 4))
        sf = F.randn((2, 3))
    
>       h = conv(g, {'user': uf, 'game': gf, 'store': sf})

tests\pytorch\test_nn.py:1178: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\hetero.py:205: in forward
    **mod_kwargs.get(etype, {}))
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:428: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF072B5D88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.5986,  0.0135, -0.9486],
        [-0.7777,  1.2475,  0.5733]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_____________________ test_hetero_conv[True-max-idtype0] ______________________

agg = 'max', idtype = torch.int32, canonical_keys = True

    @parametrize_idtype
    @pytest.mark.parametrize('agg', ['sum', 'max', 'min', 'mean', 'stack', myagg])
    @pytest.mark.parametrize('canonical_keys', [False, True])
    def test_hetero_conv(agg, idtype, canonical_keys):
        g = dgl.heterograph({
            ('user', 'follows', 'user'): ([0, 0, 2, 1], [1, 2, 1, 3]),
            ('user', 'plays', 'game'): ([0, 0, 0, 1, 2], [0, 2, 3, 0, 2]),
            ('store', 'sells', 'game'): ([0, 0, 1, 1], [0, 3, 1, 2])},
            idtype=idtype, device=F.ctx())
        if not canonical_keys:
            conv = nn.HeteroGraphConv({
                'follows': nn.GraphConv(2, 3, allow_zero_in_degree=True),
                'plays': nn.GraphConv(2, 4, allow_zero_in_degree=True),
                'sells': nn.GraphConv(3, 4, allow_zero_in_degree=True)},
                agg)
        else:
            conv = nn.HeteroGraphConv({
                ('user', 'follows', 'user'): nn.GraphConv(2, 3, allow_zero_in_degree=True),
                ('user', 'plays', 'game'): nn.GraphConv(2, 4, allow_zero_in_degree=True),
                ('store', 'sells', 'game'): nn.GraphConv(3, 4, allow_zero_in_degree=True)},
                agg)
    
        conv = conv.to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        uf = F.randn((4, 2))
        gf = F.randn((4, 4))
        sf = F.randn((2, 3))
    
>       h = conv(g, {'user': uf, 'game': gf, 'store': sf})

tests\pytorch\test_nn.py:1178: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\hetero.py:205: in forward
    **mod_kwargs.get(etype, {}))
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:428: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1225A808>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.7883, -1.6155, -0.2600],
        [-0.9716, -0.0626,  0.3580]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_____________________ test_hetero_conv[True-max-idtype1] ______________________

agg = 'max', idtype = torch.int64, canonical_keys = True

    @parametrize_idtype
    @pytest.mark.parametrize('agg', ['sum', 'max', 'min', 'mean', 'stack', myagg])
    @pytest.mark.parametrize('canonical_keys', [False, True])
    def test_hetero_conv(agg, idtype, canonical_keys):
        g = dgl.heterograph({
            ('user', 'follows', 'user'): ([0, 0, 2, 1], [1, 2, 1, 3]),
            ('user', 'plays', 'game'): ([0, 0, 0, 1, 2], [0, 2, 3, 0, 2]),
            ('store', 'sells', 'game'): ([0, 0, 1, 1], [0, 3, 1, 2])},
            idtype=idtype, device=F.ctx())
        if not canonical_keys:
            conv = nn.HeteroGraphConv({
                'follows': nn.GraphConv(2, 3, allow_zero_in_degree=True),
                'plays': nn.GraphConv(2, 4, allow_zero_in_degree=True),
                'sells': nn.GraphConv(3, 4, allow_zero_in_degree=True)},
                agg)
        else:
            conv = nn.HeteroGraphConv({
                ('user', 'follows', 'user'): nn.GraphConv(2, 3, allow_zero_in_degree=True),
                ('user', 'plays', 'game'): nn.GraphConv(2, 4, allow_zero_in_degree=True),
                ('store', 'sells', 'game'): nn.GraphConv(3, 4, allow_zero_in_degree=True)},
                agg)
    
        conv = conv.to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        uf = F.randn((4, 2))
        gf = F.randn((4, 4))
        sf = F.randn((2, 3))
    
>       h = conv(g, {'user': uf, 'game': gf, 'store': sf})

tests\pytorch\test_nn.py:1178: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\hetero.py:205: in forward
    **mod_kwargs.get(etype, {}))
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:428: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF071729C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.5541, -0.2472,  0.7340],
        [-0.1801,  0.2593, -1.6536]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_____________________ test_hetero_conv[True-min-idtype0] ______________________

agg = 'min', idtype = torch.int32, canonical_keys = True

    @parametrize_idtype
    @pytest.mark.parametrize('agg', ['sum', 'max', 'min', 'mean', 'stack', myagg])
    @pytest.mark.parametrize('canonical_keys', [False, True])
    def test_hetero_conv(agg, idtype, canonical_keys):
        g = dgl.heterograph({
            ('user', 'follows', 'user'): ([0, 0, 2, 1], [1, 2, 1, 3]),
            ('user', 'plays', 'game'): ([0, 0, 0, 1, 2], [0, 2, 3, 0, 2]),
            ('store', 'sells', 'game'): ([0, 0, 1, 1], [0, 3, 1, 2])},
            idtype=idtype, device=F.ctx())
        if not canonical_keys:
            conv = nn.HeteroGraphConv({
                'follows': nn.GraphConv(2, 3, allow_zero_in_degree=True),
                'plays': nn.GraphConv(2, 4, allow_zero_in_degree=True),
                'sells': nn.GraphConv(3, 4, allow_zero_in_degree=True)},
                agg)
        else:
            conv = nn.HeteroGraphConv({
                ('user', 'follows', 'user'): nn.GraphConv(2, 3, allow_zero_in_degree=True),
                ('user', 'plays', 'game'): nn.GraphConv(2, 4, allow_zero_in_degree=True),
                ('store', 'sells', 'game'): nn.GraphConv(3, 4, allow_zero_in_degree=True)},
                agg)
    
        conv = conv.to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        uf = F.randn((4, 2))
        gf = F.randn((4, 4))
        sf = F.randn((2, 3))
    
>       h = conv(g, {'user': uf, 'game': gf, 'store': sf})

tests\pytorch\test_nn.py:1178: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\hetero.py:205: in forward
    **mod_kwargs.get(etype, {}))
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:428: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07163548>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.0768, -0.2894,  1.5621],
        [ 1.3744,  0.1391,  0.1514]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_____________________ test_hetero_conv[True-min-idtype1] ______________________

agg = 'min', idtype = torch.int64, canonical_keys = True

    @parametrize_idtype
    @pytest.mark.parametrize('agg', ['sum', 'max', 'min', 'mean', 'stack', myagg])
    @pytest.mark.parametrize('canonical_keys', [False, True])
    def test_hetero_conv(agg, idtype, canonical_keys):
        g = dgl.heterograph({
            ('user', 'follows', 'user'): ([0, 0, 2, 1], [1, 2, 1, 3]),
            ('user', 'plays', 'game'): ([0, 0, 0, 1, 2], [0, 2, 3, 0, 2]),
            ('store', 'sells', 'game'): ([0, 0, 1, 1], [0, 3, 1, 2])},
            idtype=idtype, device=F.ctx())
        if not canonical_keys:
            conv = nn.HeteroGraphConv({
                'follows': nn.GraphConv(2, 3, allow_zero_in_degree=True),
                'plays': nn.GraphConv(2, 4, allow_zero_in_degree=True),
                'sells': nn.GraphConv(3, 4, allow_zero_in_degree=True)},
                agg)
        else:
            conv = nn.HeteroGraphConv({
                ('user', 'follows', 'user'): nn.GraphConv(2, 3, allow_zero_in_degree=True),
                ('user', 'plays', 'game'): nn.GraphConv(2, 4, allow_zero_in_degree=True),
                ('store', 'sells', 'game'): nn.GraphConv(3, 4, allow_zero_in_degree=True)},
                agg)
    
        conv = conv.to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        uf = F.randn((4, 2))
        gf = F.randn((4, 4))
        sf = F.randn((2, 3))
    
>       h = conv(g, {'user': uf, 'game': gf, 'store': sf})

tests\pytorch\test_nn.py:1178: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\hetero.py:205: in forward
    **mod_kwargs.get(etype, {}))
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:428: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE146AB3C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.9012, -0.1916, -1.3540],
        [-0.7802, -1.0068, -0.9238]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_____________________ test_hetero_conv[True-mean-idtype0] _____________________

agg = 'mean', idtype = torch.int32, canonical_keys = True

    @parametrize_idtype
    @pytest.mark.parametrize('agg', ['sum', 'max', 'min', 'mean', 'stack', myagg])
    @pytest.mark.parametrize('canonical_keys', [False, True])
    def test_hetero_conv(agg, idtype, canonical_keys):
        g = dgl.heterograph({
            ('user', 'follows', 'user'): ([0, 0, 2, 1], [1, 2, 1, 3]),
            ('user', 'plays', 'game'): ([0, 0, 0, 1, 2], [0, 2, 3, 0, 2]),
            ('store', 'sells', 'game'): ([0, 0, 1, 1], [0, 3, 1, 2])},
            idtype=idtype, device=F.ctx())
        if not canonical_keys:
            conv = nn.HeteroGraphConv({
                'follows': nn.GraphConv(2, 3, allow_zero_in_degree=True),
                'plays': nn.GraphConv(2, 4, allow_zero_in_degree=True),
                'sells': nn.GraphConv(3, 4, allow_zero_in_degree=True)},
                agg)
        else:
            conv = nn.HeteroGraphConv({
                ('user', 'follows', 'user'): nn.GraphConv(2, 3, allow_zero_in_degree=True),
                ('user', 'plays', 'game'): nn.GraphConv(2, 4, allow_zero_in_degree=True),
                ('store', 'sells', 'game'): nn.GraphConv(3, 4, allow_zero_in_degree=True)},
                agg)
    
        conv = conv.to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        uf = F.randn((4, 2))
        gf = F.randn((4, 4))
        sf = F.randn((2, 3))
    
>       h = conv(g, {'user': uf, 'game': gf, 'store': sf})

tests\pytorch\test_nn.py:1178: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\hetero.py:205: in forward
    **mod_kwargs.get(etype, {}))
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:428: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1226B8C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.1672,  1.3837, -1.2017],
        [ 0.2214, -0.7929,  0.1168]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_____________________ test_hetero_conv[True-mean-idtype1] _____________________

agg = 'mean', idtype = torch.int64, canonical_keys = True

    @parametrize_idtype
    @pytest.mark.parametrize('agg', ['sum', 'max', 'min', 'mean', 'stack', myagg])
    @pytest.mark.parametrize('canonical_keys', [False, True])
    def test_hetero_conv(agg, idtype, canonical_keys):
        g = dgl.heterograph({
            ('user', 'follows', 'user'): ([0, 0, 2, 1], [1, 2, 1, 3]),
            ('user', 'plays', 'game'): ([0, 0, 0, 1, 2], [0, 2, 3, 0, 2]),
            ('store', 'sells', 'game'): ([0, 0, 1, 1], [0, 3, 1, 2])},
            idtype=idtype, device=F.ctx())
        if not canonical_keys:
            conv = nn.HeteroGraphConv({
                'follows': nn.GraphConv(2, 3, allow_zero_in_degree=True),
                'plays': nn.GraphConv(2, 4, allow_zero_in_degree=True),
                'sells': nn.GraphConv(3, 4, allow_zero_in_degree=True)},
                agg)
        else:
            conv = nn.HeteroGraphConv({
                ('user', 'follows', 'user'): nn.GraphConv(2, 3, allow_zero_in_degree=True),
                ('user', 'plays', 'game'): nn.GraphConv(2, 4, allow_zero_in_degree=True),
                ('store', 'sells', 'game'): nn.GraphConv(3, 4, allow_zero_in_degree=True)},
                agg)
    
        conv = conv.to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        uf = F.randn((4, 2))
        gf = F.randn((4, 4))
        sf = F.randn((2, 3))
    
>       h = conv(g, {'user': uf, 'game': gf, 'store': sf})

tests\pytorch\test_nn.py:1178: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\hetero.py:205: in forward
    **mod_kwargs.get(etype, {}))
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:428: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03E1B908>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.1348,  1.4200,  0.4892],
        [-0.5011, -0.9187, -0.3083]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
____________________ test_hetero_conv[True-stack-idtype0] _____________________

agg = 'stack', idtype = torch.int32, canonical_keys = True

    @parametrize_idtype
    @pytest.mark.parametrize('agg', ['sum', 'max', 'min', 'mean', 'stack', myagg])
    @pytest.mark.parametrize('canonical_keys', [False, True])
    def test_hetero_conv(agg, idtype, canonical_keys):
        g = dgl.heterograph({
            ('user', 'follows', 'user'): ([0, 0, 2, 1], [1, 2, 1, 3]),
            ('user', 'plays', 'game'): ([0, 0, 0, 1, 2], [0, 2, 3, 0, 2]),
            ('store', 'sells', 'game'): ([0, 0, 1, 1], [0, 3, 1, 2])},
            idtype=idtype, device=F.ctx())
        if not canonical_keys:
            conv = nn.HeteroGraphConv({
                'follows': nn.GraphConv(2, 3, allow_zero_in_degree=True),
                'plays': nn.GraphConv(2, 4, allow_zero_in_degree=True),
                'sells': nn.GraphConv(3, 4, allow_zero_in_degree=True)},
                agg)
        else:
            conv = nn.HeteroGraphConv({
                ('user', 'follows', 'user'): nn.GraphConv(2, 3, allow_zero_in_degree=True),
                ('user', 'plays', 'game'): nn.GraphConv(2, 4, allow_zero_in_degree=True),
                ('store', 'sells', 'game'): nn.GraphConv(3, 4, allow_zero_in_degree=True)},
                agg)
    
        conv = conv.to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        uf = F.randn((4, 2))
        gf = F.randn((4, 4))
        sf = F.randn((2, 3))
    
>       h = conv(g, {'user': uf, 'game': gf, 'store': sf})

tests\pytorch\test_nn.py:1178: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\hetero.py:205: in forward
    **mod_kwargs.get(etype, {}))
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:428: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07960A08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.5550, -1.0242,  0.1190],
        [ 0.7740,  1.5207, -0.3976]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
____________________ test_hetero_conv[True-stack-idtype1] _____________________

agg = 'stack', idtype = torch.int64, canonical_keys = True

    @parametrize_idtype
    @pytest.mark.parametrize('agg', ['sum', 'max', 'min', 'mean', 'stack', myagg])
    @pytest.mark.parametrize('canonical_keys', [False, True])
    def test_hetero_conv(agg, idtype, canonical_keys):
        g = dgl.heterograph({
            ('user', 'follows', 'user'): ([0, 0, 2, 1], [1, 2, 1, 3]),
            ('user', 'plays', 'game'): ([0, 0, 0, 1, 2], [0, 2, 3, 0, 2]),
            ('store', 'sells', 'game'): ([0, 0, 1, 1], [0, 3, 1, 2])},
            idtype=idtype, device=F.ctx())
        if not canonical_keys:
            conv = nn.HeteroGraphConv({
                'follows': nn.GraphConv(2, 3, allow_zero_in_degree=True),
                'plays': nn.GraphConv(2, 4, allow_zero_in_degree=True),
                'sells': nn.GraphConv(3, 4, allow_zero_in_degree=True)},
                agg)
        else:
            conv = nn.HeteroGraphConv({
                ('user', 'follows', 'user'): nn.GraphConv(2, 3, allow_zero_in_degree=True),
                ('user', 'plays', 'game'): nn.GraphConv(2, 4, allow_zero_in_degree=True),
                ('store', 'sells', 'game'): nn.GraphConv(3, 4, allow_zero_in_degree=True)},
                agg)
    
        conv = conv.to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        uf = F.randn((4, 2))
        gf = F.randn((4, 4))
        sf = F.randn((2, 3))
    
>       h = conv(g, {'user': uf, 'game': gf, 'store': sf})

tests\pytorch\test_nn.py:1178: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\hetero.py:205: in forward
    **mod_kwargs.get(etype, {}))
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:428: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07988DC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.3978, -0.1320,  1.3649],
        [ 0.2784,  0.1259, -1.0251]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
____________________ test_hetero_conv[True-myagg-idtype0] _____________________

agg = <function myagg at 0x000000AE126FB620>, idtype = torch.int32
canonical_keys = True

    @parametrize_idtype
    @pytest.mark.parametrize('agg', ['sum', 'max', 'min', 'mean', 'stack', myagg])
    @pytest.mark.parametrize('canonical_keys', [False, True])
    def test_hetero_conv(agg, idtype, canonical_keys):
        g = dgl.heterograph({
            ('user', 'follows', 'user'): ([0, 0, 2, 1], [1, 2, 1, 3]),
            ('user', 'plays', 'game'): ([0, 0, 0, 1, 2], [0, 2, 3, 0, 2]),
            ('store', 'sells', 'game'): ([0, 0, 1, 1], [0, 3, 1, 2])},
            idtype=idtype, device=F.ctx())
        if not canonical_keys:
            conv = nn.HeteroGraphConv({
                'follows': nn.GraphConv(2, 3, allow_zero_in_degree=True),
                'plays': nn.GraphConv(2, 4, allow_zero_in_degree=True),
                'sells': nn.GraphConv(3, 4, allow_zero_in_degree=True)},
                agg)
        else:
            conv = nn.HeteroGraphConv({
                ('user', 'follows', 'user'): nn.GraphConv(2, 3, allow_zero_in_degree=True),
                ('user', 'plays', 'game'): nn.GraphConv(2, 4, allow_zero_in_degree=True),
                ('store', 'sells', 'game'): nn.GraphConv(3, 4, allow_zero_in_degree=True)},
                agg)
    
        conv = conv.to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        uf = F.randn((4, 2))
        gf = F.randn((4, 4))
        sf = F.randn((2, 3))
    
>       h = conv(g, {'user': uf, 'game': gf, 'store': sf})

tests\pytorch\test_nn.py:1178: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\hetero.py:205: in forward
    **mod_kwargs.get(etype, {}))
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:428: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03E26248>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.1605,  0.0172,  0.0549],
        [-0.2916, -0.1524, -0.9087]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
____________________ test_hetero_conv[True-myagg-idtype1] _____________________

agg = <function myagg at 0x000000AE126FB620>, idtype = torch.int64
canonical_keys = True

    @parametrize_idtype
    @pytest.mark.parametrize('agg', ['sum', 'max', 'min', 'mean', 'stack', myagg])
    @pytest.mark.parametrize('canonical_keys', [False, True])
    def test_hetero_conv(agg, idtype, canonical_keys):
        g = dgl.heterograph({
            ('user', 'follows', 'user'): ([0, 0, 2, 1], [1, 2, 1, 3]),
            ('user', 'plays', 'game'): ([0, 0, 0, 1, 2], [0, 2, 3, 0, 2]),
            ('store', 'sells', 'game'): ([0, 0, 1, 1], [0, 3, 1, 2])},
            idtype=idtype, device=F.ctx())
        if not canonical_keys:
            conv = nn.HeteroGraphConv({
                'follows': nn.GraphConv(2, 3, allow_zero_in_degree=True),
                'plays': nn.GraphConv(2, 4, allow_zero_in_degree=True),
                'sells': nn.GraphConv(3, 4, allow_zero_in_degree=True)},
                agg)
        else:
            conv = nn.HeteroGraphConv({
                ('user', 'follows', 'user'): nn.GraphConv(2, 3, allow_zero_in_degree=True),
                ('user', 'plays', 'game'): nn.GraphConv(2, 4, allow_zero_in_degree=True),
                ('store', 'sells', 'game'): nn.GraphConv(3, 4, allow_zero_in_degree=True)},
                agg)
    
        conv = conv.to(F.ctx())
    
        # test pickle
        th.save(conv, tmp_buffer)
    
        uf = F.randn((4, 2))
        gf = F.randn((4, 4))
        sf = F.randn((2, 3))
    
>       h = conv(g, {'user': uf, 'game': gf, 'store': sf})

tests\pytorch\test_nn.py:1178: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\hetero.py:205: in forward
    **mod_kwargs.get(etype, {}))
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:428: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE122F8D08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.5603,  0.0219, -0.7322],
        [ 0.1819, -0.7850, -0.3265]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_gnnexplainer[1-g0-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
idtype = torch.int32, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_gnnexplainer(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        feat = F.randn((g.num_nodes(), 5))
    
        class Model(th.nn.Module):
            def __init__(self, in_feats, out_feats, graph=False):
                super(Model, self).__init__()
                self.linear = th.nn.Linear(in_feats, out_feats)
                if graph:
                    self.pool = nn.AvgPooling()
                else:
                    self.pool = None
    
            def forward(self, graph, feat, eweight=None):
                with graph.local_scope():
                    feat = self.linear(feat)
                    graph.ndata['h'] = feat
                    if eweight is None:
                        graph.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
                    else:
                        graph.edata['w'] = eweight
                        graph.update_all(fn.u_mul_e('h', 'w', 'm'), fn.sum('m', 'h'))
    
                    if self.pool:
                        return self.pool(graph, graph.ndata['h'])
                    else:
                        return graph.ndata['h']
    
        # Explain node prediction
        model = Model(5, out_dim)
        model = model.to(F.ctx())
        explainer = nn.GNNExplainer(model, num_hops=1)
>       new_center, sg, feat_mask, edge_mask = explainer.explain_node(0, g, feat)

tests\pytorch\test_nn.py:1321: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
python\dgl\nn\pytorch\explain\gnnexplainer.py:275: in explain_node
    logits = self.model(graph=sg, feat=feat, **kwargs)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
tests\pytorch\test_nn.py:1307: in forward
    graph.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF078069C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.2950],
        [-0.1006],
        [ 0.4147]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_gnnexplainer[1-g0-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
idtype = torch.int64, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_gnnexplainer(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        feat = F.randn((g.num_nodes(), 5))
    
        class Model(th.nn.Module):
            def __init__(self, in_feats, out_feats, graph=False):
                super(Model, self).__init__()
                self.linear = th.nn.Linear(in_feats, out_feats)
                if graph:
                    self.pool = nn.AvgPooling()
                else:
                    self.pool = None
    
            def forward(self, graph, feat, eweight=None):
                with graph.local_scope():
                    feat = self.linear(feat)
                    graph.ndata['h'] = feat
                    if eweight is None:
                        graph.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
                    else:
                        graph.edata['w'] = eweight
                        graph.update_all(fn.u_mul_e('h', 'w', 'm'), fn.sum('m', 'h'))
    
                    if self.pool:
                        return self.pool(graph, graph.ndata['h'])
                    else:
                        return graph.ndata['h']
    
        # Explain node prediction
        model = Model(5, out_dim)
        model = model.to(F.ctx())
        explainer = nn.GNNExplainer(model, num_hops=1)
>       new_center, sg, feat_mask, edge_mask = explainer.explain_node(0, g, feat)

tests\pytorch\test_nn.py:1321: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
python\dgl\nn\pytorch\explain\gnnexplainer.py:275: in explain_node
    logits = self.model(graph=sg, feat=feat, **kwargs)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
tests\pytorch\test_nn.py:1307: in forward
    graph.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF077AE308>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0.6360],
        [0.3688],
        [0.3020]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_gnnexplainer[1-g1-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={}
      edata_schemes={})
idtype = torch.int32, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_gnnexplainer(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        feat = F.randn((g.num_nodes(), 5))
    
        class Model(th.nn.Module):
            def __init__(self, in_feats, out_feats, graph=False):
                super(Model, self).__init__()
                self.linear = th.nn.Linear(in_feats, out_feats)
                if graph:
                    self.pool = nn.AvgPooling()
                else:
                    self.pool = None
    
            def forward(self, graph, feat, eweight=None):
                with graph.local_scope():
                    feat = self.linear(feat)
                    graph.ndata['h'] = feat
                    if eweight is None:
                        graph.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
                    else:
                        graph.edata['w'] = eweight
                        graph.update_all(fn.u_mul_e('h', 'w', 'm'), fn.sum('m', 'h'))
    
                    if self.pool:
                        return self.pool(graph, graph.ndata['h'])
                    else:
                        return graph.ndata['h']
    
        # Explain node prediction
        model = Model(5, out_dim)
        model = model.to(F.ctx())
        explainer = nn.GNNExplainer(model, num_hops=1)
>       new_center, sg, feat_mask, edge_mask = explainer.explain_node(0, g, feat)

tests\pytorch\test_nn.py:1321: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
python\dgl\nn\pytorch\explain\gnnexplainer.py:275: in explain_node
    logits = self.model(graph=sg, feat=feat, **kwargs)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
tests\pytorch\test_nn.py:1307: in forward
    graph.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE121FD088>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.3144],
        [-0.7383],
        [ 0.2948]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_gnnexplainer[1-g1-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={}
      edata_schemes={})
idtype = torch.int64, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_gnnexplainer(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        feat = F.randn((g.num_nodes(), 5))
    
        class Model(th.nn.Module):
            def __init__(self, in_feats, out_feats, graph=False):
                super(Model, self).__init__()
                self.linear = th.nn.Linear(in_feats, out_feats)
                if graph:
                    self.pool = nn.AvgPooling()
                else:
                    self.pool = None
    
            def forward(self, graph, feat, eweight=None):
                with graph.local_scope():
                    feat = self.linear(feat)
                    graph.ndata['h'] = feat
                    if eweight is None:
                        graph.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
                    else:
                        graph.edata['w'] = eweight
                        graph.update_all(fn.u_mul_e('h', 'w', 'm'), fn.sum('m', 'h'))
    
                    if self.pool:
                        return self.pool(graph, graph.ndata['h'])
                    else:
                        return graph.ndata['h']
    
        # Explain node prediction
        model = Model(5, out_dim)
        model = model.to(F.ctx())
        explainer = nn.GNNExplainer(model, num_hops=1)
>       new_center, sg, feat_mask, edge_mask = explainer.explain_node(0, g, feat)

tests\pytorch\test_nn.py:1321: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
python\dgl\nn\pytorch\explain\gnnexplainer.py:275: in explain_node
    logits = self.model(graph=sg, feat=feat, **kwargs)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
tests\pytorch\test_nn.py:1307: in forward
    graph.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0749A0C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.9873],
        [-0.5664],
        [-0.3057]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_gnnexplainer[1-g2-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int32, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_gnnexplainer(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        feat = F.randn((g.num_nodes(), 5))
    
        class Model(th.nn.Module):
            def __init__(self, in_feats, out_feats, graph=False):
                super(Model, self).__init__()
                self.linear = th.nn.Linear(in_feats, out_feats)
                if graph:
                    self.pool = nn.AvgPooling()
                else:
                    self.pool = None
    
            def forward(self, graph, feat, eweight=None):
                with graph.local_scope():
                    feat = self.linear(feat)
                    graph.ndata['h'] = feat
                    if eweight is None:
                        graph.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
                    else:
                        graph.edata['w'] = eweight
                        graph.update_all(fn.u_mul_e('h', 'w', 'm'), fn.sum('m', 'h'))
    
                    if self.pool:
                        return self.pool(graph, graph.ndata['h'])
                    else:
                        return graph.ndata['h']
    
        # Explain node prediction
        model = Model(5, out_dim)
        model = model.to(F.ctx())
        explainer = nn.GNNExplainer(model, num_hops=1)
>       new_center, sg, feat_mask, edge_mask = explainer.explain_node(0, g, feat)

tests\pytorch\test_nn.py:1321: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
python\dgl\nn\pytorch\explain\gnnexplainer.py:275: in explain_node
    logits = self.model(graph=sg, feat=feat, **kwargs)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
tests\pytorch\test_nn.py:1307: in forward
    graph.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1230DCC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.2969],
        [ 0.4126],
        [-0.0673]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_gnnexplainer[1-g2-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int64, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_gnnexplainer(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        feat = F.randn((g.num_nodes(), 5))
    
        class Model(th.nn.Module):
            def __init__(self, in_feats, out_feats, graph=False):
                super(Model, self).__init__()
                self.linear = th.nn.Linear(in_feats, out_feats)
                if graph:
                    self.pool = nn.AvgPooling()
                else:
                    self.pool = None
    
            def forward(self, graph, feat, eweight=None):
                with graph.local_scope():
                    feat = self.linear(feat)
                    graph.ndata['h'] = feat
                    if eweight is None:
                        graph.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
                    else:
                        graph.edata['w'] = eweight
                        graph.update_all(fn.u_mul_e('h', 'w', 'm'), fn.sum('m', 'h'))
    
                    if self.pool:
                        return self.pool(graph, graph.ndata['h'])
                    else:
                        return graph.ndata['h']
    
        # Explain node prediction
        model = Model(5, out_dim)
        model = model.to(F.ctx())
        explainer = nn.GNNExplainer(model, num_hops=1)
>       new_center, sg, feat_mask, edge_mask = explainer.explain_node(0, g, feat)

tests\pytorch\test_nn.py:1321: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
python\dgl\nn\pytorch\explain\gnnexplainer.py:275: in explain_node
    logits = self.model(graph=sg, feat=feat, **kwargs)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
tests\pytorch\test_nn.py:1307: in forward
    graph.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07810808>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[1.2024],
        [1.0371],
        [0.8049]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_gnnexplainer[1-g3-idtype0] _______________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={}
      edata_schemes={})
idtype = torch.int32, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_gnnexplainer(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        feat = F.randn((g.num_nodes(), 5))
    
        class Model(th.nn.Module):
            def __init__(self, in_feats, out_feats, graph=False):
                super(Model, self).__init__()
                self.linear = th.nn.Linear(in_feats, out_feats)
                if graph:
                    self.pool = nn.AvgPooling()
                else:
                    self.pool = None
    
            def forward(self, graph, feat, eweight=None):
                with graph.local_scope():
                    feat = self.linear(feat)
                    graph.ndata['h'] = feat
                    if eweight is None:
                        graph.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
                    else:
                        graph.edata['w'] = eweight
                        graph.update_all(fn.u_mul_e('h', 'w', 'm'), fn.sum('m', 'h'))
    
                    if self.pool:
                        return self.pool(graph, graph.ndata['h'])
                    else:
                        return graph.ndata['h']
    
        # Explain node prediction
        model = Model(5, out_dim)
        model = model.to(F.ctx())
        explainer = nn.GNNExplainer(model, num_hops=1)
>       new_center, sg, feat_mask, edge_mask = explainer.explain_node(0, g, feat)

tests\pytorch\test_nn.py:1321: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
python\dgl\nn\pytorch\explain\gnnexplainer.py:275: in explain_node
    logits = self.model(graph=sg, feat=feat, **kwargs)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
tests\pytorch\test_nn.py:1307: in forward
    graph.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07832588>
op = 'copy_lhs', reduce_op = 'sum', lhs_data = tensor([[0.3815]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_gnnexplainer[1-g3-idtype1] _______________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={}
      edata_schemes={})
idtype = torch.int64, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_gnnexplainer(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        feat = F.randn((g.num_nodes(), 5))
    
        class Model(th.nn.Module):
            def __init__(self, in_feats, out_feats, graph=False):
                super(Model, self).__init__()
                self.linear = th.nn.Linear(in_feats, out_feats)
                if graph:
                    self.pool = nn.AvgPooling()
                else:
                    self.pool = None
    
            def forward(self, graph, feat, eweight=None):
                with graph.local_scope():
                    feat = self.linear(feat)
                    graph.ndata['h'] = feat
                    if eweight is None:
                        graph.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
                    else:
                        graph.edata['w'] = eweight
                        graph.update_all(fn.u_mul_e('h', 'w', 'm'), fn.sum('m', 'h'))
    
                    if self.pool:
                        return self.pool(graph, graph.ndata['h'])
                    else:
                        return graph.ndata['h']
    
        # Explain node prediction
        model = Model(5, out_dim)
        model = model.to(F.ctx())
        explainer = nn.GNNExplainer(model, num_hops=1)
>       new_center, sg, feat_mask, edge_mask = explainer.explain_node(0, g, feat)

tests\pytorch\test_nn.py:1321: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
python\dgl\nn\pytorch\explain\gnnexplainer.py:275: in explain_node
    logits = self.model(graph=sg, feat=feat, **kwargs)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
tests\pytorch\test_nn.py:1307: in forward
    graph.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF077F7B48>
op = 'copy_lhs', reduce_op = 'sum', lhs_data = tensor([[-0.4402]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_gnnexplainer[1-g4-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={}
      edata_schemes={})
idtype = torch.int32, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_gnnexplainer(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        feat = F.randn((g.num_nodes(), 5))
    
        class Model(th.nn.Module):
            def __init__(self, in_feats, out_feats, graph=False):
                super(Model, self).__init__()
                self.linear = th.nn.Linear(in_feats, out_feats)
                if graph:
                    self.pool = nn.AvgPooling()
                else:
                    self.pool = None
    
            def forward(self, graph, feat, eweight=None):
                with graph.local_scope():
                    feat = self.linear(feat)
                    graph.ndata['h'] = feat
                    if eweight is None:
                        graph.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
                    else:
                        graph.edata['w'] = eweight
                        graph.update_all(fn.u_mul_e('h', 'w', 'm'), fn.sum('m', 'h'))
    
                    if self.pool:
                        return self.pool(graph, graph.ndata['h'])
                    else:
                        return graph.ndata['h']
    
        # Explain node prediction
        model = Model(5, out_dim)
        model = model.to(F.ctx())
        explainer = nn.GNNExplainer(model, num_hops=1)
>       new_center, sg, feat_mask, edge_mask = explainer.explain_node(0, g, feat)

tests\pytorch\test_nn.py:1321: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
python\dgl\nn\pytorch\explain\gnnexplainer.py:275: in explain_node
    logits = self.model(graph=sg, feat=feat, **kwargs)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
tests\pytorch\test_nn.py:1307: in forward
    graph.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03CF8A48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.1401],
        [-0.6405],
        [ 0.1352]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_gnnexplainer[1-g4-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={}
      edata_schemes={})
idtype = torch.int64, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_gnnexplainer(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        feat = F.randn((g.num_nodes(), 5))
    
        class Model(th.nn.Module):
            def __init__(self, in_feats, out_feats, graph=False):
                super(Model, self).__init__()
                self.linear = th.nn.Linear(in_feats, out_feats)
                if graph:
                    self.pool = nn.AvgPooling()
                else:
                    self.pool = None
    
            def forward(self, graph, feat, eweight=None):
                with graph.local_scope():
                    feat = self.linear(feat)
                    graph.ndata['h'] = feat
                    if eweight is None:
                        graph.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
                    else:
                        graph.edata['w'] = eweight
                        graph.update_all(fn.u_mul_e('h', 'w', 'm'), fn.sum('m', 'h'))
    
                    if self.pool:
                        return self.pool(graph, graph.ndata['h'])
                    else:
                        return graph.ndata['h']
    
        # Explain node prediction
        model = Model(5, out_dim)
        model = model.to(F.ctx())
        explainer = nn.GNNExplainer(model, num_hops=1)
>       new_center, sg, feat_mask, edge_mask = explainer.explain_node(0, g, feat)

tests\pytorch\test_nn.py:1321: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
python\dgl\nn\pytorch\explain\gnnexplainer.py:275: in explain_node
    logits = self.model(graph=sg, feat=feat, **kwargs)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
tests\pytorch\test_nn.py:1307: in forward
    graph.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07991588>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 1.0955],
        [ 0.5437],
        [-0.7968]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_gnnexplainer[1-g5-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={}
      edata_schemes={})
idtype = torch.int32, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_gnnexplainer(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        feat = F.randn((g.num_nodes(), 5))
    
        class Model(th.nn.Module):
            def __init__(self, in_feats, out_feats, graph=False):
                super(Model, self).__init__()
                self.linear = th.nn.Linear(in_feats, out_feats)
                if graph:
                    self.pool = nn.AvgPooling()
                else:
                    self.pool = None
    
            def forward(self, graph, feat, eweight=None):
                with graph.local_scope():
                    feat = self.linear(feat)
                    graph.ndata['h'] = feat
                    if eweight is None:
                        graph.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
                    else:
                        graph.edata['w'] = eweight
                        graph.update_all(fn.u_mul_e('h', 'w', 'm'), fn.sum('m', 'h'))
    
                    if self.pool:
                        return self.pool(graph, graph.ndata['h'])
                    else:
                        return graph.ndata['h']
    
        # Explain node prediction
        model = Model(5, out_dim)
        model = model.to(F.ctx())
        explainer = nn.GNNExplainer(model, num_hops=1)
>       new_center, sg, feat_mask, edge_mask = explainer.explain_node(0, g, feat)

tests\pytorch\test_nn.py:1321: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
python\dgl\nn\pytorch\explain\gnnexplainer.py:275: in explain_node
    logits = self.model(graph=sg, feat=feat, **kwargs)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
tests\pytorch\test_nn.py:1307: in forward
    graph.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07770908>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0.0682],
        [0.1455],
        [0.4357]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_gnnexplainer[1-g5-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={}
      edata_schemes={})
idtype = torch.int64, out_dim = 1

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_gnnexplainer(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        feat = F.randn((g.num_nodes(), 5))
    
        class Model(th.nn.Module):
            def __init__(self, in_feats, out_feats, graph=False):
                super(Model, self).__init__()
                self.linear = th.nn.Linear(in_feats, out_feats)
                if graph:
                    self.pool = nn.AvgPooling()
                else:
                    self.pool = None
    
            def forward(self, graph, feat, eweight=None):
                with graph.local_scope():
                    feat = self.linear(feat)
                    graph.ndata['h'] = feat
                    if eweight is None:
                        graph.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
                    else:
                        graph.edata['w'] = eweight
                        graph.update_all(fn.u_mul_e('h', 'w', 'm'), fn.sum('m', 'h'))
    
                    if self.pool:
                        return self.pool(graph, graph.ndata['h'])
                    else:
                        return graph.ndata['h']
    
        # Explain node prediction
        model = Model(5, out_dim)
        model = model.to(F.ctx())
        explainer = nn.GNNExplainer(model, num_hops=1)
>       new_center, sg, feat_mask, edge_mask = explainer.explain_node(0, g, feat)

tests\pytorch\test_nn.py:1321: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
python\dgl\nn\pytorch\explain\gnnexplainer.py:275: in explain_node
    logits = self.model(graph=sg, feat=feat, **kwargs)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
tests\pytorch\test_nn.py:1307: in forward
    graph.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07078C08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[1.0359],
        [0.2529],
        [0.1073]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_gnnexplainer[2-g0-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
idtype = torch.int32, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_gnnexplainer(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        feat = F.randn((g.num_nodes(), 5))
    
        class Model(th.nn.Module):
            def __init__(self, in_feats, out_feats, graph=False):
                super(Model, self).__init__()
                self.linear = th.nn.Linear(in_feats, out_feats)
                if graph:
                    self.pool = nn.AvgPooling()
                else:
                    self.pool = None
    
            def forward(self, graph, feat, eweight=None):
                with graph.local_scope():
                    feat = self.linear(feat)
                    graph.ndata['h'] = feat
                    if eweight is None:
                        graph.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
                    else:
                        graph.edata['w'] = eweight
                        graph.update_all(fn.u_mul_e('h', 'w', 'm'), fn.sum('m', 'h'))
    
                    if self.pool:
                        return self.pool(graph, graph.ndata['h'])
                    else:
                        return graph.ndata['h']
    
        # Explain node prediction
        model = Model(5, out_dim)
        model = model.to(F.ctx())
        explainer = nn.GNNExplainer(model, num_hops=1)
>       new_center, sg, feat_mask, edge_mask = explainer.explain_node(0, g, feat)

tests\pytorch\test_nn.py:1321: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
python\dgl\nn\pytorch\explain\gnnexplainer.py:275: in explain_node
    logits = self.model(graph=sg, feat=feat, **kwargs)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
tests\pytorch\test_nn.py:1307: in forward
    graph.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF073AAFC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.1020, -0.6491],
        [-0.4660,  0.1965],
        [ 0.2674,  1.0945]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_gnnexplainer[2-g0-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32)})
idtype = torch.int64, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_gnnexplainer(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        feat = F.randn((g.num_nodes(), 5))
    
        class Model(th.nn.Module):
            def __init__(self, in_feats, out_feats, graph=False):
                super(Model, self).__init__()
                self.linear = th.nn.Linear(in_feats, out_feats)
                if graph:
                    self.pool = nn.AvgPooling()
                else:
                    self.pool = None
    
            def forward(self, graph, feat, eweight=None):
                with graph.local_scope():
                    feat = self.linear(feat)
                    graph.ndata['h'] = feat
                    if eweight is None:
                        graph.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
                    else:
                        graph.edata['w'] = eweight
                        graph.update_all(fn.u_mul_e('h', 'w', 'm'), fn.sum('m', 'h'))
    
                    if self.pool:
                        return self.pool(graph, graph.ndata['h'])
                    else:
                        return graph.ndata['h']
    
        # Explain node prediction
        model = Model(5, out_dim)
        model = model.to(F.ctx())
        explainer = nn.GNNExplainer(model, num_hops=1)
>       new_center, sg, feat_mask, edge_mask = explainer.explain_node(0, g, feat)

tests\pytorch\test_nn.py:1321: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
python\dgl\nn\pytorch\explain\gnnexplainer.py:275: in explain_node
    logits = self.model(graph=sg, feat=feat, **kwargs)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
tests\pytorch\test_nn.py:1307: in forward
    graph.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all

python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07948288>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.4145,  0.2244],
        [-0.7063,  0.4699],
        [-1.1474,  1.1299]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_gnnexplainer[2-g1-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={}
      edata_schemes={})
idtype = torch.int32, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_gnnexplainer(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        feat = F.randn((g.num_nodes(), 5))
    
        class Model(th.nn.Module):
            def __init__(self, in_feats, out_feats, graph=False):
                super(Model, self).__init__()
                self.linear = th.nn.Linear(in_feats, out_feats)
                if graph:
                    self.pool = nn.AvgPooling()
                else:
                    self.pool = None
    
            def forward(self, graph, feat, eweight=None):
                with graph.local_scope():
                    feat = self.linear(feat)
                    graph.ndata['h'] = feat
                    if eweight is None:
                        graph.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
                    else:
                        graph.edata['w'] = eweight
                        graph.update_all(fn.u_mul_e('h', 'w', 'm'), fn.sum('m', 'h'))
    
                    if self.pool:
                        return self.pool(graph, graph.ndata['h'])
                    else:
                        return graph.ndata['h']
    
        # Explain node prediction
        model = Model(5, out_dim)
        model = model.to(F.ctx())
        explainer = nn.GNNExplainer(model, num_hops=1)
>       new_center, sg, feat_mask, edge_mask = explainer.explain_node(0, g, feat)

tests\pytorch\test_nn.py:1321: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
python\dgl\nn\pytorch\explain\gnnexplainer.py:275: in explain_node
    logits = self.model(graph=sg, feat=feat, **kwargs)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
tests\pytorch\test_nn.py:1307: in forward
    graph.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0749C648>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.1387, -0.0750],
        [ 0.0180,  0.0081],
        [ 1.2263, -1.0636]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_gnnexplainer[2-g1-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={}
      edata_schemes={})
idtype = torch.int64, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_gnnexplainer(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        feat = F.randn((g.num_nodes(), 5))
    
        class Model(th.nn.Module):
            def __init__(self, in_feats, out_feats, graph=False):
                super(Model, self).__init__()
                self.linear = th.nn.Linear(in_feats, out_feats)
                if graph:
                    self.pool = nn.AvgPooling()
                else:
                    self.pool = None
    
            def forward(self, graph, feat, eweight=None):
                with graph.local_scope():
                    feat = self.linear(feat)
                    graph.ndata['h'] = feat
                    if eweight is None:
                        graph.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
                    else:
                        graph.edata['w'] = eweight
                        graph.update_all(fn.u_mul_e('h', 'w', 'm'), fn.sum('m', 'h'))
    
                    if self.pool:
                        return self.pool(graph, graph.ndata['h'])
                    else:
                        return graph.ndata['h']
    
        # Explain node prediction
        model = Model(5, out_dim)
        model = model.to(F.ctx())
        explainer = nn.GNNExplainer(model, num_hops=1)
>       new_center, sg, feat_mask, edge_mask = explainer.explain_node(0, g, feat)

tests\pytorch\test_nn.py:1321: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
python\dgl\nn\pytorch\explain\gnnexplainer.py:275: in explain_node
    logits = self.model(graph=sg, feat=feat, **kwargs)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
tests\pytorch\test_nn.py:1307: in forward
    graph.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE123A4988>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.7391, -0.7305],
        [ 0.7804,  0.4287],
        [ 0.9455, -0.2275]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_gnnexplainer[2-g2-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int32, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_gnnexplainer(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        feat = F.randn((g.num_nodes(), 5))
    
        class Model(th.nn.Module):
            def __init__(self, in_feats, out_feats, graph=False):
                super(Model, self).__init__()
                self.linear = th.nn.Linear(in_feats, out_feats)
                if graph:
                    self.pool = nn.AvgPooling()
                else:
                    self.pool = None
    
            def forward(self, graph, feat, eweight=None):
                with graph.local_scope():
                    feat = self.linear(feat)
                    graph.ndata['h'] = feat
                    if eweight is None:
                        graph.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
                    else:
                        graph.edata['w'] = eweight
                        graph.update_all(fn.u_mul_e('h', 'w', 'm'), fn.sum('m', 'h'))
    
                    if self.pool:
                        return self.pool(graph, graph.ndata['h'])
                    else:
                        return graph.ndata['h']
    
        # Explain node prediction
        model = Model(5, out_dim)
        model = model.to(F.ctx())
        explainer = nn.GNNExplainer(model, num_hops=1)
>       new_center, sg, feat_mask, edge_mask = explainer.explain_node(0, g, feat)

tests\pytorch\test_nn.py:1321: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
python\dgl\nn\pytorch\explain\gnnexplainer.py:275: in explain_node
    logits = self.model(graph=sg, feat=feat, **kwargs)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
tests\pytorch\test_nn.py:1307: in forward
    graph.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07979788>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.6991, -0.0329],
        [ 0.7869, -0.1756],
        [ 0.4590, -1.3577]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_gnnexplainer[2-g2-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})
idtype = torch.int64, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_gnnexplainer(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        feat = F.randn((g.num_nodes(), 5))
    
        class Model(th.nn.Module):
            def __init__(self, in_feats, out_feats, graph=False):
                super(Model, self).__init__()
                self.linear = th.nn.Linear(in_feats, out_feats)
                if graph:
                    self.pool = nn.AvgPooling()
                else:
                    self.pool = None
    
            def forward(self, graph, feat, eweight=None):
                with graph.local_scope():
                    feat = self.linear(feat)
                    graph.ndata['h'] = feat
                    if eweight is None:
                        graph.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
                    else:
                        graph.edata['w'] = eweight
                        graph.update_all(fn.u_mul_e('h', 'w', 'm'), fn.sum('m', 'h'))
    
                    if self.pool:
                        return self.pool(graph, graph.ndata['h'])
                    else:
                        return graph.ndata['h']
    
        # Explain node prediction
        model = Model(5, out_dim)
        model = model.to(F.ctx())
        explainer = nn.GNNExplainer(model, num_hops=1)
>       new_center, sg, feat_mask, edge_mask = explainer.explain_node(0, g, feat)

tests\pytorch\test_nn.py:1321: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
python\dgl\nn\pytorch\explain\gnnexplainer.py:275: in explain_node
    logits = self.model(graph=sg, feat=feat, **kwargs)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
tests\pytorch\test_nn.py:1307: in forward
    graph.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF079D86C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.6777,  1.9223],
        [-0.1629, -0.5304],
        [ 0.4873,  0.6564]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_gnnexplainer[2-g3-idtype0] _______________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={}
      edata_schemes={})
idtype = torch.int32, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_gnnexplainer(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        feat = F.randn((g.num_nodes(), 5))
    
        class Model(th.nn.Module):
            def __init__(self, in_feats, out_feats, graph=False):
                super(Model, self).__init__()
                self.linear = th.nn.Linear(in_feats, out_feats)
                if graph:
                    self.pool = nn.AvgPooling()
                else:
                    self.pool = None
    
            def forward(self, graph, feat, eweight=None):
                with graph.local_scope():
                    feat = self.linear(feat)
                    graph.ndata['h'] = feat
                    if eweight is None:
                        graph.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
                    else:
                        graph.edata['w'] = eweight
                        graph.update_all(fn.u_mul_e('h', 'w', 'm'), fn.sum('m', 'h'))
    
                    if self.pool:
                        return self.pool(graph, graph.ndata['h'])
                    else:
                        return graph.ndata['h']
    
        # Explain node prediction
        model = Model(5, out_dim)
        model = model.to(F.ctx())
        explainer = nn.GNNExplainer(model, num_hops=1)
>       new_center, sg, feat_mask, edge_mask = explainer.explain_node(0, g, feat)

tests\pytorch\test_nn.py:1321: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
python\dgl\nn\pytorch\explain\gnnexplainer.py:275: in explain_node
    logits = self.model(graph=sg, feat=feat, **kwargs)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
tests\pytorch\test_nn.py:1307: in forward
    graph.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF074CCE08>
op = 'copy_lhs', reduce_op = 'sum', lhs_data = tensor([[ 0.8349, -0.4577]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_gnnexplainer[2-g3-idtype1] _______________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={}
      edata_schemes={})
idtype = torch.int64, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_gnnexplainer(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        feat = F.randn((g.num_nodes(), 5))
    
        class Model(th.nn.Module):
            def __init__(self, in_feats, out_feats, graph=False):
                super(Model, self).__init__()
                self.linear = th.nn.Linear(in_feats, out_feats)
                if graph:
                    self.pool = nn.AvgPooling()
                else:
                    self.pool = None
    
            def forward(self, graph, feat, eweight=None):
                with graph.local_scope():
                    feat = self.linear(feat)
                    graph.ndata['h'] = feat
                    if eweight is None:
                        graph.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
                    else:
                        graph.edata['w'] = eweight
                        graph.update_all(fn.u_mul_e('h', 'w', 'm'), fn.sum('m', 'h'))
    
                    if self.pool:
                        return self.pool(graph, graph.ndata['h'])
                    else:
                        return graph.ndata['h']
    
        # Explain node prediction
        model = Model(5, out_dim)
        model = model.to(F.ctx())
        explainer = nn.GNNExplainer(model, num_hops=1)
>       new_center, sg, feat_mask, edge_mask = explainer.explain_node(0, g, feat)

tests\pytorch\test_nn.py:1321: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
python\dgl\nn\pytorch\explain\gnnexplainer.py:275: in explain_node
    logits = self.model(graph=sg, feat=feat, **kwargs)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
tests\pytorch\test_nn.py:1307: in forward
    graph.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE122C5408>
op = 'copy_lhs', reduce_op = 'sum', lhs_data = tensor([[ 0.4103, -0.3403]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_gnnexplainer[2-g4-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={}
      edata_schemes={})
idtype = torch.int32, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_gnnexplainer(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        feat = F.randn((g.num_nodes(), 5))
    
        class Model(th.nn.Module):
            def __init__(self, in_feats, out_feats, graph=False):
                super(Model, self).__init__()
                self.linear = th.nn.Linear(in_feats, out_feats)
                if graph:
                    self.pool = nn.AvgPooling()
                else:
                    self.pool = None
    
            def forward(self, graph, feat, eweight=None):
                with graph.local_scope():
                    feat = self.linear(feat)
                    graph.ndata['h'] = feat
                    if eweight is None:
                        graph.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
                    else:
                        graph.edata['w'] = eweight
                        graph.update_all(fn.u_mul_e('h', 'w', 'm'), fn.sum('m', 'h'))
    
                    if self.pool:
                        return self.pool(graph, graph.ndata['h'])
                    else:
                        return graph.ndata['h']
    
        # Explain node prediction
        model = Model(5, out_dim)
        model = model.to(F.ctx())
        explainer = nn.GNNExplainer(model, num_hops=1)
>       new_center, sg, feat_mask, edge_mask = explainer.explain_node(0, g, feat)

tests\pytorch\test_nn.py:1321: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
python\dgl\nn\pytorch\explain\gnnexplainer.py:275: in explain_node
    logits = self.model(graph=sg, feat=feat, **kwargs)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
tests\pytorch\test_nn.py:1307: in forward
    graph.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF079E2208>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.2174,  0.3459],
        [ 0.0814, -0.0709],
        [-0.4938,  0.4150]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_gnnexplainer[2-g4-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={}
      edata_schemes={})
idtype = torch.int64, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_gnnexplainer(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        feat = F.randn((g.num_nodes(), 5))
    
        class Model(th.nn.Module):
            def __init__(self, in_feats, out_feats, graph=False):
                super(Model, self).__init__()
                self.linear = th.nn.Linear(in_feats, out_feats)
                if graph:
                    self.pool = nn.AvgPooling()
                else:
                    self.pool = None
    
            def forward(self, graph, feat, eweight=None):
                with graph.local_scope():
                    feat = self.linear(feat)
                    graph.ndata['h'] = feat
                    if eweight is None:
                        graph.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
                    else:
                        graph.edata['w'] = eweight
                        graph.update_all(fn.u_mul_e('h', 'w', 'm'), fn.sum('m', 'h'))
    
                    if self.pool:
                        return self.pool(graph, graph.ndata['h'])
                    else:
                        return graph.ndata['h']
    
        # Explain node prediction
        model = Model(5, out_dim)
        model = model.to(F.ctx())
        explainer = nn.GNNExplainer(model, num_hops=1)
>       new_center, sg, feat_mask, edge_mask = explainer.explain_node(0, g, feat)

tests\pytorch\test_nn.py:1321: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
python\dgl\nn\pytorch\explain\gnnexplainer.py:275: in explain_node
    logits = self.model(graph=sg, feat=feat, **kwargs)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
tests\pytorch\test_nn.py:1307: in forward
    graph.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07A10E48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.8571,  0.1137],
        [ 0.0281,  1.6331],
        [-0.1166,  0.0753]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_gnnexplainer[2-g5-idtype0] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={}
      edata_schemes={})
idtype = torch.int32, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_gnnexplainer(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        feat = F.randn((g.num_nodes(), 5))
    
        class Model(th.nn.Module):
            def __init__(self, in_feats, out_feats, graph=False):
                super(Model, self).__init__()
                self.linear = th.nn.Linear(in_feats, out_feats)
                if graph:
                    self.pool = nn.AvgPooling()
                else:
                    self.pool = None
    
            def forward(self, graph, feat, eweight=None):
                with graph.local_scope():
                    feat = self.linear(feat)
                    graph.ndata['h'] = feat
                    if eweight is None:
                        graph.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
                    else:
                        graph.edata['w'] = eweight
                        graph.update_all(fn.u_mul_e('h', 'w', 'm'), fn.sum('m', 'h'))
    
                    if self.pool:
                        return self.pool(graph, graph.ndata['h'])
                    else:
                        return graph.ndata['h']
    
        # Explain node prediction
        model = Model(5, out_dim)
        model = model.to(F.ctx())
        explainer = nn.GNNExplainer(model, num_hops=1)
>       new_center, sg, feat_mask, edge_mask = explainer.explain_node(0, g, feat)

tests\pytorch\test_nn.py:1321: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
python\dgl\nn\pytorch\explain\gnnexplainer.py:275: in explain_node
    logits = self.model(graph=sg, feat=feat, **kwargs)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
tests\pytorch\test_nn.py:1307: in forward
    graph.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0752C408>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.3272,  0.0935],
        [ 0.2792, -0.4684],
        [ 0.2756, -0.2019]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_______________________ test_gnnexplainer[2-g5-idtype1] _______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={}
      edata_schemes={})
idtype = torch.int64, out_dim = 2

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['zero-degree']))
    @pytest.mark.parametrize('out_dim', [1, 2])
    def test_gnnexplainer(g, idtype, out_dim):
        g = g.astype(idtype).to(F.ctx())
        feat = F.randn((g.num_nodes(), 5))
    
        class Model(th.nn.Module):
            def __init__(self, in_feats, out_feats, graph=False):
                super(Model, self).__init__()
                self.linear = th.nn.Linear(in_feats, out_feats)
                if graph:
                    self.pool = nn.AvgPooling()
                else:
                    self.pool = None
    
            def forward(self, graph, feat, eweight=None):
                with graph.local_scope():
                    feat = self.linear(feat)
                    graph.ndata['h'] = feat
                    if eweight is None:
                        graph.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
                    else:
                        graph.edata['w'] = eweight
                        graph.update_all(fn.u_mul_e('h', 'w', 'm'), fn.sum('m', 'h'))
    
                    if self.pool:
                        return self.pool(graph, graph.ndata['h'])
                    else:
                        return graph.ndata['h']
    
        # Explain node prediction
        model = Model(5, out_dim)
        model = model.to(F.ctx())
        explainer = nn.GNNExplainer(model, num_hops=1)
>       new_center, sg, feat_mask, edge_mask = explainer.explain_node(0, g, feat)

tests\pytorch\test_nn.py:1321: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
python\dgl\nn\pytorch\explain\gnnexplainer.py:275: in explain_node
    logits = self.model(graph=sg, feat=feat, **kwargs)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
tests\pytorch\test_nn.py:1307: in forward
    graph.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03E32A88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.8294, -0.0232],
        [ 0.2900, -0.1152],
        [ 0.1495, -0.7319]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________________ test_twirls _________________________________

    def test_twirls():
        g = dgl.graph(([0,1,2,3,2,5], [1,2,3,4,0,3]))
        feat = th.ones(6, 10)
        conv = nn.TWIRLSConv(10, 2, 128, prop_step = 64)
>       res = conv(g , feat)

tests\pytorch\test_nn.py:1392: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\twirlsconv.py:194: in forward
    x = self.unfolding(graph, x)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\twirlsconv.py:561: in forward
    Y = layer(g, Y, X, self.alp, self.lam)
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\twirlsconv.py:249: in forward
    + alp * D_power_bias_X(graph, X, -1, lam, 1 - lam)
python\dgl\nn\pytorch\conv\twirlsconv.py:218: in _prop
    Y = AX(graph, Y)
python\dgl\nn\pytorch\conv\twirlsconv.py:413: in AX
    fn.u_mul_e("h", "w", "m"), fn.sum("m", "h"),
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python\dgl\ops\spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07162688>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[ 0.3929,  0.0664,  0.2063,  0.1665, -0.0047,  0.0481,  0.2710,  0.1409,
         -0.2799,  0.0362,  0.3737,  ...2714,
          0.1337,  0.2547,  0.5175, -0.8311,  0.1049, -0.2532, -0.4237,  0.7530]],
       grad_fn=<MulBackward0>)
rhs_data = tensor([[1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.],
        [1.]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
____________________________ test_hgt[1-4-idtype0] ____________________________

idtype = torch.int32, in_size = 4, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('in_size', [4])
    @pytest.mark.parametrize('num_heads', [1])
    def test_hgt(idtype, in_size, num_heads):
        dev = F.ctx()
        num_etypes = 5
        num_ntypes = 2
        head_size = in_size // num_heads
    
        g = dgl.from_scipy(sp.sparse.random(100, 100, density=0.01))
        g = g.astype(idtype).to(dev)
        etype = th.tensor([i % num_etypes for i in range(g.num_edges())]).to(dev)
        ntype = th.tensor([i % num_ntypes for i in range(g.num_nodes())]).to(dev)
        x = th.randn(g.num_nodes(), in_size).to(dev)
    
        m = nn.HGTConv(in_size, head_size, num_heads, num_ntypes, num_etypes).to(dev)
    
>       y = m(g, x, ntype, etype)

tests\pytorch\test_nn.py:1434: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\hgtconv.py:136: in forward
    g.edata['m'] = g.edata['m'] * edge_softmax(g, g.edata['a']).unsqueeze(-1)
python\dgl\ops\edge_softmax.py:135: in edge_softmax
    eids=eids, norm_by=norm_by)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF072471C8>
logits = tensor([[-0.1500],
        [ 0.1758],
        [-0.0766],
        [-0.2495],
        [-0.0076],
        [-0.1007],
    ...       [-0.2692],
        [-0.1459],
        [ 0.1354],
        [ 0.0796],
        [ 0.1519]], grad_fn=<StackBackward>)
eids = '__ALL__', norm_by = 'dst'

    def edge_softmax(gidx, logits, eids=ALL, norm_by='dst'):
        args = _cast_if_autocast_enabled(gidx, logits, eids, norm_by)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:770: TypeError
____________________________ test_hgt[1-4-idtype1] ____________________________

idtype = torch.int64, in_size = 4, num_heads = 1

    @parametrize_idtype
    @pytest.mark.parametrize('in_size', [4])
    @pytest.mark.parametrize('num_heads', [1])
    def test_hgt(idtype, in_size, num_heads):
        dev = F.ctx()
        num_etypes = 5
        num_ntypes = 2
        head_size = in_size // num_heads
    
        g = dgl.from_scipy(sp.sparse.random(100, 100, density=0.01))
        g = g.astype(idtype).to(dev)
        etype = th.tensor([i % num_etypes for i in range(g.num_edges())]).to(dev)
        ntype = th.tensor([i % num_ntypes for i in range(g.num_nodes())]).to(dev)
        x = th.randn(g.num_nodes(), in_size).to(dev)
    
        m = nn.HGTConv(in_size, head_size, num_heads, num_ntypes, num_etypes).to(dev)
    
>       y = m(g, x, ntype, etype)

tests\pytorch\test_nn.py:1434: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\hgtconv.py:136: in forward
    g.edata['m'] = g.edata['m'] * edge_softmax(g, g.edata['a']).unsqueeze(-1)
python\dgl\ops\edge_softmax.py:135: in edge_softmax
    eids=eids, norm_by=norm_by)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0743C408>
logits = tensor([[-1.1611e-01],
        [ 1.0064e-01],
        [ 1.2103e-01],
        [ 1.0987e-01],
        [ 1.8929e-03],
   ...,
        [-1.0278e-01],
        [ 2.1266e-01],
        [ 6.4873e-02],
        [ 2.8267e-02]], grad_fn=<StackBackward>)
eids = '__ALL__', norm_by = 'dst'

    def edge_softmax(gidx, logits, eids=ALL, norm_by='dst'):
        args = _cast_if_autocast_enabled(gidx, logits, eids, norm_by)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:770: TypeError
_________________________ test_group_rev_res[idtype0] _________________________

idtype = torch.int32

    @parametrize_idtype
    def test_group_rev_res(idtype):
        dev = F.ctx()
    
        num_nodes = 5
        num_edges = 20
        feats = 32
        groups = 2
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        h = th.randn(num_nodes, feats).to(dev)
        conv = nn.GraphConv(feats // groups, feats // groups)
        model = nn.GroupRevRes(conv, groups).to(dev)
>       result = model(g, h)

tests\pytorch\test_nn.py:1545: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\grouprevres.py:235: in forward
    *(args + tuple([p for p in self.parameters() if p.requires_grad])))
python\dgl\nn\pytorch\conv\grouprevres.py:29: in forward
    outputs = ctx.fn(*x).detach_()
python\dgl\nn\pytorch\conv\grouprevres.py:182: in _forward
    y_in = xs[i] + self.gnn_modules[i](g, y_in, *args_chunks[i])
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:428: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07859FC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.5902,  0.0630,  0.3305,  0.9670, -0.3103, -0.0413,  0.3273,  2.3613,
         -0.2182,  0.3764,  0.0723,  ...  0.2139, -0.6448,  0.0698, -0.0153,
          1.0399, -0.7878, -0.4574,  0.1666,  0.0027, -0.1664,  0.0547,  0.2055]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_group_rev_res[idtype1] _________________________

idtype = torch.int64

    @parametrize_idtype
    def test_group_rev_res(idtype):
        dev = F.ctx()
    
        num_nodes = 5
        num_edges = 20
        feats = 32
        groups = 2
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        h = th.randn(num_nodes, feats).to(dev)
        conv = nn.GraphConv(feats // groups, feats // groups)
        model = nn.GroupRevRes(conv, groups).to(dev)
>       result = model(g, h)

tests\pytorch\test_nn.py:1545: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\grouprevres.py:235: in forward
    *(args + tuple([p for p in self.parameters() if p.requires_grad])))
python\dgl\nn\pytorch\conv\grouprevres.py:29: in forward
    outputs = ctx.fn(*x).detach_()
python\dgl\nn\pytorch\conv\grouprevres.py:182: in _forward
    y_in = xs[i] + self.gnn_modules[i](g, y_in, *args_chunks[i])
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\graphconv.py:428: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE12258208>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.1361,  0.6922, -0.7377, -0.7217,  0.5052, -0.0251, -0.4911,  0.6332,
         -0.0894,  0.0715,  0.1784,  ...  0.0028,  0.8519, -0.9331, -0.2509,
          0.0136,  0.1051,  0.6944,  0.6106, -0.4397,  0.6314,  0.3690, -0.3548]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________________ test_egnn_conv[16-16-16-16] _________________________

in_size = 16, hidden_size = 16, out_size = 16, edge_feat_size = 16

    @pytest.mark.parametrize('in_size', [16, 32])
    @pytest.mark.parametrize('hidden_size', [16, 32])
    @pytest.mark.parametrize('out_size', [16, 32])
    @pytest.mark.parametrize('edge_feat_size', [16, 10, 0])
    def test_egnn_conv(in_size, hidden_size, out_size, edge_feat_size):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        h = th.randn(num_nodes, in_size).to(dev)
        x = th.randn(num_nodes, 3).to(dev)
        e = th.randn(num_edges, edge_feat_size).to(dev)
        model = nn.EGNNConv(in_size, hidden_size, out_size, edge_feat_size).to(dev)
>       model(g, h, x, e)

tests\pytorch\test_nn.py:1561: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egnnconv.py:137: in forward
    graph.apply_edges(fn.u_sub_v('x', 'x', 'x_diff'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03DA6608>
op = 'add'
lhs_data = tensor([[ 0.1555,  1.3800, -1.3861],
        [ 0.6016, -0.5539, -1.4905],
        [ 0.5590, -0.8129,  0.7952],
        [-0.1751,  0.4680, -0.3326],
        [ 1.4519, -0.1267,  0.9008]])
rhs_data = tensor([[-0.1555, -1.3800,  1.3861],
        [-0.6016,  0.5539,  1.4905],
        [-0.5590,  0.8129, -0.7952],
        [ 0.1751, -0.4680,  0.3326],
        [-1.4519,  0.1267, -0.9008]])
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_________________________ test_egnn_conv[16-16-16-32] _________________________

in_size = 32, hidden_size = 16, out_size = 16, edge_feat_size = 16

    @pytest.mark.parametrize('in_size', [16, 32])
    @pytest.mark.parametrize('hidden_size', [16, 32])
    @pytest.mark.parametrize('out_size', [16, 32])
    @pytest.mark.parametrize('edge_feat_size', [16, 10, 0])
    def test_egnn_conv(in_size, hidden_size, out_size, edge_feat_size):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        h = th.randn(num_nodes, in_size).to(dev)
        x = th.randn(num_nodes, 3).to(dev)
        e = th.randn(num_edges, edge_feat_size).to(dev)
        model = nn.EGNNConv(in_size, hidden_size, out_size, edge_feat_size).to(dev)
>       model(g, h, x, e)

tests\pytorch\test_nn.py:1561: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egnnconv.py:137: in forward
    graph.apply_edges(fn.u_sub_v('x', 'x', 'x_diff'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF072D3B88>
op = 'add'
lhs_data = tensor([[ 0.5551,  0.1841, -0.4555],
        [ 1.0061,  0.7136,  0.6413],
        [ 0.1150, -0.9536,  0.2384],
        [-0.2700, -1.7394, -0.4870],
        [-1.6799,  1.3412, -1.0390]])
rhs_data = tensor([[-0.5551, -0.1841,  0.4555],
        [-1.0061, -0.7136, -0.6413],
        [-0.1150,  0.9536, -0.2384],
        [ 0.2700,  1.7394,  0.4870],
        [ 1.6799, -1.3412,  1.0390]])
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_________________________ test_egnn_conv[16-16-32-16] _________________________

in_size = 16, hidden_size = 32, out_size = 16, edge_feat_size = 16

    @pytest.mark.parametrize('in_size', [16, 32])
    @pytest.mark.parametrize('hidden_size', [16, 32])
    @pytest.mark.parametrize('out_size', [16, 32])
    @pytest.mark.parametrize('edge_feat_size', [16, 10, 0])
    def test_egnn_conv(in_size, hidden_size, out_size, edge_feat_size):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        h = th.randn(num_nodes, in_size).to(dev)
        x = th.randn(num_nodes, 3).to(dev)
        e = th.randn(num_edges, edge_feat_size).to(dev)
        model = nn.EGNNConv(in_size, hidden_size, out_size, edge_feat_size).to(dev)
>       model(g, h, x, e)

tests\pytorch\test_nn.py:1561: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egnnconv.py:137: in forward
    graph.apply_edges(fn.u_sub_v('x', 'x', 'x_diff'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF070A5B08>
op = 'add'
lhs_data = tensor([[-5.7356e-01, -1.1111e+00, -2.0557e-01],
        [ 1.4431e+00, -1.0450e+00,  1.1679e-03],
        [-5.0074e-01...4e+00, -3.7470e-01],
        [ 8.6229e-01, -2.4098e-01,  7.6695e-01],
        [-9.1489e-01, -7.4053e-01, -2.6023e-01]])
rhs_data = tensor([[ 5.7356e-01,  1.1111e+00,  2.0557e-01],
        [-1.4431e+00,  1.0450e+00, -1.1679e-03],
        [ 5.0074e-01...4e+00,  3.7470e-01],
        [-8.6229e-01,  2.4098e-01, -7.6695e-01],
        [ 9.1489e-01,  7.4053e-01,  2.6023e-01]])
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_________________________ test_egnn_conv[16-16-32-32] _________________________

in_size = 32, hidden_size = 32, out_size = 16, edge_feat_size = 16

    @pytest.mark.parametrize('in_size', [16, 32])
    @pytest.mark.parametrize('hidden_size', [16, 32])
    @pytest.mark.parametrize('out_size', [16, 32])
    @pytest.mark.parametrize('edge_feat_size', [16, 10, 0])
    def test_egnn_conv(in_size, hidden_size, out_size, edge_feat_size):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        h = th.randn(num_nodes, in_size).to(dev)
        x = th.randn(num_nodes, 3).to(dev)
        e = th.randn(num_edges, edge_feat_size).to(dev)
        model = nn.EGNNConv(in_size, hidden_size, out_size, edge_feat_size).to(dev)
>       model(g, h, x, e)

tests\pytorch\test_nn.py:1561: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egnnconv.py:137: in forward
    graph.apply_edges(fn.u_sub_v('x', 'x', 'x_diff'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0743F488>
op = 'add'
lhs_data = tensor([[ 0.1025,  0.3959,  1.3502],
        [ 0.4283, -1.9424, -0.3929],
        [ 0.1560, -0.2212,  0.0080],
        [-0.6130,  0.3075,  0.4483],
        [ 0.1573, -0.1212,  1.2261]])
rhs_data = tensor([[-0.1025, -0.3959, -1.3502],
        [-0.4283,  1.9424,  0.3929],
        [-0.1560,  0.2212, -0.0080],
        [ 0.6130, -0.3075, -0.4483],
        [-0.1573,  0.1212, -1.2261]])
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_________________________ test_egnn_conv[16-32-16-16] _________________________

in_size = 16, hidden_size = 16, out_size = 32, edge_feat_size = 16

    @pytest.mark.parametrize('in_size', [16, 32])
    @pytest.mark.parametrize('hidden_size', [16, 32])
    @pytest.mark.parametrize('out_size', [16, 32])
    @pytest.mark.parametrize('edge_feat_size', [16, 10, 0])
    def test_egnn_conv(in_size, hidden_size, out_size, edge_feat_size):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        h = th.randn(num_nodes, in_size).to(dev)
        x = th.randn(num_nodes, 3).to(dev)
        e = th.randn(num_edges, edge_feat_size).to(dev)
        model = nn.EGNNConv(in_size, hidden_size, out_size, edge_feat_size).to(dev)
>       model(g, h, x, e)

tests\pytorch\test_nn.py:1561: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egnnconv.py:137: in forward
    graph.apply_edges(fn.u_sub_v('x', 'x', 'x_diff'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF077CAE08>
op = 'add'
lhs_data = tensor([[-0.2653,  1.2890,  0.6706],
        [ 0.6051,  1.2266, -0.2979],
        [-1.1657,  0.0117, -0.2497],
        [ 0.8012,  1.4983,  0.5824],
        [-0.2324, -1.0041,  0.7958]])
rhs_data = tensor([[ 0.2653, -1.2890, -0.6706],
        [-0.6051, -1.2266,  0.2979],
        [ 1.1657, -0.0117,  0.2497],
        [-0.8012, -1.4983, -0.5824],
        [ 0.2324,  1.0041, -0.7958]])
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_________________________ test_egnn_conv[16-32-16-32] _________________________

in_size = 32, hidden_size = 16, out_size = 32, edge_feat_size = 16

    @pytest.mark.parametrize('in_size', [16, 32])
    @pytest.mark.parametrize('hidden_size', [16, 32])
    @pytest.mark.parametrize('out_size', [16, 32])
    @pytest.mark.parametrize('edge_feat_size', [16, 10, 0])
    def test_egnn_conv(in_size, hidden_size, out_size, edge_feat_size):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        h = th.randn(num_nodes, in_size).to(dev)
        x = th.randn(num_nodes, 3).to(dev)
        e = th.randn(num_edges, edge_feat_size).to(dev)
        model = nn.EGNNConv(in_size, hidden_size, out_size, edge_feat_size).to(dev)
>       model(g, h, x, e)

tests\pytorch\test_nn.py:1561: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egnnconv.py:137: in forward
    graph.apply_edges(fn.u_sub_v('x', 'x', 'x_diff'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0749B748>
op = 'add'
lhs_data = tensor([[-0.3298,  0.2859,  0.9299],
        [ 1.4075, -1.2248,  0.3947],
        [ 0.5814,  0.3582, -1.2485],
        [-0.1334,  1.6807, -1.0467],
        [ 0.2541,  0.4382,  0.5471]])
rhs_data = tensor([[ 0.3298, -0.2859, -0.9299],
        [-1.4075,  1.2248, -0.3947],
        [-0.5814, -0.3582,  1.2485],
        [ 0.1334, -1.6807,  1.0467],
        [-0.2541, -0.4382, -0.5471]])
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_________________________ test_egnn_conv[16-32-32-16] _________________________

in_size = 16, hidden_size = 32, out_size = 32, edge_feat_size = 16

    @pytest.mark.parametrize('in_size', [16, 32])
    @pytest.mark.parametrize('hidden_size', [16, 32])
    @pytest.mark.parametrize('out_size', [16, 32])
    @pytest.mark.parametrize('edge_feat_size', [16, 10, 0])
    def test_egnn_conv(in_size, hidden_size, out_size, edge_feat_size):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        h = th.randn(num_nodes, in_size).to(dev)
        x = th.randn(num_nodes, 3).to(dev)
        e = th.randn(num_edges, edge_feat_size).to(dev)
        model = nn.EGNNConv(in_size, hidden_size, out_size, edge_feat_size).to(dev)
>       model(g, h, x, e)

tests\pytorch\test_nn.py:1561: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egnnconv.py:137: in forward
    graph.apply_edges(fn.u_sub_v('x', 'x', 'x_diff'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01885EC8>
op = 'add'
lhs_data = tensor([[-0.6349, -0.2510, -0.9359],
        [-0.1156, -0.9172,  1.0636],
        [-0.0240,  1.0405, -0.1474],
        [-1.1110,  1.1568,  0.4864],
        [-0.8727, -0.0653, -0.1393]])
rhs_data = tensor([[ 0.6349,  0.2510,  0.9359],
        [ 0.1156,  0.9172, -1.0636],
        [ 0.0240, -1.0405,  0.1474],
        [ 1.1110, -1.1568, -0.4864],
        [ 0.8727,  0.0653,  0.1393]])
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_________________________ test_egnn_conv[16-32-32-32] _________________________

in_size = 32, hidden_size = 32, out_size = 32, edge_feat_size = 16

    @pytest.mark.parametrize('in_size', [16, 32])
    @pytest.mark.parametrize('hidden_size', [16, 32])
    @pytest.mark.parametrize('out_size', [16, 32])
    @pytest.mark.parametrize('edge_feat_size', [16, 10, 0])
    def test_egnn_conv(in_size, hidden_size, out_size, edge_feat_size):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        h = th.randn(num_nodes, in_size).to(dev)
        x = th.randn(num_nodes, 3).to(dev)
        e = th.randn(num_edges, edge_feat_size).to(dev)
        model = nn.EGNNConv(in_size, hidden_size, out_size, edge_feat_size).to(dev)
>       model(g, h, x, e)

tests\pytorch\test_nn.py:1561: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egnnconv.py:137: in forward
    graph.apply_edges(fn.u_sub_v('x', 'x', 'x_diff'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE123C4508>
op = 'add'
lhs_data = tensor([[ 1.0763,  0.6033,  1.4741],
        [-0.7192,  0.4424,  1.2418],
        [-0.0951, -0.3115, -1.4940],
        [ 1.1853,  0.1006, -0.5396],
        [ 0.3217, -0.4816,  0.4737]])
rhs_data = tensor([[-1.0763, -0.6033, -1.4741],
        [ 0.7192, -0.4424, -1.2418],
        [ 0.0951,  0.3115,  1.4940],
        [-1.1853, -0.1006,  0.5396],
        [-0.3217,  0.4816, -0.4737]])
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_________________________ test_egnn_conv[10-16-16-16] _________________________

in_size = 16, hidden_size = 16, out_size = 16, edge_feat_size = 10

    @pytest.mark.parametrize('in_size', [16, 32])
    @pytest.mark.parametrize('hidden_size', [16, 32])
    @pytest.mark.parametrize('out_size', [16, 32])
    @pytest.mark.parametrize('edge_feat_size', [16, 10, 0])
    def test_egnn_conv(in_size, hidden_size, out_size, edge_feat_size):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        h = th.randn(num_nodes, in_size).to(dev)
        x = th.randn(num_nodes, 3).to(dev)
        e = th.randn(num_edges, edge_feat_size).to(dev)
        model = nn.EGNNConv(in_size, hidden_size, out_size, edge_feat_size).to(dev)
>       model(g, h, x, e)

tests\pytorch\test_nn.py:1561: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egnnconv.py:137: in forward
    graph.apply_edges(fn.u_sub_v('x', 'x', 'x_diff'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE13AFDCC8>
op = 'add'
lhs_data = tensor([[ 0.4903,  0.9804, -0.8172],
        [-0.0522, -1.0841,  0.7529],
        [ 0.3059, -1.4338,  0.9355],
        [-1.2628,  2.2440, -0.2803],
        [-0.6934,  1.0953, -0.3488]])
rhs_data = tensor([[-0.4903, -0.9804,  0.8172],
        [ 0.0522,  1.0841, -0.7529],
        [-0.3059,  1.4338, -0.9355],
        [ 1.2628, -2.2440,  0.2803],
        [ 0.6934, -1.0953,  0.3488]])
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_________________________ test_egnn_conv[10-16-16-32] _________________________

in_size = 32, hidden_size = 16, out_size = 16, edge_feat_size = 10

    @pytest.mark.parametrize('in_size', [16, 32])
    @pytest.mark.parametrize('hidden_size', [16, 32])
    @pytest.mark.parametrize('out_size', [16, 32])
    @pytest.mark.parametrize('edge_feat_size', [16, 10, 0])
    def test_egnn_conv(in_size, hidden_size, out_size, edge_feat_size):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        h = th.randn(num_nodes, in_size).to(dev)
        x = th.randn(num_nodes, 3).to(dev)
        e = th.randn(num_edges, edge_feat_size).to(dev)
        model = nn.EGNNConv(in_size, hidden_size, out_size, edge_feat_size).to(dev)
>       model(g, h, x, e)

tests\pytorch\test_nn.py:1561: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egnnconv.py:137: in forward
    graph.apply_edges(fn.u_sub_v('x', 'x', 'x_diff'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF030E7308>
op = 'add'
lhs_data = tensor([[-0.7751, -2.3237,  1.1517],
        [ 2.0998,  1.8434,  0.7605],
        [ 0.5685,  0.1157, -0.9222],
        [ 0.7898, -0.5610,  1.1371],
        [-0.9465,  2.0115, -0.0580]])
rhs_data = tensor([[ 0.7751,  2.3237, -1.1517],
        [-2.0998, -1.8434, -0.7605],
        [-0.5685, -0.1157,  0.9222],
        [-0.7898,  0.5610, -1.1371],
        [ 0.9465, -2.0115,  0.0580]])
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_________________________ test_egnn_conv[10-16-32-16] _________________________

in_size = 16, hidden_size = 32, out_size = 16, edge_feat_size = 10

    @pytest.mark.parametrize('in_size', [16, 32])
    @pytest.mark.parametrize('hidden_size', [16, 32])
    @pytest.mark.parametrize('out_size', [16, 32])
    @pytest.mark.parametrize('edge_feat_size', [16, 10, 0])
    def test_egnn_conv(in_size, hidden_size, out_size, edge_feat_size):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        h = th.randn(num_nodes, in_size).to(dev)
        x = th.randn(num_nodes, 3).to(dev)
        e = th.randn(num_edges, edge_feat_size).to(dev)
        model = nn.EGNNConv(in_size, hidden_size, out_size, edge_feat_size).to(dev)
>       model(g, h, x, e)

tests\pytorch\test_nn.py:1561: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egnnconv.py:137: in forward
    graph.apply_edges(fn.u_sub_v('x', 'x', 'x_diff'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03CF8208>
op = 'add'
lhs_data = tensor([[ 0.9564, -0.4323, -1.0107],
        [-0.2245, -0.9657,  1.4202],
        [-0.4110, -2.6702, -1.1381],
        [-0.2941, -0.2387, -2.0593],
        [ 0.2100,  0.2850, -0.8630]])
rhs_data = tensor([[-0.9564,  0.4323,  1.0107],
        [ 0.2245,  0.9657, -1.4202],
        [ 0.4110,  2.6702,  1.1381],
        [ 0.2941,  0.2387,  2.0593],
        [-0.2100, -0.2850,  0.8630]])
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_________________________ test_egnn_conv[10-16-32-32] _________________________

in_size = 32, hidden_size = 32, out_size = 16, edge_feat_size = 10

    @pytest.mark.parametrize('in_size', [16, 32])
    @pytest.mark.parametrize('hidden_size', [16, 32])
    @pytest.mark.parametrize('out_size', [16, 32])
    @pytest.mark.parametrize('edge_feat_size', [16, 10, 0])
    def test_egnn_conv(in_size, hidden_size, out_size, edge_feat_size):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        h = th.randn(num_nodes, in_size).to(dev)
        x = th.randn(num_nodes, 3).to(dev)
        e = th.randn(num_edges, edge_feat_size).to(dev)
        model = nn.EGNNConv(in_size, hidden_size, out_size, edge_feat_size).to(dev)
>       model(g, h, x, e)

tests\pytorch\test_nn.py:1561: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egnnconv.py:137: in forward
    graph.apply_edges(fn.u_sub_v('x', 'x', 'x_diff'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07999708>
op = 'add'
lhs_data = tensor([[-1.4590,  1.5826,  0.8992],
        [-1.5265,  0.2053, -0.3400],
        [ 0.1323, -0.1550, -1.8860],
        [ 0.4969,  0.2304,  1.4859],
        [-0.1446, -0.0718,  0.6177]])
rhs_data = tensor([[ 1.4590, -1.5826, -0.8992],
        [ 1.5265, -0.2053,  0.3400],
        [-0.1323,  0.1550,  1.8860],
        [-0.4969, -0.2304, -1.4859],
        [ 0.1446,  0.0718, -0.6177]])
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_________________________ test_egnn_conv[10-32-16-16] _________________________

in_size = 16, hidden_size = 16, out_size = 32, edge_feat_size = 10

    @pytest.mark.parametrize('in_size', [16, 32])
    @pytest.mark.parametrize('hidden_size', [16, 32])
    @pytest.mark.parametrize('out_size', [16, 32])
    @pytest.mark.parametrize('edge_feat_size', [16, 10, 0])
    def test_egnn_conv(in_size, hidden_size, out_size, edge_feat_size):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        h = th.randn(num_nodes, in_size).to(dev)
        x = th.randn(num_nodes, 3).to(dev)
        e = th.randn(num_edges, edge_feat_size).to(dev)
        model = nn.EGNNConv(in_size, hidden_size, out_size, edge_feat_size).to(dev)
>       model(g, h, x, e)

tests\pytorch\test_nn.py:1561: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egnnconv.py:137: in forward
    graph.apply_edges(fn.u_sub_v('x', 'x', 'x_diff'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0793DE08>
op = 'add'
lhs_data = tensor([[-0.1856,  0.2390,  0.3081],
        [ 0.7310, -1.2469, -0.0110],
        [ 0.3058,  0.2911,  0.0168],
        [ 1.4814,  0.2427, -0.1365],
        [ 0.2853,  1.7843,  0.3019]])
rhs_data = tensor([[ 0.1856, -0.2390, -0.3081],
        [-0.7310,  1.2469,  0.0110],
        [-0.3058, -0.2911, -0.0168],
        [-1.4814, -0.2427,  0.1365],
        [-0.2853, -1.7843, -0.3019]])
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_________________________ test_egnn_conv[10-32-16-32] _________________________

in_size = 32, hidden_size = 16, out_size = 32, edge_feat_size = 10

    @pytest.mark.parametrize('in_size', [16, 32])
    @pytest.mark.parametrize('hidden_size', [16, 32])
    @pytest.mark.parametrize('out_size', [16, 32])
    @pytest.mark.parametrize('edge_feat_size', [16, 10, 0])
    def test_egnn_conv(in_size, hidden_size, out_size, edge_feat_size):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        h = th.randn(num_nodes, in_size).to(dev)
        x = th.randn(num_nodes, 3).to(dev)
        e = th.randn(num_edges, edge_feat_size).to(dev)
        model = nn.EGNNConv(in_size, hidden_size, out_size, edge_feat_size).to(dev)
>       model(g, h, x, e)

tests\pytorch\test_nn.py:1561: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egnnconv.py:137: in forward
    graph.apply_edges(fn.u_sub_v('x', 'x', 'x_diff'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF070CAEC8>
op = 'add'
lhs_data = tensor([[-0.9435,  0.0035,  2.0401],
        [ 1.2487, -0.1685,  1.6460],
        [ 0.5154,  0.9192, -0.2428],
        [ 0.6818,  0.6948, -0.2279],
        [-0.5829,  0.1165,  0.6121]])
rhs_data = tensor([[ 0.9435, -0.0035, -2.0401],
        [-1.2487,  0.1685, -1.6460],
        [-0.5154, -0.9192,  0.2428],
        [-0.6818, -0.6948,  0.2279],
        [ 0.5829, -0.1165, -0.6121]])
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_________________________ test_egnn_conv[10-32-32-16] _________________________

in_size = 16, hidden_size = 32, out_size = 32, edge_feat_size = 10

    @pytest.mark.parametrize('in_size', [16, 32])
    @pytest.mark.parametrize('hidden_size', [16, 32])
    @pytest.mark.parametrize('out_size', [16, 32])
    @pytest.mark.parametrize('edge_feat_size', [16, 10, 0])
    def test_egnn_conv(in_size, hidden_size, out_size, edge_feat_size):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        h = th.randn(num_nodes, in_size).to(dev)
        x = th.randn(num_nodes, 3).to(dev)
        e = th.randn(num_edges, edge_feat_size).to(dev)
        model = nn.EGNNConv(in_size, hidden_size, out_size, edge_feat_size).to(dev)
>       model(g, h, x, e)

tests\pytorch\test_nn.py:1561: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egnnconv.py:137: in forward
    graph.apply_edges(fn.u_sub_v('x', 'x', 'x_diff'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07437FC8>
op = 'add'
lhs_data = tensor([[ 0.8146, -1.2644,  0.9078],
        [ 1.2332, -0.2055, -0.0797],
        [-1.8553,  0.6014, -0.2219],
        [ 1.1794,  0.9758, -1.7981],
        [ 2.3522,  0.2859,  0.8472]])
rhs_data = tensor([[-0.8146,  1.2644, -0.9078],
        [-1.2332,  0.2055,  0.0797],
        [ 1.8553, -0.6014,  0.2219],
        [-1.1794, -0.9758,  1.7981],
        [-2.3522, -0.2859, -0.8472]])
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_________________________ test_egnn_conv[10-32-32-32] _________________________

in_size = 32, hidden_size = 32, out_size = 32, edge_feat_size = 10

    @pytest.mark.parametrize('in_size', [16, 32])
    @pytest.mark.parametrize('hidden_size', [16, 32])
    @pytest.mark.parametrize('out_size', [16, 32])
    @pytest.mark.parametrize('edge_feat_size', [16, 10, 0])
    def test_egnn_conv(in_size, hidden_size, out_size, edge_feat_size):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        h = th.randn(num_nodes, in_size).to(dev)
        x = th.randn(num_nodes, 3).to(dev)
        e = th.randn(num_edges, edge_feat_size).to(dev)
        model = nn.EGNNConv(in_size, hidden_size, out_size, edge_feat_size).to(dev)
>       model(g, h, x, e)

tests\pytorch\test_nn.py:1561: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egnnconv.py:137: in forward
    graph.apply_edges(fn.u_sub_v('x', 'x', 'x_diff'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF071B0388>
op = 'add'
lhs_data = tensor([[ 0.6394, -0.9289,  1.4670],
        [-1.7976,  0.8076,  1.2068],
        [-0.7799, -0.0431, -0.1128],
        [-0.7640,  0.3676, -1.4259],
        [-0.8048, -0.1220, -0.5056]])
rhs_data = tensor([[-0.6394,  0.9289, -1.4670],
        [ 1.7976, -0.8076, -1.2068],
        [ 0.7799,  0.0431,  0.1128],
        [ 0.7640, -0.3676,  1.4259],
        [ 0.8048,  0.1220,  0.5056]])
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_________________________ test_egnn_conv[0-16-16-16] __________________________

in_size = 16, hidden_size = 16, out_size = 16, edge_feat_size = 0

    @pytest.mark.parametrize('in_size', [16, 32])
    @pytest.mark.parametrize('hidden_size', [16, 32])
    @pytest.mark.parametrize('out_size', [16, 32])
    @pytest.mark.parametrize('edge_feat_size', [16, 10, 0])
    def test_egnn_conv(in_size, hidden_size, out_size, edge_feat_size):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        h = th.randn(num_nodes, in_size).to(dev)
        x = th.randn(num_nodes, 3).to(dev)
        e = th.randn(num_edges, edge_feat_size).to(dev)
        model = nn.EGNNConv(in_size, hidden_size, out_size, edge_feat_size).to(dev)
>       model(g, h, x, e)

tests\pytorch\test_nn.py:1561: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egnnconv.py:137: in forward
    graph.apply_edges(fn.u_sub_v('x', 'x', 'x_diff'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE122BC648>
op = 'add'
lhs_data = tensor([[ 0.8452, -1.3729,  0.1664],
        [-0.0555,  0.6882,  2.4198],
        [ 0.1352, -0.5518, -0.0207],
        [ 0.1508,  2.4068,  0.1535],
        [-1.5185,  1.3602,  0.1981]])
rhs_data = tensor([[-0.8452,  1.3729, -0.1664],
        [ 0.0555, -0.6882, -2.4198],
        [-0.1352,  0.5518,  0.0207],
        [-0.1508, -2.4068, -0.1535],
        [ 1.5185, -1.3602, -0.1981]])
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_________________________ test_egnn_conv[0-16-16-32] __________________________

in_size = 32, hidden_size = 16, out_size = 16, edge_feat_size = 0

    @pytest.mark.parametrize('in_size', [16, 32])
    @pytest.mark.parametrize('hidden_size', [16, 32])
    @pytest.mark.parametrize('out_size', [16, 32])
    @pytest.mark.parametrize('edge_feat_size', [16, 10, 0])
    def test_egnn_conv(in_size, hidden_size, out_size, edge_feat_size):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        h = th.randn(num_nodes, in_size).to(dev)
        x = th.randn(num_nodes, 3).to(dev)
        e = th.randn(num_edges, edge_feat_size).to(dev)
        model = nn.EGNNConv(in_size, hidden_size, out_size, edge_feat_size).to(dev)
>       model(g, h, x, e)

tests\pytorch\test_nn.py:1561: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egnnconv.py:137: in forward
    graph.apply_edges(fn.u_sub_v('x', 'x', 'x_diff'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1225F9C8>
op = 'add'
lhs_data = tensor([[-1.1820,  0.4520, -0.1227],
        [-0.7415, -0.5571,  3.1673],
        [ 1.4870, -0.1696, -0.3700],
        [-1.0529,  1.6969, -0.0381],
        [ 0.2572, -0.7983, -0.7362]])
rhs_data = tensor([[ 1.1820, -0.4520,  0.1227],
        [ 0.7415,  0.5571, -3.1673],
        [-1.4870,  0.1696,  0.3700],
        [ 1.0529, -1.6969,  0.0381],
        [-0.2572,  0.7983,  0.7362]])
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_________________________ test_egnn_conv[0-16-32-16] __________________________

in_size = 16, hidden_size = 32, out_size = 16, edge_feat_size = 0

    @pytest.mark.parametrize('in_size', [16, 32])
    @pytest.mark.parametrize('hidden_size', [16, 32])
    @pytest.mark.parametrize('out_size', [16, 32])
    @pytest.mark.parametrize('edge_feat_size', [16, 10, 0])
    def test_egnn_conv(in_size, hidden_size, out_size, edge_feat_size):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        h = th.randn(num_nodes, in_size).to(dev)
        x = th.randn(num_nodes, 3).to(dev)
        e = th.randn(num_edges, edge_feat_size).to(dev)
        model = nn.EGNNConv(in_size, hidden_size, out_size, edge_feat_size).to(dev)
>       model(g, h, x, e)

tests\pytorch\test_nn.py:1561: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egnnconv.py:137: in forward
    graph.apply_edges(fn.u_sub_v('x', 'x', 'x_diff'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF070A57C8>
op = 'add'
lhs_data = tensor([[-0.5555,  0.4494, -0.0995],
        [-1.1580, -1.2035, -1.6830],
        [ 1.4791, -0.9554,  0.6536],
        [ 1.6538, -0.5048,  2.4620],
        [ 0.5669,  0.9360, -0.1988]])
rhs_data = tensor([[ 0.5555, -0.4494,  0.0995],
        [ 1.1580,  1.2035,  1.6830],
        [-1.4791,  0.9554, -0.6536],
        [-1.6538,  0.5048, -2.4620],
        [-0.5669, -0.9360,  0.1988]])
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_________________________ test_egnn_conv[0-16-32-32] __________________________

in_size = 32, hidden_size = 32, out_size = 16, edge_feat_size = 0

    @pytest.mark.parametrize('in_size', [16, 32])
    @pytest.mark.parametrize('hidden_size', [16, 32])
    @pytest.mark.parametrize('out_size', [16, 32])
    @pytest.mark.parametrize('edge_feat_size', [16, 10, 0])
    def test_egnn_conv(in_size, hidden_size, out_size, edge_feat_size):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        h = th.randn(num_nodes, in_size).to(dev)
        x = th.randn(num_nodes, 3).to(dev)
        e = th.randn(num_edges, edge_feat_size).to(dev)
        model = nn.EGNNConv(in_size, hidden_size, out_size, edge_feat_size).to(dev)
>       model(g, h, x, e)

tests\pytorch\test_nn.py:1561: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egnnconv.py:137: in forward
    graph.apply_edges(fn.u_sub_v('x', 'x', 'x_diff'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE123D1CC8>
op = 'add'
lhs_data = tensor([[-0.3340,  0.1061, -0.3145],
        [ 1.6711, -0.5607,  0.6925],
        [ 0.5874,  1.1827,  0.6793],
        [-0.7498, -0.7626,  0.5196],
        [-0.1976, -0.8008, -0.0455]])
rhs_data = tensor([[ 0.3340, -0.1061,  0.3145],
        [-1.6711,  0.5607, -0.6925],
        [-0.5874, -1.1827, -0.6793],
        [ 0.7498,  0.7626, -0.5196],
        [ 0.1976,  0.8008,  0.0455]])
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_________________________ test_egnn_conv[0-32-16-16] __________________________

in_size = 16, hidden_size = 16, out_size = 32, edge_feat_size = 0

    @pytest.mark.parametrize('in_size', [16, 32])
    @pytest.mark.parametrize('hidden_size', [16, 32])
    @pytest.mark.parametrize('out_size', [16, 32])
    @pytest.mark.parametrize('edge_feat_size', [16, 10, 0])
    def test_egnn_conv(in_size, hidden_size, out_size, edge_feat_size):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        h = th.randn(num_nodes, in_size).to(dev)
        x = th.randn(num_nodes, 3).to(dev)
        e = th.randn(num_edges, edge_feat_size).to(dev)
        model = nn.EGNNConv(in_size, hidden_size, out_size, edge_feat_size).to(dev)
>       model(g, h, x, e)

tests\pytorch\test_nn.py:1561: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egnnconv.py:137: in forward
    graph.apply_edges(fn.u_sub_v('x', 'x', 'x_diff'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF077F7308>
op = 'add'
lhs_data = tensor([[-1.8576, -0.0346,  0.2792],
        [ 0.0050,  0.3662,  0.1368],
        [-0.0447, -0.7263, -0.2495],
        [ 1.5286,  0.6768, -0.0672],
        [ 0.1181, -2.4297,  1.3914]])
rhs_data = tensor([[ 1.8576,  0.0346, -0.2792],
        [-0.0050, -0.3662, -0.1368],
        [ 0.0447,  0.7263,  0.2495],
        [-1.5286, -0.6768,  0.0672],
        [-0.1181,  2.4297, -1.3914]])
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_________________________ test_egnn_conv[0-32-16-32] __________________________

in_size = 32, hidden_size = 16, out_size = 32, edge_feat_size = 0

    @pytest.mark.parametrize('in_size', [16, 32])
    @pytest.mark.parametrize('hidden_size', [16, 32])
    @pytest.mark.parametrize('out_size', [16, 32])
    @pytest.mark.parametrize('edge_feat_size', [16, 10, 0])
    def test_egnn_conv(in_size, hidden_size, out_size, edge_feat_size):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        h = th.randn(num_nodes, in_size).to(dev)
        x = th.randn(num_nodes, 3).to(dev)
        e = th.randn(num_edges, edge_feat_size).to(dev)
        model = nn.EGNNConv(in_size, hidden_size, out_size, edge_feat_size).to(dev)
>       model(g, h, x, e)

tests\pytorch\test_nn.py:1561: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egnnconv.py:137: in forward
    graph.apply_edges(fn.u_sub_v('x', 'x', 'x_diff'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF072FE9C8>
op = 'add'
lhs_data = tensor([[ 0.1698, -0.6871, -0.1356],
        [ 0.6061, -0.8150, -0.0862],
        [-0.2394, -0.1918,  0.4498],
        [ 0.7842,  0.2546, -2.6836],
        [ 1.3082,  0.4661, -0.8588]])
rhs_data = tensor([[-0.1698,  0.6871,  0.1356],
        [-0.6061,  0.8150,  0.0862],
        [ 0.2394,  0.1918, -0.4498],
        [-0.7842, -0.2546,  2.6836],
        [-1.3082, -0.4661,  0.8588]])
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_________________________ test_egnn_conv[0-32-32-16] __________________________

in_size = 16, hidden_size = 32, out_size = 32, edge_feat_size = 0

    @pytest.mark.parametrize('in_size', [16, 32])
    @pytest.mark.parametrize('hidden_size', [16, 32])
    @pytest.mark.parametrize('out_size', [16, 32])
    @pytest.mark.parametrize('edge_feat_size', [16, 10, 0])
    def test_egnn_conv(in_size, hidden_size, out_size, edge_feat_size):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        h = th.randn(num_nodes, in_size).to(dev)
        x = th.randn(num_nodes, 3).to(dev)
        e = th.randn(num_edges, edge_feat_size).to(dev)
        model = nn.EGNNConv(in_size, hidden_size, out_size, edge_feat_size).to(dev)
>       model(g, h, x, e)

tests\pytorch\test_nn.py:1561: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egnnconv.py:137: in forward
    graph.apply_edges(fn.u_sub_v('x', 'x', 'x_diff'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF074CC148>
op = 'add'
lhs_data = tensor([[-0.5666,  0.2358,  1.8300],
        [ 1.5686, -0.9704,  0.1244],
        [-0.7257, -1.6196, -0.7985],
        [-1.6828,  0.7457,  0.7924],
        [-0.1769, -1.4847,  0.8365]])
rhs_data = tensor([[ 0.5666, -0.2358, -1.8300],
        [-1.5686,  0.9704, -0.1244],
        [ 0.7257,  1.6196,  0.7985],
        [ 1.6828, -0.7457, -0.7924],
        [ 0.1769,  1.4847, -0.8365]])
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
_________________________ test_egnn_conv[0-32-32-32] __________________________

in_size = 32, hidden_size = 32, out_size = 32, edge_feat_size = 0

    @pytest.mark.parametrize('in_size', [16, 32])
    @pytest.mark.parametrize('hidden_size', [16, 32])
    @pytest.mark.parametrize('out_size', [16, 32])
    @pytest.mark.parametrize('edge_feat_size', [16, 10, 0])
    def test_egnn_conv(in_size, hidden_size, out_size, edge_feat_size):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        h = th.randn(num_nodes, in_size).to(dev)
        x = th.randn(num_nodes, 3).to(dev)
        e = th.randn(num_edges, edge_feat_size).to(dev)
        model = nn.EGNNConv(in_size, hidden_size, out_size, edge_feat_size).to(dev)
>       model(g, h, x, e)

tests\pytorch\test_nn.py:1561: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\conv\egnnconv.py:137: in forward
    graph.apply_edges(fn.u_sub_v('x', 'x', 'x_diff'))
python\dgl\heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python\dgl\core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python\dgl\ops\sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python\dgl\ops\sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF079627C8>
op = 'add'
lhs_data = tensor([[-1.9018, -0.4783,  2.6359],
        [ 0.5444,  1.3759, -1.3350],
        [ 0.0092,  1.0113, -1.2648],
        [ 0.6986,  0.0140, -0.2936],
        [-0.3703,  0.8547, -0.1592]])
rhs_data = tensor([[ 1.9018,  0.4783, -2.6359],
        [-0.5444, -1.3759,  1.3350],
        [-0.0092, -1.0113,  1.2648],
        [-0.6986, -0.0140,  0.2936],
        [ 0.3703, -0.8547,  0.1592]])
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:731: TypeError
__________________ test_label_prop[True-True-True-sym-0.0-3] __________________

k = 3, alpha = 0.0, norm_type = 'sym', clamp = True, normalize = True
reset = True

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0793D7C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0.0000, 0.0000, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.4472, 0.0000],
        [0.0000, 0.5000, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.0000, 0.5774],
        [0.0000, 0.0000, 0.0000, 0.0000]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_label_prop[True-True-True-sym-0.0-5] __________________

k = 5, alpha = 0.0, norm_type = 'sym', clamp = True, normalize = True
reset = True

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03B8EF88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0.0000, 0.0000, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.5774, 0.0000],
        [0.0000, 0.5774, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.0000, 0.5000],
        [0.0000, 0.0000, 0.0000, 0.0000]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_label_prop[True-True-True-sym-0.5-3] __________________

k = 3, alpha = 0.5, norm_type = 'sym', clamp = True, normalize = True
reset = True

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF078ED408>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0.0000, 0.0000, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.4472, 0.0000],
        [0.0000, 1.0000, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.0000, 0.4472],
        [0.0000, 0.0000, 0.0000, 0.0000]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_label_prop[True-True-True-sym-0.5-5] __________________

k = 5, alpha = 0.5, norm_type = 'sym', clamp = True, normalize = True
reset = True

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF078E2D88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0.0000, 0.0000, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.5774, 0.0000],
        [0.0000, 0.5000, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.0000, 0.5000],
        [0.0000, 0.0000, 0.0000, 0.0000]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_label_prop[True-True-True-sym-1.0-3] __________________

k = 3, alpha = 1.0, norm_type = 'sym', clamp = True, normalize = True
reset = True

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07F30B08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0.0000, 0.0000, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.4472, 0.0000],
        [0.0000, 0.5774, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.0000, 0.5000],
        [0.0000, 0.0000, 0.0000, 0.0000]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_label_prop[True-True-True-sym-1.0-5] __________________

k = 5, alpha = 1.0, norm_type = 'sym', clamp = True, normalize = True
reset = True

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07F10E48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0.0000, 0.0000, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.5000, 0.0000],
        [0.0000, 0.5000, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.0000, 0.5000],
        [0.0000, 0.0000, 0.0000, 0.0000]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_label_prop[True-True-True-row-0.0-3] __________________

k = 3, alpha = 0.0, norm_type = 'row', clamp = True, normalize = True
reset = True

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE122C52C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0., 0., 0., 0.],
        [0., 0., 1., 0.],
        [0., 1., 0., 0.],
        [0., 0., 0., 1.],
        [0., 0., 0., 0.]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_label_prop[True-True-True-row-0.0-5] __________________

k = 5, alpha = 0.0, norm_type = 'row', clamp = True, normalize = True
reset = True

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03BCACC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0., 0., 0., 0.],
        [0., 0., 1., 0.],
        [0., 1., 0., 0.],
        [0., 0., 0., 1.],
        [0., 0., 0., 0.]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_label_prop[True-True-True-row-0.5-3] __________________

k = 3, alpha = 0.5, norm_type = 'row', clamp = True, normalize = True
reset = True

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF073DE848>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0., 0., 0., 0.],
        [0., 0., 1., 0.],
        [0., 1., 0., 0.],
        [0., 0., 0., 1.],
        [0., 0., 0., 0.]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_label_prop[True-True-True-row-0.5-5] __________________

k = 5, alpha = 0.5, norm_type = 'row', clamp = True, normalize = True
reset = True

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07E1B608>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0., 0., 0., 0.],
        [0., 0., 1., 0.],
        [0., 1., 0., 0.],
        [0., 0., 0., 1.],
        [0., 0., 0., 0.]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_label_prop[True-True-True-row-1.0-3] __________________

k = 3, alpha = 1.0, norm_type = 'row', clamp = True, normalize = True
reset = True

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07249F88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0., 0., 0., 0.],
        [0., 0., 1., 0.],
        [0., 1., 0., 0.],
        [0., 0., 0., 1.],
        [0., 0., 0., 0.]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
__________________ test_label_prop[True-True-True-row-1.0-5] __________________

k = 5, alpha = 1.0, norm_type = 'row', clamp = True, normalize = True
reset = True

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03CF8B48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0., 0., 0., 0.],
        [0., 0., 1., 0.],
        [0., 1., 0., 0.],
        [0., 0., 0., 1.],
        [0., 0., 0., 0.]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_label_prop[True-True-False-sym-0.0-3] __________________

k = 3, alpha = 0.0, norm_type = 'sym', clamp = False, normalize = True
reset = True

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0714B4C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0.0000, 0.0000, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.4472, 0.0000],
        [0.0000, 0.5774, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.0000, 0.5000],
        [0.0000, 0.0000, 0.0000, 0.0000]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_label_prop[True-True-False-sym-0.0-5] __________________

k = 5, alpha = 0.0, norm_type = 'sym', clamp = False, normalize = True
reset = True

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07120808>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0.0000, 0.0000, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.5774, 0.0000],
        [0.0000, 0.4472, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.0000, 0.5774],
        [0.0000, 0.0000, 0.0000, 0.0000]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_label_prop[True-True-False-sym-0.5-3] __________________

k = 3, alpha = 0.5, norm_type = 'sym', clamp = False, normalize = True
reset = True

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03AD9E88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0.0000, 0.0000, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.5774, 0.0000],
        [0.0000, 0.5000, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.0000, 0.4472],
        [0.0000, 0.0000, 0.0000, 0.0000]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_label_prop[True-True-False-sym-0.5-5] __________________

k = 5, alpha = 0.5, norm_type = 'sym', clamp = False, normalize = True
reset = True

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07848D08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0.0000, 0.0000, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.5000, 0.0000],
        [0.0000, 0.5000, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.0000, 0.5000],
        [0.0000, 0.0000, 0.0000, 0.0000]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_label_prop[True-True-False-sym-1.0-3] __________________

k = 3, alpha = 1.0, norm_type = 'sym', clamp = False, normalize = True
reset = True

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07E73CC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0.0000, 0.0000, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.4472, 0.0000],
        [0.0000, 0.5000, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.0000, 0.7071],
        [0.0000, 0.0000, 0.0000, 0.0000]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_label_prop[True-True-False-sym-1.0-5] __________________

k = 5, alpha = 1.0, norm_type = 'sym', clamp = False, normalize = True
reset = True

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF01555D88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0.0000, 0.0000, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.5000, 0.0000],
        [0.0000, 0.4472, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.0000, 0.5774],
        [0.0000, 0.0000, 0.0000, 0.0000]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_label_prop[True-True-False-row-0.0-3] __________________

k = 3, alpha = 0.0, norm_type = 'row', clamp = False, normalize = True
reset = True

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0793DAC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0., 0., 0., 0.],
        [0., 0., 1., 0.],
        [0., 1., 0., 0.],
        [0., 0., 0., 1.],
        [0., 0., 0., 0.]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_label_prop[True-True-False-row-0.0-5] __________________

k = 5, alpha = 0.0, norm_type = 'row', clamp = False, normalize = True
reset = True

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03BFBE48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0., 0., 0., 0.],
        [0., 0., 1., 0.],
        [0., 1., 0., 0.],
        [0., 0., 0., 1.],
        [0., 0., 0., 0.]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_label_prop[True-True-False-row-0.5-3] __________________

k = 3, alpha = 0.5, norm_type = 'row', clamp = False, normalize = True
reset = True

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07565B48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0., 0., 0., 0.],
        [0., 0., 1., 0.],
        [0., 1., 0., 0.],
        [0., 0., 0., 1.],
        [0., 0., 0., 0.]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_label_prop[True-True-False-row-0.5-5] __________________

k = 5, alpha = 0.5, norm_type = 'row', clamp = False, normalize = True
reset = True

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03132288>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0., 0., 0., 0.],
        [0., 0., 1., 0.],
        [0., 1., 0., 0.],
        [0., 0., 0., 1.],
        [0., 0., 0., 0.]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_label_prop[True-True-False-row-1.0-3] __________________

k = 3, alpha = 1.0, norm_type = 'row', clamp = False, normalize = True
reset = True

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03E26B88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0., 0., 0., 0.],
        [0., 0., 1., 0.],
        [0., 1., 0., 0.],
        [0., 0., 0., 1.],
        [0., 0., 0., 0.]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_label_prop[True-True-False-row-1.0-5] __________________

k = 5, alpha = 1.0, norm_type = 'row', clamp = False, normalize = True
reset = True

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE122424C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0., 0., 0., 0.],
        [0., 0., 1., 0.],
        [0., 1., 0., 0.],
        [0., 0., 0., 1.],
        [0., 0., 0., 0.]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_label_prop[True-False-True-sym-0.0-3] __________________

k = 3, alpha = 0.0, norm_type = 'sym', clamp = True, normalize = False
reset = True

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03B99508>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0.0000, 0.0000, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.5000, 0.0000],
        [0.0000, 0.5000, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.0000, 0.5000],
        [0.0000, 0.0000, 0.0000, 0.0000]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_label_prop[True-False-True-sym-0.0-5] __________________

k = 5, alpha = 0.0, norm_type = 'sym', clamp = True, normalize = False
reset = True

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03AE6A88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0.0000, 0.0000, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.5774, 0.0000],
        [0.0000, 0.4472, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.0000, 0.5000],
        [0.0000, 0.0000, 0.0000, 0.0000]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_label_prop[True-False-True-sym-0.5-3] __________________

k = 3, alpha = 0.5, norm_type = 'sym', clamp = True, normalize = False
reset = True

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF031C1F08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0.0000, 0.0000, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.5774, 0.0000],
        [0.0000, 0.5774, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.0000, 0.5000],
        [0.0000, 0.0000, 0.0000, 0.0000]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_label_prop[True-False-True-sym-0.5-5] __________________

k = 5, alpha = 0.5, norm_type = 'sym', clamp = True, normalize = False
reset = True

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03B66D08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0.0000, 0.0000, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.4472, 0.0000],
        [0.0000, 0.5000, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.0000, 0.5774],
        [0.0000, 0.0000, 0.0000, 0.0000]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_label_prop[True-False-True-sym-1.0-3] __________________

k = 3, alpha = 1.0, norm_type = 'sym', clamp = True, normalize = False
reset = True

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07490748>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0.0000, 0.0000, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.4472, 0.0000],
        [0.0000, 0.5000, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.0000, 0.5000],
        [0.0000, 0.0000, 0.0000, 0.0000]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_label_prop[True-False-True-sym-1.0-5] __________________

k = 5, alpha = 1.0, norm_type = 'sym', clamp = True, normalize = False
reset = True

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07A108C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0.0000, 0.0000, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.5000, 0.0000],
        [0.0000, 0.5774, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.0000, 0.4472],
        [0.0000, 0.0000, 0.0000, 0.0000]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_label_prop[True-False-True-row-0.0-3] __________________

k = 3, alpha = 0.0, norm_type = 'row', clamp = True, normalize = False
reset = True

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0752C0C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0., 0., 0., 0.],
        [0., 0., 1., 0.],
        [0., 1., 0., 0.],
        [0., 0., 0., 1.],
        [0., 0., 0., 0.]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_label_prop[True-False-True-row-0.0-5] __________________

k = 5, alpha = 0.0, norm_type = 'row', clamp = True, normalize = False
reset = True

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07483588>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0., 0., 0., 0.],
        [0., 0., 1., 0.],
        [0., 1., 0., 0.],
        [0., 0., 0., 1.],
        [0., 0., 0., 0.]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_label_prop[True-False-True-row-0.5-3] __________________

k = 3, alpha = 0.5, norm_type = 'row', clamp = True, normalize = False
reset = True

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07E3BE88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0., 0., 0., 0.],
        [0., 0., 1., 0.],
        [0., 1., 0., 0.],
        [0., 0., 0., 1.],
        [0., 0., 0., 0.]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_label_prop[True-False-True-row-0.5-5] __________________

k = 5, alpha = 0.5, norm_type = 'row', clamp = True, normalize = False
reset = True

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF079E2088>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0., 0., 0., 0.],
        [0., 0., 1., 0.],
        [0., 1., 0., 0.],
        [0., 0., 0., 1.],
        [0., 0., 0., 0.]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_label_prop[True-False-True-row-1.0-3] __________________

k = 3, alpha = 1.0, norm_type = 'row', clamp = True, normalize = False
reset = True

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0754B108>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0., 0., 0., 0.],
        [0., 0., 1., 0.],
        [0., 1., 0., 0.],
        [0., 0., 0., 1.],
        [0., 0., 0., 0.]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_label_prop[True-False-True-row-1.0-5] __________________

k = 5, alpha = 1.0, norm_type = 'row', clamp = True, normalize = False
reset = True

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE123DC4C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0., 0., 0., 0.],
        [0., 0., 1., 0.],
        [0., 1., 0., 0.],
        [0., 0., 0., 1.],
        [0., 0., 0., 0.]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_label_prop[True-False-False-sym-0.0-3] _________________

k = 3, alpha = 0.0, norm_type = 'sym', clamp = False, normalize = False
reset = True

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03D83C88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0.0000, 0.0000, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.5774, 0.0000],
        [0.0000, 0.5000, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.0000, 0.5000],
        [0.0000, 0.0000, 0.0000, 0.0000]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_label_prop[True-False-False-sym-0.0-5] _________________

k = 5, alpha = 0.0, norm_type = 'sym', clamp = False, normalize = False
reset = True

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF073B6148>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0.0000, 0.0000, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.5000, 0.0000],
        [0.0000, 0.4472, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.0000, 0.5774],
        [0.0000, 0.0000, 0.0000, 0.0000]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_label_prop[True-False-False-sym-0.5-3] _________________

k = 3, alpha = 0.5, norm_type = 'sym', clamp = False, normalize = False
reset = True

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl

python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03113C48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0.0000, 0.0000, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.7071, 0.0000],
        [0.0000, 0.5000, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.0000, 0.5000],
        [0.0000, 0.0000, 0.0000, 0.0000]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_label_prop[True-False-False-sym-0.5-5] _________________

k = 5, alpha = 0.5, norm_type = 'sym', clamp = False, normalize = False
reset = True

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07A12B08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0.0000, 0.0000, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.5000, 0.0000],
        [0.0000, 0.5774, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.0000, 0.5000],
        [0.0000, 0.0000, 0.0000, 0.0000]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_label_prop[True-False-False-sym-1.0-3] _________________

k = 3, alpha = 1.0, norm_type = 'sym', clamp = False, normalize = False
reset = True

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF031BECC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0.0000, 0.0000, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.5774, 0.0000],
        [0.0000, 0.5000, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.0000, 0.5000],
        [0.0000, 0.0000, 0.0000, 0.0000]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_label_prop[True-False-False-sym-1.0-5] _________________

k = 5, alpha = 1.0, norm_type = 'sym', clamp = False, normalize = False
reset = True

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF070E7B08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0.0000, 0.0000, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.5000, 0.0000],
        [0.0000, 0.5000, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.0000, 0.5000],
        [0.0000, 0.0000, 0.0000, 0.0000]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_label_prop[True-False-False-row-0.0-3] _________________

k = 3, alpha = 0.0, norm_type = 'row', clamp = False, normalize = False
reset = True

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03B925C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0., 0., 0., 0.],
        [0., 0., 1., 0.],
        [0., 1., 0., 0.],
        [0., 0., 0., 1.],
        [0., 0., 0., 0.]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_label_prop[True-False-False-row-0.0-5] _________________

k = 5, alpha = 0.0, norm_type = 'row', clamp = False, normalize = False
reset = True

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03185CC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0., 0., 0., 0.],
        [0., 0., 1., 0.],
        [0., 1., 0., 0.],
        [0., 0., 0., 1.],
        [0., 0., 0., 0.]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_label_prop[True-False-False-row-0.5-3] _________________

k = 3, alpha = 0.5, norm_type = 'row', clamp = False, normalize = False
reset = True

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE121EB988>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0., 0., 0., 0.],
        [0., 0., 1., 0.],
        [0., 1., 0., 0.],
        [0., 0., 0., 1.],
        [0., 0., 0., 0.]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_label_prop[True-False-False-row-0.5-5] _________________

k = 5, alpha = 0.5, norm_type = 'row', clamp = False, normalize = False
reset = True

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1228F348>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0., 0., 0., 0.],
        [0., 0., 1., 0.],
        [0., 1., 0., 0.],
        [0., 0., 0., 1.],
        [0., 0., 0., 0.]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_label_prop[True-False-False-row-1.0-3] _________________

k = 3, alpha = 1.0, norm_type = 'row', clamp = False, normalize = False
reset = True

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03C3BB48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0., 0., 0., 0.],
        [0., 0., 1., 0.],
        [0., 1., 0., 0.],
        [0., 0., 0., 1.],
        [0., 0., 0., 0.]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_label_prop[True-False-False-row-1.0-5] _________________

k = 5, alpha = 1.0, norm_type = 'row', clamp = False, normalize = False
reset = True

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07E0CF08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0., 0., 0., 0.],
        [0., 0., 1., 0.],
        [0., 1., 0., 0.],
        [0., 0., 0., 1.],
        [0., 0., 0., 0.]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_label_prop[False-True-True-sym-0.0-3] __________________

k = 3, alpha = 0.0, norm_type = 'sym', clamp = True, normalize = True
reset = False

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF070E7088>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0.0000, 0.0000, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.5000, 0.0000],
        [0.0000, 0.4472, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.0000, 0.5000],
        [0.0000, 0.0000, 0.0000, 0.0000]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_label_prop[False-True-True-sym-0.0-5] __________________

k = 5, alpha = 0.0, norm_type = 'sym', clamp = True, normalize = True
reset = False

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03E32808>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0.0000, 0.0000, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.7071, 0.0000],
        [0.0000, 0.4472, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.0000, 0.5000],
        [0.0000, 0.0000, 0.0000, 0.0000]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_label_prop[False-True-True-sym-0.5-3] __________________

k = 3, alpha = 0.5, norm_type = 'sym', clamp = True, normalize = True
reset = False

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03113188>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0.0000, 0.0000, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.5000, 0.0000],
        [0.0000, 0.4472, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.0000, 0.5000],
        [0.0000, 0.0000, 0.0000, 0.0000]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_label_prop[False-True-True-sym-0.5-5] __________________

k = 5, alpha = 0.5, norm_type = 'sym', clamp = True, normalize = True
reset = False

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07410708>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0.0000, 0.0000, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.4472, 0.0000],
        [0.0000, 0.4472, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.0000, 0.5774],
        [0.0000, 0.0000, 0.0000, 0.0000]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_label_prop[False-True-True-sym-1.0-3] __________________

k = 3, alpha = 1.0, norm_type = 'sym', clamp = True, normalize = True
reset = False

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF078F1E88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0.0000, 0.0000, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.5000, 0.0000],
        [0.0000, 0.5774, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.0000, 0.4472],
        [0.0000, 0.0000, 0.0000, 0.0000]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_label_prop[False-True-True-sym-1.0-5] __________________

k = 5, alpha = 1.0, norm_type = 'sym', clamp = True, normalize = True
reset = False

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03E322C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0.0000, 0.0000, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.4472, 0.0000],
        [0.0000, 0.4472, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.0000, 0.7071],
        [0.0000, 0.0000, 0.0000, 0.0000]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_label_prop[False-True-True-row-0.0-3] __________________

k = 3, alpha = 0.0, norm_type = 'row', clamp = True, normalize = True
reset = False

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF079D8CC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0., 0., 0., 0.],
        [0., 0., 1., 0.],
        [0., 1., 0., 0.],
        [0., 0., 0., 1.],
        [0., 0., 0., 0.]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_label_prop[False-True-True-row-0.0-5] __________________

k = 5, alpha = 0.0, norm_type = 'row', clamp = True, normalize = True
reset = False

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF031ED5C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0., 0., 0., 0.],
        [0., 0., 1., 0.],
        [0., 1., 0., 0.],
        [0., 0., 0., 1.],
        [0., 0., 0., 0.]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_label_prop[False-True-True-row-0.5-3] __________________

k = 3, alpha = 0.5, norm_type = 'row', clamp = True, normalize = True
reset = False

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0791A108>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0., 0., 0., 0.],
        [0., 0., 1., 0.],
        [0., 1., 0., 0.],
        [0., 0., 0., 1.],
        [0., 0., 0., 0.]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_label_prop[False-True-True-row-0.5-5] __________________

k = 5, alpha = 0.5, norm_type = 'row', clamp = True, normalize = True
reset = False

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE12371B88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0., 0., 0., 0.],
        [0., 0., 1., 0.],
        [0., 1., 0., 0.],
        [0., 0., 0., 1.],
        [0., 0., 0., 0.]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_label_prop[False-True-True-row-1.0-3] __________________

k = 3, alpha = 1.0, norm_type = 'row', clamp = True, normalize = True
reset = False

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03DBCE48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0., 0., 0., 0.],
        [0., 0., 1., 0.],
        [0., 1., 0., 0.],
        [0., 0., 0., 1.],
        [0., 0., 0., 0.]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_label_prop[False-True-True-row-1.0-5] __________________

k = 5, alpha = 1.0, norm_type = 'row', clamp = True, normalize = True
reset = False

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE1239F4C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0., 0., 0., 0.],
        [0., 0., 1., 0.],
        [0., 1., 0., 0.],
        [0., 0., 0., 1.],
        [0., 0., 0., 0.]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_label_prop[False-True-False-sym-0.0-3] _________________

k = 3, alpha = 0.0, norm_type = 'sym', clamp = False, normalize = True
reset = False

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF073E7D48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0.0000, 0.0000, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.5000, 0.0000],
        [0.0000, 0.5000, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.0000, 0.5000],
        [0.0000, 0.0000, 0.0000, 0.0000]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_label_prop[False-True-False-sym-0.0-5] _________________

k = 5, alpha = 0.0, norm_type = 'sym', clamp = False, normalize = True
reset = False

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07E5FB08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0.0000, 0.0000, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.5000, 0.0000],
        [0.0000, 0.5000, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.0000, 0.5000],
        [0.0000, 0.0000, 0.0000, 0.0000]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_label_prop[False-True-False-sym-0.5-3] _________________

k = 3, alpha = 0.5, norm_type = 'sym', clamp = False, normalize = True
reset = False

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0791B8C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0.0000, 0.0000, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.5774, 0.0000],
        [0.0000, 0.5000, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.0000, 0.5000],
        [0.0000, 0.0000, 0.0000, 0.0000]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_label_prop[False-True-False-sym-0.5-5] _________________

k = 5, alpha = 0.5, norm_type = 'sym', clamp = False, normalize = True
reset = False

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF074CCD48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0.0000, 0.0000, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.4472, 0.0000],
        [0.0000, 0.5000, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.0000, 0.5774],
        [0.0000, 0.0000, 0.0000, 0.0000]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_label_prop[False-True-False-sym-1.0-3] _________________

k = 3, alpha = 1.0, norm_type = 'sym', clamp = False, normalize = True
reset = False

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07E45CC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0.0000, 0.0000, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.5000, 0.0000],
        [0.0000, 0.5000, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.0000, 0.4472],
        [0.0000, 0.0000, 0.0000, 0.0000]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_label_prop[False-True-False-sym-1.0-5] _________________

k = 5, alpha = 1.0, norm_type = 'sym', clamp = False, normalize = True
reset = False

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF031CB548>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0.0000, 0.0000, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.5000, 0.0000],
        [0.0000, 0.5000, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.0000, 0.5000],
        [0.0000, 0.0000, 0.0000, 0.0000]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_label_prop[False-True-False-row-0.0-3] _________________

k = 3, alpha = 0.0, norm_type = 'row', clamp = False, normalize = True
reset = False

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF075AA388>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0., 0., 0., 0.],
        [0., 0., 1., 0.],
        [0., 1., 0., 0.],
        [0., 0., 0., 1.],
        [0., 0., 0., 0.]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_label_prop[False-True-False-row-0.0-5] _________________

k = 5, alpha = 0.0, norm_type = 'row', clamp = False, normalize = True
reset = False

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07E59CC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0., 0., 0., 0.],
        [0., 0., 1., 0.],
        [0., 1., 0., 0.],
        [0., 0., 0., 1.],
        [0., 0., 0., 0.]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_label_prop[False-True-False-row-0.5-3] _________________

k = 3, alpha = 0.5, norm_type = 'row', clamp = False, normalize = True
reset = False

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF077A8B08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0., 0., 0., 0.],
        [0., 0., 1., 0.],
        [0., 1., 0., 0.],
        [0., 0., 0., 1.],
        [0., 0., 0., 0.]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_label_prop[False-True-False-row-0.5-5] _________________

k = 5, alpha = 0.5, norm_type = 'row', clamp = False, normalize = True
reset = False

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AE122B40C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0., 0., 0., 0.],
        [0., 0., 1., 0.],
        [0., 1., 0., 0.],
        [0., 0., 0., 1.],
        [0., 0., 0., 0.]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_label_prop[False-True-False-row-1.0-3] _________________

k = 3, alpha = 1.0, norm_type = 'row', clamp = False, normalize = True
reset = False

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07E68CC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0., 0., 0., 0.],
        [0., 0., 1., 0.],
        [0., 1., 0., 0.],
        [0., 0., 0., 1.],
        [0., 0., 0., 0.]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_label_prop[False-True-False-row-1.0-5] _________________

k = 5, alpha = 1.0, norm_type = 'row', clamp = False, normalize = True
reset = False

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF031F2188>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0., 0., 0., 0.],
        [0., 0., 1., 0.],
        [0., 1., 0., 0.],
        [0., 0., 0., 1.],
        [0., 0., 0., 0.]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_label_prop[False-False-True-sym-0.0-3] _________________

k = 3, alpha = 0.0, norm_type = 'sym', clamp = True, normalize = False
reset = False

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07F2A088>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0.0000, 0.0000, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.5000, 0.0000],
        [0.0000, 0.5000, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.0000, 0.5774],
        [0.0000, 0.0000, 0.0000, 0.0000]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_label_prop[False-False-True-sym-0.0-5] _________________

k = 5, alpha = 0.0, norm_type = 'sym', clamp = True, normalize = False
reset = False

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07883B48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0.0000, 0.0000, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.5000, 0.0000],
        [0.0000, 0.5000, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.0000, 0.4472],
        [0.0000, 0.0000, 0.0000, 0.0000]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_label_prop[False-False-True-sym-0.5-3] _________________

k = 3, alpha = 0.5, norm_type = 'sym', clamp = True, normalize = False
reset = False

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0749C6C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0.0000, 0.0000, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.7071, 0.0000],
        [0.0000, 0.4472, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.0000, 0.5000],
        [0.0000, 0.0000, 0.0000, 0.0000]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_label_prop[False-False-True-sym-0.5-5] _________________

k = 5, alpha = 0.5, norm_type = 'sym', clamp = True, normalize = False
reset = False

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0752C488>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0.0000, 0.0000, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.5000, 0.0000],
        [0.0000, 0.5000, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.0000, 0.5000],
        [0.0000, 0.0000, 0.0000, 0.0000]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_label_prop[False-False-True-sym-1.0-3] _________________

k = 3, alpha = 1.0, norm_type = 'sym', clamp = True, normalize = False
reset = False

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03C44108>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0.0000, 0.0000, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.5774, 0.0000],
        [0.0000, 0.5000, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.0000, 0.5000],
        [0.0000, 0.0000, 0.0000, 0.0000]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_label_prop[False-False-True-sym-1.0-5] _________________

k = 5, alpha = 1.0, norm_type = 'sym', clamp = True, normalize = False
reset = False

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0781F3C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0.0000, 0.0000, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.5000, 0.0000],
        [0.0000, 0.5000, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.0000, 0.5000],
        [0.0000, 0.0000, 0.0000, 0.0000]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_label_prop[False-False-True-row-0.0-3] _________________

k = 3, alpha = 0.0, norm_type = 'row', clamp = True, normalize = False
reset = False

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03C3A6C8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0., 0., 0., 0.],
        [0., 0., 1., 0.],
        [0., 1., 0., 0.],
        [0., 0., 0., 1.],
        [0., 0., 0., 0.]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_label_prop[False-False-True-row-0.0-5] _________________

k = 5, alpha = 0.0, norm_type = 'row', clamp = True, normalize = False
reset = False

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03C6FE88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0., 0., 0., 0.],
        [0., 0., 1., 0.],
        [0., 1., 0., 0.],
        [0., 0., 0., 1.],
        [0., 0., 0., 0.]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_label_prop[False-False-True-row-0.5-3] _________________

k = 3, alpha = 0.5, norm_type = 'row', clamp = True, normalize = False
reset = False

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF072E6708>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0., 0., 0., 0.],
        [0., 0., 1., 0.],
        [0., 1., 0., 0.],
        [0., 0., 0., 1.],
        [0., 0., 0., 0.]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_label_prop[False-False-True-row-0.5-5] _________________

k = 5, alpha = 0.5, norm_type = 'row', clamp = True, normalize = False
reset = False

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03C6FD08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0., 0., 0., 0.],
        [0., 0., 1., 0.],
        [0., 1., 0., 0.],
        [0., 0., 0., 1.],
        [0., 0., 0., 0.]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_label_prop[False-False-True-row-1.0-3] _________________

k = 3, alpha = 1.0, norm_type = 'row', clamp = True, normalize = False
reset = False

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03138488>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0., 0., 0., 0.],
        [0., 0., 1., 0.],
        [0., 1., 0., 0.],
        [0., 0., 0., 1.],
        [0., 0., 0., 0.]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
_________________ test_label_prop[False-False-True-row-1.0-5] _________________

k = 5, alpha = 1.0, norm_type = 'row', clamp = True, normalize = False
reset = False

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07FF3E88>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0., 0., 0., 0.],
        [0., 0., 1., 0.],
        [0., 1., 0., 0.],
        [0., 0., 0., 1.],
        [0., 0., 0., 0.]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_label_prop[False-False-False-sym-0.0-3] _________________

k = 3, alpha = 0.0, norm_type = 'sym', clamp = False, normalize = False
reset = False

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07E0F788>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0.0000, 0.0000, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.5000, 0.0000],
        [0.0000, 0.5000, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.0000, 0.5000],
        [0.0000, 0.0000, 0.0000, 0.0000]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_label_prop[False-False-False-sym-0.0-5] _________________

k = 5, alpha = 0.0, norm_type = 'sym', clamp = False, normalize = False
reset = False

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07E03588>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0.0000, 0.0000, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.4472, 0.0000],
        [0.0000, 0.5000, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.0000, 0.5000],
        [0.0000, 0.0000, 0.0000, 0.0000]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_label_prop[False-False-False-sym-0.5-3] _________________

k = 3, alpha = 0.5, norm_type = 'sym', clamp = False, normalize = False
reset = False

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AEFF95AF48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0.0000, 0.0000, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.5000, 0.0000],
        [0.0000, 0.4472, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.0000, 0.5000],
        [0.0000, 0.0000, 0.0000, 0.0000]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_label_prop[False-False-False-sym-0.5-5] _________________

k = 5, alpha = 0.5, norm_type = 'sym', clamp = False, normalize = False
reset = False

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03C39E08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0.0000, 0.0000, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.5774, 0.0000],
        [0.0000, 0.5000, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.0000, 0.4472],
        [0.0000, 0.0000, 0.0000, 0.0000]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_label_prop[False-False-False-sym-1.0-3] _________________

k = 3, alpha = 1.0, norm_type = 'sym', clamp = False, normalize = False
reset = False

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF03128488>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0.0000, 0.0000, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.5774, 0.0000],
        [0.0000, 0.4472, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.0000, 0.4472],
        [0.0000, 0.0000, 0.0000, 0.0000]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_label_prop[False-False-False-sym-1.0-5] _________________

k = 5, alpha = 1.0, norm_type = 'sym', clamp = False, normalize = False
reset = False

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07F0ECC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0.0000, 0.0000, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.5774, 0.0000],
        [0.0000, 0.4472, 0.0000, 0.0000],
        [0.0000, 0.0000, 0.0000, 0.5000],
        [0.0000, 0.0000, 0.0000, 0.0000]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_label_prop[False-False-False-row-0.0-3] _________________

k = 3, alpha = 0.0, norm_type = 'row', clamp = False, normalize = False
reset = False

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07F05E48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0., 0., 0., 0.],
        [0., 0., 1., 0.],
        [0., 1., 0., 0.],
        [0., 0., 0., 1.],
        [0., 0., 0., 0.]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_label_prop[False-False-False-row-0.0-5] _________________

k = 5, alpha = 0.0, norm_type = 'row', clamp = False, normalize = False
reset = False

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF031BBB08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0., 0., 0., 0.],
        [0., 0., 1., 0.],
        [0., 1., 0., 0.],
        [0., 0., 0., 1.],
        [0., 0., 0., 0.]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_label_prop[False-False-False-row-0.5-3] _________________

k = 3, alpha = 0.5, norm_type = 'row', clamp = False, normalize = False
reset = False

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07F2BC48>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0., 0., 0., 0.],
        [0., 0., 1., 0.],
        [0., 1., 0., 0.],
        [0., 0., 0., 1.],
        [0., 0., 0., 0.]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_label_prop[False-False-False-row-0.5-5] _________________

k = 5, alpha = 0.5, norm_type = 'row', clamp = False, normalize = False
reset = False

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07F0E048>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0., 0., 0., 0.],
        [0., 0., 1., 0.],
        [0., 1., 0., 0.],
        [0., 0., 0., 1.],
        [0., 0., 0., 0.]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_label_prop[False-False-False-row-1.0-3] _________________

k = 3, alpha = 1.0, norm_type = 'row', clamp = False, normalize = False
reset = False

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF07883A08>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0., 0., 0., 0.],
        [0., 0., 1., 0.],
        [0., 1., 0., 0.],
        [0., 0., 0., 1.],
        [0., 0., 0., 0.]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
________________ test_label_prop[False-False-False-row-1.0-5] _________________

k = 5, alpha = 1.0, norm_type = 'row', clamp = False, normalize = False
reset = False

    @pytest.mark.parametrize('k', [3, 5])
    @pytest.mark.parametrize('alpha', [0., 0.5, 1.])
    @pytest.mark.parametrize('norm_type', ['sym', 'row'])
    @pytest.mark.parametrize('clamp', [True, False])
    @pytest.mark.parametrize('normalize', [True, False])
    @pytest.mark.parametrize('reset', [True, False])
    def test_label_prop(k, alpha, norm_type, clamp, normalize, reset):
        dev = F.ctx()
        num_nodes = 5
        num_edges = 20
        num_classes = 4
        g = dgl.rand_graph(num_nodes, num_edges).to(dev)
        labels = th.tensor([0, 2, 1, 3, 0]).long().to(dev)
        ml_labels = th.rand(num_nodes, num_classes).to(dev) > 0.7
        mask = th.tensor([0, 1, 1, 1, 0]).bool().to(dev)
        model = nn.LabelPropagation(k, alpha, norm_type, clamp, normalize, reset)
>       model(g, labels, mask)

tests\pytorch\test_nn.py:1601: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
c:\program files\python36\lib\site-packages\torch\nn\modules\module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python\dgl\nn\pytorch\utils.py:511: in forward
    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python\dgl\heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python\dgl\core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python\dgl\core.py:332: in invoke_gspmm
    z = op(graph, x)
python\dgl\ops\spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python\dgl\ops\spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x000000AF0754DBC8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0., 0., 0., 0.],
        [0., 0., 1., 0.],
        [0., 1., 0., 0.],
        [0., 0., 0., 1.],
        [0., 0., 0., 0.]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python\dgl\backend\pytorch\sparse.py:720: TypeError
============================== warnings summary ===============================
python\dgl\backend\backend.py:1741
  C:\Jenkins\workspace\dgl_PR-4648\python\dgl\backend\backend.py:1741: DeprecationWarning: invalid escape sequence \P
    """

tests/pytorch/test_dataloader.py::test_shadow[0]
  c:\program files\python36\lib\site-packages\numpy\matrixlib\defmatrix.py:69: PendingDeprecationWarning: the matrix subclass is not the recommended way to represent matrices or deal with linear algebra (see https://docs.scipy.org/doc/numpy/user/numpy-for-matlab-users.html). Please adjust your code to use regular ndarray.
    return matrix(data, dtype=dtype, copy=False)

tests/pytorch/test_dataloader.py: 80 warnings
  C:\Jenkins\workspace\dgl_PR-4648\python\dgl\dataloading\dataloader.py:863: DGLWarning: Dataloader CPU affinity opt is not enabled, consider switching it on (see enable_cpu_affinity() or CPU best practices for DGL [https://docs.dgl.ai/tutorials/cpu/cpu_best_practises.html])
    dgl_warning(f'Dataloader CPU affinity opt is not enabled, consider switching it on '

tests/pytorch/test_dataloader.py: 48 warnings
  C:\Jenkins\workspace\dgl_PR-4648\python\dgl\dataloading\dataloader.py:83: DGLWarning: The current output_nodes are out of order even if set shuffle to False in Dataloader, the reason is that the current version of torch dose not support stable sort. Please update torch to 1.10.0 or higher to fix it.
    'The current output_nodes are out of order even if set shuffle '

tests/pytorch/test_nn.py: 11 warnings
tests/pytorch/test_pickle.py: 1 warning
  C:\Jenkins\workspace\dgl_PR-4648\python\dgl\heterograph.py:72: DGLWarning: Recommend creating graphs by `dgl.graph(data)` instead of `dgl.DGLGraph(data)`.
    dgl_warning('Recommend creating graphs by `dgl.graph(data)`'

tests/pytorch/test_nn.py::test_set_trans
  C:\Jenkins\workspace\dgl_PR-4648\python\dgl\nn\pytorch\glob.py:700: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).
    lengths_x = th.tensor(lengths_x, dtype=th.int64, device=device)

tests/pytorch/test_nn.py::test_set_trans
  C:\Jenkins\workspace\dgl_PR-4648\python\dgl\nn\pytorch\glob.py:701: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).
    lengths_mem = th.tensor(lengths_mem, dtype=th.int64, device=device)

tests/pytorch/test_nn.py::test_dense_cheb_conv[1]
tests/pytorch/test_nn.py::test_dense_cheb_conv[2]
  C:\Jenkins\workspace\dgl_PR-4648\python\dgl\heterograph.py:84: DGLWarning: Keyword arguments ['readonly'] are deprecated in v0.5, and can be safely removed in all cases.
    ' removed in all cases.'.format(list(deprecate_kwargs.keys())))

-- Docs: https://docs.pytest.org/en/stable/how-to/capture-warnings.html
--- generated xml file: C:\Jenkins\workspace\dgl_PR-4648\pytest_backend.xml ---
============================ slowest 100 durations ============================
10.97s call     tests/pytorch/test_dataloader.py::test_edge_dataloader_excludes[1-sampler1-None-False]
10.86s call     tests/pytorch/test_dataloader.py::test_edge_dataloader_excludes[1-sampler0-None-False]
10.77s call     tests/pytorch/test_dataloader.py::test_node_dataloader[True-cpu-neighbor2-idtype0]
10.75s call     tests/pytorch/test_dataloader.py::test_node_dataloader[True-cpu-neighbor2-idtype1]
10.39s call     tests/pytorch/test_dataloader.py::test_node_dataloader[False-cpu-full-idtype0]
9.99s call     tests/pytorch/test_dataloader.py::test_node_dataloader[False-cpu-full-idtype1]
9.63s call     tests/pytorch/test_dataloader.py::test_node_dataloader[False-cpu-neighbor-idtype1]
9.55s call     tests/pytorch/test_dataloader.py::test_node_dataloader[False-cpu-neighbor2-idtype0]
9.43s call     tests/pytorch/test_dataloader.py::test_node_dataloader[True-cpu-full-idtype0]
9.33s call     tests/pytorch/test_dataloader.py::test_node_dataloader[True-cpu-neighbor-idtype1]
9.32s call     tests/pytorch/test_dataloader.py::test_node_dataloader[True-cpu-full-idtype1]
9.27s call     tests/pytorch/test_dataloader.py::test_node_dataloader[True-cpu-neighbor-idtype0]
9.20s call     tests/pytorch/test_dataloader.py::test_node_dataloader[False-cpu-neighbor2-idtype1]
9.18s call     tests/pytorch/test_dataloader.py::test_neighbor_nonuniform[False-cpu-idtype1]
9.17s call     tests/pytorch/test_dataloader.py::test_node_dataloader[False-cpu-neighbor-idtype0]
9.00s call     tests/pytorch/test_dataloader.py::test_shadow[0]
7.32s call     tests/pytorch/test_dataloader.py::test_neighbor_nonuniform[True-cpu-idtype0]
6.69s call     tests/pytorch/test_dataloader.py::test_neighbor_nonuniform[True-cpu-idtype1]
6.26s call     tests/pytorch/test_dataloader.py::test_neighbor_nonuniform[False-cpu-idtype0]
3.16s call     tests/pytorch/test_dataloader.py::test_shadow[4]
2.84s call     tests/pytorch/test_dataloader.py::test_saint[node-4]
2.83s call     tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-7.4-scalers1-aggregators1-32-32]
2.82s call     tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-7.4-scalers1-aggregators1-32-16]
2.82s call     tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-7.4-scalers1-aggregators0-16-16]
2.81s call     tests/pytorch/test_dataloader.py::test_saint[walk-4]
2.81s call     tests/pytorch/test_dataloader.py::test_saint[edge-4]
2.65s call     tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-7.4-scalers0-aggregators1-32-16]
2.64s call     tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-7.4-scalers0-aggregators1-32-32]
2.46s call     tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-7.4-scalers1-aggregators0-32-16]
2.08s call     tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-7.4-scalers1-aggregators0-16-32]
2.08s call     tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-7.4-scalers1-aggregators0-32-32]
1.91s call     tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-7.4-scalers1-aggregators1-16-32]
1.89s call     tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-7.4-scalers1-aggregators1-16-16]
1.81s call     tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-7.4-scalers0-aggregators1-16-32]
1.73s call     tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-7.4-scalers1-aggregators2-16-16]
1.71s call     tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-7.4-scalers1-aggregators2-32-16]
1.56s call     tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-2.5-scalers0-aggregators0-32-32]
1.54s call     tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-7.4-scalers0-aggregators2-32-32]
1.48s call     tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-7.4-scalers1-aggregators2-16-32]
1.43s call     tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-7.4-scalers0-aggregators2-16-32]
1.35s call     tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-2.5-scalers0-aggregators0-16-16]
1.35s call     tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-7.4-scalers1-aggregators2-32-32]
1.30s call     tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-2.5-scalers0-aggregators1-16-16]
1.29s call     tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-2.5-scalers0-aggregators1-16-32]
1.26s call     tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-7.4-scalers0-aggregators2-16-16]
1.24s call     tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-7.4-scalers0-aggregators0-16-32]
1.21s call     tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-2.5-scalers0-aggregators0-32-16]
1.21s call     tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-2.5-scalers1-aggregators1-32-16]
1.20s call     tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-2.5-scalers0-aggregators0-16-32]
1.12s call     tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-7.4-scalers1-aggregators1-16-32]
1.08s call     tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-7.4-scalers1-aggregators0-16-32]
1.01s call     tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-2.5-scalers1-aggregators0-16-32]
0.96s call     tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-2.5-scalers1-aggregators1-16-16]
0.95s call     tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-7.4-scalers0-aggregators0-16-16]
0.94s call     tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-2.5-scalers0-aggregators0-32-16]
0.93s call     tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-2.5-scalers1-aggregators0-16-16]
0.93s call     tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-2.5-scalers1-aggregators1-16-32]
0.93s call     tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-7.4-scalers0-aggregators0-16-16]
0.93s call     tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-7.4-scalers1-aggregators0-32-32]
0.91s call     tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-2.5-scalers0-aggregators0-16-16]
0.85s call     tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-7.4-scalers0-aggregators0-32-32]
0.83s call     tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-7.4-scalers1-aggregators2-32-32]
0.82s call     tests/pytorch/test_dataloader.py::test_saint[edge-0]
0.80s call     tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-7.4-scalers1-aggregators1-32-16]
0.79s call     tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-7.4-scalers0-aggregators1-16-32]
0.76s call     tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-2.5-scalers1-aggregators0-16-32]
0.76s call     tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-2.5-scalers0-aggregators0-32-32]
0.76s call     tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-2.5-scalers0-aggregators1-32-16]
0.75s call     tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-2.5-scalers0-aggregators1-16-32]
0.75s call     tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-2.5-scalers1-aggregators1-32-16]
0.75s call     tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-2.5-scalers1-aggregators0-32-16]
0.74s call     tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-7.4-scalers0-aggregators1-16-16]
0.74s call     tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-7.4-scalers0-aggregators0-32-32]
0.74s call     tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-7.4-scalers1-aggregators0-16-16]
0.73s call     tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-7.4-scalers0-aggregators2-32-16]
0.73s call     tests/pytorch/test_dataloader.py::test_saint[walk-0]
0.69s call     tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-2.5-scalers0-aggregators0-32-32]
0.69s call     tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-7.4-scalers1-aggregators2-32-16]
0.69s call     tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-2.5-scalers0-aggregators2-32-16]
0.67s call     tests/pytorch/test_dataloader.py::test_edge_dataloader[True-cpu-neg_sampler0-full-idtype1]
0.67s call     tests/pytorch/test_dataloader.py::test_edge_dataloader[True-cpu-neg_sampler0-full-idtype0]
0.67s call     tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-7.4-scalers1-aggregators1-32-32]
0.65s call     tests/pytorch/test_dataloader.py::test_edge_dataloader[False-cpu-neg_sampler0-full-idtype1]
0.65s call     tests/pytorch/test_dataloader.py::test_edge_dataloader[True-cpu-neg_sampler2-full-idtype0]
0.64s call     tests/pytorch/test_dataloader.py::test_saint[node-0]
0.64s call     tests/pytorch/test_dataloader.py::test_edge_dataloader[False-cpu-neg_sampler0-full-idtype0]
0.63s call     tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-2.5-scalers1-aggregators1-16-32]
0.63s call     tests/pytorch/test_dataloader.py::test_edge_dataloader[True-cpu-neg_sampler2-full-idtype1]
0.63s call     tests/pytorch/test_dataloader.py::test_edge_dataloader[True-cpu-neg_sampler1-full-idtype0]
0.62s call     tests/pytorch/test_dataloader.py::test_edge_dataloader[False-cpu-neg_sampler1-full-idtype0]
0.60s call     tests/pytorch/test_dataloader.py::test_edge_dataloader[True-cpu-neg_sampler1-full-idtype1]
0.60s call     tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-2.5-scalers1-aggregators1-32-32]
0.59s call     tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-2.5-scalers0-aggregators1-16-32]
0.58s call     tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-7.4-scalers0-aggregators1-16-32]
0.58s call     tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-2.5-scalers1-aggregators1-32-16]
0.58s call     tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-2.5-scalers1-aggregators2-16-32]
0.58s call     tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-7.4-scalers0-aggregators1-16-16]
0.57s call     tests/pytorch/test_dataloader.py::test_edge_dataloader[False-cpu-neg_sampler2-full-idtype0]
0.57s call     tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-2.5-scalers0-aggregators2-16-32]
0.57s call     tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-2.5-scalers1-aggregators0-32-16]
=========================== short test summary info ===========================
FAILED tests/pytorch/test_geometry.py::test_knn_cpu[True-euclidean-bruteforce-blas]
FAILED tests/pytorch/test_geometry.py::test_knn_cpu[True-euclidean-bruteforce]
FAILED tests/pytorch/test_geometry.py::test_knn_cpu[True-euclidean-kd-tree]
FAILED tests/pytorch/test_geometry.py::test_knn_cpu[True-cosine-bruteforce-blas]
FAILED tests/pytorch/test_geometry.py::test_knn_cpu[True-cosine-bruteforce]
FAILED tests/pytorch/test_geometry.py::test_knn_cpu[True-cosine-kd-tree] - Ty...
FAILED tests/pytorch/test_nn.py::test_graph_conv0[1] - TypeError: empty_conte...
FAILED tests/pytorch/test_nn.py::test_graph_conv0[2] - TypeError: empty_conte...
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-True-none-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-True-none-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-True-none-g1-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-True-none-g1-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-True-none-g2-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-True-none-g2-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-True-none-g3-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-True-none-g3-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-True-none-g4-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-True-none-g4-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-True-none-g5-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-True-none-g5-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-True-none-g6-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-True-none-g6-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-True-none-g7-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-True-none-g7-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-True-both-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-True-both-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-True-both-g1-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-True-both-g1-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-True-both-g2-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-True-both-g2-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-True-both-g3-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-True-both-g3-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-True-both-g4-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-True-both-g4-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-True-both-g5-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-True-both-g5-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-True-both-g6-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-True-both-g6-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-True-both-g7-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-True-both-g7-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-True-right-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-True-right-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-True-right-g1-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-True-right-g1-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-True-right-g2-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-True-right-g2-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-True-right-g3-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-True-right-g3-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-True-right-g4-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-True-right-g4-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-True-right-g5-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-True-right-g5-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-True-right-g6-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-True-right-g6-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-True-right-g7-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-True-right-g7-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-True-left-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-True-left-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-True-left-g1-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-True-left-g1-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-True-left-g2-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-True-left-g2-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-True-left-g3-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-True-left-g3-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-True-left-g4-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-True-left-g4-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-True-left-g5-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-True-left-g5-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-True-left-g6-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-True-left-g6-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-True-left-g7-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-True-left-g7-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-False-none-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-False-none-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-False-none-g1-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-False-none-g1-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-False-none-g2-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-False-none-g2-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-False-none-g3-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-False-none-g3-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-False-none-g4-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-False-none-g4-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-False-none-g5-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-False-none-g5-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-False-none-g6-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-False-none-g6-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-False-none-g7-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-False-none-g7-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-False-both-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-False-both-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-False-both-g1-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-False-both-g1-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-False-both-g2-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-False-both-g2-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-False-both-g3-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-False-both-g3-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-False-both-g4-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-False-both-g4-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-False-both-g5-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-False-both-g5-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-False-both-g6-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-False-both-g6-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-False-both-g7-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-False-both-g7-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-False-right-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-False-right-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-False-right-g1-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-False-right-g1-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-False-right-g2-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-False-right-g2-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-False-right-g3-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-False-right-g3-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-False-right-g4-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-False-right-g4-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-False-right-g5-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-False-right-g5-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-False-right-g6-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-False-right-g6-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-False-right-g7-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-False-right-g7-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-False-left-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-False-left-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-False-left-g1-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-False-left-g1-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-False-left-g2-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-False-left-g2-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-False-left-g3-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-False-left-g3-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-False-left-g4-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-False-left-g4-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-False-left-g5-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-False-left-g5-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-False-left-g6-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-False-left-g6-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-False-left-g7-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-True-False-left-g7-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-True-none-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-True-none-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-True-none-g1-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-True-none-g1-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-True-none-g2-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-True-none-g2-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-True-none-g3-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-True-none-g3-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-True-none-g4-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-True-none-g4-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-True-none-g5-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-True-none-g5-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-True-none-g6-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-True-none-g6-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-True-none-g7-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-True-none-g7-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-True-both-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-True-both-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-True-both-g1-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-True-both-g1-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-True-both-g2-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-True-both-g2-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-True-both-g3-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-True-both-g3-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-True-both-g4-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-True-both-g4-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-True-both-g5-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-True-both-g5-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-True-both-g6-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-True-both-g6-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-True-both-g7-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-True-both-g7-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-True-right-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-True-right-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-True-right-g1-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-True-right-g1-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-True-right-g2-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-True-right-g2-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-True-right-g3-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-True-right-g3-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-True-right-g4-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-True-right-g4-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-True-right-g5-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-True-right-g5-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-True-right-g6-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-True-right-g6-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-True-right-g7-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-True-right-g7-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-True-left-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-True-left-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-True-left-g1-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-True-left-g1-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-True-left-g2-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-True-left-g2-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-True-left-g3-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-True-left-g3-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-True-left-g4-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-True-left-g4-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-True-left-g5-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-True-left-g5-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-True-left-g6-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-True-left-g6-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-True-left-g7-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-True-left-g7-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-False-none-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-False-none-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-False-none-g1-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-False-none-g1-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-False-none-g2-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-False-none-g2-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-False-none-g3-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-False-none-g3-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-False-none-g4-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-False-none-g4-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-False-none-g5-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-False-none-g5-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-False-none-g6-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-False-none-g6-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-False-none-g7-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-False-none-g7-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-False-both-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-False-both-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-False-both-g1-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-False-both-g1-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-False-both-g2-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-False-both-g2-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-False-both-g3-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-False-both-g3-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-False-both-g4-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-False-both-g4-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-False-both-g5-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-False-both-g5-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-False-both-g6-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-False-both-g6-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-False-both-g7-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-False-both-g7-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-False-right-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-False-right-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-False-right-g1-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-False-right-g1-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-False-right-g2-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-False-right-g2-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-False-right-g3-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-False-right-g3-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-False-right-g4-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-False-right-g4-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-False-right-g5-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-False-right-g5-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-False-right-g6-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-False-right-g6-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-False-right-g7-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-False-right-g7-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-False-left-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-False-left-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-False-left-g1-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-False-left-g1-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-False-left-g2-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-False-left-g2-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-False-left-g3-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-False-left-g3-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-False-left-g4-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-False-left-g4-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-False-left-g5-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-False-left-g5-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-False-left-g6-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-False-left-g6-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-False-left-g7-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[1-False-False-left-g7-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-True-none-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-True-none-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-True-none-g1-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-True-none-g1-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-True-none-g2-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-True-none-g2-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-True-none-g3-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-True-none-g3-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-True-none-g4-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-True-none-g4-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-True-none-g5-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-True-none-g5-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-True-none-g6-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-True-none-g6-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-True-none-g7-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-True-none-g7-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-True-both-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-True-both-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-True-both-g1-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-True-both-g1-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-True-both-g2-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-True-both-g2-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-True-both-g3-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-True-both-g3-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-True-both-g4-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-True-both-g4-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-True-both-g5-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-True-both-g5-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-True-both-g6-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-True-both-g6-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-True-both-g7-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-True-both-g7-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-True-right-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-True-right-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-True-right-g1-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-True-right-g1-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-True-right-g2-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-True-right-g2-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-True-right-g3-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-True-right-g3-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-True-right-g4-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-True-right-g4-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-True-right-g5-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-True-right-g5-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-True-right-g6-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-True-right-g6-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-True-right-g7-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-True-right-g7-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-True-left-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-True-left-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-True-left-g1-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-True-left-g1-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-True-left-g2-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-True-left-g2-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-True-left-g3-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-True-left-g3-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-True-left-g4-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-True-left-g4-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-True-left-g5-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-True-left-g5-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-True-left-g6-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-True-left-g6-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-True-left-g7-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-True-left-g7-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-False-none-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-False-none-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-False-none-g1-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-False-none-g1-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-False-none-g2-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-False-none-g2-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-False-none-g3-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-False-none-g3-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-False-none-g4-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-False-none-g4-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-False-none-g5-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-False-none-g5-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-False-none-g6-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-False-none-g6-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-False-none-g7-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-False-none-g7-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-False-both-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-False-both-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-False-both-g1-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-False-both-g1-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-False-both-g2-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-False-both-g2-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-False-both-g3-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-False-both-g3-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-False-both-g4-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-False-both-g4-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-False-both-g5-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-False-both-g5-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-False-both-g6-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-False-both-g6-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-False-both-g7-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-False-both-g7-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-False-right-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-False-right-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-False-right-g1-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-False-right-g1-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-False-right-g2-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-False-right-g2-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-False-right-g3-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-False-right-g3-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-False-right-g4-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-False-right-g4-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-False-right-g5-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-False-right-g5-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-False-right-g6-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-False-right-g6-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-False-right-g7-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-False-right-g7-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-False-left-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-False-left-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-False-left-g1-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-False-left-g1-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-False-left-g2-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-False-left-g2-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-False-left-g3-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-False-left-g3-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-False-left-g4-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-False-left-g4-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-False-left-g5-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-False-left-g5-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-False-left-g6-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-False-left-g6-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-False-left-g7-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-True-False-left-g7-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-True-none-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-True-none-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-True-none-g1-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-True-none-g1-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-True-none-g2-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-True-none-g2-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-True-none-g3-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-True-none-g3-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-True-none-g4-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-True-none-g4-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-True-none-g5-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-True-none-g5-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-True-none-g6-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-True-none-g6-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-True-none-g7-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-True-none-g7-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-True-both-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-True-both-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-True-both-g1-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-True-both-g1-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-True-both-g2-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-True-both-g2-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-True-both-g3-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-True-both-g3-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-True-both-g4-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-True-both-g4-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-True-both-g5-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-True-both-g5-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-True-both-g6-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-True-both-g6-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-True-both-g7-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-True-both-g7-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-True-right-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-True-right-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-True-right-g1-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-True-right-g1-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-True-right-g2-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-True-right-g2-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-True-right-g3-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-True-right-g3-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-True-right-g4-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-True-right-g4-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-True-right-g5-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-True-right-g5-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-True-right-g6-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-True-right-g6-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-True-right-g7-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-True-right-g7-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-True-left-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-True-left-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-True-left-g1-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-True-left-g1-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-True-left-g2-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-True-left-g2-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-True-left-g3-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-True-left-g3-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-True-left-g4-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-True-left-g4-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-True-left-g5-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-True-left-g5-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-True-left-g6-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-True-left-g6-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-True-left-g7-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-True-left-g7-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-False-none-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-False-none-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-False-none-g1-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-False-none-g1-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-False-none-g2-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-False-none-g2-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-False-none-g3-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-False-none-g3-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-False-none-g4-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-False-none-g4-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-False-none-g5-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-False-none-g5-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-False-none-g6-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-False-none-g6-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-False-none-g7-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-False-none-g7-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-False-both-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-False-both-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-False-both-g1-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-False-both-g1-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-False-both-g2-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-False-both-g2-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-False-both-g3-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-False-both-g3-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-False-both-g4-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-False-both-g4-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-False-both-g5-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-False-both-g5-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-False-both-g6-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-False-both-g6-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-False-both-g7-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-False-both-g7-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-False-right-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-False-right-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-False-right-g1-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-False-right-g1-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-False-right-g2-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-False-right-g2-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-False-right-g3-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-False-right-g3-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-False-right-g4-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-False-right-g4-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-False-right-g5-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-False-right-g5-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-False-right-g6-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-False-right-g6-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-False-right-g7-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-False-right-g7-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-False-left-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-False-left-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-False-left-g1-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-False-left-g1-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-False-left-g2-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-False-left-g2-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-False-left-g3-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-False-left-g3-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-False-left-g4-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-False-left-g4-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-False-left-g5-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-False-left-g5-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-False-left-g6-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-False-left-g6-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-False-left-g7-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv[2-False-False-left-g7-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight[1-True-True-none-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight[1-True-True-none-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight[1-True-True-both-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight[1-True-True-both-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight[1-True-True-right-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight[1-True-True-right-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight[1-True-False-none-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight[1-True-False-none-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight[1-True-False-both-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight[1-True-False-both-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight[1-True-False-right-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight[1-True-False-right-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight[1-False-True-none-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight[1-False-True-none-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight[1-False-True-both-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight[1-False-True-both-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight[1-False-True-right-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight[1-False-True-right-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight[1-False-False-none-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight[1-False-False-none-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight[1-False-False-both-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight[1-False-False-both-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight[1-False-False-right-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight[1-False-False-right-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight[2-True-True-none-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight[2-True-True-none-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight[2-True-True-both-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight[2-True-True-both-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight[2-True-True-right-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight[2-True-True-right-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight[2-True-False-none-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight[2-True-False-none-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight[2-True-False-both-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight[2-True-False-both-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight[2-True-False-right-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight[2-True-False-right-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight[2-False-True-none-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight[2-False-True-none-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight[2-False-True-both-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight[2-False-True-both-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight[2-False-True-right-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight[2-False-True-right-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight[2-False-False-none-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight[2-False-False-none-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight[2-False-False-both-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight[2-False-False-both-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight[2-False-False-right-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight[2-False-False-right-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[1-True-True-none-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[1-True-True-none-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[1-True-True-both-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[1-True-True-both-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[1-True-True-right-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[1-True-True-right-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[1-True-False-none-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[1-True-False-none-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[1-True-False-both-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[1-True-False-both-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[1-True-False-right-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[1-True-False-right-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[1-False-True-none-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[1-False-True-none-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[1-False-True-both-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[1-False-True-both-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[1-False-True-right-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[1-False-True-right-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[1-False-False-none-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[1-False-False-none-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[1-False-False-both-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[1-False-False-both-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[1-False-False-right-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[1-False-False-right-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[2-True-True-none-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[2-True-True-none-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[2-True-True-both-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[2-True-True-both-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[2-True-True-right-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[2-True-True-right-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[2-True-False-none-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[2-True-False-none-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[2-True-False-both-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[2-True-False-both-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[2-True-False-right-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[2-True-False-right-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[2-False-True-none-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[2-False-True-none-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[2-False-True-both-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[2-False-True-both-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[2-False-True-right-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[2-False-True-right-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[2-False-False-none-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[2-False-False-none-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[2-False-False-both-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[2-False-False-both-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[2-False-False-right-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[2-False-False-right-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[1-True-True-none-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[1-True-True-none-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[1-True-True-none-g1-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[1-True-True-none-g1-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[1-True-True-both-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[1-True-True-both-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[1-True-True-both-g1-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[1-True-True-both-g1-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[1-True-True-right-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[1-True-True-right-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[1-True-True-right-g1-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[1-True-True-right-g1-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[1-True-False-none-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[1-True-False-none-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[1-True-False-none-g1-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[1-True-False-none-g1-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[1-True-False-both-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[1-True-False-both-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[1-True-False-both-g1-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[1-True-False-both-g1-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[1-True-False-right-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[1-True-False-right-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[1-True-False-right-g1-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[1-True-False-right-g1-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[1-False-True-none-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[1-False-True-none-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[1-False-True-none-g1-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[1-False-True-none-g1-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[1-False-True-both-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[1-False-True-both-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[1-False-True-both-g1-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[1-False-True-both-g1-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[1-False-True-right-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[1-False-True-right-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[1-False-True-right-g1-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[1-False-True-right-g1-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[1-False-False-none-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[1-False-False-none-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[1-False-False-none-g1-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[1-False-False-none-g1-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[1-False-False-both-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[1-False-False-both-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[1-False-False-both-g1-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[1-False-False-both-g1-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[1-False-False-right-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[1-False-False-right-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[1-False-False-right-g1-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[1-False-False-right-g1-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[2-True-True-none-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[2-True-True-none-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[2-True-True-none-g1-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[2-True-True-none-g1-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[2-True-True-both-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[2-True-True-both-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[2-True-True-both-g1-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[2-True-True-both-g1-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[2-True-True-right-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[2-True-True-right-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[2-True-True-right-g1-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[2-True-True-right-g1-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[2-True-False-none-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[2-True-False-none-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[2-True-False-none-g1-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[2-True-False-none-g1-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[2-True-False-both-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[2-True-False-both-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[2-True-False-both-g1-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[2-True-False-both-g1-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[2-True-False-right-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[2-True-False-right-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[2-True-False-right-g1-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[2-True-False-right-g1-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[2-False-True-none-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[2-False-True-none-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[2-False-True-none-g1-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[2-False-True-none-g1-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[2-False-True-both-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[2-False-True-both-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[2-False-True-both-g1-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[2-False-True-both-g1-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[2-False-True-right-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[2-False-True-right-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[2-False-True-right-g1-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[2-False-True-right-g1-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[2-False-False-none-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[2-False-False-none-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[2-False-False-none-g1-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[2-False-False-none-g1-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[2-False-False-both-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[2-False-False-both-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[2-False-False-both-g1-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[2-False-False-both-g1-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[2-False-False-right-g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[2-False-False-right-g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[2-False-False-right-g1-idtype0]
FAILED tests/pytorch/test_nn.py::test_graph_conv_bi[2-False-False-right-g1-idtype1]
FAILED tests/pytorch/test_nn.py::test_tagconv[1] - TypeError: empty_context()...
FAILED tests/pytorch/test_nn.py::test_tagconv[2] - TypeError: empty_context()...
FAILED tests/pytorch/test_nn.py::test_set2set - TypeError: empty_context() go...
FAILED tests/pytorch/test_nn.py::test_glob_att_pool - TypeError: empty_contex...
FAILED tests/pytorch/test_nn.py::test_simple_pool - TypeError: empty_context(...
FAILED tests/pytorch/test_nn.py::test_rgcn[1-idtype0] - TypeError: empty_cont...
FAILED tests/pytorch/test_nn.py::test_rgcn[1-idtype1] - TypeError: empty_cont...
FAILED tests/pytorch/test_nn.py::test_rgcn[8-idtype0] - TypeError: empty_cont...
FAILED tests/pytorch/test_nn.py::test_rgcn[8-idtype1] - TypeError: empty_cont...
FAILED tests/pytorch/test_nn.py::test_rgcn[32-idtype0] - TypeError: empty_con...
FAILED tests/pytorch/test_nn.py::test_rgcn[32-idtype1] - TypeError: empty_con...
FAILED tests/pytorch/test_nn.py::test_rgcn_default_nbasis[1-idtype0] - TypeEr...
FAILED tests/pytorch/test_nn.py::test_rgcn_default_nbasis[1-idtype1] - TypeEr...
FAILED tests/pytorch/test_nn.py::test_rgcn_default_nbasis[10-idtype0] - TypeE...
FAILED tests/pytorch/test_nn.py::test_rgcn_default_nbasis[10-idtype1] - TypeE...
FAILED tests/pytorch/test_nn.py::test_rgcn_default_nbasis[40-idtype0] - TypeE...
FAILED tests/pytorch/test_nn.py::test_rgcn_default_nbasis[40-idtype1] - TypeE...
FAILED tests/pytorch/test_nn.py::test_gat_conv[1-1-g0-idtype0] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gat_conv[1-1-g0-idtype1] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gat_conv[1-1-g1-idtype0] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gat_conv[1-1-g1-idtype1] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gat_conv[1-1-g2-idtype0] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gat_conv[1-1-g2-idtype1] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gat_conv[1-1-g3-idtype0] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gat_conv[1-1-g3-idtype1] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gat_conv[1-1-g4-idtype0] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gat_conv[1-1-g4-idtype1] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gat_conv[1-1-g5-idtype0] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gat_conv[1-1-g5-idtype1] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gat_conv[1-1-g6-idtype0] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gat_conv[1-1-g6-idtype1] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gat_conv[1-5-g0-idtype0] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gat_conv[1-5-g0-idtype1] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gat_conv[1-5-g1-idtype0] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gat_conv[1-5-g1-idtype1] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gat_conv[1-5-g2-idtype0] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gat_conv[1-5-g2-idtype1] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gat_conv[1-5-g3-idtype0] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gat_conv[1-5-g3-idtype1] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gat_conv[1-5-g4-idtype0] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gat_conv[1-5-g4-idtype1] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gat_conv[1-5-g5-idtype0] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gat_conv[1-5-g5-idtype1] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gat_conv[1-5-g6-idtype0] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gat_conv[1-5-g6-idtype1] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gat_conv[4-1-g0-idtype0] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gat_conv[4-1-g0-idtype1] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gat_conv[4-1-g1-idtype0] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gat_conv[4-1-g1-idtype1] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gat_conv[4-1-g2-idtype0] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gat_conv[4-1-g2-idtype1] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gat_conv[4-1-g3-idtype0] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gat_conv[4-1-g3-idtype1] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gat_conv[4-1-g4-idtype0] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gat_conv[4-1-g4-idtype1] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gat_conv[4-1-g5-idtype0] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gat_conv[4-1-g5-idtype1] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gat_conv[4-1-g6-idtype0] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gat_conv[4-1-g6-idtype1] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gat_conv[4-5-g0-idtype0] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gat_conv[4-5-g0-idtype1] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gat_conv[4-5-g1-idtype0] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gat_conv[4-5-g1-idtype1] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gat_conv[4-5-g2-idtype0] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gat_conv[4-5-g2-idtype1] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gat_conv[4-5-g3-idtype0] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gat_conv[4-5-g3-idtype1] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gat_conv[4-5-g4-idtype0] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gat_conv[4-5-g4-idtype1] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gat_conv[4-5-g5-idtype0] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gat_conv[4-5-g5-idtype1] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gat_conv[4-5-g6-idtype0] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gat_conv[4-5-g6-idtype1] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gat_conv_bi[1-1-g0-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_gat_conv_bi[1-1-g0-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_gat_conv_bi[1-1-g1-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_gat_conv_bi[1-1-g1-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_gat_conv_bi[1-2-g0-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_gat_conv_bi[1-2-g0-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_gat_conv_bi[1-2-g1-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_gat_conv_bi[1-2-g1-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_gat_conv_bi[4-1-g0-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_gat_conv_bi[4-1-g0-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_gat_conv_bi[4-1-g1-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_gat_conv_bi[4-1-g1-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_gat_conv_bi[4-2-g0-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_gat_conv_bi[4-2-g0-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_gat_conv_bi[4-2-g1-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_gat_conv_bi[4-2-g1-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_gatv2_conv[1-1-g0-idtype0] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gatv2_conv[1-1-g0-idtype1] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gatv2_conv[1-1-g1-idtype0] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gatv2_conv[1-1-g1-idtype1] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gatv2_conv[1-1-g2-idtype0] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gatv2_conv[1-1-g2-idtype1] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gatv2_conv[1-1-g3-idtype0] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gatv2_conv[1-1-g3-idtype1] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gatv2_conv[1-1-g4-idtype0] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gatv2_conv[1-1-g4-idtype1] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gatv2_conv[1-1-g5-idtype0] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gatv2_conv[1-1-g5-idtype1] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gatv2_conv[1-1-g6-idtype0] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gatv2_conv[1-1-g6-idtype1] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gatv2_conv[1-5-g0-idtype0] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gatv2_conv[1-5-g0-idtype1] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gatv2_conv[1-5-g1-idtype0] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gatv2_conv[1-5-g1-idtype1] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gatv2_conv[1-5-g2-idtype0] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gatv2_conv[1-5-g2-idtype1] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gatv2_conv[1-5-g3-idtype0] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gatv2_conv[1-5-g3-idtype1] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gatv2_conv[1-5-g4-idtype0] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gatv2_conv[1-5-g4-idtype1] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gatv2_conv[1-5-g5-idtype0] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gatv2_conv[1-5-g5-idtype1] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gatv2_conv[1-5-g6-idtype0] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gatv2_conv[1-5-g6-idtype1] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gatv2_conv[4-1-g0-idtype0] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gatv2_conv[4-1-g0-idtype1] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gatv2_conv[4-1-g1-idtype0] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gatv2_conv[4-1-g1-idtype1] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gatv2_conv[4-1-g2-idtype0] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gatv2_conv[4-1-g2-idtype1] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gatv2_conv[4-1-g3-idtype0] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gatv2_conv[4-1-g3-idtype1] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gatv2_conv[4-1-g4-idtype0] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gatv2_conv[4-1-g4-idtype1] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gatv2_conv[4-1-g5-idtype0] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gatv2_conv[4-1-g5-idtype1] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gatv2_conv[4-1-g6-idtype0] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gatv2_conv[4-1-g6-idtype1] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gatv2_conv[4-5-g0-idtype0] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gatv2_conv[4-5-g0-idtype1] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gatv2_conv[4-5-g1-idtype0] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gatv2_conv[4-5-g1-idtype1] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gatv2_conv[4-5-g2-idtype0] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gatv2_conv[4-5-g2-idtype1] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gatv2_conv[4-5-g3-idtype0] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gatv2_conv[4-5-g3-idtype1] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gatv2_conv[4-5-g4-idtype0] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gatv2_conv[4-5-g4-idtype1] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gatv2_conv[4-5-g5-idtype0] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gatv2_conv[4-5-g5-idtype1] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gatv2_conv[4-5-g6-idtype0] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gatv2_conv[4-5-g6-idtype1] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gatv2_conv_bi[1-1-g0-idtype0] - TypeErr...
FAILED tests/pytorch/test_nn.py::test_gatv2_conv_bi[1-1-g0-idtype1] - TypeErr...
FAILED tests/pytorch/test_nn.py::test_gatv2_conv_bi[1-1-g1-idtype0] - TypeErr...
FAILED tests/pytorch/test_nn.py::test_gatv2_conv_bi[1-1-g1-idtype1] - TypeErr...
FAILED tests/pytorch/test_nn.py::test_gatv2_conv_bi[1-2-g0-idtype0] - TypeErr...
FAILED tests/pytorch/test_nn.py::test_gatv2_conv_bi[1-2-g0-idtype1] - TypeErr...
FAILED tests/pytorch/test_nn.py::test_gatv2_conv_bi[1-2-g1-idtype0] - TypeErr...
FAILED tests/pytorch/test_nn.py::test_gatv2_conv_bi[1-2-g1-idtype1] - TypeErr...
FAILED tests/pytorch/test_nn.py::test_gatv2_conv_bi[4-1-g0-idtype0] - TypeErr...
FAILED tests/pytorch/test_nn.py::test_gatv2_conv_bi[4-1-g0-idtype1] - TypeErr...
FAILED tests/pytorch/test_nn.py::test_gatv2_conv_bi[4-1-g1-idtype0] - TypeErr...
FAILED tests/pytorch/test_nn.py::test_gatv2_conv_bi[4-1-g1-idtype1] - TypeErr...
FAILED tests/pytorch/test_nn.py::test_gatv2_conv_bi[4-2-g0-idtype0] - TypeErr...
FAILED tests/pytorch/test_nn.py::test_gatv2_conv_bi[4-2-g0-idtype1] - TypeErr...
FAILED tests/pytorch/test_nn.py::test_gatv2_conv_bi[4-2-g1-idtype0] - TypeErr...
FAILED tests/pytorch/test_nn.py::test_gatv2_conv_bi[4-2-g1-idtype1] - TypeErr...
FAILED tests/pytorch/test_nn.py::test_egat_conv[1-1-1-g0-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[1-1-1-g0-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[1-1-1-g1-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[1-1-1-g1-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[1-1-1-g2-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[1-1-1-g2-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[1-1-1-g3-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[1-1-1-g3-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[1-1-1-g4-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[1-1-1-g4-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[1-1-1-g5-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[1-1-1-g5-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[1-1-5-g0-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[1-1-5-g0-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[1-1-5-g1-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[1-1-5-g1-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[1-1-5-g2-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[1-1-5-g2-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[1-1-5-g3-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[1-1-5-g3-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[1-1-5-g4-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[1-1-5-g4-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[1-1-5-g5-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[1-1-5-g5-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[1-5-1-g0-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[1-5-1-g0-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[1-5-1-g1-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[1-5-1-g1-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[1-5-1-g2-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[1-5-1-g2-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[1-5-1-g3-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[1-5-1-g3-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[1-5-1-g4-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[1-5-1-g4-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[1-5-1-g5-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[1-5-1-g5-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[1-5-5-g0-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[1-5-5-g0-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[1-5-5-g1-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[1-5-5-g1-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[1-5-5-g2-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[1-5-5-g2-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[1-5-5-g3-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[1-5-5-g3-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[1-5-5-g4-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[1-5-5-g4-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[1-5-5-g5-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[1-5-5-g5-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[4-1-1-g0-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[4-1-1-g0-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[4-1-1-g1-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[4-1-1-g1-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[4-1-1-g2-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[4-1-1-g2-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[4-1-1-g3-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[4-1-1-g3-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[4-1-1-g4-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[4-1-1-g4-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[4-1-1-g5-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[4-1-1-g5-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[4-1-5-g0-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[4-1-5-g0-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[4-1-5-g1-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[4-1-5-g1-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[4-1-5-g2-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[4-1-5-g2-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[4-1-5-g3-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[4-1-5-g3-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[4-1-5-g4-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[4-1-5-g4-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[4-1-5-g5-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[4-1-5-g5-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[4-5-1-g0-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[4-5-1-g0-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[4-5-1-g1-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[4-5-1-g1-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[4-5-1-g2-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[4-5-1-g2-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[4-5-1-g3-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[4-5-1-g3-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[4-5-1-g4-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[4-5-1-g4-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[4-5-1-g5-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[4-5-1-g5-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[4-5-5-g0-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[4-5-5-g0-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[4-5-5-g1-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[4-5-5-g1-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[4-5-5-g2-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[4-5-5-g2-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[4-5-5-g3-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[4-5-5-g3-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[4-5-5-g4-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[4-5-5-g4-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[4-5-5-g5-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv[4-5-5-g5-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_egat_conv_bi[1-1-1-g0-idtype0] - TypeEr...
FAILED tests/pytorch/test_nn.py::test_egat_conv_bi[1-1-1-g0-idtype1] - TypeEr...
FAILED tests/pytorch/test_nn.py::test_egat_conv_bi[1-1-1-g1-idtype0] - TypeEr...
FAILED tests/pytorch/test_nn.py::test_egat_conv_bi[1-1-1-g1-idtype1] - TypeEr...
FAILED tests/pytorch/test_nn.py::test_egat_conv_bi[1-1-5-g0-idtype0] - TypeEr...
FAILED tests/pytorch/test_nn.py::test_egat_conv_bi[1-1-5-g0-idtype1] - TypeEr...
FAILED tests/pytorch/test_nn.py::test_egat_conv_bi[1-1-5-g1-idtype0] - TypeEr...
FAILED tests/pytorch/test_nn.py::test_egat_conv_bi[1-1-5-g1-idtype1] - TypeEr...
FAILED tests/pytorch/test_nn.py::test_egat_conv_bi[1-5-1-g0-idtype0] - TypeEr...
FAILED tests/pytorch/test_nn.py::test_egat_conv_bi[1-5-1-g0-idtype1] - TypeEr...
FAILED tests/pytorch/test_nn.py::test_egat_conv_bi[1-5-1-g1-idtype0] - TypeEr...
FAILED tests/pytorch/test_nn.py::test_egat_conv_bi[1-5-1-g1-idtype1] - TypeEr...
FAILED tests/pytorch/test_nn.py::test_egat_conv_bi[1-5-5-g0-idtype0] - TypeEr...
FAILED tests/pytorch/test_nn.py::test_egat_conv_bi[1-5-5-g0-idtype1] - TypeEr...
FAILED tests/pytorch/test_nn.py::test_egat_conv_bi[1-5-5-g1-idtype0] - TypeEr...
FAILED tests/pytorch/test_nn.py::test_egat_conv_bi[1-5-5-g1-idtype1] - TypeEr...
FAILED tests/pytorch/test_nn.py::test_egat_conv_bi[4-1-1-g0-idtype0] - TypeEr...
FAILED tests/pytorch/test_nn.py::test_egat_conv_bi[4-1-1-g0-idtype1] - TypeEr...
FAILED tests/pytorch/test_nn.py::test_egat_conv_bi[4-1-1-g1-idtype0] - TypeEr...
FAILED tests/pytorch/test_nn.py::test_egat_conv_bi[4-1-1-g1-idtype1] - TypeEr...
FAILED tests/pytorch/test_nn.py::test_egat_conv_bi[4-1-5-g0-idtype0] - TypeEr...
FAILED tests/pytorch/test_nn.py::test_egat_conv_bi[4-1-5-g0-idtype1] - TypeEr...
FAILED tests/pytorch/test_nn.py::test_egat_conv_bi[4-1-5-g1-idtype0] - TypeEr...
FAILED tests/pytorch/test_nn.py::test_egat_conv_bi[4-1-5-g1-idtype1] - TypeEr...
FAILED tests/pytorch/test_nn.py::test_egat_conv_bi[4-5-1-g0-idtype0] - TypeEr...
FAILED tests/pytorch/test_nn.py::test_egat_conv_bi[4-5-1-g0-idtype1] - TypeEr...
FAILED tests/pytorch/test_nn.py::test_egat_conv_bi[4-5-1-g1-idtype0] - TypeEr...
FAILED tests/pytorch/test_nn.py::test_egat_conv_bi[4-5-1-g1-idtype1] - TypeEr...
FAILED tests/pytorch/test_nn.py::test_egat_conv_bi[4-5-5-g0-idtype0] - TypeEr...
FAILED tests/pytorch/test_nn.py::test_egat_conv_bi[4-5-5-g0-idtype1] - TypeEr...
FAILED tests/pytorch/test_nn.py::test_egat_conv_bi[4-5-5-g1-idtype0] - TypeEr...
FAILED tests/pytorch/test_nn.py::test_egat_conv_bi[4-5-5-g1-idtype1] - TypeEr...
FAILED tests/pytorch/test_nn.py::test_sage_conv[mean-g0-idtype0] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_sage_conv[mean-g0-idtype1] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_sage_conv[mean-g1-idtype0] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_sage_conv[mean-g1-idtype1] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_sage_conv[mean-g2-idtype0] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_sage_conv[mean-g2-idtype1] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_sage_conv[mean-g3-idtype0] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_sage_conv[mean-g3-idtype1] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_sage_conv[mean-g4-idtype0] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_sage_conv[mean-g4-idtype1] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_sage_conv[mean-g5-idtype0] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_sage_conv[mean-g5-idtype1] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_sage_conv[mean-g6-idtype0] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_sage_conv[mean-g6-idtype1] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_sage_conv[mean-g7-idtype0] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_sage_conv[mean-g7-idtype1] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_sage_conv[pool-g0-idtype0] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_sage_conv[pool-g0-idtype1] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_sage_conv[pool-g1-idtype0] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_sage_conv[pool-g1-idtype1] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_sage_conv[pool-g2-idtype0] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_sage_conv[pool-g2-idtype1] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_sage_conv[pool-g3-idtype0] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_sage_conv[pool-g3-idtype1] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_sage_conv[pool-g4-idtype0] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_sage_conv[pool-g4-idtype1] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_sage_conv[pool-g5-idtype0] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_sage_conv[pool-g5-idtype1] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_sage_conv[pool-g6-idtype0] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_sage_conv[pool-g6-idtype1] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_sage_conv[pool-g7-idtype0] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_sage_conv[pool-g7-idtype1] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_sage_conv[gcn-g0-idtype0] - TypeError: ...
FAILED tests/pytorch/test_nn.py::test_sage_conv[gcn-g0-idtype1] - TypeError: ...
FAILED tests/pytorch/test_nn.py::test_sage_conv[gcn-g1-idtype0] - TypeError: ...
FAILED tests/pytorch/test_nn.py::test_sage_conv[gcn-g1-idtype1] - TypeError: ...
FAILED tests/pytorch/test_nn.py::test_sage_conv[gcn-g2-idtype0] - TypeError: ...
FAILED tests/pytorch/test_nn.py::test_sage_conv[gcn-g2-idtype1] - TypeError: ...
FAILED tests/pytorch/test_nn.py::test_sage_conv[gcn-g3-idtype0] - TypeError: ...
FAILED tests/pytorch/test_nn.py::test_sage_conv[gcn-g3-idtype1] - TypeError: ...
FAILED tests/pytorch/test_nn.py::test_sage_conv[gcn-g4-idtype0] - TypeError: ...
FAILED tests/pytorch/test_nn.py::test_sage_conv[gcn-g4-idtype1] - TypeError: ...
FAILED tests/pytorch/test_nn.py::test_sage_conv[gcn-g5-idtype0] - TypeError: ...
FAILED tests/pytorch/test_nn.py::test_sage_conv[gcn-g5-idtype1] - TypeError: ...
FAILED tests/pytorch/test_nn.py::test_sage_conv[gcn-g6-idtype0] - TypeError: ...
FAILED tests/pytorch/test_nn.py::test_sage_conv[gcn-g6-idtype1] - TypeError: ...
FAILED tests/pytorch/test_nn.py::test_sage_conv[gcn-g7-idtype0] - TypeError: ...
FAILED tests/pytorch/test_nn.py::test_sage_conv[gcn-g7-idtype1] - TypeError: ...
FAILED tests/pytorch/test_nn.py::test_sage_conv[lstm-g0-idtype0] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_sage_conv[lstm-g0-idtype1] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_sage_conv[lstm-g1-idtype0] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_sage_conv[lstm-g1-idtype1] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_sage_conv[lstm-g2-idtype0] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_sage_conv[lstm-g2-idtype1] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_sage_conv[lstm-g3-idtype0] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_sage_conv[lstm-g3-idtype1] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_sage_conv[lstm-g4-idtype0] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_sage_conv[lstm-g4-idtype1] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_sage_conv[lstm-g5-idtype0] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_sage_conv[lstm-g5-idtype1] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_sage_conv[lstm-g6-idtype0] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_sage_conv[lstm-g6-idtype1] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_sage_conv[lstm-g7-idtype0] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_sage_conv[lstm-g7-idtype1] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_sage_conv_bi[1-mean-g0-idtype0] - TypeE...
FAILED tests/pytorch/test_nn.py::test_sage_conv_bi[1-mean-g0-idtype1] - TypeE...
FAILED tests/pytorch/test_nn.py::test_sage_conv_bi[1-mean-g1-idtype0] - TypeE...
FAILED tests/pytorch/test_nn.py::test_sage_conv_bi[1-mean-g1-idtype1] - TypeE...
FAILED tests/pytorch/test_nn.py::test_sage_conv_bi[1-mean-g2-idtype0] - TypeE...
FAILED tests/pytorch/test_nn.py::test_sage_conv_bi[1-mean-g2-idtype1] - TypeE...
FAILED tests/pytorch/test_nn.py::test_sage_conv_bi[1-pool-g0-idtype0] - TypeE...
FAILED tests/pytorch/test_nn.py::test_sage_conv_bi[1-pool-g0-idtype1] - TypeE...
FAILED tests/pytorch/test_nn.py::test_sage_conv_bi[1-pool-g1-idtype0] - TypeE...
FAILED tests/pytorch/test_nn.py::test_sage_conv_bi[1-pool-g1-idtype1] - TypeE...
FAILED tests/pytorch/test_nn.py::test_sage_conv_bi[1-pool-g2-idtype0] - TypeE...
FAILED tests/pytorch/test_nn.py::test_sage_conv_bi[1-pool-g2-idtype1] - TypeE...
FAILED tests/pytorch/test_nn.py::test_sage_conv_bi[1-gcn-g0-idtype0] - TypeEr...
FAILED tests/pytorch/test_nn.py::test_sage_conv_bi[1-gcn-g0-idtype1] - TypeEr...
FAILED tests/pytorch/test_nn.py::test_sage_conv_bi[1-gcn-g1-idtype0] - TypeEr...
FAILED tests/pytorch/test_nn.py::test_sage_conv_bi[1-gcn-g1-idtype1] - TypeEr...
FAILED tests/pytorch/test_nn.py::test_sage_conv_bi[1-gcn-g2-idtype0] - TypeEr...
FAILED tests/pytorch/test_nn.py::test_sage_conv_bi[1-gcn-g2-idtype1] - TypeEr...
FAILED tests/pytorch/test_nn.py::test_sage_conv_bi[1-lstm-g0-idtype0] - TypeE...
FAILED tests/pytorch/test_nn.py::test_sage_conv_bi[1-lstm-g0-idtype1] - TypeE...
FAILED tests/pytorch/test_nn.py::test_sage_conv_bi[1-lstm-g1-idtype0] - TypeE...
FAILED tests/pytorch/test_nn.py::test_sage_conv_bi[1-lstm-g1-idtype1] - TypeE...
FAILED tests/pytorch/test_nn.py::test_sage_conv_bi[1-lstm-g2-idtype0] - TypeE...
FAILED tests/pytorch/test_nn.py::test_sage_conv_bi[1-lstm-g2-idtype1] - TypeE...
FAILED tests/pytorch/test_nn.py::test_sage_conv_bi[2-mean-g0-idtype0] - TypeE...
FAILED tests/pytorch/test_nn.py::test_sage_conv_bi[2-mean-g0-idtype1] - TypeE...
FAILED tests/pytorch/test_nn.py::test_sage_conv_bi[2-mean-g1-idtype0] - TypeE...
FAILED tests/pytorch/test_nn.py::test_sage_conv_bi[2-mean-g1-idtype1] - TypeE...
FAILED tests/pytorch/test_nn.py::test_sage_conv_bi[2-mean-g2-idtype0] - TypeE...
FAILED tests/pytorch/test_nn.py::test_sage_conv_bi[2-mean-g2-idtype1] - TypeE...
FAILED tests/pytorch/test_nn.py::test_sage_conv_bi[2-pool-g0-idtype0] - TypeE...
FAILED tests/pytorch/test_nn.py::test_sage_conv_bi[2-pool-g0-idtype1] - TypeE...
FAILED tests/pytorch/test_nn.py::test_sage_conv_bi[2-pool-g1-idtype0] - TypeE...
FAILED tests/pytorch/test_nn.py::test_sage_conv_bi[2-pool-g1-idtype1] - TypeE...
FAILED tests/pytorch/test_nn.py::test_sage_conv_bi[2-pool-g2-idtype0] - TypeE...
FAILED tests/pytorch/test_nn.py::test_sage_conv_bi[2-pool-g2-idtype1] - TypeE...
FAILED tests/pytorch/test_nn.py::test_sage_conv_bi[2-gcn-g0-idtype0] - TypeEr...
FAILED tests/pytorch/test_nn.py::test_sage_conv_bi[2-gcn-g0-idtype1] - TypeEr...
FAILED tests/pytorch/test_nn.py::test_sage_conv_bi[2-gcn-g1-idtype0] - TypeEr...
FAILED tests/pytorch/test_nn.py::test_sage_conv_bi[2-gcn-g1-idtype1] - TypeEr...
FAILED tests/pytorch/test_nn.py::test_sage_conv_bi[2-gcn-g2-idtype0] - TypeEr...
FAILED tests/pytorch/test_nn.py::test_sage_conv_bi[2-gcn-g2-idtype1] - TypeEr...
FAILED tests/pytorch/test_nn.py::test_sage_conv_bi[2-lstm-g0-idtype0] - TypeE...
FAILED tests/pytorch/test_nn.py::test_sage_conv_bi[2-lstm-g0-idtype1] - TypeE...
FAILED tests/pytorch/test_nn.py::test_sage_conv_bi[2-lstm-g1-idtype0] - TypeE...
FAILED tests/pytorch/test_nn.py::test_sage_conv_bi[2-lstm-g1-idtype1] - TypeE...
FAILED tests/pytorch/test_nn.py::test_sage_conv_bi[2-lstm-g2-idtype0] - TypeE...
FAILED tests/pytorch/test_nn.py::test_sage_conv_bi[2-lstm-g2-idtype1] - TypeE...
FAILED tests/pytorch/test_nn.py::test_sage_conv2[1-idtype0] - TypeError: empt...
FAILED tests/pytorch/test_nn.py::test_sage_conv2[1-idtype1] - TypeError: empt...
FAILED tests/pytorch/test_nn.py::test_sage_conv2[2-idtype0] - TypeError: empt...
FAILED tests/pytorch/test_nn.py::test_sage_conv2[2-idtype1] - TypeError: empt...
FAILED tests/pytorch/test_nn.py::test_sgc_conv[1-g0-idtype0] - TypeError: emp...
FAILED tests/pytorch/test_nn.py::test_sgc_conv[1-g0-idtype1] - TypeError: emp...
FAILED tests/pytorch/test_nn.py::test_sgc_conv[1-g1-idtype0] - TypeError: emp...
FAILED tests/pytorch/test_nn.py::test_sgc_conv[1-g1-idtype1] - TypeError: emp...
FAILED tests/pytorch/test_nn.py::test_sgc_conv[1-g2-idtype0] - TypeError: emp...
FAILED tests/pytorch/test_nn.py::test_sgc_conv[1-g2-idtype1] - TypeError: emp...
FAILED tests/pytorch/test_nn.py::test_sgc_conv[1-g3-idtype0] - TypeError: emp...
FAILED tests/pytorch/test_nn.py::test_sgc_conv[1-g3-idtype1] - TypeError: emp...
FAILED tests/pytorch/test_nn.py::test_sgc_conv[1-g4-idtype0] - TypeError: emp...
FAILED tests/pytorch/test_nn.py::test_sgc_conv[1-g4-idtype1] - TypeError: emp...
FAILED tests/pytorch/test_nn.py::test_sgc_conv[1-g5-idtype0] - TypeError: emp...
FAILED tests/pytorch/test_nn.py::test_sgc_conv[1-g5-idtype1] - TypeError: emp...
FAILED tests/pytorch/test_nn.py::test_sgc_conv[2-g0-idtype0] - TypeError: emp...
FAILED tests/pytorch/test_nn.py::test_sgc_conv[2-g0-idtype1] - TypeError: emp...
FAILED tests/pytorch/test_nn.py::test_sgc_conv[2-g1-idtype0] - TypeError: emp...
FAILED tests/pytorch/test_nn.py::test_sgc_conv[2-g1-idtype1] - TypeError: emp...
FAILED tests/pytorch/test_nn.py::test_sgc_conv[2-g2-idtype0] - TypeError: emp...
FAILED tests/pytorch/test_nn.py::test_sgc_conv[2-g2-idtype1] - TypeError: emp...
FAILED tests/pytorch/test_nn.py::test_sgc_conv[2-g3-idtype0] - TypeError: emp...
FAILED tests/pytorch/test_nn.py::test_sgc_conv[2-g3-idtype1] - TypeError: emp...
FAILED tests/pytorch/test_nn.py::test_sgc_conv[2-g4-idtype0] - TypeError: emp...
FAILED tests/pytorch/test_nn.py::test_sgc_conv[2-g4-idtype1] - TypeError: emp...
FAILED tests/pytorch/test_nn.py::test_sgc_conv[2-g5-idtype0] - TypeError: emp...
FAILED tests/pytorch/test_nn.py::test_sgc_conv[2-g5-idtype1] - TypeError: emp...
FAILED tests/pytorch/test_nn.py::test_appnp_conv[g0-idtype0] - TypeError: emp...
FAILED tests/pytorch/test_nn.py::test_appnp_conv[g0-idtype1] - TypeError: emp...
FAILED tests/pytorch/test_nn.py::test_appnp_conv[g1-idtype0] - TypeError: emp...
FAILED tests/pytorch/test_nn.py::test_appnp_conv[g1-idtype1] - TypeError: emp...
FAILED tests/pytorch/test_nn.py::test_appnp_conv[g2-idtype0] - TypeError: emp...
FAILED tests/pytorch/test_nn.py::test_appnp_conv[g2-idtype1] - TypeError: emp...
FAILED tests/pytorch/test_nn.py::test_appnp_conv[g3-idtype0] - TypeError: emp...
FAILED tests/pytorch/test_nn.py::test_appnp_conv[g3-idtype1] - TypeError: emp...
FAILED tests/pytorch/test_nn.py::test_appnp_conv[g4-idtype0] - TypeError: emp...
FAILED tests/pytorch/test_nn.py::test_appnp_conv[g4-idtype1] - TypeError: emp...
FAILED tests/pytorch/test_nn.py::test_appnp_conv[g5-idtype0] - TypeError: emp...
FAILED tests/pytorch/test_nn.py::test_appnp_conv[g5-idtype1] - TypeError: emp...
FAILED tests/pytorch/test_nn.py::test_appnp_conv_e_weight[g0-idtype0] - TypeE...
FAILED tests/pytorch/test_nn.py::test_appnp_conv_e_weight[g0-idtype1] - TypeE...
FAILED tests/pytorch/test_nn.py::test_appnp_conv_e_weight[g1-idtype0] - TypeE...
FAILED tests/pytorch/test_nn.py::test_appnp_conv_e_weight[g1-idtype1] - TypeE...
FAILED tests/pytorch/test_nn.py::test_appnp_conv_e_weight[g2-idtype0] - TypeE...
FAILED tests/pytorch/test_nn.py::test_appnp_conv_e_weight[g2-idtype1] - TypeE...
FAILED tests/pytorch/test_nn.py::test_appnp_conv_e_weight[g3-idtype0] - TypeE...
FAILED tests/pytorch/test_nn.py::test_appnp_conv_e_weight[g3-idtype1] - TypeE...
FAILED tests/pytorch/test_nn.py::test_appnp_conv_e_weight[g4-idtype0] - TypeE...
FAILED tests/pytorch/test_nn.py::test_appnp_conv_e_weight[g4-idtype1] - TypeE...
FAILED tests/pytorch/test_nn.py::test_appnp_conv_e_weight[g5-idtype0] - TypeE...
FAILED tests/pytorch/test_nn.py::test_appnp_conv_e_weight[g5-idtype1] - TypeE...
FAILED tests/pytorch/test_nn.py::test_gcn2conv_e_weight[True-g0-idtype0] - Ty...
FAILED tests/pytorch/test_nn.py::test_gcn2conv_e_weight[True-g0-idtype1] - Ty...
FAILED tests/pytorch/test_nn.py::test_gcn2conv_e_weight[True-g1-idtype0] - Ty...
FAILED tests/pytorch/test_nn.py::test_gcn2conv_e_weight[True-g1-idtype1] - Ty...
FAILED tests/pytorch/test_nn.py::test_gcn2conv_e_weight[True-g2-idtype0] - Ty...
FAILED tests/pytorch/test_nn.py::test_gcn2conv_e_weight[True-g2-idtype1] - Ty...
FAILED tests/pytorch/test_nn.py::test_gcn2conv_e_weight[True-g3-idtype0] - Ty...
FAILED tests/pytorch/test_nn.py::test_gcn2conv_e_weight[True-g3-idtype1] - Ty...
FAILED tests/pytorch/test_nn.py::test_gcn2conv_e_weight[True-g4-idtype0] - Ty...
FAILED tests/pytorch/test_nn.py::test_gcn2conv_e_weight[True-g4-idtype1] - Ty...
FAILED tests/pytorch/test_nn.py::test_gcn2conv_e_weight[True-g5-idtype0] - Ty...
FAILED tests/pytorch/test_nn.py::test_gcn2conv_e_weight[True-g5-idtype1] - Ty...
FAILED tests/pytorch/test_nn.py::test_gcn2conv_e_weight[False-g0-idtype0] - T...
FAILED tests/pytorch/test_nn.py::test_gcn2conv_e_weight[False-g0-idtype1] - T...
FAILED tests/pytorch/test_nn.py::test_gcn2conv_e_weight[False-g1-idtype0] - T...
FAILED tests/pytorch/test_nn.py::test_gcn2conv_e_weight[False-g1-idtype1] - T...
FAILED tests/pytorch/test_nn.py::test_gcn2conv_e_weight[False-g2-idtype0] - T...
FAILED tests/pytorch/test_nn.py::test_gcn2conv_e_weight[False-g2-idtype1] - T...
FAILED tests/pytorch/test_nn.py::test_gcn2conv_e_weight[False-g3-idtype0] - T...
FAILED tests/pytorch/test_nn.py::test_gcn2conv_e_weight[False-g3-idtype1] - T...
FAILED tests/pytorch/test_nn.py::test_gcn2conv_e_weight[False-g4-idtype0] - T...
FAILED tests/pytorch/test_nn.py::test_gcn2conv_e_weight[False-g4-idtype1] - T...
FAILED tests/pytorch/test_nn.py::test_gcn2conv_e_weight[False-g5-idtype0] - T...
FAILED tests/pytorch/test_nn.py::test_gcn2conv_e_weight[False-g5-idtype1] - T...
FAILED tests/pytorch/test_nn.py::test_sgconv_e_weight[g0-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_sgconv_e_weight[g0-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_sgconv_e_weight[g1-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_sgconv_e_weight[g1-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_sgconv_e_weight[g2-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_sgconv_e_weight[g2-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_sgconv_e_weight[g3-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_sgconv_e_weight[g3-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_sgconv_e_weight[g4-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_sgconv_e_weight[g4-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_sgconv_e_weight[g5-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_sgconv_e_weight[g5-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_tagconv_e_weight[g0-idtype0] - TypeErro...
FAILED tests/pytorch/test_nn.py::test_tagconv_e_weight[g0-idtype1] - TypeErro...
FAILED tests/pytorch/test_nn.py::test_tagconv_e_weight[g1-idtype0] - TypeErro...
FAILED tests/pytorch/test_nn.py::test_tagconv_e_weight[g1-idtype1] - TypeErro...
FAILED tests/pytorch/test_nn.py::test_tagconv_e_weight[g2-idtype0] - TypeErro...
FAILED tests/pytorch/test_nn.py::test_tagconv_e_weight[g2-idtype1] - TypeErro...
FAILED tests/pytorch/test_nn.py::test_tagconv_e_weight[g3-idtype0] - TypeErro...
FAILED tests/pytorch/test_nn.py::test_tagconv_e_weight[g3-idtype1] - TypeErro...
FAILED tests/pytorch/test_nn.py::test_tagconv_e_weight[g4-idtype0] - TypeErro...
FAILED tests/pytorch/test_nn.py::test_tagconv_e_weight[g4-idtype1] - TypeErro...
FAILED tests/pytorch/test_nn.py::test_tagconv_e_weight[g5-idtype0] - TypeErro...
FAILED tests/pytorch/test_nn.py::test_tagconv_e_weight[g5-idtype1] - TypeErro...
FAILED tests/pytorch/test_nn.py::test_gin_conv[mean-g0-idtype0] - TypeError: ...
FAILED tests/pytorch/test_nn.py::test_gin_conv[mean-g0-idtype1] - TypeError: ...
FAILED tests/pytorch/test_nn.py::test_gin_conv[mean-g1-idtype0] - TypeError: ...
FAILED tests/pytorch/test_nn.py::test_gin_conv[mean-g1-idtype1] - TypeError: ...
FAILED tests/pytorch/test_nn.py::test_gin_conv[mean-g2-idtype0] - TypeError: ...
FAILED tests/pytorch/test_nn.py::test_gin_conv[mean-g2-idtype1] - TypeError: ...
FAILED tests/pytorch/test_nn.py::test_gin_conv[mean-g3-idtype0] - TypeError: ...
FAILED tests/pytorch/test_nn.py::test_gin_conv[mean-g3-idtype1] - TypeError: ...
FAILED tests/pytorch/test_nn.py::test_gin_conv[mean-g4-idtype0] - TypeError: ...
FAILED tests/pytorch/test_nn.py::test_gin_conv[mean-g4-idtype1] - TypeError: ...
FAILED tests/pytorch/test_nn.py::test_gin_conv[mean-g5-idtype0] - TypeError: ...
FAILED tests/pytorch/test_nn.py::test_gin_conv[mean-g5-idtype1] - TypeError: ...
FAILED tests/pytorch/test_nn.py::test_gin_conv[mean-g6-idtype0] - TypeError: ...
FAILED tests/pytorch/test_nn.py::test_gin_conv[mean-g6-idtype1] - TypeError: ...
FAILED tests/pytorch/test_nn.py::test_gin_conv[max-g0-idtype0] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gin_conv[max-g0-idtype1] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gin_conv[max-g1-idtype0] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gin_conv[max-g1-idtype1] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gin_conv[max-g2-idtype0] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gin_conv[max-g2-idtype1] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gin_conv[max-g3-idtype0] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gin_conv[max-g3-idtype1] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gin_conv[max-g4-idtype0] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gin_conv[max-g4-idtype1] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gin_conv[max-g5-idtype0] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gin_conv[max-g5-idtype1] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gin_conv[max-g6-idtype0] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gin_conv[max-g6-idtype1] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gin_conv[sum-g0-idtype0] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gin_conv[sum-g0-idtype1] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gin_conv[sum-g1-idtype0] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gin_conv[sum-g1-idtype1] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gin_conv[sum-g2-idtype0] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gin_conv[sum-g2-idtype1] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gin_conv[sum-g3-idtype0] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gin_conv[sum-g3-idtype1] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gin_conv[sum-g4-idtype0] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gin_conv[sum-g4-idtype1] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gin_conv[sum-g5-idtype0] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gin_conv[sum-g5-idtype1] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gin_conv[sum-g6-idtype0] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gin_conv[sum-g6-idtype1] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gine_conv[g0-idtype0] - TypeError: empt...
FAILED tests/pytorch/test_nn.py::test_gine_conv[g0-idtype1] - TypeError: empt...
FAILED tests/pytorch/test_nn.py::test_gine_conv[g1-idtype0] - TypeError: empt...
FAILED tests/pytorch/test_nn.py::test_gine_conv[g1-idtype1] - TypeError: empt...
FAILED tests/pytorch/test_nn.py::test_gine_conv[g2-idtype0] - TypeError: empt...
FAILED tests/pytorch/test_nn.py::test_gine_conv[g2-idtype1] - TypeError: empt...
FAILED tests/pytorch/test_nn.py::test_gine_conv[g3-idtype0] - TypeError: empt...
FAILED tests/pytorch/test_nn.py::test_gine_conv[g3-idtype1] - TypeError: empt...
FAILED tests/pytorch/test_nn.py::test_gine_conv[g4-idtype0] - TypeError: empt...
FAILED tests/pytorch/test_nn.py::test_gine_conv[g4-idtype1] - TypeError: empt...
FAILED tests/pytorch/test_nn.py::test_gine_conv[g5-idtype0] - TypeError: empt...
FAILED tests/pytorch/test_nn.py::test_gine_conv[g5-idtype1] - TypeError: empt...
FAILED tests/pytorch/test_nn.py::test_gine_conv[g6-idtype0] - TypeError: empt...
FAILED tests/pytorch/test_nn.py::test_gine_conv[g6-idtype1] - TypeError: empt...
FAILED tests/pytorch/test_nn.py::test_gine_conv[g7-idtype0] - TypeError: empt...
FAILED tests/pytorch/test_nn.py::test_gine_conv[g7-idtype1] - TypeError: empt...
FAILED tests/pytorch/test_nn.py::test_gin_conv_bi[mean-g0-idtype0] - TypeErro...
FAILED tests/pytorch/test_nn.py::test_gin_conv_bi[mean-g0-idtype1] - TypeErro...
FAILED tests/pytorch/test_nn.py::test_gin_conv_bi[mean-g1-idtype0] - TypeErro...
FAILED tests/pytorch/test_nn.py::test_gin_conv_bi[mean-g1-idtype1] - TypeErro...
FAILED tests/pytorch/test_nn.py::test_gin_conv_bi[max-g0-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_gin_conv_bi[max-g0-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_gin_conv_bi[max-g1-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_gin_conv_bi[max-g1-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_gin_conv_bi[sum-g0-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_gin_conv_bi[sum-g0-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_gin_conv_bi[sum-g1-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_gin_conv_bi[sum-g1-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_agnn_conv[g0-idtype0] - TypeError: empt...
FAILED tests/pytorch/test_nn.py::test_agnn_conv[g0-idtype1] - TypeError: empt...
FAILED tests/pytorch/test_nn.py::test_agnn_conv[g1-idtype0] - TypeError: empt...
FAILED tests/pytorch/test_nn.py::test_agnn_conv[g1-idtype1] - TypeError: empt...
FAILED tests/pytorch/test_nn.py::test_agnn_conv[g2-idtype0] - TypeError: empt...
FAILED tests/pytorch/test_nn.py::test_agnn_conv[g2-idtype1] - TypeError: empt...
FAILED tests/pytorch/test_nn.py::test_agnn_conv[g3-idtype0] - TypeError: empt...
FAILED tests/pytorch/test_nn.py::test_agnn_conv[g3-idtype1] - TypeError: empt...

FAILED tests/pytorch/test_nn.py::test_agnn_conv[g4-idtype1] - TypeError: empt...
FAILED tests/pytorch/test_nn.py::test_agnn_conv[g5-idtype0] - TypeError: empt...
FAILED tests/pytorch/test_nn.py::test_agnn_conv[g5-idtype1] - TypeError: empt...
FAILED tests/pytorch/test_nn.py::test_agnn_conv[g6-idtype0] - TypeError: empt...
FAILED tests/pytorch/test_nn.py::test_agnn_conv[g6-idtype1] - TypeError: empt...
FAILED tests/pytorch/test_nn.py::test_agnn_conv_bi[g0-idtype0] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_agnn_conv_bi[g0-idtype1] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_agnn_conv_bi[g1-idtype0] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_agnn_conv_bi[g1-idtype1] - TypeError: e...
FAILED tests/pytorch/test_nn.py::test_gated_graph_conv[g0-idtype0] - TypeErro...
FAILED tests/pytorch/test_nn.py::test_gated_graph_conv[g0-idtype1] - TypeErro...
FAILED tests/pytorch/test_nn.py::test_gated_graph_conv[g1-idtype0] - TypeErro...
FAILED tests/pytorch/test_nn.py::test_gated_graph_conv[g1-idtype1] - TypeErro...
FAILED tests/pytorch/test_nn.py::test_gated_graph_conv[g2-idtype0] - TypeErro...
FAILED tests/pytorch/test_nn.py::test_gated_graph_conv[g2-idtype1] - TypeErro...
FAILED tests/pytorch/test_nn.py::test_gated_graph_conv[g3-idtype0] - TypeErro...
FAILED tests/pytorch/test_nn.py::test_gated_graph_conv[g3-idtype1] - TypeErro...
FAILED tests/pytorch/test_nn.py::test_gated_graph_conv[g4-idtype0] - TypeErro...
FAILED tests/pytorch/test_nn.py::test_gated_graph_conv[g4-idtype1] - TypeErro...
FAILED tests/pytorch/test_nn.py::test_gated_graph_conv[g5-idtype0] - TypeErro...
FAILED tests/pytorch/test_nn.py::test_gated_graph_conv[g5-idtype1] - TypeErro...
FAILED tests/pytorch/test_nn.py::test_gated_graph_conv_one_etype[g0-idtype0]
FAILED tests/pytorch/test_nn.py::test_gated_graph_conv_one_etype[g0-idtype1]
FAILED tests/pytorch/test_nn.py::test_gated_graph_conv_one_etype[g1-idtype0]
FAILED tests/pytorch/test_nn.py::test_gated_graph_conv_one_etype[g1-idtype1]
FAILED tests/pytorch/test_nn.py::test_gated_graph_conv_one_etype[g2-idtype0]
FAILED tests/pytorch/test_nn.py::test_gated_graph_conv_one_etype[g2-idtype1]
FAILED tests/pytorch/test_nn.py::test_gated_graph_conv_one_etype[g3-idtype0]
FAILED tests/pytorch/test_nn.py::test_gated_graph_conv_one_etype[g3-idtype1]
FAILED tests/pytorch/test_nn.py::test_gated_graph_conv_one_etype[g4-idtype0]
FAILED tests/pytorch/test_nn.py::test_gated_graph_conv_one_etype[g4-idtype1]
FAILED tests/pytorch/test_nn.py::test_gated_graph_conv_one_etype[g5-idtype0]
FAILED tests/pytorch/test_nn.py::test_gated_graph_conv_one_etype[g5-idtype1]
FAILED tests/pytorch/test_nn.py::test_nn_conv[g0-idtype0] - TypeError: empty_...
FAILED tests/pytorch/test_nn.py::test_nn_conv[g0-idtype1] - TypeError: empty_...
FAILED tests/pytorch/test_nn.py::test_nn_conv[g1-idtype0] - TypeError: empty_...
FAILED tests/pytorch/test_nn.py::test_nn_conv[g1-idtype1] - TypeError: empty_...
FAILED tests/pytorch/test_nn.py::test_nn_conv[g2-idtype0] - TypeError: empty_...
FAILED tests/pytorch/test_nn.py::test_nn_conv[g2-idtype1] - TypeError: empty_...
FAILED tests/pytorch/test_nn.py::test_nn_conv[g3-idtype0] - TypeError: empty_...
FAILED tests/pytorch/test_nn.py::test_nn_conv[g3-idtype1] - TypeError: empty_...
FAILED tests/pytorch/test_nn.py::test_nn_conv[g4-idtype0] - TypeError: empty_...
FAILED tests/pytorch/test_nn.py::test_nn_conv[g4-idtype1] - TypeError: empty_...
FAILED tests/pytorch/test_nn.py::test_nn_conv[g5-idtype0] - TypeError: empty_...
FAILED tests/pytorch/test_nn.py::test_nn_conv[g5-idtype1] - TypeError: empty_...
FAILED tests/pytorch/test_nn.py::test_nn_conv[g6-idtype0] - TypeError: empty_...
FAILED tests/pytorch/test_nn.py::test_nn_conv[g6-idtype1] - TypeError: empty_...
FAILED tests/pytorch/test_nn.py::test_nn_conv_bi[g0-idtype0] - TypeError: emp...
FAILED tests/pytorch/test_nn.py::test_nn_conv_bi[g0-idtype1] - TypeError: emp...
FAILED tests/pytorch/test_nn.py::test_nn_conv_bi[g1-idtype0] - TypeError: emp...
FAILED tests/pytorch/test_nn.py::test_nn_conv_bi[g1-idtype1] - TypeError: emp...
FAILED tests/pytorch/test_nn.py::test_gmm_conv[g0-idtype0] - TypeError: empty...
FAILED tests/pytorch/test_nn.py::test_gmm_conv[g0-idtype1] - TypeError: empty...
FAILED tests/pytorch/test_nn.py::test_gmm_conv[g1-idtype0] - TypeError: empty...
FAILED tests/pytorch/test_nn.py::test_gmm_conv[g1-idtype1] - TypeError: empty...
FAILED tests/pytorch/test_nn.py::test_gmm_conv[g2-idtype0] - TypeError: empty...
FAILED tests/pytorch/test_nn.py::test_gmm_conv[g2-idtype1] - TypeError: empty...
FAILED tests/pytorch/test_nn.py::test_gmm_conv[g3-idtype0] - TypeError: empty...
FAILED tests/pytorch/test_nn.py::test_gmm_conv[g3-idtype1] - TypeError: empty...
FAILED tests/pytorch/test_nn.py::test_gmm_conv[g4-idtype0] - TypeError: empty...
FAILED tests/pytorch/test_nn.py::test_gmm_conv[g4-idtype1] - TypeError: empty...
FAILED tests/pytorch/test_nn.py::test_gmm_conv[g5-idtype0] - TypeError: empty...
FAILED tests/pytorch/test_nn.py::test_gmm_conv[g5-idtype1] - TypeError: empty...
FAILED tests/pytorch/test_nn.py::test_gmm_conv_bi[g0-idtype0] - TypeError: em...
FAILED tests/pytorch/test_nn.py::test_gmm_conv_bi[g0-idtype1] - TypeError: em...
FAILED tests/pytorch/test_nn.py::test_gmm_conv_bi[g1-idtype0] - TypeError: em...
FAILED tests/pytorch/test_nn.py::test_gmm_conv_bi[g1-idtype1] - TypeError: em...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[1-g0-both-idtype0] - T...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[1-g0-both-idtype1] - T...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[1-g0-right-idtype0] - ...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[1-g0-right-idtype1] - ...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[1-g0-none-idtype0] - T...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[1-g0-none-idtype1] - T...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[1-g1-both-idtype0] - T...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[1-g1-both-idtype1] - T...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[1-g1-right-idtype0] - ...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[1-g1-right-idtype1] - ...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[1-g1-none-idtype0] - T...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[1-g1-none-idtype1] - T...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[1-g2-both-idtype0] - T...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[1-g2-both-idtype1] - T...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[1-g2-right-idtype0] - ...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[1-g2-right-idtype1] - ...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[1-g2-none-idtype0] - T...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[1-g2-none-idtype1] - T...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[1-g3-both-idtype0] - T...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[1-g3-both-idtype1] - T...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[1-g3-right-idtype0] - ...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[1-g3-right-idtype1] - ...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[1-g3-none-idtype0] - T...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[1-g3-none-idtype1] - T...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[1-g4-both-idtype0] - T...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[1-g4-both-idtype1] - T...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[1-g4-right-idtype0] - ...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[1-g4-right-idtype1] - ...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[1-g4-none-idtype0] - T...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[1-g4-none-idtype1] - T...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[1-g5-both-idtype0] - T...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[1-g5-both-idtype1] - T...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[1-g5-right-idtype0] - ...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[1-g5-right-idtype1] - ...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[1-g5-none-idtype0] - T...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[1-g5-none-idtype1] - T...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[1-g6-both-idtype0] - T...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[1-g6-both-idtype1] - T...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[1-g6-right-idtype0] - ...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[1-g6-right-idtype1] - ...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[1-g6-none-idtype0] - T...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[1-g6-none-idtype1] - T...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[1-g7-both-idtype0] - T...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[1-g7-both-idtype1] - T...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[1-g7-right-idtype0] - ...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[1-g7-right-idtype1] - ...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[1-g7-none-idtype0] - T...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[1-g7-none-idtype1] - T...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[2-g0-both-idtype0] - T...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[2-g0-both-idtype1] - T...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[2-g0-right-idtype0] - ...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[2-g0-right-idtype1] - ...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[2-g0-none-idtype0] - T...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[2-g0-none-idtype1] - T...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[2-g1-both-idtype0] - T...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[2-g1-both-idtype1] - T...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[2-g1-right-idtype0] - ...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[2-g1-right-idtype1] - ...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[2-g1-none-idtype0] - T...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[2-g1-none-idtype1] - T...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[2-g2-both-idtype0] - T...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[2-g2-both-idtype1] - T...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[2-g2-right-idtype0] - ...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[2-g2-right-idtype1] - ...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[2-g2-none-idtype0] - T...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[2-g2-none-idtype1] - T...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[2-g3-both-idtype0] - T...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[2-g3-both-idtype1] - T...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[2-g3-right-idtype0] - ...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[2-g3-right-idtype1] - ...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[2-g3-none-idtype0] - T...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[2-g3-none-idtype1] - T...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[2-g4-both-idtype0] - T...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[2-g4-both-idtype1] - T...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[2-g4-right-idtype0] - ...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[2-g4-right-idtype1] - ...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[2-g4-none-idtype0] - T...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[2-g4-none-idtype1] - T...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[2-g5-both-idtype0] - T...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[2-g5-both-idtype1] - T...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[2-g5-right-idtype0] - ...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[2-g5-right-idtype1] - ...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[2-g5-none-idtype0] - T...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[2-g5-none-idtype1] - T...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[2-g6-both-idtype0] - T...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[2-g6-both-idtype1] - T...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[2-g6-right-idtype0] - ...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[2-g6-right-idtype1] - ...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[2-g6-none-idtype0] - T...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[2-g6-none-idtype1] - T...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[2-g7-both-idtype0] - T...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[2-g7-both-idtype1] - T...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[2-g7-right-idtype0] - ...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[2-g7-right-idtype1] - ...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[2-g7-none-idtype0] - T...
FAILED tests/pytorch/test_nn.py::test_dense_graph_conv[2-g7-none-idtype1] - T...
FAILED tests/pytorch/test_nn.py::test_dense_sage_conv[1-g0-idtype0] - TypeErr...
FAILED tests/pytorch/test_nn.py::test_dense_sage_conv[1-g0-idtype1] - TypeErr...
FAILED tests/pytorch/test_nn.py::test_dense_sage_conv[1-g1-idtype0] - TypeErr...
FAILED tests/pytorch/test_nn.py::test_dense_sage_conv[1-g1-idtype1] - TypeErr...
FAILED tests/pytorch/test_nn.py::test_dense_sage_conv[1-g2-idtype0] - TypeErr...
FAILED tests/pytorch/test_nn.py::test_dense_sage_conv[1-g2-idtype1] - TypeErr...
FAILED tests/pytorch/test_nn.py::test_dense_sage_conv[1-g3-idtype0] - TypeErr...
FAILED tests/pytorch/test_nn.py::test_dense_sage_conv[1-g3-idtype1] - TypeErr...
FAILED tests/pytorch/test_nn.py::test_dense_sage_conv[1-g4-idtype0] - TypeErr...
FAILED tests/pytorch/test_nn.py::test_dense_sage_conv[1-g4-idtype1] - TypeErr...
FAILED tests/pytorch/test_nn.py::test_dense_sage_conv[1-g5-idtype0] - TypeErr...
FAILED tests/pytorch/test_nn.py::test_dense_sage_conv[1-g5-idtype1] - TypeErr...
FAILED tests/pytorch/test_nn.py::test_dense_sage_conv[1-g6-idtype0] - TypeErr...
FAILED tests/pytorch/test_nn.py::test_dense_sage_conv[1-g6-idtype1] - TypeErr...
FAILED tests/pytorch/test_nn.py::test_dense_sage_conv[1-g7-idtype0] - TypeErr...
FAILED tests/pytorch/test_nn.py::test_dense_sage_conv[1-g7-idtype1] - TypeErr...
FAILED tests/pytorch/test_nn.py::test_dense_sage_conv[1-g8-idtype0] - TypeErr...
FAILED tests/pytorch/test_nn.py::test_dense_sage_conv[1-g8-idtype1] - TypeErr...
FAILED tests/pytorch/test_nn.py::test_dense_sage_conv[1-g9-idtype0] - TypeErr...
FAILED tests/pytorch/test_nn.py::test_dense_sage_conv[1-g9-idtype1] - TypeErr...
FAILED tests/pytorch/test_nn.py::test_dense_sage_conv[2-g0-idtype0] - TypeErr...
FAILED tests/pytorch/test_nn.py::test_dense_sage_conv[2-g0-idtype1] - TypeErr...
FAILED tests/pytorch/test_nn.py::test_dense_sage_conv[2-g1-idtype0] - TypeErr...
FAILED tests/pytorch/test_nn.py::test_dense_sage_conv[2-g1-idtype1] - TypeErr...
FAILED tests/pytorch/test_nn.py::test_dense_sage_conv[2-g2-idtype0] - TypeErr...
FAILED tests/pytorch/test_nn.py::test_dense_sage_conv[2-g2-idtype1] - TypeErr...
FAILED tests/pytorch/test_nn.py::test_dense_sage_conv[2-g3-idtype0] - TypeErr...
FAILED tests/pytorch/test_nn.py::test_dense_sage_conv[2-g3-idtype1] - TypeErr...
FAILED tests/pytorch/test_nn.py::test_dense_sage_conv[2-g4-idtype0] - TypeErr...
FAILED tests/pytorch/test_nn.py::test_dense_sage_conv[2-g4-idtype1] - TypeErr...
FAILED tests/pytorch/test_nn.py::test_dense_sage_conv[2-g5-idtype0] - TypeErr...
FAILED tests/pytorch/test_nn.py::test_dense_sage_conv[2-g5-idtype1] - TypeErr...
FAILED tests/pytorch/test_nn.py::test_dense_sage_conv[2-g6-idtype0] - TypeErr...
FAILED tests/pytorch/test_nn.py::test_dense_sage_conv[2-g6-idtype1] - TypeErr...
FAILED tests/pytorch/test_nn.py::test_dense_sage_conv[2-g7-idtype0] - TypeErr...
FAILED tests/pytorch/test_nn.py::test_dense_sage_conv[2-g7-idtype1] - TypeErr...
FAILED tests/pytorch/test_nn.py::test_dense_sage_conv[2-g8-idtype0] - TypeErr...
FAILED tests/pytorch/test_nn.py::test_dense_sage_conv[2-g8-idtype1] - TypeErr...
FAILED tests/pytorch/test_nn.py::test_dense_sage_conv[2-g9-idtype0] - TypeErr...
FAILED tests/pytorch/test_nn.py::test_dense_sage_conv[2-g9-idtype1] - TypeErr...
FAILED tests/pytorch/test_nn.py::test_edge_conv[1-g0-idtype0] - TypeError: em...
FAILED tests/pytorch/test_nn.py::test_edge_conv[1-g0-idtype1] - TypeError: em...
FAILED tests/pytorch/test_nn.py::test_edge_conv[1-g1-idtype0] - TypeError: em...
FAILED tests/pytorch/test_nn.py::test_edge_conv[1-g1-idtype1] - TypeError: em...
FAILED tests/pytorch/test_nn.py::test_edge_conv[1-g2-idtype0] - TypeError: em...
FAILED tests/pytorch/test_nn.py::test_edge_conv[1-g2-idtype1] - TypeError: em...
FAILED tests/pytorch/test_nn.py::test_edge_conv[1-g3-idtype0] - TypeError: em...
FAILED tests/pytorch/test_nn.py::test_edge_conv[1-g3-idtype1] - TypeError: em...
FAILED tests/pytorch/test_nn.py::test_edge_conv[1-g4-idtype0] - TypeError: em...
FAILED tests/pytorch/test_nn.py::test_edge_conv[1-g4-idtype1] - TypeError: em...
FAILED tests/pytorch/test_nn.py::test_edge_conv[1-g5-idtype0] - TypeError: em...
FAILED tests/pytorch/test_nn.py::test_edge_conv[1-g5-idtype1] - TypeError: em...
FAILED tests/pytorch/test_nn.py::test_edge_conv[1-g6-idtype0] - TypeError: em...
FAILED tests/pytorch/test_nn.py::test_edge_conv[1-g6-idtype1] - TypeError: em...
FAILED tests/pytorch/test_nn.py::test_edge_conv[2-g0-idtype0] - TypeError: em...
FAILED tests/pytorch/test_nn.py::test_edge_conv[2-g0-idtype1] - TypeError: em...
FAILED tests/pytorch/test_nn.py::test_edge_conv[2-g1-idtype0] - TypeError: em...
FAILED tests/pytorch/test_nn.py::test_edge_conv[2-g1-idtype1] - TypeError: em...
FAILED tests/pytorch/test_nn.py::test_edge_conv[2-g2-idtype0] - TypeError: em...
FAILED tests/pytorch/test_nn.py::test_edge_conv[2-g2-idtype1] - TypeError: em...
FAILED tests/pytorch/test_nn.py::test_edge_conv[2-g3-idtype0] - TypeError: em...
FAILED tests/pytorch/test_nn.py::test_edge_conv[2-g3-idtype1] - TypeError: em...
FAILED tests/pytorch/test_nn.py::test_edge_conv[2-g4-idtype0] - TypeError: em...
FAILED tests/pytorch/test_nn.py::test_edge_conv[2-g4-idtype1] - TypeError: em...
FAILED tests/pytorch/test_nn.py::test_edge_conv[2-g5-idtype0] - TypeError: em...
FAILED tests/pytorch/test_nn.py::test_edge_conv[2-g5-idtype1] - TypeError: em...
FAILED tests/pytorch/test_nn.py::test_edge_conv[2-g6-idtype0] - TypeError: em...
FAILED tests/pytorch/test_nn.py::test_edge_conv[2-g6-idtype1] - TypeError: em...
FAILED tests/pytorch/test_nn.py::test_edge_conv_bi[1-g0-idtype0] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_edge_conv_bi[1-g0-idtype1] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_edge_conv_bi[1-g1-idtype0] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_edge_conv_bi[1-g1-idtype1] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_edge_conv_bi[2-g0-idtype0] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_edge_conv_bi[2-g0-idtype1] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_edge_conv_bi[2-g1-idtype0] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_edge_conv_bi[2-g1-idtype1] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_dotgat_conv[1-1-g0-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_dotgat_conv[1-1-g0-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_dotgat_conv[1-1-g1-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_dotgat_conv[1-1-g1-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_dotgat_conv[1-1-g2-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_dotgat_conv[1-1-g2-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_dotgat_conv[1-1-g3-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_dotgat_conv[1-1-g3-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_dotgat_conv[1-1-g4-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_dotgat_conv[1-1-g4-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_dotgat_conv[1-1-g5-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_dotgat_conv[1-1-g5-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_dotgat_conv[1-1-g6-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_dotgat_conv[1-1-g6-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_dotgat_conv[1-2-g0-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_dotgat_conv[1-2-g0-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_dotgat_conv[1-2-g1-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_dotgat_conv[1-2-g1-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_dotgat_conv[1-2-g2-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_dotgat_conv[1-2-g2-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_dotgat_conv[1-2-g3-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_dotgat_conv[1-2-g3-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_dotgat_conv[1-2-g4-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_dotgat_conv[1-2-g4-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_dotgat_conv[1-2-g5-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_dotgat_conv[1-2-g5-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_dotgat_conv[1-2-g6-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_dotgat_conv[1-2-g6-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_dotgat_conv[4-1-g0-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_dotgat_conv[4-1-g0-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_dotgat_conv[4-1-g1-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_dotgat_conv[4-1-g1-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_dotgat_conv[4-1-g2-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_dotgat_conv[4-1-g2-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_dotgat_conv[4-1-g3-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_dotgat_conv[4-1-g3-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_dotgat_conv[4-1-g4-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_dotgat_conv[4-1-g4-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_dotgat_conv[4-1-g5-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_dotgat_conv[4-1-g5-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_dotgat_conv[4-1-g6-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_dotgat_conv[4-1-g6-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_dotgat_conv[4-2-g0-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_dotgat_conv[4-2-g0-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_dotgat_conv[4-2-g1-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_dotgat_conv[4-2-g1-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_dotgat_conv[4-2-g2-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_dotgat_conv[4-2-g2-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_dotgat_conv[4-2-g3-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_dotgat_conv[4-2-g3-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_dotgat_conv[4-2-g4-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_dotgat_conv[4-2-g4-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_dotgat_conv[4-2-g5-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_dotgat_conv[4-2-g5-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_dotgat_conv[4-2-g6-idtype0] - TypeError...
FAILED tests/pytorch/test_nn.py::test_dotgat_conv[4-2-g6-idtype1] - TypeError...
FAILED tests/pytorch/test_nn.py::test_dotgat_conv_bi[1-1-g0-idtype0] - TypeEr...
FAILED tests/pytorch/test_nn.py::test_dotgat_conv_bi[1-1-g0-idtype1] - TypeEr...
FAILED tests/pytorch/test_nn.py::test_dotgat_conv_bi[1-1-g1-idtype0] - TypeEr...
FAILED tests/pytorch/test_nn.py::test_dotgat_conv_bi[1-1-g1-idtype1] - TypeEr...
FAILED tests/pytorch/test_nn.py::test_dotgat_conv_bi[1-2-g0-idtype0] - TypeEr...
FAILED tests/pytorch/test_nn.py::test_dotgat_conv_bi[1-2-g0-idtype1] - TypeEr...
FAILED tests/pytorch/test_nn.py::test_dotgat_conv_bi[1-2-g1-idtype0] - TypeEr...
FAILED tests/pytorch/test_nn.py::test_dotgat_conv_bi[1-2-g1-idtype1] - TypeEr...
FAILED tests/pytorch/test_nn.py::test_dotgat_conv_bi[4-1-g0-idtype0] - TypeEr...
FAILED tests/pytorch/test_nn.py::test_dotgat_conv_bi[4-1-g0-idtype1] - TypeEr...
FAILED tests/pytorch/test_nn.py::test_dotgat_conv_bi[4-1-g1-idtype0] - TypeEr...
FAILED tests/pytorch/test_nn.py::test_dotgat_conv_bi[4-1-g1-idtype1] - TypeEr...
FAILED tests/pytorch/test_nn.py::test_dotgat_conv_bi[4-2-g0-idtype0] - TypeEr...
FAILED tests/pytorch/test_nn.py::test_dotgat_conv_bi[4-2-g0-idtype1] - TypeEr...
FAILED tests/pytorch/test_nn.py::test_dotgat_conv_bi[4-2-g1-idtype0] - TypeEr...
FAILED tests/pytorch/test_nn.py::test_dotgat_conv_bi[4-2-g1-idtype1] - TypeEr...
FAILED tests/pytorch/test_nn.py::test_dense_cheb_conv[1] - TypeError: empty_c...
FAILED tests/pytorch/test_nn.py::test_dense_cheb_conv[2] - TypeError: empty_c...
FAILED tests/pytorch/test_nn.py::test_sequential - TypeError: empty_context()...
FAILED tests/pytorch/test_nn.py::test_cf_conv[1-g0-idtype0] - TypeError: empt...
FAILED tests/pytorch/test_nn.py::test_cf_conv[1-g0-idtype1] - TypeError: empt...
FAILED tests/pytorch/test_nn.py::test_cf_conv[1-g1-idtype0] - TypeError: empt...
FAILED tests/pytorch/test_nn.py::test_cf_conv[1-g1-idtype1] - TypeError: empt...
FAILED tests/pytorch/test_nn.py::test_cf_conv[1-g2-idtype0] - TypeError: empt...
FAILED tests/pytorch/test_nn.py::test_cf_conv[1-g2-idtype1] - TypeError: empt...
FAILED tests/pytorch/test_nn.py::test_cf_conv[1-g3-idtype0] - TypeError: empt...
FAILED tests/pytorch/test_nn.py::test_cf_conv[1-g3-idtype1] - TypeError: empt...
FAILED tests/pytorch/test_nn.py::test_cf_conv[1-g4-idtype0] - TypeError: empt...
FAILED tests/pytorch/test_nn.py::test_cf_conv[1-g4-idtype1] - TypeError: empt...
FAILED tests/pytorch/test_nn.py::test_cf_conv[1-g5-idtype0] - TypeError: empt...
FAILED tests/pytorch/test_nn.py::test_cf_conv[1-g5-idtype1] - TypeError: empt...
FAILED tests/pytorch/test_nn.py::test_cf_conv[1-g6-idtype0] - TypeError: empt...
FAILED tests/pytorch/test_nn.py::test_cf_conv[1-g6-idtype1] - TypeError: empt...
FAILED tests/pytorch/test_nn.py::test_cf_conv[1-g7-idtype0] - TypeError: empt...
FAILED tests/pytorch/test_nn.py::test_cf_conv[1-g7-idtype1] - TypeError: empt...
FAILED tests/pytorch/test_nn.py::test_cf_conv[3-g0-idtype0] - TypeError: empt...
FAILED tests/pytorch/test_nn.py::test_cf_conv[3-g0-idtype1] - TypeError: empt...
FAILED tests/pytorch/test_nn.py::test_cf_conv[3-g1-idtype0] - TypeError: empt...
FAILED tests/pytorch/test_nn.py::test_cf_conv[3-g1-idtype1] - TypeError: empt...
FAILED tests/pytorch/test_nn.py::test_cf_conv[3-g2-idtype0] - TypeError: empt...
FAILED tests/pytorch/test_nn.py::test_cf_conv[3-g2-idtype1] - TypeError: empt...
FAILED tests/pytorch/test_nn.py::test_cf_conv[3-g3-idtype0] - TypeError: empt...
FAILED tests/pytorch/test_nn.py::test_cf_conv[3-g3-idtype1] - TypeError: empt...
FAILED tests/pytorch/test_nn.py::test_cf_conv[3-g4-idtype0] - TypeError: empt...
FAILED tests/pytorch/test_nn.py::test_cf_conv[3-g4-idtype1] - TypeError: empt...
FAILED tests/pytorch/test_nn.py::test_cf_conv[3-g5-idtype0] - TypeError: empt...
FAILED tests/pytorch/test_nn.py::test_cf_conv[3-g5-idtype1] - TypeError: empt...
FAILED tests/pytorch/test_nn.py::test_cf_conv[3-g6-idtype0] - TypeError: empt...
FAILED tests/pytorch/test_nn.py::test_cf_conv[3-g6-idtype1] - TypeError: empt...
FAILED tests/pytorch/test_nn.py::test_cf_conv[3-g7-idtype0] - TypeError: empt...
FAILED tests/pytorch/test_nn.py::test_cf_conv[3-g7-idtype1] - TypeError: empt...
FAILED tests/pytorch/test_nn.py::test_hetero_conv[False-sum-idtype0] - TypeEr...
FAILED tests/pytorch/test_nn.py::test_hetero_conv[False-sum-idtype1] - TypeEr...
FAILED tests/pytorch/test_nn.py::test_hetero_conv[False-max-idtype0] - TypeEr...
FAILED tests/pytorch/test_nn.py::test_hetero_conv[False-max-idtype1] - TypeEr...
FAILED tests/pytorch/test_nn.py::test_hetero_conv[False-min-idtype0] - TypeEr...
FAILED tests/pytorch/test_nn.py::test_hetero_conv[False-min-idtype1] - TypeEr...
FAILED tests/pytorch/test_nn.py::test_hetero_conv[False-mean-idtype0] - TypeE...
FAILED tests/pytorch/test_nn.py::test_hetero_conv[False-mean-idtype1] - TypeE...
FAILED tests/pytorch/test_nn.py::test_hetero_conv[False-stack-idtype0] - Type...
FAILED tests/pytorch/test_nn.py::test_hetero_conv[False-stack-idtype1] - Type...
FAILED tests/pytorch/test_nn.py::test_hetero_conv[False-myagg-idtype0] - Type...
FAILED tests/pytorch/test_nn.py::test_hetero_conv[False-myagg-idtype1] - Type...
FAILED tests/pytorch/test_nn.py::test_hetero_conv[True-sum-idtype0] - TypeErr...
FAILED tests/pytorch/test_nn.py::test_hetero_conv[True-sum-idtype1] - TypeErr...
FAILED tests/pytorch/test_nn.py::test_hetero_conv[True-max-idtype0] - TypeErr...
FAILED tests/pytorch/test_nn.py::test_hetero_conv[True-max-idtype1] - TypeErr...
FAILED tests/pytorch/test_nn.py::test_hetero_conv[True-min-idtype0] - TypeErr...
FAILED tests/pytorch/test_nn.py::test_hetero_conv[True-min-idtype1] - TypeErr...
FAILED tests/pytorch/test_nn.py::test_hetero_conv[True-mean-idtype0] - TypeEr...
FAILED tests/pytorch/test_nn.py::test_hetero_conv[True-mean-idtype1] - TypeEr...
FAILED tests/pytorch/test_nn.py::test_hetero_conv[True-stack-idtype0] - TypeE...
FAILED tests/pytorch/test_nn.py::test_hetero_conv[True-stack-idtype1] - TypeE...
FAILED tests/pytorch/test_nn.py::test_hetero_conv[True-myagg-idtype0] - TypeE...
FAILED tests/pytorch/test_nn.py::test_hetero_conv[True-myagg-idtype1] - TypeE...
FAILED tests/pytorch/test_nn.py::test_gnnexplainer[1-g0-idtype0] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gnnexplainer[1-g0-idtype1] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gnnexplainer[1-g1-idtype0] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gnnexplainer[1-g1-idtype1] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gnnexplainer[1-g2-idtype0] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gnnexplainer[1-g2-idtype1] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gnnexplainer[1-g3-idtype0] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gnnexplainer[1-g3-idtype1] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gnnexplainer[1-g4-idtype0] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gnnexplainer[1-g4-idtype1] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gnnexplainer[1-g5-idtype0] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gnnexplainer[1-g5-idtype1] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gnnexplainer[2-g0-idtype0] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gnnexplainer[2-g0-idtype1] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gnnexplainer[2-g1-idtype0] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gnnexplainer[2-g1-idtype1] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gnnexplainer[2-g2-idtype0] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gnnexplainer[2-g2-idtype1] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gnnexplainer[2-g3-idtype0] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gnnexplainer[2-g3-idtype1] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gnnexplainer[2-g4-idtype0] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gnnexplainer[2-g4-idtype1] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gnnexplainer[2-g5-idtype0] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_gnnexplainer[2-g5-idtype1] - TypeError:...
FAILED tests/pytorch/test_nn.py::test_twirls - TypeError: empty_context() got...
FAILED tests/pytorch/test_nn.py::test_hgt[1-4-idtype0] - TypeError: empty_con...
FAILED tests/pytorch/test_nn.py::test_hgt[1-4-idtype1] - TypeError: empty_con...
FAILED tests/pytorch/test_nn.py::test_group_rev_res[idtype0] - TypeError: emp...
FAILED tests/pytorch/test_nn.py::test_group_rev_res[idtype1] - TypeError: emp...
FAILED tests/pytorch/test_nn.py::test_egnn_conv[16-16-16-16] - TypeError: emp...
FAILED tests/pytorch/test_nn.py::test_egnn_conv[16-16-16-32] - TypeError: emp...
FAILED tests/pytorch/test_nn.py::test_egnn_conv[16-16-32-16] - TypeError: emp...
FAILED tests/pytorch/test_nn.py::test_egnn_conv[16-16-32-32] - TypeError: emp...
FAILED tests/pytorch/test_nn.py::test_egnn_conv[16-32-16-16] - TypeError: emp...
FAILED tests/pytorch/test_nn.py::test_egnn_conv[16-32-16-32] - TypeError: emp...
FAILED tests/pytorch/test_nn.py::test_egnn_conv[16-32-32-16] - TypeError: emp...
FAILED tests/pytorch/test_nn.py::test_egnn_conv[16-32-32-32] - TypeError: emp...
FAILED tests/pytorch/test_nn.py::test_egnn_conv[10-16-16-16] - TypeError: emp...
FAILED tests/pytorch/test_nn.py::test_egnn_conv[10-16-16-32] - TypeError: emp...
FAILED tests/pytorch/test_nn.py::test_egnn_conv[10-16-32-16] - TypeError: emp...
FAILED tests/pytorch/test_nn.py::test_egnn_conv[10-16-32-32] - TypeError: emp...
FAILED tests/pytorch/test_nn.py::test_egnn_conv[10-32-16-16] - TypeError: emp...
FAILED tests/pytorch/test_nn.py::test_egnn_conv[10-32-16-32] - TypeError: emp...
FAILED tests/pytorch/test_nn.py::test_egnn_conv[10-32-32-16] - TypeError: emp...
FAILED tests/pytorch/test_nn.py::test_egnn_conv[10-32-32-32] - TypeError: emp...
FAILED tests/pytorch/test_nn.py::test_egnn_conv[0-16-16-16] - TypeError: empt...
FAILED tests/pytorch/test_nn.py::test_egnn_conv[0-16-16-32] - TypeError: empt...
FAILED tests/pytorch/test_nn.py::test_egnn_conv[0-16-32-16] - TypeError: empt...
FAILED tests/pytorch/test_nn.py::test_egnn_conv[0-16-32-32] - TypeError: empt...
FAILED tests/pytorch/test_nn.py::test_egnn_conv[0-32-16-16] - TypeError: empt...
FAILED tests/pytorch/test_nn.py::test_egnn_conv[0-32-16-32] - TypeError: empt...
FAILED tests/pytorch/test_nn.py::test_egnn_conv[0-32-32-16] - TypeError: empt...
FAILED tests/pytorch/test_nn.py::test_egnn_conv[0-32-32-32] - TypeError: empt...
FAILED tests/pytorch/test_nn.py::test_label_prop[True-True-True-sym-0.0-3] - ...
FAILED tests/pytorch/test_nn.py::test_label_prop[True-True-True-sym-0.0-5] - ...
FAILED tests/pytorch/test_nn.py::test_label_prop[True-True-True-sym-0.5-3] - ...
FAILED tests/pytorch/test_nn.py::test_label_prop[True-True-True-sym-0.5-5] - ...
FAILED tests/pytorch/test_nn.py::test_label_prop[True-True-True-sym-1.0-3] - ...
FAILED tests/pytorch/test_nn.py::test_label_prop[True-True-True-sym-1.0-5] - ...
FAILED tests/pytorch/test_nn.py::test_label_prop[True-True-True-row-0.0-3] - ...
FAILED tests/pytorch/test_nn.py::test_label_prop[True-True-True-row-0.0-5] - ...
FAILED tests/pytorch/test_nn.py::test_label_prop[True-True-True-row-0.5-3] - ...
FAILED tests/pytorch/test_nn.py::test_label_prop[True-True-True-row-0.5-5] - ...
FAILED tests/pytorch/test_nn.py::test_label_prop[True-True-True-row-1.0-3] - ...
FAILED tests/pytorch/test_nn.py::test_label_prop[True-True-True-row-1.0-5] - ...
FAILED tests/pytorch/test_nn.py::test_label_prop[True-True-False-sym-0.0-3]
FAILED tests/pytorch/test_nn.py::test_label_prop[True-True-False-sym-0.0-5]
FAILED tests/pytorch/test_nn.py::test_label_prop[True-True-False-sym-0.5-3]
FAILED tests/pytorch/test_nn.py::test_label_prop[True-True-False-sym-0.5-5]
FAILED tests/pytorch/test_nn.py::test_label_prop[True-True-False-sym-1.0-3]
FAILED tests/pytorch/test_nn.py::test_label_prop[True-True-False-sym-1.0-5]
FAILED tests/pytorch/test_nn.py::test_label_prop[True-True-False-row-0.0-3]
FAILED tests/pytorch/test_nn.py::test_label_prop[True-True-False-row-0.0-5]
FAILED tests/pytorch/test_nn.py::test_label_prop[True-True-False-row-0.5-3]
FAILED tests/pytorch/test_nn.py::test_label_prop[True-True-False-row-0.5-5]
FAILED tests/pytorch/test_nn.py::test_label_prop[True-True-False-row-1.0-3]
FAILED tests/pytorch/test_nn.py::test_label_prop[True-True-False-row-1.0-5]
FAILED tests/pytorch/test_nn.py::test_label_prop[True-False-True-sym-0.0-3]
FAILED tests/pytorch/test_nn.py::test_label_prop[True-False-True-sym-0.0-5]
FAILED tests/pytorch/test_nn.py::test_label_prop[True-False-True-sym-0.5-3]
FAILED tests/pytorch/test_nn.py::test_label_prop[True-False-True-sym-0.5-5]
FAILED tests/pytorch/test_nn.py::test_label_prop[True-False-True-sym-1.0-3]
FAILED tests/pytorch/test_nn.py::test_label_prop[True-False-True-sym-1.0-5]
FAILED tests/pytorch/test_nn.py::test_label_prop[True-False-True-row-0.0-3]
FAILED tests/pytorch/test_nn.py::test_label_prop[True-False-True-row-0.0-5]
FAILED tests/pytorch/test_nn.py::test_label_prop[True-False-True-row-0.5-3]
FAILED tests/pytorch/test_nn.py::test_label_prop[True-False-True-row-0.5-5]
FAILED tests/pytorch/test_nn.py::test_label_prop[True-False-True-row-1.0-3]
FAILED tests/pytorch/test_nn.py::test_label_prop[True-False-True-row-1.0-5]
FAILED tests/pytorch/test_nn.py::test_label_prop[True-False-False-sym-0.0-3]
FAILED tests/pytorch/test_nn.py::test_label_prop[True-False-False-sym-0.0-5]
FAILED tests/pytorch/test_nn.py::test_label_prop[True-False-False-sym-0.5-3]
FAILED tests/pytorch/test_nn.py::test_label_prop[True-False-False-sym-0.5-5]
FAILED tests/pytorch/test_nn.py::test_label_prop[True-False-False-sym-1.0-3]
FAILED tests/pytorch/test_nn.py::test_label_prop[True-False-False-sym-1.0-5]
FAILED tests/pytorch/test_nn.py::test_label_prop[True-False-False-row-0.0-3]
FAILED tests/pytorch/test_nn.py::test_label_prop[True-False-False-row-0.0-5]
FAILED tests/pytorch/test_nn.py::test_label_prop[True-False-False-row-0.5-3]
FAILED tests/pytorch/test_nn.py::test_label_prop[True-False-False-row-0.5-5]
FAILED tests/pytorch/test_nn.py::test_label_prop[True-False-False-row-1.0-3]
FAILED tests/pytorch/test_nn.py::test_label_prop[True-False-False-row-1.0-5]
FAILED tests/pytorch/test_nn.py::test_label_prop[False-True-True-sym-0.0-3]
FAILED tests/pytorch/test_nn.py::test_label_prop[False-True-True-sym-0.0-5]
FAILED tests/pytorch/test_nn.py::test_label_prop[False-True-True-sym-0.5-3]
FAILED tests/pytorch/test_nn.py::test_label_prop[False-True-True-sym-0.5-5]
FAILED tests/pytorch/test_nn.py::test_label_prop[False-True-True-sym-1.0-3]
FAILED tests/pytorch/test_nn.py::test_label_prop[False-True-True-sym-1.0-5]
FAILED tests/pytorch/test_nn.py::test_label_prop[False-True-True-row-0.0-3]
FAILED tests/pytorch/test_nn.py::test_label_prop[False-True-True-row-0.0-5]
FAILED tests/pytorch/test_nn.py::test_label_prop[False-True-True-row-0.5-3]
FAILED tests/pytorch/test_nn.py::test_label_prop[False-True-True-row-0.5-5]
FAILED tests/pytorch/test_nn.py::test_label_prop[False-True-True-row-1.0-3]
FAILED tests/pytorch/test_nn.py::test_label_prop[False-True-True-row-1.0-5]
FAILED tests/pytorch/test_nn.py::test_label_prop[False-True-False-sym-0.0-3]
FAILED tests/pytorch/test_nn.py::test_label_prop[False-True-False-sym-0.0-5]
FAILED tests/pytorch/test_nn.py::test_label_prop[False-True-False-sym-0.5-3]
FAILED tests/pytorch/test_nn.py::test_label_prop[False-True-False-sym-0.5-5]
FAILED tests/pytorch/test_nn.py::test_label_prop[False-True-False-sym-1.0-3]
FAILED tests/pytorch/test_nn.py::test_label_prop[False-True-False-sym-1.0-5]
FAILED tests/pytorch/test_nn.py::test_label_prop[False-True-False-row-0.0-3]
FAILED tests/pytorch/test_nn.py::test_label_prop[False-True-False-row-0.0-5]
FAILED tests/pytorch/test_nn.py::test_label_prop[False-True-False-row-0.5-3]
FAILED tests/pytorch/test_nn.py::test_label_prop[False-True-False-row-0.5-5]
FAILED tests/pytorch/test_nn.py::test_label_prop[False-True-False-row-1.0-3]
FAILED tests/pytorch/test_nn.py::test_label_prop[False-True-False-row-1.0-5]
FAILED tests/pytorch/test_nn.py::test_label_prop[False-False-True-sym-0.0-3]
FAILED tests/pytorch/test_nn.py::test_label_prop[False-False-True-sym-0.0-5]
FAILED tests/pytorch/test_nn.py::test_label_prop[False-False-True-sym-0.5-3]
FAILED tests/pytorch/test_nn.py::test_label_prop[False-False-True-sym-0.5-5]
FAILED tests/pytorch/test_nn.py::test_label_prop[False-False-True-sym-1.0-3]
FAILED tests/pytorch/test_nn.py::test_label_prop[False-False-True-sym-1.0-5]
FAILED tests/pytorch/test_nn.py::test_label_prop[False-False-True-row-0.0-3]
FAILED tests/pytorch/test_nn.py::test_label_prop[False-False-True-row-0.0-5]
FAILED tests/pytorch/test_nn.py::test_label_prop[False-False-True-row-0.5-3]
FAILED tests/pytorch/test_nn.py::test_label_prop[False-False-True-row-0.5-5]
FAILED tests/pytorch/test_nn.py::test_label_prop[False-False-True-row-1.0-3]
FAILED tests/pytorch/test_nn.py::test_label_prop[False-False-True-row-1.0-5]
FAILED tests/pytorch/test_nn.py::test_label_prop[False-False-False-sym-0.0-3]
FAILED tests/pytorch/test_nn.py::test_label_prop[False-False-False-sym-0.0-5]
FAILED tests/pytorch/test_nn.py::test_label_prop[False-False-False-sym-0.5-3]
FAILED tests/pytorch/test_nn.py::test_label_prop[False-False-False-sym-0.5-5]
FAILED tests/pytorch/test_nn.py::test_label_prop[False-False-False-sym-1.0-3]
FAILED tests/pytorch/test_nn.py::test_label_prop[False-False-False-sym-1.0-5]
FAILED tests/pytorch/test_nn.py::test_label_prop[False-False-False-row-0.0-3]
FAILED tests/pytorch/test_nn.py::test_label_prop[False-False-False-row-0.0-5]
FAILED tests/pytorch/test_nn.py::test_label_prop[False-False-False-row-0.5-3]
FAILED tests/pytorch/test_nn.py::test_label_prop[False-False-False-row-0.5-5]
FAILED tests/pytorch/test_nn.py::test_label_prop[False-False-False-row-1.0-3]
FAILED tests/pytorch/test_nn.py::test_label_prop[False-False-False-row-1.0-5]
== 1811 failed, 1812 passed, 355 skipped, 146 warnings in 928.17s (0:15:28) ===
Submodule path 'third_party/libxsmm': checked out 'fa687556130b6298430f1c0555a14cf79ab6101c'
Submodule path 'third_party/nanoflann': checked out '4c47ca200209550c5628c89803591f8a753c8181'
Submodule path 'third_party/nccl': checked out 'e11238b3029795d33f958b5868d47c90c4f22628'
Submodule path 'third_party/phmap': checked out '25293cefd8b85491b45600c03fe8edf07647553f'
Submodule path 'third_party/tensorpipe': checked out '6042f1a4cbce8eef997f11ed0012de137b317361'
Submodule 'third_party/googletest' (https://github.com/google/googletest.git) registered for path 'third_party/tensorpipe/third_party/googletest'
Submodule 'third_party/libnop' (https://github.com/google/libnop.git) registered for path 'third_party/tensorpipe/third_party/libnop'
Submodule 'third_party/libuv' (https://github.com/libuv/libuv.git) registered for path 'third_party/tensorpipe/third_party/libuv'
Submodule 'third_party/pybind11' (https://github.com/pybind/pybind11.git) registered for path 'third_party/tensorpipe/third_party/pybind11'
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tensorpipe/third_party/googletest'...
[Pipeline] }
[Pipeline] // timeout
[Pipeline] }
[Pipeline] // stage
[Pipeline] stage
[Pipeline] { (Torch CPU (Win64) Example test)
Stage "Torch CPU (Win64) Example test" skipped due to earlier failure(s)
[Pipeline] }
[Pipeline] // stage
Post stage
[Pipeline] cleanWs
[WS-CLEANUP] Deleting project workspace...
[WS-CLEANUP] Deferred wipeout is disabled by the job configuration...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tensorpipe/third_party/libnop'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tensorpipe/third_party/libuv'...
[WS-CLEANUP] done
[Pipeline] }
[Pipeline] // withEnv
[Pipeline] }
[Pipeline] // node
[Pipeline] }
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tensorpipe/third_party/pybind11'...
[Pipeline] // stage
[Pipeline] }
Failed in branch Torch CPU (Win64)
Submodule path 'third_party/tensorpipe/third_party/googletest': checked out 'aee0f9d9b5b87796ee8a0ab26b7587ec30e8858e'
Submodule path 'third_party/tensorpipe/third_party/libnop': checked out 'aa95422ea8c409e3f078d2ee7708a5f59a8b9fa2'
Submodule path 'third_party/tensorpipe/third_party/libuv': checked out '1dff88e5161cba5c59276d2070d2e304e4dcb242'
Submodule path 'third_party/tensorpipe/third_party/pybind11': checked out 'a23996fce38ff6ccfbcdc09f1e63f2c4be5ea2ef'
Submodule 'tools/clang' (https://github.com/wjakob/clang-cindex-python3) registered for path 'third_party/tensorpipe/third_party/pybind11/tools/clang'
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tensorpipe/third_party/pybind11/tools/clang'...
Submodule path 'third_party/tensorpipe/third_party/pybind11/tools/clang': checked out '6a00cbc4a9b8e68b71caf7f774b3f9c753ae84d5'
Submodule path 'third_party/thrust': checked out '6a3078c64cab0e2f276340fa5dcafa0d758ed890'
Submodule 'cub' (https://github.com/NVIDIA/cub.git) registered for path 'third_party/thrust/dependencies/cub'
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/thrust/dependencies/cub'...
Submodule path 'third_party/thrust/dependencies/cub': checked out 'cdaa9558a85e45d849016e5fe7b6e4ee79113f95'
Submodule path 'third_party/tvm': checked out 'b2e418cb109df4cd1f17a2cf2894a1b396a6b838'
Submodule 'dlpack' (https://github.com/dmlc/dlpack) registered for path 'third_party/tvm/3rdparty/dlpack'
Submodule 'dmlc-core' (https://github.com/dmlc/dmlc-core) registered for path 'third_party/tvm/3rdparty/dmlc-core'
Submodule '3rdparty/rang' (https://github.com/agauniyal/rang) registered for path 'third_party/tvm/3rdparty/rang'
Submodule '3rdparty/vta-hw' (https://github.com/apache/incubator-tvm-vta) registered for path 'third_party/tvm/3rdparty/vta-hw'
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tvm/3rdparty/dlpack'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tvm/3rdparty/dmlc-core'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tvm/3rdparty/rang'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tvm/3rdparty/vta-hw'...
Submodule path 'third_party/tvm/3rdparty/dlpack': checked out '3ec04430e89a6834e5a1b99471f415fa939bf642'
Submodule path 'third_party/tvm/3rdparty/dmlc-core': checked out '6c401e242c59a1f4c913918246591bb13fd714e7'
Submodule path 'third_party/tvm/3rdparty/rang': checked out 'cabe04d6d6b05356fa8f9741704924788f0dd762'
Submodule path 'third_party/tvm/3rdparty/vta-hw': checked out '87ce9acfae550d1a487746e9d06c2e250076e54c'
Submodule path 'third_party/xbyak': checked out '757e4063f6464740b8ff4a2cae9136d2f8458020'
[Pipeline] unstash
[Pipeline] echo
Unpacked build/libdgl.so, build/runUnitTests, python/dgl/_ffi/_cy3/core.cpython-*-x86_64-linux-gnu.so, build/tensoradapter/pytorch/*.so from dgl-gpu-linux
[Pipeline] timeout
Timeout set to expire in 30 min
[Pipeline] {
[Pipeline] sh
+ bash tests/scripts/task_unit_test.sh tensorflow gpu
Requirement already satisfied: pytest in /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages (7.1.2)
Collecting psutil
  Downloading psutil-5.9.2-cp37-cp37m-manylinux_2_12_x86_64.manylinux2010_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl (281 kB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 281.3/281.3 kB 9.8 MB/s eta 0:00:00
Collecting pyyaml
  Downloading PyYAML-6.0-cp37-cp37m-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_12_x86_64.manylinux2010_x86_64.whl (596 kB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 596.3/596.3 kB 38.5 MB/s eta 0:00:00
Collecting pydantic
  Downloading pydantic-1.10.2-cp37-cp37m-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (11.8 MB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 11.8/11.8 MB 116.7 MB/s eta 0:00:00
Collecting pandas
  Using cached pandas-1.1.5-cp37-cp37m-manylinux1_x86_64.whl (9.5 MB)
Collecting rdflib
  Downloading rdflib-6.2.0-py3-none-any.whl (500 kB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 500.3/500.3 kB 66.1 MB/s eta 0:00:00
Collecting ogb
  Downloading ogb-1.3.4-py3-none-any.whl (78 kB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 78.6/78.6 kB 17.8 MB/s eta 0:00:00
Requirement already satisfied: attrs>=19.2.0 in /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages (from pytest) (22.1.0)
Requirement already satisfied: pluggy<2.0,>=0.12 in /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages (from pytest) (1.0.0)
Requirement already satisfied: py>=1.8.2 in /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages (from pytest) (1.11.0)
Requirement already satisfied: importlib-metadata>=0.12 in /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages (from pytest) (4.12.0)
Requirement already satisfied: iniconfig in /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages (from pytest) (1.1.1)
Requirement already satisfied: packaging in /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages (from pytest) (21.3)
Requirement already satisfied: tomli>=1.0.0 in /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages (from pytest) (2.0.1)
Requirement already satisfied: typing-extensions>=4.1.0 in /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages (from pydantic) (4.3.0)
Requirement already satisfied: numpy>=1.15.4 in /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages (from pandas) (1.18.5)
Collecting pytz>=2017.2
  Using cached pytz-2022.2.1-py2.py3-none-any.whl (500 kB)
Requirement already satisfied: python-dateutil>=2.7.3 in /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages (from pandas) (2.8.2)
Requirement already satisfied: setuptools in /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages (from rdflib) (61.2.0)
Requirement already satisfied: pyparsing in /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages (from rdflib) (3.0.9)
Collecting isodate
  Downloading isodate-0.6.1-py2.py3-none-any.whl (41 kB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 41.7/41.7 kB 9.0 MB/s eta 0:00:00
Collecting scikit-learn>=0.20.0
  Using cached scikit_learn-1.0.2-cp37-cp37m-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (24.8 MB)
Requirement already satisfied: six>=1.12.0 in /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages (from ogb) (1.16.0)
Requirement already satisfied: tqdm>=4.29.0 in /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages (from ogb) (4.64.0)
Requirement already satisfied: urllib3>=1.24.0 in /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages (from ogb) (1.26.11)
Collecting torch>=1.6.0
  Downloading torch-1.12.1-cp37-cp37m-manylinux1_x86_64.whl (776.3 MB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 776.3/776.3 MB 4.3 MB/s eta 0:00:00
Collecting outdated>=0.2.0
  Using cached outdated-0.2.1-py3-none-any.whl (7.5 kB)
Requirement already satisfied: zipp>=0.5 in /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages (from importlib-metadata>=0.12->pytest) (3.8.1)
Collecting littleutils
  Using cached littleutils-0.2.2-py3-none-any.whl
Requirement already satisfied: requests in /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages (from outdated>=0.2.0->ogb) (2.28.1)
Collecting threadpoolctl>=2.0.0
  Using cached threadpoolctl-3.1.0-py3-none-any.whl (14 kB)
Requirement already satisfied: scipy>=1.1.0 in /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages (from scikit-learn>=0.20.0->ogb) (1.4.1)
Requirement already satisfied: joblib>=0.11 in /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages (from scikit-learn>=0.20.0->ogb) (1.1.0)
Requirement already satisfied: charset-normalizer<3,>=2 in /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages (from requests->outdated>=0.2.0->ogb) (2.1.0)
Requirement already satisfied: idna<4,>=2.5 in /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages (from requests->outdated>=0.2.0->ogb) (3.3)
Requirement already satisfied: certifi>=2017.4.17 in /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages (from requests->outdated>=0.2.0->ogb) (2022.6.15)
Installing collected packages: pytz, littleutils, torch, threadpoolctl, pyyaml, pydantic, psutil, isodate, scikit-learn, rdflib, pandas, outdated, ogb
Successfully installed isodate-0.6.1 littleutils-0.2.2 ogb-1.3.4 outdated-0.2.1 pandas-1.1.5 psutil-5.9.2 pydantic-1.10.2 pytz-2022.2.1 pyyaml-6.0 rdflib-6.2.0 scikit-learn-1.0.2 threadpoolctl-3.1.0 torch-1.12.1
WARNING: Running pip as the 'root' user can result in broken permissions and conflicting behaviour with the system package manager. It is recommended to use a virtual environment instead: https://pip.pypa.io/warnings/venv
============================= test session starts ==============================
platform linux -- Python 3.7.0, pytest-7.1.2, pluggy-1.0.0 -- /opt/conda/envs/tensorflow-ci/bin/python3
cachedir: .pytest_cache
rootdir: /root/jenkins/workspace/dgl_PR-4648
collecting ... collected 2458 items

tests/compute/test_apply_edges_hetero.py::test_unary_copy_u[idtype0] SKIPPED [  0%]
tests/compute/test_apply_edges_hetero.py::test_unary_copy_u[idtype1] SKIPPED [  0%]
tests/compute/test_apply_edges_hetero.py::test_unary_copy_e[idtype0] SKIPPED [  0%]
tests/compute/test_apply_edges_hetero.py::test_unary_copy_e[idtype1] SKIPPED [  0%]
tests/compute/test_apply_edges_hetero.py::test_binary_op[idtype0] SKIPPED [  0%]
tests/compute/test_apply_edges_hetero.py::test_binary_op[idtype1] SKIPPED [  0%]
tests/compute/test_backend.py::test_set_default_backend PASSED           [  0%]
tests/compute/test_basics.py::test_compatible PASSED                     [  0%]
tests/compute/test_basics.py::test_batch_setter_getter[idtype0] PASSED   [  0%]
tests/compute/test_basics.py::test_batch_setter_getter[idtype1] PASSED   [  0%]
tests/compute/test_basics.py::test_batch_setter_autograd[idtype0] PASSED [  0%]
tests/compute/test_basics.py::test_batch_setter_autograd[idtype1] PASSED [  0%]
tests/compute/test_basics.py::test_apply_nodes[idtype0] PASSED           [  0%]
tests/compute/test_basics.py::test_apply_nodes[idtype1] PASSED           [  0%]
tests/compute/test_basics.py::test_apply_edges[idtype0] PASSED           [  0%]
tests/compute/test_basics.py::test_apply_edges[idtype1] PASSED           [  0%]
tests/compute/test_basics.py::test_update_routines[idtype0] PASSED       [  0%]
tests/compute/test_basics.py::test_update_routines[idtype1] PASSED       [  0%]
tests/compute/test_basics.py::test_update_all_0deg[idtype0] PASSED       [  0%]
tests/compute/test_basics.py::test_update_all_0deg[idtype1] PASSED       [  0%]
tests/compute/test_basics.py::test_pull_0deg[idtype0] PASSED             [  0%]
tests/compute/test_basics.py::test_pull_0deg[idtype1] PASSED             [  0%]
tests/compute/test_basics.py::test_dynamic_addition PASSED               [  0%]
tests/compute/test_basics.py::test_repr[idtype0] PASSED                  [  0%]
tests/compute/test_basics.py::test_repr[idtype1] PASSED                  [  1%]
tests/compute/test_basics.py::test_local_var[idtype0] PASSED             [  1%]
tests/compute/test_basics.py::test_local_var[idtype1] PASSED             [  1%]
tests/compute/test_basics.py::test_local_scope[idtype0] PASSED           [  1%]
tests/compute/test_basics.py::test_local_scope[idtype1] PASSED           [  1%]
tests/compute/test_basics.py::test_isolated_nodes[idtype0] PASSED        [  1%]
tests/compute/test_basics.py::test_isolated_nodes[idtype1] PASSED        [  1%]
tests/compute/test_basics.py::test_send_multigraph[idtype0] PASSED       [  1%]
tests/compute/test_basics.py::test_send_multigraph[idtype1] PASSED       [  1%]
tests/compute/test_basics.py::test_issue_1088[idtype0] PASSED            [  1%]
tests/compute/test_basics.py::test_issue_1088[idtype1] PASSED            [  1%]
tests/compute/test_basics.py::test_degree_bucket_edge_ordering[idtype0] PASSED [  1%]
tests/compute/test_basics.py::test_degree_bucket_edge_ordering[idtype1] PASSED [  1%]
tests/compute/test_basics.py::test_issue_2484[idtype0] PASSED            [  1%]
tests/compute/test_basics.py::test_issue_2484[idtype1] PASSED            [  1%]
tests/compute/test_batched_graph.py::test_batch_unbatch[idtype0] PASSED  [  1%]
tests/compute/test_batched_graph.py::test_batch_unbatch[idtype1] PASSED  [  1%]
tests/compute/test_batched_graph.py::test_batch_unbatch1[idtype0] PASSED [  1%]
tests/compute/test_batched_graph.py::test_batch_unbatch1[idtype1] PASSED [  1%]
tests/compute/test_batched_graph.py::test_batch_unbatch_frame[idtype0] SKIPPED [  1%]
tests/compute/test_batched_graph.py::test_batch_unbatch_frame[idtype1] SKIPPED [  1%]
tests/compute/test_batched_graph.py::test_batch_unbatch2[idtype0] PASSED [  1%]
tests/compute/test_batched_graph.py::test_batch_unbatch2[idtype1] PASSED [  1%]
tests/compute/test_batched_graph.py::test_batch_send_and_recv[idtype0] PASSED [  1%]
tests/compute/test_batched_graph.py::test_batch_send_and_recv[idtype1] PASSED [  1%]
tests/compute/test_batched_graph.py::test_batch_propagate[idtype0] PASSED [  2%]
tests/compute/test_batched_graph.py::test_batch_propagate[idtype1] PASSED [  2%]
tests/compute/test_batched_graph.py::test_batched_edge_ordering[idtype0] PASSED [  2%]
tests/compute/test_batched_graph.py::test_batched_edge_ordering[idtype1] PASSED [  2%]
tests/compute/test_batched_graph.py::test_batch_no_edge[idtype0] PASSED  [  2%]
tests/compute/test_batched_graph.py::test_batch_no_edge[idtype1] PASSED  [  2%]
tests/compute/test_batched_graph.py::test_batch_keeps_empty_data[idtype0] PASSED [  2%]
tests/compute/test_batched_graph.py::test_batch_keeps_empty_data[idtype1] PASSED [  2%]
tests/compute/test_batched_graph.py::test_set_batch_info[idtype0] PASSED [  2%]
tests/compute/test_batched_graph.py::test_set_batch_info[idtype1] PASSED [  2%]
tests/compute/test_batched_heterograph.py::test_topology[idtype0-gs0] PASSED [  2%]
tests/compute/test_batched_heterograph.py::test_topology[idtype0-gs1] PASSED [  2%]
tests/compute/test_batched_heterograph.py::test_topology[idtype1-gs0] PASSED [  2%]
tests/compute/test_batched_heterograph.py::test_topology[idtype1-gs1] PASSED [  2%]
tests/compute/test_batched_heterograph.py::test_batching_batched[idtype0] PASSED [  2%]
tests/compute/test_batched_heterograph.py::test_batching_batched[idtype1] PASSED [  2%]
tests/compute/test_batched_heterograph.py::test_features[idtype0] PASSED [  2%]
tests/compute/test_batched_heterograph.py::test_features[idtype1] PASSED [  2%]
tests/compute/test_batched_heterograph.py::test_empty_relation[idtype0] PASSED [  2%]
tests/compute/test_batched_heterograph.py::test_empty_relation[idtype1] PASSED [  2%]
tests/compute/test_batched_heterograph.py::test_unbatch2[idtype0] PASSED [  2%]
tests/compute/test_batched_heterograph.py::test_unbatch2[idtype1] PASSED [  2%]
tests/compute/test_batched_heterograph.py::test_slice_batch[idtype0] PASSED [  2%]
tests/compute/test_batched_heterograph.py::test_slice_batch[idtype1] PASSED [  2%]
tests/compute/test_batched_heterograph.py::test_batch_keeps_empty_data[idtype0] PASSED [  3%]
tests/compute/test_batched_heterograph.py::test_batch_keeps_empty_data[idtype1] PASSED [  3%]
tests/compute/test_batched_heterograph.py::test_batch_netypes SKIPPED    [  3%]
tests/compute/test_csrmm.py::test_csrmm[dtype0-idtype0] PASSED           [  3%]
tests/compute/test_csrmm.py::test_csrmm[dtype0-idtype1] PASSED           [  3%]
tests/compute/test_csrmm.py::test_csrmm[dtype1-idtype0] PASSED           [  3%]
tests/compute/test_csrmm.py::test_csrmm[dtype1-idtype1] PASSED           [  3%]
tests/compute/test_csrmm.py::test_csrmm_backward[1-dtype0-idtype0] PASSED [  3%]
tests/compute/test_csrmm.py::test_csrmm_backward[1-dtype0-idtype1] PASSED [  3%]
tests/compute/test_csrmm.py::test_csrmm_backward[1-dtype1-idtype0] PASSED [  3%]
tests/compute/test_csrmm.py::test_csrmm_backward[1-dtype1-idtype1] PASSED [  3%]
tests/compute/test_csrmm.py::test_csrmm_backward[2-dtype0-idtype0] PASSED [  3%]
tests/compute/test_csrmm.py::test_csrmm_backward[2-dtype0-idtype1] PASSED [  3%]
tests/compute/test_csrmm.py::test_csrmm_backward[2-dtype1-idtype0] PASSED [  3%]
tests/compute/test_csrmm.py::test_csrmm_backward[2-dtype1-idtype1] PASSED [  3%]
tests/compute/test_csrmm.py::test_csrsum[dtype0-idtype0] PASSED          [  3%]
tests/compute/test_csrmm.py::test_csrsum[dtype0-idtype1] PASSED          [  3%]
tests/compute/test_csrmm.py::test_csrsum[dtype1-idtype0] PASSED          [  3%]
tests/compute/test_csrmm.py::test_csrsum[dtype1-idtype1] PASSED          [  3%]
tests/compute/test_csrmm.py::test_csrsum_backward[1-dtype0-idtype0] PASSED [  3%]
tests/compute/test_csrmm.py::test_csrsum_backward[1-dtype0-idtype1] PASSED [  3%]
tests/compute/test_csrmm.py::test_csrsum_backward[1-dtype1-idtype0] PASSED [  3%]
tests/compute/test_csrmm.py::test_csrsum_backward[1-dtype1-idtype1] PASSED [  3%]
tests/compute/test_csrmm.py::test_csrsum_backward[2-dtype0-idtype0] PASSED [  3%]
tests/compute/test_csrmm.py::test_csrsum_backward[2-dtype0-idtype1] PASSED [  3%]
tests/compute/test_csrmm.py::test_csrsum_backward[2-dtype1-idtype0] PASSED [  4%]
tests/compute/test_csrmm.py::test_csrsum_backward[2-dtype1-idtype1] PASSED [  4%]
tests/compute/test_csrmm.py::test_csrmask[9000-9000-dtype0-idtype0] PASSED [  4%]
tests/compute/test_csrmm.py::test_csrmask[9000-9000-dtype0-idtype1] PASSED [  4%]
tests/compute/test_csrmm.py::test_csrmask[9000-9000-dtype1-idtype0] PASSED [  4%]
tests/compute/test_csrmm.py::test_csrmask[9000-9000-dtype1-idtype1] PASSED [  4%]
tests/compute/test_csrmm.py::test_csrmask[9000-0-dtype0-idtype0] PASSED  [  4%]
tests/compute/test_csrmm.py::test_csrmask[9000-0-dtype0-idtype1] PASSED  [  4%]
tests/compute/test_csrmm.py::test_csrmask[9000-0-dtype1-idtype0] PASSED  [  4%]
tests/compute/test_csrmm.py::test_csrmask[9000-0-dtype1-idtype1] PASSED  [  4%]
tests/compute/test_csrmm.py::test_csrmask[0-9000-dtype0-idtype0] PASSED  [  4%]
tests/compute/test_csrmm.py::test_csrmask[0-9000-dtype0-idtype1] PASSED  [  4%]
tests/compute/test_csrmm.py::test_csrmask[0-9000-dtype1-idtype0] PASSED  [  4%]
tests/compute/test_csrmm.py::test_csrmask[0-9000-dtype1-idtype1] PASSED  [  4%]
tests/compute/test_csrmm.py::test_csrmask[0-0-dtype0-idtype0] PASSED     [  4%]
tests/compute/test_csrmm.py::test_csrmask[0-0-dtype0-idtype1] PASSED     [  4%]
tests/compute/test_csrmm.py::test_csrmask[0-0-dtype1-idtype0] PASSED     [  4%]
tests/compute/test_csrmm.py::test_csrmask[0-0-dtype1-idtype1] PASSED     [  4%]
tests/compute/test_csrmm.py::test_csrmask_backward[dtype0-idtype0] PASSED [  4%]
tests/compute/test_csrmm.py::test_csrmask_backward[dtype0-idtype1] PASSED [  4%]
tests/compute/test_csrmm.py::test_csrmask_backward[dtype1-idtype0] PASSED [  4%]
tests/compute/test_csrmm.py::test_csrmask_backward[dtype1-idtype1] PASSED [  4%]
tests/compute/test_data.py::test_minigc SKIPPED (Datasets don't need...) [  4%]
tests/compute/test_data.py::test_gin SKIPPED (Datasets don't need to...) [  4%]
tests/compute/test_data.py::test_fraud SKIPPED (Datasets don't need ...) [  5%]
tests/compute/test_data.py::test_fakenews SKIPPED (Datasets don't ne...) [  5%]
tests/compute/test_data.py::test_tudataset_regression SKIPPED (Datas...) [  5%]
tests/compute/test_data.py::test_data_hash SKIPPED (Datasets don't n...) [  5%]
tests/compute/test_data.py::test_citation_graph SKIPPED (Datasets do...) [  5%]
tests/compute/test_data.py::test_gnn_benchmark SKIPPED (Datasets don...) [  5%]
tests/compute/test_data.py::test_reddit SKIPPED (Datasets don't need...) [  5%]
tests/compute/test_data.py::test_explain_syn SKIPPED (Datasets don't...) [  5%]
tests/compute/test_data.py::test_wiki_cs SKIPPED (Datasets don't nee...) [  5%]
tests/compute/test_data.py::test_yelp SKIPPED (Dataset too large to ...) [  5%]
tests/compute/test_data.py::test_flickr SKIPPED (Datasets don't need...) [  5%]
tests/compute/test_data.py::test_extract_archive SKIPPED (Datasets d...) [  5%]
tests/compute/test_data.py::test_csvdataset SKIPPED (Datasets don't ...) [  5%]
tests/compute/test_data.py::test_add_nodepred_split SKIPPED (Dataset...) [  5%]
tests/compute/test_data.py::test_as_nodepred1 SKIPPED (Datasets don'...) [  5%]
tests/compute/test_data.py::test_as_nodepred2 SKIPPED (Datasets don'...) [  5%]
tests/compute/test_data.py::test_as_nodepred_ogb SKIPPED (ogb only s...) [  5%]
tests/compute/test_data.py::test_as_linkpred SKIPPED (Datasets don't...) [  5%]
tests/compute/test_data.py::test_as_linkpred_ogb SKIPPED (ogb only s...) [  5%]
tests/compute/test_data.py::test_as_nodepred_csvdataset SKIPPED (Dat...) [  5%]
tests/compute/test_data.py::test_as_graphpred SKIPPED (Datasets don'...) [  5%]
tests/compute/test_data.py::test_as_graphpred_reprocess SKIPPED (Dat...) [  5%]
tests/compute/test_data.py::test_as_graphpred_ogb SKIPPED (ogb only ...) [  5%]
tests/compute/test_dataloader.py::test_edge_prediction_sampler[idtype0] PASSED [  5%]
tests/compute/test_dataloader.py::test_edge_prediction_sampler[idtype1] PASSED [  5%]
tests/compute/test_edge_softmax_hetero.py::test_edge_softmax_unidirectional SKIPPED [  6%]
tests/compute/test_edge_softmax_hetero.py::test_edge_softmax[idtype0-src-g0] SKIPPED [  6%]
tests/compute/test_edge_softmax_hetero.py::test_edge_softmax[idtype0-dst-g0] SKIPPED [  6%]
tests/compute/test_edge_softmax_hetero.py::test_edge_softmax[idtype1-src-g0] SKIPPED [  6%]
tests/compute/test_edge_softmax_hetero.py::test_edge_softmax[idtype1-dst-g0] SKIPPED [  6%]
tests/compute/test_ffi.py::test_cython PASSED                            [  6%]
tests/compute/test_ffi.py::test_callback[1] PASSED                       [  6%]
tests/compute/test_ffi.py::test_callback[2.3] PASSED                     [  6%]
tests/compute/test_ffi.py::test_callback_thread[1] PASSED                [  6%]
tests/compute/test_ffi.py::test_callback_thread[2.3] PASSED              [  6%]
tests/compute/test_filter.py::test_graph_filter PASSED                   [  6%]
tests/compute/test_filter.py::test_array_filter[idtype0] PASSED          [  6%]
tests/compute/test_filter.py::test_array_filter[idtype1] PASSED          [  6%]
tests/compute/test_frame.py::test_column_subcolumn PASSED                [  6%]
tests/compute/test_frame.py::test_serialize_deserialize_plain PASSED     [  6%]
tests/compute/test_frame.py::test_serialize_deserialize_subcolumn PASSED [  6%]
tests/compute/test_frame.py::test_serialize_deserialize_dtype PASSED     [  6%]
tests/compute/test_generators.py::test_rand_graph SKIPPED (GPU rando...) [  6%]
tests/compute/test_graph.py::test_query PASSED                           [  6%]
tests/compute/test_graph.py::test_mutation PASSED                        [  6%]
tests/compute/test_graph.py::test_scipy_adjmat PASSED                    [  6%]
tests/compute/test_graph.py::test_incmat PASSED                          [  6%]
tests/compute/test_graph.py::test_find_edges PASSED                      [  6%]
tests/compute/test_graph.py::test_ismultigraph PASSED                    [  6%]
tests/compute/test_graph.py::test_hypersparse_query PASSED               [  6%]
tests/compute/test_graph.py::test_empty_data_initialized PASSED          [  7%]
tests/compute/test_graph.py::test_is_sorted PASSED                       [  7%]
tests/compute/test_graph.py::test_default_types PASSED                   [  7%]
tests/compute/test_graph.py::test_formats PASSED                         [  7%]
tests/compute/test_heterograph.py::test_create[idtype0] PASSED           [  7%]
tests/compute/test_heterograph.py::test_create[idtype1] PASSED           [  7%]
tests/compute/test_heterograph.py::test_create2 PASSED                   [  7%]
tests/compute/test_heterograph.py::test_query[idtype0] PASSED            [  7%]
tests/compute/test_heterograph.py::test_query[idtype1] PASSED            [  7%]
tests/compute/test_heterograph.py::test_empty_query[idtype0] PASSED      [  7%]
tests/compute/test_heterograph.py::test_empty_query[idtype1] PASSED      [  7%]
tests/compute/test_heterograph.py::test_adj[idtype0] PASSED              [  7%]
tests/compute/test_heterograph.py::test_adj[idtype1] PASSED              [  7%]
tests/compute/test_heterograph.py::test_inc[idtype0] PASSED              [  7%]
tests/compute/test_heterograph.py::test_inc[idtype1] PASSED              [  7%]
tests/compute/test_heterograph.py::test_view[idtype0] PASSED             [  7%]
tests/compute/test_heterograph.py::test_view[idtype1] PASSED             [  7%]
tests/compute/test_heterograph.py::test_view1[idtype0] PASSED            [  7%]
tests/compute/test_heterograph.py::test_view1[idtype1] PASSED            [  7%]
tests/compute/test_heterograph.py::test_flatten[idtype0] PASSED          [  7%]
tests/compute/test_heterograph.py::test_flatten[idtype1] PASSED          [  7%]
tests/compute/test_heterograph.py::test_to_device[idtype0] PASSED        [  7%]
tests/compute/test_heterograph.py::test_to_device[idtype1] PASSED        [  7%]
tests/compute/test_heterograph.py::test_to_device2[g0-idtype0] PASSED    [  7%]
tests/compute/test_heterograph.py::test_to_device2[g0-idtype1] PASSED    [  8%]
tests/compute/test_heterograph.py::test_to_device2[g1-idtype0] PASSED    [  8%]
tests/compute/test_heterograph.py::test_to_device2[g1-idtype1] PASSED    [  8%]
tests/compute/test_heterograph.py::test_pin_memory_[idtype0] SKIPPED     [  8%]
tests/compute/test_heterograph.py::test_pin_memory_[idtype1] SKIPPED     [  8%]
tests/compute/test_heterograph.py::test_convert_bound[idtype0] PASSED    [  8%]
tests/compute/test_heterograph.py::test_convert_bound[idtype1] PASSED    [  8%]
tests/compute/test_heterograph.py::test_convert[idtype0] PASSED          [  8%]
tests/compute/test_heterograph.py::test_convert[idtype1] PASSED          [  8%]
tests/compute/test_heterograph.py::test_to_homo_zero_nodes[idtype0] SKIPPED [  8%]
tests/compute/test_heterograph.py::test_to_homo_zero_nodes[idtype1] SKIPPED [  8%]
tests/compute/test_heterograph.py::test_to_homo2[idtype0] PASSED         [  8%]
tests/compute/test_heterograph.py::test_to_homo2[idtype1] PASSED         [  8%]
tests/compute/test_heterograph.py::test_invertible_conversion[idtype0] PASSED [  8%]
tests/compute/test_heterograph.py::test_invertible_conversion[idtype1] PASSED [  8%]
tests/compute/test_heterograph.py::test_metagraph_reachable[idtype0] PASSED [  8%]
tests/compute/test_heterograph.py::test_metagraph_reachable[idtype1] PASSED [  8%]
tests/compute/test_heterograph.py::test_subgraph_mask[idtype0] PASSED    [  8%]
tests/compute/test_heterograph.py::test_subgraph_mask[idtype1] PASSED    [  8%]
tests/compute/test_heterograph.py::test_subgraph[idtype0] PASSED         [  8%]
tests/compute/test_heterograph.py::test_subgraph[idtype1] PASSED         [  8%]
tests/compute/test_heterograph.py::test_apply[idtype0] PASSED            [  8%]
tests/compute/test_heterograph.py::test_apply[idtype1] PASSED            [  8%]
tests/compute/test_heterograph.py::test_level2[idtype0] PASSED           [  8%]
tests/compute/test_heterograph.py::test_level2[idtype1] PASSED           [  8%]
tests/compute/test_heterograph.py::test_more_nnz[idtype0] PASSED         [  9%]
tests/compute/test_heterograph.py::test_more_nnz[idtype1] PASSED         [  9%]
tests/compute/test_heterograph.py::test_updates[idtype0] PASSED          [  9%]
tests/compute/test_heterograph.py::test_updates[idtype1] PASSED          [  9%]
tests/compute/test_heterograph.py::test_backward[idtype0] PASSED         [  9%]
tests/compute/test_heterograph.py::test_backward[idtype1] PASSED         [  9%]
tests/compute/test_heterograph.py::test_empty_heterograph[idtype0] PASSED [  9%]
tests/compute/test_heterograph.py::test_empty_heterograph[idtype1] PASSED [  9%]
tests/compute/test_heterograph.py::test_types_in_function[idtype0] PASSED [  9%]
tests/compute/test_heterograph.py::test_types_in_function[idtype1] PASSED [  9%]
tests/compute/test_heterograph.py::test_stack_reduce[idtype0] PASSED     [  9%]
tests/compute/test_heterograph.py::test_stack_reduce[idtype1] PASSED     [  9%]
tests/compute/test_heterograph.py::test_isolated_ntype[idtype0] PASSED   [  9%]
tests/compute/test_heterograph.py::test_isolated_ntype[idtype1] PASSED   [  9%]
tests/compute/test_heterograph.py::test_ismultigraph[idtype0] PASSED     [  9%]
tests/compute/test_heterograph.py::test_ismultigraph[idtype1] PASSED     [  9%]
tests/compute/test_heterograph.py::test_graph_index_is_unibipartite[idtype0] PASSED [  9%]
tests/compute/test_heterograph.py::test_graph_index_is_unibipartite[idtype1] PASSED [  9%]
tests/compute/test_heterograph.py::test_bipartite[idtype0] PASSED        [  9%]
tests/compute/test_heterograph.py::test_bipartite[idtype1] PASSED        [  9%]
tests/compute/test_heterograph.py::test_dtype_cast[idtype0] PASSED       [  9%]
tests/compute/test_heterograph.py::test_dtype_cast[idtype1] PASSED       [  9%]
tests/compute/test_heterograph.py::test_float_cast PASSED                [  9%]
tests/compute/test_heterograph.py::test_format[idtype0] PASSED           [  9%]
tests/compute/test_heterograph.py::test_format[idtype1] PASSED           [ 10%]
tests/compute/test_heterograph.py::test_edges_order[idtype0] PASSED      [ 10%]
tests/compute/test_heterograph.py::test_edges_order[idtype1] PASSED      [ 10%]
tests/compute/test_heterograph.py::test_reverse[idtype0] PASSED          [ 10%]
tests/compute/test_heterograph.py::test_reverse[idtype1] PASSED          [ 10%]
tests/compute/test_heterograph.py::test_clone[idtype0] PASSED            [ 10%]
tests/compute/test_heterograph.py::test_clone[idtype1] PASSED            [ 10%]
tests/compute/test_heterograph.py::test_add_edges[idtype0] PASSED        [ 10%]
tests/compute/test_heterograph.py::test_add_edges[idtype1] PASSED        [ 10%]
tests/compute/test_heterograph.py::test_add_nodes[idtype0] PASSED        [ 10%]
tests/compute/test_heterograph.py::test_add_nodes[idtype1] PASSED        [ 10%]
tests/compute/test_heterograph.py::test_remove_edges[idtype0] PASSED     [ 10%]
tests/compute/test_heterograph.py::test_remove_edges[idtype1] PASSED     [ 10%]
tests/compute/test_heterograph.py::test_remove_nodes[idtype0] PASSED     [ 10%]
tests/compute/test_heterograph.py::test_remove_nodes[idtype1] PASSED     [ 10%]
tests/compute/test_heterograph.py::test_frame[idtype0] PASSED            [ 10%]
tests/compute/test_heterograph.py::test_frame[idtype1] PASSED            [ 10%]
tests/compute/test_heterograph.py::test_frame_device[idtype0] SKIPPED    [ 10%]
tests/compute/test_heterograph.py::test_frame_device[idtype1] SKIPPED    [ 10%]
tests/compute/test_heterograph.py::test_create_block[idtype0] PASSED     [ 10%]
tests/compute/test_heterograph.py::test_create_block[idtype1] PASSED     [ 10%]
tests/compute/test_heterograph.py::test_adj_sparse[coo-idtype0] PASSED   [ 10%]
tests/compute/test_heterograph.py::test_adj_sparse[coo-idtype1] PASSED   [ 10%]
tests/compute/test_heterograph.py::test_adj_sparse[csr-idtype0] PASSED   [ 10%]
tests/compute/test_heterograph.py::test_adj_sparse[csr-idtype1] PASSED   [ 10%]
tests/compute/test_heterograph.py::test_adj_sparse[csc-idtype0] PASSED   [ 11%]
tests/compute/test_heterograph.py::test_adj_sparse[csc-idtype1] PASSED   [ 11%]
tests/compute/test_heterograph.py::test_forking_pickler PASSED           [ 11%]
tests/compute/test_index.py::test_dlpack SKIPPED (TF doesn't support...) [ 11%]
tests/compute/test_kernel.py::test_copy_src_reduce PASSED                [ 11%]
tests/compute/test_kernel.py::test_copy_edge_reduce PASSED               [ 11%]
tests/compute/test_kernel.py::test_all_binary_builtins PASSED            [ 11%]
tests/compute/test_kernel.py::test_mean_zero_degree[g0-idtype0] PASSED   [ 11%]
tests/compute/test_kernel.py::test_mean_zero_degree[g0-idtype1] PASSED   [ 11%]
tests/compute/test_merge.py::test_heterograph_merge[idtype0] PASSED      [ 11%]
tests/compute/test_merge.py::test_heterograph_merge[idtype1] PASSED      [ 11%]
tests/compute/test_nccl.py::test_nccl_id PASSED                          [ 11%]
tests/compute/test_nccl.py::test_nccl_sparse_push_single_remainder PASSED [ 11%]
tests/compute/test_nccl.py::test_nccl_sparse_pull_single_remainder PASSED [ 11%]
tests/compute/test_nccl.py::test_nccl_sparse_push_single_range PASSED    [ 11%]
tests/compute/test_nccl.py::test_nccl_sparse_pull_single_range PASSED    [ 11%]
tests/compute/test_nccl.py::test_nccl_support PASSED                     [ 11%]
tests/compute/test_new_update_all_hetero.py::test_unary_copy_u[idtype0] SKIPPED [ 11%]
tests/compute/test_new_update_all_hetero.py::test_unary_copy_u[idtype1] SKIPPED [ 11%]
tests/compute/test_new_update_all_hetero.py::test_unary_copy_e[idtype0] SKIPPED [ 11%]
tests/compute/test_new_update_all_hetero.py::test_unary_copy_e[idtype1] SKIPPED [ 11%]
tests/compute/test_new_update_all_hetero.py::test_binary_op[idtype0] SKIPPED [ 11%]
tests/compute/test_new_update_all_hetero.py::test_binary_op[idtype1] SKIPPED [ 11%]
tests/compute/test_partition.py::test_get_node_partition_from_book[idtype0] PASSED [ 11%]
tests/compute/test_partition.py::test_get_node_partition_from_book[idtype1] PASSED [ 12%]
tests/compute/test_pickle.py::test_pickling_index PASSED                 [ 12%]
tests/compute/test_pickle.py::test_pickling_graph_index PASSED           [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g0-idtype0] SKIPPED    [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g0-idtype1] SKIPPED    [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g1-idtype0] SKIPPED    [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g1-idtype1] SKIPPED    [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g2-idtype0] SKIPPED    [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g2-idtype1] SKIPPED    [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g3-idtype0] SKIPPED    [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g3-idtype1] SKIPPED    [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g4-idtype0] SKIPPED    [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g4-idtype1] SKIPPED    [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g5-idtype0] SKIPPED    [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g5-idtype1] SKIPPED    [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g6-idtype0] SKIPPED    [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g6-idtype1] SKIPPED    [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g7-idtype0] SKIPPED    [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g7-idtype1] SKIPPED    [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g8-idtype0] SKIPPED    [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g8-idtype1] SKIPPED    [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g9-idtype0] SKIPPED    [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g9-idtype1] SKIPPED    [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g10-idtype0] SKIPPED   [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g10-idtype1] SKIPPED   [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g11-idtype0] SKIPPED   [ 13%]
tests/compute/test_pickle.py::test_pickling_graph[g11-idtype1] SKIPPED   [ 13%]
tests/compute/test_pickle.py::test_pickling_graph[g12-idtype0] SKIPPED   [ 13%]
tests/compute/test_pickle.py::test_pickling_graph[g12-idtype1] SKIPPED   [ 13%]
tests/compute/test_pickle.py::test_pickling_batched_heterograph SKIPPED  [ 13%]
tests/compute/test_pickle.py::test_pickling_subgraph SKIPPED (GPU ed...) [ 13%]
tests/compute/test_pickle.py::test_pickling_is_pinned[idtype0] SKIPPED   [ 13%]
tests/compute/test_pickle.py::test_pickling_is_pinned[idtype1] SKIPPED   [ 13%]
tests/compute/test_pin_memory.py::test_pin_unpin PASSED                  [ 13%]
tests/compute/test_propagate.py::test_prop_nodes_bfs[idtype0] SKIPPED    [ 13%]
tests/compute/test_propagate.py::test_prop_nodes_bfs[idtype1] SKIPPED    [ 13%]
tests/compute/test_propagate.py::test_prop_edges_dfs[idtype0] SKIPPED    [ 13%]
tests/compute/test_propagate.py::test_prop_edges_dfs[idtype1] SKIPPED    [ 13%]
tests/compute/test_propagate.py::test_prop_nodes_topo[idtype0] SKIPPED   [ 13%]
tests/compute/test_propagate.py::test_prop_nodes_topo[idtype1] SKIPPED   [ 13%]
tests/compute/test_random.py::test_random_choice SKIPPED (GPU random...) [ 13%]
tests/compute/test_readout.py::test_sum_case1[idtype0] PASSED            [ 13%]
tests/compute/test_readout.py::test_sum_case1[idtype1] PASSED            [ 13%]
tests/compute/test_readout.py::test_reduce_readout[sum-g0-idtype0] PASSED [ 13%]
tests/compute/test_readout.py::test_reduce_readout[sum-g0-idtype1] PASSED [ 13%]
tests/compute/test_readout.py::test_reduce_readout[sum-g1-idtype0] PASSED [ 13%]
tests/compute/test_readout.py::test_reduce_readout[sum-g1-idtype1] PASSED [ 13%]
tests/compute/test_readout.py::test_reduce_readout[sum-g2-idtype0] PASSED [ 13%]
tests/compute/test_readout.py::test_reduce_readout[sum-g2-idtype1] PASSED [ 13%]
tests/compute/test_readout.py::test_reduce_readout[sum-g3-idtype0] PASSED [ 13%]
tests/compute/test_readout.py::test_reduce_readout[sum-g3-idtype1] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[sum-g4-idtype0] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[sum-g4-idtype1] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[sum-g5-idtype0] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[sum-g5-idtype1] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[sum-g6-idtype0] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[sum-g6-idtype1] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g0-idtype0] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g0-idtype1] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g1-idtype0] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g1-idtype1] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g2-idtype0] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g2-idtype1] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g3-idtype0] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g3-idtype1] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g4-idtype0] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g4-idtype1] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g5-idtype0] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g5-idtype1] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g6-idtype0] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g6-idtype1] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[mean-g0-idtype0] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[mean-g0-idtype1] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[mean-g1-idtype0] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[mean-g1-idtype1] PASSED [ 15%]
tests/compute/test_readout.py::test_reduce_readout[mean-g2-idtype0] PASSED [ 15%]
tests/compute/test_readout.py::test_reduce_readout[mean-g2-idtype1] PASSED [ 15%]
tests/compute/test_readout.py::test_reduce_readout[mean-g3-idtype0] PASSED [ 15%]
tests/compute/test_readout.py::test_reduce_readout[mean-g3-idtype1] PASSED [ 15%]
tests/compute/test_readout.py::test_reduce_readout[mean-g4-idtype0] PASSED [ 15%]
tests/compute/test_readout.py::test_reduce_readout[mean-g4-idtype1] PASSED [ 15%]
tests/compute/test_readout.py::test_reduce_readout[mean-g5-idtype0] PASSED [ 15%]
tests/compute/test_readout.py::test_reduce_readout[mean-g5-idtype1] PASSED [ 15%]
tests/compute/test_readout.py::test_reduce_readout[mean-g6-idtype0] PASSED [ 15%]
tests/compute/test_readout.py::test_reduce_readout[mean-g6-idtype1] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g0-idtype0] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g0-idtype1] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g1-idtype0] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g1-idtype1] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g2-idtype0] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g2-idtype1] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g3-idtype0] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g3-idtype1] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g4-idtype0] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g4-idtype1] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g5-idtype0] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g5-idtype1] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g6-idtype0] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g6-idtype1] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g0-idtype0] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g0-idtype1] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g1-idtype0] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g1-idtype1] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g2-idtype0] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g2-idtype1] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g3-idtype0] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g3-idtype1] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g4-idtype0] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g4-idtype1] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g5-idtype0] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g5-idtype1] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g6-idtype0] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g6-idtype1] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g0-idtype0] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g0-idtype1] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g1-idtype0] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g1-idtype1] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g2-idtype0] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g2-idtype1] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g3-idtype0] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g3-idtype1] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g4-idtype0] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g4-idtype1] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g5-idtype0] PASSED [ 17%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g5-idtype1] PASSED [ 17%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g6-idtype0] PASSED [ 17%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g6-idtype1] PASSED [ 17%]
tests/compute/test_readout.py::test_topk[True-g0-idtype0] PASSED         [ 17%]
tests/compute/test_readout.py::test_topk[True-g0-idtype1] PASSED         [ 17%]
tests/compute/test_readout.py::test_topk[True-g1-idtype0] PASSED         [ 17%]
tests/compute/test_readout.py::test_topk[True-g1-idtype1] PASSED         [ 17%]
tests/compute/test_readout.py::test_topk[True-g2-idtype0] PASSED         [ 17%]
tests/compute/test_readout.py::test_topk[True-g2-idtype1] PASSED         [ 17%]
tests/compute/test_readout.py::test_topk[True-g3-idtype0] PASSED         [ 17%]
tests/compute/test_readout.py::test_topk[True-g3-idtype1] PASSED         [ 17%]
tests/compute/test_readout.py::test_topk[True-g4-idtype0] PASSED         [ 17%]
tests/compute/test_readout.py::test_topk[True-g4-idtype1] PASSED         [ 17%]
tests/compute/test_readout.py::test_topk[True-g5-idtype0] PASSED         [ 17%]
tests/compute/test_readout.py::test_topk[True-g5-idtype1] PASSED         [ 17%]
tests/compute/test_readout.py::test_topk[True-g6-idtype0] PASSED         [ 17%]
tests/compute/test_readout.py::test_topk[True-g6-idtype1] PASSED         [ 17%]
tests/compute/test_readout.py::test_topk[False-g0-idtype0] PASSED        [ 17%]
tests/compute/test_readout.py::test_topk[False-g0-idtype1] PASSED        [ 17%]
tests/compute/test_readout.py::test_topk[False-g1-idtype0] PASSED        [ 17%]
tests/compute/test_readout.py::test_topk[False-g1-idtype1] PASSED        [ 17%]
tests/compute/test_readout.py::test_topk[False-g2-idtype0] PASSED        [ 17%]
tests/compute/test_readout.py::test_topk[False-g2-idtype1] PASSED        [ 17%]
tests/compute/test_readout.py::test_topk[False-g3-idtype0] PASSED        [ 17%]
tests/compute/test_readout.py::test_topk[False-g3-idtype1] PASSED        [ 18%]
tests/compute/test_readout.py::test_topk[False-g4-idtype0] PASSED        [ 18%]
tests/compute/test_readout.py::test_topk[False-g4-idtype1] PASSED        [ 18%]
tests/compute/test_readout.py::test_topk[False-g5-idtype0] PASSED        [ 18%]
tests/compute/test_readout.py::test_topk[False-g5-idtype1] PASSED        [ 18%]
tests/compute/test_readout.py::test_topk[False-g6-idtype0] PASSED        [ 18%]
tests/compute/test_readout.py::test_topk[False-g6-idtype1] PASSED        [ 18%]
tests/compute/test_readout.py::test_softmax[g0-idtype0] PASSED           [ 18%]
tests/compute/test_readout.py::test_softmax[g0-idtype1] PASSED           [ 18%]
tests/compute/test_readout.py::test_softmax[g1-idtype0] PASSED           [ 18%]
tests/compute/test_readout.py::test_softmax[g1-idtype1] PASSED           [ 18%]
tests/compute/test_readout.py::test_softmax[g2-idtype0] PASSED           [ 18%]
tests/compute/test_readout.py::test_softmax[g2-idtype1] PASSED           [ 18%]
tests/compute/test_readout.py::test_softmax[g3-idtype0] PASSED           [ 18%]
tests/compute/test_readout.py::test_softmax[g3-idtype1] PASSED           [ 18%]
tests/compute/test_readout.py::test_softmax[g4-idtype0] PASSED           [ 18%]
tests/compute/test_readout.py::test_softmax[g4-idtype1] PASSED           [ 18%]
tests/compute/test_readout.py::test_softmax[g5-idtype0] PASSED           [ 18%]
tests/compute/test_readout.py::test_softmax[g5-idtype1] PASSED           [ 18%]
tests/compute/test_readout.py::test_softmax[g6-idtype0] PASSED           [ 18%]
tests/compute/test_readout.py::test_softmax[g6-idtype1] PASSED           [ 18%]
tests/compute/test_readout.py::test_broadcast[g0-idtype0] PASSED         [ 18%]
tests/compute/test_readout.py::test_broadcast[g0-idtype1] PASSED         [ 18%]
tests/compute/test_readout.py::test_broadcast[g1-idtype0] PASSED         [ 18%]
tests/compute/test_readout.py::test_broadcast[g1-idtype1] PASSED         [ 18%]
tests/compute/test_readout.py::test_broadcast[g2-idtype0] PASSED         [ 19%]
tests/compute/test_readout.py::test_broadcast[g2-idtype1] PASSED         [ 19%]
tests/compute/test_readout.py::test_broadcast[g3-idtype0] PASSED         [ 19%]
tests/compute/test_readout.py::test_broadcast[g3-idtype1] PASSED         [ 19%]
tests/compute/test_readout.py::test_broadcast[g4-idtype0] PASSED         [ 19%]
tests/compute/test_readout.py::test_broadcast[g4-idtype1] PASSED         [ 19%]
tests/compute/test_readout.py::test_broadcast[g5-idtype0] PASSED         [ 19%]
tests/compute/test_readout.py::test_broadcast[g5-idtype1] PASSED         [ 19%]
tests/compute/test_readout.py::test_broadcast[g6-idtype0] PASSED         [ 19%]
tests/compute/test_readout.py::test_broadcast[g6-idtype1] PASSED         [ 19%]
tests/compute/test_removal.py::test_node_removal[idtype0] PASSED         [ 19%]
tests/compute/test_removal.py::test_node_removal[idtype1] PASSED         [ 19%]
tests/compute/test_removal.py::test_multigraph_node_removal[idtype0] PASSED [ 19%]
tests/compute/test_removal.py::test_multigraph_node_removal[idtype1] PASSED [ 19%]
tests/compute/test_removal.py::test_multigraph_edge_removal[idtype0] PASSED [ 19%]
tests/compute/test_removal.py::test_multigraph_edge_removal[idtype1] PASSED [ 19%]
tests/compute/test_removal.py::test_edge_removal[idtype0] PASSED         [ 19%]
tests/compute/test_removal.py::test_edge_removal[idtype1] PASSED         [ 19%]
tests/compute/test_removal.py::test_node_and_edge_removal[idtype0] PASSED [ 19%]
tests/compute/test_removal.py::test_node_and_edge_removal[idtype1] PASSED [ 19%]
tests/compute/test_removal.py::test_node_frame[idtype0] PASSED           [ 19%]
tests/compute/test_removal.py::test_node_frame[idtype1] PASSED           [ 19%]
tests/compute/test_removal.py::test_edge_frame[idtype0] PASSED           [ 19%]
tests/compute/test_removal.py::test_edge_frame[idtype1] PASSED           [ 19%]
tests/compute/test_removal.py::test_issue1287[idtype0] PASSED            [ 20%]
tests/compute/test_removal.py::test_issue1287[idtype1] PASSED            [ 20%]
tests/compute/test_sampler.py::test_create_full PASSED                   [ 20%]
tests/compute/test_sampler.py::test_1neighbor_sampler_all PASSED         [ 20%]
tests/compute/test_sampler.py::test_1neighbor_sampler PASSED             [ 20%]
tests/compute/test_sampler.py::test_prefetch_neighbor_sampler PASSED     [ 20%]
tests/compute/test_sampler.py::test_10neighbor_sampler_all PASSED        [ 20%]
tests/compute/test_sampler.py::test_10neighbor_sampler PASSED            [ 20%]
tests/compute/test_sampler.py::test_layer_sampler PASSED                 [ 20%]
tests/compute/test_sampler.py::test_nonuniform_neighbor_sampler SKIPPED  [ 20%]
tests/compute/test_sampler.py::test_setseed PASSED                       [ 20%]
tests/compute/test_sampler.py::test_negative_sampler SKIPPED (TF doe...) [ 20%]
tests/compute/test_sampling.py::test_non_uniform_random_walk[True] SKIPPED [ 20%]
tests/compute/test_sampling.py::test_non_uniform_random_walk[False] PASSED [ 20%]
tests/compute/test_sampling.py::test_uniform_random_walk[True] PASSED    [ 20%]
tests/compute/test_sampling.py::test_uniform_random_walk[False] PASSED   [ 20%]
tests/compute/test_sampling.py::test_node2vec SKIPPED (GPU random wa...) [ 20%]
tests/compute/test_sampling.py::test_pack_traces SKIPPED (GPU pack t...) [ 20%]
tests/compute/test_sampling.py::test_pinsage_sampling[True] PASSED       [ 20%]
tests/compute/test_sampling.py::test_pinsage_sampling[False] PASSED      [ 20%]
tests/compute/test_sampling.py::test_sample_neighbors_noprob PASSED      [ 20%]
tests/compute/test_sampling.py::test_sample_neighbors_prob PASSED        [ 20%]
tests/compute/test_sampling.py::test_sample_neighbors_outedge PASSED     [ 20%]
tests/compute/test_sampling.py::test_sample_neighbors_topk SKIPPED (...) [ 20%]
tests/compute/test_sampling.py::test_sample_neighbors_topk_outedge SKIPPED [ 20%]
tests/compute/test_sampling.py::test_sample_neighbors_with_0deg PASSED   [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_biased_homogeneous SKIPPED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_biased_bipartite SKIPPED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_homogeneous[False-in-coo] SKIPPED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_homogeneous[False-in-csr] SKIPPED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_homogeneous[False-in-csc] SKIPPED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_homogeneous[False-out-coo] SKIPPED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_homogeneous[False-out-csr] SKIPPED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_homogeneous[False-out-csc] SKIPPED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_homogeneous[True-in-coo] SKIPPED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_homogeneous[True-in-csr] SKIPPED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_homogeneous[True-in-csc] SKIPPED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_homogeneous[True-out-coo] SKIPPED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_homogeneous[True-out-csr] SKIPPED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_homogeneous[True-out-csc] SKIPPED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_sorted_homogeneous[in-csr] SKIPPED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_sorted_homogeneous[in-csc] SKIPPED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_sorted_homogeneous[out-csr] SKIPPED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_sorted_homogeneous[out-csc] SKIPPED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_exclude_edges_heteroG[int32] PASSED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_exclude_edges_heteroG[int64] PASSED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_exclude_edges_homoG[int32] PASSED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_exclude_edges_homoG[int64] PASSED [ 21%]
tests/compute/test_sampling.py::test_global_uniform_negative_sampling[int32] PASSED [ 21%]
tests/compute/test_sampling.py::test_global_uniform_negative_sampling[int64] PASSED [ 22%]
tests/compute/test_serialize.py::test_graph_serialize_with_feature[True] SKIPPED [ 22%]
tests/compute/test_serialize.py::test_graph_serialize_with_feature[False] SKIPPED [ 22%]
tests/compute/test_serialize.py::test_graph_serialize_without_feature[True] SKIPPED [ 22%]
tests/compute/test_serialize.py::test_graph_serialize_without_feature[False] SKIPPED [ 22%]
tests/compute/test_serialize.py::test_graph_serialize_with_labels[True] SKIPPED [ 22%]
tests/compute/test_serialize.py::test_graph_serialize_with_labels[False] SKIPPED [ 22%]
tests/compute/test_serialize.py::test_serialize_tensors PASSED           [ 22%]
tests/compute/test_serialize.py::test_serialize_empty_dict PASSED        [ 22%]
tests/compute/test_serialize.py::test_load_old_files1 PASSED             [ 22%]
tests/compute/test_serialize.py::test_load_old_files2 PASSED             [ 22%]
tests/compute/test_serialize.py::test_deserialize_old_heterograph_file PASSED [ 22%]
tests/compute/test_serialize.py::test_serialize_heterograph SKIPPED      [ 22%]
tests/compute/test_serialize.py::test_serialize_heterograph_s3 SKIPPED   [ 22%]
tests/compute/test_shared_mem.py::test_single_process[idtype0] SKIPPED   [ 22%]
tests/compute/test_shared_mem.py::test_single_process[idtype1] SKIPPED   [ 22%]
tests/compute/test_shared_mem.py::test_multi_process[idtype0] SKIPPED    [ 22%]
tests/compute/test_shared_mem.py::test_multi_process[idtype1] SKIPPED    [ 22%]
tests/compute/test_shared_mem.py::test_copy_from_gpu SKIPPED (Not su...) [ 22%]
tests/compute/test_sort.py::test_sort_with_tag[idtype0] SKIPPED (GPU...) [ 22%]
tests/compute/test_sort.py::test_sort_with_tag[idtype1] SKIPPED (GPU...) [ 22%]
tests/compute/test_sort.py::test_sort_with_tag_bipartite[idtype0] SKIPPED [ 22%]
tests/compute/test_sort.py::test_sort_with_tag_bipartite[idtype1] SKIPPED [ 22%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-add-shp0-g0] PASSED  [ 22%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-add-shp0-g1] PASSED  [ 22%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-add-shp1-g0] PASSED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-add-shp1-g1] PASSED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-add-shp2-g0] PASSED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-add-shp2-g1] PASSED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-add-shp3-g0] PASSED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-add-shp3-g1] PASSED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-add-shp4-g0] PASSED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-add-shp4-g1] PASSED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-add-shp5-g0] PASSED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-add-shp5-g1] PASSED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-sub-shp0-g0] PASSED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-sub-shp0-g1] PASSED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-sub-shp1-g0] PASSED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-sub-shp1-g1] PASSED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-sub-shp2-g0] PASSED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-sub-shp2-g1] PASSED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-sub-shp3-g0] PASSED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-sub-shp3-g1] PASSED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-sub-shp4-g0] PASSED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-sub-shp4-g1] PASSED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-sub-shp5-g0] PASSED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-sub-shp5-g1] PASSED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-mul-shp0-g0] PASSED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-mul-shp0-g1] PASSED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-mul-shp1-g0] PASSED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-mul-shp1-g1] PASSED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-mul-shp2-g0] PASSED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-mul-shp2-g1] PASSED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-mul-shp3-g0] PASSED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-mul-shp3-g1] PASSED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-mul-shp4-g0] PASSED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-mul-shp4-g1] PASSED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-mul-shp5-g0] PASSED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-mul-shp5-g1] PASSED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-div-shp0-g0] PASSED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-div-shp0-g1] PASSED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-div-shp1-g0] PASSED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-div-shp1-g1] PASSED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-div-shp2-g0] PASSED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-div-shp2-g1] PASSED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-div-shp3-g0] PASSED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-div-shp3-g1] PASSED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-div-shp4-g0] PASSED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-div-shp4-g1] PASSED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-div-shp5-g0] PASSED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-div-shp5-g1] PASSED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_lhs-shp0-g0] PASSED [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_lhs-shp0-g1] PASSED [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_lhs-shp1-g0] PASSED [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_lhs-shp1-g1] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_lhs-shp2-g0] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_lhs-shp2-g1] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_lhs-shp3-g0] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_lhs-shp3-g1] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_lhs-shp4-g0] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_lhs-shp4-g1] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_lhs-shp5-g0] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_lhs-shp5-g1] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_rhs-shp0-g0] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_rhs-shp0-g1] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_rhs-shp1-g0] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_rhs-shp1-g1] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_rhs-shp2-g0] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_rhs-shp2-g1] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_rhs-shp3-g0] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_rhs-shp3-g1] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_rhs-shp4-g0] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_rhs-shp4-g1] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_rhs-shp5-g0] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_rhs-shp5-g1] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-add-shp0-g0] PASSED  [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-add-shp0-g1] PASSED  [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-add-shp1-g0] PASSED  [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-add-shp1-g1] PASSED  [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-add-shp2-g0] PASSED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-add-shp2-g1] PASSED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-add-shp3-g0] PASSED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-add-shp3-g1] PASSED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-add-shp4-g0] PASSED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-add-shp4-g1] PASSED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-add-shp5-g0] PASSED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-add-shp5-g1] PASSED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-sub-shp0-g0] PASSED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-sub-shp0-g1] PASSED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-sub-shp1-g0] PASSED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-sub-shp1-g1] PASSED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-sub-shp2-g0] PASSED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-sub-shp2-g1] PASSED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-sub-shp3-g0] PASSED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-sub-shp3-g1] PASSED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-sub-shp4-g0] PASSED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-sub-shp4-g1] PASSED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-sub-shp5-g0] PASSED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-sub-shp5-g1] PASSED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-mul-shp0-g0] PASSED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-mul-shp0-g1] PASSED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-mul-shp1-g0] PASSED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-mul-shp1-g1] PASSED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-mul-shp2-g0] PASSED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-mul-shp2-g1] PASSED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-mul-shp3-g0] PASSED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-mul-shp3-g1] PASSED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-mul-shp4-g0] PASSED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-mul-shp4-g1] PASSED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-mul-shp5-g0] PASSED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-mul-shp5-g1] PASSED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-div-shp0-g0] PASSED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-div-shp0-g1] PASSED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-div-shp1-g0] PASSED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-div-shp1-g1] PASSED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-div-shp2-g0] PASSED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-div-shp2-g1] PASSED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-div-shp3-g0] PASSED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-div-shp3-g1] PASSED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-div-shp4-g0] PASSED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-div-shp4-g1] PASSED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-div-shp5-g0] PASSED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-div-shp5-g1] PASSED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_lhs-shp0-g0] PASSED [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_lhs-shp0-g1] PASSED [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_lhs-shp1-g0] PASSED [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_lhs-shp1-g1] PASSED [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_lhs-shp2-g0] PASSED [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_lhs-shp2-g1] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_lhs-shp3-g0] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_lhs-shp3-g1] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_lhs-shp4-g0] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_lhs-shp4-g1] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_lhs-shp5-g0] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_lhs-shp5-g1] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_rhs-shp0-g0] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_rhs-shp0-g1] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_rhs-shp1-g0] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_rhs-shp1-g1] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_rhs-shp2-g0] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_rhs-shp2-g1] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_rhs-shp3-g0] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_rhs-shp3-g1] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_rhs-shp4-g0] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_rhs-shp4-g1] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_rhs-shp5-g0] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_rhs-shp5-g1] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-add-shp0-g0] PASSED  [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-add-shp0-g1] PASSED  [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-add-shp1-g0] PASSED  [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-add-shp1-g1] PASSED  [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-add-shp2-g0] PASSED  [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-add-shp2-g1] PASSED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-add-shp3-g0] PASSED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-add-shp3-g1] PASSED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-add-shp4-g0] PASSED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-add-shp4-g1] PASSED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-add-shp5-g0] PASSED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-add-shp5-g1] PASSED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-sub-shp0-g0] PASSED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-sub-shp0-g1] PASSED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-sub-shp1-g0] PASSED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-sub-shp1-g1] PASSED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-sub-shp2-g0] PASSED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-sub-shp2-g1] PASSED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-sub-shp3-g0] PASSED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-sub-shp3-g1] PASSED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-sub-shp4-g0] PASSED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-sub-shp4-g1] PASSED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-sub-shp5-g0] PASSED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-sub-shp5-g1] PASSED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-mul-shp0-g0] PASSED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-mul-shp0-g1] PASSED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-mul-shp1-g0] PASSED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-mul-shp1-g1] PASSED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-mul-shp2-g0] PASSED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-mul-shp2-g1] PASSED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-mul-shp3-g0] PASSED  [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-mul-shp3-g1] PASSED  [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-mul-shp4-g0] PASSED  [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-mul-shp4-g1] PASSED  [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-mul-shp5-g0] PASSED  [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-mul-shp5-g1] PASSED  [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-div-shp0-g0] PASSED  [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-div-shp0-g1] PASSED  [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-div-shp1-g0] PASSED  [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-div-shp1-g1] PASSED  [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-div-shp2-g0] PASSED  [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-div-shp2-g1] PASSED  [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-div-shp3-g0] PASSED  [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-div-shp3-g1] PASSED  [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-div-shp4-g0] PASSED  [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-div-shp4-g1] PASSED  [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-div-shp5-g0] PASSED  [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-div-shp5-g1] PASSED  [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_lhs-shp0-g0] PASSED [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_lhs-shp0-g1] PASSED [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_lhs-shp1-g0] PASSED [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_lhs-shp1-g1] PASSED [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_lhs-shp2-g0] PASSED [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_lhs-shp2-g1] PASSED [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_lhs-shp3-g0] PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_lhs-shp3-g1] PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_lhs-shp4-g0] PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_lhs-shp4-g1] PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_lhs-shp5-g0] PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_lhs-shp5-g1] PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_rhs-shp0-g0] PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_rhs-shp0-g1] PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_rhs-shp1-g0] PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_rhs-shp1-g1] PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_rhs-shp2-g0] PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_rhs-shp2-g1] PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_rhs-shp3-g0] PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_rhs-shp3-g1] PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_rhs-shp4-g0] PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_rhs-shp4-g1] PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_rhs-shp5-g0] PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_rhs-shp5-g1] PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-add-shp0-g0] PASSED  [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-add-shp0-g1] PASSED  [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-add-shp1-g0] PASSED  [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-add-shp1-g1] PASSED  [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-add-shp2-g0] PASSED  [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-add-shp2-g1] PASSED  [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-add-shp3-g0] PASSED  [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-add-shp3-g1] PASSED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-add-shp4-g0] PASSED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-add-shp4-g1] PASSED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-add-shp5-g0] PASSED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-add-shp5-g1] PASSED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-sub-shp0-g0] PASSED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-sub-shp0-g1] PASSED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-sub-shp1-g0] PASSED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-sub-shp1-g1] PASSED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-sub-shp2-g0] PASSED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-sub-shp2-g1] PASSED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-sub-shp3-g0] PASSED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-sub-shp3-g1] PASSED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-sub-shp4-g0] PASSED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-sub-shp4-g1] PASSED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-sub-shp5-g0] PASSED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-sub-shp5-g1] PASSED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-mul-shp0-g0] PASSED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-mul-shp0-g1] PASSED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-mul-shp1-g0] PASSED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-mul-shp1-g1] PASSED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-mul-shp2-g0] PASSED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-mul-shp2-g1] PASSED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-mul-shp3-g0] PASSED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-mul-shp3-g1] PASSED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-mul-shp4-g0] PASSED  [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-mul-shp4-g1] PASSED  [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-mul-shp5-g0] PASSED  [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-mul-shp5-g1] PASSED  [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-div-shp0-g0] PASSED  [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-div-shp0-g1] PASSED  [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-div-shp1-g0] PASSED  [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-div-shp1-g1] PASSED  [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-div-shp2-g0] PASSED  [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-div-shp2-g1] PASSED  [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-div-shp3-g0] PASSED  [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-div-shp3-g1] PASSED  [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-div-shp4-g0] PASSED  [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-div-shp4-g1] PASSED  [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-div-shp5-g0] PASSED  [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-div-shp5-g1] PASSED  [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_lhs-shp0-g0] PASSED [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_lhs-shp0-g1] PASSED [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_lhs-shp1-g0] PASSED [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_lhs-shp1-g1] PASSED [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_lhs-shp2-g0] PASSED [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_lhs-shp2-g1] PASSED [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_lhs-shp3-g0] PASSED [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_lhs-shp3-g1] PASSED [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_lhs-shp4-g0] PASSED [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_lhs-shp4-g1] PASSED [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_lhs-shp5-g0] PASSED [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_lhs-shp5-g1] PASSED [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_rhs-shp0-g0] PASSED [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_rhs-shp0-g1] PASSED [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_rhs-shp1-g0] PASSED [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_rhs-shp1-g1] PASSED [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_rhs-shp2-g0] PASSED [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_rhs-shp2-g1] PASSED [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_rhs-shp3-g0] PASSED [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_rhs-shp3-g1] PASSED [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_rhs-shp4-g0] PASSED [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_rhs-shp4-g1] PASSED [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_rhs-shp5-g0] PASSED [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_rhs-shp5-g1] PASSED [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-add-shp0-g0] PASSED  [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-add-shp0-g1] PASSED  [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-add-shp1-g0] PASSED  [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-add-shp1-g1] PASSED  [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-add-shp2-g0] PASSED  [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-add-shp2-g1] PASSED  [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-add-shp3-g0] PASSED  [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-add-shp3-g1] PASSED  [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-add-shp4-g0] PASSED  [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-add-shp4-g1] PASSED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-add-shp5-g0] PASSED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-add-shp5-g1] PASSED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-sub-shp0-g0] PASSED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-sub-shp0-g1] PASSED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-sub-shp1-g0] PASSED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-sub-shp1-g1] PASSED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-sub-shp2-g0] PASSED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-sub-shp2-g1] PASSED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-sub-shp3-g0] PASSED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-sub-shp3-g1] PASSED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-sub-shp4-g0] PASSED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-sub-shp4-g1] PASSED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-sub-shp5-g0] PASSED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-sub-shp5-g1] PASSED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-mul-shp0-g0] PASSED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-mul-shp0-g1] PASSED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-mul-shp1-g0] PASSED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-mul-shp1-g1] PASSED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-mul-shp2-g0] PASSED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-mul-shp2-g1] PASSED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-mul-shp3-g0] PASSED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-mul-shp3-g1] PASSED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-mul-shp4-g0] PASSED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-mul-shp4-g1] PASSED  [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-mul-shp5-g0] PASSED  [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-mul-shp5-g1] PASSED  [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-div-shp0-g0] PASSED  [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-div-shp0-g1] PASSED  [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-div-shp1-g0] PASSED  [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-div-shp1-g1] PASSED  [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-div-shp2-g0] PASSED  [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-div-shp2-g1] PASSED  [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-div-shp3-g0] PASSED  [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-div-shp3-g1] PASSED  [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-div-shp4-g0] PASSED  [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-div-shp4-g1] PASSED  [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-div-shp5-g0] PASSED  [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-div-shp5-g1] PASSED  [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_lhs-shp0-g0] PASSED [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_lhs-shp0-g1] PASSED [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_lhs-shp1-g0] PASSED [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_lhs-shp1-g1] PASSED [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_lhs-shp2-g0] PASSED [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_lhs-shp2-g1] PASSED [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_lhs-shp3-g0] PASSED [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_lhs-shp3-g1] PASSED [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_lhs-shp4-g0] PASSED [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_lhs-shp4-g1] PASSED [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_lhs-shp5-g0] PASSED [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_lhs-shp5-g1] PASSED [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_rhs-shp0-g0] PASSED [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_rhs-shp0-g1] PASSED [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_rhs-shp1-g0] PASSED [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_rhs-shp1-g1] PASSED [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_rhs-shp2-g0] PASSED [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_rhs-shp2-g1] PASSED [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_rhs-shp3-g0] PASSED [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_rhs-shp3-g1] PASSED [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_rhs-shp4-g0] PASSED [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_rhs-shp4-g1] PASSED [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_rhs-shp5-g0] PASSED [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_rhs-shp5-g1] PASSED [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-add-shp0-g0] PASSED  [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-add-shp0-g1] PASSED  [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-add-shp1-g0] PASSED  [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-add-shp1-g1] PASSED  [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-add-shp2-g0] PASSED  [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-add-shp2-g1] PASSED  [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-add-shp3-g0] PASSED  [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-add-shp3-g1] PASSED  [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-add-shp4-g0] PASSED  [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-add-shp4-g1] PASSED  [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-add-shp5-g0] PASSED  [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-add-shp5-g1] PASSED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-sub-shp0-g0] PASSED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-sub-shp0-g1] PASSED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-sub-shp1-g0] PASSED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-sub-shp1-g1] PASSED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-sub-shp2-g0] PASSED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-sub-shp2-g1] PASSED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-sub-shp3-g0] PASSED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-sub-shp3-g1] PASSED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-sub-shp4-g0] PASSED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-sub-shp4-g1] PASSED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-sub-shp5-g0] PASSED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-sub-shp5-g1] PASSED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-mul-shp0-g0] PASSED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-mul-shp0-g1] PASSED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-mul-shp1-g0] PASSED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-mul-shp1-g1] PASSED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-mul-shp2-g0] PASSED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-mul-shp2-g1] PASSED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-mul-shp3-g0] PASSED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-mul-shp3-g1] PASSED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-mul-shp4-g0] PASSED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-mul-shp4-g1] PASSED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-mul-shp5-g0] PASSED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-mul-shp5-g1] PASSED  [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-div-shp0-g0] PASSED  [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-div-shp0-g1] PASSED  [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-div-shp1-g0] PASSED  [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-div-shp1-g1] PASSED  [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-div-shp2-g0] PASSED  [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-div-shp2-g1] PASSED  [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-div-shp3-g0] PASSED  [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-div-shp3-g1] PASSED  [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-div-shp4-g0] PASSED  [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-div-shp4-g1] PASSED  [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-div-shp5-g0] PASSED  [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-div-shp5-g1] PASSED  [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_lhs-shp0-g0] PASSED [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_lhs-shp0-g1] PASSED [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_lhs-shp1-g0] PASSED [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_lhs-shp1-g1] PASSED [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_lhs-shp2-g0] PASSED [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_lhs-shp2-g1] PASSED [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_lhs-shp3-g0] PASSED [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_lhs-shp3-g1] PASSED [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_lhs-shp4-g0] PASSED [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_lhs-shp4-g1] PASSED [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_lhs-shp5-g0] PASSED [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_lhs-shp5-g1] PASSED [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_rhs-shp0-g0] PASSED [ 40%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_rhs-shp0-g1] PASSED [ 40%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_rhs-shp1-g0] PASSED [ 40%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_rhs-shp1-g1] PASSED [ 40%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_rhs-shp2-g0] PASSED [ 40%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_rhs-shp2-g1] PASSED [ 40%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_rhs-shp3-g0] PASSED [ 40%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_rhs-shp3-g1] PASSED [ 40%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_rhs-shp4-g0] PASSED [ 40%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_rhs-shp4-g1] PASSED [ 40%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_rhs-shp5-g0] PASSED [ 40%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_rhs-shp5-g1] PASSED [ 40%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-u-shp0-g0] PASSED [ 40%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-u-shp0-g1] PASSED [ 40%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-u-shp1-g0] PASSED [ 40%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-u-shp1-g1] PASSED [ 40%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-u-shp2-g0] PASSED [ 40%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-u-shp2-g1] PASSED [ 40%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-u-shp3-g0] PASSED [ 40%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-u-shp3-g1] PASSED [ 40%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-u-shp4-g0] PASSED [ 40%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-u-shp4-g1] PASSED [ 40%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-v-shp0-g0] PASSED [ 40%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-v-shp0-g1] PASSED [ 40%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-v-shp1-g0] PASSED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-v-shp1-g1] PASSED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-v-shp2-g0] PASSED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-v-shp2-g1] PASSED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-v-shp3-g0] PASSED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-v-shp3-g1] PASSED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-v-shp4-g0] PASSED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-v-shp4-g1] PASSED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-e-shp0-g0] PASSED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-e-shp0-g1] PASSED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-e-shp1-g0] PASSED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-e-shp1-g1] PASSED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-e-shp2-g0] PASSED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-e-shp2-g1] PASSED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-e-shp3-g0] PASSED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-e-shp3-g1] PASSED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-e-shp4-g0] PASSED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-e-shp4-g1] PASSED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-u-shp0-g0] PASSED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-u-shp0-g1] PASSED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-u-shp1-g0] PASSED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-u-shp1-g1] PASSED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-u-shp2-g0] PASSED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-u-shp2-g1] PASSED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-u-shp3-g0] PASSED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-u-shp3-g1] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-u-shp4-g0] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-u-shp4-g1] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-v-shp0-g0] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-v-shp0-g1] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-v-shp1-g0] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-v-shp1-g1] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-v-shp2-g0] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-v-shp2-g1] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-v-shp3-g0] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-v-shp3-g1] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-v-shp4-g0] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-v-shp4-g1] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-e-shp0-g0] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-e-shp0-g1] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-e-shp1-g0] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-e-shp1-g1] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-e-shp2-g0] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-e-shp2-g1] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-e-shp3-g0] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-e-shp3-g1] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-e-shp4-g0] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-e-shp4-g1] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-u-shp0-g0] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-u-shp0-g1] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-u-shp1-g0] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-u-shp1-g1] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-u-shp2-g0] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-u-shp2-g1] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-u-shp3-g0] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-u-shp3-g1] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-u-shp4-g0] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-u-shp4-g1] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-v-shp0-g0] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-v-shp0-g1] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-v-shp1-g0] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-v-shp1-g1] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-v-shp2-g0] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-v-shp2-g1] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-v-shp3-g0] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-v-shp3-g1] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-v-shp4-g0] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-v-shp4-g1] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-e-shp0-g0] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-e-shp0-g1] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-e-shp1-g0] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-e-shp1-g1] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-e-shp2-g0] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-e-shp2-g1] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-e-shp3-g0] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-e-shp3-g1] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-e-shp4-g0] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-e-shp4-g1] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-u-shp0-g0] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-u-shp0-g1] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-u-shp1-g0] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-u-shp1-g1] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-u-shp2-g0] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-u-shp2-g1] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-u-shp3-g0] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-u-shp3-g1] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-u-shp4-g0] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-u-shp4-g1] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-v-shp0-g0] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-v-shp0-g1] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-v-shp1-g0] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-v-shp1-g1] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-v-shp2-g0] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-v-shp2-g1] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-v-shp3-g0] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-v-shp3-g1] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-v-shp4-g0] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-v-shp4-g1] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-e-shp0-g0] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-e-shp0-g1] PASSED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-e-shp1-g0] PASSED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-e-shp1-g1] PASSED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-e-shp2-g0] PASSED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-e-shp2-g1] PASSED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-e-shp3-g0] PASSED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-e-shp3-g1] PASSED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-e-shp4-g0] PASSED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-e-shp4-g1] PASSED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-u-shp0-g0] PASSED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-u-shp0-g1] PASSED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-u-shp1-g0] PASSED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-u-shp1-g1] PASSED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-u-shp2-g0] PASSED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-u-shp2-g1] PASSED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-u-shp3-g0] PASSED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-u-shp3-g1] PASSED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-u-shp4-g0] PASSED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-u-shp4-g1] PASSED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-v-shp0-g0] PASSED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-v-shp0-g1] PASSED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-v-shp1-g0] PASSED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-v-shp1-g1] PASSED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-v-shp2-g0] PASSED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-v-shp2-g1] PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-v-shp3-g0] PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-v-shp3-g1] PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-v-shp4-g0] PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-v-shp4-g1] PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-e-shp0-g0] PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-e-shp0-g1] PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-e-shp1-g0] PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-e-shp1-g1] PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-e-shp2-g0] PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-e-shp2-g1] PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-e-shp3-g0] PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-e-shp3-g1] PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-e-shp4-g0] PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-e-shp4-g1] PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-u-shp0-g0] PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-u-shp0-g1] PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-u-shp1-g0] PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-u-shp1-g1] PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-u-shp2-g0] PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-u-shp2-g1] PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-u-shp3-g0] PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-u-shp3-g1] PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-u-shp4-g0] PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-u-shp4-g1] PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-v-shp0-g0] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-v-shp0-g1] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-v-shp1-g0] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-v-shp1-g1] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-v-shp2-g0] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-v-shp2-g1] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-v-shp3-g0] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-v-shp3-g1] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-v-shp4-g0] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-v-shp4-g1] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-e-shp0-g0] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-e-shp0-g1] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-e-shp1-g0] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-e-shp1-g1] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-e-shp2-g0] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-e-shp2-g1] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-e-shp3-g0] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-e-shp3-g1] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-e-shp4-g0] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-e-shp4-g1] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-u-shp0-g0] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-u-shp0-g1] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-u-shp1-g0] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-u-shp1-g1] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-u-shp2-g0] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-u-shp2-g1] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-u-shp3-g0] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-u-shp3-g1] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-u-shp4-g0] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-u-shp4-g1] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-v-shp0-g0] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-v-shp0-g1] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-v-shp1-g0] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-v-shp1-g1] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-v-shp2-g0] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-v-shp2-g1] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-v-shp3-g0] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-v-shp3-g1] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-v-shp4-g0] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-v-shp4-g1] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-e-shp0-g0] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-e-shp0-g1] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-e-shp1-g0] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-e-shp1-g1] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-e-shp2-g0] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-e-shp2-g1] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-e-shp3-g0] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-e-shp3-g1] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-e-shp4-g0] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-e-shp4-g1] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-u-shp0-g0] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-u-shp0-g1] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-u-shp1-g0] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-u-shp1-g1] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-u-shp2-g0] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-u-shp2-g1] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-u-shp3-g0] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-u-shp3-g1] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-u-shp4-g0] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-u-shp4-g1] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-v-shp0-g0] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-v-shp0-g1] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-v-shp1-g0] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-v-shp1-g1] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-v-shp2-g0] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-v-shp2-g1] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-v-shp3-g0] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-v-shp3-g1] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-v-shp4-g0] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-v-shp4-g1] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-e-shp0-g0] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-e-shp0-g1] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-e-shp1-g0] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-e-shp1-g1] PASSED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-e-shp2-g0] PASSED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-e-shp2-g1] PASSED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-e-shp3-g0] PASSED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-e-shp3-g1] PASSED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-e-shp4-g0] PASSED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-e-shp4-g1] PASSED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-u-shp0-g0] PASSED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-u-shp0-g1] PASSED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-u-shp1-g0] PASSED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-u-shp1-g1] PASSED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-u-shp2-g0] PASSED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-u-shp2-g1] PASSED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-u-shp3-g0] PASSED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-u-shp3-g1] PASSED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-u-shp4-g0] PASSED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-u-shp4-g1] PASSED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-v-shp0-g0] PASSED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-v-shp0-g1] PASSED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-v-shp1-g0] PASSED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-v-shp1-g1] PASSED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-v-shp2-g0] PASSED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-v-shp2-g1] PASSED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-v-shp3-g0] PASSED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-v-shp3-g1] PASSED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-v-shp4-g0] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-v-shp4-g1] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-e-shp0-g0] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-e-shp0-g1] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-e-shp1-g0] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-e-shp1-g1] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-e-shp2-g0] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-e-shp2-g1] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-e-shp3-g0] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-e-shp3-g1] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-e-shp4-g0] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-e-shp4-g1] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-u-shp0-g0] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-u-shp0-g1] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-u-shp1-g0] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-u-shp1-g1] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-u-shp2-g0] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-u-shp2-g1] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-u-shp3-g0] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-u-shp3-g1] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-u-shp4-g0] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-u-shp4-g1] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-v-shp0-g0] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-v-shp0-g1] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-v-shp1-g0] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-v-shp1-g1] PASSED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-v-shp2-g0] PASSED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-v-shp2-g1] PASSED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-v-shp3-g0] PASSED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-v-shp3-g1] PASSED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-v-shp4-g0] PASSED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-v-shp4-g1] PASSED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-e-shp0-g0] PASSED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-e-shp0-g1] PASSED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-e-shp1-g0] PASSED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-e-shp1-g1] PASSED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-e-shp2-g0] PASSED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-e-shp2-g1] PASSED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-e-shp3-g0] PASSED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-e-shp3-g1] PASSED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-e-shp4-g0] PASSED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-e-shp4-g1] PASSED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-u-shp0-g0] PASSED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-u-shp0-g1] PASSED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-u-shp1-g0] PASSED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-u-shp1-g1] PASSED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-u-shp2-g0] PASSED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-u-shp2-g1] PASSED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-u-shp3-g0] PASSED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-u-shp3-g1] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-u-shp4-g0] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-u-shp4-g1] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-v-shp0-g0] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-v-shp0-g1] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-v-shp1-g0] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-v-shp1-g1] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-v-shp2-g0] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-v-shp2-g1] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-v-shp3-g0] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-v-shp3-g1] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-v-shp4-g0] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-v-shp4-g1] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-e-shp0-g0] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-e-shp0-g1] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-e-shp1-g0] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-e-shp1-g1] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-e-shp2-g0] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-e-shp2-g1] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-e-shp3-g0] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-e-shp3-g1] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-e-shp4-g0] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-e-shp4-g1] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-u-shp0-g0] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-u-shp0-g1] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-u-shp1-g0] PASSED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-u-shp1-g1] PASSED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-u-shp2-g0] PASSED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-u-shp2-g1] PASSED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-u-shp3-g0] PASSED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-u-shp3-g1] PASSED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-u-shp4-g0] PASSED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-u-shp4-g1] PASSED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-v-shp0-g0] PASSED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-v-shp0-g1] PASSED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-v-shp1-g0] PASSED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-v-shp1-g1] PASSED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-v-shp2-g0] PASSED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-v-shp2-g1] PASSED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-v-shp3-g0] PASSED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-v-shp3-g1] PASSED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-v-shp4-g0] PASSED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-v-shp4-g1] PASSED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-e-shp0-g0] PASSED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-e-shp0-g1] PASSED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-e-shp1-g0] PASSED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-e-shp1-g1] PASSED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-e-shp2-g0] PASSED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-e-shp2-g1] PASSED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-e-shp3-g0] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-e-shp3-g1] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-e-shp4-g0] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-e-shp4-g1] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-u-shp0-g0] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-u-shp0-g1] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-u-shp1-g0] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-u-shp1-g1] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-u-shp2-g0] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-u-shp2-g1] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-u-shp3-g0] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-u-shp3-g1] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-u-shp4-g0] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-u-shp4-g1] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-v-shp0-g0] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-v-shp0-g1] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-v-shp1-g0] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-v-shp1-g1] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-v-shp2-g0] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-v-shp2-g1] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-v-shp3-g0] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-v-shp3-g1] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-v-shp4-g0] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-v-shp4-g1] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-e-shp0-g0] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-e-shp0-g1] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-e-shp1-g0] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-e-shp1-g1] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-e-shp2-g0] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-e-shp2-g1] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-e-shp3-g0] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-e-shp3-g1] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-e-shp4-g0] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-e-shp4-g1] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-u-shp0-g0] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-u-shp0-g1] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-u-shp1-g0] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-u-shp1-g1] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-u-shp2-g0] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-u-shp2-g1] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-u-shp3-g0] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-u-shp3-g1] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-u-shp4-g0] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-u-shp4-g1] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-v-shp0-g0] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-v-shp0-g1] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-v-shp1-g0] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-v-shp1-g1] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-v-shp2-g0] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-v-shp2-g1] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-v-shp3-g0] PASSED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-v-shp3-g1] PASSED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-v-shp4-g0] PASSED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-v-shp4-g1] PASSED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-e-shp0-g0] PASSED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-e-shp0-g1] PASSED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-e-shp1-g0] PASSED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-e-shp1-g1] PASSED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-e-shp2-g0] PASSED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-e-shp2-g1] PASSED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-e-shp3-g0] PASSED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-e-shp3-g1] PASSED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-e-shp4-g0] PASSED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-e-shp4-g1] PASSED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-u-shp0-g0] PASSED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-u-shp0-g1] PASSED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-u-shp1-g0] PASSED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-u-shp1-g1] PASSED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-u-shp2-g0] PASSED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-u-shp2-g1] PASSED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-u-shp3-g0] PASSED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-u-shp3-g1] PASSED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-u-shp4-g0] PASSED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-u-shp4-g1] PASSED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-v-shp0-g0] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-v-shp0-g1] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-v-shp1-g0] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-v-shp1-g1] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-v-shp2-g0] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-v-shp2-g1] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-v-shp3-g0] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-v-shp3-g1] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-v-shp4-g0] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-v-shp4-g1] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-e-shp0-g0] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-e-shp0-g1] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-e-shp1-g0] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-e-shp1-g1] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-e-shp2-g0] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-e-shp2-g1] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-e-shp3-g0] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-e-shp3-g1] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-e-shp4-g0] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-e-shp4-g1] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-u-shp0-g0] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-u-shp0-g1] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-u-shp1-g0] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-u-shp1-g1] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-u-shp2-g0] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-u-shp2-g1] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-u-shp3-g0] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-u-shp3-g1] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-u-shp4-g0] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-u-shp4-g1] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-v-shp0-g0] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-v-shp0-g1] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-v-shp1-g0] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-v-shp1-g1] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-v-shp2-g0] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-v-shp2-g1] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-v-shp3-g0] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-v-shp3-g1] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-v-shp4-g0] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-v-shp4-g1] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-e-shp0-g0] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-e-shp0-g1] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-e-shp1-g0] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-e-shp1-g1] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-e-shp2-g0] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-e-shp2-g1] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-e-shp3-g0] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-e-shp3-g1] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-e-shp4-g0] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-e-shp4-g1] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-u-shp0-g0] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-u-shp0-g1] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-u-shp1-g0] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-u-shp1-g1] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-u-shp2-g0] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-u-shp2-g1] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-u-shp3-g0] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-u-shp3-g1] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-u-shp4-g0] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-u-shp4-g1] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-v-shp0-g0] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-v-shp0-g1] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-v-shp1-g0] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-v-shp1-g1] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-v-shp2-g0] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-v-shp2-g1] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-v-shp3-g0] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-v-shp3-g1] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-v-shp4-g0] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-v-shp4-g1] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-e-shp0-g0] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-e-shp0-g1] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-e-shp1-g0] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-e-shp1-g1] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-e-shp2-g0] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-e-shp2-g1] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-e-shp3-g0] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-e-shp3-g1] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-e-shp4-g0] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-e-shp4-g1] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-u-shp0-g0] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-u-shp0-g1] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-u-shp1-g0] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-u-shp1-g1] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-u-shp2-g0] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-u-shp2-g1] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-u-shp3-g0] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-u-shp3-g1] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-u-shp4-g0] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-u-shp4-g1] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-v-shp0-g0] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-v-shp0-g1] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-v-shp1-g0] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-v-shp1-g1] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-v-shp2-g0] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-v-shp2-g1] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-v-shp3-g0] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-v-shp3-g1] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-v-shp4-g0] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-v-shp4-g1] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-e-shp0-g0] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-e-shp0-g1] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-e-shp1-g0] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-e-shp1-g1] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-e-shp2-g0] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-e-shp2-g1] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-e-shp3-g0] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-e-shp3-g1] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-e-shp4-g0] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-e-shp4-g1] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-u-shp0-g0] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-u-shp0-g1] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-u-shp1-g0] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-u-shp1-g1] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-u-shp2-g0] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-u-shp2-g1] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-u-shp3-g0] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-u-shp3-g1] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-u-shp4-g0] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-u-shp4-g1] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-v-shp0-g0] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-v-shp0-g1] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-v-shp1-g0] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-v-shp1-g1] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-v-shp2-g0] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-v-shp2-g1] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-v-shp3-g0] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-v-shp3-g1] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-v-shp4-g0] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-v-shp4-g1] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-e-shp0-g0] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-e-shp0-g1] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-e-shp1-g0] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-e-shp1-g1] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-e-shp2-g0] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-e-shp2-g1] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-e-shp3-g0] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-e-shp3-g1] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-e-shp4-g0] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-e-shp4-g1] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-u-shp0-g0] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-u-shp0-g1] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-u-shp1-g0] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-u-shp1-g1] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-u-shp2-g0] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-u-shp2-g1] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-u-shp3-g0] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-u-shp3-g1] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-u-shp4-g0] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-u-shp4-g1] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-v-shp0-g0] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-v-shp0-g1] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-v-shp1-g0] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-v-shp1-g1] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-v-shp2-g0] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-v-shp2-g1] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-v-shp3-g0] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-v-shp3-g1] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-v-shp4-g0] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-v-shp4-g1] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-e-shp0-g0] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-e-shp0-g1] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-e-shp1-g0] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-e-shp1-g1] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-e-shp2-g0] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-e-shp2-g1] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-e-shp3-g0] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-e-shp3-g1] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-e-shp4-g0] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-e-shp4-g1] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-u-shp0-g0] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-u-shp0-g1] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-u-shp1-g0] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-u-shp1-g1] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-u-shp2-g0] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-u-shp2-g1] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-u-shp3-g0] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-u-shp3-g1] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-u-shp4-g0] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-u-shp4-g1] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-v-shp0-g0] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-v-shp0-g1] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-v-shp1-g0] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-v-shp1-g1] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-v-shp2-g0] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-v-shp2-g1] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-v-shp3-g0] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-v-shp3-g1] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-v-shp4-g0] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-v-shp4-g1] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-e-shp0-g0] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-e-shp0-g1] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-e-shp1-g0] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-e-shp1-g1] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-e-shp2-g0] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-e-shp2-g1] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-e-shp3-g0] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-e-shp3-g1] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-e-shp4-g0] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-e-shp4-g1] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-u-shp0-g0] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-u-shp0-g1] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-u-shp1-g0] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-u-shp1-g1] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-u-shp2-g0] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-u-shp2-g1] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-u-shp3-g0] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-u-shp3-g1] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-u-shp4-g0] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-u-shp4-g1] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-v-shp0-g0] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-v-shp0-g1] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-v-shp1-g0] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-v-shp1-g1] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-v-shp2-g0] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-v-shp2-g1] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-v-shp3-g0] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-v-shp3-g1] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-v-shp4-g0] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-v-shp4-g1] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-e-shp0-g0] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-e-shp0-g1] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-e-shp1-g0] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-e-shp1-g1] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-e-shp2-g0] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-e-shp2-g1] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-e-shp3-g0] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-e-shp3-g1] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-e-shp4-g0] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-e-shp4-g1] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-u-shp0-g0] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-u-shp0-g1] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-u-shp1-g0] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-u-shp1-g1] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-u-shp2-g0] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-u-shp2-g1] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-u-shp3-g0] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-u-shp3-g1] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-u-shp4-g0] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-u-shp4-g1] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-v-shp0-g0] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-v-shp0-g1] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-v-shp1-g0] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-v-shp1-g1] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-v-shp2-g0] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-v-shp2-g1] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-v-shp3-g0] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-v-shp3-g1] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-v-shp4-g0] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-v-shp4-g1] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-e-shp0-g0] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-e-shp0-g1] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-e-shp1-g0] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-e-shp1-g1] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-e-shp2-g0] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-e-shp2-g1] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-e-shp3-g0] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-e-shp3-g1] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-e-shp4-g0] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-e-shp4-g1] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-u-shp0-g0] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-u-shp0-g1] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-u-shp1-g0] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-u-shp1-g1] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-u-shp2-g0] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-u-shp2-g1] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-u-shp3-g0] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-u-shp3-g1] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-u-shp4-g0] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-u-shp4-g1] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-v-shp0-g0] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-v-shp0-g1] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-v-shp1-g0] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-v-shp1-g1] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-v-shp2-g0] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-v-shp2-g1] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-v-shp3-g0] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-v-shp3-g1] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-v-shp4-g0] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-v-shp4-g1] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-e-shp0-g0] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-e-shp0-g1] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-e-shp1-g0] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-e-shp1-g1] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-e-shp2-g0] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-e-shp2-g1] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-e-shp3-g0] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-e-shp3-g1] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-e-shp4-g0] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-e-shp4-g1] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-u-shp0-g0] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-u-shp0-g1] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-u-shp1-g0] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-u-shp1-g1] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-u-shp2-g0] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-u-shp2-g1] PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-u-shp3-g0] PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-u-shp3-g1] PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-u-shp4-g0] PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-u-shp4-g1] PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-v-shp0-g0] PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-v-shp0-g1] PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-v-shp1-g0] PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-v-shp1-g1] PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-v-shp2-g0] PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-v-shp2-g1] PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-v-shp3-g0] PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-v-shp3-g1] PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-v-shp4-g0] PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-v-shp4-g1] PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-e-shp0-g0] PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-e-shp0-g1] PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-e-shp1-g0] PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-e-shp1-g1] PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-e-shp2-g0] PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-e-shp2-g1] PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-e-shp3-g0] PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-e-shp3-g1] PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-e-shp4-g0] PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-e-shp4-g1] PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-u-shp0-g0] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-u-shp0-g1] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-u-shp1-g0] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-u-shp1-g1] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-u-shp2-g0] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-u-shp2-g1] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-u-shp3-g0] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-u-shp3-g1] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-u-shp4-g0] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-u-shp4-g1] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-v-shp0-g0] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-v-shp0-g1] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-v-shp1-g0] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-v-shp1-g1] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-v-shp2-g0] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-v-shp2-g1] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-v-shp3-g0] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-v-shp3-g1] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-v-shp4-g0] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-v-shp4-g1] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-e-shp0-g0] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-e-shp0-g1] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-e-shp1-g0] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-e-shp1-g1] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-e-shp2-g0] PASSED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-e-shp2-g1] PASSED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-e-shp3-g0] PASSED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-e-shp3-g1] PASSED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-e-shp4-g0] PASSED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-e-shp4-g1] PASSED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-u-shp0-g0] PASSED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-u-shp0-g1] PASSED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-u-shp1-g0] PASSED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-u-shp1-g1] PASSED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-u-shp2-g0] PASSED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-u-shp2-g1] PASSED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-u-shp3-g0] PASSED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-u-shp3-g1] PASSED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-u-shp4-g0] PASSED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-u-shp4-g1] PASSED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-v-shp0-g0] PASSED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-v-shp0-g1] PASSED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-v-shp1-g0] PASSED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-v-shp1-g1] PASSED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-v-shp2-g0] PASSED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-v-shp2-g1] PASSED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-v-shp3-g0] PASSED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-v-shp3-g1] PASSED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-v-shp4-g0] PASSED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-v-shp4-g1] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-e-shp0-g0] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-e-shp0-g1] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-e-shp1-g0] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-e-shp1-g1] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-e-shp2-g0] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-e-shp2-g1] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-e-shp3-g0] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-e-shp3-g1] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-e-shp4-g0] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-e-shp4-g1] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-u-shp0-g0] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-u-shp0-g1] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-u-shp1-g0] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-u-shp1-g1] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-u-shp2-g0] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-u-shp2-g1] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-u-shp3-g0] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-u-shp3-g1] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-u-shp4-g0] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-u-shp4-g1] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-v-shp0-g0] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-v-shp0-g1] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-v-shp1-g0] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-v-shp1-g1] PASSED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-v-shp2-g0] PASSED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-v-shp2-g1] PASSED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-v-shp3-g0] PASSED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-v-shp3-g1] PASSED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-v-shp4-g0] PASSED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-v-shp4-g1] PASSED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-e-shp0-g0] PASSED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-e-shp0-g1] PASSED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-e-shp1-g0] PASSED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-e-shp1-g1] PASSED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-e-shp2-g0] PASSED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-e-shp2-g1] PASSED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-e-shp3-g0] PASSED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-e-shp3-g1] PASSED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-e-shp4-g0] PASSED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-e-shp4-g1] PASSED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-u-shp0-g0] PASSED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-u-shp0-g1] PASSED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-u-shp1-g0] PASSED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-u-shp1-g1] PASSED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-u-shp2-g0] PASSED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-u-shp2-g1] PASSED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-u-shp3-g0] PASSED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-u-shp3-g1] PASSED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-u-shp4-g0] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-u-shp4-g1] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-v-shp0-g0] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-v-shp0-g1] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-v-shp1-g0] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-v-shp1-g1] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-v-shp2-g0] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-v-shp2-g1] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-v-shp3-g0] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-v-shp3-g1] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-v-shp4-g0] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-v-shp4-g1] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-e-shp0-g0] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-e-shp0-g1] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-e-shp1-g0] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-e-shp1-g1] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-e-shp2-g0] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-e-shp2-g1] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-e-shp3-g0] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-e-shp3-g1] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-e-shp4-g0] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-e-shp4-g1] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-u-shp0-g0] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-u-shp0-g1] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-u-shp1-g0] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-u-shp1-g1] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-u-shp2-g0] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-u-shp2-g1] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-u-shp3-g0] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-u-shp3-g1] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-u-shp4-g0] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-u-shp4-g1] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-v-shp0-g0] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-v-shp0-g1] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-v-shp1-g0] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-v-shp1-g1] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-v-shp2-g0] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-v-shp2-g1] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-v-shp3-g0] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-v-shp3-g1] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-v-shp4-g0] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-v-shp4-g1] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-e-shp0-g0] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-e-shp0-g1] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-e-shp1-g0] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-e-shp1-g1] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-e-shp2-g0] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-e-shp2-g1] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-e-shp3-g0] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-e-shp3-g1] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-e-shp4-g0] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-e-shp4-g1] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-u-shp0-g0] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-u-shp0-g1] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-u-shp1-g0] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-u-shp1-g1] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-u-shp2-g0] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-u-shp2-g1] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-u-shp3-g0] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-u-shp3-g1] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-u-shp4-g0] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-u-shp4-g1] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-v-shp0-g0] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-v-shp0-g1] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-v-shp1-g0] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-v-shp1-g1] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-v-shp2-g0] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-v-shp2-g1] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-v-shp3-g0] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-v-shp3-g1] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-v-shp4-g0] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-v-shp4-g1] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-e-shp0-g0] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-e-shp0-g1] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-e-shp1-g0] PASSED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-e-shp1-g1] PASSED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-e-shp2-g0] PASSED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-e-shp2-g1] PASSED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-e-shp3-g0] PASSED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-e-shp3-g1] PASSED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-e-shp4-g0] PASSED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-e-shp4-g1] PASSED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-u-shp0-g0] PASSED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-u-shp0-g1] PASSED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-u-shp1-g0] PASSED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-u-shp1-g1] PASSED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-u-shp2-g0] PASSED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-u-shp2-g1] PASSED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-u-shp3-g0] PASSED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-u-shp3-g1] PASSED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-u-shp4-g0] PASSED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-u-shp4-g1] PASSED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-v-shp0-g0] PASSED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-v-shp0-g1] PASSED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-v-shp1-g0] PASSED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-v-shp1-g1] PASSED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-v-shp2-g0] PASSED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-v-shp2-g1] PASSED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-v-shp3-g0] PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-v-shp3-g1] PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-v-shp4-g0] PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-v-shp4-g1] PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-e-shp0-g0] PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-e-shp0-g1] PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-e-shp1-g0] PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-e-shp1-g1] Running on dgl-manual-large-cpu in /root/jenkins/workspace/dgl_PR-4648
[Pipeline] {
PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-e-shp2-g0] PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-e-shp2-g1] [Pipeline] checkout
The recommended git tool is: git
using credential 150de63f-189c-4717-bcaf-010460d2f51a
Cloning the remote Git repository
Cloning with configured refspecs honoured and without tags
PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-e-shp3-g0] PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-e-shp3-g1] PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-e-shp4-g0] PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-e-shp4-g1] PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-u-shp0-g0] PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-u-shp0-g1] Cloning repository https://github.com/dmlc/dgl.git
 > git init /root/jenkins/workspace/dgl_PR-4648 # timeout=10
Fetching upstream changes from https://github.com/dmlc/dgl.git
 > git --version # timeout=10
 > git --version # 'git version 2.17.1'
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git fetch --no-tags --progress -- https://github.com/dmlc/dgl.git +refs/pull/4648/head:refs/remotes/origin/PR-4648 +refs/heads/master:refs/remotes/origin/master # timeout=10
PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-u-shp1-g0] PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-u-shp1-g1] PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-u-shp2-g0] PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-u-shp2-g1] Cleaning workspace
Fetching without tags
PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-u-shp3-g0] PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-u-shp3-g1]  > git config remote.origin.url https://github.com/dmlc/dgl.git # timeout=10
 > git config --add remote.origin.fetch +refs/pull/4648/head:refs/remotes/origin/PR-4648 # timeout=10
 > git config --add remote.origin.fetch +refs/heads/master:refs/remotes/origin/master # timeout=10
 > git config remote.origin.url https://github.com/dmlc/dgl.git # timeout=10
 > git rev-parse --verify HEAD # timeout=10
No valid HEAD. Skipping the resetting
 > git clean -fdx # timeout=10
Fetching upstream changes from https://github.com/dmlc/dgl.git
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git fetch --no-tags --progress -- https://github.com/dmlc/dgl.git +refs/pull/4648/head:refs/remotes/origin/PR-4648 +refs/heads/master:refs/remotes/origin/master # timeout=10
Merging remotes/origin/master commit d78a3a4baf611b90871a849f58647160c7cd9ab4 into PR head commit 37ed78035f406940ed88a4b60cadd2b7cfde5fea
PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-u-shp4-g0] Merge succeeded, producing 37ed78035f406940ed88a4b60cadd2b7cfde5fea
Checking out Revision 37ed78035f406940ed88a4b60cadd2b7cfde5fea (PR-4648)
Running on dgl-manual-large-cpu in /root/jenkins/workspace/dgl_PR-4648@2
[Pipeline] {
[Pipeline] checkout
The recommended git tool is: git
using credential 150de63f-189c-4717-bcaf-010460d2f51a
Cloning the remote Git repository
Cloning with configured refspecs honoured and without tags
PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-u-shp4-g1] PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-v-shp0-g0] Commit message: "fix for pytorch < 1.12"
Cleaning workspace
PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-v-shp0-g1] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-v-shp1-g0] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-v-shp1-g1]  > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git remote # timeout=10
 > git config --get remote.origin.url # timeout=10
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git merge d78a3a4baf611b90871a849f58647160c7cd9ab4 # timeout=10
 > git rev-parse HEAD^{commit} # timeout=10
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git branch -a -v --no-abbrev # timeout=10
 > git checkout -b PR-4648 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git rev-parse --verify HEAD # timeout=10
Resetting working tree
 > git reset --hard # timeout=10
 > git clean -fdx # timeout=10
PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-v-shp2-g0] [Pipeline] withEnv
[Pipeline] {
Cloning repository https://github.com/dmlc/dgl.git
 > git init /root/jenkins/workspace/dgl_PR-4648@2 # timeout=10
Fetching upstream changes from https://github.com/dmlc/dgl.git
 > git --version # timeout=10
 > git --version # 'git version 2.17.1'
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git fetch --no-tags --progress -- https://github.com/dmlc/dgl.git +refs/pull/4648/head:refs/remotes/origin/PR-4648 +refs/heads/master:refs/remotes/origin/master # timeout=10
[Pipeline] isUnix
[Pipeline] withEnv
[Pipeline] {
[Pipeline] sh
PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-v-shp2-g1] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-v-shp3-g0] + docker pull dgllib/dgl-ci-cpu:v220816
PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-v-shp3-g1] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-v-shp4-g0] Cleaning workspace
Fetching without tags
PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-v-shp4-g1] Merging remotes/origin/master commit d78a3a4baf611b90871a849f58647160c7cd9ab4 into PR head commit 37ed78035f406940ed88a4b60cadd2b7cfde5fea
Merge succeeded, producing 37ed78035f406940ed88a4b60cadd2b7cfde5fea
Checking out Revision 37ed78035f406940ed88a4b60cadd2b7cfde5fea (PR-4648)
PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-e-shp0-g0] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-e-shp0-g1] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-e-shp1-g0] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-e-shp1-g1] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-e-shp2-g0] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-e-shp2-g1] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-e-shp3-g0] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-e-shp3-g1] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-e-shp4-g0] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-e-shp4-g1] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-u-shp0-g0] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-u-shp0-g1] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-u-shp1-g0] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-u-shp1-g1] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-u-shp2-g0] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-u-shp2-g1] PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-u-shp3-g0] PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-u-shp3-g1] PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-u-shp4-g0] PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-u-shp4-g1] PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-v-shp0-g0] PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-v-shp0-g1] Commit message: "fix for pytorch < 1.12"
Cleaning workspace
 > git config remote.origin.url https://github.com/dmlc/dgl.git # timeout=10
 > git config --add remote.origin.fetch +refs/pull/4648/head:refs/remotes/origin/PR-4648 # timeout=10
 > git config --add remote.origin.fetch +refs/heads/master:refs/remotes/origin/master # timeout=10
 > git config remote.origin.url https://github.com/dmlc/dgl.git # timeout=10
 > git rev-parse --verify HEAD # timeout=10
No valid HEAD. Skipping the resetting
 > git clean -fdx # timeout=10
Fetching upstream changes from https://github.com/dmlc/dgl.git
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git fetch --no-tags --progress -- https://github.com/dmlc/dgl.git +refs/pull/4648/head:refs/remotes/origin/PR-4648 +refs/heads/master:refs/remotes/origin/master # timeout=10
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git remote # timeout=10
 > git config --get remote.origin.url # timeout=10
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git merge d78a3a4baf611b90871a849f58647160c7cd9ab4 # timeout=10
 > git rev-parse HEAD^{commit} # timeout=10
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git branch -a -v --no-abbrev # timeout=10
 > git checkout -b PR-4648 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git rev-parse --verify HEAD # timeout=10
Resetting working tree
 > git reset --hard # timeout=10
PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-v-shp1-g0] v220816: Pulling from dgllib/dgl-ci-cpu
Digest: sha256:64b385c33b44dc57cb96ff264a84d8dfb8ced0caa9b30fbc4cec6d5ee511b099
Status: Image is up to date for dgllib/dgl-ci-cpu:v220816
docker.io/dgllib/dgl-ci-cpu:v220816
[Pipeline] }
[Pipeline] // withEnv
[Pipeline] isUnix
[Pipeline] withEnv
[Pipeline] {
[Pipeline] sh
PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-v-shp1-g1] PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-v-shp2-g0] + docker inspect -f . dgllib/dgl-ci-cpu:v220816
.
PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-v-shp2-g1] [Pipeline] }
[Pipeline] // withEnv
[Pipeline] withDockerContainer
dgl-manual-large-cpu does not seem to be running inside a container
$ docker run -t -d -u 0:0 -w /root/jenkins/workspace/dgl_PR-4648 -v /root/jenkins/workspace/dgl_PR-4648:/root/jenkins/workspace/dgl_PR-4648:rw,z -v /root/jenkins/workspace/dgl_PR-4648@tmp:/root/jenkins/workspace/dgl_PR-4648@tmp:rw,z -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** dgllib/dgl-ci-cpu:v220816 cat
PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-v-shp3-g0] PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-v-shp3-g1]  > git clean -fdx # timeout=10
PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-v-shp4-g0] $ docker top 280b39495ad3fad21dbb9762a178907987f011c3e8f474eb1caee524b0ff882c -eo pid,comm
PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-v-shp4-g1] PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-e-shp0-g0] [Pipeline] {
[Pipeline] withEnv
[Pipeline] {
[Pipeline] isUnix
[Pipeline] withEnv
[Pipeline] {
[Pipeline] sh
[Pipeline] sh
PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-e-shp0-g1] + docker pull dgllib/dgl-ci-cpu:v220816
PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-e-shp1-g0] + rm -rf CMakeLists.txt CONTRIBUTORS.md Jenkinsfile LICENSE NEWS.md README.md apps benchmarks cmake conda dglgo docker docs examples featgraph include pyproject.toml python readthedocs.yml src tensoradapter tests third_party tools tutorials
[Pipeline] checkout
The recommended git tool is: git
using credential 150de63f-189c-4717-bcaf-010460d2f51a
Warning: JENKINS-30600: special launcher org.jenkinsci.plugins.docker.workflow.WithContainerStep$Decorator$1@7c2abf69; decorates RemoteLauncher[hudson.remoting.Channel@1d9b9638:dgl-manual-large-cpu] will be ignored (a typical symptom is the Git executable not being run inside a designated container)
Fetching changes from the remote Git repository
Cleaning workspace
PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-e-shp1-g1] PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-e-shp2-g0] Fetching without tags
PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-e-shp2-g1] Merging remotes/origin/master commit d78a3a4baf611b90871a849f58647160c7cd9ab4 into PR head commit 37ed78035f406940ed88a4b60cadd2b7cfde5fea
Merge succeeded, producing 37ed78035f406940ed88a4b60cadd2b7cfde5fea
Checking out Revision 37ed78035f406940ed88a4b60cadd2b7cfde5fea (PR-4648)
PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-e-shp3-g0] PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-e-shp3-g1] v220816: Pulling from dgllib/dgl-ci-cpu
Digest: sha256:64b385c33b44dc57cb96ff264a84d8dfb8ced0caa9b30fbc4cec6d5ee511b099
Status: Image is up to date for dgllib/dgl-ci-cpu:v220816
docker.io/dgllib/dgl-ci-cpu:v220816
Commit message: "fix for pytorch < 1.12"
Cleaning workspace
[Pipeline] }
[Pipeline] // withEnv
[Pipeline] isUnix
PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-e-shp4-g0] PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-e-shp4-g1] [Pipeline] withEnv
[Pipeline] {
 > git rev-parse --resolve-git-dir /root/jenkins/workspace/dgl_PR-4648/.git # timeout=10
 > git config remote.origin.url https://github.com/dmlc/dgl.git # timeout=10
 > git rev-parse --verify HEAD # timeout=10
Resetting working tree
 > git reset --hard # timeout=10
 > git clean -fdx # timeout=10
Fetching upstream changes from https://github.com/dmlc/dgl.git
 > git --version # timeout=10
 > git --version # 'git version 2.17.1'
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git fetch --no-tags --progress -- https://github.com/dmlc/dgl.git +refs/pull/4648/head:refs/remotes/origin/PR-4648 +refs/heads/master:refs/remotes/origin/master # timeout=10
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git remote # timeout=10
 > git config --get remote.origin.url # timeout=10
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git merge d78a3a4baf611b90871a849f58647160c7cd9ab4 # timeout=10
 > git rev-parse HEAD^{commit} # timeout=10
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git branch -a -v --no-abbrev # timeout=10
 > git branch -D PR-4648 # timeout=10
 > git checkout -b PR-4648 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git rev-parse --verify HEAD # timeout=10
Resetting working tree
 > git reset --hard # timeout=10
 > git clean -fdx # timeout=10
[Pipeline] sh
PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-u-shp0-g0] + docker inspect -f . dgllib/dgl-ci-cpu:v220816
.
[Pipeline] }
[Pipeline] // withEnv
[Pipeline] withDockerContainer
dgl-manual-large-cpu does not seem to be running inside a container
$ docker run -t -d -u 0:0 -w /root/jenkins/workspace/dgl_PR-4648@2 -v /root/jenkins/workspace/dgl_PR-4648@2:/root/jenkins/workspace/dgl_PR-4648@2:rw,z -v /root/jenkins/workspace/dgl_PR-4648@2@tmp:/root/jenkins/workspace/dgl_PR-4648@2@tmp:rw,z -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** dgllib/dgl-ci-cpu:v220816 cat
PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-u-shp0-g1] PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-u-shp1-g0] PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-u-shp1-g1] $ docker top 7ac09922888ca0e0d8459513540702ad50c9384fbb02f538e426a164f5d98757 -eo pid,comm
[Pipeline] {
PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-u-shp2-g0] PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-u-shp2-g1] [Pipeline] sh
[Pipeline] stage
[Pipeline] { (Tensorflow CPU Unit test)
[Pipeline] sh
PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-u-shp3-g0] + git submodule update --recursive --init
Submodule 'third_party/METIS' (https://github.com/KarypisLab/METIS.git) registered for path 'third_party/METIS'
Submodule 'third_party/dlpack' (https://github.com/dmlc/dlpack.git) registered for path 'third_party/dlpack'
Submodule 'third_party/dmlc-core' (https://github.com/dmlc/dmlc-core.git) registered for path 'third_party/dmlc-core'
Submodule 'third_party/googletest' (https://github.com/google/googletest.git) registered for path 'third_party/googletest'
Submodule 'third_party/libxsmm' (https://github.com/hfp/libxsmm.git) registered for path 'third_party/libxsmm'
Submodule 'third_party/nanoflann' (https://github.com/jlblancoc/nanoflann) registered for path 'third_party/nanoflann'
Submodule 'third_party/nccl' (https://github.com/nvidia/nccl) registered for path 'third_party/nccl'
Submodule 'third_party/phmap' (https://github.com/greg7mdp/parallel-hashmap.git) registered for path 'third_party/phmap'
Submodule 'third_party/tensorpipe' (https://github.com/pytorch/tensorpipe) registered for path 'third_party/tensorpipe'
Submodule 'third_party/thrust' (https://github.com/NVIDIA/thrust.git) registered for path 'third_party/thrust'
Submodule 'third_party/tvm' (https://github.com/apache/incubator-tvm) registered for path 'third_party/tvm'
Submodule 'third_party/xbyak' (https://github.com/herumi/xbyak) registered for path 'third_party/xbyak'
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/METIS'...
+ rm -rf CMakeLists.txt CONTRIBUTORS.md Jenkinsfile LICENSE NEWS.md README.md apps benchmarks cmake conda dglgo docker docs examples featgraph include pyproject.toml python readthedocs.yml src tensoradapter tests third_party tools tutorials
[Pipeline] checkout
The recommended git tool is: git
using credential 150de63f-189c-4717-bcaf-010460d2f51a
Warning: JENKINS-30600: special launcher org.jenkinsci.plugins.docker.workflow.WithContainerStep$Decorator$1@101e882; decorates RemoteLauncher[hudson.remoting.Channel@1d9b9638:dgl-manual-large-cpu] will be ignored (a typical symptom is the Git executable not being run inside a designated container)
Fetching changes from the remote Git repository
Cleaning workspace
PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-u-shp3-g1] PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-u-shp4-g0] Fetching without tags
PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-u-shp4-g1] Merging remotes/origin/master commit d78a3a4baf611b90871a849f58647160c7cd9ab4 into PR head commit 37ed78035f406940ed88a4b60cadd2b7cfde5fea
PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-v-shp0-g0] PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-v-shp0-g1] PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-v-shp1-g0] PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-v-shp1-g1] PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-v-shp2-g0] PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-v-shp2-g1] PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-v-shp3-g0] PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-v-shp3-g1] PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-v-shp4-g0] PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-v-shp4-g1] PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-e-shp0-g0] Merge succeeded, producing 37ed78035f406940ed88a4b60cadd2b7cfde5fea
Checking out Revision 37ed78035f406940ed88a4b60cadd2b7cfde5fea (PR-4648)
Commit message: "fix for pytorch < 1.12"
Cleaning workspace
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/dlpack'...
PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-e-shp0-g1] PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-e-shp1-g0]  > git rev-parse --resolve-git-dir /root/jenkins/workspace/dgl_PR-4648@2/.git # timeout=10
 > git config remote.origin.url https://github.com/dmlc/dgl.git # timeout=10
 > git rev-parse --verify HEAD # timeout=10
Resetting working tree
 > git reset --hard # timeout=10
 > git clean -fdx # timeout=10
Fetching upstream changes from https://github.com/dmlc/dgl.git
 > git --version # timeout=10
 > git --version # 'git version 2.17.1'
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git fetch --no-tags --progress -- https://github.com/dmlc/dgl.git +refs/pull/4648/head:refs/remotes/origin/PR-4648 +refs/heads/master:refs/remotes/origin/master # timeout=10
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git remote # timeout=10
 > git config --get remote.origin.url # timeout=10
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git merge d78a3a4baf611b90871a849f58647160c7cd9ab4 # timeout=10
 > git rev-parse HEAD^{commit} # timeout=10
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git branch -a -v --no-abbrev # timeout=10
 > git branch -D PR-4648 # timeout=10
 > git checkout -b PR-4648 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git rev-parse --verify HEAD # timeout=10
Resetting working tree
 > git reset --hard # timeout=10
 > git clean -fdx # timeout=10
PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-e-shp1-g1] [Pipeline] sh
PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-e-shp2-g0] PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-e-shp2-g1] Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/dmlc-core'...
+ git submodule update --recursive --init
Submodule 'third_party/METIS' (https://github.com/KarypisLab/METIS.git) registered for path 'third_party/METIS'
Submodule 'third_party/dlpack' (https://github.com/dmlc/dlpack.git) registered for path 'third_party/dlpack'
Submodule 'third_party/dmlc-core' (https://github.com/dmlc/dmlc-core.git) registered for path 'third_party/dmlc-core'
Submodule 'third_party/googletest' (https://github.com/google/googletest.git) registered for path 'third_party/googletest'
Submodule 'third_party/libxsmm' (https://github.com/hfp/libxsmm.git) registered for path 'third_party/libxsmm'
Submodule 'third_party/nanoflann' (https://github.com/jlblancoc/nanoflann) registered for path 'third_party/nanoflann'
Submodule 'third_party/nccl' (https://github.com/nvidia/nccl) registered for path 'third_party/nccl'
Submodule 'third_party/phmap' (https://github.com/greg7mdp/parallel-hashmap.git) registered for path 'third_party/phmap'
Submodule 'third_party/tensorpipe' (https://github.com/pytorch/tensorpipe) registered for path 'third_party/tensorpipe'
Submodule 'third_party/thrust' (https://github.com/NVIDIA/thrust.git) registered for path 'third_party/thrust'
Submodule 'third_party/tvm' (https://github.com/apache/incubator-tvm) registered for path 'third_party/tvm'
Submodule 'third_party/xbyak' (https://github.com/herumi/xbyak) registered for path 'third_party/xbyak'
Cloning into '/root/jenkins/workspace/dgl_PR-4648@2/third_party/METIS'...
PASSED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-e-shp3-g0] PASSED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-e-shp3-g1] PASSED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-e-shp4-g0] Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/googletest'...
PASSED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-e-shp4-g1] Cloning into '/root/jenkins/workspace/dgl_PR-4648@2/third_party/dlpack'...
PASSED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-u-shp0-g0] PASSED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-u-shp0-g1] PASSED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-u-shp1-g0] PASSED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-u-shp1-g1] PASSED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-u-shp2-g0] Cloning into '/root/jenkins/workspace/dgl_PR-4648@2/third_party/dmlc-core'...
PASSED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-u-shp2-g1] Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/libxsmm'...
PASSED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-u-shp3-g0] PASSED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-u-shp3-g1] Cloning into '/root/jenkins/workspace/dgl_PR-4648@2/third_party/googletest'...
PASSED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-u-shp4-g0] PASSED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-u-shp4-g1] PASSED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-v-shp0-g0] PASSED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-v-shp0-g1] PASSED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-v-shp1-g0] PASSED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-v-shp1-g1] PASSED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-v-shp2-g0] PASSED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-v-shp2-g1] PASSED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-v-shp3-g0] PASSED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-v-shp3-g1] Cloning into '/root/jenkins/workspace/dgl_PR-4648@2/third_party/libxsmm'...
PASSED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-v-shp4-g0] PASSED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-v-shp4-g1] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-e-shp0-g0] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-e-shp0-g1] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-e-shp1-g0] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-e-shp1-g1] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-e-shp2-g0] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-e-shp2-g1] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-e-shp3-g0] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-e-shp3-g1] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-e-shp4-g0] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-e-shp4-g1] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-u-shp0-g0] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-u-shp0-g1] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-u-shp1-g0] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-u-shp1-g1] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-u-shp2-g0] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-u-shp2-g1] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-u-shp3-g0] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-u-shp3-g1] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-u-shp4-g0] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-u-shp4-g1] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-v-shp0-g0] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-v-shp0-g1] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-v-shp1-g0] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-v-shp1-g1] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-v-shp2-g0] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-v-shp2-g1] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-v-shp3-g0] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-v-shp3-g1] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-v-shp4-g0] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-v-shp4-g1] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-e-shp0-g0] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-e-shp0-g1] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-e-shp1-g0] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-e-shp1-g1] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-e-shp2-g0] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-e-shp2-g1] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-e-shp3-g0] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-e-shp3-g1] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-e-shp4-g0] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-e-shp4-g1] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-u-shp0-g0] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-u-shp0-g1] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-u-shp1-g0] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-u-shp1-g1] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-u-shp2-g0] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-u-shp2-g1] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-u-shp3-g0] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-u-shp3-g1] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-u-shp4-g0] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-u-shp4-g1] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-v-shp0-g0] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-v-shp0-g1] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-v-shp1-g0] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-v-shp1-g1] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-v-shp2-g0] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-v-shp2-g1] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-v-shp3-g0] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-v-shp3-g1] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-v-shp4-g0] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-v-shp4-g1] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-e-shp0-g0] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-e-shp0-g1] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-e-shp1-g0] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-e-shp1-g1] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-e-shp2-g0] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-e-shp2-g1] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-e-shp3-g0] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-e-shp3-g1] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-e-shp4-g0] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-e-shp4-g1] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-u-shp0-g0] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-u-shp0-g1] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-u-shp1-g0] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-u-shp1-g1] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-u-shp2-g0] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-u-shp2-g1] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-u-shp3-g0] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-u-shp3-g1] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-u-shp4-g0] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-u-shp4-g1] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-v-shp0-g0] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-v-shp0-g1] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-v-shp1-g0] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-v-shp1-g1] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-v-shp2-g0] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-v-shp2-g1] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-v-shp3-g0] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-v-shp3-g1] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-v-shp4-g0] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-v-shp4-g1] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-e-shp0-g0] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-e-shp0-g1] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-e-shp1-g0] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-e-shp1-g1] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-e-shp2-g0] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-e-shp2-g1] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-e-shp3-g0] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-e-shp3-g1] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-e-shp4-g0] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-e-shp4-g1] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-u-shp0-g0] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-u-shp0-g1] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-u-shp1-g0] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-u-shp1-g1] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-u-shp2-g0] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-u-shp2-g1] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-u-shp3-g0] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-u-shp3-g1] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-u-shp4-g0] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-u-shp4-g1] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-v-shp0-g0] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-v-shp0-g1] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-v-shp1-g0] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-v-shp1-g1] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-v-shp2-g0] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-v-shp2-g1] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-v-shp3-g0] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-v-shp3-g1] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-v-shp4-g0] Cloning into '/root/jenkins/workspace/dgl_PR-4648@2/third_party/nanoflann'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648@2/third_party/nccl'...
PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-v-shp4-g1] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-e-shp0-g0] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-e-shp0-g1] Cloning into '/root/jenkins/workspace/dgl_PR-4648@2/third_party/phmap'...
PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-e-shp1-g0] Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/nanoflann'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/nccl'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/phmap'...
PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-e-shp1-g1] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-e-shp2-g0] Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tensorpipe'...
PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-e-shp2-g1] Cloning into '/root/jenkins/workspace/dgl_PR-4648@2/third_party/tensorpipe'...
PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-e-shp3-g0] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-e-shp3-g1] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-e-shp4-g0] Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/thrust'...
PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-e-shp4-g1] Cloning into '/root/jenkins/workspace/dgl_PR-4648@2/third_party/thrust'...
PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-u-shp0-g0] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-u-shp0-g1] Running on dgl-manual-large-cpu in /root/jenkins/workspace/dgl_PR-4648@3
[Pipeline] {
[Pipeline] checkout
The recommended git tool is: git
using credential 150de63f-189c-4717-bcaf-010460d2f51a
Cloning the remote Git repository
Cloning with configured refspecs honoured and without tags
PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-u-shp1-g0] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-u-shp1-g1] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-u-shp2-g0] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-u-shp2-g1] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-u-shp3-g0] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-u-shp3-g1] Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tvm'...
Cloning repository https://github.com/dmlc/dgl.git
 > git init /root/jenkins/workspace/dgl_PR-4648@3 # timeout=10
Fetching upstream changes from https://github.com/dmlc/dgl.git
 > git --version # timeout=10
 > git --version # 'git version 2.17.1'
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git fetch --no-tags --progress -- https://github.com/dmlc/dgl.git +refs/pull/4648/head:refs/remotes/origin/PR-4648 +refs/heads/master:refs/remotes/origin/master # timeout=10
PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-u-shp4-g0] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-u-shp4-g1] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-v-shp0-g0] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-v-shp0-g1] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-v-shp1-g0] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-v-shp1-g1] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-v-shp2-g0] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-v-shp2-g1] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-v-shp3-g0] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-v-shp3-g1] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-v-shp4-g0] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-v-shp4-g1] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-e-shp0-g0] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-e-shp0-g1] Cloning into '/root/jenkins/workspace/dgl_PR-4648@2/third_party/tvm'...
Cleaning workspace
Fetching without tags
PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-e-shp1-g0] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-e-shp1-g1] Merging remotes/origin/master commit d78a3a4baf611b90871a849f58647160c7cd9ab4 into PR head commit 37ed78035f406940ed88a4b60cadd2b7cfde5fea
 > git config remote.origin.url https://github.com/dmlc/dgl.git # timeout=10
 > git config --add remote.origin.fetch +refs/pull/4648/head:refs/remotes/origin/PR-4648 # timeout=10
 > git config --add remote.origin.fetch +refs/heads/master:refs/remotes/origin/master # timeout=10
 > git config remote.origin.url https://github.com/dmlc/dgl.git # timeout=10
 > git rev-parse --verify HEAD # timeout=10
No valid HEAD. Skipping the resetting
 > git clean -fdx # timeout=10
Fetching upstream changes from https://github.com/dmlc/dgl.git
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git fetch --no-tags --progress -- https://github.com/dmlc/dgl.git +refs/pull/4648/head:refs/remotes/origin/PR-4648 +refs/heads/master:refs/remotes/origin/master # timeout=10
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-e-shp2-g0] Merge succeeded, producing 37ed78035f406940ed88a4b60cadd2b7cfde5fea
Checking out Revision 37ed78035f406940ed88a4b60cadd2b7cfde5fea (PR-4648)
PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-e-shp2-g1] Commit message: "fix for pytorch < 1.12"
Cleaning workspace
PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-e-shp3-g0] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-e-shp3-g1] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-e-shp4-g0] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-e-shp4-g1] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-u-shp0-g0]  > git remote # timeout=10
 > git config --get remote.origin.url # timeout=10
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git merge d78a3a4baf611b90871a849f58647160c7cd9ab4 # timeout=10
 > git rev-parse HEAD^{commit} # timeout=10
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git branch -a -v --no-abbrev # timeout=10
 > git checkout -b PR-4648 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git rev-parse --verify HEAD # timeout=10
Resetting working tree
 > git reset --hard # timeout=10
 > git clean -fdx # timeout=10
[Pipeline] withEnv
[Pipeline] {
[Pipeline] isUnix
[Pipeline] withEnv
[Pipeline] {
[Pipeline] sh
PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-u-shp0-g1] + docker pull dgllib/dgl-ci-cpu:v220816
PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-u-shp1-g0] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-u-shp1-g1] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-u-shp2-g0] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-u-shp2-g1] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-u-shp3-g0] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-u-shp3-g1] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-u-shp4-g0] v220816: Pulling from dgllib/dgl-ci-cpu
Digest: sha256:64b385c33b44dc57cb96ff264a84d8dfb8ced0caa9b30fbc4cec6d5ee511b099
Status: Image is up to date for dgllib/dgl-ci-cpu:v220816
docker.io/dgllib/dgl-ci-cpu:v220816
PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-u-shp4-g1] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-v-shp0-g0] [Pipeline] }
[Pipeline] // withEnv
[Pipeline] isUnix
[Pipeline] withEnv
[Pipeline] {
[Pipeline] sh
PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-v-shp0-g1] + docker inspect -f . dgllib/dgl-ci-cpu:v220816
.
[Pipeline] }
[Pipeline] // withEnv
[Pipeline] withDockerContainer
PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-v-shp1-g0] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-v-shp1-g1] dgl-manual-large-cpu does not seem to be running inside a container
$ docker run -t -d -u 0:0 --shm-size=4gb -w /root/jenkins/workspace/dgl_PR-4648@3 -v /root/jenkins/workspace/dgl_PR-4648@3:/root/jenkins/workspace/dgl_PR-4648@3:rw,z -v /root/jenkins/workspace/dgl_PR-4648@3@tmp:/root/jenkins/workspace/dgl_PR-4648@3@tmp:rw,z -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** dgllib/dgl-ci-cpu:v220816 cat
PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-v-shp2-g0] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-v-shp2-g1] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-v-shp3-g0] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-v-shp3-g1] $ docker top af5aa5eeb7433ab669abbd955ede40baf0b228908472383bbed72d33561f7092 -eo pid,comm
PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-v-shp4-g0] [Pipeline] {
[Pipeline] stage
[Pipeline] { (Torch CPU Unit test)
[Pipeline] sh
PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-v-shp4-g1] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-e-shp0-g0] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-e-shp0-g1] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-e-shp1-g0] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-e-shp1-g1] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-e-shp2-g0] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-e-shp2-g1] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-e-shp3-g0] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-e-shp3-g1] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-e-shp4-g0] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-e-shp4-g1] PASSED [ 91%]
tests/compute/test_sparse.py::test_edge_softmax[idtype0-shp0-src-g0] + rm -rf CMakeLists.txt CONTRIBUTORS.md Jenkinsfile LICENSE NEWS.md README.md apps benchmarks cmake conda dglgo docker docs examples featgraph include pyproject.toml python readthedocs.yml src tensoradapter tests third_party tools tutorials
PASSED [ 91%]
tests/compute/test_sparse.py::test_edge_softmax[idtype0-shp0-dst-g0] [Pipeline] checkout
The recommended git tool is: git
using credential 150de63f-189c-4717-bcaf-010460d2f51a
Warning: JENKINS-30600: special launcher org.jenkinsci.plugins.docker.workflow.WithContainerStep$Decorator$1@78894af6; decorates RemoteLauncher[hudson.remoting.Channel@1d9b9638:dgl-manual-large-cpu] will be ignored (a typical symptom is the Git executable not being run inside a designated container)
Fetching changes from the remote Git repository
Cleaning workspace
PASSED [ 91%]
tests/compute/test_sparse.py::test_edge_softmax[idtype0-shp1-src-g0] PASSED [ 91%]
tests/compute/test_sparse.py::test_edge_softmax[idtype0-shp1-dst-g0] Fetching without tags
PASSED [ 91%]
tests/compute/test_sparse.py::test_edge_softmax[idtype0-shp2-src-g0] Merging remotes/origin/master commit d78a3a4baf611b90871a849f58647160c7cd9ab4 into PR head commit 37ed78035f406940ed88a4b60cadd2b7cfde5fea
PASSED [ 91%]
tests/compute/test_sparse.py::test_edge_softmax[idtype0-shp2-dst-g0] Merge succeeded, producing 37ed78035f406940ed88a4b60cadd2b7cfde5fea
Checking out Revision 37ed78035f406940ed88a4b60cadd2b7cfde5fea (PR-4648)
Commit message: "fix for pytorch < 1.12"
Cleaning workspace
PASSED [ 91%]
tests/compute/test_sparse.py::test_edge_softmax[idtype1-shp0-src-g0] PASSED [ 92%]
tests/compute/test_sparse.py::test_edge_softmax[idtype1-shp0-dst-g0]  > git rev-parse --resolve-git-dir /root/jenkins/workspace/dgl_PR-4648@3/.git # timeout=10
 > git config remote.origin.url https://github.com/dmlc/dgl.git # timeout=10
 > git rev-parse --verify HEAD # timeout=10
Resetting working tree
 > git reset --hard # timeout=10
 > git clean -fdx # timeout=10
Fetching upstream changes from https://github.com/dmlc/dgl.git
 > git --version # timeout=10
 > git --version # 'git version 2.17.1'
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git fetch --no-tags --progress -- https://github.com/dmlc/dgl.git +refs/pull/4648/head:refs/remotes/origin/PR-4648 +refs/heads/master:refs/remotes/origin/master # timeout=10
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git remote # timeout=10
 > git config --get remote.origin.url # timeout=10
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git merge d78a3a4baf611b90871a849f58647160c7cd9ab4 # timeout=10
 > git rev-parse HEAD^{commit} # timeout=10
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git branch -a -v --no-abbrev # timeout=10
 > git branch -D PR-4648 # timeout=10
 > git checkout -b PR-4648 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git rev-parse --verify HEAD # timeout=10
Resetting working tree
 > git reset --hard # timeout=10
 > git clean -fdx # timeout=10
PASSED [ 92%]
tests/compute/test_sparse.py::test_edge_softmax[idtype1-shp1-src-g0] Cloning into '/root/jenkins/workspace/dgl_PR-4648@2/third_party/xbyak'...
Submodule path 'third_party/METIS': checked out '10603482cc055626ff99d2ac3ab7e73d1119439d'
Submodule 'GKlib' (https://github.com/KarypisLab/GKlib.git) registered for path 'third_party/METIS/GKlib'
Cloning into '/root/jenkins/workspace/dgl_PR-4648@2/third_party/METIS/GKlib'...
PASSED [ 92%]
tests/compute/test_sparse.py::test_edge_softmax[idtype1-shp1-dst-g0] PASSED [ 92%]
tests/compute/test_sparse.py::test_edge_softmax[idtype1-shp2-src-g0] [Pipeline] sh
PASSED [ 92%]
tests/compute/test_sparse.py::test_edge_softmax[idtype1-shp2-dst-g0] Submodule path 'third_party/METIS/GKlib': checked out '62de20c7f97c719abdc96d7c3e266a1f7cb52bc5'
Submodule path 'third_party/dlpack': checked out 'e2bdd3bee8cb6501558042633fa59144cc8b7f5f'
Submodule path 'third_party/dmlc-core': checked out 'bfad207b448480783a1f428ae3d93d87032d8349'
PASSED [ 92%]
tests/compute/test_sparse.py::test_segment_reduce[sum] PASSED            [ 92%]
tests/compute/test_sparse.py::test_segment_reduce[max] Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/xbyak'...
Submodule path 'third_party/googletest': checked out 'f71fb4f9a912ec945401cc49a287a759b6131026'
+ git submodule update --recursive --init
Submodule 'third_party/METIS' (https://github.com/KarypisLab/METIS.git) registered for path 'third_party/METIS'
Submodule 'third_party/dlpack' (https://github.com/dmlc/dlpack.git) registered for path 'third_party/dlpack'
Submodule 'third_party/dmlc-core' (https://github.com/dmlc/dmlc-core.git) registered for path 'third_party/dmlc-core'
Submodule 'third_party/googletest' (https://github.com/google/googletest.git) registered for path 'third_party/googletest'
Submodule 'third_party/libxsmm' (https://github.com/hfp/libxsmm.git) registered for path 'third_party/libxsmm'
Submodule 'third_party/nanoflann' (https://github.com/jlblancoc/nanoflann) registered for path 'third_party/nanoflann'
Submodule 'third_party/nccl' (https://github.com/nvidia/nccl) registered for path 'third_party/nccl'
Submodule 'third_party/phmap' (https://github.com/greg7mdp/parallel-hashmap.git) registered for path 'third_party/phmap'
Submodule 'third_party/tensorpipe' (https://github.com/pytorch/tensorpipe) registered for path 'third_party/tensorpipe'
Submodule 'third_party/thrust' (https://github.com/NVIDIA/thrust.git) registered for path 'third_party/thrust'
Submodule 'third_party/tvm' (https://github.com/apache/incubator-tvm) registered for path 'third_party/tvm'
Submodule 'third_party/xbyak' (https://github.com/herumi/xbyak) registered for path 'third_party/xbyak'
Cloning into '/root/jenkins/workspace/dgl_PR-4648@3/third_party/METIS'...
PASSED            [ 92%]
tests/compute/test_sparse.py::test_segment_reduce[min] Submodule path 'third_party/METIS': checked out '10603482cc055626ff99d2ac3ab7e73d1119439d'
Submodule 'GKlib' (https://github.com/KarypisLab/GKlib.git) registered for path 'third_party/METIS/GKlib'
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/METIS/GKlib'...
PASSED            [ 92%]
tests/compute/test_sparse.py::test_segment_reduce[mean] Cloning into '/root/jenkins/workspace/dgl_PR-4648@3/third_party/dlpack'...
PASSED           [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype0-0.01-1-idtype0] SKIPPED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype0-0.01-1-idtype1] SKIPPED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype0-0.01-8-idtype0] SKIPPED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype0-0.01-8-idtype1] SKIPPED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype0-0.01-16-idtype0] SKIPPED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype0-0.01-16-idtype1] SKIPPED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype0-0.01-64-idtype0] SKIPPED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype0-0.01-64-idtype1] SKIPPED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype0-0.01-256-idtype0] SKIPPED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype0-0.01-256-idtype1] SKIPPED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype1-0.01-1-idtype0] SKIPPED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype1-0.01-1-idtype1] SKIPPED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype1-0.01-8-idtype0] SKIPPED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype1-0.01-8-idtype1] SKIPPED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype1-0.01-16-idtype0] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype1-0.01-16-idtype1] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype1-0.01-64-idtype0] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype1-0.01-64-idtype1] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype1-0.01-256-idtype0] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype1-0.01-256-idtype1] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype2-0.003-1-idtype0] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype2-0.003-1-idtype1] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype2-0.003-8-idtype0] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype2-0.003-8-idtype1] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype2-0.003-16-idtype0] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype2-0.003-16-idtype1] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype2-0.003-64-idtype0] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype2-0.003-64-idtype1] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype2-0.003-256-idtype0] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype2-0.003-256-idtype1] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype3-0.0001-1-idtype0] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype3-0.0001-1-idtype1] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype3-0.0001-8-idtype0] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype3-0.0001-8-idtype1] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype3-0.0001-16-idtype0] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype3-0.0001-16-idtype1] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype3-0.0001-64-idtype0] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype3-0.0001-64-idtype1] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype3-0.0001-256-idtype0] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype3-0.0001-256-idtype1] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype0-0.01-1] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype0-0.01-8] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype0-0.01-16] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype0-0.01-64] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype0-0.01-256] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype1-0.02-1] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype1-0.02-8] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype1-0.02-16] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype1-0.02-64] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype1-0.02-256] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype2-0.003-1] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype2-0.003-8] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype2-0.003-16] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype2-0.003-64] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype2-0.003-256] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype3-0.0001-1] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype3-0.0001-8] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype3-0.0001-16] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype3-0.0001-64] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype3-0.0001-256] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_use_libxsmm_switch SKIPPED (Only ...) [ 94%]
tests/compute/test_specialization.py::test_v2v_update_all[idtype0] PASSED [ 94%]
tests/compute/test_specialization.py::test_v2v_update_all[idtype1] PASSED [ 94%]
tests/compute/test_specialization.py::test_v2v_snr[idtype0] Submodule path 'third_party/METIS/GKlib': checked out '62de20c7f97c719abdc96d7c3e266a1f7cb52bc5'
Submodule path 'third_party/dlpack': checked out 'e2bdd3bee8cb6501558042633fa59144cc8b7f5f'
PASSED       [ 94%]
tests/compute/test_specialization.py::test_v2v_snr[idtype1] PASSED       [ 95%]
tests/compute/test_specialization.py::test_v2v_pull[idtype0] PASSED      [ 95%]
tests/compute/test_specialization.py::test_v2v_pull[idtype1] Submodule path 'third_party/libxsmm': checked out 'fa687556130b6298430f1c0555a14cf79ab6101c'
Submodule path 'third_party/dmlc-core': checked out 'bfad207b448480783a1f428ae3d93d87032d8349'
Cloning into '/root/jenkins/workspace/dgl_PR-4648@3/third_party/dmlc-core'...
PASSED      [ 95%]
tests/compute/test_specialization.py::test_update_all_multi_fallback[idtype0] PASSED [ 95%]
tests/compute/test_specialization.py::test_update_all_multi_fallback[idtype1] PASSED [ 95%]
tests/compute/test_specialization.py::test_pull_multi_fallback[idtype0] PASSED [ 95%]
tests/compute/test_specialization.py::test_pull_multi_fallback[idtype1] Submodule path 'third_party/nanoflann': checked out '4c47ca200209550c5628c89803591f8a753c8181'
Submodule path 'third_party/nccl': checked out 'e11238b3029795d33f958b5868d47c90c4f22628'
Submodule path 'third_party/phmap': checked out '25293cefd8b85491b45600c03fe8edf07647553f'
Submodule path 'third_party/googletest': checked out 'f71fb4f9a912ec945401cc49a287a759b6131026'
PASSED [ 95%]
tests/compute/test_specialization.py::test_spmv_3d_feat[idtype0] PASSED  [ 95%]
tests/compute/test_specialization.py::test_spmv_3d_feat[idtype1] PASSED  [ 95%]
tests/compute/test_subgraph.py::test_edge_subgraph PASSED                [ 95%]
tests/compute/test_subgraph.py::test_subgraph PASSED                     [ 95%]
tests/compute/test_subgraph.py::test_subgraph_mask[idtype0] PASSED       [ 95%]
tests/compute/test_subgraph.py::test_subgraph_mask[idtype1] PASSED       [ 95%]
tests/compute/test_subgraph.py::test_subgraph1[idtype0] Submodule path 'third_party/tensorpipe': checked out '6042f1a4cbce8eef997f11ed0012de137b317361'
Submodule 'third_party/googletest' (https://github.com/google/googletest.git) registered for path 'third_party/tensorpipe/third_party/googletest'
Submodule 'third_party/libnop' (https://github.com/google/libnop.git) registered for path 'third_party/tensorpipe/third_party/libnop'
Submodule 'third_party/libuv' (https://github.com/libuv/libuv.git) registered for path 'third_party/tensorpipe/third_party/libuv'
Submodule 'third_party/pybind11' (https://github.com/pybind/pybind11.git) registered for path 'third_party/tensorpipe/third_party/pybind11'
Cloning into '/root/jenkins/workspace/dgl_PR-4648@2/third_party/tensorpipe/third_party/googletest'...
PASSED           [ 95%]
tests/compute/test_subgraph.py::test_subgraph1[idtype1] PASSED           [ 95%]
tests/compute/test_subgraph.py::test_in_subgraph[idtype0] PASSED         [ 95%]
tests/compute/test_subgraph.py::test_in_subgraph[idtype1] PASSED         [ 95%]
tests/compute/test_subgraph.py::test_out_subgraph[idtype0] PASSED        [ 95%]
tests/compute/test_subgraph.py::test_out_subgraph[idtype1] PASSED        [ 95%]
tests/compute/test_subgraph.py::test_subgraph_message_passing PASSED     [ 95%]
tests/compute/test_subgraph.py::test_khop_in_subgraph[idtype0] PASSED    [ 95%]
tests/compute/test_subgraph.py::test_khop_in_subgraph[idtype1] PASSED    [ 95%]
tests/compute/test_subgraph.py::test_khop_out_subgraph[idtype0] Cloning into '/root/jenkins/workspace/dgl_PR-4648@3/third_party/googletest'...
PASSED   [ 95%]
tests/compute/test_subgraph.py::test_khop_out_subgraph[idtype1] PASSED   [ 95%]
tests/compute/test_subgraph.py::test_subframes[/cpu:0-parent_idx_device0] PASSED [ 96%]
tests/compute/test_subgraph.py::test_subframes[/cpu:0-parent_idx_device1] PASSED [ 96%]
tests/compute/test_subgraph.py::test_subframes[/cpu:0-parent_idx_device2] SKIPPED [ 96%]
tests/compute/test_subgraph.py::test_subframes[/cpu:0-parent_idx_device3] SKIPPED [ 96%]
tests/compute/test_subgraph.py::test_subframes[/gpu:0-parent_idx_device0] PASSED [ 96%]
tests/compute/test_subgraph.py::test_subframes[/gpu:0-parent_idx_device1] PASSED [ 96%]
tests/compute/test_subgraph.py::test_subframes[/gpu:0-parent_idx_device2] SKIPPED [ 96%]
tests/compute/test_subgraph.py::test_subframes[/gpu:0-parent_idx_device3] SKIPPED [ 96%]
tests/compute/test_subgraph.py::test_uva_subgraph[idtype0-/cpu:0] SKIPPED [ 96%]
tests/compute/test_subgraph.py::test_uva_subgraph[idtype0-/gpu:0] SKIPPED [ 96%]
tests/compute/test_subgraph.py::test_uva_subgraph[idtype1-/cpu:0] SKIPPED [ 96%]
tests/compute/test_subgraph.py::test_uva_subgraph[idtype1-/gpu:0] SKIPPED [ 96%]
tests/compute/test_transform.py::test_line_graph1 PASSED                 [ 96%]
tests/compute/test_transform.py::test_line_graph2[idtype0] PASSED        [ 96%]
tests/compute/test_transform.py::test_line_graph2[idtype1] PASSED        [ 96%]
tests/compute/test_transform.py::test_no_backtracking PASSED             [ 96%]
tests/compute/test_transform.py::test_reverse[idtype0] PASSED            [ 96%]
tests/compute/test_transform.py::test_reverse[idtype1] PASSED            [ 96%]
tests/compute/test_transform.py::test_reverse_shared_frames[idtype0] PASSED [ 96%]
tests/compute/test_transform.py::test_reverse_shared_frames[idtype1] PASSED [ 96%]
tests/compute/test_transform.py::test_to_bidirected SKIPPED (GPU not...) [ 96%]
tests/compute/test_transform.py::test_add_reverse_edges PASSED           [ 96%]
tests/compute/test_transform.py::test_simple_graph SKIPPED (GPU not ...) [ 96%]
tests/compute/test_transform.py::test_khop_graph SKIPPED (GPU not im...) [ 96%]
tests/compute/test_transform.py::test_khop_adj SKIPPED (GPU not impl...) [ 96%]
tests/compute/test_transform.py::test_laplacian_lambda_max SKIPPED (...) [ 97%]
tests/compute/test_transform.py::test_partition_with_halo SKIPPED (G...) [ 97%]
tests/compute/test_transform.py::test_metis_partition[idtype0] SKIPPED   [ 97%]
tests/compute/test_transform.py::test_metis_partition[idtype1] SKIPPED   [ 97%]
tests/compute/test_transform.py::test_reorder_nodes SKIPPED (It does...) [ 97%]
tests/compute/test_transform.py::test_compact[idtype0] PASSED            [ 97%]
tests/compute/test_transform.py::test_compact[idtype1] PASSED            [ 97%]
tests/compute/test_transform.py::test_to_simple[idtype0] SKIPPED (GP...) [ 97%]
tests/compute/test_transform.py::test_to_simple[idtype1] SKIPPED (GP...) [ 97%]
tests/compute/test_transform.py::test_to_block[idtype0] PASSED           [ 97%]
tests/compute/test_transform.py::test_to_block[idtype1] PASSED           [ 97%]
tests/compute/test_transform.py::test_remove_edges[idtype0] PASSED       [ 97%]
tests/compute/test_transform.py::test_remove_edges[idtype1] Submodule path 'third_party/libxsmm': checked out 'fa687556130b6298430f1c0555a14cf79ab6101c'
Submodule path 'third_party/nanoflann': checked out '4c47ca200209550c5628c89803591f8a753c8181'
Submodule path 'third_party/nccl': checked out 'e11238b3029795d33f958b5868d47c90c4f22628'
PASSED       [ 97%]
tests/compute/test_transform.py::test_add_edges[idtype0] PASSED          [ 97%]
tests/compute/test_transform.py::test_add_edges[idtype1] PASSED          [ 97%]
tests/compute/test_transform.py::test_add_nodes[idtype0] PASSED          [ 97%]
tests/compute/test_transform.py::test_add_nodes[idtype1] PASSED          [ 97%]
tests/compute/test_transform.py::test_remove_nodes[idtype0] PASSED       [ 97%]
tests/compute/test_transform.py::test_remove_nodes[idtype1] Submodule path 'third_party/phmap': checked out '25293cefd8b85491b45600c03fe8edf07647553f'
Submodule path 'third_party/tensorpipe': checked out '6042f1a4cbce8eef997f11ed0012de137b317361'
Submodule 'third_party/googletest' (https://github.com/google/googletest.git) registered for path 'third_party/tensorpipe/third_party/googletest'
Submodule 'third_party/libnop' (https://github.com/google/libnop.git) registered for path 'third_party/tensorpipe/third_party/libnop'
Submodule 'third_party/libuv' (https://github.com/libuv/libuv.git) registered for path 'third_party/tensorpipe/third_party/libuv'
Submodule 'third_party/pybind11' (https://github.com/pybind/pybind11.git) registered for path 'third_party/tensorpipe/third_party/pybind11'
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tensorpipe/third_party/googletest'...
PASSED       [ 97%]
tests/compute/test_transform.py::test_add_selfloop[idtype0] PASSED       [ 97%]
tests/compute/test_transform.py::test_add_selfloop[idtype1] PASSED       [ 97%]
tests/compute/test_transform.py::test_remove_selfloop[idtype0] PASSED    [ 97%]
tests/compute/test_transform.py::test_remove_selfloop[idtype1] PASSED    [ 97%]
tests/compute/test_transform.py::test_reorder_graph[idtype0] PASSED      [ 97%]
tests/compute/test_transform.py::test_reorder_graph[idtype1] PASSED      [ 98%]
tests/compute/test_transform.py::test_norm_by_dst[idtype0] SKIPPED (...) [ 98%]
tests/compute/test_transform.py::test_norm_by_dst[idtype1] SKIPPED (...) [ 98%]
tests/compute/test_transform.py::test_module_add_self_loop[idtype0] PASSED [ 98%]
tests/compute/test_transform.py::test_module_add_self_loop[idtype1] PASSED [ 98%]
tests/compute/test_transform.py::test_module_remove_self_loop[idtype0] PASSED [ 98%]
tests/compute/test_transform.py::test_module_remove_self_loop[idtype1] PASSED [ 98%]
tests/compute/test_transform.py::test_module_add_reverse[idtype0] PASSED [ 98%]
tests/compute/test_transform.py::test_module_add_reverse[idtype1] PASSED [ 98%]
tests/compute/test_transform.py::test_module_to_simple[idtype0] SKIPPED  [ 98%]
tests/compute/test_transform.py::test_module_to_simple[idtype1] SKIPPED  [ 98%]
tests/compute/test_transform.py::test_module_line_graph[idtype0] PASSED  [ 98%]
tests/compute/test_transform.py::test_module_line_graph[idtype1] PASSED  [ 98%]
tests/compute/test_transform.py::test_module_khop_graph[idtype0] PASSED  [ 98%]
tests/compute/test_transform.py::test_module_khop_graph[idtype1] PASSED  [ 98%]
tests/compute/test_transform.py::test_module_add_metapaths[idtype0] PASSED [ 98%]
tests/compute/test_transform.py::test_module_add_metapaths[idtype1] PASSED [ 98%]
tests/compute/test_transform.py::test_module_compose[idtype0] PASSED     [ 98%]
tests/compute/test_transform.py::test_module_compose[idtype1] PASSED     [ 98%]
tests/compute/test_transform.py::test_module_gcnnorm[idtype0] PASSED     [ 98%]
tests/compute/test_transform.py::test_module_gcnnorm[idtype1] PASSED     [ 98%]
tests/compute/test_transform.py::test_module_ppr[idtype0] SKIPPED (O...) [ 98%]
tests/compute/test_transform.py::test_module_ppr[idtype1] SKIPPED (O...) [ 98%]
tests/compute/test_transform.py::test_module_heat_kernel[idtype0] SKIPPED [ 98%]
tests/compute/test_transform.py::test_module_heat_kernel[idtype1] SKIPPED [ 98%]
tests/compute/test_transform.py::test_module_gdc[idtype0] SKIPPED (O...) [ 99%]
tests/compute/test_transform.py::test_module_gdc[idtype1] SKIPPED (O...) [ 99%]
tests/compute/test_transform.py::test_module_node_shuffle[idtype0] PASSED [ 99%]
tests/compute/test_transform.py::test_module_node_shuffle[idtype1] PASSED [ 99%]
tests/compute/test_transform.py::test_module_drop_node[idtype0] SKIPPED  [ 99%]
tests/compute/test_transform.py::test_module_drop_node[idtype1] SKIPPED  [ 99%]
tests/compute/test_transform.py::test_module_drop_edge[idtype0] SKIPPED  [ 99%]
tests/compute/test_transform.py::test_module_drop_edge[idtype1] SKIPPED  [ 99%]
tests/compute/test_transform.py::test_module_add_edge[idtype0] PASSED    [ 99%]
tests/compute/test_transform.py::test_module_add_edge[idtype1] PASSED    [ 99%]
tests/compute/test_transform.py::test_module_random_walk_pe[idtype0] PASSED [ 99%]
tests/compute/test_transform.py::test_module_random_walk_pe[idtype1] PASSED [ 99%]
tests/compute/test_transform.py::test_module_laplacian_pe[idtype0] PASSED [ 99%]
tests/compute/test_transform.py::test_module_laplacian_pe[idtype1] PASSED [ 99%]
tests/compute/test_transform.py::test_module_sign[g0] SKIPPED (Only ...) [ 99%]
tests/compute/test_transform.py::test_module_row_feat_normalizer[idtype0] SKIPPED [ 99%]
tests/compute/test_transform.py::test_module_row_feat_normalizer[idtype1] SKIPPED [ 99%]
tests/compute/test_transform.py::test_module_feat_mask[idtype0] SKIPPED  [ 99%]
tests/compute/test_transform.py::test_module_feat_mask[idtype1] SKIPPED  [ 99%]
tests/compute/test_traversal.py::test_bfs[idtype0] Cloning into '/root/jenkins/workspace/dgl_PR-4648@3/third_party/libxsmm'...
PASSED                [ 99%]
tests/compute/test_traversal.py::test_bfs[idtype1] Cloning into '/root/jenkins/workspace/dgl_PR-4648@2/third_party/tensorpipe/third_party/libnop'...
PASSED                [ 99%]
tests/compute/test_traversal.py::test_topological_nodes[idtype0] PASSED  [ 99%]
tests/compute/test_traversal.py::test_topological_nodes[idtype1] PASSED  [ 99%]
tests/compute/test_traversal.py::test_dfs_labeled_edges[idtype0] PASSED  [ 99%]
tests/compute/test_traversal.py::test_dfs_labeled_edges[idtype1] PASSED  [100%]

=============================== warnings summary ===============================
python/dgl/backend/backend.py:1741
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/backend/backend.py:1741: DeprecationWarning: invalid escape sequence \P
    """

../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/tensorflow/python/data/ops/iterator_ops.py:546
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/tensorflow/python/data/ops/iterator_ops.py:546: DeprecationWarning: Using or importing the ABCs from 'collections' instead of from 'collections.abc' is deprecated, and in 3.8 it will stop working
    class IteratorBase(collections.Iterator, trackable.Trackable,

../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/tensorflow/python/data/ops/dataset_ops.py:106
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/tensorflow/python/data/ops/dataset_ops.py:106: DeprecationWarning: Using or importing the ABCs from 'collections' instead of from 'collections.abc' is deprecated, and in 3.8 it will stop working
    class DatasetV2(collections.Iterable, tracking_base.Trackable,

../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/tensorflow/python/autograph/utils/testing.py:21
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/tensorflow/python/autograph/utils/testing.py:21: DeprecationWarning: the imp module is deprecated in favour of importlib; see the module's documentation for alternative uses
    import imp

../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:10
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:10: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    _nlv = LooseVersion(_np_version)

../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:11
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:11: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    _np_version_under1p16 = _nlv < LooseVersion("1.16")

../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:12
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:12: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    _np_version_under1p17 = _nlv < LooseVersion("1.17")

../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:13
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:13: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    _np_version_under1p18 = _nlv < LooseVersion("1.18")

../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:14
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:14: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    _np_version_under1p19 = _nlv < LooseVersion("1.19")

../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:15
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:15: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    _np_version_under1p20 = _nlv < LooseVersion("1.20")

../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/setuptools/_distutils/version.py:351
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/setuptools/_distutils/version.py:351: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    other = LooseVersion(other)

../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/pandas/compat/numpy/function.py:125
../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/pandas/compat/numpy/function.py:125
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/pandas/compat/numpy/function.py:125: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    if LooseVersion(_np_version) >= LooseVersion("1.17.0"):

../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/keras_preprocessing/image/utils.py:23
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/keras_preprocessing/image/utils.py:23: DeprecationWarning: NEAREST is deprecated and will be removed in Pillow 10 (2023-07-01). Use Resampling.NEAREST or Dither.NONE instead.
    'nearest': pil_image.NEAREST,

../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/keras_preprocessing/image/utils.py:24
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/keras_preprocessing/image/utils.py:24: DeprecationWarning: BILINEAR is deprecated and will be removed in Pillow 10 (2023-07-01). Use Resampling.BILINEAR instead.
    'bilinear': pil_image.BILINEAR,

../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/keras_preprocessing/image/utils.py:25
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/keras_preprocessing/image/utils.py:25: DeprecationWarning: BICUBIC is deprecated and will be removed in Pillow 10 (2023-07-01). Use Resampling.BICUBIC instead.
    'bicubic': pil_image.BICUBIC,

../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/keras_preprocessing/image/utils.py:28
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/keras_preprocessing/image/utils.py:28: DeprecationWarning: HAMMING is deprecated and will be removed in Pillow 10 (2023-07-01). Use Resampling.HAMMING instead.
    if hasattr(pil_image, 'HAMMING'):

../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/keras_preprocessing/image/utils.py:29
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/keras_preprocessing/image/utils.py:29: DeprecationWarning: HAMMING is deprecated and will be removed in Pillow 10 (2023-07-01). Use Resampling.HAMMING instead.
    _PIL_INTERPOLATION_METHODS['hamming'] = pil_image.HAMMING

../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/keras_preprocessing/image/utils.py:30
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/keras_preprocessing/image/utils.py:30: DeprecationWarning: BOX is deprecated and will be removed in Pillow 10 (2023-07-01). Use Resampling.BOX instead.
    if hasattr(pil_image, 'BOX'):

../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/keras_preprocessing/image/utils.py:31
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/keras_preprocessing/image/utils.py:31: DeprecationWarning: BOX is deprecated and will be removed in Pillow 10 (2023-07-01). Use Resampling.BOX instead.
    _PIL_INTERPOLATION_METHODS['box'] = pil_image.BOX

../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/keras_preprocessing/image/utils.py:33
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/keras_preprocessing/image/utils.py:33: DeprecationWarning: LANCZOS is deprecated and will be removed in Pillow 10 (2023-07-01). Use Resampling.LANCZOS instead.
    if hasattr(pil_image, 'LANCZOS'):

../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/keras_preprocessing/image/utils.py:34
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/keras_preprocessing/image/utils.py:34: DeprecationWarning: LANCZOS is deprecated and will be removed in Pillow 10 (2023-07-01). Use Resampling.LANCZOS instead.
    _PIL_INTERPOLATION_METHODS['lanczos'] = pil_image.LANCZOS

python/dgl/backend/tensorflow/tensor.py:15
python/dgl/backend/tensorflow/tensor.py:15
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/backend/tensorflow/tensor.py:15: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    if LooseVersion(tf.__version__) < LooseVersion("2.3.0"):

tests/compute/test_basics.py: 2 warnings
tests/compute/test_filter.py: 1 warning
tests/compute/test_graph.py: 9 warnings
tests/compute/test_kernel.py: 3 warnings
tests/compute/test_removal.py: 16 warnings
tests/compute/test_specialization.py: 12 warnings
tests/compute/test_subgraph.py: 2 warnings
tests/compute/test_transform.py: 6 warnings
tests/compute/test_traversal.py: 2 warnings
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/heterograph.py:72: DGLWarning: Recommend creating graphs by `dgl.graph(data)` instead of `dgl.DGLGraph(data)`.
    dgl_warning('Recommend creating graphs by `dgl.graph(data)`'

tests/compute/test_basics.py: 2 warnings
tests/compute/test_batched_graph.py: 8 warnings
tests/compute/test_graph.py: 2 warnings
tests/compute/test_kernel.py: 1 warning
tests/compute/test_removal.py: 10 warnings
tests/compute/test_specialization.py: 10 warnings
tests/compute/test_subgraph.py: 2 warnings
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/heterograph.py:354: DGLWarning: DGLGraph.add_edge is deprecated. Please use DGLGraph.add_edges
    dgl_warning("DGLGraph.add_edge is deprecated. Please use DGLGraph.add_edges")

tests/compute/test_basics.py::test_update_all_0deg[idtype0]
tests/compute/test_basics.py::test_update_all_0deg[idtype1]
tests/compute/test_basics.py::test_pull_0deg[idtype0]
tests/compute/test_basics.py::test_pull_0deg[idtype1]
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/core.py:79: DGLWarning: The input graph for the user-defined edge function does not contain valid edges
    dgl_warning('The input graph for the user-defined edge function ' \

tests/compute/test_batched_graph.py::test_batched_edge_ordering[idtype0]
tests/compute/test_batched_graph.py::test_batched_edge_ordering[idtype1]
tests/compute/test_graph.py::test_query
tests/compute/test_graph.py::test_hypersparse_query
tests/compute/test_transform.py::test_no_backtracking
tests/compute/test_transform.py::test_reverse[idtype0]
tests/compute/test_transform.py::test_reverse[idtype1]
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/heterograph.py:2978: DGLWarning: DGLGraph.edge_id is deprecated. Please use DGLGraph.edge_ids.
    dgl_warning("DGLGraph.edge_id is deprecated. Please use DGLGraph.edge_ids.")

tests/compute/test_batched_heterograph.py::test_features[idtype0]
tests/compute/test_batched_heterograph.py::test_features[idtype1]
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/batch.py:159: DGLWarning: Arguments edge_attrs has been deprecated. Please use edata instead.
    dgl_warning('Arguments edge_attrs has been deprecated. Please use'

tests/compute/test_csrmm.py::test_csrmm[dtype0-idtype0]
tests/compute/test_csrmm.py::test_csrmm[dtype0-idtype1]
tests/compute/test_csrmm.py::test_csrmm[dtype1-idtype0]
tests/compute/test_csrmm.py::test_csrmm[dtype1-idtype1]
tests/compute/test_csrmm.py::test_csrsum[dtype0-idtype0]
tests/compute/test_csrmm.py::test_csrsum[dtype0-idtype1]
tests/compute/test_csrmm.py::test_csrsum[dtype1-idtype0]
tests/compute/test_csrmm.py::test_csrsum[dtype1-idtype1]
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/heterograph_index.py:797: FutureWarning: Adjacency matrix by default currently returns edge IDs.  As a result there is one 0 entry which is not eliminated.  In the next release it will return 1s by default, and 0 will be eliminated otherwise.
    FutureWarning)

tests/compute/test_graph.py::test_query
tests/compute/test_graph.py::test_hypersparse_query
tests/compute/test_heterograph.py::test_query[idtype0]
tests/compute/test_heterograph.py::test_query[idtype1]
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/heterograph.py:2753: DGLWarning: DGLGraph.has_node is deprecated. Please use DGLGraph.has_nodes
    dgl_warning("DGLGraph.has_node is deprecated. Please use DGLGraph.has_nodes")

tests/compute/test_graph.py::test_query
tests/compute/test_graph.py::test_hypersparse_query
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/heterograph.py:2687: DGLWarning: DGLGraph.__contains__ is deprecated. Please directly call has_nodes.
    dgl_warning('DGLGraph.__contains__ is deprecated.'

tests/compute/test_graph.py::test_query
tests/compute/test_sampling.py::test_non_uniform_random_walk[False]
tests/compute/test_sampling.py::test_uniform_random_walk[True]
tests/compute/test_sampling.py::test_uniform_random_walk[False]
tests/compute/test_transform.py::test_no_backtracking
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/heterograph.py:2851: DGLWarning: DGLGraph.has_edge_between is deprecated. Please use DGLGraph.has_edges_between
    dgl_warning("DGLGraph.has_edge_between is deprecated. "

tests/compute/test_graph.py::test_query
tests/compute/test_graph.py::test_hypersparse_query
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/heterograph.py:3432: DGLWarning: DGLGraph.in_degree is deprecated. Please use DGLGraph.in_degrees
    dgl_warning("DGLGraph.in_degree is deprecated. Please use DGLGraph.in_degrees")

tests/compute/test_graph.py::test_query
tests/compute/test_graph.py::test_hypersparse_query
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/heterograph.py:3516: DGLWarning: DGLGraph.out_degree is deprecated. Please use DGLGraph.out_degrees
    dgl_warning("DGLGraph.out_degree is deprecated. Please use DGLGraph.out_degrees")

tests/compute/test_graph.py::test_query
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/heterograph.py:84: DGLWarning: Keyword arguments ['readonly', 'sort_csr'] are deprecated in v0.5, and can be safely removed in all cases.
    ' removed in all cases.'.format(list(deprecate_kwargs.keys())))

tests/compute/test_heterograph.py: 20 warnings
  /root/jenkins/workspace/dgl_PR-4648/tests/compute/test_heterograph.py:1128: DeprecationWarning: np.asscalar(a) is deprecated since NumPy v1.16, use a.item() instead
    assert np.asscalar(F.asnumpy(src_i)) == nid[src[i]]

tests/compute/test_heterograph.py: 20 warnings
  /root/jenkins/workspace/dgl_PR-4648/tests/compute/test_heterograph.py:1129: DeprecationWarning: np.asscalar(a) is deprecated since NumPy v1.16, use a.item() instead
    assert np.asscalar(F.asnumpy(dst_i)) == nid[dst[i]]

tests/compute/test_heterograph.py::test_invertible_conversion[idtype0]
tests/compute/test_heterograph.py::test_invertible_conversion[idtype1]
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/heterograph.py:2635: DGLWarning: DGLGraph.is_readonly is deprecated in v0.5.
  DGLGraph now always supports mutable operations like add_nodes and add_edges.
    dgl_warning('DGLGraph.is_readonly is deprecated in v0.5.\n'

tests/compute/test_partition.py::test_get_node_partition_from_book[idtype0]
tests/compute/test_partition.py::test_get_node_partition_from_book[idtype1]
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
    "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")

tests/compute/test_sampler.py::test_create_full
tests/compute/test_sampler.py::test_1neighbor_sampler_all
tests/compute/test_sampler.py::test_1neighbor_sampler
tests/compute/test_sampler.py::test_prefetch_neighbor_sampler
tests/compute/test_sampler.py::test_10neighbor_sampler_all
tests/compute/test_sampler.py::test_10neighbor_sampler
tests/compute/test_sampler.py::test_setseed
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/contrib/sampling/sampler.py:317: DGLWarning: dgl.contrib.sampling.NeighborSampler is deprecated starting from v0.5. Please read our guide<link> for how to use the new sampling APIs.
    dgl_warning('dgl.contrib.sampling.NeighborSampler is deprecated starting from v0.5.'

tests/compute/test_sampler.py::test_create_full
tests/compute/test_sampler.py::test_1neighbor_sampler_all
tests/compute/test_sampler.py::test_1neighbor_sampler
tests/compute/test_sampler.py::test_prefetch_neighbor_sampler
tests/compute/test_sampler.py::test_10neighbor_sampler_all
tests/compute/test_sampler.py::test_10neighbor_sampler
tests/compute/test_sampler.py::test_layer_sampler
tests/compute/test_sampler.py::test_setseed
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/_deprecate/nodeflow.py:99: DGLWarning: NodeFlow APIs are deprecated starting from v0.5. Please read our guide<link> for how to use the new sampling APIs.
    dgl_warning('NodeFlow APIs are deprecated starting from v0.5. Please read our'

tests/compute/test_sampling.py::test_global_uniform_negative_sampling[int32]
tests/compute/test_sampling.py::test_global_uniform_negative_sampling[int64]
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/sampling/negative.py:102: ComplexWarning: Casting complex values to real discards the imaginary part
    g._graph, etype_id, num_samples, 3, exclude_self_loops, replace, redundancy)

tests/compute/test_serialize.py::test_load_old_files1
tests/compute/test_serialize.py::test_load_old_files2
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/data/graph_serialize.py:179: DGLWarning: You are loading a graph file saved by old version of dgl.              Please consider saving it again with the current format.
    Please consider saving it again with the current format.")

tests/compute/test_transform.py::test_reverse_shared_frames[idtype0]
tests/compute/test_transform.py::test_reverse_shared_frames[idtype1]
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/transforms/functional.py:1267: DGLWarning: share_ndata argument has been renamed to copy_ndata.
    dgl_warning('share_ndata argument has been renamed to copy_ndata.')

tests/compute/test_transform.py::test_reverse_shared_frames[idtype0]
tests/compute/test_transform.py::test_reverse_shared_frames[idtype1]
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/transforms/functional.py:1270: DGLWarning: share_edata argument has been renamed to copy_edata.
    dgl_warning('share_edata argument has been renamed to copy_edata.')

-- Docs: https://docs.pytest.org/en/stable/how-to/capture-warnings.html
-- generated xml file: /root/jenkins/workspace/dgl_PR-4648/pytest_compute.xml --
============================ slowest 100 durations =============================
30.41s call     tests/compute/test_kernel.py::test_all_binary_builtins
3.52s call     tests/compute/test_heterograph.py::test_forking_pickler
0.97s call     tests/compute/test_sampling.py::test_non_uniform_random_walk[False]
0.84s call     tests/compute/test_graph.py::test_query
0.67s call     tests/compute/test_sampling.py::test_uniform_random_walk[False]
0.59s call     tests/compute/test_heterograph.py::test_query[idtype1]
0.53s call     tests/compute/test_heterograph.py::test_query[idtype0]
0.49s call     tests/compute/test_sampling.py::test_uniform_random_walk[True]
0.43s call     tests/compute/test_sampling.py::test_sample_neighbors_outedge
0.33s call     tests/compute/test_basics.py::test_batch_setter_getter[idtype0]
0.33s call     tests/compute/test_kernel.py::test_copy_edge_reduce
0.31s call     tests/compute/test_kernel.py::test_copy_src_reduce
0.27s call     tests/compute/test_traversal.py::test_bfs[idtype1]
0.27s call     tests/compute/test_traversal.py::test_bfs[idtype0]
0.26s call     tests/compute/test_heterograph.py::test_view1[idtype1]
0.24s call     tests/compute/test_sampler.py::test_10neighbor_sampler
0.24s call     tests/compute/test_sampling.py::test_sample_neighbors_prob
0.24s call     tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-e-shp1-g0]
0.23s call     tests/compute/test_sampler.py::test_prefetch_neighbor_sampler
0.23s call     tests/compute/test_sampler.py::test_1neighbor_sampler
0.23s call     tests/compute/test_sampling.py::test_sample_neighbors_noprob
0.23s call     tests/compute/test_heterograph.py::test_view1[idtype0]
0.22s call     tests/compute/test_basics.py::test_issue_1088[idtype0]
0.18s call     tests/compute/test_sparse.py::test_segment_reduce[mean]
0.18s call     tests/compute/test_sparse.py::test_segment_reduce[min]
0.18s call     tests/compute/test_sparse.py::test_segment_reduce[max]
0.18s call     tests/compute/test_sparse.py::test_segment_reduce[sum]
0.18s call     tests/compute/test_heterograph.py::test_updates[idtype1]
0.18s call     tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-v-shp4-g1]
0.18s call     tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-u-shp1-g1]
0.18s call     tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-v-shp1-g1]
0.18s call     tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-v-shp4-g0]
0.18s call     tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-v-shp2-g1]
0.18s call     tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-e-shp1-g1]
0.18s call     tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-v-shp1-g0]
0.18s call     tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-u-shp4-g1]
0.18s call     tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-v-shp0-g1]
0.18s call     tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-e-shp0-g0]
0.18s call     tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-v-shp0-g0]
0.18s call     tests/compute/test_sparse.py::test_edge_softmax[idtype1-shp0-src-g0]
0.18s call     tests/compute/test_sparse.py::test_edge_softmax[idtype0-shp0-src-g0]
0.18s call     tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-e-shp4-g0]
0.18s call     tests/compute/test_sparse.py::test_edge_softmax[idtype1-shp1-src-g0]
0.18s call     tests/compute/test_sparse.py::test_edge_softmax[idtype1-shp2-src-g0]
0.18s call     tests/compute/test_sparse.py::test_edge_softmax[idtype1-shp2-dst-g0]
0.18s call     tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-e-shp4-g1]
0.18s call     tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-v-shp2-g0]
0.18s call     tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-v-shp4-g0]
0.18s call     tests/compute/test_sparse.py::test_edge_softmax[idtype1-shp1-dst-g0]
0.18s call     tests/compute/test_sparse.py::test_edge_softmax[idtype0-shp2-src-g0]
0.18s call     tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-v-shp3-g1]
0.18s call     tests/compute/test_sparse.py::test_edge_softmax[idtype0-shp2-dst-g0]
0.18s call     tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-v-shp3-g0]
0.18s call     tests/compute/test_sparse.py::test_edge_softmax[idtype0-shp1-dst-g0]
0.18s call     tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-e-shp3-g1]
0.18s call     tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-e-shp1-g0]
0.18s call     tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-u-shp2-g0]
0.18s call     tests/compute/test_sparse.py::test_edge_softmax[idtype0-shp0-dst-g0]
0.18s call     tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-v-shp4-g1]
0.18s call     tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-e-shp3-g0]
0.18s call     tests/compute/test_sparse.py::test_edge_softmax[idtype0-shp1-src-g0]
0.18s call     tests/compute/test_sparse.py::test_edge_softmax[idtype1-shp0-dst-g0]
0.18s call     tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-u-shp1-g0]
0.18s call     tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-u-shp2-g1]
0.18s call     tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-u-shp3-g1]
0.18s call     tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-u-shp3-g1]
0.18s call     tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-e-shp0-g1]
0.18s call     tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-e-shp2-g0]
0.18s call     tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-u-shp1-g1]
0.18s call     tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-e-shp2-g1]
0.18s call     tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-u-shp3-g0]
0.18s call     tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-u-shp4-g0]
0.18s call     tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-u-shp0-g1]
0.18s call     tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-e-shp0-g0]
0.18s call     tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-u-shp4-g1]
0.17s call     tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-e-shp1-g1]
0.17s call     tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-u-shp2-g1]
0.17s call     tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-u-shp0-g1]
0.17s call     tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-u-shp4-g0]
0.17s call     tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-u-shp1-g0]
0.17s call     tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-u-shp0-g0]
0.17s call     tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-u-shp2-g0]
0.17s call     tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-u-shp3-g0]
0.17s call     tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-v-shp1-g1]
0.17s call     tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-u-shp0-g0]
0.17s call     tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-e-shp4-g0]
0.17s call     tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-e-shp2-g1]
0.17s call     tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-e-shp0-g1]
0.17s call     tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-e-shp3-g1]
0.17s call     tests/compute/test_heterograph.py::test_updates[idtype0]
0.17s call     tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-v-shp3-g1]
0.17s call     tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-e-shp0-g1]
0.17s call     tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-e-shp1-g0]
0.17s call     tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-e-shp4-g1]
0.17s call     tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-v-shp0-g0]
0.17s call     tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-v-shp4-g1]
0.17s call     tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-v-shp1-g0]
0.17s call     tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-e-shp2-g0]
0.17s call     tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-e-shp0-g0]
0.17s call     tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-v-shp3-g1]
========= 2227 passed, 231 skipped, 216 warnings in 233.36s (0:03:53) ==========
Cloning into '/root/jenkins/workspace/dgl_PR-4648@2/third_party/tensorpipe/third_party/libuv'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tensorpipe/third_party/libnop'...
============================= test session starts ==============================
platform linux -- Python 3.7.0, pytest-7.1.2, pluggy-1.0.0 -- /opt/conda/envs/tensorflow-ci/bin/python3
cachedir: .pytest_cache
rootdir: /root/jenkins/workspace/dgl_PR-4648
collecting ... Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tensorpipe/third_party/libuv'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648@2/third_party/tensorpipe/third_party/pybind11'...
Submodule path 'third_party/tensorpipe/third_party/googletest': checked out 'aee0f9d9b5b87796ee8a0ab26b7587ec30e8858e'
Submodule path 'third_party/tensorpipe/third_party/libnop': checked out 'aa95422ea8c409e3f078d2ee7708a5f59a8b9fa2'
Submodule path 'third_party/tensorpipe/third_party/libuv': checked out '1dff88e5161cba5c59276d2070d2e304e4dcb242'
Submodule path 'third_party/tensorpipe/third_party/pybind11': checked out 'a23996fce38ff6ccfbcdc09f1e63f2c4be5ea2ef'
Submodule 'tools/clang' (https://github.com/wjakob/clang-cindex-python3) registered for path 'third_party/tensorpipe/third_party/pybind11/tools/clang'
Cloning into '/root/jenkins/workspace/dgl_PR-4648@2/third_party/tensorpipe/third_party/pybind11/tools/clang'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tensorpipe/third_party/pybind11'...
Submodule path 'third_party/tensorpipe/third_party/pybind11/tools/clang': checked out '6a00cbc4a9b8e68b71caf7f774b3f9c753ae84d5'
Submodule path 'third_party/thrust': checked out '6a3078c64cab0e2f276340fa5dcafa0d758ed890'
Submodule 'cub' (https://github.com/NVIDIA/cub.git) registered for path 'third_party/thrust/dependencies/cub'
Cloning into '/root/jenkins/workspace/dgl_PR-4648@2/third_party/thrust/dependencies/cub'...
Submodule path 'third_party/tensorpipe/third_party/googletest': checked out 'aee0f9d9b5b87796ee8a0ab26b7587ec30e8858e'
Submodule path 'third_party/tensorpipe/third_party/libnop': checked out 'aa95422ea8c409e3f078d2ee7708a5f59a8b9fa2'
collected 920 items

tests/tensorflow/test_basic.py::test PASSED                              [  0%]
tests/tensorflow/test_nn.py::test_graph_conv[1] PASSED                   [  0%]
tests/tensorflow/test_nn.py::test_graph_conv[2] PASSED                   [  0%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-none-g0-idtype0] Submodule path 'third_party/tensorpipe/third_party/libuv': checked out '1dff88e5161cba5c59276d2070d2e304e4dcb242'
Submodule path 'third_party/tensorpipe/third_party/pybind11': checked out 'a23996fce38ff6ccfbcdc09f1e63f2c4be5ea2ef'
Submodule 'tools/clang' (https://github.com/wjakob/clang-cindex-python3) registered for path 'third_party/tensorpipe/third_party/pybind11/tools/clang'
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tensorpipe/third_party/pybind11/tools/clang'...
PASSED [  0%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-none-g0-idtype1] PASSED [  0%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-none-g1-idtype0] PASSED [  0%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-none-g1-idtype1] PASSED [  0%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-none-g2-idtype0] PASSED [  0%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-none-g2-idtype1] PASSED [  0%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-none-g3-idtype0] PASSED [  1%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-none-g3-idtype1] PASSED [  1%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-none-g4-idtype0] PASSED [  1%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-none-g4-idtype1] PASSED [  1%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-none-g5-idtype0] PASSED [  1%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-none-g5-idtype1] PASSED [  1%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-none-g6-idtype0] PASSED [  1%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-none-g6-idtype1] PASSED [  1%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-both-g0-idtype0] PASSED [  1%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-both-g0-idtype1] PASSED [  2%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-both-g1-idtype0] PASSED [  2%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-both-g1-idtype1] PASSED [  2%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-both-g2-idtype0] PASSED [  2%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-both-g2-idtype1] PASSED [  2%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-both-g3-idtype0] PASSED [  2%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-both-g3-idtype1] PASSED [  2%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-both-g4-idtype0] PASSED [  2%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-both-g4-idtype1] PASSED [  2%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-both-g5-idtype0] PASSED [  3%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-both-g5-idtype1] PASSED [  3%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-both-g6-idtype0] PASSED [  3%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-both-g6-idtype1] PASSED [  3%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-right-g0-idtype0] PASSED [  3%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-right-g0-idtype1] PASSED [  3%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-right-g1-idtype0] PASSED [  3%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-right-g1-idtype1] PASSED [  3%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-right-g2-idtype0] PASSED [  3%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-right-g2-idtype1] PASSED [  4%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-right-g3-idtype0] PASSED [  4%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-right-g3-idtype1] PASSED [  4%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-right-g4-idtype0] PASSED [  4%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-right-g4-idtype1] PASSED [  4%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-right-g5-idtype0] PASSED [  4%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-right-g5-idtype1] PASSED [  4%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-right-g6-idtype0] PASSED [  4%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-right-g6-idtype1] PASSED [  4%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-left-g0-idtype0] PASSED [  5%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-left-g0-idtype1] PASSED [  5%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-left-g1-idtype0] PASSED [  5%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-left-g1-idtype1] PASSED [  5%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-left-g2-idtype0] PASSED [  5%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-left-g2-idtype1] PASSED [  5%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-left-g3-idtype0] PASSED [  5%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-left-g3-idtype1] PASSED [  5%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-left-g4-idtype0] PASSED [  5%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-left-g4-idtype1] PASSED [  5%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-left-g5-idtype0] PASSED [  6%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-left-g5-idtype1] PASSED [  6%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-left-g6-idtype0] PASSED [  6%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-left-g6-idtype1] PASSED [  6%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-none-g0-idtype0] PASSED [  6%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-none-g0-idtype1] PASSED [  6%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-none-g1-idtype0] PASSED [  6%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-none-g1-idtype1] PASSED [  6%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-none-g2-idtype0] PASSED [  6%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-none-g2-idtype1] PASSED [  7%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-none-g3-idtype0] PASSED [  7%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-none-g3-idtype1] PASSED [  7%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-none-g4-idtype0] PASSED [  7%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-none-g4-idtype1] PASSED [  7%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-none-g5-idtype0] PASSED [  7%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-none-g5-idtype1] PASSED [  7%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-none-g6-idtype0] PASSED [  7%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-none-g6-idtype1] PASSED [  7%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-both-g0-idtype0] PASSED [  8%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-both-g0-idtype1] PASSED [  8%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-both-g1-idtype0] PASSED [  8%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-both-g1-idtype1] PASSED [  8%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-both-g2-idtype0] PASSED [  8%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-both-g2-idtype1] PASSED [  8%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-both-g3-idtype0] PASSED [  8%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-both-g3-idtype1] Submodule path 'third_party/tensorpipe/third_party/pybind11/tools/clang': checked out '6a00cbc4a9b8e68b71caf7f774b3f9c753ae84d5'
PASSED [  8%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-both-g4-idtype0] PASSED [  8%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-both-g4-idtype1] PASSED [  9%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-both-g5-idtype0] PASSED [  9%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-both-g5-idtype1] PASSED [  9%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-both-g6-idtype0] PASSED [  9%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-both-g6-idtype1] PASSED [  9%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-right-g0-idtype0] PASSED [  9%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-right-g0-idtype1] PASSED [  9%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-right-g1-idtype0] PASSED [  9%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-right-g1-idtype1] PASSED [  9%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-right-g2-idtype0] PASSED [ 10%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-right-g2-idtype1] PASSED [ 10%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-right-g3-idtype0] PASSED [ 10%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-right-g3-idtype1] PASSED [ 10%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-right-g4-idtype0] PASSED [ 10%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-right-g4-idtype1] PASSED [ 10%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-right-g5-idtype0] PASSED [ 10%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-right-g5-idtype1] PASSED [ 10%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-right-g6-idtype0] PASSED [ 10%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-right-g6-idtype1] PASSED [ 10%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-left-g0-idtype0] PASSED [ 11%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-left-g0-idtype1] PASSED [ 11%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-left-g1-idtype0] PASSED [ 11%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-left-g1-idtype1] PASSED [ 11%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-left-g2-idtype0] PASSED [ 11%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-left-g2-idtype1] PASSED [ 11%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-left-g3-idtype0] PASSED [ 11%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-left-g3-idtype1] PASSED [ 11%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-left-g4-idtype0] PASSED [ 11%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-left-g4-idtype1] PASSED [ 12%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-left-g5-idtype0] PASSED [ 12%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-left-g5-idtype1] PASSED [ 12%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-left-g6-idtype0] PASSED [ 12%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-left-g6-idtype1] PASSED [ 12%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-none-g0-idtype0] PASSED [ 12%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-none-g0-idtype1] PASSED [ 12%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-none-g1-idtype0] PASSED [ 12%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-none-g1-idtype1] PASSED [ 12%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-none-g2-idtype0] PASSED [ 13%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-none-g2-idtype1] PASSED [ 13%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-none-g3-idtype0] PASSED [ 13%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-none-g3-idtype1] PASSED [ 13%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-none-g4-idtype0] PASSED [ 13%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-none-g4-idtype1] Submodule path 'third_party/thrust': checked out '6a3078c64cab0e2f276340fa5dcafa0d758ed890'
Submodule 'cub' (https://github.com/NVIDIA/cub.git) registered for path 'third_party/thrust/dependencies/cub'
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/thrust/dependencies/cub'...
PASSED [ 13%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-none-g5-idtype0] PASSED [ 13%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-none-g5-idtype1] PASSED [ 13%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-none-g6-idtype0] PASSED [ 13%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-none-g6-idtype1] PASSED [ 14%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-both-g0-idtype0] PASSED [ 14%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-both-g0-idtype1] PASSED [ 14%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-both-g1-idtype0] PASSED [ 14%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-both-g1-idtype1] PASSED [ 14%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-both-g2-idtype0] PASSED [ 14%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-both-g2-idtype1] PASSED [ 14%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-both-g3-idtype0] PASSED [ 14%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-both-g3-idtype1] PASSED [ 14%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-both-g4-idtype0] PASSED [ 15%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-both-g4-idtype1] PASSED [ 15%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-both-g5-idtype0] PASSED [ 15%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-both-g5-idtype1] PASSED [ 15%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-both-g6-idtype0] PASSED [ 15%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-both-g6-idtype1] PASSED [ 15%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-right-g0-idtype0] PASSED [ 15%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-right-g0-idtype1] PASSED [ 15%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-right-g1-idtype0] PASSED [ 15%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-right-g1-idtype1] PASSED [ 15%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-right-g2-idtype0] PASSED [ 16%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-right-g2-idtype1] PASSED [ 16%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-right-g3-idtype0] PASSED [ 16%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-right-g3-idtype1] PASSED [ 16%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-right-g4-idtype0] PASSED [ 16%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-right-g4-idtype1] PASSED [ 16%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-right-g5-idtype0] PASSED [ 16%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-right-g5-idtype1] PASSED [ 16%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-right-g6-idtype0] PASSED [ 16%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-right-g6-idtype1] PASSED [ 17%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-left-g0-idtype0] PASSED [ 17%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-left-g0-idtype1] PASSED [ 17%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-left-g1-idtype0] PASSED [ 17%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-left-g1-idtype1] PASSED [ 17%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-left-g2-idtype0] PASSED [ 17%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-left-g2-idtype1] PASSED [ 17%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-left-g3-idtype0] PASSED [ 17%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-left-g3-idtype1] PASSED [ 17%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-left-g4-idtype0] PASSED [ 18%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-left-g4-idtype1] PASSED [ 18%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-left-g5-idtype0] PASSED [ 18%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-left-g5-idtype1] PASSED [ 18%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-left-g6-idtype0] PASSED [ 18%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-left-g6-idtype1] PASSED [ 18%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-none-g0-idtype0] PASSED [ 18%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-none-g0-idtype1] PASSED [ 18%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-none-g1-idtype0] PASSED [ 18%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-none-g1-idtype1] PASSED [ 19%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-none-g2-idtype0] PASSED [ 19%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-none-g2-idtype1] PASSED [ 19%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-none-g3-idtype0] PASSED [ 19%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-none-g3-idtype1] PASSED [ 19%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-none-g4-idtype0] PASSED [ 19%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-none-g4-idtype1] PASSED [ 19%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-none-g5-idtype0] PASSED [ 19%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-none-g5-idtype1] PASSED [ 19%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-none-g6-idtype0] PASSED [ 20%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-none-g6-idtype1] PASSED [ 20%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-both-g0-idtype0] PASSED [ 20%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-both-g0-idtype1] PASSED [ 20%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-both-g1-idtype0] PASSED [ 20%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-both-g1-idtype1] PASSED [ 20%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-both-g2-idtype0] PASSED [ 20%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-both-g2-idtype1] PASSED [ 20%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-both-g3-idtype0] PASSED [ 20%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-both-g3-idtype1] PASSED [ 20%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-both-g4-idtype0] PASSED [ 21%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-both-g4-idtype1] PASSED [ 21%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-both-g5-idtype0] PASSED [ 21%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-both-g5-idtype1] PASSED [ 21%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-both-g6-idtype0] PASSED [ 21%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-both-g6-idtype1] PASSED [ 21%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-right-g0-idtype0] PASSED [ 21%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-right-g0-idtype1] PASSED [ 21%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-right-g1-idtype0] PASSED [ 21%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-right-g1-idtype1] PASSED [ 22%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-right-g2-idtype0] PASSED [ 22%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-right-g2-idtype1] PASSED [ 22%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-right-g3-idtype0] PASSED [ 22%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-right-g3-idtype1] PASSED [ 22%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-right-g4-idtype0] PASSED [ 22%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-right-g4-idtype1] PASSED [ 22%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-right-g5-idtype0] PASSED [ 22%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-right-g5-idtype1] PASSED [ 22%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-right-g6-idtype0] PASSED [ 23%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-right-g6-idtype1] PASSED [ 23%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-left-g0-idtype0] PASSED [ 23%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-left-g0-idtype1] PASSED [ 23%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-left-g1-idtype0] PASSED [ 23%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-left-g1-idtype1] PASSED [ 23%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-left-g2-idtype0] PASSED [ 23%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-left-g2-idtype1] PASSED [ 23%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-left-g3-idtype0] PASSED [ 23%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-left-g3-idtype1] PASSED [ 24%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-left-g4-idtype0] PASSED [ 24%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-left-g4-idtype1] PASSED [ 24%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-left-g5-idtype0] PASSED [ 24%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-left-g5-idtype1] PASSED [ 24%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-left-g6-idtype0] PASSED [ 24%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-left-g6-idtype1] PASSED [ 24%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-none-g0-idtype0] PASSED [ 24%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-none-g0-idtype1] PASSED [ 24%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-none-g1-idtype0] PASSED [ 25%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-none-g1-idtype1] PASSED [ 25%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-none-g2-idtype0] PASSED [ 25%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-none-g2-idtype1] PASSED [ 25%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-none-g3-idtype0] PASSED [ 25%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-none-g3-idtype1] PASSED [ 25%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-none-g4-idtype0] PASSED [ 25%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-none-g4-idtype1] PASSED [ 25%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-none-g5-idtype0] PASSED [ 25%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-none-g5-idtype1] PASSED [ 25%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-none-g6-idtype0] PASSED [ 26%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-none-g6-idtype1] PASSED [ 26%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-both-g0-idtype0] PASSED [ 26%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-both-g0-idtype1] PASSED [ 26%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-both-g1-idtype0] PASSED [ 26%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-both-g1-idtype1] PASSED [ 26%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-both-g2-idtype0] PASSED [ 26%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-both-g2-idtype1] PASSED [ 26%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-both-g3-idtype0] PASSED [ 26%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-both-g3-idtype1] PASSED [ 27%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-both-g4-idtype0] PASSED [ 27%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-both-g4-idtype1] PASSED [ 27%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-both-g5-idtype0] PASSED [ 27%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-both-g5-idtype1] PASSED [ 27%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-both-g6-idtype0] PASSED [ 27%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-both-g6-idtype1] PASSED [ 27%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-right-g0-idtype0] PASSED [ 27%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-right-g0-idtype1] PASSED [ 27%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-right-g1-idtype0] PASSED [ 28%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-right-g1-idtype1] PASSED [ 28%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-right-g2-idtype0] PASSED [ 28%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-right-g2-idtype1] PASSED [ 28%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-right-g3-idtype0] PASSED [ 28%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-right-g3-idtype1] PASSED [ 28%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-right-g4-idtype0] PASSED [ 28%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-right-g4-idtype1] PASSED [ 28%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-right-g5-idtype0] PASSED [ 28%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-right-g5-idtype1] PASSED [ 29%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-right-g6-idtype0] PASSED [ 29%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-right-g6-idtype1] PASSED [ 29%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-left-g0-idtype0] PASSED [ 29%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-left-g0-idtype1] PASSED [ 29%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-left-g1-idtype0] PASSED [ 29%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-left-g1-idtype1] PASSED [ 29%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-left-g2-idtype0] PASSED [ 29%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-left-g2-idtype1] PASSED [ 29%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-left-g3-idtype0] PASSED [ 30%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-left-g3-idtype1] PASSED [ 30%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-left-g4-idtype0] PASSED [ 30%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-left-g4-idtype1] PASSED [ 30%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-left-g5-idtype0] PASSED [ 30%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-left-g5-idtype1] PASSED [ 30%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-left-g6-idtype0] PASSED [ 30%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-left-g6-idtype1] PASSED [ 30%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-none-g0-idtype0] PASSED [ 30%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-none-g0-idtype1] PASSED [ 30%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-none-g1-idtype0] PASSED [ 31%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-none-g1-idtype1] PASSED [ 31%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-none-g2-idtype0] PASSED [ 31%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-none-g2-idtype1] PASSED [ 31%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-none-g3-idtype0] PASSED [ 31%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-none-g3-idtype1] PASSED [ 31%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-none-g4-idtype0] PASSED [ 31%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-none-g4-idtype1] PASSED [ 31%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-none-g5-idtype0] PASSED [ 31%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-none-g5-idtype1] PASSED [ 32%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-none-g6-idtype0] PASSED [ 32%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-none-g6-idtype1] PASSED [ 32%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-both-g0-idtype0] PASSED [ 32%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-both-g0-idtype1] PASSED [ 32%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-both-g1-idtype0] PASSED [ 32%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-both-g1-idtype1] PASSED [ 32%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-both-g2-idtype0] PASSED [ 32%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-both-g2-idtype1] PASSED [ 32%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-both-g3-idtype0] PASSED [ 33%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-both-g3-idtype1] PASSED [ 33%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-both-g4-idtype0] PASSED [ 33%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-both-g4-idtype1] PASSED [ 33%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-both-g5-idtype0] PASSED [ 33%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-both-g5-idtype1] PASSED [ 33%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-both-g6-idtype0] PASSED [ 33%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-both-g6-idtype1] PASSED [ 33%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-right-g0-idtype0] PASSED [ 33%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-right-g0-idtype1] PASSED [ 34%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-right-g1-idtype0] PASSED [ 34%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-right-g1-idtype1] PASSED [ 34%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-right-g2-idtype0] PASSED [ 34%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-right-g2-idtype1] PASSED [ 34%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-right-g3-idtype0] PASSED [ 34%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-right-g3-idtype1] PASSED [ 34%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-right-g4-idtype0] PASSED [ 34%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-right-g4-idtype1] PASSED [ 34%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-right-g5-idtype0] PASSED [ 35%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-right-g5-idtype1] PASSED [ 35%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-right-g6-idtype0] PASSED [ 35%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-right-g6-idtype1] PASSED [ 35%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-left-g0-idtype0] PASSED [ 35%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-left-g0-idtype1] PASSED [ 35%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-left-g1-idtype0] PASSED [ 35%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-left-g1-idtype1] PASSED [ 35%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-left-g2-idtype0] PASSED [ 35%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-left-g2-idtype1] PASSED [ 35%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-left-g3-idtype0] PASSED [ 36%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-left-g3-idtype1] PASSED [ 36%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-left-g4-idtype0] PASSED [ 36%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-left-g4-idtype1] PASSED [ 36%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-left-g5-idtype0] PASSED [ 36%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-left-g5-idtype1] PASSED [ 36%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-left-g6-idtype0] PASSED [ 36%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-left-g6-idtype1] PASSED [ 36%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-none-g0-idtype0] PASSED [ 36%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-none-g0-idtype1] PASSED [ 37%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-none-g1-idtype0] PASSED [ 37%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-none-g1-idtype1] PASSED [ 37%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-none-g2-idtype0] PASSED [ 37%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-none-g2-idtype1] PASSED [ 37%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-none-g3-idtype0] PASSED [ 37%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-none-g3-idtype1] PASSED [ 37%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-none-g4-idtype0] PASSED [ 37%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-none-g4-idtype1] PASSED [ 37%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-none-g5-idtype0] PASSED [ 38%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-none-g5-idtype1] PASSED [ 38%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-none-g6-idtype0] PASSED [ 38%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-none-g6-idtype1] PASSED [ 38%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-both-g0-idtype0] PASSED [ 38%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-both-g0-idtype1] PASSED [ 38%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-both-g1-idtype0] PASSED [ 38%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-both-g1-idtype1] PASSED [ 38%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-both-g2-idtype0] PASSED [ 38%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-both-g2-idtype1] PASSED [ 39%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-both-g3-idtype0] PASSED [ 39%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-both-g3-idtype1] PASSED [ 39%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-both-g4-idtype0] PASSED [ 39%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-both-g4-idtype1] PASSED [ 39%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-both-g5-idtype0] PASSED [ 39%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-both-g5-idtype1] PASSED [ 39%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-both-g6-idtype0] PASSED [ 39%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-both-g6-idtype1] PASSED [ 39%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-right-g0-idtype0] PASSED [ 40%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-right-g0-idtype1] PASSED [ 40%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-right-g1-idtype0] PASSED [ 40%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-right-g1-idtype1] PASSED [ 40%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-right-g2-idtype0] PASSED [ 40%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-right-g2-idtype1] PASSED [ 40%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-right-g3-idtype0] PASSED [ 40%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-right-g3-idtype1] PASSED [ 40%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-right-g4-idtype0] PASSED [ 40%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-right-g4-idtype1] PASSED [ 40%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-right-g5-idtype0] PASSED [ 41%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-right-g5-idtype1] PASSED [ 41%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-right-g6-idtype0] PASSED [ 41%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-right-g6-idtype1] PASSED [ 41%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-left-g0-idtype0] PASSED [ 41%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-left-g0-idtype1] PASSED [ 41%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-left-g1-idtype0] PASSED [ 41%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-left-g1-idtype1] PASSED [ 41%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-left-g2-idtype0] PASSED [ 41%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-left-g2-idtype1] PASSED [ 42%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-left-g3-idtype0] PASSED [ 42%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-left-g3-idtype1] PASSED [ 42%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-left-g4-idtype0] PASSED [ 42%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-left-g4-idtype1] PASSED [ 42%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-left-g5-idtype0] PASSED [ 42%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-left-g5-idtype1] PASSED [ 42%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-left-g6-idtype0] PASSED [ 42%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-left-g6-idtype1] PASSED [ 42%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-none-g0-idtype0] PASSED [ 43%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-none-g0-idtype1] PASSED [ 43%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-none-g1-idtype0] PASSED [ 43%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-none-g1-idtype1] PASSED [ 43%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-none-g2-idtype0] PASSED [ 43%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-none-g2-idtype1] PASSED [ 43%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-none-g3-idtype0] PASSED [ 43%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-none-g3-idtype1] PASSED [ 43%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-none-g4-idtype0] PASSED [ 43%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-none-g4-idtype1] PASSED [ 44%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-none-g5-idtype0] PASSED [ 44%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-none-g5-idtype1] PASSED [ 44%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-none-g6-idtype0] PASSED [ 44%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-none-g6-idtype1] PASSED [ 44%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-both-g0-idtype0] PASSED [ 44%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-both-g0-idtype1] PASSED [ 44%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-both-g1-idtype0] PASSED [ 44%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-both-g1-idtype1] PASSED [ 44%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-both-g2-idtype0] PASSED [ 45%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-both-g2-idtype1] PASSED [ 45%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-both-g3-idtype0] PASSED [ 45%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-both-g3-idtype1] PASSED [ 45%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-both-g4-idtype0] PASSED [ 45%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-both-g4-idtype1] PASSED [ 45%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-both-g5-idtype0] PASSED [ 45%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-both-g5-idtype1] PASSED [ 45%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-both-g6-idtype0] PASSED [ 45%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-both-g6-idtype1] PASSED [ 45%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-right-g0-idtype0] PASSED [ 46%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-right-g0-idtype1] PASSED [ 46%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-right-g1-idtype0] PASSED [ 46%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-right-g1-idtype1] PASSED [ 46%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-right-g2-idtype0] PASSED [ 46%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-right-g2-idtype1] PASSED [ 46%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-right-g3-idtype0] PASSED [ 46%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-right-g3-idtype1] PASSED [ 46%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-right-g4-idtype0] PASSED [ 46%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-right-g4-idtype1] PASSED [ 47%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-right-g5-idtype0] PASSED [ 47%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-right-g5-idtype1] PASSED [ 47%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-right-g6-idtype0] PASSED [ 47%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-right-g6-idtype1] PASSED [ 47%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-left-g0-idtype0] PASSED [ 47%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-left-g0-idtype1] PASSED [ 47%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-left-g1-idtype0] PASSED [ 47%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-left-g1-idtype1] PASSED [ 47%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-left-g2-idtype0] PASSED [ 48%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-left-g2-idtype1] PASSED [ 48%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-left-g3-idtype0] PASSED [ 48%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-left-g3-idtype1] PASSED [ 48%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-left-g4-idtype0] PASSED [ 48%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-left-g4-idtype1] PASSED [ 48%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-left-g5-idtype0] PASSED [ 48%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-left-g5-idtype1] PASSED [ 48%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-left-g6-idtype0] PASSED [ 48%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-left-g6-idtype1] PASSED [ 49%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-True-True-none-g0-idtype0] PASSED [ 49%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-True-True-none-g0-idtype1] PASSED [ 49%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-True-True-none-g1-idtype0] PASSED [ 49%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-True-True-none-g1-idtype1] PASSED [ 49%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-True-True-both-g0-idtype0] PASSED [ 49%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-True-True-both-g0-idtype1] PASSED [ 49%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-True-True-both-g1-idtype0] PASSED [ 49%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-True-True-both-g1-idtype1] PASSED [ 49%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-True-True-right-g0-idtype0] PASSED [ 50%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-True-True-right-g0-idtype1] PASSED [ 50%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-True-True-right-g1-idtype0] PASSED [ 50%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-True-True-right-g1-idtype1] PASSED [ 50%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-True-False-none-g0-idtype0] PASSED [ 50%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-True-False-none-g0-idtype1] PASSED [ 50%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-True-False-none-g1-idtype0] PASSED [ 50%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-True-False-none-g1-idtype1] PASSED [ 50%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-True-False-both-g0-idtype0] PASSED [ 50%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-True-False-both-g0-idtype1] PASSED [ 50%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-True-False-both-g1-idtype0] PASSED [ 51%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-True-False-both-g1-idtype1] PASSED [ 51%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-True-False-right-g0-idtype0] PASSED [ 51%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-True-False-right-g0-idtype1] PASSED [ 51%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-True-False-right-g1-idtype0] PASSED [ 51%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-True-False-right-g1-idtype1] PASSED [ 51%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-False-True-none-g0-idtype0] PASSED [ 51%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-False-True-none-g0-idtype1] PASSED [ 51%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-False-True-none-g1-idtype0] PASSED [ 51%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-False-True-none-g1-idtype1] Submodule path 'third_party/thrust/dependencies/cub': checked out 'cdaa9558a85e45d849016e5fe7b6e4ee79113f95'
PASSED [ 52%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-False-True-both-g0-idtype0] PASSED [ 52%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-False-True-both-g0-idtype1] PASSED [ 52%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-False-True-both-g1-idtype0] PASSED [ 52%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-False-True-both-g1-idtype1] PASSED [ 52%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-False-True-right-g0-idtype0] PASSED [ 52%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-False-True-right-g0-idtype1] PASSED [ 52%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-False-True-right-g1-idtype0] PASSED [ 52%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-False-True-right-g1-idtype1] PASSED [ 52%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-False-False-none-g0-idtype0] PASSED [ 53%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-False-False-none-g0-idtype1] PASSED [ 53%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-False-False-none-g1-idtype0] PASSED [ 53%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-False-False-none-g1-idtype1] PASSED [ 53%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-False-False-both-g0-idtype0] PASSED [ 53%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-False-False-both-g0-idtype1] PASSED [ 53%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-False-False-both-g1-idtype0] PASSED [ 53%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-False-False-both-g1-idtype1] PASSED [ 53%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-False-False-right-g0-idtype0] PASSED [ 53%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-False-False-right-g0-idtype1] PASSED [ 54%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-False-False-right-g1-idtype0] PASSED [ 54%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-False-False-right-g1-idtype1] PASSED [ 54%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-True-True-none-g0-idtype0] PASSED [ 54%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-True-True-none-g0-idtype1] PASSED [ 54%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-True-True-none-g1-idtype0] PASSED [ 54%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-True-True-none-g1-idtype1] PASSED [ 54%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-True-True-both-g0-idtype0] PASSED [ 54%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-True-True-both-g0-idtype1] PASSED [ 54%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-True-True-both-g1-idtype0] PASSED [ 55%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-True-True-both-g1-idtype1] PASSED [ 55%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-True-True-right-g0-idtype0] PASSED [ 55%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-True-True-right-g0-idtype1] PASSED [ 55%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-True-True-right-g1-idtype0] PASSED [ 55%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-True-True-right-g1-idtype1] PASSED [ 55%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-True-False-none-g0-idtype0] PASSED [ 55%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-True-False-none-g0-idtype1] PASSED [ 55%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-True-False-none-g1-idtype0] PASSED [ 55%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-True-False-none-g1-idtype1] PASSED [ 55%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-True-False-both-g0-idtype0] PASSED [ 56%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-True-False-both-g0-idtype1] PASSED [ 56%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-True-False-both-g1-idtype0] PASSED [ 56%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-True-False-both-g1-idtype1] PASSED [ 56%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-True-False-right-g0-idtype0] PASSED [ 56%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-True-False-right-g0-idtype1] Submodule path 'third_party/tvm': checked out 'b2e418cb109df4cd1f17a2cf2894a1b396a6b838'
Submodule 'dlpack' (https://github.com/dmlc/dlpack) registered for path 'third_party/tvm/3rdparty/dlpack'
Submodule 'dmlc-core' (https://github.com/dmlc/dmlc-core) registered for path 'third_party/tvm/3rdparty/dmlc-core'
Submodule '3rdparty/rang' (https://github.com/agauniyal/rang) registered for path 'third_party/tvm/3rdparty/rang'
Submodule '3rdparty/vta-hw' (https://github.com/apache/incubator-tvm-vta) registered for path 'third_party/tvm/3rdparty/vta-hw'
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tvm/3rdparty/dlpack'...
PASSED [ 56%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-True-False-right-g1-idtype0] PASSED [ 56%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-True-False-right-g1-idtype1] PASSED [ 56%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-False-True-none-g0-idtype0] PASSED [ 56%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-False-True-none-g0-idtype1] PASSED [ 57%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-False-True-none-g1-idtype0] PASSED [ 57%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-False-True-none-g1-idtype1] PASSED [ 57%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-False-True-both-g0-idtype0] PASSED [ 57%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-False-True-both-g0-idtype1] PASSED [ 57%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-False-True-both-g1-idtype0] PASSED [ 57%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-False-True-both-g1-idtype1] PASSED [ 57%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-False-True-right-g0-idtype0] PASSED [ 57%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-False-True-right-g0-idtype1] PASSED [ 57%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-False-True-right-g1-idtype0] PASSED [ 58%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-False-True-right-g1-idtype1] PASSED [ 58%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-False-False-none-g0-idtype0] PASSED [ 58%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-False-False-none-g0-idtype1] PASSED [ 58%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-False-False-none-g1-idtype0] PASSED [ 58%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-False-False-none-g1-idtype1] PASSED [ 58%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-False-False-both-g0-idtype0] PASSED [ 58%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-False-False-both-g0-idtype1] PASSED [ 58%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-False-False-both-g1-idtype0] PASSED [ 58%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-False-False-both-g1-idtype1] PASSED [ 59%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-False-False-right-g0-idtype0] PASSED [ 59%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-False-False-right-g0-idtype1] PASSED [ 59%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-False-False-right-g1-idtype0] PASSED [ 59%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-False-False-right-g1-idtype1] PASSED [ 59%]
tests/tensorflow/test_nn.py::test_simple_pool PASSED                     [ 59%]
tests/tensorflow/test_nn.py::test_glob_att_pool PASSED                   [ 59%]
tests/tensorflow/test_nn.py::test_rgcn[1] PASSED                         [ 59%]
tests/tensorflow/test_nn.py::test_rgcn[2] PASSED                         [ 59%]
tests/tensorflow/test_nn.py::test_rgcn[8] PASSED                         [ 60%]
tests/tensorflow/test_nn.py::test_gat_conv[1-1-g0-idtype0] Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tvm/3rdparty/dmlc-core'...
PASSED        [ 60%]
tests/tensorflow/test_nn.py::test_gat_conv[1-1-g0-idtype1] PASSED        [ 60%]
tests/tensorflow/test_nn.py::test_gat_conv[1-1-g1-idtype0] PASSED        [ 60%]
tests/tensorflow/test_nn.py::test_gat_conv[1-1-g1-idtype1] PASSED        [ 60%]
tests/tensorflow/test_nn.py::test_gat_conv[1-1-g2-idtype0] PASSED        [ 60%]
tests/tensorflow/test_nn.py::test_gat_conv[1-1-g2-idtype1] PASSED        [ 60%]
tests/tensorflow/test_nn.py::test_gat_conv[1-1-g3-idtype0] PASSED        [ 60%]
tests/tensorflow/test_nn.py::test_gat_conv[1-1-g3-idtype1] PASSED        [ 60%]
tests/tensorflow/test_nn.py::test_gat_conv[1-1-g4-idtype0] PASSED        [ 60%]
tests/tensorflow/test_nn.py::test_gat_conv[1-1-g4-idtype1] PASSED        [ 61%]
tests/tensorflow/test_nn.py::test_gat_conv[1-1-g5-idtype0] PASSED        [ 61%]
tests/tensorflow/test_nn.py::test_gat_conv[1-1-g5-idtype1] PASSED        [ 61%]
tests/tensorflow/test_nn.py::test_gat_conv[1-1-g6-idtype0] PASSED        [ 61%]
tests/tensorflow/test_nn.py::test_gat_conv[1-1-g6-idtype1] PASSED        [ 61%]
tests/tensorflow/test_nn.py::test_gat_conv[1-2-g0-idtype0] PASSED        [ 61%]
tests/tensorflow/test_nn.py::test_gat_conv[1-2-g0-idtype1] PASSED        [ 61%]
tests/tensorflow/test_nn.py::test_gat_conv[1-2-g1-idtype0] PASSED        [ 61%]
tests/tensorflow/test_nn.py::test_gat_conv[1-2-g1-idtype1] PASSED        [ 61%]
tests/tensorflow/test_nn.py::test_gat_conv[1-2-g2-idtype0] PASSED        [ 62%]
tests/tensorflow/test_nn.py::test_gat_conv[1-2-g2-idtype1] PASSED        [ 62%]
tests/tensorflow/test_nn.py::test_gat_conv[1-2-g3-idtype0] Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tvm/3rdparty/rang'...
PASSED        [ 62%]
tests/tensorflow/test_nn.py::test_gat_conv[1-2-g3-idtype1] PASSED        [ 62%]
tests/tensorflow/test_nn.py::test_gat_conv[1-2-g4-idtype0] PASSED        [ 62%]
tests/tensorflow/test_nn.py::test_gat_conv[1-2-g4-idtype1] PASSED        [ 62%]
tests/tensorflow/test_nn.py::test_gat_conv[1-2-g5-idtype0] PASSED        [ 62%]
tests/tensorflow/test_nn.py::test_gat_conv[1-2-g5-idtype1] PASSED        [ 62%]
tests/tensorflow/test_nn.py::test_gat_conv[1-2-g6-idtype0] PASSED        [ 62%]
tests/tensorflow/test_nn.py::test_gat_conv[1-2-g6-idtype1] PASSED        [ 63%]
tests/tensorflow/test_nn.py::test_gat_conv[4-1-g0-idtype0] PASSED        [ 63%]
tests/tensorflow/test_nn.py::test_gat_conv[4-1-g0-idtype1] PASSED        [ 63%]
tests/tensorflow/test_nn.py::test_gat_conv[4-1-g1-idtype0] PASSED        [ 63%]
tests/tensorflow/test_nn.py::test_gat_conv[4-1-g1-idtype1] PASSED        [ 63%]
tests/tensorflow/test_nn.py::test_gat_conv[4-1-g2-idtype0] PASSED        [ 63%]
tests/tensorflow/test_nn.py::test_gat_conv[4-1-g2-idtype1] PASSED        [ 63%]
tests/tensorflow/test_nn.py::test_gat_conv[4-1-g3-idtype0] PASSED        [ 63%]
tests/tensorflow/test_nn.py::test_gat_conv[4-1-g3-idtype1] PASSED        [ 63%]
tests/tensorflow/test_nn.py::test_gat_conv[4-1-g4-idtype0] PASSED        [ 64%]
tests/tensorflow/test_nn.py::test_gat_conv[4-1-g4-idtype1] PASSED        [ 64%]
tests/tensorflow/test_nn.py::test_gat_conv[4-1-g5-idtype0] PASSED        [ 64%]
tests/tensorflow/test_nn.py::test_gat_conv[4-1-g5-idtype1] PASSED        [ 64%]
tests/tensorflow/test_nn.py::test_gat_conv[4-1-g6-idtype0] PASSED        [ 64%]
tests/tensorflow/test_nn.py::test_gat_conv[4-1-g6-idtype1] PASSED        [ 64%]
tests/tensorflow/test_nn.py::test_gat_conv[4-2-g0-idtype0] PASSED        [ 64%]
tests/tensorflow/test_nn.py::test_gat_conv[4-2-g0-idtype1] PASSED        [ 64%]
tests/tensorflow/test_nn.py::test_gat_conv[4-2-g1-idtype0] PASSED        [ 64%]
tests/tensorflow/test_nn.py::test_gat_conv[4-2-g1-idtype1] PASSED        [ 65%]
tests/tensorflow/test_nn.py::test_gat_conv[4-2-g2-idtype0] PASSED        [ 65%]
tests/tensorflow/test_nn.py::test_gat_conv[4-2-g2-idtype1] PASSED        [ 65%]
tests/tensorflow/test_nn.py::test_gat_conv[4-2-g3-idtype0] PASSED        [ 65%]
tests/tensorflow/test_nn.py::test_gat_conv[4-2-g3-idtype1] PASSED        [ 65%]
tests/tensorflow/test_nn.py::test_gat_conv[4-2-g4-idtype0] PASSED        [ 65%]
tests/tensorflow/test_nn.py::test_gat_conv[4-2-g4-idtype1] Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tvm/3rdparty/vta-hw'...
PASSED        [ 65%]
tests/tensorflow/test_nn.py::test_gat_conv[4-2-g5-idtype0] PASSED        [ 65%]
tests/tensorflow/test_nn.py::test_gat_conv[4-2-g5-idtype1] PASSED        [ 65%]
tests/tensorflow/test_nn.py::test_gat_conv[4-2-g6-idtype0] PASSED        [ 65%]
tests/tensorflow/test_nn.py::test_gat_conv[4-2-g6-idtype1] PASSED        [ 66%]
tests/tensorflow/test_nn.py::test_gat_conv_bi[1-1-g0-idtype0] PASSED     [ 66%]
tests/tensorflow/test_nn.py::test_gat_conv_bi[1-1-g0-idtype1] PASSED     [ 66%]
tests/tensorflow/test_nn.py::test_gat_conv_bi[1-1-g1-idtype0] PASSED     [ 66%]
tests/tensorflow/test_nn.py::test_gat_conv_bi[1-1-g1-idtype1] PASSED     [ 66%]
tests/tensorflow/test_nn.py::test_gat_conv_bi[1-2-g0-idtype0] PASSED     [ 66%]
tests/tensorflow/test_nn.py::test_gat_conv_bi[1-2-g0-idtype1] PASSED     [ 66%]
tests/tensorflow/test_nn.py::test_gat_conv_bi[1-2-g1-idtype0] PASSED     [ 66%]
tests/tensorflow/test_nn.py::test_gat_conv_bi[1-2-g1-idtype1] PASSED     [ 66%]
tests/tensorflow/test_nn.py::test_gat_conv_bi[4-1-g0-idtype0] PASSED     [ 67%]
tests/tensorflow/test_nn.py::test_gat_conv_bi[4-1-g0-idtype1] PASSED     [ 67%]
tests/tensorflow/test_nn.py::test_gat_conv_bi[4-1-g1-idtype0] PASSED     [ 67%]
tests/tensorflow/test_nn.py::test_gat_conv_bi[4-1-g1-idtype1] PASSED     [ 67%]
tests/tensorflow/test_nn.py::test_gat_conv_bi[4-2-g0-idtype0] PASSED     [ 67%]
tests/tensorflow/test_nn.py::test_gat_conv_bi[4-2-g0-idtype1] PASSED     [ 67%]
tests/tensorflow/test_nn.py::test_gat_conv_bi[4-2-g1-idtype0] PASSED     [ 67%]
tests/tensorflow/test_nn.py::test_gat_conv_bi[4-2-g1-idtype1] PASSED     [ 67%]
tests/tensorflow/test_nn.py::test_sage_conv[1-mean-g0-idtype0] PASSED    [ 67%]
tests/tensorflow/test_nn.py::test_sage_conv[1-mean-g0-idtype1] PASSED    [ 68%]
tests/tensorflow/test_nn.py::test_sage_conv[1-mean-g1-idtype0] PASSED    [ 68%]
tests/tensorflow/test_nn.py::test_sage_conv[1-mean-g1-idtype1] PASSED    [ 68%]
tests/tensorflow/test_nn.py::test_sage_conv[1-mean-g2-idtype0] PASSED    [ 68%]
tests/tensorflow/test_nn.py::test_sage_conv[1-mean-g2-idtype1] PASSED    [ 68%]
tests/tensorflow/test_nn.py::test_sage_conv[1-mean-g3-idtype0] PASSED    [ 68%]
tests/tensorflow/test_nn.py::test_sage_conv[1-mean-g3-idtype1] PASSED    [ 68%]
tests/tensorflow/test_nn.py::test_sage_conv[1-mean-g4-idtype0] PASSED    [ 68%]
tests/tensorflow/test_nn.py::test_sage_conv[1-mean-g4-idtype1] PASSED    [ 68%]
tests/tensorflow/test_nn.py::test_sage_conv[1-mean-g5-idtype0] PASSED    [ 69%]
tests/tensorflow/test_nn.py::test_sage_conv[1-mean-g5-idtype1] PASSED    [ 69%]
tests/tensorflow/test_nn.py::test_sage_conv[1-mean-g6-idtype0] PASSED    [ 69%]
tests/tensorflow/test_nn.py::test_sage_conv[1-mean-g6-idtype1] PASSED    [ 69%]
tests/tensorflow/test_nn.py::test_sage_conv[1-mean-g7-idtype0] PASSED    [ 69%]
tests/tensorflow/test_nn.py::test_sage_conv[1-mean-g7-idtype1] PASSED    [ 69%]
tests/tensorflow/test_nn.py::test_sage_conv[1-pool-g0-idtype0] PASSED    [ 69%]
tests/tensorflow/test_nn.py::test_sage_conv[1-pool-g0-idtype1] Submodule path 'third_party/tvm/3rdparty/dlpack': checked out '3ec04430e89a6834e5a1b99471f415fa939bf642'
PASSED    [ 69%]
tests/tensorflow/test_nn.py::test_sage_conv[1-pool-g1-idtype0] PASSED    [ 69%]
tests/tensorflow/test_nn.py::test_sage_conv[1-pool-g1-idtype1] PASSED    [ 70%]
tests/tensorflow/test_nn.py::test_sage_conv[1-pool-g2-idtype0] PASSED    [ 70%]
tests/tensorflow/test_nn.py::test_sage_conv[1-pool-g2-idtype1] PASSED    [ 70%]
tests/tensorflow/test_nn.py::test_sage_conv[1-pool-g3-idtype0] PASSED    [ 70%]
tests/tensorflow/test_nn.py::test_sage_conv[1-pool-g3-idtype1] PASSED    [ 70%]
tests/tensorflow/test_nn.py::test_sage_conv[1-pool-g4-idtype0] PASSED    [ 70%]
tests/tensorflow/test_nn.py::test_sage_conv[1-pool-g4-idtype1] PASSED    [ 70%]
tests/tensorflow/test_nn.py::test_sage_conv[1-pool-g5-idtype0] PASSED    [ 70%]
tests/tensorflow/test_nn.py::test_sage_conv[1-pool-g5-idtype1] PASSED    [ 70%]
tests/tensorflow/test_nn.py::test_sage_conv[1-pool-g6-idtype0] PASSED    [ 70%]
tests/tensorflow/test_nn.py::test_sage_conv[1-pool-g6-idtype1] PASSED    [ 71%]
tests/tensorflow/test_nn.py::test_sage_conv[1-pool-g7-idtype0] PASSED    [ 71%]
tests/tensorflow/test_nn.py::test_sage_conv[1-pool-g7-idtype1] PASSED    [ 71%]
tests/tensorflow/test_nn.py::test_sage_conv[1-gcn-g0-idtype0] PASSED     [ 71%]
tests/tensorflow/test_nn.py::test_sage_conv[1-gcn-g0-idtype1] PASSED     [ 71%]
tests/tensorflow/test_nn.py::test_sage_conv[1-gcn-g1-idtype0] PASSED     [ 71%]
tests/tensorflow/test_nn.py::test_sage_conv[1-gcn-g1-idtype1] PASSED     [ 71%]
tests/tensorflow/test_nn.py::test_sage_conv[1-gcn-g2-idtype0] PASSED     [ 71%]
tests/tensorflow/test_nn.py::test_sage_conv[1-gcn-g2-idtype1] PASSED     [ 71%]
tests/tensorflow/test_nn.py::test_sage_conv[1-gcn-g3-idtype0] PASSED     [ 72%]
tests/tensorflow/test_nn.py::test_sage_conv[1-gcn-g3-idtype1] PASSED     [ 72%]
tests/tensorflow/test_nn.py::test_sage_conv[1-gcn-g4-idtype0] PASSED     [ 72%]
tests/tensorflow/test_nn.py::test_sage_conv[1-gcn-g4-idtype1] PASSED     [ 72%]
tests/tensorflow/test_nn.py::test_sage_conv[1-gcn-g5-idtype0] PASSED     [ 72%]
tests/tensorflow/test_nn.py::test_sage_conv[1-gcn-g5-idtype1] PASSED     [ 72%]
tests/tensorflow/test_nn.py::test_sage_conv[1-gcn-g6-idtype0] PASSED     [ 72%]
tests/tensorflow/test_nn.py::test_sage_conv[1-gcn-g6-idtype1] PASSED     [ 72%]
tests/tensorflow/test_nn.py::test_sage_conv[1-gcn-g7-idtype0] PASSED     [ 72%]
tests/tensorflow/test_nn.py::test_sage_conv[1-gcn-g7-idtype1] PASSED     [ 73%]
tests/tensorflow/test_nn.py::test_sage_conv[10-mean-g0-idtype0] Submodule path 'third_party/tvm/3rdparty/dmlc-core': checked out '6c401e242c59a1f4c913918246591bb13fd714e7'
Submodule path 'third_party/tvm/3rdparty/rang': checked out 'cabe04d6d6b05356fa8f9741704924788f0dd762'
Submodule path 'third_party/tvm/3rdparty/vta-hw': checked out '87ce9acfae550d1a487746e9d06c2e250076e54c'
PASSED   [ 73%]
tests/tensorflow/test_nn.py::test_sage_conv[10-mean-g0-idtype1] PASSED   [ 73%]
tests/tensorflow/test_nn.py::test_sage_conv[10-mean-g1-idtype0] PASSED   [ 73%]
tests/tensorflow/test_nn.py::test_sage_conv[10-mean-g1-idtype1] PASSED   [ 73%]
tests/tensorflow/test_nn.py::test_sage_conv[10-mean-g2-idtype0] PASSED   [ 73%]
tests/tensorflow/test_nn.py::test_sage_conv[10-mean-g2-idtype1] PASSED   [ 73%]
tests/tensorflow/test_nn.py::test_sage_conv[10-mean-g3-idtype0] PASSED   [ 73%]
tests/tensorflow/test_nn.py::test_sage_conv[10-mean-g3-idtype1] PASSED   [ 73%]
tests/tensorflow/test_nn.py::test_sage_conv[10-mean-g4-idtype0] PASSED   [ 74%]
tests/tensorflow/test_nn.py::test_sage_conv[10-mean-g4-idtype1] PASSED   [ 74%]
tests/tensorflow/test_nn.py::test_sage_conv[10-mean-g5-idtype0] PASSED   [ 74%]
tests/tensorflow/test_nn.py::test_sage_conv[10-mean-g5-idtype1] PASSED   [ 74%]
tests/tensorflow/test_nn.py::test_sage_conv[10-mean-g6-idtype0] PASSED   [ 74%]
tests/tensorflow/test_nn.py::test_sage_conv[10-mean-g6-idtype1] PASSED   [ 74%]
tests/tensorflow/test_nn.py::test_sage_conv[10-mean-g7-idtype0] PASSED   [ 74%]
tests/tensorflow/test_nn.py::test_sage_conv[10-mean-g7-idtype1] PASSED   [ 74%]
tests/tensorflow/test_nn.py::test_sage_conv[10-pool-g0-idtype0] PASSED   [ 74%]
tests/tensorflow/test_nn.py::test_sage_conv[10-pool-g0-idtype1] PASSED   [ 75%]
tests/tensorflow/test_nn.py::test_sage_conv[10-pool-g1-idtype0] PASSED   [ 75%]
tests/tensorflow/test_nn.py::test_sage_conv[10-pool-g1-idtype1] PASSED   [ 75%]
tests/tensorflow/test_nn.py::test_sage_conv[10-pool-g2-idtype0] PASSED   [ 75%]
tests/tensorflow/test_nn.py::test_sage_conv[10-pool-g2-idtype1] PASSED   [ 75%]
tests/tensorflow/test_nn.py::test_sage_conv[10-pool-g3-idtype0] PASSED   [ 75%]
tests/tensorflow/test_nn.py::test_sage_conv[10-pool-g3-idtype1] PASSED   [ 75%]
tests/tensorflow/test_nn.py::test_sage_conv[10-pool-g4-idtype0] PASSED   [ 75%]
tests/tensorflow/test_nn.py::test_sage_conv[10-pool-g4-idtype1] PASSED   [ 75%]
tests/tensorflow/test_nn.py::test_sage_conv[10-pool-g5-idtype0] PASSED   [ 75%]
tests/tensorflow/test_nn.py::test_sage_conv[10-pool-g5-idtype1] PASSED   [ 76%]
tests/tensorflow/test_nn.py::test_sage_conv[10-pool-g6-idtype0] Submodule path 'third_party/xbyak': checked out '757e4063f6464740b8ff4a2cae9136d2f8458020'
[Pipeline] unstash
PASSED   [ 76%]
tests/tensorflow/test_nn.py::test_sage_conv[10-pool-g6-idtype1] PASSED   [ 76%]
tests/tensorflow/test_nn.py::test_sage_conv[10-pool-g7-idtype0] PASSED   [ 76%]
tests/tensorflow/test_nn.py::test_sage_conv[10-pool-g7-idtype1] PASSED   [ 76%]
tests/tensorflow/test_nn.py::test_sage_conv[10-gcn-g0-idtype0] PASSED    [ 76%]
tests/tensorflow/test_nn.py::test_sage_conv[10-gcn-g0-idtype1] PASSED    [ 76%]
tests/tensorflow/test_nn.py::test_sage_conv[10-gcn-g1-idtype0] PASSED    [ 76%]
tests/tensorflow/test_nn.py::test_sage_conv[10-gcn-g1-idtype1] PASSED    [ 76%]
tests/tensorflow/test_nn.py::test_sage_conv[10-gcn-g2-idtype0] PASSED    [ 77%]
tests/tensorflow/test_nn.py::test_sage_conv[10-gcn-g2-idtype1] PASSED    [ 77%]
tests/tensorflow/test_nn.py::test_sage_conv[10-gcn-g3-idtype0] PASSED    [ 77%]
tests/tensorflow/test_nn.py::test_sage_conv[10-gcn-g3-idtype1] PASSED    [ 77%]
tests/tensorflow/test_nn.py::test_sage_conv[10-gcn-g4-idtype0] PASSED    [ 77%]
tests/tensorflow/test_nn.py::test_sage_conv[10-gcn-g4-idtype1] PASSED    [ 77%]
tests/tensorflow/test_nn.py::test_sage_conv[10-gcn-g5-idtype0] PASSED    [ 77%]
tests/tensorflow/test_nn.py::test_sage_conv[10-gcn-g5-idtype1] PASSED    [ 77%]
tests/tensorflow/test_nn.py::test_sage_conv[10-gcn-g6-idtype0] PASSED    [ 77%]
tests/tensorflow/test_nn.py::test_sage_conv[10-gcn-g6-idtype1] PASSED    [ 78%]
tests/tensorflow/test_nn.py::test_sage_conv[10-gcn-g7-idtype0] PASSED    [ 78%]
tests/tensorflow/test_nn.py::test_sage_conv[10-gcn-g7-idtype1] PASSED    [ 78%]
tests/tensorflow/test_nn.py::test_sage_conv_bi[1-mean-g0-idtype0] PASSED [ 78%]
tests/tensorflow/test_nn.py::test_sage_conv_bi[1-mean-g0-idtype1] PASSED [ 78%]
tests/tensorflow/test_nn.py::test_sage_conv_bi[1-mean-g1-idtype0] PASSED [ 78%]
tests/tensorflow/test_nn.py::test_sage_conv_bi[1-mean-g1-idtype1] PASSED [ 78%]
tests/tensorflow/test_nn.py::test_sage_conv_bi[1-mean-g2-idtype0] PASSED [ 78%]
tests/tensorflow/test_nn.py::test_sage_conv_bi[1-mean-g2-idtype1] PASSED [ 78%]
tests/tensorflow/test_nn.py::test_sage_conv_bi[1-pool-g0-idtype0] PASSED [ 79%]
tests/tensorflow/test_nn.py::test_sage_conv_bi[1-pool-g0-idtype1] PASSED [ 79%]
tests/tensorflow/test_nn.py::test_sage_conv_bi[1-pool-g1-idtype0] PASSED [ 79%]
tests/tensorflow/test_nn.py::test_sage_conv_bi[1-pool-g1-idtype1] PASSED [ 79%]
tests/tensorflow/test_nn.py::test_sage_conv_bi[1-pool-g2-idtype0] PASSED [ 79%]
tests/tensorflow/test_nn.py::test_sage_conv_bi[1-pool-g2-idtype1] [Pipeline] echo
Unpacked build/libdgl.so, build/runUnitTests, python/dgl/_ffi/_cy3/core.cpython-*-x86_64-linux-gnu.so, build/tensoradapter/pytorch/*.so from dgl-cpu-linux
[Pipeline] sh
PASSED [ 79%]
tests/tensorflow/test_nn.py::test_sage_conv_bi[1-gcn-g0-idtype0] PASSED  [ 79%]
tests/tensorflow/test_nn.py::test_sage_conv_bi[1-gcn-g0-idtype1] PASSED  [ 79%]
tests/tensorflow/test_nn.py::test_sage_conv_bi[1-gcn-g1-idtype0] PASSED  [ 79%]
tests/tensorflow/test_nn.py::test_sage_conv_bi[1-gcn-g1-idtype1] PASSED  [ 80%]
tests/tensorflow/test_nn.py::test_sage_conv_bi[1-gcn-g2-idtype0] PASSED  [ 80%]
tests/tensorflow/test_nn.py::test_sage_conv_bi[1-gcn-g2-idtype1] PASSED  [ 80%]
tests/tensorflow/test_nn.py::test_sage_conv_bi[2-mean-g0-idtype0] PASSED [ 80%]
tests/tensorflow/test_nn.py::test_sage_conv_bi[2-mean-g0-idtype1] PASSED [ 80%]
tests/tensorflow/test_nn.py::test_sage_conv_bi[2-mean-g1-idtype0] PASSED [ 80%]
tests/tensorflow/test_nn.py::test_sage_conv_bi[2-mean-g1-idtype1] PASSED [ 80%]
tests/tensorflow/test_nn.py::test_sage_conv_bi[2-mean-g2-idtype0] PASSED [ 80%]
tests/tensorflow/test_nn.py::test_sage_conv_bi[2-mean-g2-idtype1] PASSED [ 80%]
tests/tensorflow/test_nn.py::test_sage_conv_bi[2-pool-g0-idtype0] PASSED [ 80%]
tests/tensorflow/test_nn.py::test_sage_conv_bi[2-pool-g0-idtype1] PASSED [ 81%]
tests/tensorflow/test_nn.py::test_sage_conv_bi[2-pool-g1-idtype0] PASSED [ 81%]
tests/tensorflow/test_nn.py::test_sage_conv_bi[2-pool-g1-idtype1] PASSED [ 81%]
tests/tensorflow/test_nn.py::test_sage_conv_bi[2-pool-g2-idtype0] PASSED [ 81%]
tests/tensorflow/test_nn.py::test_sage_conv_bi[2-pool-g2-idtype1] PASSED [ 81%]
tests/tensorflow/test_nn.py::test_sage_conv_bi[2-gcn-g0-idtype0] PASSED  [ 81%]
tests/tensorflow/test_nn.py::test_sage_conv_bi[2-gcn-g0-idtype1] PASSED  [ 81%]
tests/tensorflow/test_nn.py::test_sage_conv_bi[2-gcn-g1-idtype0] PASSED  [ 81%]
tests/tensorflow/test_nn.py::test_sage_conv_bi[2-gcn-g1-idtype1] PASSED  [ 81%]
tests/tensorflow/test_nn.py::test_sage_conv_bi[2-gcn-g2-idtype0] PASSED  [ 82%]
tests/tensorflow/test_nn.py::test_sage_conv_bi[2-gcn-g2-idtype1] PASSED  [ 82%]
tests/tensorflow/test_nn.py::test_sage_conv_bi_empty[1-mean-idtype0] PASSED [ 82%]
tests/tensorflow/test_nn.py::test_sage_conv_bi_empty[1-mean-idtype1] + bash tests/scripts/task_cpp_unit_test.sh
/root/jenkins/workspace/dgl_PR-4648
~/jenkins/workspace/dgl_PR-4648/build ~/jenkins/workspace/dgl_PR-4648
total 14M
-rwxr-xr-x 1 root root  12M Sep 27 04:43 libdgl.so
-rwxr-xr-x 1 root root 2.4M Sep 27 04:44 runUnitTests
drwxr-xr-x 3 root root 4.0K Sep 27 05:23 tensoradapter
Running main() from /root/jenkins/workspace/dgl_PR-4648/third_party/googletest/googletest/src/gtest_main.cc
[==========] Running 101 tests from 23 test suites.
[----------] Global test environment set-up.
[----------] 1 test from GraphTest
[ RUN      ] GraphTest.TestNumVertices
[       OK ] GraphTest.TestNumVertices (0 ms)
[----------] 1 test from GraphTest (0 ms total)

[----------] 3 tests from MessageQueueTest
[ RUN      ] MessageQueueTest.AddRemove
[05:23:48] /root/jenkins/workspace/dgl_PR-4648/src/rpc/network/msg_queue.cc:27: Message is larger than the queue.
[       OK ] MessageQueueTest.AddRemove (0 ms)
[ RUN      ] MessageQueueTest.EmptyAndNoMoreAdd
[       OK ] MessageQueueTest.EmptyAndNoMoreAdd (0 ms)
[ RUN      ] MessageQueueTest.MultiThread
PASSED [ 82%]
tests/tensorflow/test_nn.py::test_sage_conv_bi_empty[1-pool-idtype0] PASSED [ 82%]
tests/tensorflow/test_nn.py::test_sage_conv_bi_empty[1-pool-idtype1] PASSED [ 82%]
tests/tensorflow/test_nn.py::test_sage_conv_bi_empty[1-gcn-idtype0] PASSED [ 82%]
tests/tensorflow/test_nn.py::test_sage_conv_bi_empty[1-gcn-idtype1] PASSED [ 82%]
tests/tensorflow/test_nn.py::test_sage_conv_bi_empty[2-mean-idtype0] PASSED [ 82%]
tests/tensorflow/test_nn.py::test_sage_conv_bi_empty[2-mean-idtype1] PASSED [ 83%]
tests/tensorflow/test_nn.py::test_sage_conv_bi_empty[2-pool-idtype0] [       OK ] MessageQueueTest.MultiThread (11 ms)
[----------] 3 tests from MessageQueueTest (11 ms total)

[----------] 3 tests from SocketCommunicatorTest
[ RUN      ] SocketCommunicatorTest.SendAndRecv
PASSED [ 83%]
tests/tensorflow/test_nn.py::test_sage_conv_bi_empty[2-pool-idtype1] PASSED [ 83%]
tests/tensorflow/test_nn.py::test_sage_conv_bi_empty[2-gcn-idtype0] PASSED [ 83%]
tests/tensorflow/test_nn.py::test_sage_conv_bi_empty[2-gcn-idtype1] PASSED [ 83%]
tests/tensorflow/test_nn.py::test_sgc_conv[1-g0-idtype0] PASSED          [ 83%]
tests/tensorflow/test_nn.py::test_sgc_conv[1-g0-idtype1] PASSED          [ 83%]
tests/tensorflow/test_nn.py::test_sgc_conv[1-g1-idtype0] PASSED          [ 83%]
tests/tensorflow/test_nn.py::test_sgc_conv[1-g1-idtype1] PASSED          [ 83%]
tests/tensorflow/test_nn.py::test_sgc_conv[1-g2-idtype0] PASSED          [ 84%]
tests/tensorflow/test_nn.py::test_sgc_conv[1-g2-idtype1] PASSED          [ 84%]
tests/tensorflow/test_nn.py::test_sgc_conv[1-g3-idtype0] PASSED          [ 84%]
tests/tensorflow/test_nn.py::test_sgc_conv[1-g3-idtype1] PASSED          [ 84%]
tests/tensorflow/test_nn.py::test_sgc_conv[1-g4-idtype0] PASSED          [ 84%]
tests/tensorflow/test_nn.py::test_sgc_conv[1-g4-idtype1] PASSED          [ 84%]
tests/tensorflow/test_nn.py::test_sgc_conv[1-g5-idtype0] PASSED          [ 84%]
tests/tensorflow/test_nn.py::test_sgc_conv[1-g5-idtype1] PASSED          [ 84%]
tests/tensorflow/test_nn.py::test_sgc_conv[2-g0-idtype0] PASSED          [ 84%]
tests/tensorflow/test_nn.py::test_sgc_conv[2-g0-idtype1] PASSED          [ 85%]
tests/tensorflow/test_nn.py::test_sgc_conv[2-g1-idtype0] PASSED          [ 85%]
tests/tensorflow/test_nn.py::test_sgc_conv[2-g1-idtype1] PASSED          [ 85%]
tests/tensorflow/test_nn.py::test_sgc_conv[2-g2-idtype0] PASSED          [ 85%]
tests/tensorflow/test_nn.py::test_sgc_conv[2-g2-idtype1] PASSED          [ 85%]
tests/tensorflow/test_nn.py::test_sgc_conv[2-g3-idtype0] PASSED          [ 85%]
tests/tensorflow/test_nn.py::test_sgc_conv[2-g3-idtype1] PASSED          [ 85%]
tests/tensorflow/test_nn.py::test_sgc_conv[2-g4-idtype0] PASSED          [ 85%]
tests/tensorflow/test_nn.py::test_sgc_conv[2-g4-idtype1] PASSED          [ 85%]
tests/tensorflow/test_nn.py::test_sgc_conv[2-g5-idtype0] PASSED          [ 85%]
tests/tensorflow/test_nn.py::test_sgc_conv[2-g5-idtype1] PASSED          [ 86%]
tests/tensorflow/test_nn.py::test_appnp_conv[g0-idtype0] PASSED          [ 86%]
tests/tensorflow/test_nn.py::test_appnp_conv[g0-idtype1] PASSED          [ 86%]
tests/tensorflow/test_nn.py::test_appnp_conv[g1-idtype0] PASSED          [ 86%]
tests/tensorflow/test_nn.py::test_appnp_conv[g1-idtype1] PASSED          [ 86%]
tests/tensorflow/test_nn.py::test_appnp_conv[g2-idtype0] PASSED          [ 86%]
tests/tensorflow/test_nn.py::test_appnp_conv[g2-idtype1] PASSED          [ 86%]
tests/tensorflow/test_nn.py::test_appnp_conv[g3-idtype0] PASSED          [ 86%]
tests/tensorflow/test_nn.py::test_appnp_conv[g3-idtype1] PASSED          [ 86%]
tests/tensorflow/test_nn.py::test_appnp_conv[g4-idtype0] PASSED          [ 87%]
tests/tensorflow/test_nn.py::test_appnp_conv[g4-idtype1] PASSED          [ 87%]
tests/tensorflow/test_nn.py::test_appnp_conv[g5-idtype0] PASSED          [ 87%]
tests/tensorflow/test_nn.py::test_appnp_conv[g5-idtype1] PASSED          [ 87%]
tests/tensorflow/test_nn.py::test_gin_conv[mean-g0-idtype0] PASSED       [ 87%]
tests/tensorflow/test_nn.py::test_gin_conv[mean-g0-idtype1] PASSED       [ 87%]
tests/tensorflow/test_nn.py::test_gin_conv[mean-g1-idtype0] PASSED       [ 87%]
tests/tensorflow/test_nn.py::test_gin_conv[mean-g1-idtype1] PASSED       [ 87%]
tests/tensorflow/test_nn.py::test_gin_conv[mean-g2-idtype0] PASSED       [ 87%]
tests/tensorflow/test_nn.py::test_gin_conv[mean-g2-idtype1] PASSED       [ 88%]
tests/tensorflow/test_nn.py::test_gin_conv[mean-g3-idtype0] PASSED       [ 88%]
tests/tensorflow/test_nn.py::test_gin_conv[mean-g3-idtype1] PASSED       [ 88%]
tests/tensorflow/test_nn.py::test_gin_conv[mean-g4-idtype0] PASSED       [ 88%]
tests/tensorflow/test_nn.py::test_gin_conv[mean-g4-idtype1] PASSED       [ 88%]
tests/tensorflow/test_nn.py::test_gin_conv[mean-g5-idtype0] PASSED       [ 88%]
tests/tensorflow/test_nn.py::test_gin_conv[mean-g5-idtype1] PASSED       [ 88%]
tests/tensorflow/test_nn.py::test_gin_conv[mean-g6-idtype0] PASSED       [ 88%]
tests/tensorflow/test_nn.py::test_gin_conv[mean-g6-idtype1] PASSED       [ 88%]
tests/tensorflow/test_nn.py::test_gin_conv[mean-g7-idtype0] PASSED       [ 89%]
tests/tensorflow/test_nn.py::test_gin_conv[mean-g7-idtype1] PASSED       [ 89%]
tests/tensorflow/test_nn.py::test_gin_conv[max-g0-idtype0] PASSED        [ 89%]
tests/tensorflow/test_nn.py::test_gin_conv[max-g0-idtype1] Submodule path 'third_party/thrust/dependencies/cub': checked out 'cdaa9558a85e45d849016e5fe7b6e4ee79113f95'
Submodule path 'third_party/tvm': checked out 'b2e418cb109df4cd1f17a2cf2894a1b396a6b838'
Submodule 'dlpack' (https://github.com/dmlc/dlpack) registered for path 'third_party/tvm/3rdparty/dlpack'
Submodule 'dmlc-core' (https://github.com/dmlc/dmlc-core) registered for path 'third_party/tvm/3rdparty/dmlc-core'
Submodule '3rdparty/rang' (https://github.com/agauniyal/rang) registered for path 'third_party/tvm/3rdparty/rang'
Submodule '3rdparty/vta-hw' (https://github.com/apache/incubator-tvm-vta) registered for path 'third_party/tvm/3rdparty/vta-hw'
Cloning into '/root/jenkins/workspace/dgl_PR-4648@2/third_party/tvm/3rdparty/dlpack'...
PASSED        [ 89%]
tests/tensorflow/test_nn.py::test_gin_conv[max-g1-idtype0] PASSED        [ 89%]
tests/tensorflow/test_nn.py::test_gin_conv[max-g1-idtype1] PASSED        [ 89%]
tests/tensorflow/test_nn.py::test_gin_conv[max-g2-idtype0] PASSED        [ 89%]
tests/tensorflow/test_nn.py::test_gin_conv[max-g2-idtype1] PASSED        [ 89%]
tests/tensorflow/test_nn.py::test_gin_conv[max-g3-idtype0] PASSED        [ 89%]
tests/tensorflow/test_nn.py::test_gin_conv[max-g3-idtype1] PASSED        [ 90%]
tests/tensorflow/test_nn.py::test_gin_conv[max-g4-idtype0] PASSED        [ 90%]
tests/tensorflow/test_nn.py::test_gin_conv[max-g4-idtype1] PASSED        [ 90%]
tests/tensorflow/test_nn.py::test_gin_conv[max-g5-idtype0] PASSED        [ 90%]
tests/tensorflow/test_nn.py::test_gin_conv[max-g5-idtype1] PASSED        [ 90%]
tests/tensorflow/test_nn.py::test_gin_conv[max-g6-idtype0] PASSED        [ 90%]
tests/tensorflow/test_nn.py::test_gin_conv[max-g6-idtype1] PASSED        [ 90%]
tests/tensorflow/test_nn.py::test_gin_conv[max-g7-idtype0] PASSED        [ 90%]
tests/tensorflow/test_nn.py::test_gin_conv[max-g7-idtype1] PASSED        [ 90%]
tests/tensorflow/test_nn.py::test_gin_conv[sum-g0-idtype0] PASSED        [ 90%]
tests/tensorflow/test_nn.py::test_gin_conv[sum-g0-idtype1] PASSED        [ 91%]
tests/tensorflow/test_nn.py::test_gin_conv[sum-g1-idtype0] PASSED        [ 91%]
tests/tensorflow/test_nn.py::test_gin_conv[sum-g1-idtype1] PASSED        [ 91%]
tests/tensorflow/test_nn.py::test_gin_conv[sum-g2-idtype0] PASSED        [ 91%]
tests/tensorflow/test_nn.py::test_gin_conv[sum-g2-idtype1] PASSED        [ 91%]
tests/tensorflow/test_nn.py::test_gin_conv[sum-g3-idtype0] PASSED        [ 91%]
tests/tensorflow/test_nn.py::test_gin_conv[sum-g3-idtype1] PASSED        [ 91%]
tests/tensorflow/test_nn.py::test_gin_conv[sum-g4-idtype0] PASSED        [ 91%]
tests/tensorflow/test_nn.py::test_gin_conv[sum-g4-idtype1] PASSED        [ 91%]
tests/tensorflow/test_nn.py::test_gin_conv[sum-g5-idtype0] PASSED        [ 92%]
tests/tensorflow/test_nn.py::test_gin_conv[sum-g5-idtype1] PASSED        [ 92%]
tests/tensorflow/test_nn.py::test_gin_conv[sum-g6-idtype0] PASSED        [ 92%]
tests/tensorflow/test_nn.py::test_gin_conv[sum-g6-idtype1] PASSED        [ 92%]
tests/tensorflow/test_nn.py::test_gin_conv[sum-g7-idtype0] PASSED        [ 92%]
tests/tensorflow/test_nn.py::test_gin_conv[sum-g7-idtype1] PASSED        [ 92%]
tests/tensorflow/test_nn.py::test_gin_conv_bi[mean-g0-idtype0] PASSED    [ 92%]
tests/tensorflow/test_nn.py::test_gin_conv_bi[mean-g0-idtype1] PASSED    [ 92%]
tests/tensorflow/test_nn.py::test_gin_conv_bi[mean-g1-idtype0] PASSED    [ 92%]
tests/tensorflow/test_nn.py::test_gin_conv_bi[mean-g1-idtype1] PASSED    [ 93%]
tests/tensorflow/test_nn.py::test_gin_conv_bi[mean-g2-idtype0] PASSED    [ 93%]
tests/tensorflow/test_nn.py::test_gin_conv_bi[mean-g2-idtype1] PASSED    [ 93%]
tests/tensorflow/test_nn.py::test_gin_conv_bi[max-g0-idtype0] PASSED     [ 93%]
tests/tensorflow/test_nn.py::test_gin_conv_bi[max-g0-idtype1] PASSED     [ 93%]
tests/tensorflow/test_nn.py::test_gin_conv_bi[max-g1-idtype0] PASSED     [ 93%]
tests/tensorflow/test_nn.py::test_gin_conv_bi[max-g1-idtype1] PASSED     [ 93%]
tests/tensorflow/test_nn.py::test_gin_conv_bi[max-g2-idtype0] PASSED     [ 93%]
tests/tensorflow/test_nn.py::test_gin_conv_bi[max-g2-idtype1] Cloning into '/root/jenkins/workspace/dgl_PR-4648@2/third_party/tvm/3rdparty/dmlc-core'...
PASSED     [ 93%]
tests/tensorflow/test_nn.py::test_gin_conv_bi[sum-g0-idtype0] PASSED     [ 94%]
tests/tensorflow/test_nn.py::test_gin_conv_bi[sum-g0-idtype1] PASSED     [ 94%]
tests/tensorflow/test_nn.py::test_gin_conv_bi[sum-g1-idtype0] PASSED     [ 94%]
tests/tensorflow/test_nn.py::test_gin_conv_bi[sum-g1-idtype1] PASSED     [ 94%]
tests/tensorflow/test_nn.py::test_gin_conv_bi[sum-g2-idtype0] PASSED     [ 94%]
tests/tensorflow/test_nn.py::test_gin_conv_bi[sum-g2-idtype1] PASSED     [ 94%]
tests/tensorflow/test_nn.py::test_edge_conv[1-g0-idtype0] PASSED         [ 94%]
tests/tensorflow/test_nn.py::test_edge_conv[1-g0-idtype1] PASSED         [ 94%]
tests/tensorflow/test_nn.py::test_edge_conv[1-g1-idtype0] PASSED         [ 94%]
tests/tensorflow/test_nn.py::test_edge_conv[1-g1-idtype1] PASSED         [ 95%]
tests/tensorflow/test_nn.py::test_edge_conv[1-g2-idtype0] PASSED         [ 95%]
tests/tensorflow/test_nn.py::test_edge_conv[1-g2-idtype1] PASSED         [ 95%]
tests/tensorflow/test_nn.py::test_edge_conv[1-g3-idtype0] PASSED         [ 95%]
tests/tensorflow/test_nn.py::test_edge_conv[1-g3-idtype1] PASSED         [ 95%]
tests/tensorflow/test_nn.py::test_edge_conv[1-g4-idtype0] PASSED         [ 95%]
tests/tensorflow/test_nn.py::test_edge_conv[1-g4-idtype1] PASSED         [ 95%]
tests/tensorflow/test_nn.py::test_edge_conv[1-g5-idtype0] PASSED         [ 95%]
tests/tensorflow/test_nn.py::test_edge_conv[1-g5-idtype1] PASSED         [ 95%]
tests/tensorflow/test_nn.py::test_edge_conv[1-g6-idtype0] PASSED         [ 95%]
tests/tensorflow/test_nn.py::test_edge_conv[1-g6-idtype1] PASSED         [ 96%]
tests/tensorflow/test_nn.py::test_edge_conv[2-g0-idtype0] PASSED         [ 96%]
tests/tensorflow/test_nn.py::test_edge_conv[2-g0-idtype1] PASSED         [ 96%]
tests/tensorflow/test_nn.py::test_edge_conv[2-g1-idtype0] PASSED         [ 96%]
tests/tensorflow/test_nn.py::test_edge_conv[2-g1-idtype1] PASSED         [ 96%]
tests/tensorflow/test_nn.py::test_edge_conv[2-g2-idtype0] PASSED         [ 96%]
tests/tensorflow/test_nn.py::test_edge_conv[2-g2-idtype1] PASSED         [ 96%]
tests/tensorflow/test_nn.py::test_edge_conv[2-g3-idtype0] PASSED         [ 96%]
tests/tensorflow/test_nn.py::test_edge_conv[2-g3-idtype1] PASSED         [ 96%]
tests/tensorflow/test_nn.py::test_edge_conv[2-g4-idtype0] PASSED         [ 97%]
tests/tensorflow/test_nn.py::test_edge_conv[2-g4-idtype1] PASSED         [ 97%]
tests/tensorflow/test_nn.py::test_edge_conv[2-g5-idtype0] PASSED         [ 97%]
tests/tensorflow/test_nn.py::test_edge_conv[2-g5-idtype1] PASSED         [ 97%]
tests/tensorflow/test_nn.py::test_edge_conv[2-g6-idtype0] PASSED         [ 97%]
tests/tensorflow/test_nn.py::test_edge_conv[2-g6-idtype1] PASSED         [ 97%]
tests/tensorflow/test_nn.py::test_edge_conv_bi[1-g0-idtype0] PASSED      [ 97%]
tests/tensorflow/test_nn.py::test_edge_conv_bi[1-g0-idtype1] PASSED      [ 97%]
tests/tensorflow/test_nn.py::test_edge_conv_bi[1-g1-idtype0] PASSED      [ 97%]
tests/tensorflow/test_nn.py::test_edge_conv_bi[1-g1-idtype1] PASSED      [ 98%]
tests/tensorflow/test_nn.py::test_edge_conv_bi[2-g0-idtype0] PASSED      [ 98%]
tests/tensorflow/test_nn.py::test_edge_conv_bi[2-g0-idtype1] PASSED      [ 98%]
tests/tensorflow/test_nn.py::test_edge_conv_bi[2-g1-idtype0] PASSED      [ 98%]
tests/tensorflow/test_nn.py::test_edge_conv_bi[2-g1-idtype1] PASSED      [ 98%]
tests/tensorflow/test_nn.py::test_hetero_conv[sum-idtype0] PASSED        [ 98%]
tests/tensorflow/test_nn.py::test_hetero_conv[sum-idtype1] PASSED        [ 98%]
tests/tensorflow/test_nn.py::test_hetero_conv[max-idtype0] PASSED        [ 98%]
tests/tensorflow/test_nn.py::test_hetero_conv[max-idtype1] PASSED        [ 98%]
tests/tensorflow/test_nn.py::test_hetero_conv[min-idtype0] PASSED        [ 99%]
tests/tensorflow/test_nn.py::test_hetero_conv[min-idtype1] PASSED        [ 99%]
tests/tensorflow/test_nn.py::test_hetero_conv[mean-idtype0] PASSED       [ 99%]
tests/tensorflow/test_nn.py::test_hetero_conv[mean-idtype1] Cloning into '/root/jenkins/workspace/dgl_PR-4648@2/third_party/tvm/3rdparty/rang'...
PASSED       [ 99%]
tests/tensorflow/test_nn.py::test_hetero_conv[stack-idtype0] PASSED      [ 99%]
tests/tensorflow/test_nn.py::test_hetero_conv[stack-idtype1] PASSED      [ 99%]
tests/tensorflow/test_nn.py::test_hetero_conv[myagg-idtype0] PASSED      [ 99%]
tests/tensorflow/test_nn.py::test_hetero_conv[myagg-idtype1] PASSED      [ 99%]
tests/tensorflow/test_nn.py::test_dense_cheb_conv[1] PASSED              [ 99%]
tests/tensorflow/test_nn.py::test_dense_cheb_conv[2] PASSED              [100%]

=============================== warnings summary ===============================
../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/tensorflow/python/data/ops/iterator_ops.py:546
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/tensorflow/python/data/ops/iterator_ops.py:546: DeprecationWarning: Using or importing the ABCs from 'collections' instead of from 'collections.abc' is deprecated, and in 3.8 it will stop working
    class IteratorBase(collections.Iterator, trackable.Trackable,

../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/tensorflow/python/data/ops/dataset_ops.py:106
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/tensorflow/python/data/ops/dataset_ops.py:106: DeprecationWarning: Using or importing the ABCs from 'collections' instead of from 'collections.abc' is deprecated, and in 3.8 it will stop working
    class DatasetV2(collections.Iterable, tracking_base.Trackable,

../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/tensorflow/python/autograph/utils/testing.py:21
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/tensorflow/python/autograph/utils/testing.py:21: DeprecationWarning: the imp module is deprecated in favour of importlib; see the module's documentation for alternative uses
    import imp

../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:10
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:10: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    _nlv = LooseVersion(_np_version)

../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:11
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:11: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    _np_version_under1p16 = _nlv < LooseVersion("1.16")

../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:12
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:12: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    _np_version_under1p17 = _nlv < LooseVersion("1.17")

../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:13
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:13: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    _np_version_under1p18 = _nlv < LooseVersion("1.18")

../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:14
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:14: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    _np_version_under1p19 = _nlv < LooseVersion("1.19")

../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:15
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:15: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    _np_version_under1p20 = _nlv < LooseVersion("1.20")

../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/setuptools/_distutils/version.py:351
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/setuptools/_distutils/version.py:351: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    other = LooseVersion(other)

../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/pandas/compat/numpy/function.py:125
../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/pandas/compat/numpy/function.py:125
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/pandas/compat/numpy/function.py:125: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    if LooseVersion(_np_version) >= LooseVersion("1.17.0"):

../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/keras_preprocessing/image/utils.py:23
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/keras_preprocessing/image/utils.py:23: DeprecationWarning: NEAREST is deprecated and will be removed in Pillow 10 (2023-07-01). Use Resampling.NEAREST or Dither.NONE instead.
    'nearest': pil_image.NEAREST,

../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/keras_preprocessing/image/utils.py:24
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/keras_preprocessing/image/utils.py:24: DeprecationWarning: BILINEAR is deprecated and will be removed in Pillow 10 (2023-07-01). Use Resampling.BILINEAR instead.
    'bilinear': pil_image.BILINEAR,

../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/keras_preprocessing/image/utils.py:25
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/keras_preprocessing/image/utils.py:25: DeprecationWarning: BICUBIC is deprecated and will be removed in Pillow 10 (2023-07-01). Use Resampling.BICUBIC instead.
    'bicubic': pil_image.BICUBIC,

../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/keras_preprocessing/image/utils.py:28
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/keras_preprocessing/image/utils.py:28: DeprecationWarning: HAMMING is deprecated and will be removed in Pillow 10 (2023-07-01). Use Resampling.HAMMING instead.
    if hasattr(pil_image, 'HAMMING'):

../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/keras_preprocessing/image/utils.py:29
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/keras_preprocessing/image/utils.py:29: DeprecationWarning: HAMMING is deprecated and will be removed in Pillow 10 (2023-07-01). Use Resampling.HAMMING instead.
    _PIL_INTERPOLATION_METHODS['hamming'] = pil_image.HAMMING

../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/keras_preprocessing/image/utils.py:30
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/keras_preprocessing/image/utils.py:30: DeprecationWarning: BOX is deprecated and will be removed in Pillow 10 (2023-07-01). Use Resampling.BOX instead.
    if hasattr(pil_image, 'BOX'):

../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/keras_preprocessing/image/utils.py:31
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/keras_preprocessing/image/utils.py:31: DeprecationWarning: BOX is deprecated and will be removed in Pillow 10 (2023-07-01). Use Resampling.BOX instead.
    _PIL_INTERPOLATION_METHODS['box'] = pil_image.BOX

../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/keras_preprocessing/image/utils.py:33
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/keras_preprocessing/image/utils.py:33: DeprecationWarning: LANCZOS is deprecated and will be removed in Pillow 10 (2023-07-01). Use Resampling.LANCZOS instead.
    if hasattr(pil_image, 'LANCZOS'):

../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/keras_preprocessing/image/utils.py:34
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/keras_preprocessing/image/utils.py:34: DeprecationWarning: LANCZOS is deprecated and will be removed in Pillow 10 (2023-07-01). Use Resampling.LANCZOS instead.
    _PIL_INTERPOLATION_METHODS['lanczos'] = pil_image.LANCZOS

python/dgl/backend/tensorflow/tensor.py:15
python/dgl/backend/tensorflow/tensor.py:15
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/backend/tensorflow/tensor.py:15: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    if LooseVersion(tf.__version__) < LooseVersion("2.3.0"):

tests/tensorflow/test_nn.py::test_graph_conv[1]
tests/tensorflow/test_nn.py::test_graph_conv[2]
tests/tensorflow/test_nn.py::test_simple_pool
tests/tensorflow/test_nn.py::test_glob_att_pool
tests/tensorflow/test_nn.py::test_rgcn[1]
tests/tensorflow/test_nn.py::test_rgcn[2]
tests/tensorflow/test_nn.py::test_rgcn[8]
tests/tensorflow/test_nn.py::test_dense_cheb_conv[1]
tests/tensorflow/test_nn.py::test_dense_cheb_conv[2]
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/heterograph.py:72: DGLWarning: Recommend creating graphs by `dgl.graph(data)` instead of `dgl.DGLGraph(data)`.
    dgl_warning('Recommend creating graphs by `dgl.graph(data)`'

tests/tensorflow/test_nn.py::test_rgcn[1]
tests/tensorflow/test_nn.py::test_rgcn[2]
tests/tensorflow/test_nn.py::test_rgcn[8]
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/heterograph.py:84: DGLWarning: Keyword arguments ['readonly'] are deprecated in v0.5, and can be safely removed in all cases.
    ' removed in all cases.'.format(list(deprecate_kwargs.keys())))

-- Docs: https://docs.pytest.org/en/stable/how-to/capture-warnings.html
-- generated xml file: /root/jenkins/workspace/dgl_PR-4648/pytest_backend.xml --
============================ slowest 100 durations =============================
0.32s call     tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-none-g0-idtype0]
0.26s call     tests/tensorflow/test_nn.py::test_graph_conv[1]
0.12s call     tests/tensorflow/test_nn.py::test_rgcn[8]
0.11s call     tests/tensorflow/test_nn.py::test_rgcn[2]
0.06s call     tests/tensorflow/test_nn.py::test_rgcn[1]
0.06s call     tests/tensorflow/test_nn.py::test_hetero_conv[sum-idtype0]
0.05s call     tests/tensorflow/test_nn.py::test_hetero_conv[sum-idtype1]
0.05s call     tests/tensorflow/test_nn.py::test_hetero_conv[mean-idtype1]
0.05s call     tests/tensorflow/test_nn.py::test_hetero_conv[min-idtype1]
0.05s call     tests/tensorflow/test_nn.py::test_hetero_conv[max-idtype1]
0.05s call     tests/tensorflow/test_nn.py::test_hetero_conv[min-idtype0]
0.05s call     tests/tensorflow/test_nn.py::test_hetero_conv[myagg-idtype1]
0.05s call     tests/tensorflow/test_nn.py::test_hetero_conv[stack-idtype1]
0.05s call     tests/tensorflow/test_nn.py::test_hetero_conv[mean-idtype0]
0.05s call     tests/tensorflow/test_nn.py::test_hetero_conv[max-idtype0]
0.05s call     tests/tensorflow/test_nn.py::test_hetero_conv[myagg-idtype0]
0.05s call     tests/tensorflow/test_nn.py::test_hetero_conv[stack-idtype0]
0.04s call     tests/tensorflow/test_nn.py::test_sage_conv_bi_empty[1-mean-idtype0]
0.03s call     tests/tensorflow/test_nn.py::test_sage_conv_bi_empty[1-gcn-idtype0]
0.03s call     tests/tensorflow/test_nn.py::test_sage_conv_bi_empty[2-mean-idtype0]
0.03s call     tests/tensorflow/test_nn.py::test_sage_conv_bi_empty[2-gcn-idtype0]
0.03s call     tests/tensorflow/test_nn.py::test_sage_conv_bi_empty[1-pool-idtype0]
0.03s call     tests/tensorflow/test_nn.py::test_sage_conv_bi_empty[2-pool-idtype0]
0.03s call     tests/tensorflow/test_nn.py::test_sage_conv_bi_empty[1-gcn-idtype1]
0.03s call     tests/tensorflow/test_nn.py::test_sage_conv_bi_empty[1-mean-idtype1]
0.03s call     tests/tensorflow/test_nn.py::test_sage_conv_bi_empty[1-pool-idtype1]
0.03s call     tests/tensorflow/test_nn.py::test_sage_conv_bi_empty[2-pool-idtype1]
0.03s call     tests/tensorflow/test_nn.py::test_sage_conv_bi_empty[2-mean-idtype1]
0.03s call     tests/tensorflow/test_nn.py::test_sage_conv_bi_empty[2-gcn-idtype1]
0.03s call     tests/tensorflow/test_nn.py::test_simple_pool
0.03s call     tests/tensorflow/test_nn.py::test_gat_conv[1-1-g0-idtype0]
0.03s call     tests/tensorflow/test_nn.py::test_graph_conv[2]
0.02s call     tests/tensorflow/test_nn.py::test_gat_conv[1-2-g4-idtype0]
0.02s call     tests/tensorflow/test_nn.py::test_gat_conv[1-1-g4-idtype0]
0.02s call     tests/tensorflow/test_nn.py::test_gat_conv[4-1-g4-idtype0]
0.02s call     tests/tensorflow/test_nn.py::test_gat_conv[4-2-g4-idtype0]
0.02s call     tests/tensorflow/test_nn.py::test_gat_conv[1-2-g0-idtype0]
0.02s call     tests/tensorflow/test_nn.py::test_gat_conv[1-2-g2-idtype0]
0.02s call     tests/tensorflow/test_nn.py::test_gat_conv[1-2-g3-idtype0]
0.02s call     tests/tensorflow/test_nn.py::test_gat_conv[1-2-g4-idtype1]
0.02s call     tests/tensorflow/test_nn.py::test_gat_conv[4-1-g0-idtype0]
0.02s call     tests/tensorflow/test_nn.py::test_gat_conv[1-2-g6-idtype0]
0.02s call     tests/tensorflow/test_nn.py::test_gat_conv[1-1-g6-idtype0]
0.02s call     tests/tensorflow/test_nn.py::test_gat_conv[1-2-g5-idtype0]
0.02s call     tests/tensorflow/test_nn.py::test_gat_conv[1-1-g2-idtype0]
0.02s call     tests/tensorflow/test_nn.py::test_gat_conv[4-2-g0-idtype0]
0.02s call     tests/tensorflow/test_nn.py::test_gat_conv[4-2-g2-idtype0]
0.02s call     tests/tensorflow/test_nn.py::test_gat_conv[4-2-g4-idtype1]
0.02s call     tests/tensorflow/test_nn.py::test_gat_conv[4-1-g4-idtype1]
0.02s call     tests/tensorflow/test_nn.py::test_gat_conv[1-1-g4-idtype1]
0.02s call     tests/tensorflow/test_nn.py::test_gat_conv[1-1-g3-idtype0]
0.02s call     tests/tensorflow/test_nn.py::test_gat_conv[1-1-g5-idtype0]
0.02s call     tests/tensorflow/test_nn.py::test_gat_conv[1-1-g1-idtype0]
0.02s call     tests/tensorflow/test_nn.py::test_gat_conv[1-2-g1-idtype0]
0.02s call     tests/tensorflow/test_nn.py::test_gat_conv[4-2-g6-idtype0]
0.02s call     tests/tensorflow/test_nn.py::test_gat_conv[4-1-g5-idtype0]
0.02s call     tests/tensorflow/test_nn.py::test_gat_conv[4-1-g6-idtype0]
0.02s call     tests/tensorflow/test_nn.py::test_gat_conv[4-2-g3-idtype0]
0.02s call     tests/tensorflow/test_nn.py::test_gat_conv[4-1-g3-idtype0]
0.02s call     tests/tensorflow/test_nn.py::test_gat_conv[4-1-g2-idtype0]
0.02s call     tests/tensorflow/test_nn.py::test_gat_conv[4-2-g5-idtype0]
0.02s call     tests/tensorflow/test_nn.py::test_gat_conv[4-1-g1-idtype0]
0.02s call     tests/tensorflow/test_nn.py::test_gat_conv[4-2-g1-idtype0]
0.02s call     tests/tensorflow/test_nn.py::test_gat_conv[1-2-g6-idtype1]
0.02s call     tests/tensorflow/test_nn.py::test_gat_conv[1-2-g0-idtype1]
0.02s call     tests/tensorflow/test_nn.py::test_gat_conv[4-2-g6-idtype1]
0.02s call     tests/tensorflow/test_nn.py::test_gat_conv[1-1-g6-idtype1]
0.02s call     tests/tensorflow/test_nn.py::test_gat_conv[1-2-g3-idtype1]
0.02s call     tests/tensorflow/test_nn.py::test_gat_conv[4-1-g6-idtype1]
0.02s call     tests/tensorflow/test_nn.py::test_gat_conv[1-2-g2-idtype1]
0.02s call     tests/tensorflow/test_nn.py::test_gat_conv[1-1-g0-idtype1]
0.02s call     tests/tensorflow/test_nn.py::test_gat_conv[4-1-g0-idtype1]
0.02s call     tests/tensorflow/test_nn.py::test_gat_conv[1-2-g5-idtype1]
0.02s call     tests/tensorflow/test_nn.py::test_gat_conv[1-1-g3-idtype1]
0.02s call     tests/tensorflow/test_nn.py::test_gat_conv[4-2-g2-idtype1]
0.02s call     tests/tensorflow/test_nn.py::test_gat_conv[4-2-g5-idtype1]
0.02s call     tests/tensorflow/test_nn.py::test_gat_conv[1-1-g2-idtype1]
0.02s call     tests/tensorflow/test_nn.py::test_gat_conv[4-2-g3-idtype1]
0.02s call     tests/tensorflow/test_nn.py::test_gat_conv[4-1-g5-idtype1]
0.02s call     tests/tensorflow/test_nn.py::test_gat_conv[4-2-g0-idtype1]
0.02s call     tests/tensorflow/test_nn.py::test_gat_conv[1-1-g5-idtype1]
0.02s call     tests/tensorflow/test_nn.py::test_gat_conv[4-1-g3-idtype1]
0.02s call     tests/tensorflow/test_nn.py::test_gat_conv[4-1-g2-idtype1]
0.02s call     tests/tensorflow/test_nn.py::test_gat_conv[1-2-g1-idtype1]
0.02s call     tests/tensorflow/test_nn.py::test_gat_conv[1-1-g1-idtype1]
0.02s call     tests/tensorflow/test_nn.py::test_gat_conv[4-2-g1-idtype1]
0.02s call     tests/tensorflow/test_nn.py::test_gat_conv[4-1-g1-idtype1]
0.02s call     tests/tensorflow/test_nn.py::test_glob_att_pool
0.01s call     tests/tensorflow/test_nn.py::test_gat_conv_bi[1-2-g1-idtype0]
0.01s call     tests/tensorflow/test_nn.py::test_gat_conv_bi[1-1-g1-idtype0]
0.01s call     tests/tensorflow/test_nn.py::test_gat_conv_bi[4-2-g1-idtype0]
0.01s call     tests/tensorflow/test_nn.py::test_gat_conv_bi[4-1-g1-idtype0]
0.01s call     tests/tensorflow/test_nn.py::test_gat_conv_bi[4-1-g0-idtype0]
0.01s call     tests/tensorflow/test_nn.py::test_gat_conv_bi[1-1-g0-idtype0]
0.01s call     tests/tensorflow/test_nn.py::test_gat_conv_bi[4-2-g0-idtype0]
0.01s call     tests/tensorflow/test_nn.py::test_gat_conv_bi[1-2-g0-idtype0]
0.01s call     tests/tensorflow/test_nn.py::test_appnp_conv[g0-idtype0]
0.01s call     tests/tensorflow/test_nn.py::test_sgc_conv[1-g0-idtype0]
0.01s call     tests/tensorflow/test_nn.py::test_sgc_conv[1-g5-idtype0]
0.01s call     tests/tensorflow/test_nn.py::test_sgc_conv[2-g5-idtype0]
====================== 920 passed, 35 warnings in 13.00s =======================
Cloning into '/root/jenkins/workspace/dgl_PR-4648@2/third_party/tvm/3rdparty/vta-hw'...
[Pipeline] }
[Pipeline] // timeout
[Pipeline] }
[Pipeline] // stage
Post stage
[Pipeline] cleanWs
[WS-CLEANUP] Deleting project workspace...
[WS-CLEANUP] Deferred wipeout is disabled by the job configuration...
Submodule path 'third_party/tvm/3rdparty/dlpack': checked out '3ec04430e89a6834e5a1b99471f415fa939bf642'
Submodule path 'third_party/tvm/3rdparty/dmlc-core': checked out '6c401e242c59a1f4c913918246591bb13fd714e7'
Submodule path 'third_party/tvm/3rdparty/rang': checked out 'cabe04d6d6b05356fa8f9741704924788f0dd762'
Submodule path 'third_party/tvm/3rdparty/vta-hw': checked out '87ce9acfae550d1a487746e9d06c2e250076e54c'
[WS-CLEANUP] done
[Pipeline] }
$ docker stop --time=1 accf251fd0b56b1d65207f2fd238e906ccb6fddb8b2532d85c1dfc3329fb23a3
Submodule path 'third_party/xbyak': checked out '757e4063f6464740b8ff4a2cae9136d2f8458020'
$ docker rm -f accf251fd0b56b1d65207f2fd238e906ccb6fddb8b2532d85c1dfc3329fb23a3
[Pipeline] // withDockerContainer
[Pipeline] }
[Pipeline] unstash
[Pipeline] // withEnv
[Pipeline] }
Running on dglci-manual-gpu-worker in /root/jenkins/workspace/dgl_PR-4648
[Pipeline] // node
[Pipeline] {
[Pipeline] }
[Pipeline] // stage
[Pipeline] }
[Pipeline] checkout
The recommended git tool is: git
Cloning into '/root/jenkins/workspace/dgl_PR-4648@3/third_party/nanoflann'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648@3/third_party/nccl'...
using credential 150de63f-189c-4717-bcaf-010460d2f51a
Cloning the remote Git repository
Cloning with configured refspecs honoured and without tags
[Pipeline] echo
Unpacked build/libdgl.so, build/runUnitTests, python/dgl/_ffi/_cy3/core.cpython-*-x86_64-linux-gnu.so, build/tensoradapter/pytorch/*.so from dgl-cpu-linux
[Pipeline] timeout
Timeout set to expire in 30 min
[Pipeline] {
[Pipeline] sh
Cloning into '/root/jenkins/workspace/dgl_PR-4648@3/third_party/phmap'...
+ bash tests/scripts/task_unit_test.sh tensorflow cpu
Cloning into '/root/jenkins/workspace/dgl_PR-4648@3/third_party/tensorpipe'...
Cloning repository https://github.com/dmlc/dgl.git
 > git init /root/jenkins/workspace/dgl_PR-4648 # timeout=10
Fetching upstream changes from https://github.com/dmlc/dgl.git
 > git --version # timeout=10
 > git --version # 'git version 2.17.1'
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git fetch --no-tags --progress -- https://github.com/dmlc/dgl.git +refs/pull/4648/head:refs/remotes/origin/PR-4648 +refs/heads/master:refs/remotes/origin/master # timeout=10
Requirement already satisfied: pytest in /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages (7.1.2)
Collecting psutil
  Downloading psutil-5.9.2-cp37-cp37m-manylinux_2_12_x86_64.manylinux2010_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl (281 kB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 281.3/281.3 kB 11.7 MB/s eta 0:00:00
Collecting pyyaml
  Downloading PyYAML-6.0-cp37-cp37m-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_12_x86_64.manylinux2010_x86_64.whl (596 kB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 596.3/596.3 kB 27.7 MB/s eta 0:00:00
Cloning into '/root/jenkins/workspace/dgl_PR-4648@3/third_party/thrust'...
Cleaning workspace
Fetching without tags
Collecting pydantic
  Downloading pydantic-1.10.2-cp37-cp37m-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (11.8 MB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 11.8/11.8 MB 115.9 MB/s eta 0:00:00
Merging remotes/origin/master commit d78a3a4baf611b90871a849f58647160c7cd9ab4 into PR head commit 37ed78035f406940ed88a4b60cadd2b7cfde5fea
 > git config remote.origin.url https://github.com/dmlc/dgl.git # timeout=10
 > git config --add remote.origin.fetch +refs/pull/4648/head:refs/remotes/origin/PR-4648 # timeout=10
 > git config --add remote.origin.fetch +refs/heads/master:refs/remotes/origin/master # timeout=10
 > git config remote.origin.url https://github.com/dmlc/dgl.git # timeout=10
 > git rev-parse --verify HEAD # timeout=10
No valid HEAD. Skipping the resetting
 > git clean -fdx # timeout=10
Fetching upstream changes from https://github.com/dmlc/dgl.git
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git fetch --no-tags --progress -- https://github.com/dmlc/dgl.git +refs/pull/4648/head:refs/remotes/origin/PR-4648 +refs/heads/master:refs/remotes/origin/master # timeout=10
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
Collecting pandas
  Using cached pandas-1.1.5-cp37-cp37m-manylinux1_x86_64.whl (9.5 MB)
Merge succeeded, producing 37ed78035f406940ed88a4b60cadd2b7cfde5fea
Checking out Revision 37ed78035f406940ed88a4b60cadd2b7cfde5fea (PR-4648)
Collecting rdflib
  Downloading rdflib-6.2.0-py3-none-any.whl (500 kB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 500.3/500.3 kB 52.2 MB/s eta 0:00:00
Collecting ogb
  Downloading ogb-1.3.4-py3-none-any.whl (78 kB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 78.6/78.6 kB 13.5 MB/s eta 0:00:00
Requirement already satisfied: iniconfig in /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages (from pytest) (1.1.1)
Requirement already satisfied: py>=1.8.2 in /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages (from pytest) (1.11.0)
Requirement already satisfied: pluggy<2.0,>=0.12 in /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages (from pytest) (1.0.0)
Requirement already satisfied: attrs>=19.2.0 in /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages (from pytest) (22.1.0)
Requirement already satisfied: packaging in /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages (from pytest) (21.3)
Requirement already satisfied: importlib-metadata>=0.12 in /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages (from pytest) (4.12.0)
Requirement already satisfied: tomli>=1.0.0 in /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages (from pytest) (2.0.1)
Requirement already satisfied: typing-extensions>=4.1.0 in /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages (from pydantic) (4.3.0)
Commit message: "fix for pytorch < 1.12"
Cleaning workspace
[       OK ] SocketCommunicatorTest.SendAndRecv (7005 ms)
[ RUN      ] SocketCommunicatorTest.SendAndRecvTimeout
Collecting pytz>=2017.2
  Using cached pytz-2022.2.1-py2.py3-none-any.whl (500 kB)
Requirement already satisfied: numpy>=1.15.4 in /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages (from pandas) (1.18.5)
Requirement already satisfied: python-dateutil>=2.7.3 in /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages (from pandas) (2.8.2)
Collecting isodate
  Downloading isodate-0.6.1-py2.py3-none-any.whl (41 kB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 41.7/41.7 kB 6.8 MB/s eta 0:00:00
Requirement already satisfied: pyparsing in /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages (from rdflib) (3.0.9)
Requirement already satisfied: setuptools in /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages (from rdflib) (61.2.0)
Collecting torch>=1.6.0
  Downloading torch-1.12.1-cp37-cp37m-manylinux1_x86_64.whl (776.3 MB)
[Pipeline] withEnv
[Pipeline] {
[Pipeline] isUnix
[Pipeline] withEnv
[Pipeline] {
[Pipeline] sh
Cloning into '/root/jenkins/workspace/dgl_PR-4648@3/third_party/tvm'...
 > git remote # timeout=10
 > git config --get remote.origin.url # timeout=10
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git merge d78a3a4baf611b90871a849f58647160c7cd9ab4 # timeout=10
 > git rev-parse HEAD^{commit} # timeout=10
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git branch -a -v --no-abbrev # timeout=10
 > git checkout -b PR-4648 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git rev-parse --verify HEAD # timeout=10
Resetting working tree
 > git reset --hard # timeout=10
 > git clean -fdx # timeout=10
+ docker pull dgllib/dgl-ci-gpu:cu101_v220816
[       OK ] SocketCommunicatorTest.SendAndRecvTimeout (2001 ms)
[ RUN      ] SocketCommunicatorTest.TCPSocketBind
[       OK ] SocketCommunicatorTest.TCPSocketBind (1 ms)
[----------] 3 tests from SocketCommunicatorTest (9007 ms total)

[----------] 3 tests from SplitStringTest
[ RUN      ] SplitStringTest.SplitStringUsingCompoundDelim
[       OK ] SplitStringTest.SplitStringUsingCompoundDelim (0 ms)
[ RUN      ] SplitStringTest.testSplitStringUsingSingleDelim
[       OK ] SplitStringTest.testSplitStringUsingSingleDelim (0 ms)
[ RUN      ] SplitStringTest.testSplitingNoDelimString
[       OK ] SplitStringTest.testSplitingNoDelimString (0 ms)
[----------] 3 tests from SplitStringTest (0 ms total)

[----------] 1 test from StringPrintf
[ RUN      ] StringPrintf.normal
[       OK ] StringPrintf.normal (0 ms)
[----------] 1 test from StringPrintf (0 ms total)

[----------] 13 tests from ArrayTest
[ RUN      ] ArrayTest.TestCreate
[       OK ] ArrayTest.TestCreate (0 ms)
[ RUN      ] ArrayTest.TestRange
[       OK ] ArrayTest.TestRange (0 ms)
[ RUN      ] ArrayTest.TestFull
[       OK ] ArrayTest.TestFull (0 ms)
[ RUN      ] ArrayTest.TestClone
[       OK ] ArrayTest.TestClone (0 ms)
[ RUN      ] ArrayTest.TestAsNumBits
[       OK ] ArrayTest.TestAsNumBits (0 ms)
[ RUN      ] ArrayTest.Arith
[       OK ] ArrayTest.Arith (0 ms)
[ RUN      ] ArrayTest.HStack
[       OK ] ArrayTest.HStack (0 ms)
[ RUN      ] ArrayTest.TestIndexSelect
[       OK ] ArrayTest.TestIndexSelect (0 ms)
[ RUN      ] ArrayTest.TestRelabel_
[       OK ] ArrayTest.TestRelabel_ (0 ms)
[ RUN      ] ArrayTest.CumSum
[       OK ] ArrayTest.CumSum (1 ms)
[ RUN      ] ArrayTest.Scatter_
[       OK ] ArrayTest.Scatter_ (0 ms)
[ RUN      ] ArrayTest.NonZero
[       OK ] ArrayTest.NonZero (0 ms)
[ RUN      ] ArrayTest.Sort
[       OK ] ArrayTest.Sort (0 ms)
[----------] 13 tests from ArrayTest (1 ms total)

[----------] 2 tests from MatrixTest
[ RUN      ] MatrixTest.TestToSimpleCsr
[       OK ] MatrixTest.TestToSimpleCsr (3 ms)
[ RUN      ] MatrixTest.TestToSimpleCoo
[       OK ] MatrixTest.TestToSimpleCoo (0 ms)
[----------] 2 tests from MatrixTest (3 ms total)

[----------] 2 tests from DisjointUnionTest
[ RUN      ] DisjointUnionTest.TestDisjointUnionPartitionCoo
[       OK ] DisjointUnionTest.TestDisjointUnionPartitionCoo (0 ms)
[ RUN      ] DisjointUnionTest.TestDisjointUnionPartitionCsr
[       OK ] DisjointUnionTest.TestDisjointUnionPartitionCsr (1 ms)
[----------] 2 tests from DisjointUnionTest (1 ms total)

[----------] 2 tests from SliceContiguousChunk
[ RUN      ] SliceContiguousChunk.TestSliceContiguousChunkCoo
[       OK ] SliceContiguousChunk.TestSliceContiguousChunkCoo (0 ms)
[ RUN      ] SliceContiguousChunk.TestSliceContiguousChunkCsr
[       OK ] SliceContiguousChunk.TestSliceContiguousChunkCsr (0 ms)
[----------] 2 tests from SliceContiguousChunk (0 ms total)

[----------] 2 tests from MatrixUnionTest
[ RUN      ] MatrixUnionTest.TestMatrixUnionCsr
[       OK ] MatrixUnionTest.TestMatrixUnionCsr (0 ms)
[ RUN      ] MatrixUnionTest.TestMatrixUnionCoo
[       OK ] MatrixUnionTest.TestMatrixUnionCoo (0 ms)
[----------] 2 tests from MatrixUnionTest (0 ms total)

[----------] 1 test from LineGraphTest
[ RUN      ] LineGraphTest.LineGraphCOO
[       OK ] LineGraphTest.LineGraphCOO (0 ms)
[----------] 1 test from LineGraphTest (0 ms total)

[----------] 3 tests from CsrmmTest
[ RUN      ] CsrmmTest.TestCsrmm
[       OK ] CsrmmTest.TestCsrmm (3 ms)
[ RUN      ] CsrmmTest.TestCsrsum
[       OK ] CsrmmTest.TestCsrsum (0 ms)
[ RUN      ] CsrmmTest.TestCsrmask
[       OK ] CsrmmTest.TestCsrmask (4 ms)
[----------] 3 tests from CsrmmTest (7 ms total)

[----------] 2 tests from PartitionTest
[ RUN      ] PartitionTest.TestRemainderPartition
[       OK ] PartitionTest.TestRemainderPartition (0 ms)
[ RUN      ] PartitionTest.TestRangePartition
[       OK ] PartitionTest.TestRangePartition (0 ms)
[----------] 2 tests from PartitionTest (0 ms total)

[----------] 11 tests from RowwiseTest
[ RUN      ] RowwiseTest.TestCSRSampling
cu101_v220816: Pulling from dgllib/dgl-ci-gpu
Digest: sha256:ca40fc52876a2563a4e904d0c271d658c1acc8e6a4f8611b578bb49f8c7fd925
Status: Image is up to date for dgllib/dgl-ci-gpu:cu101_v220816
docker.io/dgllib/dgl-ci-gpu:cu101_v220816
[Pipeline] }
[Pipeline] // withEnv
[Pipeline] isUnix
[Pipeline] withEnv
[Pipeline] {
[Pipeline] sh
+ docker inspect -f . dgllib/dgl-ci-gpu:cu101_v220816
.
[Pipeline] }
[Pipeline] // withEnv
[Pipeline] withDockerContainer
dglci-manual-gpu-worker does not seem to be running inside a container
$ docker run -t -d -u 0:0 --runtime nvidia --shm-size=8gb -w /root/jenkins/workspace/dgl_PR-4648 -v /root/jenkins/workspace/dgl_PR-4648:/root/jenkins/workspace/dgl_PR-4648:rw,z -v /root/jenkins/workspace/dgl_PR-4648@tmp:/root/jenkins/workspace/dgl_PR-4648@tmp:rw,z -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** dgllib/dgl-ci-gpu:cu101_v220816 cat
$ docker top f9b6859df1d65887fd5c1562bda6c9300bf7673dd8efba82320500053c49173b -eo pid,comm
[Pipeline] {
[Pipeline] stage
[Pipeline] { (Torch GPU Unit test)
[Pipeline] sh
+ nvidia-smi
Tue Sep 27 05:24:00 2022       
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 510.47.03    Driver Version: 510.47.03    CUDA Version: 11.6     |
|-------------------------------+----------------------+----------------------+
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|                               |                      |               MIG M. |
|===============================+======================+======================|
|   0  Tesla T4            On   | 00000000:00:1E.0 Off |                    0 |
| N/A   45C    P8    16W /  70W |      0MiB / 15360MiB |      0%      Default |
|                               |                      |                  N/A |
+-------------------------------+----------------------+----------------------+
                                                                               
+-----------------------------------------------------------------------------+
| Processes:                                                                  |
|  GPU   GI   CI        PID   Type   Process name                  GPU Memory |
|        ID   ID                                                   Usage      |
|=============================================================================|
|  No running processes found                                                 |
+-----------------------------------------------------------------------------+
[Pipeline] sh
+ rm -rf CMakeLists.txt CONTRIBUTORS.md Jenkinsfile LICENSE NEWS.md README.md apps benchmarks cmake conda dglgo docker docs examples featgraph include pyproject.toml python readthedocs.yml src tensoradapter tests third_party tools tutorials
[Pipeline] checkout
The recommended git tool is: git
using credential 150de63f-189c-4717-bcaf-010460d2f51a
Warning: JENKINS-30600: special launcher org.jenkinsci.plugins.docker.workflow.WithContainerStep$Decorator$1@65304b80; decorates RemoteLauncher[hudson.remoting.Channel@41a64267:dglci-manual-gpu-worker] will be ignored (a typical symptom is the Git executable not being run inside a designated container)
Fetching changes from the remote Git repository
Cleaning workspace
Fetching without tags
Merging remotes/origin/master commit d78a3a4baf611b90871a849f58647160c7cd9ab4 into PR head commit 37ed78035f406940ed88a4b60cadd2b7cfde5fea
Merge succeeded, producing 37ed78035f406940ed88a4b60cadd2b7cfde5fea
Checking out Revision 37ed78035f406940ed88a4b60cadd2b7cfde5fea (PR-4648)
Commit message: "fix for pytorch < 1.12"
Cleaning workspace
 > git rev-parse --resolve-git-dir /root/jenkins/workspace/dgl_PR-4648/.git # timeout=10
 > git config remote.origin.url https://github.com/dmlc/dgl.git # timeout=10
 > git rev-parse --verify HEAD # timeout=10
Resetting working tree
 > git reset --hard # timeout=10
 > git clean -fdx # timeout=10
Fetching upstream changes from https://github.com/dmlc/dgl.git
 > git --version # timeout=10
 > git --version # 'git version 2.17.1'
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git fetch --no-tags --progress -- https://github.com/dmlc/dgl.git +refs/pull/4648/head:refs/remotes/origin/PR-4648 +refs/heads/master:refs/remotes/origin/master # timeout=10
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git remote # timeout=10
 > git config --get remote.origin.url # timeout=10
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git merge d78a3a4baf611b90871a849f58647160c7cd9ab4 # timeout=10
 > git rev-parse HEAD^{commit} # timeout=10
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git branch -a -v --no-abbrev # timeout=10
 > git branch -D PR-4648 # timeout=10
 > git checkout -b PR-4648 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git rev-parse --verify HEAD # timeout=10
Resetting working tree
 > git reset --hard # timeout=10
 > git clean -fdx # timeout=10
[Pipeline] sh
+ git submodule update --recursive --init
Submodule 'third_party/METIS' (https://github.com/KarypisLab/METIS.git) registered for path 'third_party/METIS'
Submodule 'third_party/dlpack' (https://github.com/dmlc/dlpack.git) registered for path 'third_party/dlpack'
Submodule 'third_party/dmlc-core' (https://github.com/dmlc/dmlc-core.git) registered for path 'third_party/dmlc-core'
Submodule 'third_party/googletest' (https://github.com/google/googletest.git) registered for path 'third_party/googletest'
Submodule 'third_party/libxsmm' (https://github.com/hfp/libxsmm.git) registered for path 'third_party/libxsmm'
Submodule 'third_party/nanoflann' (https://github.com/jlblancoc/nanoflann) registered for path 'third_party/nanoflann'
Submodule 'third_party/nccl' (https://github.com/nvidia/nccl) registered for path 'third_party/nccl'
Submodule 'third_party/phmap' (https://github.com/greg7mdp/parallel-hashmap.git) registered for path 'third_party/phmap'
Submodule 'third_party/tensorpipe' (https://github.com/pytorch/tensorpipe) registered for path 'third_party/tensorpipe'
Submodule 'third_party/thrust' (https://github.com/NVIDIA/thrust.git) registered for path 'third_party/thrust'
Submodule 'third_party/tvm' (https://github.com/apache/incubator-tvm) registered for path 'third_party/tvm'
Submodule 'third_party/xbyak' (https://github.com/herumi/xbyak) registered for path 'third_party/xbyak'
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/METIS'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/dlpack'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648@3/third_party/xbyak'...
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 776.3/776.3 MB 5.6 MB/s eta 0:00:00
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/dmlc-core'...
Submodule path 'third_party/METIS': checked out '10603482cc055626ff99d2ac3ab7e73d1119439d'
Submodule 'GKlib' (https://github.com/KarypisLab/GKlib.git) registered for path 'third_party/METIS/GKlib'
Cloning into '/root/jenkins/workspace/dgl_PR-4648@3/third_party/METIS/GKlib'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/googletest'...
Requirement already satisfied: six>=1.12.0 in /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages (from ogb) (1.16.0)
Requirement already satisfied: tqdm>=4.29.0 in /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages (from ogb) (4.64.0)
Collecting scikit-learn>=0.20.0
  Using cached scikit_learn-1.0.2-cp37-cp37m-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (24.8 MB)
Requirement already satisfied: urllib3>=1.24.0 in /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages (from ogb) (1.26.11)
Collecting outdated>=0.2.0
  Using cached outdated-0.2.1-py3-none-any.whl (7.5 kB)
Requirement already satisfied: zipp>=0.5 in /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages (from importlib-metadata>=0.12->pytest) (3.8.1)
Requirement already satisfied: requests in /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages (from outdated>=0.2.0->ogb) (2.28.1)
Collecting littleutils
  Using cached littleutils-0.2.2-py3-none-any.whl
Collecting threadpoolctl>=2.0.0
  Using cached threadpoolctl-3.1.0-py3-none-any.whl (14 kB)
Requirement already satisfied: scipy>=1.1.0 in /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages (from scikit-learn>=0.20.0->ogb) (1.4.1)
Requirement already satisfied: joblib>=0.11 in /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages (from scikit-learn>=0.20.0->ogb) (1.1.0)
Requirement already satisfied: charset-normalizer<3,>=2 in /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages (from requests->outdated>=0.2.0->ogb) (2.1.0)
Requirement already satisfied: certifi>=2017.4.17 in /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages (from requests->outdated>=0.2.0->ogb) (2022.6.15)
Requirement already satisfied: idna<4,>=2.5 in /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages (from requests->outdated>=0.2.0->ogb) (3.3)
Submodule path 'third_party/METIS/GKlib': checked out '62de20c7f97c719abdc96d7c3e266a1f7cb52bc5'
Submodule path 'third_party/dlpack': checked out 'e2bdd3bee8cb6501558042633fa59144cc8b7f5f'
Submodule path 'third_party/dmlc-core': checked out 'bfad207b448480783a1f428ae3d93d87032d8349'
Submodule path 'third_party/googletest': checked out 'f71fb4f9a912ec945401cc49a287a759b6131026'
Installing collected packages: pytz, littleutils, torch, threadpoolctl, pyyaml, pydantic, psutil, isodate, scikit-learn, rdflib, pandas, outdated, ogb
Submodule path 'third_party/libxsmm': checked out 'fa687556130b6298430f1c0555a14cf79ab6101c'
Submodule path 'third_party/nanoflann': checked out '4c47ca200209550c5628c89803591f8a753c8181'
Submodule path 'third_party/nccl': checked out 'e11238b3029795d33f958b5868d47c90c4f22628'
Submodule path 'third_party/phmap': checked out '25293cefd8b85491b45600c03fe8edf07647553f'
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/libxsmm'...
Submodule path 'third_party/tensorpipe': checked out '6042f1a4cbce8eef997f11ed0012de137b317361'
Submodule 'third_party/googletest' (https://github.com/google/googletest.git) registered for path 'third_party/tensorpipe/third_party/googletest'
Submodule 'third_party/libnop' (https://github.com/google/libnop.git) registered for path 'third_party/tensorpipe/third_party/libnop'
Submodule 'third_party/libuv' (https://github.com/libuv/libuv.git) registered for path 'third_party/tensorpipe/third_party/libuv'
Submodule 'third_party/pybind11' (https://github.com/pybind/pybind11.git) registered for path 'third_party/tensorpipe/third_party/pybind11'
Cloning into '/root/jenkins/workspace/dgl_PR-4648@3/third_party/tensorpipe/third_party/googletest'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648@3/third_party/tensorpipe/third_party/libnop'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648@3/third_party/tensorpipe/third_party/libuv'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648@3/third_party/tensorpipe/third_party/pybind11'...
Submodule path 'third_party/tensorpipe/third_party/googletest': checked out 'aee0f9d9b5b87796ee8a0ab26b7587ec30e8858e'
Submodule path 'third_party/tensorpipe/third_party/libnop': checked out 'aa95422ea8c409e3f078d2ee7708a5f59a8b9fa2'
Submodule path 'third_party/tensorpipe/third_party/libuv': checked out '1dff88e5161cba5c59276d2070d2e304e4dcb242'
Submodule path 'third_party/tensorpipe/third_party/pybind11': checked out 'a23996fce38ff6ccfbcdc09f1e63f2c4be5ea2ef'
Submodule 'tools/clang' (https://github.com/wjakob/clang-cindex-python3) registered for path 'third_party/tensorpipe/third_party/pybind11/tools/clang'
Cloning into '/root/jenkins/workspace/dgl_PR-4648@3/third_party/tensorpipe/third_party/pybind11/tools/clang'...
Submodule path 'third_party/tensorpipe/third_party/pybind11/tools/clang': checked out '6a00cbc4a9b8e68b71caf7f774b3f9c753ae84d5'
Submodule path 'third_party/thrust': checked out '6a3078c64cab0e2f276340fa5dcafa0d758ed890'
Submodule 'cub' (https://github.com/NVIDIA/cub.git) registered for path 'third_party/thrust/dependencies/cub'
Cloning into '/root/jenkins/workspace/dgl_PR-4648@3/third_party/thrust/dependencies/cub'...
Submodule path 'third_party/thrust/dependencies/cub': checked out 'cdaa9558a85e45d849016e5fe7b6e4ee79113f95'
Submodule path 'third_party/tvm': checked out 'b2e418cb109df4cd1f17a2cf2894a1b396a6b838'
Submodule 'dlpack' (https://github.com/dmlc/dlpack) registered for path 'third_party/tvm/3rdparty/dlpack'
Submodule 'dmlc-core' (https://github.com/dmlc/dmlc-core) registered for path 'third_party/tvm/3rdparty/dmlc-core'
Submodule '3rdparty/rang' (https://github.com/agauniyal/rang) registered for path 'third_party/tvm/3rdparty/rang'
Submodule '3rdparty/vta-hw' (https://github.com/apache/incubator-tvm-vta) registered for path 'third_party/tvm/3rdparty/vta-hw'
Cloning into '/root/jenkins/workspace/dgl_PR-4648@3/third_party/tvm/3rdparty/dlpack'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648@3/third_party/tvm/3rdparty/dmlc-core'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648@3/third_party/tvm/3rdparty/rang'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648@3/third_party/tvm/3rdparty/vta-hw'...
Submodule path 'third_party/tvm/3rdparty/dlpack': checked out '3ec04430e89a6834e5a1b99471f415fa939bf642'
Submodule path 'third_party/tvm/3rdparty/dmlc-core': checked out '6c401e242c59a1f4c913918246591bb13fd714e7'
Submodule path 'third_party/tvm/3rdparty/rang': checked out 'cabe04d6d6b05356fa8f9741704924788f0dd762'
Submodule path 'third_party/tvm/3rdparty/vta-hw': checked out '87ce9acfae550d1a487746e9d06c2e250076e54c'
Submodule path 'third_party/xbyak': checked out '757e4063f6464740b8ff4a2cae9136d2f8458020'
[Pipeline] unstash
[Pipeline] echo
Unpacked build/libdgl.so, build/runUnitTests, python/dgl/_ffi/_cy3/core.cpython-*-x86_64-linux-gnu.so, build/tensoradapter/pytorch/*.so from dgl-cpu-linux
[Pipeline] timeout
Timeout set to expire in 30 min
[Pipeline] {
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/nanoflann'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/nccl'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/phmap'...
[Pipeline] sh
+ bash tests/scripts/task_unit_test.sh pytorch cpu
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tensorpipe'...
Requirement already satisfied: pytest in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (7.1.2)
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/thrust'...
Collecting psutil
  Downloading psutil-5.9.2-cp37-cp37m-manylinux_2_12_x86_64.manylinux2010_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl (281 kB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 281.3/281.3 kB 12.7 MB/s eta 0:00:00
Collecting pyyaml
  Downloading PyYAML-6.0-cp37-cp37m-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_12_x86_64.manylinux2010_x86_64.whl (596 kB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 596.3/596.3 kB 45.6 MB/s eta 0:00:00
Collecting pydantic
  Downloading pydantic-1.10.2-cp37-cp37m-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (11.8 MB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 11.8/11.8 MB 143.8 MB/s eta 0:00:00
Requirement already satisfied: pandas in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (1.1.5)
Collecting rdflib
  Downloading rdflib-6.2.0-py3-none-any.whl (500 kB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 500.3/500.3 kB 71.4 MB/s eta 0:00:00
Requirement already satisfied: ogb in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (1.3.3)
Requirement already satisfied: attrs>=19.2.0 in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from pytest) (22.1.0)
Requirement already satisfied: pluggy<2.0,>=0.12 in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from pytest) (1.0.0)
Requirement already satisfied: packaging in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from pytest) (21.3)
Requirement already satisfied: py>=1.8.2 in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from pytest) (1.11.0)
Requirement already satisfied: tomli>=1.0.0 in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from pytest) (2.0.1)
Requirement already satisfied: importlib-metadata>=0.12 in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from pytest) (4.12.0)
Requirement already satisfied: iniconfig in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from pytest) (1.1.1)
Requirement already satisfied: typing-extensions>=4.1.0 in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from pydantic) (4.3.0)
Requirement already satisfied: numpy>=1.15.4 in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from pandas) (1.21.6)
Requirement already satisfied: pytz>=2017.2 in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from pandas) (2022.2.1)
Requirement already satisfied: python-dateutil>=2.7.3 in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from pandas) (2.8.2)
Requirement already satisfied: setuptools in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from rdflib) (61.2.0)
Requirement already satisfied: pyparsing in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from rdflib) (3.0.9)
Collecting isodate
  Downloading isodate-0.6.1-py2.py3-none-any.whl (41 kB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 41.7/41.7 kB 10.0 MB/s eta 0:00:00
Requirement already satisfied: scikit-learn>=0.20.0 in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from ogb) (1.0.2)
Requirement already satisfied: torch>=1.6.0 in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from ogb) (1.9.0+cpu)
Requirement already satisfied: six>=1.12.0 in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from ogb) (1.16.0)
Requirement already satisfied: outdated>=0.2.0 in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from ogb) (0.2.1)
Requirement already satisfied: urllib3>=1.24.0 in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from ogb) (1.26.11)
Requirement already satisfied: tqdm>=4.29.0 in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from ogb) (4.64.0)
Requirement already satisfied: zipp>=0.5 in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from importlib-metadata>=0.12->pytest) (3.8.1)
Requirement already satisfied: littleutils in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from outdated>=0.2.0->ogb) (0.2.2)
Requirement already satisfied: requests in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from outdated>=0.2.0->ogb) (2.28.1)
Requirement already satisfied: threadpoolctl>=2.0.0 in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from scikit-learn>=0.20.0->ogb) (3.1.0)
Requirement already satisfied: scipy>=1.1.0 in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from scikit-learn>=0.20.0->ogb) (1.7.3)
Requirement already satisfied: joblib>=0.11 in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from scikit-learn>=0.20.0->ogb) (1.1.0)
Requirement already satisfied: certifi>=2017.4.17 in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from requests->outdated>=0.2.0->ogb) (2022.6.15)
Requirement already satisfied: idna<4,>=2.5 in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from requests->outdated>=0.2.0->ogb) (3.3)
Requirement already satisfied: charset-normalizer<3,>=2 in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from requests->outdated>=0.2.0->ogb) (2.1.0)
Installing collected packages: pyyaml, pydantic, psutil, isodate, rdflib
Successfully installed isodate-0.6.1 psutil-5.9.2 pydantic-1.10.2 pyyaml-6.0 rdflib-6.2.0
WARNING: Running pip as the 'root' user can result in broken permissions and conflicting behaviour with the system package manager. It is recommended to use a virtual environment instead: https://pip.pypa.io/warnings/venv
[       OK ] RowwiseTest.TestCSRSampling (24230 ms)
[ RUN      ] RowwiseTest.TestCSRSamplingUniform
============================= test session starts ==============================
platform linux -- Python 3.7.0, pytest-7.1.2, pluggy-1.0.0 -- /opt/conda/envs/pytorch-ci/bin/python3
cachedir: .pytest_cache
rootdir: /root/jenkins/workspace/dgl_PR-4648@3
collecting ... Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tvm'...
[       OK ] RowwiseTest.TestCSRSamplingUniform (4560 ms)
[ RUN      ] RowwiseTest.TestCSRPerEtypeSampling
[       OK ] RowwiseTest.TestCSRPerEtypeSampling (20 ms)
[ RUN      ] RowwiseTest.TestCSRPerEtypeSamplingUniform
[       OK ] RowwiseTest.TestCSRPerEtypeSamplingUniform (8 ms)
[ RUN      ] RowwiseTest.TestCOOSampling
Successfully installed isodate-0.6.1 littleutils-0.2.2 ogb-1.3.4 outdated-0.2.1 pandas-1.1.5 psutil-5.9.2 pydantic-1.10.2 pytz-2022.2.1 pyyaml-6.0 rdflib-6.2.0 scikit-learn-1.0.2 threadpoolctl-3.1.0 torch-1.12.1
WARNING: Running pip as the 'root' user can result in broken permissions and conflicting behaviour with the system package manager. It is recommended to use a virtual environment instead: https://pip.pypa.io/warnings/venv
============================= test session starts ==============================
platform linux -- Python 3.7.0, pytest-7.1.2, pluggy-1.0.0 -- /opt/conda/envs/tensorflow-ci/bin/python3
cachedir: .pytest_cache
rootdir: /root/jenkins/workspace/dgl_PR-4648@2
collecting ... collected 2458 items

tests/compute/test_apply_edges_hetero.py::test_unary_copy_u[idtype0] FAILED [  0%]
tests/compute/test_apply_edges_hetero.py::test_unary_copy_u[idtype1] FAILED [  0%]
tests/compute/test_apply_edges_hetero.py::test_unary_copy_e[idtype0] PASSED [  0%]
tests/compute/test_apply_edges_hetero.py::test_unary_copy_e[idtype1] PASSED [  0%]
tests/compute/test_apply_edges_hetero.py::test_binary_op[idtype0] FAILED [  0%]
tests/compute/test_apply_edges_hetero.py::test_binary_op[idtype1] FAILED [  0%]
tests/compute/test_backend.py::test_set_default_backend PASSED           [  0%]
tests/compute/test_basics.py::test_compatible PASSED                     [  0%]
tests/compute/test_basics.py::test_batch_setter_getter[idtype0] PASSED   [  0%]
tests/compute/test_basics.py::test_batch_setter_getter[idtype1] PASSED   [  0%]
tests/compute/test_basics.py::test_batch_setter_autograd[idtype0] PASSED [  0%]
tests/compute/test_basics.py::test_batch_setter_autograd[idtype1] PASSED [  0%]
tests/compute/test_basics.py::test_apply_nodes[idtype0] PASSED           [  0%]
tests/compute/test_basics.py::test_apply_nodes[idtype1] PASSED           [  0%]
tests/compute/test_basics.py::test_apply_edges[idtype0] PASSED           [  0%]
tests/compute/test_basics.py::test_apply_edges[idtype1] PASSED           [  0%]
tests/compute/test_basics.py::test_update_routines[idtype0] PASSED       [  0%]
tests/compute/test_basics.py::test_update_routines[idtype1] Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/xbyak'...
Submodule path 'third_party/METIS': checked out '10603482cc055626ff99d2ac3ab7e73d1119439d'
Submodule 'GKlib' (https://github.com/KarypisLab/GKlib.git) registered for path 'third_party/METIS/GKlib'
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/METIS/GKlib'...
Submodule path 'third_party/METIS/GKlib': checked out '62de20c7f97c719abdc96d7c3e266a1f7cb52bc5'
Submodule path 'third_party/dlpack': checked out 'e2bdd3bee8cb6501558042633fa59144cc8b7f5f'
Submodule path 'third_party/dmlc-core': checked out 'bfad207b448480783a1f428ae3d93d87032d8349'
PASSED       [  0%]
tests/compute/test_basics.py::test_update_all_0deg[idtype0] PASSED       [  0%]
tests/compute/test_basics.py::test_update_all_0deg[idtype1] PASSED       [  0%]
tests/compute/test_basics.py::test_pull_0deg[idtype0] Submodule path 'third_party/googletest': checked out 'f71fb4f9a912ec945401cc49a287a759b6131026'
collected 2458 items

tests/compute/test_apply_edges_hetero.py::test_unary_copy_u[idtype0] SKIPPED [  0%]
tests/compute/test_apply_edges_hetero.py::test_unary_copy_u[idtype1] SKIPPED [  0%]
tests/compute/test_apply_edges_hetero.py::test_unary_copy_e[idtype0] SKIPPED [  0%]
tests/compute/test_apply_edges_hetero.py::test_unary_copy_e[idtype1] SKIPPED [  0%]
tests/compute/test_apply_edges_hetero.py::test_binary_op[idtype0] SKIPPED [  0%]
tests/compute/test_apply_edges_hetero.py::test_binary_op[idtype1] SKIPPED [  0%]
tests/compute/test_backend.py::test_set_default_backend PASSED           [  0%]
tests/compute/test_basics.py::test_compatible PASSED                     [  0%]
tests/compute/test_basics.py::test_batch_setter_getter[idtype0] PASSED   [  0%]
tests/compute/test_basics.py::test_batch_setter_getter[idtype1] PASSED   [  0%]
tests/compute/test_basics.py::test_batch_setter_autograd[idtype0] PASSED [  0%]
tests/compute/test_basics.py::test_batch_setter_autograd[idtype1] PASSED [  0%]
tests/compute/test_basics.py::test_apply_nodes[idtype0] PASSED           [  0%]
tests/compute/test_basics.py::test_apply_nodes[idtype1] PASSED           [  0%]
tests/compute/test_basics.py::test_apply_edges[idtype0] PASSED           [  0%]
tests/compute/test_basics.py::test_apply_edges[idtype1] PASSED           [  0%]
tests/compute/test_basics.py::test_update_routines[idtype0] PASSED             [  0%]
tests/compute/test_basics.py::test_pull_0deg[idtype1] PASSED             [  0%]
tests/compute/test_basics.py::test_dynamic_addition PASSED               [  0%]
tests/compute/test_basics.py::test_repr[idtype0] PASSED                  [  0%]
tests/compute/test_basics.py::test_repr[idtype1] PASSED                  [  1%]
tests/compute/test_basics.py::test_local_var[idtype0] PASSED             [  1%]
tests/compute/test_basics.py::test_local_var[idtype1] PASSED             [  1%]
tests/compute/test_basics.py::test_local_scope[idtype0] PASSED           [  1%]
tests/compute/test_basics.py::test_local_scope[idtype1] PASSED           [  1%]
tests/compute/test_basics.py::test_isolated_nodes[idtype0] PASSED        [  1%]
tests/compute/test_basics.py::test_isolated_nodes[idtype1] PASSED        [  1%]
tests/compute/test_basics.py::test_send_multigraph[idtype0] PASSED       [  1%]
tests/compute/test_basics.py::test_send_multigraph[idtype1] PASSED       [  1%]
tests/compute/test_basics.py::test_issue_1088[idtype0] Submodule path 'third_party/libxsmm': checked out 'fa687556130b6298430f1c0555a14cf79ab6101c'
Submodule path 'third_party/nanoflann': checked out '4c47ca200209550c5628c89803591f8a753c8181'
Submodule path 'third_party/nccl': checked out 'e11238b3029795d33f958b5868d47c90c4f22628'
Submodule path 'third_party/phmap': checked out '25293cefd8b85491b45600c03fe8edf07647553f'
Submodule path 'third_party/tensorpipe': checked out '6042f1a4cbce8eef997f11ed0012de137b317361'
Submodule 'third_party/googletest' (https://github.com/google/googletest.git) registered for path 'third_party/tensorpipe/third_party/googletest'
Submodule 'third_party/libnop' (https://github.com/google/libnop.git) registered for path 'third_party/tensorpipe/third_party/libnop'
Submodule 'third_party/libuv' (https://github.com/libuv/libuv.git) registered for path 'third_party/tensorpipe/third_party/libuv'
Submodule 'third_party/pybind11' (https://github.com/pybind/pybind11.git) registered for path 'third_party/tensorpipe/third_party/pybind11'
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tensorpipe/third_party/googletest'...
FAILED            [  1%]
tests/compute/test_basics.py::test_issue_1088[idtype1] FAILED            [  1%]
tests/compute/test_basics.py::test_degree_bucket_edge_ordering[idtype0] PASSED [  1%]
tests/compute/test_basics.py::test_degree_bucket_edge_ordering[idtype1] PASSED [  1%]
tests/compute/test_basics.py::test_issue_2484[idtype0] FAILED            [  1%]
tests/compute/test_basics.py::test_issue_2484[idtype1] FAILED            [  1%]
tests/compute/test_batched_graph.py::test_batch_unbatch[idtype0] PASSED  [  1%]
tests/compute/test_batched_graph.py::test_batch_unbatch[idtype1] PASSED  [  1%]
tests/compute/test_batched_graph.py::test_batch_unbatch1[idtype0] PASSED [  1%]
tests/compute/test_batched_graph.py::test_batch_unbatch1[idtype1] PASSED [  1%]
tests/compute/test_batched_graph.py::test_batch_unbatch_frame[idtype0] PASSED [  1%]
tests/compute/test_batched_graph.py::test_batch_unbatch_frame[idtype1] PASSED [  1%]
tests/compute/test_batched_graph.py::test_batch_unbatch2[idtype0] PASSED [  1%]
tests/compute/test_batched_graph.py::test_batch_unbatch2[idtype1] PASSED [  1%]
tests/compute/test_batched_graph.py::test_batch_send_and_recv[idtype0] PASSED       [  0%]
tests/compute/test_basics.py::test_update_routines[idtype1] PASSED [  1%]
tests/compute/test_batched_graph.py::test_batch_send_and_recv[idtype1] PASSED [  1%]
tests/compute/test_batched_graph.py::test_batch_propagate[idtype0] PASSED [  2%]
tests/compute/test_batched_graph.py::test_batch_propagate[idtype1] Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tensorpipe/third_party/libnop'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tensorpipe/third_party/libuv'...
PASSED [  2%]
tests/compute/test_batched_graph.py::test_batched_edge_ordering[idtype0] PASSED [  2%]
tests/compute/test_batched_graph.py::test_batched_edge_ordering[idtype1] PASSED [  2%]
tests/compute/test_batched_graph.py::test_batch_no_edge[idtype0] PASSED  [  2%]
tests/compute/test_batched_graph.py::test_batch_no_edge[idtype1] PASSED  [  2%]
tests/compute/test_batched_graph.py::test_batch_keeps_empty_data[idtype0] PASSED [  2%]
tests/compute/test_batched_graph.py::test_batch_keeps_empty_data[idtype1] PASSED [  2%]
tests/compute/test_batched_graph.py::test_set_batch_info[idtype0] PASSED [  2%]
tests/compute/test_batched_graph.py::test_set_batch_info[idtype1] PASSED [  2%]
tests/compute/test_batched_heterograph.py::test_topology[idtype0-gs0] PASSED [  2%]
tests/compute/test_batched_heterograph.py::test_topology[idtype0-gs1] PASSED [  2%]
tests/compute/test_batched_heterograph.py::test_topology[idtype1-gs0] PASSED [  2%]
tests/compute/test_batched_heterograph.py::test_topology[idtype1-gs1] PASSED [  2%]
tests/compute/test_batched_heterograph.py::test_batching_batched[idtype0] PASSED [  2%]
tests/compute/test_batched_heterograph.py::test_batching_batched[idtype1] PASSED [  2%]
tests/compute/test_batched_heterograph.py::test_features[idtype0] PASSED [  2%]
tests/compute/test_batched_heterograph.py::test_features[idtype1] PASSED [  2%]
tests/compute/test_batched_heterograph.py::test_empty_relation[idtype0] PASSED [  2%]
tests/compute/test_batched_heterograph.py::test_empty_relation[idtype1] PASSED [  2%]
tests/compute/test_batched_heterograph.py::test_unbatch2[idtype0] PASSED [  2%]
tests/compute/test_batched_heterograph.py::test_unbatch2[idtype1] PASSED [  2%]
tests/compute/test_batched_heterograph.py::test_slice_batch[idtype0] PASSED       [  0%]
tests/compute/test_basics.py::test_update_all_0deg[idtype0] PASSED [  2%]
tests/compute/test_batched_heterograph.py::test_slice_batch[idtype1] PASSED       [  0%]
tests/compute/test_basics.py::test_update_all_0deg[idtype1] Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tensorpipe/third_party/pybind11'...
PASSED [  2%]
tests/compute/test_batched_heterograph.py::test_batch_keeps_empty_data[idtype0] PASSED [  3%]
tests/compute/test_batched_heterograph.py::test_batch_keeps_empty_data[idtype1] PASSED [  3%]
tests/compute/test_batched_heterograph.py::test_batch_netypes PASSED     [  3%]
tests/compute/test_csrmm.py::test_csrmm[dtype0-idtype0] PASSED       [  0%]
tests/compute/test_basics.py::test_pull_0deg[idtype0] PASSED           [  3%]
tests/compute/test_csrmm.py::test_csrmm[dtype0-idtype1] PASSED             [  0%]
tests/compute/test_basics.py::test_pull_0deg[idtype1] PASSED           [  3%]
tests/compute/test_csrmm.py::test_csrmm[dtype1-idtype0] PASSED           [  3%]
tests/compute/test_csrmm.py::test_csrmm[dtype1-idtype1] PASSED             [  0%]
tests/compute/test_basics.py::test_dynamic_addition PASSED               [  0%]
tests/compute/test_basics.py::test_repr[idtype0] PASSED                  [  0%]
tests/compute/test_basics.py::test_repr[idtype1] PASSED                  [  1%]
tests/compute/test_basics.py::test_local_var[idtype0] Submodule path 'third_party/tensorpipe/third_party/googletest': checked out 'aee0f9d9b5b87796ee8a0ab26b7587ec30e8858e'
Submodule path 'third_party/tensorpipe/third_party/libnop': checked out 'aa95422ea8c409e3f078d2ee7708a5f59a8b9fa2'
Submodule path 'third_party/tensorpipe/third_party/libuv': checked out '1dff88e5161cba5c59276d2070d2e304e4dcb242'
Submodule path 'third_party/tensorpipe/third_party/pybind11': checked out 'a23996fce38ff6ccfbcdc09f1e63f2c4be5ea2ef'
Submodule 'tools/clang' (https://github.com/wjakob/clang-cindex-python3) registered for path 'third_party/tensorpipe/third_party/pybind11/tools/clang'
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tensorpipe/third_party/pybind11/tools/clang'...
PASSED             [  1%]
tests/compute/test_basics.py::test_local_var[idtype1] PASSED             [  1%]
tests/compute/test_basics.py::test_local_scope[idtype0] PASSED           [  1%]
tests/compute/test_basics.py::test_local_scope[idtype1] PASSED           [  3%]
tests/compute/test_csrmm.py::test_csrmm_backward[1-dtype0-idtype0] PASSED           [  1%]
tests/compute/test_basics.py::test_isolated_nodes[idtype0] PASSED        [  1%]
tests/compute/test_basics.py::test_isolated_nodes[idtype1] PASSED        [  1%]
tests/compute/test_basics.py::test_send_multigraph[idtype0] Submodule path 'third_party/tensorpipe/third_party/pybind11/tools/clang': checked out '6a00cbc4a9b8e68b71caf7f774b3f9c753ae84d5'
Submodule path 'third_party/thrust': checked out '6a3078c64cab0e2f276340fa5dcafa0d758ed890'
Submodule 'cub' (https://github.com/NVIDIA/cub.git) registered for path 'third_party/thrust/dependencies/cub'
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/thrust/dependencies/cub'...
PASSED [  3%]
tests/compute/test_csrmm.py::test_csrmm_backward[1-dtype0-idtype1] PASSED       [  1%]
tests/compute/test_basics.py::test_send_multigraph[idtype1] PASSED [  3%]
tests/compute/test_csrmm.py::test_csrmm_backward[1-dtype1-idtype0] PASSED       [  1%]
tests/compute/test_basics.py::test_issue_1088[idtype0] PASSED            [  1%]
tests/compute/test_basics.py::test_issue_1088[idtype1] PASSED            [  1%]
tests/compute/test_basics.py::test_degree_bucket_edge_ordering[idtype0] PASSED [  3%]
tests/compute/test_csrmm.py::test_csrmm_backward[1-dtype1-idtype1] PASSED [  1%]
tests/compute/test_basics.py::test_degree_bucket_edge_ordering[idtype1] PASSED [  3%]
tests/compute/test_csrmm.py::test_csrmm_backward[2-dtype0-idtype0] PASSED [  3%]
tests/compute/test_csrmm.py::test_csrmm_backward[2-dtype0-idtype1] PASSED [  1%]
tests/compute/test_basics.py::test_issue_2484[idtype0] PASSED [  3%]
tests/compute/test_csrmm.py::test_csrmm_backward[2-dtype1-idtype0] PASSED [  3%]
tests/compute/test_csrmm.py::test_csrmm_backward[2-dtype1-idtype1] Submodule path 'third_party/thrust/dependencies/cub': checked out 'cdaa9558a85e45d849016e5fe7b6e4ee79113f95'
Submodule path 'third_party/tvm': checked out 'b2e418cb109df4cd1f17a2cf2894a1b396a6b838'
Submodule 'dlpack' (https://github.com/dmlc/dlpack) registered for path 'third_party/tvm/3rdparty/dlpack'
Submodule 'dmlc-core' (https://github.com/dmlc/dmlc-core) registered for path 'third_party/tvm/3rdparty/dmlc-core'
Submodule '3rdparty/rang' (https://github.com/agauniyal/rang) registered for path 'third_party/tvm/3rdparty/rang'
Submodule '3rdparty/vta-hw' (https://github.com/apache/incubator-tvm-vta) registered for path 'third_party/tvm/3rdparty/vta-hw'
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tvm/3rdparty/dlpack'...
PASSED [  3%]
tests/compute/test_csrmm.py::test_csrsum[dtype0-idtype0] PASSED          [  3%]
tests/compute/test_csrmm.py::test_csrsum[dtype0-idtype1] PASSED            [  1%]
tests/compute/test_basics.py::test_issue_2484[idtype1] Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tvm/3rdparty/dmlc-core'...
PASSED          [  3%]
tests/compute/test_csrmm.py::test_csrsum[dtype1-idtype0] PASSED          [  3%]
tests/compute/test_csrmm.py::test_csrsum[dtype1-idtype1] PASSED            [  1%]
tests/compute/test_batched_graph.py::test_batch_unbatch[idtype0] PASSED  [  1%]
tests/compute/test_batched_graph.py::test_batch_unbatch[idtype1] PASSED  [  1%]
tests/compute/test_batched_graph.py::test_batch_unbatch1[idtype0] PASSED [  1%]
tests/compute/test_batched_graph.py::test_batch_unbatch1[idtype1] PASSED [  1%]
tests/compute/test_batched_graph.py::test_batch_unbatch_frame[idtype0] SKIPPED [  1%]
tests/compute/test_batched_graph.py::test_batch_unbatch_frame[idtype1] SKIPPED [  1%]
tests/compute/test_batched_graph.py::test_batch_unbatch2[idtype0] PASSED [  1%]
tests/compute/test_batched_graph.py::test_batch_unbatch2[idtype1] PASSED [  1%]
tests/compute/test_batched_graph.py::test_batch_send_and_recv[idtype0] PASSED          [  3%]
tests/compute/test_csrmm.py::test_csrsum_backward[1-dtype0-idtype0] PASSED [  3%]
tests/compute/test_csrmm.py::test_csrsum_backward[1-dtype0-idtype1] PASSED [  3%]
tests/compute/test_csrmm.py::test_csrsum_backward[1-dtype1-idtype0] PASSED [  1%]
tests/compute/test_batched_graph.py::test_batch_send_and_recv[idtype1] PASSED [  3%]
tests/compute/test_csrmm.py::test_csrsum_backward[1-dtype1-idtype1] PASSED [  3%]
tests/compute/test_csrmm.py::test_csrsum_backward[2-dtype0-idtype0] PASSED [  3%]
tests/compute/test_csrmm.py::test_csrsum_backward[2-dtype0-idtype1] PASSED [  3%]
tests/compute/test_csrmm.py::test_csrsum_backward[2-dtype1-idtype0] PASSED [  1%]
tests/compute/test_batched_graph.py::test_batch_propagate[idtype0] PASSED [  4%]
tests/compute/test_csrmm.py::test_csrsum_backward[2-dtype1-idtype1] PASSED [  4%]
tests/compute/test_csrmm.py::test_csrmask[9000-9000-dtype0-idtype0] PASSED [  4%]
tests/compute/test_csrmm.py::test_csrmask[9000-9000-dtype0-idtype1] PASSED [  4%]
tests/compute/test_csrmm.py::test_csrmask[9000-9000-dtype1-idtype0] Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tvm/3rdparty/rang'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tvm/3rdparty/vta-hw'...
PASSED [  2%]
tests/compute/test_batched_graph.py::test_batch_propagate[idtype1] PASSED [  4%]
tests/compute/test_csrmm.py::test_csrmask[9000-9000-dtype1-idtype1] PASSED [  4%]
tests/compute/test_csrmm.py::test_csrmask[9000-0-dtype0-idtype0] PASSED  [  4%]
tests/compute/test_csrmm.py::test_csrmask[9000-0-dtype0-idtype1] PASSED  [  4%]
tests/compute/test_csrmm.py::test_csrmask[9000-0-dtype1-idtype0] PASSED  [  4%]
tests/compute/test_csrmm.py::test_csrmask[9000-0-dtype1-idtype1] PASSED  [  4%]
tests/compute/test_csrmm.py::test_csrmask[0-9000-dtype0-idtype0] PASSED  [  4%]
tests/compute/test_csrmm.py::test_csrmask[0-9000-dtype0-idtype1] PASSED  [  4%]
tests/compute/test_csrmm.py::test_csrmask[0-9000-dtype1-idtype0] PASSED  [  4%]
tests/compute/test_csrmm.py::test_csrmask[0-9000-dtype1-idtype1] Submodule path 'third_party/tvm/3rdparty/dlpack': checked out '3ec04430e89a6834e5a1b99471f415fa939bf642'
Submodule path 'third_party/tvm/3rdparty/dmlc-core': checked out '6c401e242c59a1f4c913918246591bb13fd714e7'
Submodule path 'third_party/tvm/3rdparty/rang': checked out 'cabe04d6d6b05356fa8f9741704924788f0dd762'
Submodule path 'third_party/tvm/3rdparty/vta-hw': checked out '87ce9acfae550d1a487746e9d06c2e250076e54c'
PASSED  [  4%]
tests/compute/test_csrmm.py::test_csrmask[0-0-dtype0-idtype0] PASSED     [  4%]
tests/compute/test_csrmm.py::test_csrmask[0-0-dtype0-idtype1] PASSED     [  4%]
tests/compute/test_csrmm.py::test_csrmask[0-0-dtype1-idtype0] PASSED     [  4%]
tests/compute/test_csrmm.py::test_csrmask[0-0-dtype1-idtype1] PASSED     [  4%]
tests/compute/test_csrmm.py::test_csrmask_backward[dtype0-idtype0] Submodule path 'third_party/xbyak': checked out '757e4063f6464740b8ff4a2cae9136d2f8458020'
[Pipeline] unstash
PASSED [  4%]
tests/compute/test_csrmm.py::test_csrmask_backward[dtype0-idtype1] PASSED [  2%]
tests/compute/test_batched_graph.py::test_batched_edge_ordering[idtype0] PASSED [  4%]
tests/compute/test_csrmm.py::test_csrmask_backward[dtype1-idtype0] PASSED [  4%]
tests/compute/test_csrmm.py::test_csrmask_backward[dtype1-idtype1] PASSED [  2%]
tests/compute/test_batched_graph.py::test_batched_edge_ordering[idtype1] PASSED [  4%]
tests/compute/test_data.py::test_minigc PASSED [  2%]
tests/compute/test_batched_graph.py::test_batch_no_edge[idtype0] PASSED  [  2%]
tests/compute/test_batched_graph.py::test_batch_no_edge[idtype1] PASSED  [  2%]
tests/compute/test_batched_graph.py::test_batch_keeps_empty_data[idtype0] PASSED [  2%]
tests/compute/test_batched_graph.py::test_batch_keeps_empty_data[idtype1] PASSED [  2%]
tests/compute/test_batched_graph.py::test_set_batch_info[idtype0] PASSED                           [  4%]
tests/compute/test_data.py::test_gin PASSED [  2%]
tests/compute/test_batched_graph.py::test_set_batch_info[idtype1] PASSED [  2%]
tests/compute/test_batched_heterograph.py::test_topology[idtype0-gs0] PASSED [  2%]
tests/compute/test_batched_heterograph.py::test_topology[idtype0-gs1] PASSED [  2%]
tests/compute/test_batched_heterograph.py::test_topology[idtype1-gs0] PASSED [  2%]
tests/compute/test_batched_heterograph.py::test_topology[idtype1-gs1] PASSED [  2%]
tests/compute/test_batched_heterograph.py::test_batching_batched[idtype0] PASSED [  2%]
tests/compute/test_batched_heterograph.py::test_batching_batched[idtype1] PASSED [  2%]
tests/compute/test_batched_heterograph.py::test_features[idtype0] PASSED [  2%]
tests/compute/test_batched_heterograph.py::test_features[idtype1] PASSED [  2%]
tests/compute/test_batched_heterograph.py::test_empty_relation[idtype0] PASSED [  2%]
tests/compute/test_batched_heterograph.py::test_empty_relation[idtype1] PASSED [  2%]
tests/compute/test_batched_heterograph.py::test_unbatch2[idtype0] PASSED [  2%]
tests/compute/test_batched_heterograph.py::test_unbatch2[idtype1] PASSED [  2%]
tests/compute/test_batched_heterograph.py::test_slice_batch[idtype0] PASSED [  2%]
tests/compute/test_batched_heterograph.py::test_slice_batch[idtype1] PASSED [  2%]
tests/compute/test_batched_heterograph.py::test_batch_keeps_empty_data[idtype0] PASSED [  3%]
tests/compute/test_batched_heterograph.py::test_batch_keeps_empty_data[idtype1] PASSED [  3%]
tests/compute/test_batched_heterograph.py::test_batch_netypes PASSED     [  3%]
tests/compute/test_csrmm.py::test_csrmm[dtype0-idtype0] PASSED           [  3%]
tests/compute/test_csrmm.py::test_csrmm[dtype0-idtype1] PASSED           [  3%]
tests/compute/test_csrmm.py::test_csrmm[dtype1-idtype0] PASSED           [  3%]
tests/compute/test_csrmm.py::test_csrmm[dtype1-idtype1] PASSED           [  3%]
tests/compute/test_csrmm.py::test_csrmm_backward[1-dtype0-idtype0] PASSED [  3%]
tests/compute/test_csrmm.py::test_csrmm_backward[1-dtype0-idtype1] PASSED [  3%]
tests/compute/test_csrmm.py::test_csrmm_backward[1-dtype1-idtype0] PASSED [  3%]
tests/compute/test_csrmm.py::test_csrmm_backward[1-dtype1-idtype1] PASSED [  3%]
tests/compute/test_csrmm.py::test_csrmm_backward[2-dtype0-idtype0] [Pipeline] echo
Unpacked build/libdgl.so, build/runUnitTests, python/dgl/_ffi/_cy3/core.cpython-*-x86_64-linux-gnu.so, build/tensoradapter/pytorch/*.so from dgl-gpu-linux
[Pipeline] timeout
Timeout set to expire in 30 min
[Pipeline] {
[Pipeline] sh
PASSED [  3%]
tests/compute/test_csrmm.py::test_csrmm_backward[2-dtype0-idtype1] + bash tests/scripts/task_unit_test.sh pytorch gpu
PASSED [  3%]
tests/compute/test_csrmm.py::test_csrmm_backward[2-dtype1-idtype0] Requirement already satisfied: pytest in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (7.1.2)
Collecting psutil
  Downloading psutil-5.9.2-cp37-cp37m-manylinux_2_12_x86_64.manylinux2010_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl (281 kB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 281.3/281.3 kB 11.5 MB/s eta 0:00:00
Collecting pyyaml
  Downloading PyYAML-6.0-cp37-cp37m-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_12_x86_64.manylinux2010_x86_64.whl (596 kB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 596.3/596.3 kB 47.9 MB/s eta 0:00:00
PASSED [  3%]
tests/compute/test_csrmm.py::test_csrmm_backward[2-dtype1-idtype1] Collecting pydantic
  Downloading pydantic-1.10.2-cp37-cp37m-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (11.8 MB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 11.8/11.8 MB 118.6 MB/s eta 0:00:00
Requirement already satisfied: pandas in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (1.1.5)
Collecting rdflib
  Downloading rdflib-6.2.0-py3-none-any.whl (500 kB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 500.3/500.3 kB 64.3 MB/s eta 0:00:00
Requirement already satisfied: ogb in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (1.3.3)
Requirement already satisfied: pluggy<2.0,>=0.12 in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from pytest) (1.0.0)
Requirement already satisfied: packaging in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from pytest) (21.3)
Requirement already satisfied: tomli>=1.0.0 in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from pytest) (2.0.1)
Requirement already satisfied: py>=1.8.2 in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from pytest) (1.11.0)
Requirement already satisfied: iniconfig in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from pytest) (1.1.1)
Requirement already satisfied: importlib-metadata>=0.12 in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from pytest) (4.12.0)
Requirement already satisfied: attrs>=19.2.0 in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from pytest) (22.1.0)
Requirement already satisfied: typing-extensions>=4.1.0 in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from pydantic) (4.3.0)
Requirement already satisfied: python-dateutil>=2.7.3 in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from pandas) (2.8.2)
Requirement already satisfied: pytz>=2017.2 in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from pandas) (2022.2.1)
Requirement already satisfied: numpy>=1.15.4 in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from pandas) (1.21.6)
Requirement already satisfied: setuptools in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from rdflib) (61.2.0)
Collecting isodate
  Downloading isodate-0.6.1-py2.py3-none-any.whl (41 kB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 41.7/41.7 kB 8.5 MB/s eta 0:00:00
Requirement already satisfied: pyparsing in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from rdflib) (3.0.9)
Requirement already satisfied: tqdm>=4.29.0 in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from ogb) (4.64.0)
Requirement already satisfied: outdated>=0.2.0 in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from ogb) (0.2.1)
Requirement already satisfied: torch>=1.6.0 in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from ogb) (1.9.0+cu102)
Requirement already satisfied: urllib3>=1.24.0 in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from ogb) (1.26.11)
Requirement already satisfied: six>=1.12.0 in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from ogb) (1.16.0)
Requirement already satisfied: scikit-learn>=0.20.0 in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from ogb) (1.0.2)
Requirement already satisfied: zipp>=0.5 in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from importlib-metadata>=0.12->pytest) (3.8.1)
Requirement already satisfied: requests in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from outdated>=0.2.0->ogb) (2.28.1)
Requirement already satisfied: littleutils in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from outdated>=0.2.0->ogb) (0.2.2)
Requirement already satisfied: scipy>=1.1.0 in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from scikit-learn>=0.20.0->ogb) (1.7.3)
Requirement already satisfied: joblib>=0.11 in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from scikit-learn>=0.20.0->ogb) (1.1.0)
Requirement already satisfied: threadpoolctl>=2.0.0 in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from scikit-learn>=0.20.0->ogb) (3.1.0)
Requirement already satisfied: certifi>=2017.4.17 in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from requests->outdated>=0.2.0->ogb) (2022.6.15)
Requirement already satisfied: charset-normalizer<3,>=2 in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from requests->outdated>=0.2.0->ogb) (2.1.0)
Requirement already satisfied: idna<4,>=2.5 in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from requests->outdated>=0.2.0->ogb) (3.3)
Installing collected packages: pyyaml, pydantic, psutil, isodate, rdflib
PASSED [  3%]
tests/compute/test_csrmm.py::test_csrsum[dtype0-idtype0] PASSED          [  3%]
tests/compute/test_csrmm.py::test_csrsum[dtype0-idtype1] PASSED          [  3%]
tests/compute/test_csrmm.py::test_csrsum[dtype1-idtype0] PASSED          [  3%]
tests/compute/test_csrmm.py::test_csrsum[dtype1-idtype1] PASSED          [  3%]
tests/compute/test_csrmm.py::test_csrsum_backward[1-dtype0-idtype0] PASSED [  3%]
tests/compute/test_csrmm.py::test_csrsum_backward[1-dtype0-idtype1] Successfully installed isodate-0.6.1 psutil-5.9.2 pydantic-1.10.2 pyyaml-6.0 rdflib-6.2.0
WARNING: Running pip as the 'root' user can result in broken permissions and conflicting behaviour with the system package manager. It is recommended to use a virtual environment instead: https://pip.pypa.io/warnings/venv
PASSED [  3%]
tests/compute/test_csrmm.py::test_csrsum_backward[1-dtype1-idtype0] PASSED [  3%]
tests/compute/test_csrmm.py::test_csrsum_backward[1-dtype1-idtype1] ============================= test session starts ==============================
platform linux -- Python 3.7.0, pytest-7.1.2, pluggy-1.0.0 -- /opt/conda/envs/pytorch-ci/bin/python3
cachedir: .pytest_cache
rootdir: /root/jenkins/workspace/dgl_PR-4648
collecting ... PASSED [  3%]
tests/compute/test_csrmm.py::test_csrsum_backward[2-dtype0-idtype0] PASSED [  3%]
tests/compute/test_csrmm.py::test_csrsum_backward[2-dtype0-idtype1] PASSED [  3%]
tests/compute/test_csrmm.py::test_csrsum_backward[2-dtype1-idtype0] PASSED [  4%]
tests/compute/test_csrmm.py::test_csrsum_backward[2-dtype1-idtype1] PASSED [  4%]
tests/compute/test_csrmm.py::test_csrmask[9000-9000-dtype0-idtype0] PASSED [  4%]
tests/compute/test_csrmm.py::test_csrmask[9000-9000-dtype0-idtype1] PASSED [  4%]
tests/compute/test_csrmm.py::test_csrmask[9000-9000-dtype1-idtype0] PASSED [  4%]
tests/compute/test_csrmm.py::test_csrmask[9000-9000-dtype1-idtype1] PASSED [  4%]
tests/compute/test_csrmm.py::test_csrmask[9000-0-dtype0-idtype0] PASSED  [  4%]
tests/compute/test_csrmm.py::test_csrmask[9000-0-dtype0-idtype1] PASSED  [  4%]
tests/compute/test_csrmm.py::test_csrmask[9000-0-dtype1-idtype0] PASSED  [  4%]
tests/compute/test_csrmm.py::test_csrmask[9000-0-dtype1-idtype1] PASSED  [  4%]
tests/compute/test_csrmm.py::test_csrmask[0-9000-dtype0-idtype0] PASSED  [  4%]
tests/compute/test_csrmm.py::test_csrmask[0-9000-dtype0-idtype1] PASSED  [  4%]
tests/compute/test_csrmm.py::test_csrmask[0-9000-dtype1-idtype0] PASSED  [  4%]
tests/compute/test_csrmm.py::test_csrmask[0-9000-dtype1-idtype1] PASSED  [  4%]
tests/compute/test_csrmm.py::test_csrmask[0-0-dtype0-idtype0] PASSED     [  4%]
tests/compute/test_csrmm.py::test_csrmask[0-0-dtype0-idtype1] PASSED     [  4%]
tests/compute/test_csrmm.py::test_csrmask[0-0-dtype1-idtype0] PASSED     [  4%]
tests/compute/test_csrmm.py::test_csrmask[0-0-dtype1-idtype1] PASSED     [  4%]
tests/compute/test_csrmm.py::test_csrmask_backward[dtype0-idtype0] PASSED [  4%]
tests/compute/test_csrmm.py::test_csrmask_backward[dtype0-idtype1] PASSED [  4%]
tests/compute/test_csrmm.py::test_csrmask_backward[dtype1-idtype0] PASSED [  4%]
tests/compute/test_csrmm.py::test_csrmask_backward[dtype1-idtype1] PASSED [  4%]
tests/compute/test_data.py::test_minigc PASSED                           [  4%]
tests/compute/test_data.py::test_gin collected 2458 items

tests/compute/test_apply_edges_hetero.py::test_unary_copy_u[idtype0] PASSED [  0%]
tests/compute/test_apply_edges_hetero.py::test_unary_copy_u[idtype1] PASSED [  0%]
tests/compute/test_apply_edges_hetero.py::test_unary_copy_e[idtype0] PASSED [  0%]
tests/compute/test_apply_edges_hetero.py::test_unary_copy_e[idtype1] PASSED [  0%]
tests/compute/test_apply_edges_hetero.py::test_binary_op[idtype0] PASSED [  0%]
tests/compute/test_apply_edges_hetero.py::test_binary_op[idtype1] PASSED [  0%]
tests/compute/test_backend.py::test_set_default_backend PASSED           [  0%]
tests/compute/test_basics.py::test_compatible PASSED                     [  0%]
tests/compute/test_basics.py::test_batch_setter_getter[idtype0] PASSED   [  0%]
tests/compute/test_basics.py::test_batch_setter_getter[idtype1] PASSED   [  0%]
tests/compute/test_basics.py::test_batch_setter_autograd[idtype0] PASSED [  0%]
tests/compute/test_basics.py::test_batch_setter_autograd[idtype1] PASSED [  0%]
tests/compute/test_basics.py::test_apply_nodes[idtype0] PASSED           [  0%]
tests/compute/test_basics.py::test_apply_nodes[idtype1] PASSED           [  0%]
tests/compute/test_basics.py::test_apply_edges[idtype0] PASSED           [  0%]
tests/compute/test_basics.py::test_apply_edges[idtype1] PASSED           [  0%]
tests/compute/test_basics.py::test_update_routines[idtype0] PASSED       [  0%]
tests/compute/test_basics.py::test_update_routines[idtype1] PASSED       [  0%]
tests/compute/test_basics.py::test_update_all_0deg[idtype0] PASSED       [  0%]
tests/compute/test_basics.py::test_update_all_0deg[idtype1] PASSED       [  0%]
tests/compute/test_basics.py::test_pull_0deg[idtype0] PASSED             [  0%]
tests/compute/test_basics.py::test_pull_0deg[idtype1] PASSED             [  0%]
tests/compute/test_basics.py::test_dynamic_addition PASSED               [  0%]
tests/compute/test_basics.py::test_repr[idtype0] PASSED                  [  0%]
tests/compute/test_basics.py::test_repr[idtype1] PASSED                  [  1%]
tests/compute/test_basics.py::test_local_var[idtype0] PASSED             [  1%]
tests/compute/test_basics.py::test_local_var[idtype1] PASSED             [  1%]
tests/compute/test_basics.py::test_local_scope[idtype0] PASSED           [  1%]
tests/compute/test_basics.py::test_local_scope[idtype1] PASSED           [  1%]
tests/compute/test_basics.py::test_isolated_nodes[idtype0] PASSED        [  1%]
tests/compute/test_basics.py::test_isolated_nodes[idtype1] PASSED        [  1%]
tests/compute/test_basics.py::test_send_multigraph[idtype0] PASSED       [  1%]
tests/compute/test_basics.py::test_send_multigraph[idtype1] PASSED       [  1%]
tests/compute/test_basics.py::test_issue_1088[idtype0] PASSED            [  1%]
tests/compute/test_basics.py::test_issue_1088[idtype1] PASSED            [  1%]
tests/compute/test_basics.py::test_degree_bucket_edge_ordering[idtype0] PASSED [  1%]
tests/compute/test_basics.py::test_degree_bucket_edge_ordering[idtype1] PASSED [  1%]
tests/compute/test_basics.py::test_issue_2484[idtype0] PASSED            [  1%]
tests/compute/test_basics.py::test_issue_2484[idtype1] PASSED            [  1%]
tests/compute/test_batched_graph.py::test_batch_unbatch[idtype0] PASSED  [  1%]
tests/compute/test_batched_graph.py::test_batch_unbatch[idtype1] PASSED  [  1%]
tests/compute/test_batched_graph.py::test_batch_unbatch1[idtype0] PASSED [  1%]
tests/compute/test_batched_graph.py::test_batch_unbatch1[idtype1] PASSED [  1%]
tests/compute/test_batched_graph.py::test_batch_unbatch_frame[idtype0] PASSED [  1%]
tests/compute/test_batched_graph.py::test_batch_unbatch_frame[idtype1] PASSED [  1%]
tests/compute/test_batched_graph.py::test_batch_unbatch2[idtype0] PASSED [  1%]
tests/compute/test_batched_graph.py::test_batch_unbatch2[idtype1] PASSED [  1%]
tests/compute/test_batched_graph.py::test_batch_send_and_recv[idtype0] PASSED [  1%]
tests/compute/test_batched_graph.py::test_batch_send_and_recv[idtype1] PASSED [  1%]
tests/compute/test_batched_graph.py::test_batch_propagate[idtype0] PASSED [  2%]
tests/compute/test_batched_graph.py::test_batch_propagate[idtype1] PASSED [  2%]
tests/compute/test_batched_graph.py::test_batched_edge_ordering[idtype0] PASSED [  2%]
tests/compute/test_batched_graph.py::test_batched_edge_ordering[idtype1] PASSED [  2%]
tests/compute/test_batched_graph.py::test_batch_no_edge[idtype0] PASSED  [  2%]
tests/compute/test_batched_graph.py::test_batch_no_edge[idtype1] PASSED  [  2%]
tests/compute/test_batched_graph.py::test_batch_keeps_empty_data[idtype0] PASSED [  2%]
tests/compute/test_batched_graph.py::test_batch_keeps_empty_data[idtype1] PASSED [  2%]
tests/compute/test_batched_graph.py::test_set_batch_info[idtype0] PASSED [  2%]
tests/compute/test_batched_graph.py::test_set_batch_info[idtype1] PASSED [  2%]
tests/compute/test_batched_heterograph.py::test_topology[idtype0-gs0] PASSED [  2%]
tests/compute/test_batched_heterograph.py::test_topology[idtype0-gs1] PASSED [  2%]
tests/compute/test_batched_heterograph.py::test_topology[idtype1-gs0] PASSED [  2%]
tests/compute/test_batched_heterograph.py::test_topology[idtype1-gs1] PASSED [  2%]
tests/compute/test_batched_heterograph.py::test_batching_batched[idtype0] PASSED [  2%]
tests/compute/test_batched_heterograph.py::test_batching_batched[idtype1] PASSED [  2%]
tests/compute/test_batched_heterograph.py::test_features[idtype0] PASSED [  2%]
tests/compute/test_batched_heterograph.py::test_features[idtype1] PASSED [  2%]
tests/compute/test_batched_heterograph.py::test_empty_relation[idtype0] PASSED [  2%]
tests/compute/test_batched_heterograph.py::test_empty_relation[idtype1] PASSED [  2%]
tests/compute/test_batched_heterograph.py::test_unbatch2[idtype0] PASSED [  2%]
tests/compute/test_batched_heterograph.py::test_unbatch2[idtype1] PASSED [  2%]
tests/compute/test_batched_heterograph.py::test_slice_batch[idtype0] PASSED [  2%]
tests/compute/test_batched_heterograph.py::test_slice_batch[idtype1] PASSED [  2%]
tests/compute/test_batched_heterograph.py::test_batch_keeps_empty_data[idtype0] PASSED [  3%]
tests/compute/test_batched_heterograph.py::test_batch_keeps_empty_data[idtype1] PASSED [  3%]
tests/compute/test_batched_heterograph.py::test_batch_netypes SKIPPED    [  3%]
tests/compute/test_csrmm.py::test_csrmm[dtype0-idtype0] PASSED           [  3%]
tests/compute/test_csrmm.py::test_csrmm[dtype0-idtype1] PASSED           [  3%]
tests/compute/test_csrmm.py::test_csrmm[dtype1-idtype0] PASSED           [  3%]
tests/compute/test_csrmm.py::test_csrmm[dtype1-idtype1] PASSED           [  3%]
tests/compute/test_csrmm.py::test_csrmm_backward[1-dtype0-idtype0] PASSED [  3%]
tests/compute/test_csrmm.py::test_csrmm_backward[1-dtype0-idtype1] PASSED [  3%]
tests/compute/test_csrmm.py::test_csrmm_backward[1-dtype1-idtype0] PASSED [  3%]
tests/compute/test_csrmm.py::test_csrmm_backward[1-dtype1-idtype1] PASSED [  3%]
tests/compute/test_csrmm.py::test_csrmm_backward[2-dtype0-idtype0] PASSED [  3%]
tests/compute/test_csrmm.py::test_csrmm_backward[2-dtype0-idtype1] PASSED [  3%]
tests/compute/test_csrmm.py::test_csrmm_backward[2-dtype1-idtype0] PASSED [  3%]
tests/compute/test_csrmm.py::test_csrmm_backward[2-dtype1-idtype1] PASSED [  3%]
tests/compute/test_csrmm.py::test_csrsum[dtype0-idtype0] PASSED          [  3%]
tests/compute/test_csrmm.py::test_csrsum[dtype0-idtype1] PASSED          [  3%]
tests/compute/test_csrmm.py::test_csrsum[dtype1-idtype0] PASSED          [  3%]
tests/compute/test_csrmm.py::test_csrsum[dtype1-idtype1] PASSED          [  3%]
tests/compute/test_csrmm.py::test_csrsum_backward[1-dtype0-idtype0] PASSED [  3%]
tests/compute/test_csrmm.py::test_csrsum_backward[1-dtype0-idtype1] PASSED [  3%]
tests/compute/test_csrmm.py::test_csrsum_backward[1-dtype1-idtype0] PASSED [  3%]
tests/compute/test_csrmm.py::test_csrsum_backward[1-dtype1-idtype1] PASSED [  3%]
tests/compute/test_csrmm.py::test_csrsum_backward[2-dtype0-idtype0] PASSED [  3%]
tests/compute/test_csrmm.py::test_csrsum_backward[2-dtype0-idtype1] PASSED [  3%]
tests/compute/test_csrmm.py::test_csrsum_backward[2-dtype1-idtype0] PASSED [  4%]
tests/compute/test_csrmm.py::test_csrsum_backward[2-dtype1-idtype1] PASSED [  4%]
tests/compute/test_csrmm.py::test_csrmask[9000-9000-dtype0-idtype0] PASSED [  4%]
tests/compute/test_csrmm.py::test_csrmask[9000-9000-dtype0-idtype1] PASSED [  4%]
tests/compute/test_csrmm.py::test_csrmask[9000-9000-dtype1-idtype0] PASSED [  4%]
tests/compute/test_csrmm.py::test_csrmask[9000-9000-dtype1-idtype1] PASSED [  4%]
tests/compute/test_csrmm.py::test_csrmask[9000-0-dtype0-idtype0] PASSED  [  4%]
tests/compute/test_csrmm.py::test_csrmask[9000-0-dtype0-idtype1] PASSED  [  4%]
tests/compute/test_csrmm.py::test_csrmask[9000-0-dtype1-idtype0] PASSED  [  4%]
tests/compute/test_csrmm.py::test_csrmask[9000-0-dtype1-idtype1] PASSED  [  4%]
tests/compute/test_csrmm.py::test_csrmask[0-9000-dtype0-idtype0] PASSED  [  4%]
tests/compute/test_csrmm.py::test_csrmask[0-9000-dtype0-idtype1] PASSED  [  4%]
tests/compute/test_csrmm.py::test_csrmask[0-9000-dtype1-idtype0] PASSED  [  4%]
tests/compute/test_csrmm.py::test_csrmask[0-9000-dtype1-idtype1] PASSED  [  4%]
tests/compute/test_csrmm.py::test_csrmask[0-0-dtype0-idtype0] PASSED     [  4%]
tests/compute/test_csrmm.py::test_csrmask[0-0-dtype0-idtype1] PASSED     [  4%]
tests/compute/test_csrmm.py::test_csrmask[0-0-dtype1-idtype0] PASSED     [  4%]
tests/compute/test_csrmm.py::test_csrmask[0-0-dtype1-idtype1] PASSED     [  4%]
tests/compute/test_csrmm.py::test_csrmask_backward[dtype0-idtype0] PASSED [  4%]
tests/compute/test_csrmm.py::test_csrmask_backward[dtype0-idtype1] PASSED [  4%]
tests/compute/test_csrmm.py::test_csrmask_backward[dtype1-idtype0] PASSED [  4%]
tests/compute/test_csrmm.py::test_csrmask_backward[dtype1-idtype1] PASSED [  4%]
tests/compute/test_data.py::test_minigc SKIPPED (Datasets don't need...) [  4%]
tests/compute/test_data.py::test_gin SKIPPED (Datasets don't need to...) [  4%]
tests/compute/test_data.py::test_fraud SKIPPED (Datasets don't need ...) [  5%]
tests/compute/test_data.py::test_fakenews SKIPPED (Datasets don't ne...) [  5%]
tests/compute/test_data.py::test_tudataset_regression SKIPPED (Datas...) [  5%]
tests/compute/test_data.py::test_data_hash SKIPPED (Datasets don't n...) [  5%]
tests/compute/test_data.py::test_citation_graph SKIPPED (Datasets do...) [  5%]
tests/compute/test_data.py::test_gnn_benchmark SKIPPED (Datasets don...) [  5%]
tests/compute/test_data.py::test_reddit SKIPPED (Datasets don't need...) [  5%]
tests/compute/test_data.py::test_explain_syn SKIPPED (Datasets don't...) [  5%]
tests/compute/test_data.py::test_wiki_cs SKIPPED (Datasets don't nee...) [  5%]
tests/compute/test_data.py::test_yelp SKIPPED (Dataset too large to ...) [  5%]
tests/compute/test_data.py::test_flickr SKIPPED (Datasets don't need...) [  5%]
tests/compute/test_data.py::test_extract_archive SKIPPED (Datasets d...) [  5%]
tests/compute/test_data.py::test_csvdataset SKIPPED (Datasets don't ...) [  5%]
tests/compute/test_data.py::test_add_nodepred_split SKIPPED (Dataset...) [  5%]
tests/compute/test_data.py::test_as_nodepred1 SKIPPED (Datasets don'...) [  5%]
tests/compute/test_data.py::test_as_nodepred2 SKIPPED (Datasets don'...) [  5%]
tests/compute/test_data.py::test_as_nodepred_ogb PASSED                  [  5%]
tests/compute/test_data.py::test_as_linkpred SKIPPED (Datasets don't...) [  5%]
tests/compute/test_data.py::test_as_linkpred_ogb PASSED                              [  4%]
tests/compute/test_data.py::test_fraud PASSED                            [  5%]
tests/compute/test_data.py::test_fakenews PASSED                  [  5%]
tests/compute/test_data.py::test_as_nodepred_csvdataset SKIPPED (Dat...) [  5%]
tests/compute/test_data.py::test_as_graphpred SKIPPED (Datasets don'...) [  5%]
tests/compute/test_data.py::test_as_graphpred_reprocess SKIPPED (Dat...) [  5%]
tests/compute/test_data.py::test_as_graphpred_ogb PASSED                 [  5%]
tests/compute/test_dataloader.py::test_edge_prediction_sampler[idtype0] PASSED [  5%]
tests/compute/test_dataloader.py::test_edge_prediction_sampler[idtype1] PASSED [  5%]
tests/compute/test_edge_softmax_hetero.py::test_edge_softmax_unidirectional PASSED [  6%]
tests/compute/test_edge_softmax_hetero.py::test_edge_softmax[idtype0-src-g0] PASSED [  6%]
tests/compute/test_edge_softmax_hetero.py::test_edge_softmax[idtype0-dst-g0] PASSED [  6%]
tests/compute/test_edge_softmax_hetero.py::test_edge_softmax[idtype1-src-g0] PASSED [  6%]
tests/compute/test_edge_softmax_hetero.py::test_edge_softmax[idtype1-dst-g0] PASSED [  6%]
tests/compute/test_ffi.py::test_cython PASSED                            [  6%]
tests/compute/test_ffi.py::test_callback[1] PASSED                       [  6%]
tests/compute/test_ffi.py::test_callback[2.3] PASSED                     [  6%]
tests/compute/test_ffi.py::test_callback_thread[1] PASSED                [  6%]
tests/compute/test_ffi.py::test_callback_thread[2.3] PASSED              [  6%]
tests/compute/test_filter.py::test_graph_filter PASSED                   [  6%]
tests/compute/test_filter.py::test_array_filter[idtype0] PASSED          [  6%]
tests/compute/test_filter.py::test_array_filter[idtype1] PASSED          [  6%]
tests/compute/test_frame.py::test_column_subcolumn PASSED                [  6%]
tests/compute/test_frame.py::test_serialize_deserialize_plain PASSED     [  6%]
tests/compute/test_frame.py::test_serialize_deserialize_subcolumn PASSED [  6%]
tests/compute/test_frame.py::test_serialize_deserialize_dtype PASSED     [  6%]
tests/compute/test_generators.py::test_rand_graph SKIPPED (GPU rando...) [  6%]
tests/compute/test_graph.py::test_query PASSED                           [  6%]
tests/compute/test_graph.py::test_mutation PASSED                        [  6%]
tests/compute/test_graph.py::test_scipy_adjmat PASSED                    [  6%]
tests/compute/test_graph.py::test_incmat PASSED                          [  6%]
tests/compute/test_graph.py::test_find_edges PASSED                      [  6%]
tests/compute/test_graph.py::test_ismultigraph PASSED                    [  6%]
tests/compute/test_graph.py::test_hypersparse_query PASSED               [  6%]
tests/compute/test_graph.py::test_empty_data_initialized PASSED          [  7%]
tests/compute/test_graph.py::test_is_sorted PASSED                       [  7%]
tests/compute/test_graph.py::test_default_types PASSED                   [  7%]
tests/compute/test_graph.py::test_formats PASSED                         [  7%]
tests/compute/test_heterograph.py::test_create[idtype0] PASSED           [  7%]
tests/compute/test_heterograph.py::test_create[idtype1] PASSED           [  7%]
tests/compute/test_heterograph.py::test_create2 PASSED                   [  7%]
tests/compute/test_heterograph.py::test_query[idtype0] PASSED            [  7%]
tests/compute/test_heterograph.py::test_query[idtype1] PASSED            [  7%]
tests/compute/test_heterograph.py::test_empty_query[idtype0] PASSED      [  7%]
tests/compute/test_heterograph.py::test_empty_query[idtype1] PASSED      [  7%]
tests/compute/test_heterograph.py::test_adj[idtype0] PASSED              [  7%]
tests/compute/test_heterograph.py::test_adj[idtype1] PASSED              [  7%]
tests/compute/test_heterograph.py::test_inc[idtype0] PASSED              [  7%]
tests/compute/test_heterograph.py::test_inc[idtype1] PASSED              [  7%]
tests/compute/test_heterograph.py::test_view[idtype0] PASSED             [  7%]
tests/compute/test_heterograph.py::test_view[idtype1] PASSED             [  7%]
tests/compute/test_heterograph.py::test_view1[idtype0] PASSED            [  7%]
tests/compute/test_heterograph.py::test_view1[idtype1] PASSED            [  7%]
tests/compute/test_heterograph.py::test_flatten[idtype0] PASSED          [  7%]
tests/compute/test_heterograph.py::test_flatten[idtype1] PASSED          [  7%]
tests/compute/test_heterograph.py::test_to_device[idtype0] PASSED        [  7%]
tests/compute/test_heterograph.py::test_to_device[idtype1] PASSED        [  7%]
tests/compute/test_heterograph.py::test_to_device2[g0-idtype0] PASSED    [  7%]
tests/compute/test_heterograph.py::test_to_device2[g0-idtype1] PASSED    [  8%]
tests/compute/test_heterograph.py::test_to_device2[g1-idtype0] PASSED    [  8%]
tests/compute/test_heterograph.py::test_to_device2[g1-idtype1] PASSED    [  8%]
tests/compute/test_heterograph.py::test_pin_memory_[idtype0] PASSED      [  8%]
tests/compute/test_heterograph.py::test_pin_memory_[idtype1] PASSED      [  8%]
tests/compute/test_heterograph.py::test_convert_bound[idtype0] PASSED    [  8%]
tests/compute/test_heterograph.py::test_convert_bound[idtype1] PASSED    [  8%]
tests/compute/test_heterograph.py::test_convert[idtype0] PASSED          [  8%]
tests/compute/test_heterograph.py::test_convert[idtype1] PASSED          [  8%]
tests/compute/test_heterograph.py::test_to_homo_zero_nodes[idtype0] SKIPPED [  8%]
tests/compute/test_heterograph.py::test_to_homo_zero_nodes[idtype1] SKIPPED [  8%]
tests/compute/test_heterograph.py::test_to_homo2[idtype0] PASSED         [  8%]
tests/compute/test_heterograph.py::test_to_homo2[idtype1] PASSED         [  8%]
tests/compute/test_heterograph.py::test_invertible_conversion[idtype0] PASSED [  8%]
tests/compute/test_heterograph.py::test_invertible_conversion[idtype1] PASSED [  8%]
tests/compute/test_heterograph.py::test_metagraph_reachable[idtype0] PASSED [  8%]
tests/compute/test_heterograph.py::test_metagraph_reachable[idtype1] PASSED [  8%]
tests/compute/test_heterograph.py::test_subgraph_mask[idtype0] PASSED    [  8%]
tests/compute/test_heterograph.py::test_subgraph_mask[idtype1] PASSED    [  8%]
tests/compute/test_heterograph.py::test_subgraph[idtype0] PASSED         [  8%]
tests/compute/test_heterograph.py::test_subgraph[idtype1] PASSED         [  8%]
tests/compute/test_heterograph.py::test_apply[idtype0] PASSED            [  8%]
tests/compute/test_heterograph.py::test_apply[idtype1] PASSED            [  8%]
tests/compute/test_heterograph.py::test_level2[idtype0] PASSED           [  8%]
tests/compute/test_heterograph.py::test_level2[idtype1] PASSED           [  8%]
tests/compute/test_heterograph.py::test_more_nnz[idtype0] PASSED         [  9%]
tests/compute/test_heterograph.py::test_more_nnz[idtype1] PASSED         [  9%]
tests/compute/test_heterograph.py::test_updates[idtype0] PASSED          [  9%]
tests/compute/test_heterograph.py::test_updates[idtype1] PASSED          [  9%]
tests/compute/test_heterograph.py::test_backward[idtype0] PASSED         [  9%]
tests/compute/test_heterograph.py::test_backward[idtype1] PASSED         [  9%]
tests/compute/test_heterograph.py::test_empty_heterograph[idtype0] PASSED [  9%]
tests/compute/test_heterograph.py::test_empty_heterograph[idtype1] PASSED [  9%]
tests/compute/test_heterograph.py::test_types_in_function[idtype0] PASSED [  9%]
tests/compute/test_heterograph.py::test_types_in_function[idtype1] PASSED [  9%]
tests/compute/test_heterograph.py::test_stack_reduce[idtype0] PASSED     [  9%]
tests/compute/test_heterograph.py::test_stack_reduce[idtype1] PASSED     [  9%]
tests/compute/test_heterograph.py::test_isolated_ntype[idtype0] PASSED   [  9%]
tests/compute/test_heterograph.py::test_isolated_ntype[idtype1] PASSED   [  9%]
tests/compute/test_heterograph.py::test_ismultigraph[idtype0] PASSED     [  9%]
tests/compute/test_heterograph.py::test_ismultigraph[idtype1] PASSED     [  9%]
tests/compute/test_heterograph.py::test_graph_index_is_unibipartite[idtype0] PASSED [  9%]
tests/compute/test_heterograph.py::test_graph_index_is_unibipartite[idtype1] PASSED [  9%]
tests/compute/test_heterograph.py::test_bipartite[idtype0] PASSED        [  9%]
tests/compute/test_heterograph.py::test_bipartite[idtype1] PASSED        [  9%]
tests/compute/test_heterograph.py::test_dtype_cast[idtype0] PASSED       [  9%]
tests/compute/test_heterograph.py::test_dtype_cast[idtype1] PASSED       [  9%]
tests/compute/test_heterograph.py::test_float_cast PASSED                [  9%]
tests/compute/test_heterograph.py::test_format[idtype0] PASSED           [  9%]
tests/compute/test_heterograph.py::test_format[idtype1] PASSED           [ 10%]
tests/compute/test_heterograph.py::test_edges_order[idtype0] PASSED      [ 10%]
tests/compute/test_heterograph.py::test_edges_order[idtype1] PASSED      [ 10%]
tests/compute/test_heterograph.py::test_reverse[idtype0] PASSED          [ 10%]
tests/compute/test_heterograph.py::test_reverse[idtype1] PASSED          [ 10%]
tests/compute/test_heterograph.py::test_clone[idtype0] PASSED            [ 10%]
tests/compute/test_heterograph.py::test_clone[idtype1] PASSED            [ 10%]
tests/compute/test_heterograph.py::test_add_edges[idtype0] PASSED        [ 10%]
tests/compute/test_heterograph.py::test_add_edges[idtype1] PASSED        [ 10%]
tests/compute/test_heterograph.py::test_add_nodes[idtype0] PASSED        [ 10%]
tests/compute/test_heterograph.py::test_add_nodes[idtype1] PASSED        [ 10%]
tests/compute/test_heterograph.py::test_remove_edges[idtype0] PASSED     [ 10%]
tests/compute/test_heterograph.py::test_remove_edges[idtype1] PASSED     [ 10%]
tests/compute/test_heterograph.py::test_remove_nodes[idtype0] PASSED     [ 10%]
tests/compute/test_heterograph.py::test_remove_nodes[idtype1] PASSED     [ 10%]
tests/compute/test_heterograph.py::test_frame[idtype0] PASSED            [ 10%]
tests/compute/test_heterograph.py::test_frame[idtype1] PASSED            [ 10%]
tests/compute/test_heterograph.py::test_frame_device[idtype0] PASSED     [ 10%]
tests/compute/test_heterograph.py::test_frame_device[idtype1] PASSED     [ 10%]
tests/compute/test_heterograph.py::test_create_block[idtype0] PASSED     [ 10%]
tests/compute/test_heterograph.py::test_create_block[idtype1] PASSED     [ 10%]
tests/compute/test_heterograph.py::test_adj_sparse[coo-idtype0] PASSED   [ 10%]
tests/compute/test_heterograph.py::test_adj_sparse[coo-idtype1] PASSED   [ 10%]
tests/compute/test_heterograph.py::test_adj_sparse[csr-idtype0] PASSED   [ 10%]
tests/compute/test_heterograph.py::test_adj_sparse[csr-idtype1] PASSED   [ 10%]
tests/compute/test_heterograph.py::test_adj_sparse[csc-idtype0] PASSED   [ 11%]
tests/compute/test_heterograph.py::test_adj_sparse[csc-idtype1] PASSED   [ 11%]
tests/compute/test_heterograph.py::test_forking_pickler PASSED           [ 11%]
tests/compute/test_index.py::test_dlpack PASSED                          [ 11%]
tests/compute/test_kernel.py::test_copy_src_reduce PASSED                [ 11%]
tests/compute/test_kernel.py::test_copy_edge_reduce PASSED               [ 11%]
tests/compute/test_kernel.py::test_all_binary_builtins [       OK ] RowwiseTest.TestCOOSampling (73878 ms)
[ RUN      ] RowwiseTest.TestCOOSamplingUniform
PASSED            [ 11%]
tests/compute/test_kernel.py::test_mean_zero_degree[g0-idtype0] PASSED   [ 11%]
tests/compute/test_kernel.py::test_mean_zero_degree[g0-idtype1] PASSED   [ 11%]
tests/compute/test_merge.py::test_heterograph_merge[idtype0] PASSED      [ 11%]
tests/compute/test_merge.py::test_heterograph_merge[idtype1] PASSED      [ 11%]
tests/compute/test_nccl.py::test_nccl_id PASSED                          [ 11%]
tests/compute/test_nccl.py::test_nccl_sparse_push_single_remainder PASSED [ 11%]
tests/compute/test_nccl.py::test_nccl_sparse_pull_single_remainder PASSED [ 11%]
tests/compute/test_nccl.py::test_nccl_sparse_push_single_range PASSED    [ 11%]
tests/compute/test_nccl.py::test_nccl_sparse_pull_single_range PASSED    [ 11%]
tests/compute/test_nccl.py::test_nccl_support PASSED                     [ 11%]
tests/compute/test_new_update_all_hetero.py::test_unary_copy_u[idtype0] PASSED [ 11%]
tests/compute/test_new_update_all_hetero.py::test_unary_copy_u[idtype1] PASSED [ 11%]
tests/compute/test_new_update_all_hetero.py::test_unary_copy_e[idtype0] PASSED [ 11%]
tests/compute/test_new_update_all_hetero.py::test_unary_copy_e[idtype1] PASSED [ 11%]
tests/compute/test_new_update_all_hetero.py::test_binary_op[idtype0] PASSED [ 11%]
tests/compute/test_new_update_all_hetero.py::test_binary_op[idtype1] PASSED [ 11%]
tests/compute/test_partition.py::test_get_node_partition_from_book[idtype0] PASSED [ 11%]
tests/compute/test_partition.py::test_get_node_partition_from_book[idtype1] PASSED [ 12%]
tests/compute/test_pickle.py::test_pickling_index PASSED                 [ 12%]
tests/compute/test_pickle.py::test_pickling_graph_index PASSED           [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g0-idtype0] SKIPPED    [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g0-idtype1] SKIPPED    [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g1-idtype0] SKIPPED    [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g1-idtype1] SKIPPED    [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g2-idtype0] SKIPPED    [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g2-idtype1] SKIPPED    [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g3-idtype0] SKIPPED    [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g3-idtype1] SKIPPED    [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g4-idtype0] SKIPPED    [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g4-idtype1] SKIPPED    [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g5-idtype0] SKIPPED    [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g5-idtype1] SKIPPED    [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g6-idtype0] SKIPPED    [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g6-idtype1] SKIPPED    [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g7-idtype0] SKIPPED    [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g7-idtype1] SKIPPED    [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g8-idtype0] SKIPPED    [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g8-idtype1] SKIPPED    [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g9-idtype0] SKIPPED    [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g9-idtype1] SKIPPED    [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g10-idtype0] SKIPPED   [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g10-idtype1] SKIPPED   [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g11-idtype0] SKIPPED   [ 13%]
tests/compute/test_pickle.py::test_pickling_graph[g11-idtype1] SKIPPED   [ 13%]
tests/compute/test_pickle.py::test_pickling_graph[g12-idtype0] SKIPPED   [ 13%]
tests/compute/test_pickle.py::test_pickling_graph[g12-idtype1] SKIPPED   [ 13%]
tests/compute/test_pickle.py::test_pickling_batched_heterograph SKIPPED  [ 13%]
tests/compute/test_pickle.py::test_pickling_subgraph SKIPPED (GPU ed...) [ 13%]
tests/compute/test_pickle.py::test_pickling_is_pinned[idtype0] PASSED    [ 13%]
tests/compute/test_pickle.py::test_pickling_is_pinned[idtype1] PASSED    [ 13%]
tests/compute/test_pin_memory.py::test_pin_unpin PASSED                  [ 13%]
tests/compute/test_propagate.py::test_prop_nodes_bfs[idtype0] SKIPPED    [ 13%]
tests/compute/test_propagate.py::test_prop_nodes_bfs[idtype1] SKIPPED    [ 13%]
tests/compute/test_propagate.py::test_prop_edges_dfs[idtype0] SKIPPED    [ 13%]
tests/compute/test_propagate.py::test_prop_edges_dfs[idtype1] SKIPPED    [ 13%]
tests/compute/test_propagate.py::test_prop_nodes_topo[idtype0] SKIPPED   [ 13%]
tests/compute/test_propagate.py::test_prop_nodes_topo[idtype1] SKIPPED   [ 13%]
tests/compute/test_random.py::test_random_choice SKIPPED (GPU random...) [ 13%]
tests/compute/test_readout.py::test_sum_case1[idtype0] PASSED            [ 13%]
tests/compute/test_readout.py::test_sum_case1[idtype1] PASSED            [ 13%]
tests/compute/test_readout.py::test_reduce_readout[sum-g0-idtype0] PASSED [ 13%]
tests/compute/test_readout.py::test_reduce_readout[sum-g0-idtype1] PASSED [ 13%]
tests/compute/test_readout.py::test_reduce_readout[sum-g1-idtype0] PASSED [ 13%]
tests/compute/test_readout.py::test_reduce_readout[sum-g1-idtype1] PASSED [ 13%]
tests/compute/test_readout.py::test_reduce_readout[sum-g2-idtype0] PASSED [ 13%]
tests/compute/test_readout.py::test_reduce_readout[sum-g2-idtype1] PASSED [ 13%]
tests/compute/test_readout.py::test_reduce_readout[sum-g3-idtype0] PASSED [ 13%]
tests/compute/test_readout.py::test_reduce_readout[sum-g3-idtype1] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[sum-g4-idtype0] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[sum-g4-idtype1] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[sum-g5-idtype0] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[sum-g5-idtype1] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[sum-g6-idtype0] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[sum-g6-idtype1] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g0-idtype0] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g0-idtype1] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g1-idtype0] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g1-idtype1] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g2-idtype0] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g2-idtype1] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g3-idtype0] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g3-idtype1] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g4-idtype0] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g4-idtype1] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g5-idtype0] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g5-idtype1] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g6-idtype0] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g6-idtype1] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[mean-g0-idtype0] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[mean-g0-idtype1] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[mean-g1-idtype0] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[mean-g1-idtype1] PASSED [ 15%]
tests/compute/test_readout.py::test_reduce_readout[mean-g2-idtype0] PASSED [ 15%]
tests/compute/test_readout.py::test_reduce_readout[mean-g2-idtype1] PASSED [ 15%]
tests/compute/test_readout.py::test_reduce_readout[mean-g3-idtype0] PASSED [ 15%]
tests/compute/test_readout.py::test_reduce_readout[mean-g3-idtype1] PASSED [ 15%]
tests/compute/test_readout.py::test_reduce_readout[mean-g4-idtype0] PASSED [ 15%]
tests/compute/test_readout.py::test_reduce_readout[mean-g4-idtype1] PASSED [ 15%]
tests/compute/test_readout.py::test_reduce_readout[mean-g5-idtype0] PASSED [ 15%]
tests/compute/test_readout.py::test_reduce_readout[mean-g5-idtype1] PASSED [ 15%]
tests/compute/test_readout.py::test_reduce_readout[mean-g6-idtype0] PASSED [ 15%]
tests/compute/test_readout.py::test_reduce_readout[mean-g6-idtype1] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g0-idtype0] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g0-idtype1] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g1-idtype0] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g1-idtype1] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g2-idtype0] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g2-idtype1] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g3-idtype0] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g3-idtype1] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g4-idtype0] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g4-idtype1] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g5-idtype0] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g5-idtype1] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g6-idtype0] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g6-idtype1] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g0-idtype0] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g0-idtype1] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g1-idtype0] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g1-idtype1] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g2-idtype0] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g2-idtype1] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g3-idtype0] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g3-idtype1] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g4-idtype0] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g4-idtype1] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g5-idtype0] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g5-idtype1] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g6-idtype0] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g6-idtype1] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g0-idtype0] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g0-idtype1] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g1-idtype0] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g1-idtype1] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g2-idtype0] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g2-idtype1] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g3-idtype0] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g3-idtype1] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g4-idtype0] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g4-idtype1] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g5-idtype0] PASSED [ 17%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g5-idtype1] PASSED [ 17%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g6-idtype0] PASSED [ 17%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g6-idtype1] PASSED [ 17%]
tests/compute/test_readout.py::test_topk[True-g0-idtype0] PASSED         [ 17%]
tests/compute/test_readout.py::test_topk[True-g0-idtype1] PASSED         [ 17%]
tests/compute/test_readout.py::test_topk[True-g1-idtype0] PASSED         [ 17%]
tests/compute/test_readout.py::test_topk[True-g1-idtype1] PASSED         [ 17%]
tests/compute/test_readout.py::test_topk[True-g2-idtype0] PASSED         [ 17%]
tests/compute/test_readout.py::test_topk[True-g2-idtype1] PASSED         [ 17%]
tests/compute/test_readout.py::test_topk[True-g3-idtype0] PASSED         [ 17%]
tests/compute/test_readout.py::test_topk[True-g3-idtype1] PASSED         [ 17%]
tests/compute/test_readout.py::test_topk[True-g4-idtype0] PASSED         [ 17%]
tests/compute/test_readout.py::test_topk[True-g4-idtype1] PASSED         [ 17%]
tests/compute/test_readout.py::test_topk[True-g5-idtype0] PASSED         [ 17%]
tests/compute/test_readout.py::test_topk[True-g5-idtype1] PASSED         [ 17%]
tests/compute/test_readout.py::test_topk[True-g6-idtype0] PASSED         [ 17%]
tests/compute/test_readout.py::test_topk[True-g6-idtype1] PASSED         [ 17%]
tests/compute/test_readout.py::test_topk[False-g0-idtype0] PASSED        [ 17%]
tests/compute/test_readout.py::test_topk[False-g0-idtype1] PASSED        [ 17%]
tests/compute/test_readout.py::test_topk[False-g1-idtype0] PASSED        [ 17%]
tests/compute/test_readout.py::test_topk[False-g1-idtype1] PASSED        [ 17%]
tests/compute/test_readout.py::test_topk[False-g2-idtype0] PASSED        [ 17%]
tests/compute/test_readout.py::test_topk[False-g2-idtype1] PASSED        [ 17%]
tests/compute/test_readout.py::test_topk[False-g3-idtype0] [       OK ] RowwiseTest.TestCOOSamplingUniform (9886 ms)
[ RUN      ] RowwiseTest.TestCOOerEtypeSampling
PASSED        [ 17%]
tests/compute/test_readout.py::test_topk[False-g3-idtype1] PASSED        [ 18%]
tests/compute/test_readout.py::test_topk[False-g4-idtype0] PASSED        [ 18%]
tests/compute/test_readout.py::test_topk[False-g4-idtype1] PASSED        [ 18%]
tests/compute/test_readout.py::test_topk[False-g5-idtype0] PASSED        [ 18%]
tests/compute/test_readout.py::test_topk[False-g5-idtype1] PASSED        [ 18%]
tests/compute/test_readout.py::test_topk[False-g6-idtype0] PASSED        [ 18%]
tests/compute/test_readout.py::test_topk[False-g6-idtype1] PASSED        [ 18%]
tests/compute/test_readout.py::test_softmax[g0-idtype0] PASSED           [ 18%]
tests/compute/test_readout.py::test_softmax[g0-idtype1] PASSED           [ 18%]
tests/compute/test_readout.py::test_softmax[g1-idtype0] PASSED           [ 18%]
tests/compute/test_readout.py::test_softmax[g1-idtype1] PASSED           [ 18%]
tests/compute/test_readout.py::test_softmax[g2-idtype0] PASSED           [ 18%]
tests/compute/test_readout.py::test_softmax[g2-idtype1] PASSED           [ 18%]
tests/compute/test_readout.py::test_softmax[g3-idtype0] PASSED           [ 18%]
tests/compute/test_readout.py::test_softmax[g3-idtype1] PASSED           [ 18%]
tests/compute/test_readout.py::test_softmax[g4-idtype0] PASSED           [ 18%]
tests/compute/test_readout.py::test_softmax[g4-idtype1] PASSED           [ 18%]
tests/compute/test_readout.py::test_softmax[g5-idtype0] PASSED           [ 18%]
tests/compute/test_readout.py::test_softmax[g5-idtype1] PASSED           [ 18%]
tests/compute/test_readout.py::test_softmax[g6-idtype0] PASSED           [ 18%]
tests/compute/test_readout.py::test_softmax[g6-idtype1] PASSED           [ 18%]
tests/compute/test_readout.py::test_broadcast[g0-idtype0] PASSED         [ 18%]
tests/compute/test_readout.py::test_broadcast[g0-idtype1] PASSED         [ 18%]
tests/compute/test_readout.py::test_broadcast[g1-idtype0] PASSED         [ 18%]
tests/compute/test_readout.py::test_broadcast[g1-idtype1] PASSED         [ 18%]
tests/compute/test_readout.py::test_broadcast[g2-idtype0] PASSED         [ 19%]
tests/compute/test_readout.py::test_broadcast[g2-idtype1] PASSED         [ 19%]
tests/compute/test_readout.py::test_broadcast[g3-idtype0] PASSED         [ 19%]
tests/compute/test_readout.py::test_broadcast[g3-idtype1] PASSED         [ 19%]
tests/compute/test_readout.py::test_broadcast[g4-idtype0] PASSED         [ 19%]
tests/compute/test_readout.py::test_broadcast[g4-idtype1] PASSED         [ 19%]
tests/compute/test_readout.py::test_broadcast[g5-idtype0] PASSED         [ 19%]
tests/compute/test_readout.py::test_broadcast[g5-idtype1] PASSED         [ 19%]
tests/compute/test_readout.py::test_broadcast[g6-idtype0] PASSED         [ 19%]
tests/compute/test_readout.py::test_broadcast[g6-idtype1] PASSED         [ 19%]
tests/compute/test_removal.py::test_node_removal[idtype0] PASSED         [ 19%]
tests/compute/test_removal.py::test_node_removal[idtype1] PASSED         [ 19%]
tests/compute/test_removal.py::test_multigraph_node_removal[idtype0] PASSED [ 19%]
tests/compute/test_removal.py::test_multigraph_node_removal[idtype1] PASSED [ 19%]
tests/compute/test_removal.py::test_multigraph_edge_removal[idtype0] PASSED [ 19%]
tests/compute/test_removal.py::test_multigraph_edge_removal[idtype1] PASSED [ 19%]
tests/compute/test_removal.py::test_edge_removal[idtype0] PASSED         [ 19%]
tests/compute/test_removal.py::test_edge_removal[idtype1] PASSED         [ 19%]
tests/compute/test_removal.py::test_node_and_edge_removal[idtype0] PASSED [ 19%]
tests/compute/test_removal.py::test_node_and_edge_removal[idtype1] PASSED [ 19%]
tests/compute/test_removal.py::test_node_frame[idtype0] PASSED           [ 19%]
tests/compute/test_removal.py::test_node_frame[idtype1] PASSED           [ 19%]
tests/compute/test_removal.py::test_edge_frame[idtype0] PASSED           [ 19%]
tests/compute/test_removal.py::test_edge_frame[idtype1] PASSED           [ 19%]
tests/compute/test_removal.py::test_issue1287[idtype0] PASSED            [ 20%]
tests/compute/test_removal.py::test_issue1287[idtype1] PASSED            [ 20%]
tests/compute/test_sampler.py::test_create_full PASSED                   [ 20%]
tests/compute/test_sampler.py::test_1neighbor_sampler_all PASSED         [ 20%]
tests/compute/test_sampler.py::test_1neighbor_sampler PASSED             [ 20%]
tests/compute/test_sampler.py::test_prefetch_neighbor_sampler PASSED     [ 20%]
tests/compute/test_sampler.py::test_10neighbor_sampler_all PASSED        [ 20%]
tests/compute/test_sampler.py::test_10neighbor_sampler PASSED            [ 20%]
tests/compute/test_sampler.py::test_layer_sampler PASSED                 [ 20%]
tests/compute/test_sampler.py::test_nonuniform_neighbor_sampler PASSED   [ 20%]
tests/compute/test_sampler.py::test_setseed PASSED                       [ 20%]
tests/compute/test_sampler.py::test_negative_sampler PASSED                              [  4%]
tests/compute/test_data.py::test_fraud PASSED                            [  5%]
tests/compute/test_data.py::test_fakenews PASSED              [ 20%]
tests/compute/test_sampling.py::test_non_uniform_random_walk[True] PASSED [ 20%]
tests/compute/test_sampling.py::test_non_uniform_random_walk[False] PASSED [ 20%]
tests/compute/test_sampling.py::test_uniform_random_walk[True] PASSED    [ 20%]
tests/compute/test_sampling.py::test_uniform_random_walk[False] PASSED   [ 20%]
tests/compute/test_sampling.py::test_node2vec SKIPPED (GPU random wa...) [ 20%]
tests/compute/test_sampling.py::test_pack_traces SKIPPED (GPU pack t...) [ 20%]
tests/compute/test_sampling.py::test_pinsage_sampling[True] PASSED       [ 20%]
tests/compute/test_sampling.py::test_pinsage_sampling[False] PASSED      [ 20%]
tests/compute/test_sampling.py::test_sample_neighbors_noprob PASSED      [ 20%]
tests/compute/test_sampling.py::test_sample_neighbors_prob PASSED        [ 20%]
tests/compute/test_sampling.py::test_sample_neighbors_outedge PASSED     [ 20%]
tests/compute/test_sampling.py::test_sample_neighbors_topk SKIPPED (...) [ 20%]
tests/compute/test_sampling.py::test_sample_neighbors_topk_outedge SKIPPED [ 20%]
tests/compute/test_sampling.py::test_sample_neighbors_with_0deg PASSED   [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_biased_homogeneous SKIPPED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_biased_bipartite SKIPPED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_homogeneous[False-in-coo] SKIPPED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_homogeneous[False-in-csr] SKIPPED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_homogeneous[False-in-csc] SKIPPED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_homogeneous[False-out-coo] SKIPPED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_homogeneous[False-out-csr] SKIPPED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_homogeneous[False-out-csc] SKIPPED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_homogeneous[True-in-coo] SKIPPED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_homogeneous[True-in-csr] SKIPPED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_homogeneous[True-in-csc] SKIPPED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_homogeneous[True-out-coo] SKIPPED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_homogeneous[True-out-csr] SKIPPED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_homogeneous[True-out-csc] SKIPPED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_sorted_homogeneous[in-csr] SKIPPED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_sorted_homogeneous[in-csc] SKIPPED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_sorted_homogeneous[out-csr] SKIPPED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_sorted_homogeneous[out-csc] SKIPPED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_exclude_edges_heteroG[int32] PASSED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_exclude_edges_heteroG[int64] PASSED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_exclude_edges_homoG[int32] PASSED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_exclude_edges_homoG[int64] PASSED [ 21%]
tests/compute/test_sampling.py::test_global_uniform_negative_sampling[int32] PASSED [ 21%]
tests/compute/test_sampling.py::test_global_uniform_negative_sampling[int64] PASSED [ 22%]
tests/compute/test_serialize.py::test_graph_serialize_with_feature[True] SKIPPED [ 22%]
tests/compute/test_serialize.py::test_graph_serialize_with_feature[False] SKIPPED [ 22%]
tests/compute/test_serialize.py::test_graph_serialize_without_feature[True] SKIPPED [ 22%]
tests/compute/test_serialize.py::test_graph_serialize_without_feature[False] SKIPPED [ 22%]
tests/compute/test_serialize.py::test_graph_serialize_with_labels[True] SKIPPED [ 22%]
tests/compute/test_serialize.py::test_graph_serialize_with_labels[False] SKIPPED [ 22%]
tests/compute/test_serialize.py::test_serialize_tensors PASSED           [ 22%]
tests/compute/test_serialize.py::test_serialize_empty_dict PASSED        [ 22%]
tests/compute/test_serialize.py::test_load_old_files1 PASSED             [ 22%]
tests/compute/test_serialize.py::test_load_old_files2 PASSED             [ 22%]
tests/compute/test_serialize.py::test_deserialize_old_heterograph_file PASSED [ 22%]
tests/compute/test_serialize.py::test_serialize_heterograph SKIPPED      [ 22%]
tests/compute/test_serialize.py::test_serialize_heterograph_s3 SKIPPED   [ 22%]
tests/compute/test_shared_mem.py::test_single_process[idtype0] PASSED    [ 22%]
tests/compute/test_shared_mem.py::test_single_process[idtype1] PASSED    [ 22%]
tests/compute/test_shared_mem.py::test_multi_process[idtype0] PASSED     [ 22%]
tests/compute/test_shared_mem.py::test_multi_process[idtype1] PASSED     [ 22%]
tests/compute/test_shared_mem.py::test_copy_from_gpu PASSED              [ 22%]
tests/compute/test_sort.py::test_sort_with_tag[idtype0] SKIPPED (GPU...) [ 22%]
tests/compute/test_sort.py::test_sort_with_tag[idtype1] SKIPPED (GPU...) [ 22%]
tests/compute/test_sort.py::test_sort_with_tag_bipartite[idtype0] SKIPPED [ 22%]
tests/compute/test_sort.py::test_sort_with_tag_bipartite[idtype1] SKIPPED [ 22%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-add-shp0-g0] PASSED  [ 22%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-add-shp0-g1] PASSED  [ 22%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-add-shp1-g0] PASSED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-add-shp1-g1] PASSED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-add-shp2-g0] PASSED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-add-shp2-g1] PASSED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-add-shp3-g0] PASSED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-add-shp3-g1] PASSED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-add-shp4-g0] PASSED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-add-shp4-g1] PASSED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-add-shp5-g0] PASSED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-add-shp5-g1] PASSED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-sub-shp0-g0] PASSED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-sub-shp0-g1] PASSED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-sub-shp1-g0] PASSED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-sub-shp1-g1] PASSED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-sub-shp2-g0] PASSED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-sub-shp2-g1] PASSED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-sub-shp3-g0] PASSED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-sub-shp3-g1] PASSED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-sub-shp4-g0] PASSED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-sub-shp4-g1] PASSED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-sub-shp5-g0] PASSED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-sub-shp5-g1] PASSED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-mul-shp0-g0] PASSED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-mul-shp0-g1] PASSED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-mul-shp1-g0] PASSED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-mul-shp1-g1] PASSED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-mul-shp2-g0] PASSED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-mul-shp2-g1] PASSED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-mul-shp3-g0] PASSED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-mul-shp3-g1] PASSED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-mul-shp4-g0] PASSED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-mul-shp4-g1] PASSED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-mul-shp5-g0] PASSED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-mul-shp5-g1] PASSED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-div-shp0-g0] PASSED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-div-shp0-g1] PASSED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-div-shp1-g0] PASSED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-div-shp1-g1] PASSED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-div-shp2-g0] PASSED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-div-shp2-g1] PASSED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-div-shp3-g0] PASSED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-div-shp3-g1] PASSED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-div-shp4-g0] PASSED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-div-shp4-g1] PASSED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-div-shp5-g0] PASSED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-div-shp5-g1] PASSED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_lhs-shp0-g0] PASSED [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_lhs-shp0-g1] PASSED [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_lhs-shp1-g0] PASSED [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_lhs-shp1-g1] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_lhs-shp2-g0] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_lhs-shp2-g1] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_lhs-shp3-g0] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_lhs-shp3-g1] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_lhs-shp4-g0] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_lhs-shp4-g1] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_lhs-shp5-g0] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_lhs-shp5-g1] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_rhs-shp0-g0] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_rhs-shp0-g1] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_rhs-shp1-g0] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_rhs-shp1-g1] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_rhs-shp2-g0] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_rhs-shp2-g1] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_rhs-shp3-g0] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_rhs-shp3-g1] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_rhs-shp4-g0] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_rhs-shp4-g1] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_rhs-shp5-g0] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_rhs-shp5-g1] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-add-shp0-g0] PASSED  [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-add-shp0-g1] PASSED  [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-add-shp1-g0] PASSED  [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-add-shp1-g1] PASSED  [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-add-shp2-g0] PASSED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-add-shp2-g1] PASSED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-add-shp3-g0] PASSED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-add-shp3-g1] PASSED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-add-shp4-g0] PASSED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-add-shp4-g1] PASSED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-add-shp5-g0] PASSED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-add-shp5-g1] PASSED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-sub-shp0-g0] PASSED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-sub-shp0-g1] PASSED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-sub-shp1-g0] PASSED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-sub-shp1-g1] PASSED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-sub-shp2-g0] PASSED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-sub-shp2-g1] PASSED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-sub-shp3-g0] PASSED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-sub-shp3-g1] PASSED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-sub-shp4-g0] PASSED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-sub-shp4-g1] PASSED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-sub-shp5-g0] PASSED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-sub-shp5-g1] PASSED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-mul-shp0-g0] PASSED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-mul-shp0-g1] PASSED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-mul-shp1-g0] PASSED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-mul-shp1-g1] PASSED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-mul-shp2-g0] PASSED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-mul-shp2-g1] PASSED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-mul-shp3-g0] PASSED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-mul-shp3-g1] PASSED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-mul-shp4-g0] PASSED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-mul-shp4-g1] PASSED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-mul-shp5-g0] PASSED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-mul-shp5-g1] PASSED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-div-shp0-g0] PASSED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-div-shp0-g1] PASSED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-div-shp1-g0] PASSED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-div-shp1-g1] PASSED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-div-shp2-g0] PASSED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-div-shp2-g1] PASSED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-div-shp3-g0] PASSED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-div-shp3-g1] PASSED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-div-shp4-g0] PASSED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-div-shp4-g1] PASSED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-div-shp5-g0] PASSED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-div-shp5-g1] PASSED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_lhs-shp0-g0] PASSED [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_lhs-shp0-g1] PASSED [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_lhs-shp1-g0] PASSED [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_lhs-shp1-g1] PASSED [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_lhs-shp2-g0] PASSED [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_lhs-shp2-g1] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_lhs-shp3-g0] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_lhs-shp3-g1] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_lhs-shp4-g0] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_lhs-shp4-g1] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_lhs-shp5-g0] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_lhs-shp5-g1] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_rhs-shp0-g0] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_rhs-shp0-g1] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_rhs-shp1-g0] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_rhs-shp1-g1] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_rhs-shp2-g0] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_rhs-shp2-g1] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_rhs-shp3-g0] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_rhs-shp3-g1] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_rhs-shp4-g0] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_rhs-shp4-g1] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_rhs-shp5-g0] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_rhs-shp5-g1] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-add-shp0-g0] PASSED  [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-add-shp0-g1] PASSED  [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-add-shp1-g0] PASSED  [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-add-shp1-g1] PASSED  [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-add-shp2-g0] PASSED  [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-add-shp2-g1] PASSED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-add-shp3-g0] PASSED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-add-shp3-g1] PASSED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-add-shp4-g0] PASSED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-add-shp4-g1] PASSED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-add-shp5-g0] PASSED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-add-shp5-g1] PASSED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-sub-shp0-g0] PASSED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-sub-shp0-g1] PASSED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-sub-shp1-g0] PASSED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-sub-shp1-g1] PASSED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-sub-shp2-g0] PASSED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-sub-shp2-g1] PASSED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-sub-shp3-g0] PASSED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-sub-shp3-g1] PASSED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-sub-shp4-g0] PASSED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-sub-shp4-g1] PASSED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-sub-shp5-g0] PASSED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-sub-shp5-g1] PASSED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-mul-shp0-g0] PASSED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-mul-shp0-g1] PASSED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-mul-shp1-g0] PASSED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-mul-shp1-g1] PASSED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-mul-shp2-g0] PASSED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-mul-shp2-g1] PASSED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-mul-shp3-g0] PASSED  [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-mul-shp3-g1] PASSED  [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-mul-shp4-g0] PASSED  [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-mul-shp4-g1] PASSED  [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-mul-shp5-g0] PASSED  [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-mul-shp5-g1] PASSED  [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-div-shp0-g0] PASSED  [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-div-shp0-g1] PASSED  [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-div-shp1-g0] PASSED  [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-div-shp1-g1] PASSED  [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-div-shp2-g0] PASSED  [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-div-shp2-g1] PASSED  [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-div-shp3-g0] PASSED  [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-div-shp3-g1] PASSED  [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-div-shp4-g0] PASSED  [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-div-shp4-g1] PASSED  [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-div-shp5-g0] PASSED  [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-div-shp5-g1] PASSED  [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_lhs-shp0-g0] PASSED [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_lhs-shp0-g1] PASSED [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_lhs-shp1-g0] PASSED [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_lhs-shp1-g1] PASSED [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_lhs-shp2-g0] PASSED [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_lhs-shp2-g1] PASSED [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_lhs-shp3-g0] PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_lhs-shp3-g1] PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_lhs-shp4-g0] PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_lhs-shp4-g1] PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_lhs-shp5-g0] PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_lhs-shp5-g1] PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_rhs-shp0-g0] PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_rhs-shp0-g1] PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_rhs-shp1-g0] PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_rhs-shp1-g1] PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_rhs-shp2-g0] PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_rhs-shp2-g1] PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_rhs-shp3-g0] PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_rhs-shp3-g1] PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_rhs-shp4-g0] PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_rhs-shp4-g1] PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_rhs-shp5-g0] PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_rhs-shp5-g1] PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-add-shp0-g0] PASSED  [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-add-shp0-g1] PASSED  [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-add-shp1-g0] PASSED  [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-add-shp1-g1] PASSED  [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-add-shp2-g0] PASSED  [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-add-shp2-g1] PASSED  [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-add-shp3-g0] PASSED  [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-add-shp3-g1] PASSED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-add-shp4-g0] PASSED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-add-shp4-g1] PASSED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-add-shp5-g0] PASSED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-add-shp5-g1] PASSED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-sub-shp0-g0] PASSED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-sub-shp0-g1] PASSED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-sub-shp1-g0] PASSED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-sub-shp1-g1] PASSED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-sub-shp2-g0] PASSED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-sub-shp2-g1] PASSED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-sub-shp3-g0] PASSED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-sub-shp3-g1] PASSED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-sub-shp4-g0] PASSED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-sub-shp4-g1] PASSED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-sub-shp5-g0] PASSED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-sub-shp5-g1] PASSED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-mul-shp0-g0] PASSED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-mul-shp0-g1] PASSED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-mul-shp1-g0] PASSED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-mul-shp1-g1] PASSED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-mul-shp2-g0] PASSED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-mul-shp2-g1] PASSED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-mul-shp3-g0] PASSED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-mul-shp3-g1] PASSED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-mul-shp4-g0] PASSED  [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-mul-shp4-g1] PASSED  [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-mul-shp5-g0] PASSED  [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-mul-shp5-g1] PASSED  [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-div-shp0-g0] PASSED  [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-div-shp0-g1] PASSED  [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-div-shp1-g0] PASSED  [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-div-shp1-g1] PASSED  [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-div-shp2-g0] PASSED  [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-div-shp2-g1] PASSED  [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-div-shp3-g0] PASSED  [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-div-shp3-g1] PASSED  [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-div-shp4-g0] PASSED  [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-div-shp4-g1] PASSED  [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-div-shp5-g0] PASSED  [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-div-shp5-g1] PASSED  [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_lhs-shp0-g0] PASSED [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_lhs-shp0-g1] PASSED [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_lhs-shp1-g0] PASSED [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_lhs-shp1-g1] PASSED [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_lhs-shp2-g0] PASSED [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_lhs-shp2-g1] PASSED [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_lhs-shp3-g0] PASSED [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_lhs-shp3-g1] PASSED [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_lhs-shp4-g0] PASSED [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_lhs-shp4-g1] PASSED [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_lhs-shp5-g0] PASSED [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_lhs-shp5-g1] PASSED [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_rhs-shp0-g0] PASSED [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_rhs-shp0-g1] PASSED [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_rhs-shp1-g0] PASSED [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_rhs-shp1-g1] PASSED [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_rhs-shp2-g0] PASSED [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_rhs-shp2-g1] PASSED [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_rhs-shp3-g0] PASSED [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_rhs-shp3-g1] PASSED [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_rhs-shp4-g0] PASSED [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_rhs-shp4-g1] PASSED [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_rhs-shp5-g0] PASSED [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_rhs-shp5-g1] PASSED [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-add-shp0-g0] PASSED  [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-add-shp0-g1] PASSED  [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-add-shp1-g0] PASSED  [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-add-shp1-g1] PASSED  [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-add-shp2-g0] PASSED  [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-add-shp2-g1] PASSED  [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-add-shp3-g0] PASSED  [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-add-shp3-g1] PASSED  [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-add-shp4-g0] PASSED  [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-add-shp4-g1] PASSED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-add-shp5-g0] PASSED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-add-shp5-g1] PASSED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-sub-shp0-g0] PASSED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-sub-shp0-g1] PASSED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-sub-shp1-g0] PASSED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-sub-shp1-g1] PASSED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-sub-shp2-g0] PASSED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-sub-shp2-g1] PASSED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-sub-shp3-g0] PASSED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-sub-shp3-g1] PASSED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-sub-shp4-g0] PASSED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-sub-shp4-g1] PASSED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-sub-shp5-g0] PASSED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-sub-shp5-g1] PASSED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-mul-shp0-g0] PASSED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-mul-shp0-g1] PASSED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-mul-shp1-g0] PASSED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-mul-shp1-g1] PASSED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-mul-shp2-g0] PASSED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-mul-shp2-g1] PASSED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-mul-shp3-g0] PASSED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-mul-shp3-g1] PASSED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-mul-shp4-g0] PASSED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-mul-shp4-g1] PASSED  [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-mul-shp5-g0] PASSED  [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-mul-shp5-g1] PASSED  [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-div-shp0-g0] PASSED  [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-div-shp0-g1] PASSED  [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-div-shp1-g0] PASSED  [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-div-shp1-g1] PASSED  [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-div-shp2-g0] PASSED  [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-div-shp2-g1] PASSED  [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-div-shp3-g0] PASSED  [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-div-shp3-g1] PASSED  [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-div-shp4-g0] PASSED  [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-div-shp4-g1] PASSED  [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-div-shp5-g0] PASSED  [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-div-shp5-g1] PASSED  [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_lhs-shp0-g0] PASSED [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_lhs-shp0-g1] PASSED [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_lhs-shp1-g0] PASSED [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_lhs-shp1-g1] PASSED [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_lhs-shp2-g0] PASSED [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_lhs-shp2-g1] PASSED [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_lhs-shp3-g0] PASSED [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_lhs-shp3-g1] PASSED [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_lhs-shp4-g0] PASSED [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_lhs-shp4-g1] PASSED [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_lhs-shp5-g0] PASSED [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_lhs-shp5-g1] PASSED [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_rhs-shp0-g0] PASSED [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_rhs-shp0-g1] PASSED [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_rhs-shp1-g0] PASSED [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_rhs-shp1-g1] PASSED [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_rhs-shp2-g0] PASSED [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_rhs-shp2-g1] PASSED [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_rhs-shp3-g0] PASSED [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_rhs-shp3-g1] PASSED [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_rhs-shp4-g0] PASSED [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_rhs-shp4-g1] PASSED [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_rhs-shp5-g0] PASSED [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_rhs-shp5-g1] PASSED [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-add-shp0-g0] PASSED  [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-add-shp0-g1] PASSED  [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-add-shp1-g0] PASSED  [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-add-shp1-g1] PASSED  [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-add-shp2-g0] PASSED  [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-add-shp2-g1] PASSED  [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-add-shp3-g0] PASSED  [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-add-shp3-g1] PASSED  [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-add-shp4-g0] PASSED  [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-add-shp4-g1] PASSED  [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-add-shp5-g0] PASSED  [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-add-shp5-g1] PASSED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-sub-shp0-g0] PASSED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-sub-shp0-g1] PASSED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-sub-shp1-g0] PASSED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-sub-shp1-g1] PASSED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-sub-shp2-g0] PASSED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-sub-shp2-g1] PASSED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-sub-shp3-g0] PASSED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-sub-shp3-g1] PASSED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-sub-shp4-g0] PASSED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-sub-shp4-g1] PASSED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-sub-shp5-g0] PASSED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-sub-shp5-g1] PASSED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-mul-shp0-g0] PASSED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-mul-shp0-g1] PASSED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-mul-shp1-g0] PASSED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-mul-shp1-g1] PASSED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-mul-shp2-g0] PASSED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-mul-shp2-g1] PASSED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-mul-shp3-g0] PASSED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-mul-shp3-g1] PASSED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-mul-shp4-g0] PASSED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-mul-shp4-g1] PASSED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-mul-shp5-g0] PASSED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-mul-shp5-g1] PASSED  [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-div-shp0-g0] PASSED  [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-div-shp0-g1] PASSED  [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-div-shp1-g0] PASSED  [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-div-shp1-g1] PASSED  [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-div-shp2-g0] PASSED  [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-div-shp2-g1] PASSED  [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-div-shp3-g0] PASSED  [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-div-shp3-g1] PASSED  [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-div-shp4-g0] PASSED  [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-div-shp4-g1] PASSED  [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-div-shp5-g0] PASSED  [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-div-shp5-g1] PASSED  [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_lhs-shp0-g0] PASSED [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_lhs-shp0-g1] PASSED [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_lhs-shp1-g0] PASSED [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_lhs-shp1-g1] PASSED [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_lhs-shp2-g0] PASSED [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_lhs-shp2-g1] PASSED [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_lhs-shp3-g0] PASSED [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_lhs-shp3-g1] PASSED [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_lhs-shp4-g0] PASSED [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_lhs-shp4-g1] PASSED [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_lhs-shp5-g0] PASSED [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_lhs-shp5-g1] PASSED [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_rhs-shp0-g0] PASSED [ 40%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_rhs-shp0-g1] PASSED [ 40%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_rhs-shp1-g0] PASSED [ 40%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_rhs-shp1-g1] PASSED [ 40%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_rhs-shp2-g0] PASSED [ 40%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_rhs-shp2-g1] PASSED [ 40%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_rhs-shp3-g0] PASSED [ 40%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_rhs-shp3-g1] PASSED [ 40%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_rhs-shp4-g0] PASSED [ 40%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_rhs-shp4-g1] PASSED [ 40%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_rhs-shp5-g0] PASSED [ 40%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_rhs-shp5-g1] PASSED [ 40%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-u-shp0-g0] PASSED [ 40%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-u-shp0-g1] PASSED [ 40%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-u-shp1-g0] PASSED [ 40%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-u-shp1-g1] PASSED [ 40%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-u-shp2-g0] PASSED [ 40%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-u-shp2-g1] PASSED [ 40%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-u-shp3-g0] PASSED [ 40%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-u-shp3-g1] PASSED [ 40%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-u-shp4-g0] PASSED [ 40%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-u-shp4-g1] PASSED [ 40%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-v-shp0-g0] PASSED [ 40%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-v-shp0-g1] PASSED [ 40%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-v-shp1-g0] PASSED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-v-shp1-g1] PASSED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-v-shp2-g0] PASSED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-v-shp2-g1] PASSED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-v-shp3-g0] PASSED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-v-shp3-g1] PASSED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-v-shp4-g0] PASSED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-v-shp4-g1] PASSED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-e-shp0-g0] PASSED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-e-shp0-g1] PASSED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-e-shp1-g0] PASSED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-e-shp1-g1] PASSED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-e-shp2-g0] PASSED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-e-shp2-g1] PASSED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-e-shp3-g0] PASSED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-e-shp3-g1] PASSED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-e-shp4-g0] PASSED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-e-shp4-g1] PASSED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-u-shp0-g0] PASSED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-u-shp0-g1] PASSED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-u-shp1-g0] PASSED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-u-shp1-g1] PASSED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-u-shp2-g0] PASSED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-u-shp2-g1] PASSED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-u-shp3-g0] PASSED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-u-shp3-g1] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-u-shp4-g0] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-u-shp4-g1] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-v-shp0-g0] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-v-shp0-g1] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-v-shp1-g0] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-v-shp1-g1] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-v-shp2-g0] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-v-shp2-g1] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-v-shp3-g0] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-v-shp3-g1] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-v-shp4-g0] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-v-shp4-g1] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-e-shp0-g0] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-e-shp0-g1] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-e-shp1-g0] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-e-shp1-g1] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-e-shp2-g0] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-e-shp2-g1] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-e-shp3-g0] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-e-shp3-g1] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-e-shp4-g0] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-e-shp4-g1] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-u-shp0-g0] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-u-shp0-g1] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-u-shp1-g0] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-u-shp1-g1] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-u-shp2-g0] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-u-shp2-g1] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-u-shp3-g0] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-u-shp3-g1] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-u-shp4-g0] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-u-shp4-g1] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-v-shp0-g0] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-v-shp0-g1] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-v-shp1-g0] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-v-shp1-g1] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-v-shp2-g0] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-v-shp2-g1] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-v-shp3-g0] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-v-shp3-g1] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-v-shp4-g0] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-v-shp4-g1] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-e-shp0-g0] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-e-shp0-g1] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-e-shp1-g0] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-e-shp1-g1] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-e-shp2-g0] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-e-shp2-g1] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-e-shp3-g0] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-e-shp3-g1] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-e-shp4-g0] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-e-shp4-g1] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-u-shp0-g0] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-u-shp0-g1] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-u-shp1-g0] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-u-shp1-g1] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-u-shp2-g0] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-u-shp2-g1] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-u-shp3-g0] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-u-shp3-g1] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-u-shp4-g0] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-u-shp4-g1] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-v-shp0-g0] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-v-shp0-g1] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-v-shp1-g0] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-v-shp1-g1] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-v-shp2-g0] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-v-shp2-g1] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-v-shp3-g0] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-v-shp3-g1] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-v-shp4-g0] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-v-shp4-g1] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-e-shp0-g0] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-e-shp0-g1] PASSED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-e-shp1-g0] PASSED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-e-shp1-g1] PASSED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-e-shp2-g0] PASSED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-e-shp2-g1] PASSED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-e-shp3-g0] PASSED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-e-shp3-g1] PASSED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-e-shp4-g0] PASSED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-e-shp4-g1] PASSED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-u-shp0-g0] PASSED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-u-shp0-g1] PASSED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-u-shp1-g0] PASSED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-u-shp1-g1] PASSED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-u-shp2-g0] PASSED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-u-shp2-g1] PASSED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-u-shp3-g0] PASSED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-u-shp3-g1] PASSED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-u-shp4-g0] PASSED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-u-shp4-g1] PASSED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-v-shp0-g0] PASSED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-v-shp0-g1] PASSED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-v-shp1-g0] PASSED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-v-shp1-g1] PASSED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-v-shp2-g0] PASSED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-v-shp2-g1] PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-v-shp3-g0] PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-v-shp3-g1] PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-v-shp4-g0] PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-v-shp4-g1] PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-e-shp0-g0] PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-e-shp0-g1] PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-e-shp1-g0] PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-e-shp1-g1] PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-e-shp2-g0] PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-e-shp2-g1] PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-e-shp3-g0] PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-e-shp3-g1] PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-e-shp4-g0] PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-e-shp4-g1] PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-u-shp0-g0] PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-u-shp0-g1] PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-u-shp1-g0] PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-u-shp1-g1] PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-u-shp2-g0] PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-u-shp2-g1] PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-u-shp3-g0] PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-u-shp3-g1] PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-u-shp4-g0] PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-u-shp4-g1] PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-v-shp0-g0] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-v-shp0-g1] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-v-shp1-g0] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-v-shp1-g1] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-v-shp2-g0] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-v-shp2-g1] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-v-shp3-g0] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-v-shp3-g1] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-v-shp4-g0] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-v-shp4-g1] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-e-shp0-g0] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-e-shp0-g1] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-e-shp1-g0] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-e-shp1-g1] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-e-shp2-g0] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-e-shp2-g1] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-e-shp3-g0] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-e-shp3-g1] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-e-shp4-g0] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-e-shp4-g1] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-u-shp0-g0] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-u-shp0-g1] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-u-shp1-g0] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-u-shp1-g1] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-u-shp2-g0] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-u-shp2-g1] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-u-shp3-g0] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-u-shp3-g1] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-u-shp4-g0] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-u-shp4-g1] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-v-shp0-g0] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-v-shp0-g1] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-v-shp1-g0] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-v-shp1-g1] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-v-shp2-g0] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-v-shp2-g1] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-v-shp3-g0] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-v-shp3-g1] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-v-shp4-g0] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-v-shp4-g1] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-e-shp0-g0] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-e-shp0-g1] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-e-shp1-g0] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-e-shp1-g1] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-e-shp2-g0] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-e-shp2-g1] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-e-shp3-g0] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-e-shp3-g1] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-e-shp4-g0] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-e-shp4-g1] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-u-shp0-g0] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-u-shp0-g1] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-u-shp1-g0] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-u-shp1-g1] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-u-shp2-g0] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-u-shp2-g1] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-u-shp3-g0] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-u-shp3-g1] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-u-shp4-g0] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-u-shp4-g1] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-v-shp0-g0] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-v-shp0-g1] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-v-shp1-g0] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-v-shp1-g1] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-v-shp2-g0] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-v-shp2-g1] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-v-shp3-g0] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-v-shp3-g1] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-v-shp4-g0] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-v-shp4-g1] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-e-shp0-g0] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-e-shp0-g1] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-e-shp1-g0] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-e-shp1-g1] PASSED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-e-shp2-g0] PASSED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-e-shp2-g1] PASSED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-e-shp3-g0] PASSED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-e-shp3-g1] PASSED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-e-shp4-g0] PASSED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-e-shp4-g1] PASSED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-u-shp0-g0] PASSED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-u-shp0-g1] PASSED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-u-shp1-g0] PASSED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-u-shp1-g1] PASSED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-u-shp2-g0] PASSED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-u-shp2-g1] PASSED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-u-shp3-g0] PASSED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-u-shp3-g1] PASSED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-u-shp4-g0] PASSED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-u-shp4-g1] PASSED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-v-shp0-g0] PASSED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-v-shp0-g1] PASSED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-v-shp1-g0] PASSED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-v-shp1-g1] PASSED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-v-shp2-g0] PASSED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-v-shp2-g1] PASSED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-v-shp3-g0] PASSED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-v-shp3-g1] PASSED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-v-shp4-g0] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-v-shp4-g1] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-e-shp0-g0] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-e-shp0-g1] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-e-shp1-g0] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-e-shp1-g1] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-e-shp2-g0] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-e-shp2-g1] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-e-shp3-g0] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-e-shp3-g1] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-e-shp4-g0] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-e-shp4-g1] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-u-shp0-g0] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-u-shp0-g1] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-u-shp1-g0] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-u-shp1-g1] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-u-shp2-g0] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-u-shp2-g1] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-u-shp3-g0] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-u-shp3-g1] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-u-shp4-g0] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-u-shp4-g1] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-v-shp0-g0] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-v-shp0-g1] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-v-shp1-g0] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-v-shp1-g1] PASSED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-v-shp2-g0] PASSED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-v-shp2-g1] PASSED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-v-shp3-g0] PASSED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-v-shp3-g1] PASSED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-v-shp4-g0] PASSED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-v-shp4-g1] PASSED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-e-shp0-g0] PASSED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-e-shp0-g1] PASSED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-e-shp1-g0] PASSED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-e-shp1-g1] PASSED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-e-shp2-g0] PASSED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-e-shp2-g1] PASSED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-e-shp3-g0] PASSED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-e-shp3-g1] PASSED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-e-shp4-g0] PASSED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-e-shp4-g1] PASSED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-u-shp0-g0] PASSED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-u-shp0-g1] PASSED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-u-shp1-g0] PASSED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-u-shp1-g1] PASSED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-u-shp2-g0] PASSED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-u-shp2-g1] PASSED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-u-shp3-g0] PASSED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-u-shp3-g1] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-u-shp4-g0] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-u-shp4-g1] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-v-shp0-g0] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-v-shp0-g1] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-v-shp1-g0] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-v-shp1-g1] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-v-shp2-g0] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-v-shp2-g1] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-v-shp3-g0] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-v-shp3-g1] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-v-shp4-g0] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-v-shp4-g1] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-e-shp0-g0] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-e-shp0-g1] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-e-shp1-g0] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-e-shp1-g1] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-e-shp2-g0] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-e-shp2-g1] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-e-shp3-g0] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-e-shp3-g1] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-e-shp4-g0] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-e-shp4-g1] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-u-shp0-g0] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-u-shp0-g1] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-u-shp1-g0] PASSED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-u-shp1-g1] PASSED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-u-shp2-g0] PASSED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-u-shp2-g1] PASSED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-u-shp3-g0] PASSED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-u-shp3-g1] PASSED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-u-shp4-g0] PASSED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-u-shp4-g1] PASSED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-v-shp0-g0] PASSED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-v-shp0-g1] PASSED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-v-shp1-g0] PASSED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-v-shp1-g1] PASSED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-v-shp2-g0] PASSED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-v-shp2-g1] PASSED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-v-shp3-g0] PASSED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-v-shp3-g1] PASSED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-v-shp4-g0] PASSED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-v-shp4-g1] PASSED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-e-shp0-g0] PASSED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-e-shp0-g1] PASSED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-e-shp1-g0] PASSED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-e-shp1-g1] PASSED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-e-shp2-g0] PASSED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-e-shp2-g1] PASSED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-e-shp3-g0] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-e-shp3-g1] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-e-shp4-g0] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-e-shp4-g1] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-u-shp0-g0] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-u-shp0-g1] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-u-shp1-g0] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-u-shp1-g1] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-u-shp2-g0] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-u-shp2-g1] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-u-shp3-g0] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-u-shp3-g1] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-u-shp4-g0] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-u-shp4-g1] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-v-shp0-g0] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-v-shp0-g1] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-v-shp1-g0] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-v-shp1-g1] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-v-shp2-g0] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-v-shp2-g1] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-v-shp3-g0] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-v-shp3-g1] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-v-shp4-g0] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-v-shp4-g1] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-e-shp0-g0] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-e-shp0-g1] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-e-shp1-g0] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-e-shp1-g1] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-e-shp2-g0] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-e-shp2-g1] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-e-shp3-g0] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-e-shp3-g1] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-e-shp4-g0] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-e-shp4-g1] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-u-shp0-g0] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-u-shp0-g1] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-u-shp1-g0] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-u-shp1-g1] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-u-shp2-g0] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-u-shp2-g1] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-u-shp3-g0] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-u-shp3-g1] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-u-shp4-g0] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-u-shp4-g1] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-v-shp0-g0] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-v-shp0-g1] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-v-shp1-g0] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-v-shp1-g1] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-v-shp2-g0] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-v-shp2-g1] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-v-shp3-g0] PASSED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-v-shp3-g1] PASSED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-v-shp4-g0] PASSED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-v-shp4-g1] PASSED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-e-shp0-g0] PASSED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-e-shp0-g1] PASSED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-e-shp1-g0] PASSED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-e-shp1-g1] PASSED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-e-shp2-g0] PASSED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-e-shp2-g1] PASSED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-e-shp3-g0] PASSED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-e-shp3-g1] PASSED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-e-shp4-g0] PASSED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-e-shp4-g1] PASSED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-u-shp0-g0] PASSED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-u-shp0-g1] PASSED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-u-shp1-g0] PASSED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-u-shp1-g1] PASSED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-u-shp2-g0] PASSED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-u-shp2-g1] PASSED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-u-shp3-g0] PASSED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-u-shp3-g1] PASSED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-u-shp4-g0] PASSED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-u-shp4-g1] PASSED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-v-shp0-g0] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-v-shp0-g1] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-v-shp1-g0] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-v-shp1-g1] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-v-shp2-g0] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-v-shp2-g1] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-v-shp3-g0] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-v-shp3-g1] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-v-shp4-g0] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-v-shp4-g1] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-e-shp0-g0] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-e-shp0-g1] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-e-shp1-g0] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-e-shp1-g1] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-e-shp2-g0] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-e-shp2-g1] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-e-shp3-g0] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-e-shp3-g1] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-e-shp4-g0] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-e-shp4-g1] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-u-shp0-g0] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-u-shp0-g1] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-u-shp1-g0] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-u-shp1-g1] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-u-shp2-g0] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-u-shp2-g1] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-u-shp3-g0] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-u-shp3-g1] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-u-shp4-g0] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-u-shp4-g1] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-v-shp0-g0] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-v-shp0-g1] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-v-shp1-g0] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-v-shp1-g1] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-v-shp2-g0] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-v-shp2-g1] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-v-shp3-g0] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-v-shp3-g1] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-v-shp4-g0] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-v-shp4-g1] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-e-shp0-g0] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-e-shp0-g1] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-e-shp1-g0] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-e-shp1-g1] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-e-shp2-g0] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-e-shp2-g1] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-e-shp3-g0] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-e-shp3-g1] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-e-shp4-g0] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-e-shp4-g1] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-u-shp0-g0] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-u-shp0-g1] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-u-shp1-g0] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-u-shp1-g1] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-u-shp2-g0] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-u-shp2-g1] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-u-shp3-g0] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-u-shp3-g1] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-u-shp4-g0] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-u-shp4-g1] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-v-shp0-g0] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-v-shp0-g1] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-v-shp1-g0] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-v-shp1-g1] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-v-shp2-g0] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-v-shp2-g1] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-v-shp3-g0] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-v-shp3-g1] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-v-shp4-g0] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-v-shp4-g1] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-e-shp0-g0] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-e-shp0-g1] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-e-shp1-g0] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-e-shp1-g1] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-e-shp2-g0] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-e-shp2-g1] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-e-shp3-g0] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-e-shp3-g1] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-e-shp4-g0] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-e-shp4-g1] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-u-shp0-g0] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-u-shp0-g1] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-u-shp1-g0] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-u-shp1-g1] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-u-shp2-g0] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-u-shp2-g1] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-u-shp3-g0] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-u-shp3-g1] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-u-shp4-g0] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-u-shp4-g1] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-v-shp0-g0] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-v-shp0-g1] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-v-shp1-g0] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-v-shp1-g1] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-v-shp2-g0] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-v-shp2-g1] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-v-shp3-g0] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-v-shp3-g1] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-v-shp4-g0] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-v-shp4-g1] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-e-shp0-g0] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-e-shp0-g1] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-e-shp1-g0] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-e-shp1-g1] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-e-shp2-g0] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-e-shp2-g1] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-e-shp3-g0] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-e-shp3-g1] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-e-shp4-g0] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-e-shp4-g1] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-u-shp0-g0] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-u-shp0-g1] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-u-shp1-g0] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-u-shp1-g1] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-u-shp2-g0] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-u-shp2-g1] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-u-shp3-g0] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-u-shp3-g1] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-u-shp4-g0] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-u-shp4-g1] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-v-shp0-g0] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-v-shp0-g1] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-v-shp1-g0] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-v-shp1-g1] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-v-shp2-g0] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-v-shp2-g1] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-v-shp3-g0] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-v-shp3-g1] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-v-shp4-g0] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-v-shp4-g1] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-e-shp0-g0] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-e-shp0-g1] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-e-shp1-g0] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-e-shp1-g1] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-e-shp2-g0] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-e-shp2-g1] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-e-shp3-g0] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-e-shp3-g1] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-e-shp4-g0] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-e-shp4-g1] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-u-shp0-g0] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-u-shp0-g1] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-u-shp1-g0] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-u-shp1-g1] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-u-shp2-g0] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-u-shp2-g1] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-u-shp3-g0] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-u-shp3-g1] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-u-shp4-g0] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-u-shp4-g1] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-v-shp0-g0] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-v-shp0-g1] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-v-shp1-g0] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-v-shp1-g1] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-v-shp2-g0] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-v-shp2-g1] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-v-shp3-g0] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-v-shp3-g1] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-v-shp4-g0] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-v-shp4-g1] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-e-shp0-g0] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-e-shp0-g1] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-e-shp1-g0] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-e-shp1-g1] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-e-shp2-g0] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-e-shp2-g1] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-e-shp3-g0] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-e-shp3-g1] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-e-shp4-g0] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-e-shp4-g1] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-u-shp0-g0] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-u-shp0-g1] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-u-shp1-g0] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-u-shp1-g1] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-u-shp2-g0] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-u-shp2-g1] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-u-shp3-g0] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-u-shp3-g1] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-u-shp4-g0] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-u-shp4-g1] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-v-shp0-g0] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-v-shp0-g1] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-v-shp1-g0] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-v-shp1-g1] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-v-shp2-g0] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-v-shp2-g1] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-v-shp3-g0] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-v-shp3-g1] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-v-shp4-g0] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-v-shp4-g1] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-e-shp0-g0] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-e-shp0-g1] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-e-shp1-g0] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-e-shp1-g1] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-e-shp2-g0] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-e-shp2-g1] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-e-shp3-g0] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-e-shp3-g1] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-e-shp4-g0] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-e-shp4-g1] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-u-shp0-g0] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-u-shp0-g1] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-u-shp1-g0] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-u-shp1-g1] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-u-shp2-g0] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-u-shp2-g1] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-u-shp3-g0] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-u-shp3-g1] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-u-shp4-g0] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-u-shp4-g1] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-v-shp0-g0] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-v-shp0-g1] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-v-shp1-g0] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-v-shp1-g1] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-v-shp2-g0] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-v-shp2-g1] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-v-shp3-g0] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-v-shp3-g1] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-v-shp4-g0] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-v-shp4-g1] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-e-shp0-g0] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-e-shp0-g1] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-e-shp1-g0] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-e-shp1-g1] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-e-shp2-g0] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-e-shp2-g1] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-e-shp3-g0] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-e-shp3-g1] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-e-shp4-g0] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-e-shp4-g1] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-u-shp0-g0] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-u-shp0-g1] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-u-shp1-g0] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-u-shp1-g1] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-u-shp2-g0] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-u-shp2-g1] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-u-shp3-g0] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-u-shp3-g1] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-u-shp4-g0] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-u-shp4-g1] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-v-shp0-g0] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-v-shp0-g1] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-v-shp1-g0] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-v-shp1-g1] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-v-shp2-g0] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-v-shp2-g1] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-v-shp3-g0] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-v-shp3-g1] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-v-shp4-g0] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-v-shp4-g1] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-e-shp0-g0] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-e-shp0-g1] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-e-shp1-g0] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-e-shp1-g1] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-e-shp2-g0] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-e-shp2-g1] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-e-shp3-g0] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-e-shp3-g1] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-e-shp4-g0] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-e-shp4-g1] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-u-shp0-g0] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-u-shp0-g1] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-u-shp1-g0] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-u-shp1-g1] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-u-shp2-g0] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-u-shp2-g1] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-u-shp3-g0] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-u-shp3-g1] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-u-shp4-g0] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-u-shp4-g1] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-v-shp0-g0] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-v-shp0-g1] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-v-shp1-g0] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-v-shp1-g1] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-v-shp2-g0] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-v-shp2-g1] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-v-shp3-g0] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-v-shp3-g1] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-v-shp4-g0] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-v-shp4-g1] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-e-shp0-g0] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-e-shp0-g1] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-e-shp1-g0] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-e-shp1-g1] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-e-shp2-g0] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-e-shp2-g1] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-e-shp3-g0] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-e-shp3-g1] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-e-shp4-g0] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-e-shp4-g1] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-u-shp0-g0] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-u-shp0-g1] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-u-shp1-g0] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-u-shp1-g1] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-u-shp2-g0] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-u-shp2-g1] PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-u-shp3-g0] PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-u-shp3-g1] PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-u-shp4-g0] PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-u-shp4-g1] PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-v-shp0-g0] PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-v-shp0-g1] PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-v-shp1-g0] PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-v-shp1-g1] PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-v-shp2-g0] PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-v-shp2-g1] PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-v-shp3-g0] PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-v-shp3-g1] PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-v-shp4-g0] PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-v-shp4-g1] PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-e-shp0-g0] PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-e-shp0-g1] PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-e-shp1-g0] PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-e-shp1-g1] PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-e-shp2-g0] PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-e-shp2-g1] PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-e-shp3-g0] PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-e-shp3-g1] PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-e-shp4-g0] PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-e-shp4-g1] PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-u-shp0-g0] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-u-shp0-g1] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-u-shp1-g0] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-u-shp1-g1] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-u-shp2-g0] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-u-shp2-g1] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-u-shp3-g0] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-u-shp3-g1] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-u-shp4-g0] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-u-shp4-g1] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-v-shp0-g0] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-v-shp0-g1] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-v-shp1-g0] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-v-shp1-g1] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-v-shp2-g0] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-v-shp2-g1] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-v-shp3-g0] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-v-shp3-g1] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-v-shp4-g0] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-v-shp4-g1] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-e-shp0-g0] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-e-shp0-g1] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-e-shp1-g0] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-e-shp1-g1] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-e-shp2-g0] PASSED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-e-shp2-g1] PASSED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-e-shp3-g0] PASSED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-e-shp3-g1] PASSED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-e-shp4-g0] PASSED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-e-shp4-g1] PASSED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-u-shp0-g0] PASSED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-u-shp0-g1] PASSED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-u-shp1-g0] PASSED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-u-shp1-g1] PASSED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-u-shp2-g0] PASSED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-u-shp2-g1] PASSED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-u-shp3-g0] PASSED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-u-shp3-g1] PASSED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-u-shp4-g0] PASSED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-u-shp4-g1] PASSED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-v-shp0-g0] PASSED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-v-shp0-g1] PASSED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-v-shp1-g0] PASSED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-v-shp1-g1] PASSED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-v-shp2-g0] PASSED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-v-shp2-g1] PASSED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-v-shp3-g0] PASSED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-v-shp3-g1] PASSED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-v-shp4-g0] PASSED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-v-shp4-g1] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-e-shp0-g0] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-e-shp0-g1] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-e-shp1-g0] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-e-shp1-g1] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-e-shp2-g0] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-e-shp2-g1] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-e-shp3-g0] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-e-shp3-g1] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-e-shp4-g0] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-e-shp4-g1] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-u-shp0-g0] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-u-shp0-g1] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-u-shp1-g0] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-u-shp1-g1] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-u-shp2-g0] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-u-shp2-g1] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-u-shp3-g0] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-u-shp3-g1] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-u-shp4-g0] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-u-shp4-g1] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-v-shp0-g0] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-v-shp0-g1] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-v-shp1-g0] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-v-shp1-g1] PASSED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-v-shp2-g0] PASSED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-v-shp2-g1] PASSED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-v-shp3-g0] PASSED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-v-shp3-g1] PASSED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-v-shp4-g0] PASSED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-v-shp4-g1] PASSED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-e-shp0-g0] PASSED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-e-shp0-g1] PASSED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-e-shp1-g0] PASSED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-e-shp1-g1] PASSED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-e-shp2-g0] PASSED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-e-shp2-g1] PASSED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-e-shp3-g0] PASSED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-e-shp3-g1] PASSED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-e-shp4-g0] PASSED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-e-shp4-g1] PASSED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-u-shp0-g0] PASSED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-u-shp0-g1] PASSED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-u-shp1-g0] PASSED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-u-shp1-g1] PASSED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-u-shp2-g0] PASSED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-u-shp2-g1] PASSED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-u-shp3-g0] PASSED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-u-shp3-g1] PASSED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-u-shp4-g0] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-u-shp4-g1] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-v-shp0-g0] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-v-shp0-g1] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-v-shp1-g0] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-v-shp1-g1] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-v-shp2-g0] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-v-shp2-g1] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-v-shp3-g0] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-v-shp3-g1] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-v-shp4-g0] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-v-shp4-g1] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-e-shp0-g0] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-e-shp0-g1] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-e-shp1-g0] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-e-shp1-g1] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-e-shp2-g0] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-e-shp2-g1] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-e-shp3-g0] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-e-shp3-g1] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-e-shp4-g0] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-e-shp4-g1] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-u-shp0-g0] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-u-shp0-g1] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-u-shp1-g0] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-u-shp1-g1] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-u-shp2-g0] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-u-shp2-g1] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-u-shp3-g0] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-u-shp3-g1] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-u-shp4-g0] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-u-shp4-g1] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-v-shp0-g0] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-v-shp0-g1] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-v-shp1-g0] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-v-shp1-g1] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-v-shp2-g0] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-v-shp2-g1] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-v-shp3-g0] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-v-shp3-g1] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-v-shp4-g0] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-v-shp4-g1] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-e-shp0-g0] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-e-shp0-g1] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-e-shp1-g0] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-e-shp1-g1] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-e-shp2-g0] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-e-shp2-g1] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-e-shp3-g0] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-e-shp3-g1] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-e-shp4-g0] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-e-shp4-g1] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-u-shp0-g0] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-u-shp0-g1] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-u-shp1-g0] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-u-shp1-g1] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-u-shp2-g0] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-u-shp2-g1] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-u-shp3-g0] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-u-shp3-g1] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-u-shp4-g0] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-u-shp4-g1] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-v-shp0-g0] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-v-shp0-g1] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-v-shp1-g0] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-v-shp1-g1] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-v-shp2-g0] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-v-shp2-g1] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-v-shp3-g0] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-v-shp3-g1] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-v-shp4-g0] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-v-shp4-g1] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-e-shp0-g0] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-e-shp0-g1] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-e-shp1-g0] PASSED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-e-shp1-g1] PASSED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-e-shp2-g0] PASSED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-e-shp2-g1] PASSED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-e-shp3-g0] PASSED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-e-shp3-g1] PASSED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-e-shp4-g0] PASSED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-e-shp4-g1] PASSED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-u-shp0-g0] PASSED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-u-shp0-g1] PASSED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-u-shp1-g0] PASSED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-u-shp1-g1] PASSED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-u-shp2-g0] PASSED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-u-shp2-g1] PASSED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-u-shp3-g0] PASSED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-u-shp3-g1] PASSED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-u-shp4-g0] PASSED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-u-shp4-g1] PASSED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-v-shp0-g0] PASSED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-v-shp0-g1] PASSED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-v-shp1-g0] PASSED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-v-shp1-g1] PASSED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-v-shp2-g0] PASSED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-v-shp2-g1] PASSED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-v-shp3-g0] PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-v-shp3-g1] PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-v-shp4-g0] PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-v-shp4-g1] PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-e-shp0-g0] PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-e-shp0-g1] PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-e-shp1-g0] PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-e-shp1-g1] PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-e-shp2-g0] PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-e-shp2-g1] PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-e-shp3-g0] PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-e-shp3-g1] PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-e-shp4-g0] PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-e-shp4-g1] PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-u-shp0-g0] PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-u-shp0-g1] PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-u-shp1-g0] PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-u-shp1-g1] PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-u-shp2-g0] PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-u-shp2-g1] PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-u-shp3-g0] PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-u-shp3-g1] PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-u-shp4-g0] PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-u-shp4-g1] PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-v-shp0-g0] PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-v-shp0-g1] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-v-shp1-g0] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-v-shp1-g1] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-v-shp2-g0] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-v-shp2-g1] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-v-shp3-g0] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-v-shp3-g1] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-v-shp4-g0] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-v-shp4-g1] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-e-shp0-g0] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-e-shp0-g1] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-e-shp1-g0] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-e-shp1-g1] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-e-shp2-g0] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-e-shp2-g1] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-e-shp3-g0] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-e-shp3-g1] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-e-shp4-g0] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-e-shp4-g1] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-u-shp0-g0] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-u-shp0-g1] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-u-shp1-g0] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-u-shp1-g1] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-u-shp2-g0] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-u-shp2-g1] PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-u-shp3-g0] PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-u-shp3-g1] PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-u-shp4-g0] PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-u-shp4-g1] PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-v-shp0-g0] PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-v-shp0-g1] PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-v-shp1-g0] PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-v-shp1-g1] PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-v-shp2-g0] PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-v-shp2-g1] PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-v-shp3-g0] PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-v-shp3-g1] PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-v-shp4-g0] PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-v-shp4-g1] PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-e-shp0-g0] PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-e-shp0-g1] PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-e-shp1-g0] PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-e-shp1-g1] PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-e-shp2-g0] PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-e-shp2-g1] PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-e-shp3-g0] PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-e-shp3-g1] PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-e-shp4-g0] PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-e-shp4-g1] PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-u-shp0-g0] PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-u-shp0-g1] PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-u-shp1-g0] PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-u-shp1-g1] PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-u-shp2-g0] PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-u-shp2-g1] PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-u-shp3-g0] PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-u-shp3-g1] PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-u-shp4-g0] PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-u-shp4-g1] PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-v-shp0-g0] PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-v-shp0-g1] PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-v-shp1-g0] PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-v-shp1-g1] PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-v-shp2-g0] PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-v-shp2-g1] PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-v-shp3-g0] PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-v-shp3-g1] PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-v-shp4-g0] PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-v-shp4-g1] PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-e-shp0-g0] PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-e-shp0-g1] PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-e-shp1-g0] PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-e-shp1-g1] PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-e-shp2-g0] PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-e-shp2-g1] PASSED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-e-shp3-g0] PASSED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-e-shp3-g1] PASSED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-e-shp4-g0] PASSED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-e-shp4-g1] PASSED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-u-shp0-g0] PASSED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-u-shp0-g1] PASSED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-u-shp1-g0] PASSED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-u-shp1-g1] PASSED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-u-shp2-g0] PASSED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-u-shp2-g1] PASSED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-u-shp3-g0] PASSED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-u-shp3-g1] PASSED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-u-shp4-g0] PASSED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-u-shp4-g1] PASSED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-v-shp0-g0] PASSED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-v-shp0-g1] PASSED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-v-shp1-g0] PASSED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-v-shp1-g1] PASSED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-v-shp2-g0] PASSED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-v-shp2-g1] PASSED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-v-shp3-g0] PASSED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-v-shp3-g1] PASSED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-v-shp4-g0] PASSED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-v-shp4-g1] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-e-shp0-g0] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-e-shp0-g1] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-e-shp1-g0] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-e-shp1-g1] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-e-shp2-g0] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-e-shp2-g1] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-e-shp3-g0] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-e-shp3-g1] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-e-shp4-g0] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-e-shp4-g1] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-u-shp0-g0] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-u-shp0-g1] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-u-shp1-g0] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-u-shp1-g1] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-u-shp2-g0] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-u-shp2-g1] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-u-shp3-g0] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-u-shp3-g1] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-u-shp4-g0] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-u-shp4-g1] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-v-shp0-g0] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-v-shp0-g1] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-v-shp1-g0] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-v-shp1-g1] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-v-shp2-g0] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-v-shp2-g1] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-v-shp3-g0] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-v-shp3-g1] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-v-shp4-g0] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-v-shp4-g1] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-e-shp0-g0] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-e-shp0-g1] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-e-shp1-g0] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-e-shp1-g1] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-e-shp2-g0] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-e-shp2-g1] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-e-shp3-g0] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-e-shp3-g1] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-e-shp4-g0] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-e-shp4-g1] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-u-shp0-g0] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-u-shp0-g1] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-u-shp1-g0] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-u-shp1-g1] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-u-shp2-g0] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-u-shp2-g1] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-u-shp3-g0] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-u-shp3-g1] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-u-shp4-g0] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-u-shp4-g1] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-v-shp0-g0] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-v-shp0-g1] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-v-shp1-g0] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-v-shp1-g1] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-v-shp2-g0] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-v-shp2-g1] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-v-shp3-g0] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-v-shp3-g1] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-v-shp4-g0] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-v-shp4-g1] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-e-shp0-g0] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-e-shp0-g1] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-e-shp1-g0] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-e-shp1-g1] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-e-shp2-g0] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-e-shp2-g1] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-e-shp3-g0] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-e-shp3-g1] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-e-shp4-g0] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-e-shp4-g1] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-u-shp0-g0] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-u-shp0-g1] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-u-shp1-g0] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-u-shp1-g1] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-u-shp2-g0] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-u-shp2-g1] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-u-shp3-g0] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-u-shp3-g1] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-u-shp4-g0] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-u-shp4-g1] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-v-shp0-g0] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-v-shp0-g1] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-v-shp1-g0] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-v-shp1-g1] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-v-shp2-g0] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-v-shp2-g1] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-v-shp3-g0] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-v-shp3-g1] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-v-shp4-g0] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-v-shp4-g1] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-e-shp0-g0] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-e-shp0-g1] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-e-shp1-g0] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-e-shp1-g1] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-e-shp2-g0] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-e-shp2-g1] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-e-shp3-g0] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-e-shp3-g1] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-e-shp4-g0] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-e-shp4-g1] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-u-shp0-g0] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-u-shp0-g1] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-u-shp1-g0] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-u-shp1-g1] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-u-shp2-g0] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-u-shp2-g1] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-u-shp3-g0] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-u-shp3-g1] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-u-shp4-g0] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-u-shp4-g1] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-v-shp0-g0] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-v-shp0-g1] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-v-shp1-g0] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-v-shp1-g1] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-v-shp2-g0] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-v-shp2-g1] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-v-shp3-g0] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-v-shp3-g1] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-v-shp4-g0] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-v-shp4-g1] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-e-shp0-g0] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-e-shp0-g1] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-e-shp1-g0] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-e-shp1-g1] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-e-shp2-g0] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-e-shp2-g1] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-e-shp3-g0] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-e-shp3-g1] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-e-shp4-g0] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-e-shp4-g1] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-u-shp0-g0] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-u-shp0-g1] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-u-shp1-g0] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-u-shp1-g1] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-u-shp2-g0] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-u-shp2-g1] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-u-shp3-g0] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-u-shp3-g1] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-u-shp4-g0] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-u-shp4-g1] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-v-shp0-g0] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-v-shp0-g1] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-v-shp1-g0] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-v-shp1-g1] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-v-shp2-g0] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-v-shp2-g1] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-v-shp3-g0] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-v-shp3-g1] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-v-shp4-g0] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-v-shp4-g1] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-e-shp0-g0] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-e-shp0-g1] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-e-shp1-g0] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-e-shp1-g1] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-e-shp2-g0] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-e-shp2-g1] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-e-shp3-g0] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-e-shp3-g1] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-e-shp4-g0] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-e-shp4-g1] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-u-shp0-g0] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-u-shp0-g1] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-u-shp1-g0] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-u-shp1-g1] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-u-shp2-g0] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-u-shp2-g1] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-u-shp3-g0] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-u-shp3-g1] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-u-shp4-g0] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-u-shp4-g1] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-v-shp0-g0] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-v-shp0-g1] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-v-shp1-g0] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-v-shp1-g1] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-v-shp2-g0] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-v-shp2-g1] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-v-shp3-g0] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-v-shp3-g1] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-v-shp4-g0] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-v-shp4-g1] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-e-shp0-g0] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-e-shp0-g1] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-e-shp1-g0] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-e-shp1-g1] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-e-shp2-g0] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-e-shp2-g1] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-e-shp3-g0] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-e-shp3-g1] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-e-shp4-g0] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-e-shp4-g1] PASSED [ 91%]
tests/compute/test_sparse.py::test_edge_softmax[idtype0-shp0-src-g0] PASSED [ 91%]
tests/compute/test_sparse.py::test_edge_softmax[idtype0-shp0-dst-g0] PASSED [ 91%]
tests/compute/test_sparse.py::test_edge_softmax[idtype0-shp1-src-g0] PASSED [ 91%]
tests/compute/test_sparse.py::test_edge_softmax[idtype0-shp1-dst-g0] PASSED [ 91%]
tests/compute/test_sparse.py::test_edge_softmax[idtype0-shp2-src-g0] PASSED [ 91%]
tests/compute/test_sparse.py::test_edge_softmax[idtype0-shp2-dst-g0] PASSED [ 91%]
tests/compute/test_sparse.py::test_edge_softmax[idtype1-shp0-src-g0] PASSED [ 92%]
tests/compute/test_sparse.py::test_edge_softmax[idtype1-shp0-dst-g0] PASSED [ 92%]
tests/compute/test_sparse.py::test_edge_softmax[idtype1-shp1-src-g0] PASSED [ 92%]
tests/compute/test_sparse.py::test_edge_softmax[idtype1-shp1-dst-g0] PASSED [ 92%]
tests/compute/test_sparse.py::test_edge_softmax[idtype1-shp2-src-g0] PASSED [ 92%]
tests/compute/test_sparse.py::test_edge_softmax[idtype1-shp2-dst-g0] PASSED [ 92%]
tests/compute/test_sparse.py::test_segment_reduce[sum] PASSED            [ 92%]
tests/compute/test_sparse.py::test_segment_reduce[max] PASSED            [ 92%]
tests/compute/test_sparse.py::test_segment_reduce[min] PASSED            [ 92%]
tests/compute/test_sparse.py::test_segment_reduce[mean] PASSED           [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype0-0.01-1-idtype0] PASSED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype0-0.01-1-idtype1] PASSED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype0-0.01-8-idtype0] PASSED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype0-0.01-8-idtype1] PASSED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype0-0.01-16-idtype0] PASSED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype0-0.01-16-idtype1] PASSED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype0-0.01-64-idtype0] PASSED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype0-0.01-64-idtype1] PASSED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype0-0.01-256-idtype0] PASSED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype0-0.01-256-idtype1] PASSED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype1-0.01-1-idtype0] SKIPPED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype1-0.01-1-idtype1] SKIPPED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype1-0.01-8-idtype0] SKIPPED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype1-0.01-8-idtype1] SKIPPED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype1-0.01-16-idtype0] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype1-0.01-16-idtype1] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype1-0.01-64-idtype0] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype1-0.01-64-idtype1] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype1-0.01-256-idtype0] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype1-0.01-256-idtype1] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype2-0.003-1-idtype0] PASSED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype2-0.003-1-idtype1] PASSED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype2-0.003-8-idtype0] PASSED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype2-0.003-8-idtype1] PASSED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype2-0.003-16-idtype0] PASSED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype2-0.003-16-idtype1] PASSED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype2-0.003-64-idtype0] PASSED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype2-0.003-64-idtype1] PASSED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype2-0.003-256-idtype0] PASSED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype2-0.003-256-idtype1] PASSED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype3-0.0001-1-idtype0] PASSED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype3-0.0001-1-idtype1] PASSED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype3-0.0001-8-idtype0] PASSED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype3-0.0001-8-idtype1] PASSED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype3-0.0001-16-idtype0] PASSED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype3-0.0001-16-idtype1] PASSED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype3-0.0001-64-idtype0] PASSED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype3-0.0001-64-idtype1] PASSED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype3-0.0001-256-idtype0] PASSED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype3-0.0001-256-idtype1] PASSED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype0-0.01-1] PASSED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype0-0.01-8] PASSED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype0-0.01-16] PASSED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype0-0.01-64] PASSED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype0-0.01-256] PASSED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype1-0.02-1] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype1-0.02-8] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype1-0.02-16] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype1-0.02-64] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype1-0.02-256] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype2-0.003-1] PASSED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype2-0.003-8] PASSED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype2-0.003-16] PASSED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype2-0.003-64] PASSED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype2-0.003-256] PASSED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype3-0.0001-1] PASSED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype3-0.0001-8] PASSED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype3-0.0001-16] PASSED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype3-0.0001-64] PASSED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype3-0.0001-256] PASSED [ 94%]
tests/compute/test_sparse.py::test_use_libxsmm_switch SKIPPED (Libxs...) [ 94%]
tests/compute/test_specialization.py::test_v2v_update_all[idtype0] PASSED [ 94%]
tests/compute/test_specialization.py::test_v2v_update_all[idtype1] PASSED [ 94%]
tests/compute/test_specialization.py::test_v2v_snr[idtype0] PASSED       [ 94%]
tests/compute/test_specialization.py::test_v2v_snr[idtype1] PASSED       [ 95%]
tests/compute/test_specialization.py::test_v2v_pull[idtype0] PASSED      [ 95%]
tests/compute/test_specialization.py::test_v2v_pull[idtype1] PASSED      [ 95%]
tests/compute/test_specialization.py::test_update_all_multi_fallback[idtype0] PASSED [ 95%]
tests/compute/test_specialization.py::test_update_all_multi_fallback[idtype1] PASSED [ 95%]
tests/compute/test_specialization.py::test_pull_multi_fallback[idtype0] PASSED [ 95%]
tests/compute/test_specialization.py::test_pull_multi_fallback[idtype1] PASSED [ 95%]
tests/compute/test_specialization.py::test_spmv_3d_feat[idtype0] PASSED  [ 95%]
tests/compute/test_specialization.py::test_spmv_3d_feat[idtype1] PASSED  [ 95%]
tests/compute/test_subgraph.py::test_edge_subgraph PASSED                [ 95%]
tests/compute/test_subgraph.py::test_subgraph PASSED                     [ 95%]
tests/compute/test_subgraph.py::test_subgraph_mask[idtype0] PASSED       [ 95%]
tests/compute/test_subgraph.py::test_subgraph_mask[idtype1] PASSED       [ 95%]
tests/compute/test_subgraph.py::test_subgraph1[idtype0] PASSED           [ 95%]
tests/compute/test_subgraph.py::test_subgraph1[idtype1] PASSED           [ 95%]
tests/compute/test_subgraph.py::test_in_subgraph[idtype0] PASSED         [ 95%]
tests/compute/test_subgraph.py::test_in_subgraph[idtype1] PASSED         [ 95%]
tests/compute/test_subgraph.py::test_out_subgraph[idtype0] PASSED        [ 95%]
tests/compute/test_subgraph.py::test_out_subgraph[idtype1] PASSED        [ 95%]
tests/compute/test_subgraph.py::test_subgraph_message_passing PASSED     [ 95%]
tests/compute/test_subgraph.py::test_khop_in_subgraph[idtype0] PASSED    [ 95%]
tests/compute/test_subgraph.py::test_khop_in_subgraph[idtype1] PASSED    [ 95%]
tests/compute/test_subgraph.py::test_khop_out_subgraph[idtype0] PASSED   [ 95%]
tests/compute/test_subgraph.py::test_khop_out_subgraph[idtype1] PASSED   [ 95%]
tests/compute/test_subgraph.py::test_subframes[child_device0-parent_idx_device0] PASSED [ 96%]
tests/compute/test_subgraph.py::test_subframes[child_device0-parent_idx_device1] PASSED [ 96%]
tests/compute/test_subgraph.py::test_subframes[child_device0-parent_idx_device2] PASSED [ 96%]
tests/compute/test_subgraph.py::test_subframes[child_device0-parent_idx_device3] PASSED [ 96%]
tests/compute/test_subgraph.py::test_subframes[child_device1-parent_idx_device0] PASSED [ 96%]
tests/compute/test_subgraph.py::test_subframes[child_device1-parent_idx_device1] PASSED [ 96%]
tests/compute/test_subgraph.py::test_subframes[child_device1-parent_idx_device2] PASSED [ 96%]
tests/compute/test_subgraph.py::test_subframes[child_device1-parent_idx_device3] PASSED [ 96%]
tests/compute/test_subgraph.py::test_uva_subgraph[idtype0-device0] PASSED [ 96%]
tests/compute/test_subgraph.py::test_uva_subgraph[idtype0-device1] PASSED [ 96%]
tests/compute/test_subgraph.py::test_uva_subgraph[idtype1-device0] PASSED [ 96%]
tests/compute/test_subgraph.py::test_uva_subgraph[idtype1-device1] PASSED [ 96%]
tests/compute/test_transform.py::test_line_graph1 PASSED                 [ 96%]
tests/compute/test_transform.py::test_line_graph2[idtype0] PASSED        [ 96%]
tests/compute/test_transform.py::test_line_graph2[idtype1] PASSED        [ 96%]
tests/compute/test_transform.py::test_no_backtracking PASSED             [ 96%]
tests/compute/test_transform.py::test_reverse[idtype0] PASSED            [ 96%]
tests/compute/test_transform.py::test_reverse[idtype1] PASSED            [ 96%]
tests/compute/test_transform.py::test_reverse_shared_frames[idtype0] PASSED [ 96%]
tests/compute/test_transform.py::test_reverse_shared_frames[idtype1] PASSED [ 96%]
tests/compute/test_transform.py::test_to_bidirected SKIPPED (GPU not...) [ 96%]
tests/compute/test_transform.py::test_add_reverse_edges PASSED           [ 96%]
tests/compute/test_transform.py::test_simple_graph SKIPPED (GPU not ...) [ 96%]
tests/compute/test_transform.py::test_khop_graph SKIPPED (GPU not im...) [ 96%]
tests/compute/test_transform.py::test_khop_adj SKIPPED (GPU not impl...) [ 96%]
tests/compute/test_transform.py::test_laplacian_lambda_max SKIPPED (...) [ 97%]
tests/compute/test_transform.py::test_partition_with_halo SKIPPED (G...) [ 97%]
tests/compute/test_transform.py::test_metis_partition[idtype0] SKIPPED   [ 97%]
tests/compute/test_transform.py::test_metis_partition[idtype1] SKIPPED   [ 97%]
tests/compute/test_transform.py::test_reorder_nodes SKIPPED (It does...) [ 97%]
tests/compute/test_transform.py::test_compact[idtype0] PASSED            [ 97%]
tests/compute/test_transform.py::test_compact[idtype1] PASSED            [ 97%]
tests/compute/test_transform.py::test_to_simple[idtype0] SKIPPED (GP...) [ 97%]
tests/compute/test_transform.py::test_to_simple[idtype1] SKIPPED (GP...) [ 97%]
tests/compute/test_transform.py::test_to_block[idtype0] PASSED           [ 97%]
tests/compute/test_transform.py::test_to_block[idtype1] PASSED           [ 97%]
tests/compute/test_transform.py::test_remove_edges[idtype0] PASSED       [ 97%]
tests/compute/test_transform.py::test_remove_edges[idtype1] PASSED       [ 97%]
tests/compute/test_transform.py::test_add_edges[idtype0] PASSED          [ 97%]
tests/compute/test_transform.py::test_add_edges[idtype1] PASSED          [ 97%]
tests/compute/test_transform.py::test_add_nodes[idtype0] PASSED          [ 97%]
tests/compute/test_transform.py::test_add_nodes[idtype1] PASSED          [ 97%]
tests/compute/test_transform.py::test_remove_nodes[idtype0] PASSED       [ 97%]
tests/compute/test_transform.py::test_remove_nodes[idtype1] PASSED       [ 97%]
tests/compute/test_transform.py::test_add_selfloop[idtype0] PASSED       [ 97%]
tests/compute/test_transform.py::test_add_selfloop[idtype1] PASSED       [ 97%]
tests/compute/test_transform.py::test_remove_selfloop[idtype0] PASSED    [ 97%]
tests/compute/test_transform.py::test_remove_selfloop[idtype1] PASSED    [ 97%]
tests/compute/test_transform.py::test_reorder_graph[idtype0] PASSED      [ 97%]
tests/compute/test_transform.py::test_reorder_graph[idtype1] PASSED      [ 98%]
tests/compute/test_transform.py::test_norm_by_dst[idtype0] PASSED        [ 98%]
tests/compute/test_transform.py::test_norm_by_dst[idtype1] PASSED        [ 98%]
tests/compute/test_transform.py::test_module_add_self_loop[idtype0] PASSED [ 98%]
tests/compute/test_transform.py::test_module_add_self_loop[idtype1] PASSED [ 98%]
tests/compute/test_transform.py::test_module_remove_self_loop[idtype0] PASSED [ 98%]
tests/compute/test_transform.py::test_module_remove_self_loop[idtype1] PASSED [ 98%]
tests/compute/test_transform.py::test_module_add_reverse[idtype0] PASSED [ 98%]
tests/compute/test_transform.py::test_module_add_reverse[idtype1] PASSED [ 98%]
tests/compute/test_transform.py::test_module_to_simple[idtype0] SKIPPED  [ 98%]
tests/compute/test_transform.py::test_module_to_simple[idtype1] SKIPPED  [ 98%]
tests/compute/test_transform.py::test_module_line_graph[idtype0] PASSED  [ 98%]
tests/compute/test_transform.py::test_module_line_graph[idtype1] PASSED  [ 98%]
tests/compute/test_transform.py::test_module_khop_graph[idtype0] PASSED  [ 98%]
tests/compute/test_transform.py::test_module_khop_graph[idtype1] PASSED  [ 98%]
tests/compute/test_transform.py::test_module_add_metapaths[idtype0] PASSED [ 98%]
tests/compute/test_transform.py::test_module_add_metapaths[idtype1] PASSED [ 98%]
tests/compute/test_transform.py::test_module_compose[idtype0] PASSED     [ 98%]
tests/compute/test_transform.py::test_module_compose[idtype1] PASSED     [ 98%]
tests/compute/test_transform.py::test_module_gcnnorm[idtype0] PASSED     [ 98%]
tests/compute/test_transform.py::test_module_gcnnorm[idtype1] PASSED     [ 98%]
tests/compute/test_transform.py::test_module_ppr[idtype0] PASSED         [ 98%]
tests/compute/test_transform.py::test_module_ppr[idtype1] PASSED         [ 98%]
tests/compute/test_transform.py::test_module_heat_kernel[idtype0] PASSED [ 98%]
tests/compute/test_transform.py::test_module_heat_kernel[idtype1] PASSED [ 98%]
tests/compute/test_transform.py::test_module_gdc[idtype0] PASSED         [ 99%]
tests/compute/test_transform.py::test_module_gdc[idtype1] PASSED         [ 99%]
tests/compute/test_transform.py::test_module_node_shuffle[idtype0] PASSED [ 99%]
tests/compute/test_transform.py::test_module_node_shuffle[idtype1] PASSED [ 99%]
tests/compute/test_transform.py::test_module_drop_node[idtype0] PASSED   [ 99%]
tests/compute/test_transform.py::test_module_drop_node[idtype1] PASSED   [ 99%]
tests/compute/test_transform.py::test_module_drop_edge[idtype0] PASSED   [ 99%]
tests/compute/test_transform.py::test_module_drop_edge[idtype1] PASSED   [ 99%]
tests/compute/test_transform.py::test_module_add_edge[idtype0] PASSED    [ 99%]
tests/compute/test_transform.py::test_module_add_edge[idtype1] PASSED    [ 99%]
tests/compute/test_transform.py::test_module_random_walk_pe[idtype0] PASSED [ 99%]
tests/compute/test_transform.py::test_module_random_walk_pe[idtype1] PASSED [ 99%]
tests/compute/test_transform.py::test_module_laplacian_pe[idtype0] PASSED [ 99%]
tests/compute/test_transform.py::test_module_laplacian_pe[idtype1] PASSED [ 99%]
tests/compute/test_transform.py::test_module_sign[g0] PASSED             [ 99%]
tests/compute/test_transform.py::test_module_row_feat_normalizer[idtype0] PASSED [ 99%]
tests/compute/test_transform.py::test_module_row_feat_normalizer[idtype1] PASSED [ 99%]
tests/compute/test_transform.py::test_module_feat_mask[idtype0] PASSED   [ 99%]
tests/compute/test_transform.py::test_module_feat_mask[idtype1] PASSED   [ 99%]
tests/compute/test_traversal.py::test_bfs[idtype0] PASSED                [ 99%]
tests/compute/test_traversal.py::test_bfs[idtype1] PASSED                [ 99%]
tests/compute/test_traversal.py::test_topological_nodes[idtype0] PASSED  [ 99%]
tests/compute/test_traversal.py::test_topological_nodes[idtype1] PASSED  [ 99%]
tests/compute/test_traversal.py::test_dfs_labeled_edges[idtype0] PASSED  [ 99%]
tests/compute/test_traversal.py::test_dfs_labeled_edges[idtype1] PASSED  [100%]

=============================== warnings summary ===============================
python/dgl/backend/backend.py:1741
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/backend/backend.py:1741: DeprecationWarning: invalid escape sequence \P
    """

python/dgl/backend/pytorch/tensor.py:16
python/dgl/backend/pytorch/tensor.py:16
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/backend/pytorch/tensor.py:16: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    if LooseVersion(th.__version__) < LooseVersion("1.9.0"):

python/dgl/backend/pytorch/tensor.py:340
python/dgl/backend/pytorch/tensor.py:340
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/backend/pytorch/tensor.py:340: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    if LooseVersion(th.__version__) >= LooseVersion("1.10.0"):

python/dgl/dataloading/dataloader.py:33
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/dataloading/dataloader.py:33: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    PYTORCH_VER = LooseVersion(torch.__version__)

python/dgl/_dataloading/pytorch/dataloader.py:23
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/_dataloading/pytorch/dataloader.py:23: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    PYTORCH_VER = LooseVersion(th.__version__)

python/dgl/_dataloading/pytorch/dataloader.py:24
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/_dataloading/pytorch/dataloader.py:24: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    PYTORCH_16 = PYTORCH_VER >= LooseVersion("1.6.0")

python/dgl/_dataloading/pytorch/dataloader.py:25
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/_dataloading/pytorch/dataloader.py:25: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    PYTORCH_17 = PYTORCH_VER >= LooseVersion("1.7.0")

../../../../opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:10
  /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:10: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    _nlv = LooseVersion(_np_version)

../../../../opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:11
  /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:11: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    _np_version_under1p16 = _nlv < LooseVersion("1.16")

../../../../opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:12
  /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:12: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    _np_version_under1p17 = _nlv < LooseVersion("1.17")

../../../../opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:13
  /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:13: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    _np_version_under1p18 = _nlv < LooseVersion("1.18")

../../../../opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:14
  /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:14: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    _np_version_under1p19 = _nlv < LooseVersion("1.19")

../../../../opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:15
  /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:15: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    _np_version_under1p20 = _nlv < LooseVersion("1.20")

../../../../opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/setuptools/_distutils/version.py:351
  /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/setuptools/_distutils/version.py:351: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    other = LooseVersion(other)

../../../../opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/pandas/compat/numpy/function.py:125
../../../../opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/pandas/compat/numpy/function.py:125
  /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/pandas/compat/numpy/function.py:125: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    if LooseVersion(_np_version) >= LooseVersion("1.17.0"):

tests/compute/test_basics.py: 2 warnings
tests/compute/test_filter.py: 1 warning
tests/compute/test_graph.py: 9 warnings
tests/compute/test_kernel.py: 3 warnings
tests/compute/test_removal.py: 16 warnings
tests/compute/test_specialization.py: 12 warnings
tests/compute/test_subgraph.py: 2 warnings
tests/compute/test_transform.py: 6 warnings
tests/compute/test_traversal.py: 2 warnings
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/heterograph.py:72: DGLWarning: Recommend creating graphs by `dgl.graph(data)` instead of `dgl.DGLGraph(data)`.
    dgl_warning('Recommend creating graphs by `dgl.graph(data)`'

tests/compute/test_basics.py: 2 warnings
tests/compute/test_batched_graph.py: 10 warnings
tests/compute/test_graph.py: 2 warnings
tests/compute/test_kernel.py: 1 warning
tests/compute/test_removal.py: 10 warnings
tests/compute/test_specialization.py: 10 warnings
tests/compute/test_subgraph.py: 2 warnings
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/heterograph.py:354: DGLWarning: DGLGraph.add_edge is deprecated. Please use DGLGraph.add_edges
    dgl_warning("DGLGraph.add_edge is deprecated. Please use DGLGraph.add_edges")

tests/compute/test_basics.py::test_update_all_0deg[idtype0]
tests/compute/test_basics.py::test_update_all_0deg[idtype1]
tests/compute/test_basics.py::test_pull_0deg[idtype0]
tests/compute/test_basics.py::test_pull_0deg[idtype1]
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/core.py:79: DGLWarning: The input graph for the user-defined edge function does not contain valid edges
    dgl_warning('The input graph for the user-defined edge function ' \

tests/compute/test_batched_graph.py::test_batched_edge_ordering[idtype0]
tests/compute/test_batched_graph.py::test_batched_edge_ordering[idtype1]
tests/compute/test_graph.py::test_query
tests/compute/test_graph.py::test_hypersparse_query
tests/compute/test_transform.py::test_no_backtracking
tests/compute/test_transform.py::test_reverse[idtype0]
tests/compute/test_transform.py::test_reverse[idtype1]
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/heterograph.py:2978: DGLWarning: DGLGraph.edge_id is deprecated. Please use DGLGraph.edge_ids.
    dgl_warning("DGLGraph.edge_id is deprecated. Please use DGLGraph.edge_ids.")

tests/compute/test_batched_heterograph.py::test_features[idtype0]
tests/compute/test_batched_heterograph.py::test_features[idtype1]
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/batch.py:159: DGLWarning: Arguments edge_attrs has been deprecated. Please use edata instead.
    dgl_warning('Arguments edge_attrs has been deprecated. Please use'

tests/compute/test_csrmm.py::test_csrmm[dtype0-idtype0]
tests/compute/test_csrmm.py::test_csrmm[dtype0-idtype1]
tests/compute/test_csrmm.py::test_csrmm[dtype1-idtype0]
tests/compute/test_csrmm.py::test_csrmm[dtype1-idtype1]
tests/compute/test_csrmm.py::test_csrsum[dtype0-idtype0]
tests/compute/test_csrmm.py::test_csrsum[dtype0-idtype1]
tests/compute/test_csrmm.py::test_csrsum[dtype1-idtype0]
tests/compute/test_csrmm.py::test_csrsum[dtype1-idtype1]
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/heterograph_index.py:797: FutureWarning: Adjacency matrix by default currently returns edge IDs.  As a result there is one 0 entry which is not eliminated.  In the next release it will return 1s by default, and 0 will be eliminated otherwise.
    FutureWarning)

tests/compute/test_graph.py::test_query
tests/compute/test_graph.py::test_hypersparse_query
tests/compute/test_heterograph.py::test_query[idtype0]
tests/compute/test_heterograph.py::test_query[idtype1]
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/heterograph.py:2753: DGLWarning: DGLGraph.has_node is deprecated. Please use DGLGraph.has_nodes
    dgl_warning("DGLGraph.has_node is deprecated. Please use DGLGraph.has_nodes")

tests/compute/test_graph.py::test_query
tests/compute/test_graph.py::test_hypersparse_query
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/heterograph.py:2687: DGLWarning: DGLGraph.__contains__ is deprecated. Please directly call has_nodes.
    dgl_warning('DGLGraph.__contains__ is deprecated.'

tests/compute/test_graph.py::test_query
tests/compute/test_sampling.py::test_non_uniform_random_walk[True]
tests/compute/test_sampling.py::test_non_uniform_random_walk[False]
tests/compute/test_sampling.py::test_uniform_random_walk[True]
tests/compute/test_sampling.py::test_uniform_random_walk[False]
tests/compute/test_transform.py::test_no_backtracking
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/heterograph.py:2851: DGLWarning: DGLGraph.has_edge_between is deprecated. Please use DGLGraph.has_edges_between
    dgl_warning("DGLGraph.has_edge_between is deprecated. "

tests/compute/test_graph.py::test_query
tests/compute/test_graph.py::test_hypersparse_query
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/heterograph.py:3432: DGLWarning: DGLGraph.in_degree is deprecated. Please use DGLGraph.in_degrees
    dgl_warning("DGLGraph.in_degree is deprecated. Please use DGLGraph.in_degrees")

tests/compute/test_graph.py::test_query
tests/compute/test_graph.py::test_hypersparse_query
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/heterograph.py:3516: DGLWarning: DGLGraph.out_degree is deprecated. Please use DGLGraph.out_degrees
    dgl_warning("DGLGraph.out_degree is deprecated. Please use DGLGraph.out_degrees")

tests/compute/test_graph.py::test_query
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/heterograph.py:84: DGLWarning: Keyword arguments ['readonly', 'sort_csr'] are deprecated in v0.5, and can be safely removed in all cases.
    ' removed in all cases.'.format(list(deprecate_kwargs.keys())))

tests/compute/test_heterograph.py: 20 warnings
  /root/jenkins/workspace/dgl_PR-4648/tests/compute/test_heterograph.py:1128: DeprecationWarning: np.asscalar(a) is deprecated since NumPy v1.16, use a.item() instead
    assert np.asscalar(F.asnumpy(src_i)) == nid[src[i]]

tests/compute/test_heterograph.py: 20 warnings
  /root/jenkins/workspace/dgl_PR-4648/tests/compute/test_heterograph.py:1129: DeprecationWarning: np.asscalar(a) is deprecated since NumPy v1.16, use a.item() instead
    assert np.asscalar(F.asnumpy(dst_i)) == nid[dst[i]]

tests/compute/test_heterograph.py::test_invertible_conversion[idtype0]
tests/compute/test_heterograph.py::test_invertible_conversion[idtype1]
tests/compute/test_shared_mem.py::test_single_process[idtype0]
tests/compute/test_shared_mem.py::test_single_process[idtype1]
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/heterograph.py:2635: DGLWarning: DGLGraph.is_readonly is deprecated in v0.5.
  DGLGraph now always supports mutable operations like add_nodes and add_edges.
    dgl_warning('DGLGraph.is_readonly is deprecated in v0.5.\n'

tests/compute/test_heterograph.py::test_types_in_function[idtype0]
tests/compute/test_heterograph.py::test_types_in_function[idtype1]
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/backend/pytorch/tensor.py:277: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).
    mask = th.tensor(mask, dtype=th.bool)

tests/compute/test_new_update_all_hetero.py::test_binary_op[idtype0]
tests/compute/test_new_update_all_hetero.py::test_binary_op[idtype1]
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/backend/pytorch/tensor.py:569: UserWarning: The .grad attribute of a Tensor that is not a leaf Tensor is being accessed. Its .grad attribute won't be populated during autograd.backward(). If you indeed want the gradient for a non-leaf Tensor, use .retain_grad() on the non-leaf Tensor. If you access the non-leaf Tensor by mistake, make sure you access the leaf Tensor instead. See github.com/pytorch/pytorch/pull/30531 for more information.
    return x.grad

tests/compute/test_partition.py::test_get_node_partition_from_book[idtype0]
tests/compute/test_partition.py::test_get_node_partition_from_book[idtype1]
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
    "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")

tests/compute/test_sampler.py::test_create_full
tests/compute/test_sampler.py::test_1neighbor_sampler_all
tests/compute/test_sampler.py::test_1neighbor_sampler
tests/compute/test_sampler.py::test_prefetch_neighbor_sampler
tests/compute/test_sampler.py::test_10neighbor_sampler_all
tests/compute/test_sampler.py::test_10neighbor_sampler
tests/compute/test_sampler.py::test_nonuniform_neighbor_sampler
tests/compute/test_sampler.py::test_setseed
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/contrib/sampling/sampler.py:317: DGLWarning: dgl.contrib.sampling.NeighborSampler is deprecated starting from v0.5. Please read our guide<link> for how to use the new sampling APIs.
    dgl_warning('dgl.contrib.sampling.NeighborSampler is deprecated starting from v0.5.'

tests/compute/test_sampler.py::test_create_full
tests/compute/test_sampler.py::test_1neighbor_sampler_all
tests/compute/test_sampler.py::test_1neighbor_sampler
tests/compute/test_sampler.py::test_prefetch_neighbor_sampler
tests/compute/test_sampler.py::test_10neighbor_sampler_all
tests/compute/test_sampler.py::test_10neighbor_sampler
tests/compute/test_sampler.py::test_layer_sampler
tests/compute/test_sampler.py::test_nonuniform_neighbor_sampler
tests/compute/test_sampler.py::test_setseed
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/_deprecate/nodeflow.py:99: DGLWarning: NodeFlow APIs are deprecated starting from v0.5. Please read our guide<link> for how to use the new sampling APIs.
    dgl_warning('NodeFlow APIs are deprecated starting from v0.5. Please read our'

tests/compute/test_sampling.py::test_global_uniform_negative_sampling[int32]
tests/compute/test_sampling.py::test_global_uniform_negative_sampling[int64]
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/sampling/negative.py:102: ComplexWarning: Casting complex values to real discards the imaginary part
    g._graph, etype_id, num_samples, 3, exclude_self_loops, replace, redundancy)

tests/compute/test_serialize.py::test_load_old_files1
tests/compute/test_serialize.py::test_load_old_files2
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/data/graph_serialize.py:179: DGLWarning: You are loading a graph file saved by old version of dgl.              Please consider saving it again with the current format.
    Please consider saving it again with the current format.")

tests/compute/test_sparse.py: 80 warnings
  /root/jenkins/workspace/dgl_PR-4648/tests/compute/test_sparse.py:299: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    if LooseVersion(torch.version.cuda) < LooseVersion("11.0") \

tests/compute/test_sparse.py: 40 warnings
  /root/jenkins/workspace/dgl_PR-4648/tests/compute/test_sparse.py:339: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    if LooseVersion(torch.version.cuda) < LooseVersion("11.0") \

tests/compute/test_transform.py::test_reverse_shared_frames[idtype0]
tests/compute/test_transform.py::test_reverse_shared_frames[idtype1]
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/transforms/functional.py:1267: DGLWarning: share_ndata argument has been renamed to copy_ndata.
    dgl_warning('share_ndata argument has been renamed to copy_ndata.')

tests/compute/test_transform.py::test_reverse_shared_frames[idtype0]
tests/compute/test_transform.py::test_reverse_shared_frames[idtype1]
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/transforms/functional.py:1270: DGLWarning: share_edata argument has been renamed to copy_edata.
    dgl_warning('share_edata argument has been renamed to copy_edata.')

tests/compute/test_transform.py::test_module_random_walk_pe[idtype0]
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/backend/pytorch/tensor.py:44: UserWarning: The given NumPy array is not writeable, and PyTorch does not support non-writeable tensors. This means you can write to the underlying (supposedly non-writeable) NumPy array using the tensor. You may want to copy the array to protect its data or make it writeable before converting it to a tensor. This type of warning will be suppressed for the rest of this program. (Triggered internally at  /pytorch/torch/csrc/utils/tensor_numpy.cpp:180.)
    return th.as_tensor(data, dtype=dtype)

-- Docs: https://docs.pytest.org/en/stable/how-to/capture-warnings.html
-- generated xml file: /root/jenkins/workspace/dgl_PR-4648/pytest_compute.xml --
============================ slowest 100 durations =============================
23.31s call     tests/compute/test_data.py::test_as_graphpred_ogb
11.00s call     tests/compute/test_data.py::test_as_linkpred_ogb
6.86s call     tests/compute/test_data.py::test_as_nodepred_ogb
6.60s call     tests/compute/test_sampler.py::test_negative_sampler
3.20s call     tests/compute/test_kernel.py::test_all_binary_builtins
1.28s call     tests/compute/test_heterograph.py::test_forking_pickler
0.67s call     tests/compute/test_apply_edges_hetero.py::test_unary_copy_u[idtype0]
0.38s call     tests/compute/test_apply_edges_hetero.py::test_binary_op[idtype0]
0.28s call     tests/compute/test_sampling.py::test_non_uniform_random_walk[False]
0.26s call     tests/compute/test_graph.py::test_query
0.19s call     tests/compute/test_heterograph.py::test_query[idtype0]
0.18s call     tests/compute/test_heterograph.py::test_query[idtype1]
0.18s call     tests/compute/test_sampling.py::test_sample_neighbors_outedge
0.18s call     tests/compute/test_sampling.py::test_uniform_random_walk[False]
0.17s call     tests/compute/test_sampling.py::test_non_uniform_random_walk[True]
0.15s call     tests/compute/test_apply_edges_hetero.py::test_binary_op[idtype1]
0.15s call     tests/compute/test_new_update_all_hetero.py::test_binary_op[idtype0]
0.15s call     tests/compute/test_new_update_all_hetero.py::test_binary_op[idtype1]
0.12s call     tests/compute/test_sampling.py::test_uniform_random_walk[True]
0.12s call     tests/compute/test_kernel.py::test_copy_edge_reduce
0.12s call     tests/compute/test_shared_mem.py::test_multi_process[idtype0]
0.11s call     tests/compute/test_shared_mem.py::test_copy_from_gpu
0.11s call     tests/compute/test_kernel.py::test_copy_src_reduce
0.11s call     tests/compute/test_shared_mem.py::test_multi_process[idtype1]
0.10s call     tests/compute/test_sampling.py::test_sample_neighbors_prob
0.10s call     tests/compute/test_sampling.py::test_sample_neighbors_noprob
0.08s call     tests/compute/test_heterograph.py::test_view1[idtype0]
0.08s call     tests/compute/test_heterograph.py::test_view1[idtype1]
0.06s call     tests/compute/test_heterograph.py::test_updates[idtype0]
0.06s call     tests/compute/test_heterograph.py::test_updates[idtype1]
0.06s call     tests/compute/test_sampler.py::test_prefetch_neighbor_sampler
0.06s call     tests/compute/test_nccl.py::test_nccl_sparse_push_single_remainder
0.05s call     tests/compute/test_traversal.py::test_bfs[idtype0]
0.05s call     tests/compute/test_removal.py::test_node_and_edge_removal[idtype0]
0.05s call     tests/compute/test_removal.py::test_node_and_edge_removal[idtype1]
0.05s call     tests/compute/test_sampler.py::test_1neighbor_sampler
0.05s call     tests/compute/test_heterograph.py::test_level2[idtype0]
0.04s call     tests/compute/test_traversal.py::test_bfs[idtype1]
0.04s call     tests/compute/test_heterograph.py::test_level2[idtype1]
0.04s call     tests/compute/test_transform.py::test_remove_nodes[idtype0]
0.04s call     tests/compute/test_specialization.py::test_v2v_snr[idtype0]
0.04s call     tests/compute/test_transform.py::test_remove_nodes[idtype1]
0.04s call     tests/compute/test_specialization.py::test_v2v_snr[idtype1]
0.04s call     tests/compute/test_sampler.py::test_10neighbor_sampler
0.04s call     tests/compute/test_nccl.py::test_nccl_sparse_pull_single_remainder
0.04s call     tests/compute/test_specialization.py::test_v2v_pull[idtype0]
0.03s call     tests/compute/test_sampler.py::test_1neighbor_sampler_all
0.03s call     tests/compute/test_nccl.py::test_nccl_sparse_pull_single_range
0.03s call     tests/compute/test_nccl.py::test_nccl_sparse_push_single_range
0.03s call     tests/compute/test_specialization.py::test_v2v_pull[idtype1]
0.03s call     tests/compute/test_transform.py::test_remove_edges[idtype0]
0.03s call     tests/compute/test_specialization.py::test_pull_multi_fallback[idtype0]
0.03s call     tests/compute/test_transform.py::test_remove_edges[idtype1]
0.03s call     tests/compute/test_specialization.py::test_pull_multi_fallback[idtype1]
0.03s call     tests/compute/test_new_update_all_hetero.py::test_unary_copy_u[idtype0]
0.03s call     tests/compute/test_new_update_all_hetero.py::test_unary_copy_e[idtype0]
0.03s call     tests/compute/test_heterograph.py::test_flatten[idtype0]
0.03s call     tests/compute/test_new_update_all_hetero.py::test_unary_copy_e[idtype1]
0.03s call     tests/compute/test_heterograph.py::test_format[idtype0]
0.03s call     tests/compute/test_specialization.py::test_v2v_update_all[idtype0]
0.03s call     tests/compute/test_sampler.py::test_nonuniform_neighbor_sampler
0.03s call     tests/compute/test_heterograph.py::test_format[idtype1]
0.03s call     tests/compute/test_new_update_all_hetero.py::test_unary_copy_u[idtype1]
0.03s call     tests/compute/test_serialize.py::test_load_old_files1
0.02s call     tests/compute/test_heterograph.py::test_flatten[idtype1]
0.02s call     tests/compute/test_sampler.py::test_setseed
0.02s call     tests/compute/test_specialization.py::test_v2v_update_all[idtype1]
0.02s call     tests/compute/test_heterograph.py::test_convert[idtype0]
0.02s call     tests/compute/test_heterograph.py::test_convert[idtype1]
0.02s call     tests/compute/test_transform.py::test_to_block[idtype0]
0.02s call     tests/compute/test_subgraph.py::test_subgraph1[idtype0]
0.02s call     tests/compute/test_subgraph.py::test_subgraph1[idtype1]
0.02s call     tests/compute/test_transform.py::test_to_block[idtype1]
0.02s call     tests/compute/test_sampler.py::test_layer_sampler
0.02s call     tests/compute/test_transform.py::test_reorder_graph[idtype0]
0.02s call     tests/compute/test_specialization.py::test_spmv_3d_feat[idtype1]
0.02s call     tests/compute/test_specialization.py::test_spmv_3d_feat[idtype0]
0.02s call     tests/compute/test_transform.py::test_add_edges[idtype0]
0.02s call     tests/compute/test_transform.py::test_add_edges[idtype1]
0.02s call     tests/compute/test_transform.py::test_reorder_graph[idtype1]
0.02s call     tests/compute/test_heterograph.py::test_types_in_function[idtype0]
0.02s call     tests/compute/test_heterograph.py::test_types_in_function[idtype1]
0.02s call     tests/compute/test_sampling.py::test_pinsage_sampling[True]
0.02s call     tests/compute/test_transform.py::test_module_sign[g0]
0.02s call     tests/compute/test_transform.py::test_add_reverse_edges
0.02s call     tests/compute/test_csrmm.py::test_csrmm[dtype0-idtype0]
0.02s call     tests/compute/test_heterograph.py::test_add_edges[idtype0]
0.02s call     tests/compute/test_subgraph.py::test_khop_out_subgraph[idtype0]
0.02s call     tests/compute/test_heterograph.py::test_remove_nodes[idtype0]
0.02s call     tests/compute/test_heterograph.py::test_float_cast
0.02s call     tests/compute/test_heterograph.py::test_add_edges[idtype1]
0.01s call     tests/compute/test_subgraph.py::test_khop_out_subgraph[idtype1]
0.01s call     tests/compute/test_heterograph.py::test_remove_nodes[idtype1]
0.01s call     tests/compute/test_specialization.py::test_update_all_multi_fallback[idtype0]
0.01s call     tests/compute/test_heterograph.py::test_remove_edges[idtype0]
0.01s call     tests/compute/test_heterograph.py::test_remove_edges[idtype1]
0.01s call     tests/compute/test_csrmm.py::test_csrsum[dtype0-idtype1]
0.01s call     tests/compute/test_transform.py::test_add_selfloop[idtype0]
0.01s call     tests/compute/test_specialization.py::test_update_all_multi_fallback[idtype1]
0.01s call     tests/compute/test_readout.py::test_topk[True-g0-idtype0]
========== 2334 passed, 124 skipped, 342 warnings in 77.18s (0:01:17) ==========
============================= test session starts ==============================
platform linux -- Python 3.7.0, pytest-7.1.2, pluggy-1.0.0 -- /opt/conda/envs/pytorch-ci/bin/python3
cachedir: .pytest_cache
rootdir: /root/jenkins/workspace/dgl_PR-4648
collecting ... collected 3978 items

tests/pytorch/test_dataloader.py::test_graph_dataloader PASSED           [  0%]
tests/pytorch/test_dataloader.py::test_cluster_gcn[0] PASSED             [  0%]
tests/pytorch/test_dataloader.py::test_cluster_gcn[4] PASSED             [  0%]
tests/pytorch/test_dataloader.py::test_shadow[0] PASSED                  [  0%]
tests/pytorch/test_dataloader.py::test_shadow[4] PASSED                  [  0%]
tests/pytorch/test_dataloader.py::test_saint[node-0] PASSED              [  0%]
tests/pytorch/test_dataloader.py::test_saint[node-4] PASSED              [  0%]
tests/pytorch/test_dataloader.py::test_saint[edge-0] PASSED              [  0%]
tests/pytorch/test_dataloader.py::test_saint[edge-4] PASSED              [  0%]
tests/pytorch/test_dataloader.py::test_saint[walk-0] PASSED              [  0%]
tests/pytorch/test_dataloader.py::test_saint[walk-4] PASSED              [  0%]
tests/pytorch/test_dataloader.py::test_neighbor_nonuniform[False-cpu-idtype0] PASSED [  0%]
tests/pytorch/test_dataloader.py::test_neighbor_nonuniform[False-cpu-idtype1] PASSED [  0%]
tests/pytorch/test_dataloader.py::test_neighbor_nonuniform[False-uva_cuda_indices-idtype0] PASSED [  0%]
tests/pytorch/test_dataloader.py::test_neighbor_nonuniform[False-uva_cuda_indices-idtype1] PASSED [  0%]
tests/pytorch/test_dataloader.py::test_neighbor_nonuniform[False-uva_cpu_indices-idtype0] PASSED [  0%]
tests/pytorch/test_dataloader.py::test_neighbor_nonuniform[False-uva_cpu_indices-idtype1] PASSED [  0%]
tests/pytorch/test_dataloader.py::test_neighbor_nonuniform[False-pure_gpu-idtype0] PASSED [  0%]
tests/pytorch/test_dataloader.py::test_neighbor_nonuniform[False-pure_gpu-idtype1] PASSED [  0%]
tests/pytorch/test_dataloader.py::test_neighbor_nonuniform[True-cpu-idtype0] PASSED [  0%]
tests/pytorch/test_dataloader.py::test_neighbor_nonuniform[True-cpu-idtype1] PASSED [  0%]
tests/pytorch/test_dataloader.py::test_neighbor_nonuniform[True-uva_cuda_indices-idtype0] PASSED [  0%]
tests/pytorch/test_dataloader.py::test_neighbor_nonuniform[True-uva_cuda_indices-idtype1] PASSED [  0%]
tests/pytorch/test_dataloader.py::test_neighbor_nonuniform[True-uva_cpu_indices-idtype0] PASSED [  0%]
tests/pytorch/test_dataloader.py::test_neighbor_nonuniform[True-uva_cpu_indices-idtype1] PASSED [  0%]
tests/pytorch/test_dataloader.py::test_neighbor_nonuniform[True-pure_gpu-idtype0] PASSED [  0%]
tests/pytorch/test_dataloader.py::test_neighbor_nonuniform[True-pure_gpu-idtype1] PASSED [  0%]
tests/pytorch/test_dataloader.py::test_node_dataloader[False-cpu-full-idtype0] PASSED [  0%]
tests/pytorch/test_dataloader.py::test_node_dataloader[False-cpu-full-idtype1] PASSED [  0%]
tests/pytorch/test_dataloader.py::test_node_dataloader[False-cpu-neighbor-idtype0] PASSED [  0%]
tests/pytorch/test_dataloader.py::test_node_dataloader[False-cpu-neighbor-idtype1] PASSED [  0%]
tests/pytorch/test_dataloader.py::test_node_dataloader[False-cpu-neighbor2-idtype0] PASSED [  0%]
tests/pytorch/test_dataloader.py::test_node_dataloader[False-cpu-neighbor2-idtype1] PASSED [  0%]
tests/pytorch/test_dataloader.py::test_node_dataloader[False-uva_cuda_indices-full-idtype0] PASSED [  0%]
tests/pytorch/test_dataloader.py::test_node_dataloader[False-uva_cuda_indices-full-idtype1] PASSED [  0%]
tests/pytorch/test_dataloader.py::test_node_dataloader[False-uva_cuda_indices-neighbor-idtype0] PASSED [  0%]
tests/pytorch/test_dataloader.py::test_node_dataloader[False-uva_cuda_indices-neighbor-idtype1] PASSED [  0%]
tests/pytorch/test_dataloader.py::test_node_dataloader[False-uva_cuda_indices-neighbor2-idtype0] PASSED [  0%]
tests/pytorch/test_dataloader.py::test_node_dataloader[False-uva_cuda_indices-neighbor2-idtype1] PASSED [  0%]
tests/pytorch/test_dataloader.py::test_node_dataloader[False-uva_cpu_indices-full-idtype0] PASSED [  1%]
tests/pytorch/test_dataloader.py::test_node_dataloader[False-uva_cpu_indices-full-idtype1] PASSED [  1%]
tests/pytorch/test_dataloader.py::test_node_dataloader[False-uva_cpu_indices-neighbor-idtype0] PASSED [  1%]
tests/pytorch/test_dataloader.py::test_node_dataloader[False-uva_cpu_indices-neighbor-idtype1] PASSED [  1%]
tests/pytorch/test_dataloader.py::test_node_dataloader[False-uva_cpu_indices-neighbor2-idtype0] PASSED [  1%]
tests/pytorch/test_dataloader.py::test_node_dataloader[False-uva_cpu_indices-neighbor2-idtype1] PASSED [  1%]
tests/pytorch/test_dataloader.py::test_node_dataloader[False-pure_gpu-full-idtype0] PASSED [  1%]
tests/pytorch/test_dataloader.py::test_node_dataloader[False-pure_gpu-full-idtype1] PASSED [  1%]
tests/pytorch/test_dataloader.py::test_node_dataloader[False-pure_gpu-neighbor-idtype0] PASSED [  1%]
tests/pytorch/test_dataloader.py::test_node_dataloader[False-pure_gpu-neighbor-idtype1] PASSED [  1%]
tests/pytorch/test_dataloader.py::test_node_dataloader[False-pure_gpu-neighbor2-idtype0] PASSED [  1%]
tests/pytorch/test_dataloader.py::test_node_dataloader[False-pure_gpu-neighbor2-idtype1] PASSED [  1%]
tests/pytorch/test_dataloader.py::test_node_dataloader[True-cpu-full-idtype0] PASSED [  1%]
tests/pytorch/test_dataloader.py::test_node_dataloader[True-cpu-full-idtype1] PASSED [  1%]
tests/pytorch/test_dataloader.py::test_node_dataloader[True-cpu-neighbor-idtype0] [       OK ] RowwiseTest.TestCOOerEtypeSampling (40265 ms)
[ RUN      ] RowwiseTest.TestCOOPerEtypeSamplingUniform
PASSED [  1%]
tests/pytorch/test_dataloader.py::test_node_dataloader[True-cpu-neighbor-idtype1] PASSED [  1%]
tests/pytorch/test_dataloader.py::test_node_dataloader[True-cpu-neighbor2-idtype0] PASSED [  1%]
tests/pytorch/test_dataloader.py::test_node_dataloader[True-cpu-neighbor2-idtype1] PASSED [  1%]
tests/pytorch/test_dataloader.py::test_node_dataloader[True-uva_cuda_indices-full-idtype0] PASSED [  1%]
tests/pytorch/test_dataloader.py::test_node_dataloader[True-uva_cuda_indices-full-idtype1] PASSED [  1%]
tests/pytorch/test_dataloader.py::test_node_dataloader[True-uva_cuda_indices-neighbor-idtype0] PASSED [  1%]
tests/pytorch/test_dataloader.py::test_node_dataloader[True-uva_cuda_indices-neighbor-idtype1] PASSED [  1%]
tests/pytorch/test_dataloader.py::test_node_dataloader[True-uva_cuda_indices-neighbor2-idtype0] PASSED [  1%]
tests/pytorch/test_dataloader.py::test_node_dataloader[True-uva_cuda_indices-neighbor2-idtype1] PASSED [  1%]
tests/pytorch/test_dataloader.py::test_node_dataloader[True-uva_cpu_indices-full-idtype0] PASSED [  1%]
tests/pytorch/test_dataloader.py::test_node_dataloader[True-uva_cpu_indices-full-idtype1] PASSED [  1%]
tests/pytorch/test_dataloader.py::test_node_dataloader[True-uva_cpu_indices-neighbor-idtype0] PASSED [  1%]
tests/pytorch/test_dataloader.py::test_node_dataloader[True-uva_cpu_indices-neighbor-idtype1] PASSED [  1%]
tests/pytorch/test_dataloader.py::test_node_dataloader[True-uva_cpu_indices-neighbor2-idtype0] PASSED [  1%]
tests/pytorch/test_dataloader.py::test_node_dataloader[True-uva_cpu_indices-neighbor2-idtype1] PASSED [  1%]
tests/pytorch/test_dataloader.py::test_node_dataloader[True-pure_gpu-full-idtype0] PASSED [  1%]
tests/pytorch/test_dataloader.py::test_node_dataloader[True-pure_gpu-full-idtype1] PASSED [  1%]
tests/pytorch/test_dataloader.py::test_node_dataloader[True-pure_gpu-neighbor-idtype0] PASSED [  1%]
tests/pytorch/test_dataloader.py::test_node_dataloader[True-pure_gpu-neighbor-idtype1] PASSED [  1%]
tests/pytorch/test_dataloader.py::test_node_dataloader[True-pure_gpu-neighbor2-idtype0] PASSED [  1%]
tests/pytorch/test_dataloader.py::test_node_dataloader[True-pure_gpu-neighbor2-idtype1] PASSED [  1%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[False-cpu-neg_sampler0-full-idtype0] PASSED [  1%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[False-cpu-neg_sampler0-full-idtype1] PASSED [  1%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[False-cpu-neg_sampler0-neighbor-idtype0] PASSED [  1%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[False-cpu-neg_sampler0-neighbor-idtype1] PASSED [  1%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[False-cpu-neg_sampler1-full-idtype0] PASSED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[False-cpu-neg_sampler1-full-idtype1] PASSED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[False-cpu-neg_sampler1-neighbor-idtype0] PASSED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[False-cpu-neg_sampler1-neighbor-idtype1] PASSED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[False-cpu-neg_sampler2-full-idtype0] PASSED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[False-cpu-neg_sampler2-full-idtype1] PASSED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[False-cpu-neg_sampler2-neighbor-idtype0] PASSED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[False-cpu-neg_sampler2-neighbor-idtype1] PASSED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[False-uva-neg_sampler0-full-idtype0] PASSED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[False-uva-neg_sampler0-full-idtype1] PASSED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[False-uva-neg_sampler0-neighbor-idtype0] PASSED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[False-uva-neg_sampler0-neighbor-idtype1] PASSED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[False-uva-neg_sampler1-full-idtype0] SKIPPED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[False-uva-neg_sampler1-full-idtype1] SKIPPED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[False-uva-neg_sampler1-neighbor-idtype0] SKIPPED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[False-uva-neg_sampler1-neighbor-idtype1] SKIPPED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[False-uva-neg_sampler2-full-idtype0] SKIPPED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[False-uva-neg_sampler2-full-idtype1] SKIPPED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[False-uva-neg_sampler2-neighbor-idtype0] SKIPPED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[False-uva-neg_sampler2-neighbor-idtype1] SKIPPED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[False-pure_gpu-neg_sampler0-full-idtype0] PASSED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[False-pure_gpu-neg_sampler0-full-idtype1] PASSED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[False-pure_gpu-neg_sampler0-neighbor-idtype0] PASSED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[False-pure_gpu-neg_sampler0-neighbor-idtype1] PASSED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[False-pure_gpu-neg_sampler1-full-idtype0] PASSED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[False-pure_gpu-neg_sampler1-full-idtype1] PASSED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[False-pure_gpu-neg_sampler1-neighbor-idtype0] PASSED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[False-pure_gpu-neg_sampler1-neighbor-idtype1] PASSED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[False-pure_gpu-neg_sampler2-full-idtype0] PASSED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[False-pure_gpu-neg_sampler2-full-idtype1] PASSED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[False-pure_gpu-neg_sampler2-neighbor-idtype0] PASSED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[False-pure_gpu-neg_sampler2-neighbor-idtype1] PASSED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[True-cpu-neg_sampler0-full-idtype0] PASSED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[True-cpu-neg_sampler0-full-idtype1] PASSED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[True-cpu-neg_sampler0-neighbor-idtype0] PASSED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[True-cpu-neg_sampler0-neighbor-idtype1] PASSED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[True-cpu-neg_sampler1-full-idtype0] PASSED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[True-cpu-neg_sampler1-full-idtype1] PASSED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[True-cpu-neg_sampler1-neighbor-idtype0] PASSED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[True-cpu-neg_sampler1-neighbor-idtype1] PASSED [  2%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[True-cpu-neg_sampler2-full-idtype0] PASSED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[True-cpu-neg_sampler2-full-idtype1] PASSED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[True-cpu-neg_sampler2-neighbor-idtype0] PASSED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[True-cpu-neg_sampler2-neighbor-idtype1] PASSED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[True-uva-neg_sampler0-full-idtype0] PASSED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[True-uva-neg_sampler0-full-idtype1] PASSED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[True-uva-neg_sampler0-neighbor-idtype0] PASSED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[True-uva-neg_sampler0-neighbor-idtype1] PASSED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[True-uva-neg_sampler1-full-idtype0] SKIPPED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[True-uva-neg_sampler1-full-idtype1] SKIPPED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[True-uva-neg_sampler1-neighbor-idtype0] SKIPPED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[True-uva-neg_sampler1-neighbor-idtype1] SKIPPED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[True-uva-neg_sampler2-full-idtype0] SKIPPED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[True-uva-neg_sampler2-full-idtype1] SKIPPED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[True-uva-neg_sampler2-neighbor-idtype0] SKIPPED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[True-uva-neg_sampler2-neighbor-idtype1] SKIPPED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[True-pure_gpu-neg_sampler0-full-idtype0] PASSED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[True-pure_gpu-neg_sampler0-full-idtype1] PASSED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[True-pure_gpu-neg_sampler0-neighbor-idtype0] PASSED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[True-pure_gpu-neg_sampler0-neighbor-idtype1] PASSED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[True-pure_gpu-neg_sampler1-full-idtype0] PASSED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[True-pure_gpu-neg_sampler1-full-idtype1] PASSED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[True-pure_gpu-neg_sampler1-neighbor-idtype0] PASSED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[True-pure_gpu-neg_sampler1-neighbor-idtype1] PASSED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[True-pure_gpu-neg_sampler2-full-idtype0] PASSED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[True-pure_gpu-neg_sampler2-full-idtype1] PASSED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[True-pure_gpu-neg_sampler2-neighbor-idtype0] PASSED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader[True-pure_gpu-neg_sampler2-neighbor-idtype1] PASSED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader_excludes[1-sampler0-None-False] PASSED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader_excludes[1-sampler0-None-True] PASSED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader_excludes[1-sampler0-self-False] PASSED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader_excludes[1-sampler0-self-True] PASSED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader_excludes[1-sampler0-reverse_id-False] PASSED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader_excludes[1-sampler0-reverse_id-True] PASSED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader_excludes[1-sampler0-reverse_types-False] PASSED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader_excludes[1-sampler0-reverse_types-True] PASSED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader_excludes[1-sampler1-None-False] PASSED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader_excludes[1-sampler1-None-True] PASSED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader_excludes[1-sampler1-self-False] PASSED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader_excludes[1-sampler1-self-True] PASSED [  3%]
tests/pytorch/test_dataloader.py::test_edge_dataloader_excludes[1-sampler1-reverse_id-False] PASSED [  4%]
tests/pytorch/test_dataloader.py::test_edge_dataloader_excludes[1-sampler1-reverse_id-True] PASSED [  4%]
tests/pytorch/test_dataloader.py::test_edge_dataloader_excludes[1-sampler1-reverse_types-False] PASSED [  4%]
tests/pytorch/test_dataloader.py::test_edge_dataloader_excludes[1-sampler1-reverse_types-True] PASSED [  4%]
tests/pytorch/test_dataloader.py::test_edge_dataloader_excludes[50-sampler0-None-False] PASSED [  4%]
tests/pytorch/test_dataloader.py::test_edge_dataloader_excludes[50-sampler0-None-True] PASSED [  4%]
tests/pytorch/test_dataloader.py::test_edge_dataloader_excludes[50-sampler0-self-False] PASSED [  4%]
tests/pytorch/test_dataloader.py::test_edge_dataloader_excludes[50-sampler0-self-True] PASSED [  4%]
tests/pytorch/test_dataloader.py::test_edge_dataloader_excludes[50-sampler0-reverse_id-False] PASSED [  4%]
tests/pytorch/test_dataloader.py::test_edge_dataloader_excludes[50-sampler0-reverse_id-True] PASSED [  4%]
tests/pytorch/test_dataloader.py::test_edge_dataloader_excludes[50-sampler0-reverse_types-False] PASSED [  4%]
tests/pytorch/test_dataloader.py::test_edge_dataloader_excludes[50-sampler0-reverse_types-True] PASSED [  4%]
tests/pytorch/test_dataloader.py::test_edge_dataloader_excludes[50-sampler1-None-False] PASSED [  4%]
tests/pytorch/test_dataloader.py::test_edge_dataloader_excludes[50-sampler1-None-True] PASSED [  4%]
tests/pytorch/test_dataloader.py::test_edge_dataloader_excludes[50-sampler1-self-False] PASSED [  4%]
tests/pytorch/test_dataloader.py::test_edge_dataloader_excludes[50-sampler1-self-True] PASSED [  4%]
tests/pytorch/test_dataloader.py::test_edge_dataloader_excludes[50-sampler1-reverse_id-False] PASSED [  4%]
tests/pytorch/test_dataloader.py::test_edge_dataloader_excludes[50-sampler1-reverse_id-True] PASSED [  4%]
tests/pytorch/test_dataloader.py::test_edge_dataloader_excludes[50-sampler1-reverse_types-False] PASSED [  4%]
tests/pytorch/test_dataloader.py::test_edge_dataloader_excludes[50-sampler1-reverse_types-True] PASSED [  4%]
tests/pytorch/test_dist_optim.py::test_sparse_opt PASSED                         [  5%]
tests/compute/test_data.py::test_tudataset_regression PASSED                         [  5%]
tests/compute/test_data.py::test_tudataset_regression [       OK ] RowwiseTest.TestCOOPerEtypeSamplingUniform (30441 ms)
[ RUN      ] RowwiseTest.TestCSRTopk
[       OK ] RowwiseTest.TestCSRTopk (381 ms)
[ RUN      ] RowwiseTest.TestCOOTopk
[       OK ] RowwiseTest.TestCOOTopk (1037 ms)
[ RUN      ] RowwiseTest.TestCSRSamplingBiased
[       OK ] RowwiseTest.TestCSRSamplingBiased (3684 ms)
[----------] 11 tests from RowwiseTest (188390 ms total)

[----------] 3 tests from SampleUtilsTest
[ RUN      ] SampleUtilsTest.TestWithReplacement
PASSED             [  5%]
tests/compute/test_data.py::test_data_hash PASSED                        [  5%]
tests/compute/test_data.py::test_citation_graph [       OK ] SampleUtilsTest.TestWithReplacement (2537 ms)
[ RUN      ] SampleUtilsTest.TestWithoutReplacementOrder
[       OK ] SampleUtilsTest.TestWithoutReplacementOrder (0 ms)
[ RUN      ] SampleUtilsTest.TestWithoutReplacementUnique
PASSED                   [  5%]
tests/compute/test_data.py::test_gnn_benchmark PASSED             [  5%]
tests/compute/test_data.py::test_data_hash PASSED                        [  5%]
tests/compute/test_data.py::test_citation_graph [       OK ] SampleUtilsTest.TestWithoutReplacementUnique (3933 ms)
[----------] 3 tests from SampleUtilsTest (6470 ms total)

[----------] 3 tests from RandomTest
[ RUN      ] RandomTest.TestChoice
[       OK ] RandomTest.TestChoice (1 ms)
[ RUN      ] RandomTest.TestUniformChoice
[       OK ] RandomTest.TestUniformChoice (0 ms)
[ RUN      ] RandomTest.TestBiasedChoice
[       OK ] RandomTest.TestBiasedChoice (11 ms)
[----------] 3 tests from RandomTest (12 ms total)

[----------] 4 tests from Serialize
[ RUN      ] Serialize.UnitGraph_COO
[       OK ] Serialize.UnitGraph_COO (59 ms)
[ RUN      ] Serialize.UnitGraph_CSR
[       OK ] Serialize.UnitGraph_CSR (49 ms)
[ RUN      ] Serialize.ImmutableGraph
[       OK ] Serialize.ImmutableGraph (42 ms)
[ RUN      ] Serialize.HeteroGraph
[       OK ] Serialize.HeteroGraph (100 ms)
[----------] 4 tests from Serialize (251 ms total)

[----------] 3 tests from SmartPtrTest/0, where TypeParam = std::shared_ptr<MyClass>
[ RUN      ] SmartPtrTest/0.Obj_Test
[       OK ] SmartPtrTest/0.Obj_Test (0 ms)
[ RUN      ] SmartPtrTest/0.Vector_Test1
[       OK ] SmartPtrTest/0.Vector_Test1 (0 ms)
[ RUN      ] SmartPtrTest/0.Vector_Test2
[       OK ] SmartPtrTest/0.Vector_Test2 (0 ms)
[----------] 3 tests from SmartPtrTest/0 (0 ms total)

[----------] 3 tests from SmartPtrTest/1, where TypeParam = std::unique_ptr<MyClass, std::default_delete<MyClass> >
[ RUN      ] SmartPtrTest/1.Obj_Test
[       OK ] SmartPtrTest/1.Obj_Test (0 ms)
[ RUN      ] SmartPtrTest/1.Vector_Test1
[       OK ] SmartPtrTest/1.Vector_Test1 (0 ms)
[ RUN      ] SmartPtrTest/1.Vector_Test2
[       OK ] SmartPtrTest/1.Vector_Test2 (0 ms)
[----------] 3 tests from SmartPtrTest/1 (0 ms total)

[----------] 19 tests from SpmatTest
[ RUN      ] SpmatTest.COOToCSR
[       OK ] SpmatTest.COOToCSR (383 ms)
[ RUN      ] SpmatTest.TestCOOHasDuplicate
[       OK ] SpmatTest.TestCOOHasDuplicate (0 ms)
[ RUN      ] SpmatTest.COOSort
[       OK ] SpmatTest.COOSort (0 ms)
[ RUN      ] SpmatTest.TestCOOReorder
[       OK ] SpmatTest.TestCOOReorder (6 ms)
[ RUN      ] SpmatTest.COOGetData
[       OK ] SpmatTest.COOGetData (55 ms)
[ RUN      ] SpmatTest.COOGetDataAndIndices
[       OK ] SpmatTest.COOGetDataAndIndices (0 ms)
[ RUN      ] SpmatTest.TestCSRIsNonZero
[       OK ] SpmatTest.TestCSRIsNonZero (12 ms)
[ RUN      ] SpmatTest.TestCSRGetRowNNZ
[       OK ] SpmatTest.TestCSRGetRowNNZ (0 ms)
[ RUN      ] SpmatTest.TestCSRGetRowColumnIndices
[       OK ] SpmatTest.TestCSRGetRowColumnIndices (0 ms)
[ RUN      ] SpmatTest.TestCSRGetRowData
[       OK ] SpmatTest.TestCSRGetRowData (0 ms)
[ RUN      ] SpmatTest.CSRGetData
[       OK ] SpmatTest.CSRGetData (0 ms)
[ RUN      ] SpmatTest.CSRGetDataAndIndices
[       OK ] SpmatTest.CSRGetDataAndIndices (0 ms)
[ RUN      ] SpmatTest.CSRTranspose
[       OK ] SpmatTest.CSRTranspose (0 ms)
[ RUN      ] SpmatTest.CSRToCOO
[       OK ] SpmatTest.CSRToCOO (0 ms)
[ RUN      ] SpmatTest.TestCSRSliceRows
[       OK ] SpmatTest.TestCSRSliceRows (601 ms)
[ RUN      ] SpmatTest.CSRSliceMatrix
[       OK ] SpmatTest.CSRSliceMatrix (2 ms)
[ RUN      ] SpmatTest.CSRHasDuplicate
[       OK ] SpmatTest.CSRHasDuplicate (0 ms)
[ RUN      ] SpmatTest.CSRSort
[       OK ] SpmatTest.CSRSort (0 ms)
[ RUN      ] SpmatTest.TestCSRReorder
[       OK ] SpmatTest.TestCSRReorder (0 ms)
[----------] 19 tests from SpmatTest (1059 ms total)

[----------] 6 tests from SpmmTest
[ RUN      ] SpmmTest.TestSpmmCopyLhs
[       OK ] SpmmTest.TestSpmmCopyLhs (0 ms)
[ RUN      ] SpmmTest.TestSpmmCopyRhs
[       OK ] SpmmTest.TestSpmmCopyRhs (1 ms)
[ RUN      ] SpmmTest.TestSpmmAdd
[       OK ] SpmmTest.TestSpmmAdd (0 ms)
[ RUN      ] SpmmTest.TestSpmmSub
[       OK ] SpmmTest.TestSpmmSub (0 ms)
[ RUN      ] SpmmTest.TestSpmmMul
[       OK ] SpmmTest.TestSpmmMul (1 ms)
[ RUN      ] SpmmTest.TestSpmmDiv
[       OK ] SpmmTest.TestSpmmDiv (0 ms)
[----------] 6 tests from SpmmTest (2 ms total)

[----------] 7 tests from UniGraphTest
[ RUN      ] UniGraphTest.TestUnitGraph_CopyTo
[       OK ] UniGraphTest.TestUnitGraph_CopyTo (0 ms)
[ RUN      ] UniGraphTest.TestUnitGraph_InOutDegrees
[       OK ] UniGraphTest.TestUnitGraph_InOutDegrees (52 ms)
[ RUN      ] UniGraphTest.TestUnitGraph_Create
[       OK ] UniGraphTest.TestUnitGraph_Create (0 ms)
[ RUN      ] UniGraphTest.TestUnitGraph_GetInCSR
[       OK ] UniGraphTest.TestUnitGraph_GetInCSR (190 ms)
[ RUN      ] UniGraphTest.TestUnitGraph_GetOutCSR
[       OK ] UniGraphTest.TestUnitGraph_GetOutCSR (97 ms)
[ RUN      ] UniGraphTest.TestUnitGraph_GetCOO
[       OK ] UniGraphTest.TestUnitGraph_GetCOO (0 ms)
[ RUN      ] UniGraphTest.TestUnitGraph_Reserve
[       OK ] UniGraphTest.TestUnitGraph_Reserve (122 ms)
[----------] 7 tests from UniGraphTest (461 ms total)

[----------] 4 tests from ZeroCopySerialize
[ RUN      ] ZeroCopySerialize.NDArray
[       OK ] ZeroCopySerialize.NDArray (0 ms)
[ RUN      ] ZeroCopySerialize.ZeroShapeNDArray
[       OK ] ZeroCopySerialize.ZeroShapeNDArray (0 ms)
[ RUN      ] ZeroCopySerialize.SharedMem
[       OK ] ZeroCopySerialize.SharedMem (0 ms)
[ RUN      ] ZeroCopySerialize.HeteroGraph
[       OK ] ZeroCopySerialize.HeteroGraph (91 ms)
[----------] 4 tests from ZeroCopySerialize (91 ms total)

[----------] Global test environment tear-down
[==========] 101 tests from 23 test suites ran. (205767 ms total)
[  PASSED  ] 101 tests.
~/jenkins/workspace/dgl_PR-4648
Post stage
[Pipeline] cleanWs
[WS-CLEANUP] Deleting project workspace...
[WS-CLEANUP] Deferred wipeout is disabled by the job configuration...
[WS-CLEANUP] done
[Pipeline] }
$ docker stop --time=1 280b39495ad3fad21dbb9762a178907987f011c3e8f474eb1caee524b0ff882c
PASSED                   [  5%]
tests/compute/test_data.py::test_gnn_benchmark PASSED                    [  5%]
tests/compute/test_data.py::test_reddit $ docker rm -f 280b39495ad3fad21dbb9762a178907987f011c3e8f474eb1caee524b0ff882c
[Pipeline] // withDockerContainer
[Pipeline] }
[Pipeline] // withEnv
[Pipeline] }
Running on dgl-manual-large-cpu in /root/jenkins/workspace/dgl_PR-4648
PASSED                 [  4%]
tests/pytorch/test_geometry.py::test_fps PASSED                          [  4%]
tests/pytorch/test_geometry.py::test_fps_start_idx PASSED                [  4%]
tests/pytorch/test_geometry.py::test_knn_cpu[False-euclidean-bruteforce-blas] PASSED [  4%]
tests/pytorch/test_geometry.py::test_knn_cpu[False-euclidean-bruteforce] PASSED [  4%]
tests/pytorch/test_geometry.py::test_knn_cpu[False-euclidean-kd-tree] PASSED [  4%]
tests/pytorch/test_geometry.py::test_knn_cpu[False-cosine-bruteforce-blas] PASSED [  4%]
tests/pytorch/test_geometry.py::test_knn_cpu[False-cosine-bruteforce] PASSED [  4%]
tests/pytorch/test_geometry.py::test_knn_cpu[False-cosine-kd-tree] PASSED [  4%]
tests/pytorch/test_geometry.py::test_knn_cpu[True-euclidean-bruteforce-blas] PASSED [  4%]
tests/pytorch/test_geometry.py::test_knn_cpu[True-euclidean-bruteforce] PASSED [  4%]
tests/pytorch/test_geometry.py::test_knn_cpu[True-euclidean-kd-tree] PASSED [  4%]
tests/pytorch/test_geometry.py::test_knn_cpu[True-cosine-bruteforce-blas] PASSED [  4%]
tests/pytorch/test_geometry.py::test_knn_cpu[True-cosine-bruteforce] PASSED [  4%]
tests/pytorch/test_geometry.py::test_knn_cpu[True-cosine-kd-tree] PASSED [  4%]
tests/pytorch/test_geometry.py::test_knn_cuda[False-euclidean-bruteforce-blas] PASSED [  4%]
tests/pytorch/test_geometry.py::test_knn_cuda[False-euclidean-bruteforce] PASSED [  4%]
tests/pytorch/test_geometry.py::test_knn_cuda[False-euclidean-bruteforce-sharemem] PASSED [  4%]
tests/pytorch/test_geometry.py::test_knn_cuda[False-cosine-bruteforce-blas] PASSED [  4%]
tests/pytorch/test_geometry.py::test_knn_cuda[False-cosine-bruteforce] PASSED [  5%]
tests/pytorch/test_geometry.py::test_knn_cuda[False-cosine-bruteforce-sharemem] PASSED [  5%]
tests/pytorch/test_geometry.py::test_knn_cuda[True-euclidean-bruteforce-blas] PASSED [  5%]
tests/pytorch/test_geometry.py::test_knn_cuda[True-euclidean-bruteforce] PASSED [  5%]
tests/pytorch/test_geometry.py::test_knn_cuda[True-euclidean-bruteforce-sharemem] PASSED [  5%]
tests/pytorch/test_geometry.py::test_knn_cuda[True-cosine-bruteforce-blas] PASSED [  5%]
tests/pytorch/test_geometry.py::test_knn_cuda[True-cosine-bruteforce] PASSED [  5%]
tests/pytorch/test_geometry.py::test_knn_cuda[True-cosine-bruteforce-sharemem] PASSED [  5%]
tests/pytorch/test_geometry.py::test_edge_coarsening[True-True-g0-idtype0] PASSED [  5%]
tests/pytorch/test_geometry.py::test_edge_coarsening[True-True-g0-idtype1] PASSED [  5%]
tests/pytorch/test_geometry.py::test_edge_coarsening[True-True-g1-idtype0] PASSED [  5%]
tests/pytorch/test_geometry.py::test_edge_coarsening[True-True-g1-idtype1] PASSED [  5%]
tests/pytorch/test_geometry.py::test_edge_coarsening[True-True-g2-idtype0] PASSED [  5%]
tests/pytorch/test_geometry.py::test_edge_coarsening[True-True-g2-idtype1] PASSED [  5%]
tests/pytorch/test_geometry.py::test_edge_coarsening[True-True-g3-idtype0] PASSED [  5%]
tests/pytorch/test_geometry.py::test_edge_coarsening[True-True-g3-idtype1] PASSED [  5%]
tests/pytorch/test_geometry.py::test_edge_coarsening[True-True-g4-idtype0] PASSED [  5%]
tests/pytorch/test_geometry.py::test_edge_coarsening[True-True-g4-idtype1] PASSED [  5%]
tests/pytorch/test_geometry.py::test_edge_coarsening[True-True-g5-idtype0] PASSED [  5%]
tests/pytorch/test_geometry.py::test_edge_coarsening[True-True-g5-idtype1] PASSED [  5%]
tests/pytorch/test_geometry.py::test_edge_coarsening[True-True-g6-idtype0] PASSED [  5%]
tests/pytorch/test_geometry.py::test_edge_coarsening[True-True-g6-idtype1] PASSED [  5%]
tests/pytorch/test_geometry.py::test_edge_coarsening[True-False-g0-idtype0] PASSED [  5%]
tests/pytorch/test_geometry.py::test_edge_coarsening[True-False-g0-idtype1] PASSED [  5%]
tests/pytorch/test_geometry.py::test_edge_coarsening[True-False-g1-idtype0] PASSED [  5%]
tests/pytorch/test_geometry.py::test_edge_coarsening[True-False-g1-idtype1] PASSED [  5%]
tests/pytorch/test_geometry.py::test_edge_coarsening[True-False-g2-idtype0] PASSED [  5%]
tests/pytorch/test_geometry.py::test_edge_coarsening[True-False-g2-idtype1] PASSED [  5%]
tests/pytorch/test_geometry.py::test_edge_coarsening[True-False-g3-idtype0] PASSED [  5%]
tests/pytorch/test_geometry.py::test_edge_coarsening[True-False-g3-idtype1] PASSED [  5%]
tests/pytorch/test_geometry.py::test_edge_coarsening[True-False-g4-idtype0] PASSED [  5%]
tests/pytorch/test_geometry.py::test_edge_coarsening[True-False-g4-idtype1] PASSED [  5%]
tests/pytorch/test_geometry.py::test_edge_coarsening[True-False-g5-idtype0] PASSED [  5%]
tests/pytorch/test_geometry.py::test_edge_coarsening[True-False-g5-idtype1] PASSED [  5%]
tests/pytorch/test_geometry.py::test_edge_coarsening[True-False-g6-idtype0] PASSED [  5%]
tests/pytorch/test_geometry.py::test_edge_coarsening[True-False-g6-idtype1] PASSED [  5%]
tests/pytorch/test_geometry.py::test_edge_coarsening[False-True-g0-idtype0] PASSED [  5%]
tests/pytorch/test_geometry.py::test_edge_coarsening[False-True-g0-idtype1] PASSED [  5%]
tests/pytorch/test_geometry.py::test_edge_coarsening[False-True-g1-idtype0] PASSED [  5%]
tests/pytorch/test_geometry.py::test_edge_coarsening[False-True-g1-idtype1] PASSED [  5%]
tests/pytorch/test_geometry.py::test_edge_coarsening[False-True-g2-idtype0] PASSED [  6%]
tests/pytorch/test_geometry.py::test_edge_coarsening[False-True-g2-idtype1] PASSED [  6%]
tests/pytorch/test_geometry.py::test_edge_coarsening[False-True-g3-idtype0] PASSED [  6%]
tests/pytorch/test_geometry.py::test_edge_coarsening[False-True-g3-idtype1] PASSED [  6%]
tests/pytorch/test_geometry.py::test_edge_coarsening[False-True-g4-idtype0] PASSED [  6%]
tests/pytorch/test_geometry.py::test_edge_coarsening[False-True-g4-idtype1] PASSED [  6%]
tests/pytorch/test_geometry.py::test_edge_coarsening[False-True-g5-idtype0] PASSED [  6%]
tests/pytorch/test_geometry.py::test_edge_coarsening[False-True-g5-idtype1] PASSED [  6%]
tests/pytorch/test_geometry.py::test_edge_coarsening[False-True-g6-idtype0] PASSED [  6%]
tests/pytorch/test_geometry.py::test_edge_coarsening[False-True-g6-idtype1] PASSED [  6%]
tests/pytorch/test_geometry.py::test_edge_coarsening[False-False-g0-idtype0] [Pipeline] // node
[Pipeline] {
[Pipeline] }
[Pipeline] // stage
[Pipeline] }
[Pipeline] checkout
The recommended git tool is: git
using credential 150de63f-189c-4717-bcaf-010460d2f51a
Cloning the remote Git repository
Cloning with configured refspecs honoured and without tags
PASSED [  6%]
tests/pytorch/test_geometry.py::test_edge_coarsening[False-False-g0-idtype1] PASSED [  6%]
tests/pytorch/test_geometry.py::test_edge_coarsening[False-False-g1-idtype0] PASSED [  6%]
tests/pytorch/test_geometry.py::test_edge_coarsening[False-False-g1-idtype1] PASSED [  6%]
tests/pytorch/test_geometry.py::test_edge_coarsening[False-False-g2-idtype0] PASSED [  6%]
tests/pytorch/test_geometry.py::test_edge_coarsening[False-False-g2-idtype1] PASSED [  6%]
tests/pytorch/test_geometry.py::test_edge_coarsening[False-False-g3-idtype0] PASSED [  6%]
tests/pytorch/test_geometry.py::test_edge_coarsening[False-False-g3-idtype1] PASSED [  6%]
tests/pytorch/test_geometry.py::test_edge_coarsening[False-False-g4-idtype0] PASSED [  6%]
tests/pytorch/test_geometry.py::test_edge_coarsening[False-False-g4-idtype1] PASSED [  6%]
tests/pytorch/test_geometry.py::test_edge_coarsening[False-False-g5-idtype0] PASSED [  6%]
tests/pytorch/test_geometry.py::test_edge_coarsening[False-False-g5-idtype1] PASSED [  6%]
tests/pytorch/test_geometry.py::test_edge_coarsening[False-False-g6-idtype0] PASSED [  6%]
tests/pytorch/test_geometry.py::test_edge_coarsening[False-False-g6-idtype1] PASSED [  6%]
tests/pytorch/test_ipc.py::test_torch_ipc Cloning repository https://github.com/dmlc/dgl.git
 > git init /root/jenkins/workspace/dgl_PR-4648 # timeout=10
Fetching upstream changes from https://github.com/dmlc/dgl.git
 > git --version # timeout=10
 > git --version # 'git version 2.17.1'
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git fetch --no-tags --progress -- https://github.com/dmlc/dgl.git +refs/pull/4648/head:refs/remotes/origin/PR-4648 +refs/heads/master:refs/remotes/origin/master # timeout=10
PASSED                         [  6%]
tests/pytorch/test_nn.py::test_graph_conv0[1] PASSED                     [  6%]
tests/pytorch/test_nn.py::test_graph_conv0[2] PASSED                     [  6%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-none-g0-idtype0] PASSED [  6%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-none-g0-idtype1] PASSED [  6%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-none-g1-idtype0] PASSED [  6%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-none-g1-idtype1] PASSED [  6%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-none-g2-idtype0] PASSED [  6%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-none-g2-idtype1] PASSED [  6%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-none-g3-idtype0] PASSED [  6%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-none-g3-idtype1] PASSED [  6%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-none-g4-idtype0] PASSED [  6%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-none-g4-idtype1] PASSED [  6%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-none-g5-idtype0] PASSED [  6%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-none-g5-idtype1] PASSED [  6%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-none-g6-idtype0] PASSED [  6%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-none-g6-idtype1] PASSED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-none-g7-idtype0] PASSED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-none-g7-idtype1] PASSED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-both-g0-idtype0] PASSED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-both-g0-idtype1] PASSED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-both-g1-idtype0] PASSED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-both-g1-idtype1] PASSED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-both-g2-idtype0] PASSED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-both-g2-idtype1] PASSED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-both-g3-idtype0] PASSED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-both-g3-idtype1] PASSED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-both-g4-idtype0] PASSED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-both-g4-idtype1] PASSED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-both-g5-idtype0] PASSED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-both-g5-idtype1] PASSED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-both-g6-idtype0] PASSED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-both-g6-idtype1] PASSED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-both-g7-idtype0] PASSED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-both-g7-idtype1] PASSED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-right-g0-idtype0] PASSED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-right-g0-idtype1] PASSED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-right-g1-idtype0] PASSED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-right-g1-idtype1] PASSED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-right-g2-idtype0] PASSED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-right-g2-idtype1] PASSED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-right-g3-idtype0] PASSED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-right-g3-idtype1] PASSED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-right-g4-idtype0] PASSED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-right-g4-idtype1] PASSED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-right-g5-idtype0] PASSED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-right-g5-idtype1] PASSED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-right-g6-idtype0] PASSED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-right-g6-idtype1] PASSED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-right-g7-idtype0] PASSED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-right-g7-idtype1] PASSED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-left-g0-idtype0] PASSED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-left-g0-idtype1] PASSED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-left-g1-idtype0] PASSED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-left-g1-idtype1] PASSED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-left-g2-idtype0] PASSED [  7%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-left-g2-idtype1] PASSED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-left-g3-idtype0] PASSED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-left-g3-idtype1] PASSED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-left-g4-idtype0] PASSED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-left-g4-idtype1] PASSED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-left-g5-idtype0] PASSED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-left-g5-idtype1] PASSED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-left-g6-idtype0] PASSED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-left-g6-idtype1] PASSED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-left-g7-idtype0] PASSED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-True-left-g7-idtype1] PASSED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-none-g0-idtype0] PASSED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-none-g0-idtype1] PASSED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-none-g1-idtype0] PASSED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-none-g1-idtype1] PASSED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-none-g2-idtype0] PASSED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-none-g2-idtype1] PASSED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-none-g3-idtype0] PASSED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-none-g3-idtype1] PASSED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-none-g4-idtype0] PASSED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-none-g4-idtype1] PASSED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-none-g5-idtype0] PASSED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-none-g5-idtype1] PASSED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-none-g6-idtype0] PASSED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-none-g6-idtype1] PASSED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-none-g7-idtype0] PASSED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-none-g7-idtype1] PASSED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-both-g0-idtype0] PASSED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-both-g0-idtype1] PASSED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-both-g1-idtype0] PASSED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-both-g1-idtype1] PASSED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-both-g2-idtype0] PASSED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-both-g2-idtype1] PASSED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-both-g3-idtype0] PASSED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-both-g3-idtype1] PASSED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-both-g4-idtype0] PASSED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-both-g4-idtype1] PASSED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-both-g5-idtype0] PASSED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-both-g5-idtype1] PASSED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-both-g6-idtype0] PASSED [  8%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-both-g6-idtype1] PASSED [  9%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-both-g7-idtype0] PASSED [  9%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-both-g7-idtype1] PASSED [  9%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-right-g0-idtype0] PASSED [  9%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-right-g0-idtype1] PASSED [  9%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-right-g1-idtype0] PASSED [  9%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-right-g1-idtype1] PASSED [  9%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-right-g2-idtype0] PASSED [  9%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-right-g2-idtype1] PASSED [  9%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-right-g3-idtype0] PASSED [  9%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-right-g3-idtype1] PASSED [  9%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-right-g4-idtype0] PASSED [  9%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-right-g4-idtype1] PASSED [  9%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-right-g5-idtype0] PASSED [  9%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-right-g5-idtype1] PASSED [  9%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-right-g6-idtype0] PASSED [  9%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-right-g6-idtype1] PASSED [  9%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-right-g7-idtype0] PASSED [  9%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-right-g7-idtype1] PASSED [  9%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-left-g0-idtype0] PASSED [  9%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-left-g0-idtype1] PASSED [  9%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-left-g1-idtype0] PASSED [  9%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-left-g1-idtype1] PASSED [  9%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-left-g2-idtype0] PASSED [  9%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-left-g2-idtype1] PASSED [  9%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-left-g3-idtype0] PASSED [  9%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-left-g3-idtype1] PASSED [  9%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-left-g4-idtype0] PASSED [  9%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-left-g4-idtype1] PASSED [  9%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-left-g5-idtype0] PASSED [  9%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-left-g5-idtype1] PASSED [  9%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-left-g6-idtype0] PASSED [  9%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-left-g6-idtype1] PASSED [  9%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-left-g7-idtype0] PASSED [  9%]
tests/pytorch/test_nn.py::test_graph_conv[1-True-False-left-g7-idtype1] PASSED [  9%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-none-g0-idtype0] PASSED [  9%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-none-g0-idtype1] PASSED [  9%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-none-g1-idtype0] PASSED [  9%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-none-g1-idtype1] PASSED [  9%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-none-g2-idtype0] PASSED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-none-g2-idtype1] PASSED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-none-g3-idtype0] PASSED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-none-g3-idtype1] PASSED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-none-g4-idtype0] PASSED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-none-g4-idtype1] PASSED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-none-g5-idtype0] PASSED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-none-g5-idtype1] PASSED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-none-g6-idtype0] PASSED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-none-g6-idtype1] PASSED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-none-g7-idtype0] PASSED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-none-g7-idtype1] PASSED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-both-g0-idtype0] PASSED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-both-g0-idtype1] PASSED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-both-g1-idtype0] PASSED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-both-g1-idtype1] PASSED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-both-g2-idtype0] PASSED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-both-g2-idtype1] PASSED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-both-g3-idtype0] PASSED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-both-g3-idtype1] PASSED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-both-g4-idtype0] PASSED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-both-g4-idtype1] PASSED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-both-g5-idtype0] PASSED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-both-g5-idtype1] PASSED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-both-g6-idtype0] PASSED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-both-g6-idtype1] PASSED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-both-g7-idtype0] PASSED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-both-g7-idtype1] PASSED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-right-g0-idtype0] PASSED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-right-g0-idtype1] PASSED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-right-g1-idtype0] PASSED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-right-g1-idtype1] PASSED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-right-g2-idtype0] PASSED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-right-g2-idtype1] PASSED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-right-g3-idtype0] PASSED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-right-g3-idtype1] PASSED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-right-g4-idtype0] PASSED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-right-g4-idtype1] PASSED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-right-g5-idtype0] PASSED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-right-g5-idtype1] PASSED [ 10%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-right-g6-idtype0] PASSED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-right-g6-idtype1] PASSED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-right-g7-idtype0] PASSED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-right-g7-idtype1] PASSED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-left-g0-idtype0] PASSED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-left-g0-idtype1] PASSED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-left-g1-idtype0] PASSED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-left-g1-idtype1] PASSED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-left-g2-idtype0] PASSED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-left-g2-idtype1] PASSED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-left-g3-idtype0] PASSED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-left-g3-idtype1] PASSED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-left-g4-idtype0] PASSED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-left-g4-idtype1] PASSED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-left-g5-idtype0] PASSED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-left-g5-idtype1] PASSED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-left-g6-idtype0] PASSED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-left-g6-idtype1] PASSED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-left-g7-idtype0] PASSED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-True-left-g7-idtype1] PASSED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-none-g0-idtype0] PASSED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-none-g0-idtype1] PASSED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-none-g1-idtype0] PASSED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-none-g1-idtype1] PASSED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-none-g2-idtype0] PASSED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-none-g2-idtype1] PASSED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-none-g3-idtype0] PASSED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-none-g3-idtype1] PASSED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-none-g4-idtype0] PASSED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-none-g4-idtype1] PASSED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-none-g5-idtype0] PASSED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-none-g5-idtype1] PASSED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-none-g6-idtype0] PASSED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-none-g6-idtype1] PASSED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-none-g7-idtype0] PASSED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-none-g7-idtype1] PASSED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-both-g0-idtype0] PASSED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-both-g0-idtype1] PASSED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-both-g1-idtype0] PASSED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-both-g1-idtype1] PASSED [ 11%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-both-g2-idtype0] PASSED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-both-g2-idtype1] PASSED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-both-g3-idtype0] PASSED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-both-g3-idtype1] PASSED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-both-g4-idtype0] PASSED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-both-g4-idtype1] PASSED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-both-g5-idtype0] PASSED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-both-g5-idtype1] PASSED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-both-g6-idtype0] PASSED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-both-g6-idtype1] PASSED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-both-g7-idtype0] PASSED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-both-g7-idtype1] PASSED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-right-g0-idtype0] PASSED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-right-g0-idtype1] PASSED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-right-g1-idtype0] PASSED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-right-g1-idtype1] PASSED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-right-g2-idtype0] PASSED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-right-g2-idtype1] PASSED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-right-g3-idtype0] PASSED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-right-g3-idtype1] PASSED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-right-g4-idtype0] PASSED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-right-g4-idtype1] PASSED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-right-g5-idtype0] PASSED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-right-g5-idtype1] PASSED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-right-g6-idtype0] PASSED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-right-g6-idtype1] PASSED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-right-g7-idtype0] PASSED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-right-g7-idtype1] PASSED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-left-g0-idtype0] PASSED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-left-g0-idtype1] PASSED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-left-g1-idtype0] PASSED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-left-g1-idtype1] PASSED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-left-g2-idtype0] PASSED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-left-g2-idtype1] PASSED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-left-g3-idtype0] PASSED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-left-g3-idtype1] PASSED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-left-g4-idtype0] PASSED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-left-g4-idtype1] PASSED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-left-g5-idtype0] PASSED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-left-g5-idtype1] PASSED [ 12%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-left-g6-idtype0] PASSED [ 13%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-left-g6-idtype1] PASSED [ 13%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-left-g7-idtype0] PASSED [ 13%]
tests/pytorch/test_nn.py::test_graph_conv[1-False-False-left-g7-idtype1] PASSED [ 13%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-none-g0-idtype0] PASSED [ 13%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-none-g0-idtype1] PASSED [ 13%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-none-g1-idtype0] PASSED [ 13%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-none-g1-idtype1] PASSED [ 13%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-none-g2-idtype0] PASSED [ 13%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-none-g2-idtype1] PASSED [ 13%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-none-g3-idtype0] PASSED [ 13%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-none-g3-idtype1] PASSED [ 13%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-none-g4-idtype0] PASSED [ 13%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-none-g4-idtype1] PASSED [ 13%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-none-g5-idtype0] PASSED [ 13%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-none-g5-idtype1] PASSED [ 13%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-none-g6-idtype0] PASSED [ 13%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-none-g6-idtype1] PASSED [ 13%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-none-g7-idtype0] PASSED [ 13%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-none-g7-idtype1] PASSED [ 13%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-both-g0-idtype0] PASSED [ 13%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-both-g0-idtype1] PASSED [ 13%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-both-g1-idtype0] PASSED [ 13%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-both-g1-idtype1] PASSED [ 13%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-both-g2-idtype0] PASSED [ 13%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-both-g2-idtype1] PASSED [ 13%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-both-g3-idtype0] PASSED [ 13%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-both-g3-idtype1] PASSED [ 13%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-both-g4-idtype0] PASSED [ 13%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-both-g4-idtype1] PASSED [ 13%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-both-g5-idtype0] PASSED [ 13%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-both-g5-idtype1] PASSED [ 13%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-both-g6-idtype0] PASSED [ 13%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-both-g6-idtype1] PASSED [ 13%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-both-g7-idtype0] PASSED [ 13%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-both-g7-idtype1] PASSED [ 13%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-right-g0-idtype0] PASSED [ 13%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-right-g0-idtype1] PASSED [ 13%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-right-g1-idtype0] PASSED [ 13%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-right-g1-idtype1] PASSED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-right-g2-idtype0] PASSED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-right-g2-idtype1] PASSED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-right-g3-idtype0] PASSED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-right-g3-idtype1] PASSED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-right-g4-idtype0] PASSED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-right-g4-idtype1] PASSED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-right-g5-idtype0] PASSED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-right-g5-idtype1] PASSED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-right-g6-idtype0] PASSED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-right-g6-idtype1] PASSED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-right-g7-idtype0] PASSED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-right-g7-idtype1] PASSED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-left-g0-idtype0] PASSED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-left-g0-idtype1] PASSED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-left-g1-idtype0] PASSED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-left-g1-idtype1] PASSED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-left-g2-idtype0] PASSED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-left-g2-idtype1] PASSED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-left-g3-idtype0] PASSED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-left-g3-idtype1] PASSED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-left-g4-idtype0] PASSED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-left-g4-idtype1] PASSED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-left-g5-idtype0] PASSED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-left-g5-idtype1] PASSED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-left-g6-idtype0] PASSED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-left-g6-idtype1] PASSED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-left-g7-idtype0] PASSED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-True-left-g7-idtype1] PASSED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-none-g0-idtype0] PASSED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-none-g0-idtype1] PASSED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-none-g1-idtype0] PASSED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-none-g1-idtype1] PASSED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-none-g2-idtype0] PASSED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-none-g2-idtype1] PASSED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-none-g3-idtype0] PASSED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-none-g3-idtype1] PASSED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-none-g4-idtype0] PASSED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-none-g4-idtype1] PASSED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-none-g5-idtype0] PASSED [ 14%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-none-g5-idtype1] PASSED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-none-g6-idtype0] PASSED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-none-g6-idtype1] PASSED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-none-g7-idtype0] PASSED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-none-g7-idtype1] PASSED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-both-g0-idtype0] PASSED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-both-g0-idtype1] PASSED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-both-g1-idtype0] PASSED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-both-g1-idtype1] PASSED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-both-g2-idtype0] PASSED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-both-g2-idtype1] PASSED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-both-g3-idtype0] PASSED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-both-g3-idtype1] PASSED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-both-g4-idtype0] PASSED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-both-g4-idtype1] PASSED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-both-g5-idtype0] PASSED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-both-g5-idtype1] PASSED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-both-g6-idtype0] PASSED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-both-g6-idtype1] PASSED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-both-g7-idtype0] PASSED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-both-g7-idtype1] PASSED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-right-g0-idtype0] PASSED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-right-g0-idtype1] PASSED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-right-g1-idtype0] PASSED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-right-g1-idtype1] PASSED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-right-g2-idtype0] PASSED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-right-g2-idtype1] PASSED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-right-g3-idtype0] PASSED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-right-g3-idtype1] PASSED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-right-g4-idtype0] PASSED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-right-g4-idtype1] PASSED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-right-g5-idtype0] PASSED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-right-g5-idtype1] PASSED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-right-g6-idtype0] PASSED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-right-g6-idtype1] PASSED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-right-g7-idtype0] PASSED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-right-g7-idtype1] PASSED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-left-g0-idtype0] PASSED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-left-g0-idtype1] PASSED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-left-g1-idtype0] PASSED [ 15%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-left-g1-idtype1] PASSED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-left-g2-idtype0] PASSED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-left-g2-idtype1] PASSED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-left-g3-idtype0] PASSED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-left-g3-idtype1] PASSED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-left-g4-idtype0] PASSED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-left-g4-idtype1] PASSED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-left-g5-idtype0] PASSED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-left-g5-idtype1] PASSED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-left-g6-idtype0] PASSED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-left-g6-idtype1] PASSED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-left-g7-idtype0] PASSED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-True-False-left-g7-idtype1] PASSED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-none-g0-idtype0] PASSED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-none-g0-idtype1] PASSED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-none-g1-idtype0] PASSED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-none-g1-idtype1] PASSED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-none-g2-idtype0] PASSED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-none-g2-idtype1] PASSED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-none-g3-idtype0] PASSED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-none-g3-idtype1] PASSED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-none-g4-idtype0] PASSED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-none-g4-idtype1] PASSED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-none-g5-idtype0] PASSED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-none-g5-idtype1] PASSED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-none-g6-idtype0] PASSED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-none-g6-idtype1] PASSED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-none-g7-idtype0] PASSED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-none-g7-idtype1] PASSED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-both-g0-idtype0] PASSED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-both-g0-idtype1] PASSED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-both-g1-idtype0] PASSED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-both-g1-idtype1] PASSED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-both-g2-idtype0] PASSED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-both-g2-idtype1] PASSED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-both-g3-idtype0] PASSED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-both-g3-idtype1] PASSED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-both-g4-idtype0] PASSED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-both-g4-idtype1] PASSED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-both-g5-idtype0] PASSED [ 16%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-both-g5-idtype1] PASSED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-both-g6-idtype0] PASSED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-both-g6-idtype1] PASSED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-both-g7-idtype0] PASSED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-both-g7-idtype1] PASSED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-right-g0-idtype0] PASSED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-right-g0-idtype1] PASSED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-right-g1-idtype0] PASSED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-right-g1-idtype1] PASSED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-right-g2-idtype0] PASSED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-right-g2-idtype1] PASSED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-right-g3-idtype0] PASSED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-right-g3-idtype1] PASSED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-right-g4-idtype0] PASSED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-right-g4-idtype1] PASSED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-right-g5-idtype0] PASSED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-right-g5-idtype1] PASSED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-right-g6-idtype0] PASSED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-right-g6-idtype1] PASSED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-right-g7-idtype0] PASSED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-right-g7-idtype1] PASSED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-left-g0-idtype0] PASSED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-left-g0-idtype1] PASSED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-left-g1-idtype0] PASSED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-left-g1-idtype1] PASSED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-left-g2-idtype0] PASSED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-left-g2-idtype1] PASSED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-left-g3-idtype0] PASSED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-left-g3-idtype1] PASSED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-left-g4-idtype0] PASSED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-left-g4-idtype1] PASSED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-left-g5-idtype0] PASSED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-left-g5-idtype1] PASSED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-left-g6-idtype0] PASSED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-left-g6-idtype1] PASSED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-left-g7-idtype0] PASSED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-True-left-g7-idtype1] PASSED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-none-g0-idtype0] PASSED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-none-g0-idtype1] PASSED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-none-g1-idtype0] PASSED [ 17%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-none-g1-idtype1] PASSED [ 18%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-none-g2-idtype0] PASSED [ 18%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-none-g2-idtype1] PASSED [ 18%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-none-g3-idtype0] PASSED [ 18%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-none-g3-idtype1] PASSED [ 18%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-none-g4-idtype0] PASSED [ 18%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-none-g4-idtype1] PASSED [ 18%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-none-g5-idtype0] PASSED [ 18%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-none-g5-idtype1] PASSED [ 18%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-none-g6-idtype0] PASSED [ 18%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-none-g6-idtype1] PASSED [ 18%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-none-g7-idtype0] PASSED [ 18%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-none-g7-idtype1] PASSED [ 18%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-both-g0-idtype0] PASSED [ 18%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-both-g0-idtype1] PASSED [ 18%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-both-g1-idtype0] PASSED [ 18%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-both-g1-idtype1] PASSED [ 18%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-both-g2-idtype0] PASSED [ 18%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-both-g2-idtype1] PASSED [ 18%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-both-g3-idtype0] PASSED [ 18%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-both-g3-idtype1] PASSED [ 18%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-both-g4-idtype0] PASSED [ 18%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-both-g4-idtype1] PASSED [ 18%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-both-g5-idtype0] PASSED [ 18%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-both-g5-idtype1] PASSED [ 18%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-both-g6-idtype0] PASSED [ 18%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-both-g6-idtype1] PASSED [ 18%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-both-g7-idtype0] PASSED [ 18%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-both-g7-idtype1] PASSED [ 18%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-right-g0-idtype0] PASSED [ 18%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-right-g0-idtype1] PASSED [ 18%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-right-g1-idtype0] PASSED [ 18%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-right-g1-idtype1] PASSED [ 18%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-right-g2-idtype0] PASSED [ 18%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-right-g2-idtype1] PASSED [ 18%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-right-g3-idtype0] PASSED [ 18%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-right-g3-idtype1] PASSED [ 18%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-right-g4-idtype0] PASSED [ 18%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-right-g4-idtype1] PASSED [ 18%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-right-g5-idtype0] PASSED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-right-g5-idtype1] PASSED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-right-g6-idtype0] PASSED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-right-g6-idtype1] PASSED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-right-g7-idtype0] PASSED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-right-g7-idtype1] PASSED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-left-g0-idtype0] PASSED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-left-g0-idtype1] PASSED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-left-g1-idtype0] PASSED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-left-g1-idtype1] PASSED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-left-g2-idtype0] PASSED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-left-g2-idtype1] PASSED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-left-g3-idtype0] PASSED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-left-g3-idtype1] PASSED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-left-g4-idtype0] PASSED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-left-g4-idtype1] PASSED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-left-g5-idtype0] PASSED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-left-g5-idtype1] PASSED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-left-g6-idtype0] PASSED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-left-g6-idtype1] PASSED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-left-g7-idtype0] PASSED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv[2-False-False-left-g7-idtype1] PASSED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[1-True-True-none-g0-idtype0] PASSED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[1-True-True-none-g0-idtype1] PASSED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[1-True-True-both-g0-idtype0] PASSED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[1-True-True-both-g0-idtype1] PASSED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[1-True-True-right-g0-idtype0] PASSED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[1-True-True-right-g0-idtype1] PASSED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[1-True-False-none-g0-idtype0] PASSED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[1-True-False-none-g0-idtype1] PASSED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[1-True-False-both-g0-idtype0] PASSED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[1-True-False-both-g0-idtype1] PASSED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[1-True-False-right-g0-idtype0] PASSED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[1-True-False-right-g0-idtype1] PASSED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[1-False-True-none-g0-idtype0] PASSED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[1-False-True-none-g0-idtype1] PASSED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[1-False-True-both-g0-idtype0] PASSED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[1-False-True-both-g0-idtype1] PASSED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[1-False-True-right-g0-idtype0] PASSED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[1-False-True-right-g0-idtype1] PASSED [ 19%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[1-False-False-none-g0-idtype0] PASSED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[1-False-False-none-g0-idtype1] PASSED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[1-False-False-both-g0-idtype0] PASSED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[1-False-False-both-g0-idtype1] PASSED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[1-False-False-right-g0-idtype0] PASSED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[1-False-False-right-g0-idtype1] PASSED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[2-True-True-none-g0-idtype0] PASSED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[2-True-True-none-g0-idtype1] PASSED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[2-True-True-both-g0-idtype0] PASSED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[2-True-True-both-g0-idtype1] PASSED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[2-True-True-right-g0-idtype0] PASSED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[2-True-True-right-g0-idtype1] PASSED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[2-True-False-none-g0-idtype0] PASSED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[2-True-False-none-g0-idtype1] PASSED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[2-True-False-both-g0-idtype0] PASSED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[2-True-False-both-g0-idtype1] PASSED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[2-True-False-right-g0-idtype0] PASSED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[2-True-False-right-g0-idtype1] PASSED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[2-False-True-none-g0-idtype0] PASSED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[2-False-True-none-g0-idtype1] PASSED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[2-False-True-both-g0-idtype0] PASSED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[2-False-True-both-g0-idtype1] PASSED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[2-False-True-right-g0-idtype0] PASSED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[2-False-True-right-g0-idtype1] PASSED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[2-False-False-none-g0-idtype0] PASSED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[2-False-False-none-g0-idtype1] PASSED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[2-False-False-both-g0-idtype0] PASSED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[2-False-False-both-g0-idtype1] PASSED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[2-False-False-right-g0-idtype0] PASSED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight[2-False-False-right-g0-idtype1] PASSED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[1-True-True-none-g0-idtype0] PASSED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[1-True-True-none-g0-idtype1] PASSED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[1-True-True-both-g0-idtype0] PASSED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[1-True-True-both-g0-idtype1] PASSED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[1-True-True-right-g0-idtype0] PASSED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[1-True-True-right-g0-idtype1] PASSED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[1-True-False-none-g0-idtype0] PASSED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[1-True-False-none-g0-idtype1] PASSED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[1-True-False-both-g0-idtype0] PASSED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[1-True-False-both-g0-idtype1] PASSED [ 20%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[1-True-False-right-g0-idtype0] PASSED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[1-True-False-right-g0-idtype1] PASSED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[1-False-True-none-g0-idtype0] PASSED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[1-False-True-none-g0-idtype1] PASSED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[1-False-True-both-g0-idtype0] PASSED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[1-False-True-both-g0-idtype1] PASSED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[1-False-True-right-g0-idtype0] PASSED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[1-False-True-right-g0-idtype1] PASSED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[1-False-False-none-g0-idtype0] PASSED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[1-False-False-none-g0-idtype1] PASSED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[1-False-False-both-g0-idtype0] PASSED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[1-False-False-both-g0-idtype1] PASSED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[1-False-False-right-g0-idtype0] PASSED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[1-False-False-right-g0-idtype1] PASSED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[2-True-True-none-g0-idtype0] PASSED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[2-True-True-none-g0-idtype1] PASSED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[2-True-True-both-g0-idtype0] PASSED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[2-True-True-both-g0-idtype1] PASSED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[2-True-True-right-g0-idtype0] PASSED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[2-True-True-right-g0-idtype1] PASSED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[2-True-False-none-g0-idtype0] PASSED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[2-True-False-none-g0-idtype1] PASSED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[2-True-False-both-g0-idtype0] PASSED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[2-True-False-both-g0-idtype1] PASSED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[2-True-False-right-g0-idtype0] PASSED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[2-True-False-right-g0-idtype1] PASSED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[2-False-True-none-g0-idtype0] PASSED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[2-False-True-none-g0-idtype1] PASSED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[2-False-True-both-g0-idtype0] PASSED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[2-False-True-both-g0-idtype1] PASSED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[2-False-True-right-g0-idtype0] PASSED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[2-False-True-right-g0-idtype1] PASSED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[2-False-False-none-g0-idtype0] PASSED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[2-False-False-none-g0-idtype1] PASSED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[2-False-False-both-g0-idtype0] PASSED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[2-False-False-both-g0-idtype1] PASSED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[2-False-False-right-g0-idtype0] PASSED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_e_weight_norm[2-False-False-right-g0-idtype1] PASSED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-True-True-none-g0-idtype0] PASSED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-True-True-none-g0-idtype1] PASSED [ 21%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-True-True-none-g1-idtype0] PASSED [ 22%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-True-True-none-g1-idtype1] PASSED [ 22%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-True-True-both-g0-idtype0] PASSED [ 22%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-True-True-both-g0-idtype1] PASSED [ 22%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-True-True-both-g1-idtype0] PASSED [ 22%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-True-True-both-g1-idtype1] PASSED [ 22%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-True-True-right-g0-idtype0] PASSED [ 22%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-True-True-right-g0-idtype1] PASSED [ 22%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-True-True-right-g1-idtype0] PASSED [ 22%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-True-True-right-g1-idtype1] PASSED [ 22%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-True-False-none-g0-idtype0] PASSED [ 22%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-True-False-none-g0-idtype1] PASSED [ 22%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-True-False-none-g1-idtype0] PASSED [ 22%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-True-False-none-g1-idtype1] PASSED [ 22%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-True-False-both-g0-idtype0] PASSED [ 22%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-True-False-both-g0-idtype1] PASSED [ 22%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-True-False-both-g1-idtype0] PASSED [ 22%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-True-False-both-g1-idtype1] PASSED [ 22%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-True-False-right-g0-idtype0] PASSED [ 22%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-True-False-right-g0-idtype1] PASSED [ 22%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-True-False-right-g1-idtype0] PASSED [ 22%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-True-False-right-g1-idtype1] PASSED [ 22%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-False-True-none-g0-idtype0] PASSED [ 22%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-False-True-none-g0-idtype1] PASSED [ 22%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-False-True-none-g1-idtype0] PASSED [ 22%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-False-True-none-g1-idtype1] PASSED [ 22%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-False-True-both-g0-idtype0] PASSED [ 22%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-False-True-both-g0-idtype1] PASSED [ 22%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-False-True-both-g1-idtype0] PASSED [ 22%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-False-True-both-g1-idtype1] PASSED [ 22%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-False-True-right-g0-idtype0] PASSED [ 22%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-False-True-right-g0-idtype1] PASSED [ 22%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-False-True-right-g1-idtype0] PASSED [ 22%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-False-True-right-g1-idtype1] PASSED [ 22%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-False-False-none-g0-idtype0] PASSED [ 22%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-False-False-none-g0-idtype1] PASSED [ 22%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-False-False-none-g1-idtype0] PASSED [ 22%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-False-False-none-g1-idtype1] PASSED [ 22%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-False-False-both-g0-idtype0] PASSED [ 22%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-False-False-both-g0-idtype1] PASSED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-False-False-both-g1-idtype0] PASSED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-False-False-both-g1-idtype1] PASSED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-False-False-right-g0-idtype0] PASSED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-False-False-right-g0-idtype1] PASSED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-False-False-right-g1-idtype0] PASSED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[1-False-False-right-g1-idtype1] PASSED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-True-True-none-g0-idtype0] PASSED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-True-True-none-g0-idtype1] PASSED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-True-True-none-g1-idtype0] PASSED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-True-True-none-g1-idtype1] PASSED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-True-True-both-g0-idtype0] PASSED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-True-True-both-g0-idtype1] PASSED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-True-True-both-g1-idtype0] PASSED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-True-True-both-g1-idtype1] PASSED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-True-True-right-g0-idtype0] PASSED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-True-True-right-g0-idtype1] PASSED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-True-True-right-g1-idtype0] PASSED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-True-True-right-g1-idtype1] PASSED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-True-False-none-g0-idtype0] PASSED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-True-False-none-g0-idtype1] PASSED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-True-False-none-g1-idtype0] PASSED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-True-False-none-g1-idtype1] PASSED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-True-False-both-g0-idtype0] PASSED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-True-False-both-g0-idtype1] PASSED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-True-False-both-g1-idtype0] PASSED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-True-False-both-g1-idtype1] PASSED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-True-False-right-g0-idtype0] PASSED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-True-False-right-g0-idtype1] PASSED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-True-False-right-g1-idtype0] PASSED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-True-False-right-g1-idtype1] PASSED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-False-True-none-g0-idtype0] PASSED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-False-True-none-g0-idtype1] PASSED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-False-True-none-g1-idtype0] PASSED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-False-True-none-g1-idtype1] PASSED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-False-True-both-g0-idtype0] PASSED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-False-True-both-g0-idtype1] PASSED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-False-True-both-g1-idtype0] PASSED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-False-True-both-g1-idtype1] PASSED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-False-True-right-g0-idtype0] PASSED [ 23%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-False-True-right-g0-idtype1] PASSED [ 24%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-False-True-right-g1-idtype0] PASSED [ 24%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-False-True-right-g1-idtype1] PASSED [ 24%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-False-False-none-g0-idtype0] PASSED [ 24%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-False-False-none-g0-idtype1] PASSED [ 24%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-False-False-none-g1-idtype0] PASSED [ 24%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-False-False-none-g1-idtype1] PASSED [ 24%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-False-False-both-g0-idtype0] PASSED [ 24%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-False-False-both-g0-idtype1] PASSED [ 24%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-False-False-both-g1-idtype0] PASSED [ 24%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-False-False-both-g1-idtype1] PASSED [ 24%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-False-False-right-g0-idtype0] PASSED [ 24%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-False-False-right-g0-idtype1] PASSED [ 24%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-False-False-right-g1-idtype0] PASSED [ 24%]
tests/pytorch/test_nn.py::test_graph_conv_bi[2-False-False-right-g1-idtype1] PASSED [ 24%]
tests/pytorch/test_nn.py::test_tagconv[1] PASSED                         [ 24%]
tests/pytorch/test_nn.py::test_tagconv[2] PASSED                         [ 24%]
tests/pytorch/test_nn.py::test_set2set PASSED                            [ 24%]
tests/pytorch/test_nn.py::test_glob_att_pool PASSED                      [ 24%]
tests/pytorch/test_nn.py::test_simple_pool PASSED                        [ 24%]
tests/pytorch/test_nn.py::test_set_trans PASSED                          [ 24%]
tests/pytorch/test_nn.py::test_rgcn[1-idtype0] PASSED                    [ 24%]
tests/pytorch/test_nn.py::test_rgcn[1-idtype1] PASSED                    [ 24%]
tests/pytorch/test_nn.py::test_rgcn[8-idtype0] PASSED                    [ 24%]
tests/pytorch/test_nn.py::test_rgcn[8-idtype1] PASSED                    [ 24%]
tests/pytorch/test_nn.py::test_rgcn[32-idtype0] PASSED                   [ 24%]
tests/pytorch/test_nn.py::test_rgcn[32-idtype1] PASSED                   [ 24%]
tests/pytorch/test_nn.py::test_rgcn_default_nbasis[1-idtype0] PASSED     [ 24%]
tests/pytorch/test_nn.py::test_rgcn_default_nbasis[1-idtype1] PASSED     [ 24%]
tests/pytorch/test_nn.py::test_rgcn_default_nbasis[10-idtype0] PASSED    [ 24%]
tests/pytorch/test_nn.py::test_rgcn_default_nbasis[10-idtype1] PASSED    [ 24%]
tests/pytorch/test_nn.py::test_rgcn_default_nbasis[40-idtype0] PASSED    [ 24%]
tests/pytorch/test_nn.py::test_rgcn_default_nbasis[40-idtype1] PASSED    [ 24%]
tests/pytorch/test_nn.py::test_gat_conv[1-1-g0-idtype0] PASSED           [ 24%]
tests/pytorch/test_nn.py::test_gat_conv[1-1-g0-idtype1] PASSED           [ 24%]
tests/pytorch/test_nn.py::test_gat_conv[1-1-g1-idtype0] PASSED           [ 24%]
tests/pytorch/test_nn.py::test_gat_conv[1-1-g1-idtype1] PASSED           [ 24%]
tests/pytorch/test_nn.py::test_gat_conv[1-1-g2-idtype0] PASSED           [ 24%]
tests/pytorch/test_nn.py::test_gat_conv[1-1-g2-idtype1] PASSED           [ 24%]
tests/pytorch/test_nn.py::test_gat_conv[1-1-g3-idtype0] PASSED           [ 24%]
tests/pytorch/test_nn.py::test_gat_conv[1-1-g3-idtype1] PASSED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[1-1-g4-idtype0] PASSED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[1-1-g4-idtype1] PASSED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[1-1-g5-idtype0] PASSED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[1-1-g5-idtype1] PASSED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[1-1-g6-idtype0] PASSED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[1-1-g6-idtype1] PASSED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[1-5-g0-idtype0] PASSED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[1-5-g0-idtype1] PASSED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[1-5-g1-idtype0] PASSED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[1-5-g1-idtype1] PASSED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[1-5-g2-idtype0] PASSED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[1-5-g2-idtype1] PASSED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[1-5-g3-idtype0] PASSED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[1-5-g3-idtype1] PASSED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[1-5-g4-idtype0] PASSED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[1-5-g4-idtype1] PASSED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[1-5-g5-idtype0] PASSED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[1-5-g5-idtype1] PASSED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[1-5-g6-idtype0] PASSED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[1-5-g6-idtype1] PASSED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[4-1-g0-idtype0] PASSED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[4-1-g0-idtype1] PASSED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[4-1-g1-idtype0] PASSED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[4-1-g1-idtype1] PASSED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[4-1-g2-idtype0] PASSED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[4-1-g2-idtype1] PASSED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[4-1-g3-idtype0] PASSED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[4-1-g3-idtype1] PASSED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[4-1-g4-idtype0] PASSED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[4-1-g4-idtype1] PASSED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[4-1-g5-idtype0] PASSED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[4-1-g5-idtype1] PASSED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[4-1-g6-idtype0] PASSED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[4-1-g6-idtype1] PASSED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[4-5-g0-idtype0] PASSED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[4-5-g0-idtype1] PASSED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[4-5-g1-idtype0] PASSED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[4-5-g1-idtype1] PASSED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[4-5-g2-idtype0] PASSED           [ 25%]
tests/pytorch/test_nn.py::test_gat_conv[4-5-g2-idtype1] PASSED           [ 26%]
tests/pytorch/test_nn.py::test_gat_conv[4-5-g3-idtype0] PASSED           [ 26%]
tests/pytorch/test_nn.py::test_gat_conv[4-5-g3-idtype1] PASSED           [ 26%]
tests/pytorch/test_nn.py::test_gat_conv[4-5-g4-idtype0] PASSED           [ 26%]
tests/pytorch/test_nn.py::test_gat_conv[4-5-g4-idtype1] PASSED           [ 26%]
tests/pytorch/test_nn.py::test_gat_conv[4-5-g5-idtype0] PASSED           [ 26%]
tests/pytorch/test_nn.py::test_gat_conv[4-5-g5-idtype1] PASSED           [ 26%]
tests/pytorch/test_nn.py::test_gat_conv[4-5-g6-idtype0] PASSED           [ 26%]
tests/pytorch/test_nn.py::test_gat_conv[4-5-g6-idtype1] PASSED           [ 26%]
tests/pytorch/test_nn.py::test_gat_conv_bi[1-1-g0-idtype0] PASSED        [ 26%]
tests/pytorch/test_nn.py::test_gat_conv_bi[1-1-g0-idtype1] PASSED        [ 26%]
tests/pytorch/test_nn.py::test_gat_conv_bi[1-1-g1-idtype0] PASSED        [ 26%]
tests/pytorch/test_nn.py::test_gat_conv_bi[1-1-g1-idtype1] PASSED        [ 26%]
tests/pytorch/test_nn.py::test_gat_conv_bi[1-2-g0-idtype0] PASSED        [ 26%]
tests/pytorch/test_nn.py::test_gat_conv_bi[1-2-g0-idtype1] PASSED        [ 26%]
tests/pytorch/test_nn.py::test_gat_conv_bi[1-2-g1-idtype0] PASSED        [ 26%]
tests/pytorch/test_nn.py::test_gat_conv_bi[1-2-g1-idtype1] PASSED        [ 26%]
tests/pytorch/test_nn.py::test_gat_conv_bi[4-1-g0-idtype0] PASSED        [ 26%]
tests/pytorch/test_nn.py::test_gat_conv_bi[4-1-g0-idtype1] PASSED        [ 26%]
tests/pytorch/test_nn.py::test_gat_conv_bi[4-1-g1-idtype0] PASSED        [ 26%]
tests/pytorch/test_nn.py::test_gat_conv_bi[4-1-g1-idtype1] PASSED        [ 26%]
tests/pytorch/test_nn.py::test_gat_conv_bi[4-2-g0-idtype0] PASSED        [ 26%]
tests/pytorch/test_nn.py::test_gat_conv_bi[4-2-g0-idtype1] PASSED        [ 26%]
tests/pytorch/test_nn.py::test_gat_conv_bi[4-2-g1-idtype0] PASSED        [ 26%]
tests/pytorch/test_nn.py::test_gat_conv_bi[4-2-g1-idtype1] PASSED        [ 26%]
tests/pytorch/test_nn.py::test_gatv2_conv[1-1-g0-idtype0] PASSED         [ 26%]
tests/pytorch/test_nn.py::test_gatv2_conv[1-1-g0-idtype1] PASSED         [ 26%]
tests/pytorch/test_nn.py::test_gatv2_conv[1-1-g1-idtype0] PASSED         [ 26%]
tests/pytorch/test_nn.py::test_gatv2_conv[1-1-g1-idtype1] PASSED         [ 26%]
tests/pytorch/test_nn.py::test_gatv2_conv[1-1-g2-idtype0] PASSED         [ 26%]
tests/pytorch/test_nn.py::test_gatv2_conv[1-1-g2-idtype1] PASSED         [ 26%]
tests/pytorch/test_nn.py::test_gatv2_conv[1-1-g3-idtype0] PASSED         [ 26%]
tests/pytorch/test_nn.py::test_gatv2_conv[1-1-g3-idtype1] PASSED         [ 26%]
tests/pytorch/test_nn.py::test_gatv2_conv[1-1-g4-idtype0] PASSED         [ 26%]
tests/pytorch/test_nn.py::test_gatv2_conv[1-1-g4-idtype1] PASSED         [ 26%]
tests/pytorch/test_nn.py::test_gatv2_conv[1-1-g5-idtype0] PASSED         [ 26%]
tests/pytorch/test_nn.py::test_gatv2_conv[1-1-g5-idtype1] PASSED         [ 26%]
tests/pytorch/test_nn.py::test_gatv2_conv[1-1-g6-idtype0] PASSED         [ 26%]
tests/pytorch/test_nn.py::test_gatv2_conv[1-1-g6-idtype1] PASSED         [ 26%]
tests/pytorch/test_nn.py::test_gatv2_conv[1-5-g0-idtype0] PASSED         [ 26%]
tests/pytorch/test_nn.py::test_gatv2_conv[1-5-g0-idtype1] PASSED         [ 27%]
tests/pytorch/test_nn.py::test_gatv2_conv[1-5-g1-idtype0] PASSED         [ 27%]
tests/pytorch/test_nn.py::test_gatv2_conv[1-5-g1-idtype1] PASSED         [ 27%]
tests/pytorch/test_nn.py::test_gatv2_conv[1-5-g2-idtype0] PASSED         [ 27%]
tests/pytorch/test_nn.py::test_gatv2_conv[1-5-g2-idtype1] PASSED         [ 27%]
tests/pytorch/test_nn.py::test_gatv2_conv[1-5-g3-idtype0] PASSED         [ 27%]
tests/pytorch/test_nn.py::test_gatv2_conv[1-5-g3-idtype1] PASSED         [ 27%]
tests/pytorch/test_nn.py::test_gatv2_conv[1-5-g4-idtype0] PASSED         [ 27%]
tests/pytorch/test_nn.py::test_gatv2_conv[1-5-g4-idtype1] PASSED         [ 27%]
tests/pytorch/test_nn.py::test_gatv2_conv[1-5-g5-idtype0] PASSED         [ 27%]
tests/pytorch/test_nn.py::test_gatv2_conv[1-5-g5-idtype1] PASSED         [ 27%]
tests/pytorch/test_nn.py::test_gatv2_conv[1-5-g6-idtype0] PASSED         [ 27%]
tests/pytorch/test_nn.py::test_gatv2_conv[1-5-g6-idtype1] PASSED         [ 27%]
tests/pytorch/test_nn.py::test_gatv2_conv[4-1-g0-idtype0] PASSED         [ 27%]
tests/pytorch/test_nn.py::test_gatv2_conv[4-1-g0-idtype1] PASSED         [ 27%]
tests/pytorch/test_nn.py::test_gatv2_conv[4-1-g1-idtype0] PASSED         [ 27%]
tests/pytorch/test_nn.py::test_gatv2_conv[4-1-g1-idtype1] PASSED         [ 27%]
tests/pytorch/test_nn.py::test_gatv2_conv[4-1-g2-idtype0] PASSED         [ 27%]
tests/pytorch/test_nn.py::test_gatv2_conv[4-1-g2-idtype1] PASSED         [ 27%]
tests/pytorch/test_nn.py::test_gatv2_conv[4-1-g3-idtype0] PASSED         [ 27%]
tests/pytorch/test_nn.py::test_gatv2_conv[4-1-g3-idtype1] PASSED         [ 27%]
tests/pytorch/test_nn.py::test_gatv2_conv[4-1-g4-idtype0] PASSED         [ 27%]
tests/pytorch/test_nn.py::test_gatv2_conv[4-1-g4-idtype1] PASSED         [ 27%]
tests/pytorch/test_nn.py::test_gatv2_conv[4-1-g5-idtype0] PASSED         [ 27%]
tests/pytorch/test_nn.py::test_gatv2_conv[4-1-g5-idtype1] PASSED         [ 27%]
tests/pytorch/test_nn.py::test_gatv2_conv[4-1-g6-idtype0] PASSED         [ 27%]
tests/pytorch/test_nn.py::test_gatv2_conv[4-1-g6-idtype1] PASSED         [ 27%]
tests/pytorch/test_nn.py::test_gatv2_conv[4-5-g0-idtype0] PASSED         [ 27%]
tests/pytorch/test_nn.py::test_gatv2_conv[4-5-g0-idtype1] PASSED         [ 27%]
tests/pytorch/test_nn.py::test_gatv2_conv[4-5-g1-idtype0] PASSED         [ 27%]
tests/pytorch/test_nn.py::test_gatv2_conv[4-5-g1-idtype1] PASSED         [ 27%]
tests/pytorch/test_nn.py::test_gatv2_conv[4-5-g2-idtype0] PASSED         [ 27%]
tests/pytorch/test_nn.py::test_gatv2_conv[4-5-g2-idtype1] PASSED         [ 27%]
tests/pytorch/test_nn.py::test_gatv2_conv[4-5-g3-idtype0] PASSED         [ 27%]
tests/pytorch/test_nn.py::test_gatv2_conv[4-5-g3-idtype1] PASSED         [ 27%]
tests/pytorch/test_nn.py::test_gatv2_conv[4-5-g4-idtype0] PASSED         [ 27%]
tests/pytorch/test_nn.py::test_gatv2_conv[4-5-g4-idtype1] PASSED         [ 27%]
tests/pytorch/test_nn.py::test_gatv2_conv[4-5-g5-idtype0] PASSED         [ 27%]
tests/pytorch/test_nn.py::test_gatv2_conv[4-5-g5-idtype1] PASSED         [ 27%]
tests/pytorch/test_nn.py::test_gatv2_conv[4-5-g6-idtype0] PASSED         [ 28%]
tests/pytorch/test_nn.py::test_gatv2_conv[4-5-g6-idtype1] PASSED         [ 28%]
tests/pytorch/test_nn.py::test_gatv2_conv_bi[1-1-g0-idtype0] PASSED      [ 28%]
tests/pytorch/test_nn.py::test_gatv2_conv_bi[1-1-g0-idtype1] PASSED      [ 28%]
tests/pytorch/test_nn.py::test_gatv2_conv_bi[1-1-g1-idtype0] PASSED      [ 28%]
tests/pytorch/test_nn.py::test_gatv2_conv_bi[1-1-g1-idtype1] PASSED      [ 28%]
tests/pytorch/test_nn.py::test_gatv2_conv_bi[1-2-g0-idtype0] PASSED      [ 28%]
tests/pytorch/test_nn.py::test_gatv2_conv_bi[1-2-g0-idtype1] PASSED      [ 28%]
tests/pytorch/test_nn.py::test_gatv2_conv_bi[1-2-g1-idtype0] PASSED      [ 28%]
tests/pytorch/test_nn.py::test_gatv2_conv_bi[1-2-g1-idtype1] PASSED      [ 28%]
tests/pytorch/test_nn.py::test_gatv2_conv_bi[4-1-g0-idtype0] PASSED      [ 28%]
tests/pytorch/test_nn.py::test_gatv2_conv_bi[4-1-g0-idtype1] PASSED      [ 28%]
tests/pytorch/test_nn.py::test_gatv2_conv_bi[4-1-g1-idtype0] PASSED      [ 28%]
tests/pytorch/test_nn.py::test_gatv2_conv_bi[4-1-g1-idtype1] PASSED      [ 28%]
tests/pytorch/test_nn.py::test_gatv2_conv_bi[4-2-g0-idtype0] PASSED      [ 28%]
tests/pytorch/test_nn.py::test_gatv2_conv_bi[4-2-g0-idtype1] PASSED      [ 28%]
tests/pytorch/test_nn.py::test_gatv2_conv_bi[4-2-g1-idtype0] PASSED      [ 28%]
tests/pytorch/test_nn.py::test_gatv2_conv_bi[4-2-g1-idtype1] PASSED      [ 28%]
tests/pytorch/test_nn.py::test_egat_conv[1-1-1-g0-idtype0] PASSED        [ 28%]
tests/pytorch/test_nn.py::test_egat_conv[1-1-1-g0-idtype1] PASSED        [ 28%]
tests/pytorch/test_nn.py::test_egat_conv[1-1-1-g1-idtype0] PASSED        [ 28%]
tests/pytorch/test_nn.py::test_egat_conv[1-1-1-g1-idtype1] PASSED        [ 28%]
tests/pytorch/test_nn.py::test_egat_conv[1-1-1-g2-idtype0] PASSED        [ 28%]
tests/pytorch/test_nn.py::test_egat_conv[1-1-1-g2-idtype1] PASSED        [ 28%]
tests/pytorch/test_nn.py::test_egat_conv[1-1-1-g3-idtype0] PASSED        [ 28%]
tests/pytorch/test_nn.py::test_egat_conv[1-1-1-g3-idtype1] PASSED        [ 28%]
tests/pytorch/test_nn.py::test_egat_conv[1-1-1-g4-idtype0] PASSED        [ 28%]
tests/pytorch/test_nn.py::test_egat_conv[1-1-1-g4-idtype1] PASSED        [ 28%]
tests/pytorch/test_nn.py::test_egat_conv[1-1-1-g5-idtype0] PASSED        [ 28%]
tests/pytorch/test_nn.py::test_egat_conv[1-1-1-g5-idtype1] PASSED        [ 28%]
tests/pytorch/test_nn.py::test_egat_conv[1-1-5-g0-idtype0] PASSED        [ 28%]
tests/pytorch/test_nn.py::test_egat_conv[1-1-5-g0-idtype1] PASSED        [ 28%]
tests/pytorch/test_nn.py::test_egat_conv[1-1-5-g1-idtype0] PASSED        [ 28%]
tests/pytorch/test_nn.py::test_egat_conv[1-1-5-g1-idtype1] PASSED        [ 28%]
tests/pytorch/test_nn.py::test_egat_conv[1-1-5-g2-idtype0] PASSED        [ 28%]
tests/pytorch/test_nn.py::test_egat_conv[1-1-5-g2-idtype1] PASSED        [ 28%]
tests/pytorch/test_nn.py::test_egat_conv[1-1-5-g3-idtype0] PASSED        [ 28%]
tests/pytorch/test_nn.py::test_egat_conv[1-1-5-g3-idtype1] PASSED        [ 28%]
tests/pytorch/test_nn.py::test_egat_conv[1-1-5-g4-idtype0] PASSED        [ 28%]
tests/pytorch/test_nn.py::test_egat_conv[1-1-5-g4-idtype1] PASSED        [ 28%]
tests/pytorch/test_nn.py::test_egat_conv[1-1-5-g5-idtype0] PASSED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[1-1-5-g5-idtype1] PASSED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[1-5-1-g0-idtype0] PASSED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[1-5-1-g0-idtype1] PASSED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[1-5-1-g1-idtype0] PASSED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[1-5-1-g1-idtype1] PASSED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[1-5-1-g2-idtype0] PASSED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[1-5-1-g2-idtype1] PASSED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[1-5-1-g3-idtype0] PASSED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[1-5-1-g3-idtype1] PASSED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[1-5-1-g4-idtype0] PASSED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[1-5-1-g4-idtype1] PASSED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[1-5-1-g5-idtype0] PASSED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[1-5-1-g5-idtype1] PASSED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[1-5-5-g0-idtype0] PASSED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[1-5-5-g0-idtype1] PASSED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[1-5-5-g1-idtype0] PASSED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[1-5-5-g1-idtype1] PASSED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[1-5-5-g2-idtype0] PASSED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[1-5-5-g2-idtype1] PASSED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[1-5-5-g3-idtype0] PASSED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[1-5-5-g3-idtype1] PASSED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[1-5-5-g4-idtype0] PASSED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[1-5-5-g4-idtype1] PASSED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[1-5-5-g5-idtype0] PASSED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[1-5-5-g5-idtype1] PASSED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[4-1-1-g0-idtype0] PASSED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[4-1-1-g0-idtype1] PASSED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[4-1-1-g1-idtype0] PASSED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[4-1-1-g1-idtype1] PASSED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[4-1-1-g2-idtype0] PASSED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[4-1-1-g2-idtype1] PASSED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[4-1-1-g3-idtype0] PASSED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[4-1-1-g3-idtype1] PASSED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[4-1-1-g4-idtype0] PASSED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[4-1-1-g4-idtype1] PASSED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[4-1-1-g5-idtype0] PASSED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[4-1-1-g5-idtype1] PASSED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[4-1-5-g0-idtype0] PASSED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[4-1-5-g0-idtype1] PASSED        [ 29%]
tests/pytorch/test_nn.py::test_egat_conv[4-1-5-g1-idtype0] PASSED        [ 30%]
tests/pytorch/test_nn.py::test_egat_conv[4-1-5-g1-idtype1] PASSED        [ 30%]
tests/pytorch/test_nn.py::test_egat_conv[4-1-5-g2-idtype0] PASSED        [ 30%]
tests/pytorch/test_nn.py::test_egat_conv[4-1-5-g2-idtype1] PASSED        [ 30%]
tests/pytorch/test_nn.py::test_egat_conv[4-1-5-g3-idtype0] PASSED        [ 30%]
tests/pytorch/test_nn.py::test_egat_conv[4-1-5-g3-idtype1] PASSED        [ 30%]
tests/pytorch/test_nn.py::test_egat_conv[4-1-5-g4-idtype0] PASSED        [ 30%]
tests/pytorch/test_nn.py::test_egat_conv[4-1-5-g4-idtype1] PASSED        [ 30%]
tests/pytorch/test_nn.py::test_egat_conv[4-1-5-g5-idtype0] PASSED        [ 30%]
tests/pytorch/test_nn.py::test_egat_conv[4-1-5-g5-idtype1] PASSED        [ 30%]
tests/pytorch/test_nn.py::test_egat_conv[4-5-1-g0-idtype0] PASSED        [ 30%]
tests/pytorch/test_nn.py::test_egat_conv[4-5-1-g0-idtype1] PASSED        [ 30%]
tests/pytorch/test_nn.py::test_egat_conv[4-5-1-g1-idtype0] PASSED        [ 30%]
tests/pytorch/test_nn.py::test_egat_conv[4-5-1-g1-idtype1] PASSED        [ 30%]
tests/pytorch/test_nn.py::test_egat_conv[4-5-1-g2-idtype0] PASSED        [ 30%]
tests/pytorch/test_nn.py::test_egat_conv[4-5-1-g2-idtype1] PASSED        [ 30%]
tests/pytorch/test_nn.py::test_egat_conv[4-5-1-g3-idtype0] PASSED        [ 30%]
tests/pytorch/test_nn.py::test_egat_conv[4-5-1-g3-idtype1] PASSED        [ 30%]
tests/pytorch/test_nn.py::test_egat_conv[4-5-1-g4-idtype0] PASSED        [ 30%]
tests/pytorch/test_nn.py::test_egat_conv[4-5-1-g4-idtype1] PASSED        [ 30%]
tests/pytorch/test_nn.py::test_egat_conv[4-5-1-g5-idtype0] PASSED        [ 30%]
tests/pytorch/test_nn.py::test_egat_conv[4-5-1-g5-idtype1] PASSED        [ 30%]
tests/pytorch/test_nn.py::test_egat_conv[4-5-5-g0-idtype0] PASSED        [ 30%]
tests/pytorch/test_nn.py::test_egat_conv[4-5-5-g0-idtype1] PASSED        [ 30%]
tests/pytorch/test_nn.py::test_egat_conv[4-5-5-g1-idtype0] PASSED        [ 30%]
tests/pytorch/test_nn.py::test_egat_conv[4-5-5-g1-idtype1] PASSED        [ 30%]
tests/pytorch/test_nn.py::test_egat_conv[4-5-5-g2-idtype0] PASSED        [ 30%]
tests/pytorch/test_nn.py::test_egat_conv[4-5-5-g2-idtype1] PASSED        [ 30%]
tests/pytorch/test_nn.py::test_egat_conv[4-5-5-g3-idtype0] PASSED        [ 30%]
tests/pytorch/test_nn.py::test_egat_conv[4-5-5-g3-idtype1] PASSED        [ 30%]
tests/pytorch/test_nn.py::test_egat_conv[4-5-5-g4-idtype0] PASSED        [ 30%]
tests/pytorch/test_nn.py::test_egat_conv[4-5-5-g4-idtype1] PASSED        [ 30%]
tests/pytorch/test_nn.py::test_egat_conv[4-5-5-g5-idtype0] PASSED        [ 30%]
tests/pytorch/test_nn.py::test_egat_conv[4-5-5-g5-idtype1] PASSED        [ 30%]
tests/pytorch/test_nn.py::test_egat_conv_bi[1-1-1-g0-idtype0] PASSED     [ 30%]
tests/pytorch/test_nn.py::test_egat_conv_bi[1-1-1-g0-idtype1] PASSED     [ 30%]
tests/pytorch/test_nn.py::test_egat_conv_bi[1-1-1-g1-idtype0] PASSED     [ 30%]
tests/pytorch/test_nn.py::test_egat_conv_bi[1-1-1-g1-idtype1] PASSED     [ 30%]
tests/pytorch/test_nn.py::test_egat_conv_bi[1-1-5-g0-idtype0] PASSED     [ 30%]
tests/pytorch/test_nn.py::test_egat_conv_bi[1-1-5-g0-idtype1] PASSED     [ 30%]
tests/pytorch/test_nn.py::test_egat_conv_bi[1-1-5-g1-idtype0] PASSED     [ 31%]
tests/pytorch/test_nn.py::test_egat_conv_bi[1-1-5-g1-idtype1] PASSED     [ 31%]
tests/pytorch/test_nn.py::test_egat_conv_bi[1-5-1-g0-idtype0] PASSED     [ 31%]
tests/pytorch/test_nn.py::test_egat_conv_bi[1-5-1-g0-idtype1] PASSED     [ 31%]
tests/pytorch/test_nn.py::test_egat_conv_bi[1-5-1-g1-idtype0] PASSED     [ 31%]
tests/pytorch/test_nn.py::test_egat_conv_bi[1-5-1-g1-idtype1] PASSED     [ 31%]
tests/pytorch/test_nn.py::test_egat_conv_bi[1-5-5-g0-idtype0] PASSED     [ 31%]
tests/pytorch/test_nn.py::test_egat_conv_bi[1-5-5-g0-idtype1] PASSED     [ 31%]
tests/pytorch/test_nn.py::test_egat_conv_bi[1-5-5-g1-idtype0] PASSED     [ 31%]
tests/pytorch/test_nn.py::test_egat_conv_bi[1-5-5-g1-idtype1] PASSED     [ 31%]
tests/pytorch/test_nn.py::test_egat_conv_bi[4-1-1-g0-idtype0] PASSED     [ 31%]
tests/pytorch/test_nn.py::test_egat_conv_bi[4-1-1-g0-idtype1] PASSED     [ 31%]
tests/pytorch/test_nn.py::test_egat_conv_bi[4-1-1-g1-idtype0] PASSED     [ 31%]
tests/pytorch/test_nn.py::test_egat_conv_bi[4-1-1-g1-idtype1] PASSED     [ 31%]
tests/pytorch/test_nn.py::test_egat_conv_bi[4-1-5-g0-idtype0] PASSED     [ 31%]
tests/pytorch/test_nn.py::test_egat_conv_bi[4-1-5-g0-idtype1] PASSED     [ 31%]
tests/pytorch/test_nn.py::test_egat_conv_bi[4-1-5-g1-idtype0] PASSED     [ 31%]
tests/pytorch/test_nn.py::test_egat_conv_bi[4-1-5-g1-idtype1] PASSED     [ 31%]
tests/pytorch/test_nn.py::test_egat_conv_bi[4-5-1-g0-idtype0] PASSED     [ 31%]
tests/pytorch/test_nn.py::test_egat_conv_bi[4-5-1-g0-idtype1] PASSED     [ 31%]
tests/pytorch/test_nn.py::test_egat_conv_bi[4-5-1-g1-idtype0] PASSED     [ 31%]
tests/pytorch/test_nn.py::test_egat_conv_bi[4-5-1-g1-idtype1] PASSED     [ 31%]
tests/pytorch/test_nn.py::test_egat_conv_bi[4-5-5-g0-idtype0] PASSED     [ 31%]
tests/pytorch/test_nn.py::test_egat_conv_bi[4-5-5-g0-idtype1] PASSED     [ 31%]
tests/pytorch/test_nn.py::test_egat_conv_bi[4-5-5-g1-idtype0] PASSED     [ 31%]
tests/pytorch/test_nn.py::test_egat_conv_bi[4-5-5-g1-idtype1] PASSED     [ 31%]
tests/pytorch/test_nn.py::test_sage_conv[mean-g0-idtype0] PASSED         [ 31%]
tests/pytorch/test_nn.py::test_sage_conv[mean-g0-idtype1] PASSED         [ 31%]
tests/pytorch/test_nn.py::test_sage_conv[mean-g1-idtype0] PASSED         [ 31%]
tests/pytorch/test_nn.py::test_sage_conv[mean-g1-idtype1] PASSED         [ 31%]
tests/pytorch/test_nn.py::test_sage_conv[mean-g2-idtype0] PASSED         [ 31%]
tests/pytorch/test_nn.py::test_sage_conv[mean-g2-idtype1] PASSED         [ 31%]
tests/pytorch/test_nn.py::test_sage_conv[mean-g3-idtype0] PASSED         [ 31%]
tests/pytorch/test_nn.py::test_sage_conv[mean-g3-idtype1] PASSED         [ 31%]
tests/pytorch/test_nn.py::test_sage_conv[mean-g4-idtype0] PASSED         [ 31%]
tests/pytorch/test_nn.py::test_sage_conv[mean-g4-idtype1] PASSED         [ 31%]
tests/pytorch/test_nn.py::test_sage_conv[mean-g5-idtype0] PASSED         [ 31%]
tests/pytorch/test_nn.py::test_sage_conv[mean-g5-idtype1] PASSED         [ 31%]
tests/pytorch/test_nn.py::test_sage_conv[mean-g6-idtype0] PASSED         [ 31%]
tests/pytorch/test_nn.py::test_sage_conv[mean-g6-idtype1] PASSED         [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[mean-g7-idtype0] PASSED         [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[mean-g7-idtype1] PASSED         [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[pool-g0-idtype0] PASSED         [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[pool-g0-idtype1] PASSED         [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[pool-g1-idtype0] PASSED         [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[pool-g1-idtype1] PASSED         [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[pool-g2-idtype0] PASSED         [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[pool-g2-idtype1] PASSED         [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[pool-g3-idtype0] PASSED         [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[pool-g3-idtype1] PASSED         [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[pool-g4-idtype0] PASSED         [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[pool-g4-idtype1] PASSED         [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[pool-g5-idtype0] PASSED         [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[pool-g5-idtype1] PASSED         [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[pool-g6-idtype0] PASSED         [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[pool-g6-idtype1] PASSED         [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[pool-g7-idtype0] PASSED         [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[pool-g7-idtype1] PASSED         [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[gcn-g0-idtype0] PASSED          [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[gcn-g0-idtype1] PASSED          [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[gcn-g1-idtype0] PASSED          [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[gcn-g1-idtype1] PASSED          [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[gcn-g2-idtype0] PASSED          [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[gcn-g2-idtype1] PASSED          [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[gcn-g3-idtype0] PASSED          [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[gcn-g3-idtype1] PASSED          [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[gcn-g4-idtype0] PASSED          [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[gcn-g4-idtype1] PASSED          [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[gcn-g5-idtype0] PASSED          [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[gcn-g5-idtype1] PASSED          [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[gcn-g6-idtype0] PASSED          [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[gcn-g6-idtype1] PASSED          [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[gcn-g7-idtype0] PASSED          [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[gcn-g7-idtype1] PASSED          [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[lstm-g0-idtype0] PASSED         [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[lstm-g0-idtype1] PASSED         [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[lstm-g1-idtype0] PASSED         [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[lstm-g1-idtype1] PASSED         [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[lstm-g2-idtype0] PASSED         [ 32%]
tests/pytorch/test_nn.py::test_sage_conv[lstm-g2-idtype1] PASSED         [ 33%]
tests/pytorch/test_nn.py::test_sage_conv[lstm-g3-idtype0] PASSED         [ 33%]
tests/pytorch/test_nn.py::test_sage_conv[lstm-g3-idtype1] PASSED         [ 33%]
tests/pytorch/test_nn.py::test_sage_conv[lstm-g4-idtype0] PASSED         [ 33%]
tests/pytorch/test_nn.py::test_sage_conv[lstm-g4-idtype1] PASSED         [ 33%]
tests/pytorch/test_nn.py::test_sage_conv[lstm-g5-idtype0] PASSED         [ 33%]
tests/pytorch/test_nn.py::test_sage_conv[lstm-g5-idtype1] PASSED         [ 33%]
tests/pytorch/test_nn.py::test_sage_conv[lstm-g6-idtype0] PASSED         [ 33%]
tests/pytorch/test_nn.py::test_sage_conv[lstm-g6-idtype1] PASSED         [ 33%]
tests/pytorch/test_nn.py::test_sage_conv[lstm-g7-idtype0] PASSED         [ 33%]
tests/pytorch/test_nn.py::test_sage_conv[lstm-g7-idtype1] PASSED         [ 33%]
tests/pytorch/test_nn.py::test_sage_conv_bi[1-mean-g0-idtype0] PASSED    [ 33%]
tests/pytorch/test_nn.py::test_sage_conv_bi[1-mean-g0-idtype1] PASSED    [ 33%]
tests/pytorch/test_nn.py::test_sage_conv_bi[1-mean-g1-idtype0] PASSED    [ 33%]
tests/pytorch/test_nn.py::test_sage_conv_bi[1-mean-g1-idtype1] PASSED    [ 33%]
tests/pytorch/test_nn.py::test_sage_conv_bi[1-mean-g2-idtype0] PASSED    [ 33%]
tests/pytorch/test_nn.py::test_sage_conv_bi[1-mean-g2-idtype1] PASSED    [ 33%]
tests/pytorch/test_nn.py::test_sage_conv_bi[1-pool-g0-idtype0] PASSED    [ 33%]
tests/pytorch/test_nn.py::test_sage_conv_bi[1-pool-g0-idtype1] PASSED    [ 33%]
tests/pytorch/test_nn.py::test_sage_conv_bi[1-pool-g1-idtype0] PASSED    [ 33%]
tests/pytorch/test_nn.py::test_sage_conv_bi[1-pool-g1-idtype1] PASSED    [ 33%]
tests/pytorch/test_nn.py::test_sage_conv_bi[1-pool-g2-idtype0] PASSED    [ 33%]
tests/pytorch/test_nn.py::test_sage_conv_bi[1-pool-g2-idtype1] PASSED    [ 33%]
tests/pytorch/test_nn.py::test_sage_conv_bi[1-gcn-g0-idtype0] PASSED     [ 33%]
tests/pytorch/test_nn.py::test_sage_conv_bi[1-gcn-g0-idtype1] PASSED     [ 33%]
tests/pytorch/test_nn.py::test_sage_conv_bi[1-gcn-g1-idtype0] PASSED     [ 33%]
tests/pytorch/test_nn.py::test_sage_conv_bi[1-gcn-g1-idtype1] PASSED     [ 33%]
tests/pytorch/test_nn.py::test_sage_conv_bi[1-gcn-g2-idtype0] PASSED     [ 33%]
tests/pytorch/test_nn.py::test_sage_conv_bi[1-gcn-g2-idtype1] PASSED     [ 33%]
tests/pytorch/test_nn.py::test_sage_conv_bi[1-lstm-g0-idtype0] PASSED    [ 33%]
tests/pytorch/test_nn.py::test_sage_conv_bi[1-lstm-g0-idtype1] PASSED    [ 33%]
tests/pytorch/test_nn.py::test_sage_conv_bi[1-lstm-g1-idtype0] PASSED    [ 33%]
tests/pytorch/test_nn.py::test_sage_conv_bi[1-lstm-g1-idtype1] PASSED    [ 33%]
tests/pytorch/test_nn.py::test_sage_conv_bi[1-lstm-g2-idtype0] PASSED    [ 33%]
tests/pytorch/test_nn.py::test_sage_conv_bi[1-lstm-g2-idtype1] PASSED    [ 33%]
tests/pytorch/test_nn.py::test_sage_conv_bi[2-mean-g0-idtype0] PASSED    [ 33%]
tests/pytorch/test_nn.py::test_sage_conv_bi[2-mean-g0-idtype1] PASSED    [ 33%]
tests/pytorch/test_nn.py::test_sage_conv_bi[2-mean-g1-idtype0] PASSED    [ 33%]
tests/pytorch/test_nn.py::test_sage_conv_bi[2-mean-g1-idtype1] PASSED    [ 33%]
tests/pytorch/test_nn.py::test_sage_conv_bi[2-mean-g2-idtype0] PASSED    [ 33%]
tests/pytorch/test_nn.py::test_sage_conv_bi[2-mean-g2-idtype1] PASSED    [ 34%]
tests/pytorch/test_nn.py::test_sage_conv_bi[2-pool-g0-idtype0] PASSED    [ 34%]
tests/pytorch/test_nn.py::test_sage_conv_bi[2-pool-g0-idtype1] PASSED    [ 34%]
tests/pytorch/test_nn.py::test_sage_conv_bi[2-pool-g1-idtype0] PASSED    [ 34%]
tests/pytorch/test_nn.py::test_sage_conv_bi[2-pool-g1-idtype1] PASSED    [ 34%]
tests/pytorch/test_nn.py::test_sage_conv_bi[2-pool-g2-idtype0] PASSED    [ 34%]
tests/pytorch/test_nn.py::test_sage_conv_bi[2-pool-g2-idtype1] PASSED    [ 34%]
tests/pytorch/test_nn.py::test_sage_conv_bi[2-gcn-g0-idtype0] PASSED     [ 34%]
tests/pytorch/test_nn.py::test_sage_conv_bi[2-gcn-g0-idtype1] PASSED     [ 34%]
tests/pytorch/test_nn.py::test_sage_conv_bi[2-gcn-g1-idtype0] PASSED     [ 34%]
tests/pytorch/test_nn.py::test_sage_conv_bi[2-gcn-g1-idtype1] PASSED     [ 34%]
tests/pytorch/test_nn.py::test_sage_conv_bi[2-gcn-g2-idtype0] PASSED     [ 34%]
tests/pytorch/test_nn.py::test_sage_conv_bi[2-gcn-g2-idtype1] PASSED     [ 34%]
tests/pytorch/test_nn.py::test_sage_conv_bi[2-lstm-g0-idtype0] PASSED    [ 34%]
tests/pytorch/test_nn.py::test_sage_conv_bi[2-lstm-g0-idtype1] PASSED    [ 34%]
tests/pytorch/test_nn.py::test_sage_conv_bi[2-lstm-g1-idtype0] PASSED    [ 34%]
tests/pytorch/test_nn.py::test_sage_conv_bi[2-lstm-g1-idtype1] PASSED    [ 34%]
tests/pytorch/test_nn.py::test_sage_conv_bi[2-lstm-g2-idtype0] PASSED    [ 34%]
tests/pytorch/test_nn.py::test_sage_conv_bi[2-lstm-g2-idtype1] PASSED    [ 34%]
tests/pytorch/test_nn.py::test_sage_conv2[1-idtype0] PASSED              [ 34%]
tests/pytorch/test_nn.py::test_sage_conv2[1-idtype1] PASSED              [ 34%]
tests/pytorch/test_nn.py::test_sage_conv2[2-idtype0] PASSED              [ 34%]
tests/pytorch/test_nn.py::test_sage_conv2[2-idtype1] PASSED              [ 34%]
tests/pytorch/test_nn.py::test_sgc_conv[1-g0-idtype0] PASSED             [ 34%]
tests/pytorch/test_nn.py::test_sgc_conv[1-g0-idtype1] PASSED             [ 34%]
tests/pytorch/test_nn.py::test_sgc_conv[1-g1-idtype0] PASSED             [ 34%]
tests/pytorch/test_nn.py::test_sgc_conv[1-g1-idtype1] PASSED             [ 34%]
tests/pytorch/test_nn.py::test_sgc_conv[1-g2-idtype0] PASSED             [ 34%]
tests/pytorch/test_nn.py::test_sgc_conv[1-g2-idtype1] PASSED             [ 34%]
tests/pytorch/test_nn.py::test_sgc_conv[1-g3-idtype0] PASSED             [ 34%]
tests/pytorch/test_nn.py::test_sgc_conv[1-g3-idtype1] PASSED             [ 34%]
tests/pytorch/test_nn.py::test_sgc_conv[1-g4-idtype0] PASSED             [ 34%]
tests/pytorch/test_nn.py::test_sgc_conv[1-g4-idtype1] PASSED             [ 34%]
tests/pytorch/test_nn.py::test_sgc_conv[1-g5-idtype0] PASSED             [ 34%]
tests/pytorch/test_nn.py::test_sgc_conv[1-g5-idtype1] PASSED             [ 34%]
tests/pytorch/test_nn.py::test_sgc_conv[2-g0-idtype0] PASSED             [ 34%]
tests/pytorch/test_nn.py::test_sgc_conv[2-g0-idtype1] PASSED             [ 34%]
tests/pytorch/test_nn.py::test_sgc_conv[2-g1-idtype0] PASSED             [ 34%]
tests/pytorch/test_nn.py::test_sgc_conv[2-g1-idtype1] PASSED             [ 34%]
tests/pytorch/test_nn.py::test_sgc_conv[2-g2-idtype0] PASSED             [ 34%]
tests/pytorch/test_nn.py::test_sgc_conv[2-g2-idtype1] PASSED             [ 35%]
tests/pytorch/test_nn.py::test_sgc_conv[2-g3-idtype0] PASSED             [ 35%]
tests/pytorch/test_nn.py::test_sgc_conv[2-g3-idtype1] PASSED             [ 35%]
tests/pytorch/test_nn.py::test_sgc_conv[2-g4-idtype0] PASSED             [ 35%]
tests/pytorch/test_nn.py::test_sgc_conv[2-g4-idtype1] PASSED             [ 35%]
tests/pytorch/test_nn.py::test_sgc_conv[2-g5-idtype0] PASSED             [ 35%]
tests/pytorch/test_nn.py::test_sgc_conv[2-g5-idtype1] PASSED             [ 35%]
tests/pytorch/test_nn.py::test_appnp_conv[g0-idtype0] PASSED             [ 35%]
tests/pytorch/test_nn.py::test_appnp_conv[g0-idtype1] PASSED             [ 35%]
tests/pytorch/test_nn.py::test_appnp_conv[g1-idtype0] PASSED             [ 35%]
tests/pytorch/test_nn.py::test_appnp_conv[g1-idtype1] PASSED             [ 35%]
tests/pytorch/test_nn.py::test_appnp_conv[g2-idtype0] PASSED             [ 35%]
tests/pytorch/test_nn.py::test_appnp_conv[g2-idtype1] PASSED             [ 35%]
tests/pytorch/test_nn.py::test_appnp_conv[g3-idtype0] PASSED             [ 35%]
tests/pytorch/test_nn.py::test_appnp_conv[g3-idtype1] PASSED             [ 35%]
tests/pytorch/test_nn.py::test_appnp_conv[g4-idtype0] PASSED             [ 35%]
tests/pytorch/test_nn.py::test_appnp_conv[g4-idtype1] PASSED             [ 35%]
tests/pytorch/test_nn.py::test_appnp_conv[g5-idtype0] PASSED             [ 35%]
tests/pytorch/test_nn.py::test_appnp_conv[g5-idtype1] PASSED             [ 35%]
tests/pytorch/test_nn.py::test_appnp_conv_e_weight[g0-idtype0] PASSED    [ 35%]
tests/pytorch/test_nn.py::test_appnp_conv_e_weight[g0-idtype1] PASSED    [ 35%]
tests/pytorch/test_nn.py::test_appnp_conv_e_weight[g1-idtype0] PASSED    [ 35%]
tests/pytorch/test_nn.py::test_appnp_conv_e_weight[g1-idtype1] PASSED    [ 35%]
tests/pytorch/test_nn.py::test_appnp_conv_e_weight[g2-idtype0] PASSED    [ 35%]
tests/pytorch/test_nn.py::test_appnp_conv_e_weight[g2-idtype1] PASSED    [ 35%]
tests/pytorch/test_nn.py::test_appnp_conv_e_weight[g3-idtype0] PASSED    [ 35%]
tests/pytorch/test_nn.py::test_appnp_conv_e_weight[g3-idtype1] PASSED    [ 35%]
tests/pytorch/test_nn.py::test_appnp_conv_e_weight[g4-idtype0] PASSED    [ 35%]
tests/pytorch/test_nn.py::test_appnp_conv_e_weight[g4-idtype1] PASSED    [ 35%]
tests/pytorch/test_nn.py::test_appnp_conv_e_weight[g5-idtype0] PASSED    [ 35%]
tests/pytorch/test_nn.py::test_appnp_conv_e_weight[g5-idtype1] PASSED    [ 35%]
tests/pytorch/test_nn.py::test_gcn2conv_e_weight[True-g0-idtype0] PASSED [ 35%]
tests/pytorch/test_nn.py::test_gcn2conv_e_weight[True-g0-idtype1] PASSED [ 35%]
tests/pytorch/test_nn.py::test_gcn2conv_e_weight[True-g1-idtype0] PASSED [ 35%]
tests/pytorch/test_nn.py::test_gcn2conv_e_weight[True-g1-idtype1] PASSED [ 35%]
tests/pytorch/test_nn.py::test_gcn2conv_e_weight[True-g2-idtype0] PASSED [ 35%]
tests/pytorch/test_nn.py::test_gcn2conv_e_weight[True-g2-idtype1] PASSED [ 35%]
tests/pytorch/test_nn.py::test_gcn2conv_e_weight[True-g3-idtype0] PASSED [ 35%]
tests/pytorch/test_nn.py::test_gcn2conv_e_weight[True-g3-idtype1] PASSED [ 35%]
tests/pytorch/test_nn.py::test_gcn2conv_e_weight[True-g4-idtype0] PASSED [ 35%]
tests/pytorch/test_nn.py::test_gcn2conv_e_weight[True-g4-idtype1] PASSED [ 36%]
tests/pytorch/test_nn.py::test_gcn2conv_e_weight[True-g5-idtype0] PASSED [ 36%]
tests/pytorch/test_nn.py::test_gcn2conv_e_weight[True-g5-idtype1] PASSED [ 36%]
tests/pytorch/test_nn.py::test_gcn2conv_e_weight[False-g0-idtype0] PASSED [ 36%]
tests/pytorch/test_nn.py::test_gcn2conv_e_weight[False-g0-idtype1] PASSED [ 36%]
tests/pytorch/test_nn.py::test_gcn2conv_e_weight[False-g1-idtype0] PASSED [ 36%]
tests/pytorch/test_nn.py::test_gcn2conv_e_weight[False-g1-idtype1] PASSED [ 36%]
tests/pytorch/test_nn.py::test_gcn2conv_e_weight[False-g2-idtype0] PASSED [ 36%]
tests/pytorch/test_nn.py::test_gcn2conv_e_weight[False-g2-idtype1] PASSED [ 36%]
tests/pytorch/test_nn.py::test_gcn2conv_e_weight[False-g3-idtype0] PASSED [ 36%]
tests/pytorch/test_nn.py::test_gcn2conv_e_weight[False-g3-idtype1] PASSED [ 36%]
tests/pytorch/test_nn.py::test_gcn2conv_e_weight[False-g4-idtype0] PASSED [ 36%]
tests/pytorch/test_nn.py::test_gcn2conv_e_weight[False-g4-idtype1] PASSED [ 36%]
tests/pytorch/test_nn.py::test_gcn2conv_e_weight[False-g5-idtype0] PASSED [ 36%]
tests/pytorch/test_nn.py::test_gcn2conv_e_weight[False-g5-idtype1] PASSED [ 36%]
tests/pytorch/test_nn.py::test_sgconv_e_weight[g0-idtype0] PASSED        [ 36%]
tests/pytorch/test_nn.py::test_sgconv_e_weight[g0-idtype1] PASSED        [ 36%]
tests/pytorch/test_nn.py::test_sgconv_e_weight[g1-idtype0] PASSED        [ 36%]
tests/pytorch/test_nn.py::test_sgconv_e_weight[g1-idtype1] PASSED        [ 36%]
tests/pytorch/test_nn.py::test_sgconv_e_weight[g2-idtype0] PASSED        [ 36%]
tests/pytorch/test_nn.py::test_sgconv_e_weight[g2-idtype1] PASSED        [ 36%]
tests/pytorch/test_nn.py::test_sgconv_e_weight[g3-idtype0] PASSED        [ 36%]
tests/pytorch/test_nn.py::test_sgconv_e_weight[g3-idtype1] PASSED        [ 36%]
tests/pytorch/test_nn.py::test_sgconv_e_weight[g4-idtype0] PASSED        [ 36%]
tests/pytorch/test_nn.py::test_sgconv_e_weight[g4-idtype1] PASSED        [ 36%]
tests/pytorch/test_nn.py::test_sgconv_e_weight[g5-idtype0] PASSED        [ 36%]
tests/pytorch/test_nn.py::test_sgconv_e_weight[g5-idtype1] PASSED        [ 36%]
tests/pytorch/test_nn.py::test_tagconv_e_weight[g0-idtype0] PASSED       [ 36%]
tests/pytorch/test_nn.py::test_tagconv_e_weight[g0-idtype1] PASSED       [ 36%]
tests/pytorch/test_nn.py::test_tagconv_e_weight[g1-idtype0] PASSED       [ 36%]
tests/pytorch/test_nn.py::test_tagconv_e_weight[g1-idtype1] PASSED       [ 36%]
tests/pytorch/test_nn.py::test_tagconv_e_weight[g2-idtype0] PASSED       [ 36%]
tests/pytorch/test_nn.py::test_tagconv_e_weight[g2-idtype1] PASSED       [ 36%]
tests/pytorch/test_nn.py::test_tagconv_e_weight[g3-idtype0] PASSED       [ 36%]
tests/pytorch/test_nn.py::test_tagconv_e_weight[g3-idtype1] PASSED       [ 36%]
tests/pytorch/test_nn.py::test_tagconv_e_weight[g4-idtype0] PASSED       [ 36%]
tests/pytorch/test_nn.py::test_tagconv_e_weight[g4-idtype1] PASSED       [ 36%]
tests/pytorch/test_nn.py::test_tagconv_e_weight[g5-idtype0] PASSED       [ 36%]
tests/pytorch/test_nn.py::test_tagconv_e_weight[g5-idtype1] PASSED       [ 36%]
tests/pytorch/test_nn.py::test_gin_conv[mean-g0-idtype0] PASSED          [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[mean-g0-idtype1] PASSED          [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[mean-g1-idtype0] PASSED          [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[mean-g1-idtype1] PASSED          [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[mean-g2-idtype0] PASSED          [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[mean-g2-idtype1] PASSED          [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[mean-g3-idtype0] PASSED          [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[mean-g3-idtype1] PASSED          [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[mean-g4-idtype0] PASSED          [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[mean-g4-idtype1] PASSED          [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[mean-g5-idtype0] PASSED          [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[mean-g5-idtype1] PASSED          [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[mean-g6-idtype0] PASSED          [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[mean-g6-idtype1] PASSED          [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[max-g0-idtype0] PASSED           [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[max-g0-idtype1] PASSED           [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[max-g1-idtype0] PASSED           [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[max-g1-idtype1] PASSED           [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[max-g2-idtype0] PASSED           [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[max-g2-idtype1] PASSED           [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[max-g3-idtype0] PASSED           [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[max-g3-idtype1] PASSED           [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[max-g4-idtype0] PASSED           [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[max-g4-idtype1] PASSED           [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[max-g5-idtype0] PASSED           [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[max-g5-idtype1] PASSED           [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[max-g6-idtype0] PASSED           [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[max-g6-idtype1] PASSED           [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[sum-g0-idtype0] PASSED           [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[sum-g0-idtype1] PASSED           [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[sum-g1-idtype0] PASSED           [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[sum-g1-idtype1] PASSED           [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[sum-g2-idtype0] PASSED           [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[sum-g2-idtype1] PASSED           [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[sum-g3-idtype0] PASSED           [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[sum-g3-idtype1] PASSED           [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[sum-g4-idtype0] PASSED           [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[sum-g4-idtype1] PASSED           [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[sum-g5-idtype0] PASSED           [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[sum-g5-idtype1] PASSED           [ 37%]
tests/pytorch/test_nn.py::test_gin_conv[sum-g6-idtype0] PASSED           [ 38%]
tests/pytorch/test_nn.py::test_gin_conv[sum-g6-idtype1] PASSED           [ 38%]
tests/pytorch/test_nn.py::test_gine_conv[g0-idtype0] PASSED              [ 38%]
tests/pytorch/test_nn.py::test_gine_conv[g0-idtype1] PASSED              [ 38%]
tests/pytorch/test_nn.py::test_gine_conv[g1-idtype0] PASSED              [ 38%]
tests/pytorch/test_nn.py::test_gine_conv[g1-idtype1] PASSED              [ 38%]
tests/pytorch/test_nn.py::test_gine_conv[g2-idtype0] PASSED              [ 38%]
tests/pytorch/test_nn.py::test_gine_conv[g2-idtype1] PASSED              [ 38%]
tests/pytorch/test_nn.py::test_gine_conv[g3-idtype0] PASSED              [ 38%]
tests/pytorch/test_nn.py::test_gine_conv[g3-idtype1] PASSED              [ 38%]
tests/pytorch/test_nn.py::test_gine_conv[g4-idtype0] PASSED              [ 38%]
tests/pytorch/test_nn.py::test_gine_conv[g4-idtype1] PASSED              [ 38%]
tests/pytorch/test_nn.py::test_gine_conv[g5-idtype0] PASSED              [ 38%]
tests/pytorch/test_nn.py::test_gine_conv[g5-idtype1] PASSED              [ 38%]
tests/pytorch/test_nn.py::test_gine_conv[g6-idtype0] PASSED              [ 38%]
tests/pytorch/test_nn.py::test_gine_conv[g6-idtype1] PASSED              [ 38%]
tests/pytorch/test_nn.py::test_gine_conv[g7-idtype0] PASSED              [ 38%]
tests/pytorch/test_nn.py::test_gine_conv[g7-idtype1] PASSED              [ 38%]
tests/pytorch/test_nn.py::test_gin_conv_bi[mean-g0-idtype0] PASSED       [ 38%]
tests/pytorch/test_nn.py::test_gin_conv_bi[mean-g0-idtype1] PASSED       [ 38%]
tests/pytorch/test_nn.py::test_gin_conv_bi[mean-g1-idtype0] PASSED       [ 38%]
tests/pytorch/test_nn.py::test_gin_conv_bi[mean-g1-idtype1] PASSED       [ 38%]
tests/pytorch/test_nn.py::test_gin_conv_bi[max-g0-idtype0] PASSED        [ 38%]
tests/pytorch/test_nn.py::test_gin_conv_bi[max-g0-idtype1] PASSED        [ 38%]
tests/pytorch/test_nn.py::test_gin_conv_bi[max-g1-idtype0] PASSED        [ 38%]
tests/pytorch/test_nn.py::test_gin_conv_bi[max-g1-idtype1] PASSED        [ 38%]
tests/pytorch/test_nn.py::test_gin_conv_bi[sum-g0-idtype0] PASSED        [ 38%]
tests/pytorch/test_nn.py::test_gin_conv_bi[sum-g0-idtype1] PASSED        [ 38%]
tests/pytorch/test_nn.py::test_gin_conv_bi[sum-g1-idtype0] PASSED        [ 38%]
tests/pytorch/test_nn.py::test_gin_conv_bi[sum-g1-idtype1] PASSED        [ 38%]
tests/pytorch/test_nn.py::test_agnn_conv[g0-idtype0] PASSED              [ 38%]
tests/pytorch/test_nn.py::test_agnn_conv[g0-idtype1] PASSED              [ 38%]
tests/pytorch/test_nn.py::test_agnn_conv[g1-idtype0] PASSED              [ 38%]
tests/pytorch/test_nn.py::test_agnn_conv[g1-idtype1] PASSED              [ 38%]
tests/pytorch/test_nn.py::test_agnn_conv[g2-idtype0] PASSED              [ 38%]
tests/pytorch/test_nn.py::test_agnn_conv[g2-idtype1] PASSED              [ 38%]
tests/pytorch/test_nn.py::test_agnn_conv[g3-idtype0] PASSED              [ 38%]
tests/pytorch/test_nn.py::test_agnn_conv[g3-idtype1] PASSED              [ 38%]
tests/pytorch/test_nn.py::test_agnn_conv[g4-idtype0] PASSED              [ 38%]
tests/pytorch/test_nn.py::test_agnn_conv[g4-idtype1] PASSED              [ 38%]
tests/pytorch/test_nn.py::test_agnn_conv[g5-idtype0] PASSED              [ 39%]
tests/pytorch/test_nn.py::test_agnn_conv[g5-idtype1] PASSED              [ 39%]
tests/pytorch/test_nn.py::test_agnn_conv[g6-idtype0] PASSED              [ 39%]
tests/pytorch/test_nn.py::test_agnn_conv[g6-idtype1] PASSED              [ 39%]
tests/pytorch/test_nn.py::test_agnn_conv_bi[g0-idtype0] PASSED           [ 39%]
tests/pytorch/test_nn.py::test_agnn_conv_bi[g0-idtype1] PASSED           [ 39%]
tests/pytorch/test_nn.py::test_agnn_conv_bi[g1-idtype0] PASSED           [ 39%]
tests/pytorch/test_nn.py::test_agnn_conv_bi[g1-idtype1] PASSED           [ 39%]
tests/pytorch/test_nn.py::test_gated_graph_conv[g0-idtype0] PASSED       [ 39%]
tests/pytorch/test_nn.py::test_gated_graph_conv[g0-idtype1] PASSED       [ 39%]
tests/pytorch/test_nn.py::test_gated_graph_conv[g1-idtype0] PASSED       [ 39%]
tests/pytorch/test_nn.py::test_gated_graph_conv[g1-idtype1] PASSED       [ 39%]
tests/pytorch/test_nn.py::test_gated_graph_conv[g2-idtype0] PASSED       [ 39%]
tests/pytorch/test_nn.py::test_gated_graph_conv[g2-idtype1] PASSED       [ 39%]
tests/pytorch/test_nn.py::test_gated_graph_conv[g3-idtype0] PASSED       [ 39%]
tests/pytorch/test_nn.py::test_gated_graph_conv[g3-idtype1] PASSED       [ 39%]
tests/pytorch/test_nn.py::test_gated_graph_conv[g4-idtype0] PASSED       [ 39%]
tests/pytorch/test_nn.py::test_gated_graph_conv[g4-idtype1] PASSED       [ 39%]
tests/pytorch/test_nn.py::test_gated_graph_conv[g5-idtype0] PASSED       [ 39%]
tests/pytorch/test_nn.py::test_gated_graph_conv[g5-idtype1] PASSED       [ 39%]
tests/pytorch/test_nn.py::test_gated_graph_conv_one_etype[g0-idtype0] PASSED [ 39%]
tests/pytorch/test_nn.py::test_gated_graph_conv_one_etype[g0-idtype1] PASSED [ 39%]
tests/pytorch/test_nn.py::test_gated_graph_conv_one_etype[g1-idtype0] PASSED [ 39%]
tests/pytorch/test_nn.py::test_gated_graph_conv_one_etype[g1-idtype1] PASSED [ 39%]
tests/pytorch/test_nn.py::test_gated_graph_conv_one_etype[g2-idtype0] PASSED [ 39%]
tests/pytorch/test_nn.py::test_gated_graph_conv_one_etype[g2-idtype1] PASSED [ 39%]
tests/pytorch/test_nn.py::test_gated_graph_conv_one_etype[g3-idtype0] PASSED [ 39%]
tests/pytorch/test_nn.py::test_gated_graph_conv_one_etype[g3-idtype1] PASSED [ 39%]
tests/pytorch/test_nn.py::test_gated_graph_conv_one_etype[g4-idtype0] PASSED [ 39%]
tests/pytorch/test_nn.py::test_gated_graph_conv_one_etype[g4-idtype1] PASSED [ 39%]
tests/pytorch/test_nn.py::test_gated_graph_conv_one_etype[g5-idtype0] PASSED [ 39%]
tests/pytorch/test_nn.py::test_gated_graph_conv_one_etype[g5-idtype1] PASSED [ 39%]
tests/pytorch/test_nn.py::test_nn_conv[g0-idtype0] PASSED                [ 39%]
tests/pytorch/test_nn.py::test_nn_conv[g0-idtype1] PASSED                [ 39%]
tests/pytorch/test_nn.py::test_nn_conv[g1-idtype0] PASSED                [ 39%]
tests/pytorch/test_nn.py::test_nn_conv[g1-idtype1] PASSED                [ 39%]
tests/pytorch/test_nn.py::test_nn_conv[g2-idtype0] PASSED                [ 39%]
tests/pytorch/test_nn.py::test_nn_conv[g2-idtype1] PASSED                [ 39%]
tests/pytorch/test_nn.py::test_nn_conv[g3-idtype0] PASSED                    [  5%]
tests/compute/test_data.py::test_reddit PASSED                [ 39%]
tests/pytorch/test_nn.py::test_nn_conv[g3-idtype1] PASSED                [ 39%]
tests/pytorch/test_nn.py::test_nn_conv[g4-idtype0] PASSED                [ 40%]
tests/pytorch/test_nn.py::test_nn_conv[g4-idtype1] PASSED                [ 40%]
tests/pytorch/test_nn.py::test_nn_conv[g5-idtype0] PASSED                [ 40%]
tests/pytorch/test_nn.py::test_nn_conv[g5-idtype1] PASSED                [ 40%]
tests/pytorch/test_nn.py::test_nn_conv[g6-idtype0] PASSED                [ 40%]
tests/pytorch/test_nn.py::test_nn_conv[g6-idtype1] PASSED                [ 40%]
tests/pytorch/test_nn.py::test_nn_conv_bi[g0-idtype0] PASSED             [ 40%]
tests/pytorch/test_nn.py::test_nn_conv_bi[g0-idtype1] PASSED             [ 40%]
tests/pytorch/test_nn.py::test_nn_conv_bi[g1-idtype0] PASSED             [ 40%]
tests/pytorch/test_nn.py::test_nn_conv_bi[g1-idtype1] PASSED             [ 40%]
tests/pytorch/test_nn.py::test_gmm_conv[g0-idtype0] PASSED               [ 40%]
tests/pytorch/test_nn.py::test_gmm_conv[g0-idtype1] PASSED               [ 40%]
tests/pytorch/test_nn.py::test_gmm_conv[g1-idtype0] PASSED               [ 40%]
tests/pytorch/test_nn.py::test_gmm_conv[g1-idtype1] PASSED               [ 40%]
tests/pytorch/test_nn.py::test_gmm_conv[g2-idtype0] PASSED               [ 40%]
tests/pytorch/test_nn.py::test_gmm_conv[g2-idtype1] PASSED               [ 40%]
tests/pytorch/test_nn.py::test_gmm_conv[g3-idtype0] PASSED               [ 40%]
tests/pytorch/test_nn.py::test_gmm_conv[g3-idtype1] PASSED               [ 40%]
tests/pytorch/test_nn.py::test_gmm_conv[g4-idtype0] PASSED               [ 40%]
tests/pytorch/test_nn.py::test_gmm_conv[g4-idtype1] PASSED               [ 40%]
tests/pytorch/test_nn.py::test_gmm_conv[g5-idtype0] PASSED               [ 40%]
tests/pytorch/test_nn.py::test_gmm_conv[g5-idtype1] PASSED               [ 40%]
tests/pytorch/test_nn.py::test_gmm_conv_bi[g0-idtype0] PASSED            [ 40%]
tests/pytorch/test_nn.py::test_gmm_conv_bi[g0-idtype1] PASSED            [ 40%]
tests/pytorch/test_nn.py::test_gmm_conv_bi[g1-idtype0] PASSED            [ 40%]
tests/pytorch/test_nn.py::test_gmm_conv_bi[g1-idtype1] PASSED            [ 40%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g0-both-idtype0] PASSED [ 40%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g0-both-idtype1] PASSED [ 40%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g0-right-idtype0] PASSED [ 40%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g0-right-idtype1] PASSED [ 40%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g0-none-idtype0] PASSED [ 40%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g0-none-idtype1] PASSED [ 40%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g1-both-idtype0] PASSED [ 40%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g1-both-idtype1] PASSED [ 40%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g1-right-idtype0] PASSED [ 40%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g1-right-idtype1] PASSED [ 40%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g1-none-idtype0] PASSED [ 40%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g1-none-idtype1] PASSED [ 40%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g2-both-idtype0] PASSED [ 40%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g2-both-idtype1] PASSED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g2-right-idtype0] PASSED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g2-right-idtype1] PASSED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g2-none-idtype0] PASSED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g2-none-idtype1] PASSED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g3-both-idtype0] PASSED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g3-both-idtype1] PASSED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g3-right-idtype0] PASSED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g3-right-idtype1] PASSED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g3-none-idtype0] PASSED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g3-none-idtype1] PASSED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g4-both-idtype0] PASSED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g4-both-idtype1] PASSED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g4-right-idtype0] PASSED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g4-right-idtype1] PASSED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g4-none-idtype0] PASSED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g4-none-idtype1] PASSED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g5-both-idtype0] PASSED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g5-both-idtype1] PASSED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g5-right-idtype0] PASSED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g5-right-idtype1] PASSED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g5-none-idtype0] PASSED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g5-none-idtype1] PASSED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g6-both-idtype0] PASSED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g6-both-idtype1] PASSED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g6-right-idtype0] PASSED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g6-right-idtype1] PASSED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g6-none-idtype0] PASSED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g6-none-idtype1] PASSED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g7-both-idtype0] PASSED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g7-both-idtype1] PASSED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g7-right-idtype0] PASSED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g7-right-idtype1] PASSED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g7-none-idtype0] PASSED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[1-g7-none-idtype1] PASSED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g0-both-idtype0] PASSED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g0-both-idtype1] PASSED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g0-right-idtype0] PASSED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g0-right-idtype1] PASSED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g0-none-idtype0] PASSED [ 41%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g0-none-idtype1] PASSED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g1-both-idtype0] PASSED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g1-both-idtype1] PASSED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g1-right-idtype0] PASSED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g1-right-idtype1] PASSED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g1-none-idtype0] PASSED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g1-none-idtype1] PASSED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g2-both-idtype0] PASSED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g2-both-idtype1] PASSED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g2-right-idtype0] PASSED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g2-right-idtype1] PASSED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g2-none-idtype0] PASSED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g2-none-idtype1] PASSED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g3-both-idtype0] PASSED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g3-both-idtype1] PASSED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g3-right-idtype0] PASSED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g3-right-idtype1] PASSED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g3-none-idtype0] PASSED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g3-none-idtype1] PASSED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g4-both-idtype0] PASSED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g4-both-idtype1] PASSED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g4-right-idtype0] PASSED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g4-right-idtype1] PASSED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g4-none-idtype0] PASSED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g4-none-idtype1] PASSED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g5-both-idtype0] PASSED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g5-both-idtype1] PASSED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g5-right-idtype0] PASSED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g5-right-idtype1] PASSED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g5-none-idtype0] PASSED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g5-none-idtype1] PASSED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g6-both-idtype0] PASSED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g6-both-idtype1] PASSED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g6-right-idtype0] PASSED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g6-right-idtype1] PASSED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g6-none-idtype0] PASSED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g6-none-idtype1] PASSED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g7-both-idtype0] PASSED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g7-both-idtype1] PASSED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g7-right-idtype0] PASSED [ 42%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g7-right-idtype1] PASSED [ 43%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g7-none-idtype0] PASSED [ 43%]
tests/pytorch/test_nn.py::test_dense_graph_conv[2-g7-none-idtype1] PASSED [ 43%]
tests/pytorch/test_nn.py::test_dense_sage_conv[1-g0-idtype0] PASSED      [ 43%]
tests/pytorch/test_nn.py::test_dense_sage_conv[1-g0-idtype1] PASSED      [ 43%]
tests/pytorch/test_nn.py::test_dense_sage_conv[1-g1-idtype0] PASSED      [ 43%]
tests/pytorch/test_nn.py::test_dense_sage_conv[1-g1-idtype1] PASSED      [ 43%]
tests/pytorch/test_nn.py::test_dense_sage_conv[1-g2-idtype0] PASSED      [ 43%]
tests/pytorch/test_nn.py::test_dense_sage_conv[1-g2-idtype1] PASSED      [ 43%]
tests/pytorch/test_nn.py::test_dense_sage_conv[1-g3-idtype0] PASSED      [ 43%]
tests/pytorch/test_nn.py::test_dense_sage_conv[1-g3-idtype1] PASSED      [ 43%]
tests/pytorch/test_nn.py::test_dense_sage_conv[1-g4-idtype0] PASSED      [ 43%]
tests/pytorch/test_nn.py::test_dense_sage_conv[1-g4-idtype1] PASSED      [ 43%]
tests/pytorch/test_nn.py::test_dense_sage_conv[1-g5-idtype0] PASSED      [ 43%]
tests/pytorch/test_nn.py::test_dense_sage_conv[1-g5-idtype1] PASSED      [ 43%]
tests/pytorch/test_nn.py::test_dense_sage_conv[1-g6-idtype0] PASSED      [ 43%]
tests/pytorch/test_nn.py::test_dense_sage_conv[1-g6-idtype1] PASSED      [ 43%]
tests/pytorch/test_nn.py::test_dense_sage_conv[1-g7-idtype0] PASSED      [ 43%]
tests/pytorch/test_nn.py::test_dense_sage_conv[1-g7-idtype1] PASSED      [ 43%]
tests/pytorch/test_nn.py::test_dense_sage_conv[1-g8-idtype0] PASSED      [ 43%]
tests/pytorch/test_nn.py::test_dense_sage_conv[1-g8-idtype1] PASSED      [ 43%]
tests/pytorch/test_nn.py::test_dense_sage_conv[1-g9-idtype0] PASSED      [ 43%]
tests/pytorch/test_nn.py::test_dense_sage_conv[1-g9-idtype1] PASSED      [ 43%]
tests/pytorch/test_nn.py::test_dense_sage_conv[2-g0-idtype0] PASSED      [ 43%]
tests/pytorch/test_nn.py::test_dense_sage_conv[2-g0-idtype1] PASSED      [ 43%]
tests/pytorch/test_nn.py::test_dense_sage_conv[2-g1-idtype0] PASSED      [ 43%]
tests/pytorch/test_nn.py::test_dense_sage_conv[2-g1-idtype1] PASSED      [ 43%]
tests/pytorch/test_nn.py::test_dense_sage_conv[2-g2-idtype0] PASSED      [ 43%]
tests/pytorch/test_nn.py::test_dense_sage_conv[2-g2-idtype1] PASSED      [ 43%]
tests/pytorch/test_nn.py::test_dense_sage_conv[2-g3-idtype0] PASSED      [ 43%]
tests/pytorch/test_nn.py::test_dense_sage_conv[2-g3-idtype1] PASSED      [ 43%]
tests/pytorch/test_nn.py::test_dense_sage_conv[2-g4-idtype0] PASSED      [ 43%]
tests/pytorch/test_nn.py::test_dense_sage_conv[2-g4-idtype1] PASSED      [ 43%]
tests/pytorch/test_nn.py::test_dense_sage_conv[2-g5-idtype0] PASSED      [ 43%]
tests/pytorch/test_nn.py::test_dense_sage_conv[2-g5-idtype1] PASSED      [ 43%]
tests/pytorch/test_nn.py::test_dense_sage_conv[2-g6-idtype0] PASSED      [ 43%]
tests/pytorch/test_nn.py::test_dense_sage_conv[2-g6-idtype1] PASSED      [ 43%]
tests/pytorch/test_nn.py::test_dense_sage_conv[2-g7-idtype0] PASSED      [ 43%]
tests/pytorch/test_nn.py::test_dense_sage_conv[2-g7-idtype1] PASSED      [ 43%]
tests/pytorch/test_nn.py::test_dense_sage_conv[2-g8-idtype0] PASSED      [ 43%]
tests/pytorch/test_nn.py::test_dense_sage_conv[2-g8-idtype1] PASSED      [ 44%]
tests/pytorch/test_nn.py::test_dense_sage_conv[2-g9-idtype0] PASSED      [ 44%]
tests/pytorch/test_nn.py::test_dense_sage_conv[2-g9-idtype1] PASSED      [ 44%]
tests/pytorch/test_nn.py::test_edge_conv[1-g0-idtype0] PASSED            [ 44%]
tests/pytorch/test_nn.py::test_edge_conv[1-g0-idtype1] PASSED            [ 44%]
tests/pytorch/test_nn.py::test_edge_conv[1-g1-idtype0] PASSED            [ 44%]
tests/pytorch/test_nn.py::test_edge_conv[1-g1-idtype1] PASSED            [ 44%]
tests/pytorch/test_nn.py::test_edge_conv[1-g2-idtype0] PASSED            [ 44%]
tests/pytorch/test_nn.py::test_edge_conv[1-g2-idtype1] PASSED            [ 44%]
tests/pytorch/test_nn.py::test_edge_conv[1-g3-idtype0] PASSED            [ 44%]
tests/pytorch/test_nn.py::test_edge_conv[1-g3-idtype1] PASSED            [ 44%]
tests/pytorch/test_nn.py::test_edge_conv[1-g4-idtype0] PASSED            [ 44%]
tests/pytorch/test_nn.py::test_edge_conv[1-g4-idtype1] PASSED            [ 44%]
tests/pytorch/test_nn.py::test_edge_conv[1-g5-idtype0] PASSED            [ 44%]
tests/pytorch/test_nn.py::test_edge_conv[1-g5-idtype1] PASSED            [ 44%]
tests/pytorch/test_nn.py::test_edge_conv[1-g6-idtype0] PASSED            [ 44%]
tests/pytorch/test_nn.py::test_edge_conv[1-g6-idtype1] PASSED            [ 44%]
tests/pytorch/test_nn.py::test_edge_conv[2-g0-idtype0] PASSED            [ 44%]
tests/pytorch/test_nn.py::test_edge_conv[2-g0-idtype1] PASSED            [ 44%]
tests/pytorch/test_nn.py::test_edge_conv[2-g1-idtype0] PASSED            [ 44%]
tests/pytorch/test_nn.py::test_edge_conv[2-g1-idtype1] PASSED            [ 44%]
tests/pytorch/test_nn.py::test_edge_conv[2-g2-idtype0] PASSED            [ 44%]
tests/pytorch/test_nn.py::test_edge_conv[2-g2-idtype1] PASSED            [ 44%]
tests/pytorch/test_nn.py::test_edge_conv[2-g3-idtype0] PASSED            [ 44%]
tests/pytorch/test_nn.py::test_edge_conv[2-g3-idtype1] PASSED            [ 44%]
tests/pytorch/test_nn.py::test_edge_conv[2-g4-idtype0] PASSED            [ 44%]
tests/pytorch/test_nn.py::test_edge_conv[2-g4-idtype1] PASSED            [ 44%]
tests/pytorch/test_nn.py::test_edge_conv[2-g5-idtype0] PASSED            [ 44%]
tests/pytorch/test_nn.py::test_edge_conv[2-g5-idtype1] PASSED            [ 44%]
tests/pytorch/test_nn.py::test_edge_conv[2-g6-idtype0] PASSED            [ 44%]
tests/pytorch/test_nn.py::test_edge_conv[2-g6-idtype1] PASSED            [ 44%]
tests/pytorch/test_nn.py::test_edge_conv_bi[1-g0-idtype0] PASSED         [ 44%]
tests/pytorch/test_nn.py::test_edge_conv_bi[1-g0-idtype1] PASSED         [ 44%]
tests/pytorch/test_nn.py::test_edge_conv_bi[1-g1-idtype0] PASSED         [ 44%]
tests/pytorch/test_nn.py::test_edge_conv_bi[1-g1-idtype1] PASSED         [ 44%]
tests/pytorch/test_nn.py::test_edge_conv_bi[2-g0-idtype0] PASSED         [ 44%]
tests/pytorch/test_nn.py::test_edge_conv_bi[2-g0-idtype1] PASSED         [ 44%]
tests/pytorch/test_nn.py::test_edge_conv_bi[2-g1-idtype0] PASSED         [ 44%]
tests/pytorch/test_nn.py::test_edge_conv_bi[2-g1-idtype1] PASSED         [ 44%]
tests/pytorch/test_nn.py::test_dotgat_conv[1-1-g0-idtype0] PASSED        [ 44%]
tests/pytorch/test_nn.py::test_dotgat_conv[1-1-g0-idtype1] PASSED        [ 45%]
tests/pytorch/test_nn.py::test_dotgat_conv[1-1-g1-idtype0] PASSED        [ 45%]
tests/pytorch/test_nn.py::test_dotgat_conv[1-1-g1-idtype1] PASSED        [ 45%]
tests/pytorch/test_nn.py::test_dotgat_conv[1-1-g2-idtype0] PASSED        [ 45%]
tests/pytorch/test_nn.py::test_dotgat_conv[1-1-g2-idtype1] PASSED        [ 45%]
tests/pytorch/test_nn.py::test_dotgat_conv[1-1-g3-idtype0] PASSED        [ 45%]
tests/pytorch/test_nn.py::test_dotgat_conv[1-1-g3-idtype1] PASSED        [ 45%]
tests/pytorch/test_nn.py::test_dotgat_conv[1-1-g4-idtype0] PASSED        [ 45%]
tests/pytorch/test_nn.py::test_dotgat_conv[1-1-g4-idtype1] PASSED        [ 45%]
tests/pytorch/test_nn.py::test_dotgat_conv[1-1-g5-idtype0] PASSED        [ 45%]
tests/pytorch/test_nn.py::test_dotgat_conv[1-1-g5-idtype1] PASSED        [ 45%]
tests/pytorch/test_nn.py::test_dotgat_conv[1-1-g6-idtype0] PASSED        [ 45%]
tests/pytorch/test_nn.py::test_dotgat_conv[1-1-g6-idtype1] PASSED        [ 45%]
tests/pytorch/test_nn.py::test_dotgat_conv[1-2-g0-idtype0] PASSED        [ 45%]
tests/pytorch/test_nn.py::test_dotgat_conv[1-2-g0-idtype1] PASSED        [ 45%]
tests/pytorch/test_nn.py::test_dotgat_conv[1-2-g1-idtype0] PASSED        [ 45%]
tests/pytorch/test_nn.py::test_dotgat_conv[1-2-g1-idtype1] PASSED        [ 45%]
tests/pytorch/test_nn.py::test_dotgat_conv[1-2-g2-idtype0] PASSED        [ 45%]
tests/pytorch/test_nn.py::test_dotgat_conv[1-2-g2-idtype1] PASSED        [ 45%]
tests/pytorch/test_nn.py::test_dotgat_conv[1-2-g3-idtype0] PASSED        [ 45%]
tests/pytorch/test_nn.py::test_dotgat_conv[1-2-g3-idtype1] PASSED        [ 45%]
tests/pytorch/test_nn.py::test_dotgat_conv[1-2-g4-idtype0] PASSED        [ 45%]
tests/pytorch/test_nn.py::test_dotgat_conv[1-2-g4-idtype1] PASSED        [ 45%]
tests/pytorch/test_nn.py::test_dotgat_conv[1-2-g5-idtype0] PASSED        [ 45%]
tests/pytorch/test_nn.py::test_dotgat_conv[1-2-g5-idtype1] PASSED        [ 45%]
tests/pytorch/test_nn.py::test_dotgat_conv[1-2-g6-idtype0] PASSED        [ 45%]
tests/pytorch/test_nn.py::test_dotgat_conv[1-2-g6-idtype1] PASSED        [ 45%]
tests/pytorch/test_nn.py::test_dotgat_conv[4-1-g0-idtype0] PASSED        [ 45%]
tests/pytorch/test_nn.py::test_dotgat_conv[4-1-g0-idtype1] PASSED        [ 45%]
tests/pytorch/test_nn.py::test_dotgat_conv[4-1-g1-idtype0] PASSED        [ 45%]
tests/pytorch/test_nn.py::test_dotgat_conv[4-1-g1-idtype1] PASSED        [ 45%]
tests/pytorch/test_nn.py::test_dotgat_conv[4-1-g2-idtype0] PASSED        [ 45%]
tests/pytorch/test_nn.py::test_dotgat_conv[4-1-g2-idtype1] PASSED        [ 45%]
tests/pytorch/test_nn.py::test_dotgat_conv[4-1-g3-idtype0] PASSED        [ 45%]
tests/pytorch/test_nn.py::test_dotgat_conv[4-1-g3-idtype1] PASSED        [ 45%]
tests/pytorch/test_nn.py::test_dotgat_conv[4-1-g4-idtype0] PASSED        [ 45%]
tests/pytorch/test_nn.py::test_dotgat_conv[4-1-g4-idtype1] PASSED        [ 45%]
tests/pytorch/test_nn.py::test_dotgat_conv[4-1-g5-idtype0] PASSED        [ 45%]
tests/pytorch/test_nn.py::test_dotgat_conv[4-1-g5-idtype1] PASSED        [ 45%]
tests/pytorch/test_nn.py::test_dotgat_conv[4-1-g6-idtype0] PASSED        [ 46%]
tests/pytorch/test_nn.py::test_dotgat_conv[4-1-g6-idtype1] PASSED        [ 46%]
tests/pytorch/test_nn.py::test_dotgat_conv[4-2-g0-idtype0] PASSED        [ 46%]
tests/pytorch/test_nn.py::test_dotgat_conv[4-2-g0-idtype1] PASSED        [ 46%]
tests/pytorch/test_nn.py::test_dotgat_conv[4-2-g1-idtype0] PASSED        [ 46%]
tests/pytorch/test_nn.py::test_dotgat_conv[4-2-g1-idtype1] PASSED        [ 46%]
tests/pytorch/test_nn.py::test_dotgat_conv[4-2-g2-idtype0] PASSED        [ 46%]
tests/pytorch/test_nn.py::test_dotgat_conv[4-2-g2-idtype1] PASSED        [ 46%]
tests/pytorch/test_nn.py::test_dotgat_conv[4-2-g3-idtype0] PASSED        [ 46%]
tests/pytorch/test_nn.py::test_dotgat_conv[4-2-g3-idtype1] PASSED        [ 46%]
tests/pytorch/test_nn.py::test_dotgat_conv[4-2-g4-idtype0] PASSED        [ 46%]
tests/pytorch/test_nn.py::test_dotgat_conv[4-2-g4-idtype1] PASSED        [ 46%]
tests/pytorch/test_nn.py::test_dotgat_conv[4-2-g5-idtype0] PASSED        [ 46%]
tests/pytorch/test_nn.py::test_dotgat_conv[4-2-g5-idtype1] PASSED        [ 46%]
tests/pytorch/test_nn.py::test_dotgat_conv[4-2-g6-idtype0] PASSED        [ 46%]
tests/pytorch/test_nn.py::test_dotgat_conv[4-2-g6-idtype1] PASSED        [ 46%]
tests/pytorch/test_nn.py::test_dotgat_conv_bi[1-1-g0-idtype0] PASSED     [ 46%]
tests/pytorch/test_nn.py::test_dotgat_conv_bi[1-1-g0-idtype1] PASSED     [ 46%]
tests/pytorch/test_nn.py::test_dotgat_conv_bi[1-1-g1-idtype0] PASSED     [ 46%]
tests/pytorch/test_nn.py::test_dotgat_conv_bi[1-1-g1-idtype1] PASSED     [ 46%]
tests/pytorch/test_nn.py::test_dotgat_conv_bi[1-2-g0-idtype0] PASSED     [ 46%]
tests/pytorch/test_nn.py::test_dotgat_conv_bi[1-2-g0-idtype1] PASSED     [ 46%]
tests/pytorch/test_nn.py::test_dotgat_conv_bi[1-2-g1-idtype0] PASSED     [ 46%]
tests/pytorch/test_nn.py::test_dotgat_conv_bi[1-2-g1-idtype1] PASSED     [ 46%]
tests/pytorch/test_nn.py::test_dotgat_conv_bi[4-1-g0-idtype0] PASSED     [ 46%]
tests/pytorch/test_nn.py::test_dotgat_conv_bi[4-1-g0-idtype1] PASSED     [ 46%]
tests/pytorch/test_nn.py::test_dotgat_conv_bi[4-1-g1-idtype0] PASSED     [ 46%]
tests/pytorch/test_nn.py::test_dotgat_conv_bi[4-1-g1-idtype1] PASSED     [ 46%]
tests/pytorch/test_nn.py::test_dotgat_conv_bi[4-2-g0-idtype0] PASSED     [ 46%]
tests/pytorch/test_nn.py::test_dotgat_conv_bi[4-2-g0-idtype1] PASSED     [ 46%]
tests/pytorch/test_nn.py::test_dotgat_conv_bi[4-2-g1-idtype0] PASSED     [ 46%]
tests/pytorch/test_nn.py::test_dotgat_conv_bi[4-2-g1-idtype1] PASSED     [ 46%]
tests/pytorch/test_nn.py::test_dense_cheb_conv[1] PASSED                 [ 46%]
tests/pytorch/test_nn.py::test_dense_cheb_conv[2] PASSED                 [ 46%]
tests/pytorch/test_nn.py::test_sequential PASSED                         [ 46%]
tests/pytorch/test_nn.py::test_atomic_conv[g0-idtype0] PASSED            [ 46%]
tests/pytorch/test_nn.py::test_atomic_conv[g0-idtype1] PASSED            [ 46%]
tests/pytorch/test_nn.py::test_atomic_conv[g1-idtype0] PASSED            [ 46%]
tests/pytorch/test_nn.py::test_atomic_conv[g1-idtype1] PASSED            [ 46%]
tests/pytorch/test_nn.py::test_atomic_conv[g2-idtype0] PASSED            [ 46%]
tests/pytorch/test_nn.py::test_atomic_conv[g2-idtype1] PASSED            [ 47%]
tests/pytorch/test_nn.py::test_atomic_conv[g3-idtype0] PASSED            [ 47%]
tests/pytorch/test_nn.py::test_atomic_conv[g3-idtype1] PASSED            [ 47%]
tests/pytorch/test_nn.py::test_atomic_conv[g4-idtype0] PASSED            [ 47%]
tests/pytorch/test_nn.py::test_atomic_conv[g4-idtype1] PASSED            [ 47%]
tests/pytorch/test_nn.py::test_atomic_conv[g5-idtype0] PASSED            [ 47%]
tests/pytorch/test_nn.py::test_atomic_conv[g5-idtype1] PASSED            [ 47%]
tests/pytorch/test_nn.py::test_cf_conv[1-g0-idtype0] PASSED              [ 47%]
tests/pytorch/test_nn.py::test_cf_conv[1-g0-idtype1] PASSED              [ 47%]
tests/pytorch/test_nn.py::test_cf_conv[1-g1-idtype0] PASSED              [ 47%]
tests/pytorch/test_nn.py::test_cf_conv[1-g1-idtype1] PASSED              [ 47%]
tests/pytorch/test_nn.py::test_cf_conv[1-g2-idtype0] PASSED              [ 47%]
tests/pytorch/test_nn.py::test_cf_conv[1-g2-idtype1] PASSED              [ 47%]
tests/pytorch/test_nn.py::test_cf_conv[1-g3-idtype0] PASSED              [ 47%]
tests/pytorch/test_nn.py::test_cf_conv[1-g3-idtype1] PASSED              [ 47%]
tests/pytorch/test_nn.py::test_cf_conv[1-g4-idtype0] PASSED              [ 47%]
tests/pytorch/test_nn.py::test_cf_conv[1-g4-idtype1] PASSED              [ 47%]
tests/pytorch/test_nn.py::test_cf_conv[1-g5-idtype0] PASSED              [ 47%]
tests/pytorch/test_nn.py::test_cf_conv[1-g5-idtype1] PASSED              [ 47%]
tests/pytorch/test_nn.py::test_cf_conv[1-g6-idtype0] PASSED              [ 47%]
tests/pytorch/test_nn.py::test_cf_conv[1-g6-idtype1] PASSED              [ 47%]
tests/pytorch/test_nn.py::test_cf_conv[1-g7-idtype0] PASSED              [ 47%]
tests/pytorch/test_nn.py::test_cf_conv[1-g7-idtype1] PASSED              [ 47%]
tests/pytorch/test_nn.py::test_cf_conv[3-g0-idtype0] PASSED              [ 47%]
tests/pytorch/test_nn.py::test_cf_conv[3-g0-idtype1] PASSED              [ 47%]
tests/pytorch/test_nn.py::test_cf_conv[3-g1-idtype0] PASSED              [ 47%]
tests/pytorch/test_nn.py::test_cf_conv[3-g1-idtype1] PASSED              [ 47%]
tests/pytorch/test_nn.py::test_cf_conv[3-g2-idtype0] PASSED              [ 47%]
tests/pytorch/test_nn.py::test_cf_conv[3-g2-idtype1] PASSED              [ 47%]
tests/pytorch/test_nn.py::test_cf_conv[3-g3-idtype0] PASSED              [ 47%]
tests/pytorch/test_nn.py::test_cf_conv[3-g3-idtype1] PASSED              [ 47%]
tests/pytorch/test_nn.py::test_cf_conv[3-g4-idtype0] PASSED              [ 47%]
tests/pytorch/test_nn.py::test_cf_conv[3-g4-idtype1] PASSED              [ 47%]
tests/pytorch/test_nn.py::test_cf_conv[3-g5-idtype0] PASSED              [ 47%]
tests/pytorch/test_nn.py::test_cf_conv[3-g5-idtype1] PASSED              [ 47%]
tests/pytorch/test_nn.py::test_cf_conv[3-g6-idtype0] PASSED              [ 47%]
tests/pytorch/test_nn.py::test_cf_conv[3-g6-idtype1] PASSED              [ 47%]
tests/pytorch/test_nn.py::test_cf_conv[3-g7-idtype0] PASSED              [ 47%]
tests/pytorch/test_nn.py::test_cf_conv[3-g7-idtype1] PASSED              [ 47%]
tests/pytorch/test_nn.py::test_hetero_conv[False-sum-idtype0] PASSED     [ 47%]
tests/pytorch/test_nn.py::test_hetero_conv[False-sum-idtype1] PASSED     [ 48%]
tests/pytorch/test_nn.py::test_hetero_conv[False-max-idtype0] PASSED     [ 48%]
tests/pytorch/test_nn.py::test_hetero_conv[False-max-idtype1] PASSED     [ 48%]
tests/pytorch/test_nn.py::test_hetero_conv[False-min-idtype0] PASSED     [ 48%]
tests/pytorch/test_nn.py::test_hetero_conv[False-min-idtype1] PASSED     [ 48%]
tests/pytorch/test_nn.py::test_hetero_conv[False-mean-idtype0] PASSED    [ 48%]
tests/pytorch/test_nn.py::test_hetero_conv[False-mean-idtype1] PASSED    [ 48%]
tests/pytorch/test_nn.py::test_hetero_conv[False-stack-idtype0] PASSED   [ 48%]
tests/pytorch/test_nn.py::test_hetero_conv[False-stack-idtype1] PASSED   [ 48%]
tests/pytorch/test_nn.py::test_hetero_conv[False-myagg-idtype0] PASSED   [ 48%]
tests/pytorch/test_nn.py::test_hetero_conv[False-myagg-idtype1] PASSED   [ 48%]
tests/pytorch/test_nn.py::test_hetero_conv[True-sum-idtype0] PASSED      [ 48%]
tests/pytorch/test_nn.py::test_hetero_conv[True-sum-idtype1] PASSED      [ 48%]
tests/pytorch/test_nn.py::test_hetero_conv[True-max-idtype0] PASSED      [ 48%]
tests/pytorch/test_nn.py::test_hetero_conv[True-max-idtype1] PASSED      [ 48%]
tests/pytorch/test_nn.py::test_hetero_conv[True-min-idtype0] PASSED      [ 48%]
tests/pytorch/test_nn.py::test_hetero_conv[True-min-idtype1] PASSED      [ 48%]
tests/pytorch/test_nn.py::test_hetero_conv[True-mean-idtype0] PASSED     [ 48%]
tests/pytorch/test_nn.py::test_hetero_conv[True-mean-idtype1] PASSED     [ 48%]
tests/pytorch/test_nn.py::test_hetero_conv[True-stack-idtype0] PASSED    [ 48%]
tests/pytorch/test_nn.py::test_hetero_conv[True-stack-idtype1] PASSED    [ 48%]
tests/pytorch/test_nn.py::test_hetero_conv[True-myagg-idtype0] PASSED    [ 48%]
tests/pytorch/test_nn.py::test_hetero_conv[True-myagg-idtype1] PASSED    [ 48%]
tests/pytorch/test_nn.py::test_hetero_linear[1] PASSED                   [ 48%]
tests/pytorch/test_nn.py::test_hetero_linear[2] PASSED                   [ 48%]
tests/pytorch/test_nn.py::test_hetero_linear[100] PASSED                 [ 48%]
tests/pytorch/test_nn.py::test_hetero_embedding[1] PASSED                [ 48%]
tests/pytorch/test_nn.py::test_hetero_embedding[2] PASSED                [ 48%]
tests/pytorch/test_nn.py::test_hetero_embedding[100] PASSED              [ 48%]
tests/pytorch/test_nn.py::test_gnnexplainer[1-g0-idtype0] PASSED         [ 48%]
tests/pytorch/test_nn.py::test_gnnexplainer[1-g0-idtype1] PASSED         [ 48%]
tests/pytorch/test_nn.py::test_gnnexplainer[1-g1-idtype0] PASSED         [ 48%]
tests/pytorch/test_nn.py::test_gnnexplainer[1-g1-idtype1] PASSED         [ 48%]
tests/pytorch/test_nn.py::test_gnnexplainer[1-g2-idtype0] Cleaning workspace
Fetching without tags
PASSED         [ 48%]
tests/pytorch/test_nn.py::test_gnnexplainer[1-g2-idtype1] Merging remotes/origin/master commit d78a3a4baf611b90871a849f58647160c7cd9ab4 into PR head commit 37ed78035f406940ed88a4b60cadd2b7cfde5fea
Merge succeeded, producing 37ed78035f406940ed88a4b60cadd2b7cfde5fea
Checking out Revision 37ed78035f406940ed88a4b60cadd2b7cfde5fea (PR-4648)
 > git config remote.origin.url https://github.com/dmlc/dgl.git # timeout=10
 > git config --add remote.origin.fetch +refs/pull/4648/head:refs/remotes/origin/PR-4648 # timeout=10
 > git config --add remote.origin.fetch +refs/heads/master:refs/remotes/origin/master # timeout=10
 > git config remote.origin.url https://github.com/dmlc/dgl.git # timeout=10
 > git rev-parse --verify HEAD # timeout=10
No valid HEAD. Skipping the resetting
 > git clean -fdx # timeout=10
Fetching upstream changes from https://github.com/dmlc/dgl.git
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git fetch --no-tags --progress -- https://github.com/dmlc/dgl.git +refs/pull/4648/head:refs/remotes/origin/PR-4648 +refs/heads/master:refs/remotes/origin/master # timeout=10
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git remote # timeout=10
 > git config --get remote.origin.url # timeout=10
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git merge d78a3a4baf611b90871a849f58647160c7cd9ab4 # timeout=10
 > git rev-parse HEAD^{commit} # timeout=10
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
PASSED         [ 48%]
tests/pytorch/test_nn.py::test_gnnexplainer[1-g3-idtype0] Commit message: "fix for pytorch < 1.12"
Cleaning workspace
PASSED         [ 48%]
tests/pytorch/test_nn.py::test_gnnexplainer[1-g3-idtype1] [Pipeline] withEnv
[Pipeline] {
[Pipeline] isUnix
[Pipeline] withEnv
[Pipeline] {
[Pipeline] sh
PASSED         [ 48%]
tests/pytorch/test_nn.py::test_gnnexplainer[1-g4-idtype0] + docker pull dgllib/dgl-ci-cpu:ssh_v220818
PASSED         [ 48%]
tests/pytorch/test_nn.py::test_gnnexplainer[1-g4-idtype1]  > git branch -a -v --no-abbrev # timeout=10
 > git checkout -b PR-4648 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git rev-parse --verify HEAD # timeout=10
Resetting working tree
 > git reset --hard # timeout=10
 > git clean -fdx # timeout=10
PASSED         [ 48%]
tests/pytorch/test_nn.py::test_gnnexplainer[1-g5-idtype0] ssh_v220818: Pulling from dgllib/dgl-ci-cpu
Digest: sha256:9c66e1df96cfae9cb9aa9ba39ef4921e152409d1447d6b123659ff4f84b4291f
Status: Image is up to date for dgllib/dgl-ci-cpu:ssh_v220818
docker.io/dgllib/dgl-ci-cpu:ssh_v220818
[Pipeline] }
[Pipeline] // withEnv
[Pipeline] isUnix
[Pipeline] withEnv
[Pipeline] {
[Pipeline] sh
PASSED         [ 48%]
tests/pytorch/test_nn.py::test_gnnexplainer[1-g5-idtype1] + docker inspect -f . dgllib/dgl-ci-cpu:ssh_v220818
.
[Pipeline] }
[Pipeline] // withEnv
[Pipeline] withDockerContainer
dgl-manual-large-cpu does not seem to be running inside a container
$ docker run -t -d -u 0:0 --shm-size=4gb -w /root/jenkins/workspace/dgl_PR-4648 -v /root/jenkins/workspace/dgl_PR-4648:/root/jenkins/workspace/dgl_PR-4648:rw,z -v /root/jenkins/workspace/dgl_PR-4648@tmp:/root/jenkins/workspace/dgl_PR-4648@tmp:rw,z -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** dgllib/dgl-ci-cpu:ssh_v220818 cat
PASSED         [ 49%]
tests/pytorch/test_nn.py::test_gnnexplainer[2-g0-idtype0] PASSED         [ 49%]
tests/pytorch/test_nn.py::test_gnnexplainer[2-g0-idtype1] PASSED         [ 49%]
tests/pytorch/test_nn.py::test_gnnexplainer[2-g1-idtype0] PASSED         [ 49%]
tests/pytorch/test_nn.py::test_gnnexplainer[2-g1-idtype1] $ docker top 9d11908e8a1623fe2cb3ee8ecb53c2f89a04f190533b91026267498f2decd76b -eo pid,comm
ERROR: The container started but didn't run the expected command. Please double check your ENTRYPOINT does execute the command passed as docker run argument, as required by official docker images (see https://github.com/docker-library/official-images#consistency for entrypoint consistency requirements).
Alternatively you can force image entrypoint to be disabled by adding option `--entrypoint=''`.
[Pipeline] {
[Pipeline] stage
[Pipeline] { (Distributed Torch CPU Unit test)
[Pipeline] sh
PASSED         [ 49%]
tests/pytorch/test_nn.py::test_gnnexplainer[2-g2-idtype0] + rm -rf CMakeLists.txt CONTRIBUTORS.md Jenkinsfile LICENSE NEWS.md README.md apps benchmarks cmake conda dglgo docker docs examples featgraph include pyproject.toml python readthedocs.yml src tensoradapter tests third_party tools tutorials
[Pipeline] checkout
The recommended git tool is: git
using credential 150de63f-189c-4717-bcaf-010460d2f51a
Warning: JENKINS-30600: special launcher org.jenkinsci.plugins.docker.workflow.WithContainerStep$Decorator$1@487ac80b; decorates RemoteLauncher[hudson.remoting.Channel@1d9b9638:dgl-manual-large-cpu] will be ignored (a typical symptom is the Git executable not being run inside a designated container)
Fetching changes from the remote Git repository
Cleaning workspace
PASSED         [ 49%]
tests/pytorch/test_nn.py::test_gnnexplainer[2-g2-idtype1] Fetching without tags
PASSED         [ 49%]
tests/pytorch/test_nn.py::test_gnnexplainer[2-g3-idtype0] Merging remotes/origin/master commit d78a3a4baf611b90871a849f58647160c7cd9ab4 into PR head commit 37ed78035f406940ed88a4b60cadd2b7cfde5fea
Merge succeeded, producing 37ed78035f406940ed88a4b60cadd2b7cfde5fea
Checking out Revision 37ed78035f406940ed88a4b60cadd2b7cfde5fea (PR-4648)
Commit message: "fix for pytorch < 1.12"
Cleaning workspace
 > git rev-parse --resolve-git-dir /root/jenkins/workspace/dgl_PR-4648/.git # timeout=10
 > git config remote.origin.url https://github.com/dmlc/dgl.git # timeout=10
 > git rev-parse --verify HEAD # timeout=10
Resetting working tree
 > git reset --hard # timeout=10
 > git clean -fdx # timeout=10
Fetching upstream changes from https://github.com/dmlc/dgl.git
 > git --version # timeout=10
 > git --version # 'git version 2.17.1'
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git fetch --no-tags --progress -- https://github.com/dmlc/dgl.git +refs/pull/4648/head:refs/remotes/origin/PR-4648 +refs/heads/master:refs/remotes/origin/master # timeout=10
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git remote # timeout=10
 > git config --get remote.origin.url # timeout=10
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git merge d78a3a4baf611b90871a849f58647160c7cd9ab4 # timeout=10
 > git rev-parse HEAD^{commit} # timeout=10
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git branch -a -v --no-abbrev # timeout=10
 > git branch -D PR-4648 # timeout=10
 > git checkout -b PR-4648 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git rev-parse --verify HEAD # timeout=10
Resetting working tree
 > git reset --hard # timeout=10
 > git clean -fdx # timeout=10
PASSED         [ 49%]
tests/pytorch/test_nn.py::test_gnnexplainer[2-g3-idtype1] [Pipeline] sh
PASSED         [ 49%]
tests/pytorch/test_nn.py::test_gnnexplainer[2-g4-idtype0] + git submodule update --recursive --init
Submodule 'third_party/METIS' (https://github.com/KarypisLab/METIS.git) registered for path 'third_party/METIS'
Submodule 'third_party/dlpack' (https://github.com/dmlc/dlpack.git) registered for path 'third_party/dlpack'
Submodule 'third_party/dmlc-core' (https://github.com/dmlc/dmlc-core.git) registered for path 'third_party/dmlc-core'
Submodule 'third_party/googletest' (https://github.com/google/googletest.git) registered for path 'third_party/googletest'
Submodule 'third_party/libxsmm' (https://github.com/hfp/libxsmm.git) registered for path 'third_party/libxsmm'
Submodule 'third_party/nanoflann' (https://github.com/jlblancoc/nanoflann) registered for path 'third_party/nanoflann'
Submodule 'third_party/nccl' (https://github.com/nvidia/nccl) registered for path 'third_party/nccl'
Submodule 'third_party/phmap' (https://github.com/greg7mdp/parallel-hashmap.git) registered for path 'third_party/phmap'
Submodule 'third_party/tensorpipe' (https://github.com/pytorch/tensorpipe) registered for path 'third_party/tensorpipe'
Submodule 'third_party/thrust' (https://github.com/NVIDIA/thrust.git) registered for path 'third_party/thrust'
Submodule 'third_party/tvm' (https://github.com/apache/incubator-tvm) registered for path 'third_party/tvm'
Submodule 'third_party/xbyak' (https://github.com/herumi/xbyak) registered for path 'third_party/xbyak'
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/METIS'...
PASSED         [ 49%]
tests/pytorch/test_nn.py::test_gnnexplainer[2-g4-idtype1] PASSED         [ 49%]
tests/pytorch/test_nn.py::test_gnnexplainer[2-g5-idtype0] Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/dlpack'...
PASSED         [ 49%]
tests/pytorch/test_nn.py::test_gnnexplainer[2-g5-idtype1] Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/dmlc-core'...
PASSED         [ 49%]
tests/pytorch/test_nn.py::test_jumping_knowledge PASSED                  [ 49%]
tests/pytorch/test_nn.py::test_edge_predictor[dot] PASSED                [ 49%]
tests/pytorch/test_nn.py::test_edge_predictor[cos] PASSED                [ 49%]
tests/pytorch/test_nn.py::test_edge_predictor[ele] PASSED                [ 49%]
tests/pytorch/test_nn.py::test_edge_predictor[cat] PASSED                [ 49%]
tests/pytorch/test_nn.py::test_ke_score_funcs PASSED                     [ 49%]
tests/pytorch/test_nn.py::test_twirls PASSED                             [ 49%]
tests/pytorch/test_nn.py::test_typed_linear[None-None-4] PASSED          [ 49%]
tests/pytorch/test_nn.py::test_typed_linear[None-None-32] PASSED         [ 49%]
tests/pytorch/test_nn.py::test_typed_linear[basis-4-4] PASSED            [ 49%]
tests/pytorch/test_nn.py::test_typed_linear[basis-4-32] PASSED           [ 49%]
tests/pytorch/test_nn.py::test_typed_linear[bdd-4-4] PASSED              [ 49%]
tests/pytorch/test_nn.py::test_typed_linear[bdd-4-32] PASSED             [ 49%]
tests/pytorch/test_nn.py::test_hgt[1-4-idtype0] PASSED                   [ 49%]
tests/pytorch/test_nn.py::test_hgt[1-4-idtype1] PASSED                   [ 49%]
tests/pytorch/test_nn.py::test_radius_graph[True-True] PASSED            [ 49%]
tests/pytorch/test_nn.py::test_radius_graph[True-False] PASSED           [ 49%]
tests/pytorch/test_nn.py::test_radius_graph[False-True] PASSED           [ 49%]
tests/pytorch/test_nn.py::test_radius_graph[False-False] PASSED          [ 49%]
tests/pytorch/test_nn.py::test_group_rev_res[idtype0] PASSED             [ 49%]
tests/pytorch/test_nn.py::test_group_rev_res[idtype1] PASSED             [ 49%]
tests/pytorch/test_nn.py::test_egnn_conv[16-16-16-16] PASSED             [ 49%]
tests/pytorch/test_nn.py::test_egnn_conv[16-16-16-32] PASSED             [ 49%]
tests/pytorch/test_nn.py::test_egnn_conv[16-16-32-16] PASSED             [ 49%]
tests/pytorch/test_nn.py::test_egnn_conv[16-16-32-32] PASSED             [ 49%]
tests/pytorch/test_nn.py::test_egnn_conv[16-32-16-16] PASSED             [ 49%]
tests/pytorch/test_nn.py::test_egnn_conv[16-32-16-32] PASSED             [ 50%]
tests/pytorch/test_nn.py::test_egnn_conv[16-32-32-16] PASSED             [ 50%]
tests/pytorch/test_nn.py::test_egnn_conv[16-32-32-32] PASSED             [ 50%]
tests/pytorch/test_nn.py::test_egnn_conv[10-16-16-16] PASSED             [ 50%]
tests/pytorch/test_nn.py::test_egnn_conv[10-16-16-32] PASSED             [ 50%]
tests/pytorch/test_nn.py::test_egnn_conv[10-16-32-16] PASSED             [ 50%]
tests/pytorch/test_nn.py::test_egnn_conv[10-16-32-32] PASSED             [ 50%]
tests/pytorch/test_nn.py::test_egnn_conv[10-32-16-16] PASSED             [ 50%]
tests/pytorch/test_nn.py::test_egnn_conv[10-32-16-32] PASSED             [ 50%]
tests/pytorch/test_nn.py::test_egnn_conv[10-32-32-16] PASSED             [ 50%]
tests/pytorch/test_nn.py::test_egnn_conv[10-32-32-32] PASSED             [ 50%]
tests/pytorch/test_nn.py::test_egnn_conv[0-16-16-16] PASSED              [ 50%]
tests/pytorch/test_nn.py::test_egnn_conv[0-16-16-32] PASSED              [ 50%]
tests/pytorch/test_nn.py::test_egnn_conv[0-16-32-16] PASSED              [ 50%]
tests/pytorch/test_nn.py::test_egnn_conv[0-16-32-32] PASSED              [ 50%]
tests/pytorch/test_nn.py::test_egnn_conv[0-32-16-16] PASSED              [ 50%]
tests/pytorch/test_nn.py::test_egnn_conv[0-32-16-32] Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/googletest'...
PASSED              [ 50%]
tests/pytorch/test_nn.py::test_egnn_conv[0-32-32-16] PASSED              [ 50%]
tests/pytorch/test_nn.py::test_egnn_conv[0-32-32-32] PASSED              [ 50%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-2.5-scalers0-aggregators0-16-16] PASSED [ 50%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-2.5-scalers0-aggregators0-16-32] PASSED [ 50%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-2.5-scalers0-aggregators0-32-16] PASSED [ 50%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-2.5-scalers0-aggregators0-32-32] PASSED [ 50%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-2.5-scalers0-aggregators1-16-16] PASSED [ 50%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-2.5-scalers0-aggregators1-16-32] PASSED [ 50%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-2.5-scalers0-aggregators1-32-16] PASSED [ 50%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-2.5-scalers0-aggregators1-32-32] PASSED [ 50%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-2.5-scalers0-aggregators2-16-16] PASSED [ 50%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-2.5-scalers0-aggregators2-16-32] PASSED [ 50%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-2.5-scalers0-aggregators2-32-16] PASSED [ 50%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-2.5-scalers0-aggregators2-32-32] PASSED [ 50%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-2.5-scalers1-aggregators0-16-16] PASSED [ 50%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-2.5-scalers1-aggregators0-16-32] PASSED [ 50%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-2.5-scalers1-aggregators0-32-16] PASSED [ 50%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-2.5-scalers1-aggregators0-32-32] PASSED [ 50%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-2.5-scalers1-aggregators1-16-16] PASSED [ 50%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-2.5-scalers1-aggregators1-16-32] PASSED [ 50%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-2.5-scalers1-aggregators1-32-16] PASSED [ 50%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-2.5-scalers1-aggregators1-32-32] PASSED [ 50%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-2.5-scalers1-aggregators2-16-16] PASSED [ 50%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-2.5-scalers1-aggregators2-16-32] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-2.5-scalers1-aggregators2-32-16] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-2.5-scalers1-aggregators2-32-32] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-7.4-scalers0-aggregators0-16-16] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-7.4-scalers0-aggregators0-16-32] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-7.4-scalers0-aggregators0-32-16] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-7.4-scalers0-aggregators0-32-32] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-7.4-scalers0-aggregators1-16-16] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-7.4-scalers0-aggregators1-16-32] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-7.4-scalers0-aggregators1-32-16] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-7.4-scalers0-aggregators1-32-32] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-7.4-scalers0-aggregators2-16-16] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-7.4-scalers0-aggregators2-16-32] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-7.4-scalers0-aggregators2-32-16] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-7.4-scalers0-aggregators2-32-32] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-7.4-scalers1-aggregators0-16-16] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-7.4-scalers1-aggregators0-16-32] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-7.4-scalers1-aggregators0-32-16] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-7.4-scalers1-aggregators0-32-32] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-7.4-scalers1-aggregators1-16-16] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-7.4-scalers1-aggregators1-16-32] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-7.4-scalers1-aggregators1-32-16] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-7.4-scalers1-aggregators1-32-32] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-7.4-scalers1-aggregators2-16-16] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-7.4-scalers1-aggregators2-16-32] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-7.4-scalers1-aggregators2-32-16] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.0-7.4-scalers1-aggregators2-32-32] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-2.5-scalers0-aggregators0-16-16] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-2.5-scalers0-aggregators0-16-32] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-2.5-scalers0-aggregators0-32-16] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-2.5-scalers0-aggregators0-32-32] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-2.5-scalers0-aggregators1-16-16] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-2.5-scalers0-aggregators1-16-32] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-2.5-scalers0-aggregators1-32-16] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-2.5-scalers0-aggregators1-32-32] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-2.5-scalers0-aggregators2-16-16] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-2.5-scalers0-aggregators2-16-32] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-2.5-scalers0-aggregators2-32-16] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-2.5-scalers0-aggregators2-32-32] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-2.5-scalers1-aggregators0-16-16] PASSED [ 51%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-2.5-scalers1-aggregators0-16-32] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-2.5-scalers1-aggregators0-32-16] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-2.5-scalers1-aggregators0-32-32] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-2.5-scalers1-aggregators1-16-16] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-2.5-scalers1-aggregators1-16-32] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-2.5-scalers1-aggregators1-32-16] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-2.5-scalers1-aggregators1-32-32] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-2.5-scalers1-aggregators2-16-16] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-2.5-scalers1-aggregators2-16-32] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-2.5-scalers1-aggregators2-32-16] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-2.5-scalers1-aggregators2-32-32] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-7.4-scalers0-aggregators0-16-16] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-7.4-scalers0-aggregators0-16-32] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-7.4-scalers0-aggregators0-32-16] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-7.4-scalers0-aggregators0-32-32] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-7.4-scalers0-aggregators1-16-16] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-7.4-scalers0-aggregators1-16-32] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-7.4-scalers0-aggregators1-32-16] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-7.4-scalers0-aggregators1-32-32] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-7.4-scalers0-aggregators2-16-16] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-7.4-scalers0-aggregators2-16-32] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-7.4-scalers0-aggregators2-32-16] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-7.4-scalers0-aggregators2-32-32] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-7.4-scalers1-aggregators0-16-16] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-7.4-scalers1-aggregators0-16-32] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-7.4-scalers1-aggregators0-32-16] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-7.4-scalers1-aggregators0-32-32] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-7.4-scalers1-aggregators1-16-16] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-7.4-scalers1-aggregators1-16-32] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-7.4-scalers1-aggregators1-32-16] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-7.4-scalers1-aggregators1-32-32] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-7.4-scalers1-aggregators2-16-16] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-7.4-scalers1-aggregators2-16-32] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-7.4-scalers1-aggregators2-32-16] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-1-0.1-7.4-scalers1-aggregators2-32-32] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-2.5-scalers0-aggregators0-16-16] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-2.5-scalers0-aggregators0-16-32] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-2.5-scalers0-aggregators0-32-16] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-2.5-scalers0-aggregators0-32-32] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-2.5-scalers0-aggregators1-16-16] PASSED [ 52%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-2.5-scalers0-aggregators1-16-32] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-2.5-scalers0-aggregators1-32-16] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-2.5-scalers0-aggregators1-32-32] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-2.5-scalers0-aggregators2-16-16] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-2.5-scalers0-aggregators2-16-32] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-2.5-scalers0-aggregators2-32-16] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-2.5-scalers0-aggregators2-32-32] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-2.5-scalers1-aggregators0-16-16] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-2.5-scalers1-aggregators0-16-32] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-2.5-scalers1-aggregators0-32-16] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-2.5-scalers1-aggregators0-32-32] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-2.5-scalers1-aggregators1-16-16] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-2.5-scalers1-aggregators1-16-32] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-2.5-scalers1-aggregators1-32-16] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-2.5-scalers1-aggregators1-32-32] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-2.5-scalers1-aggregators2-16-16] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-2.5-scalers1-aggregators2-16-32] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-2.5-scalers1-aggregators2-32-16] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-2.5-scalers1-aggregators2-32-32] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-7.4-scalers0-aggregators0-16-16] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-7.4-scalers0-aggregators0-16-32] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-7.4-scalers0-aggregators0-32-16] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-7.4-scalers0-aggregators0-32-32] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-7.4-scalers0-aggregators1-16-16] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-7.4-scalers0-aggregators1-16-32] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-7.4-scalers0-aggregators1-32-16] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-7.4-scalers0-aggregators1-32-32] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-7.4-scalers0-aggregators2-16-16] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-7.4-scalers0-aggregators2-16-32] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-7.4-scalers0-aggregators2-32-16] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-7.4-scalers0-aggregators2-32-32] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-7.4-scalers1-aggregators0-16-16] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-7.4-scalers1-aggregators0-16-32] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-7.4-scalers1-aggregators0-32-16] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-7.4-scalers1-aggregators0-32-32] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-7.4-scalers1-aggregators1-16-16] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-7.4-scalers1-aggregators1-16-32] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-7.4-scalers1-aggregators1-32-16] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-7.4-scalers1-aggregators1-32-32] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-7.4-scalers1-aggregators2-16-16] PASSED [ 53%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-7.4-scalers1-aggregators2-16-32] PASSED [ 54%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-7.4-scalers1-aggregators2-32-16] PASSED [ 54%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.0-7.4-scalers1-aggregators2-32-32] PASSED [ 54%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-2.5-scalers0-aggregators0-16-16] PASSED [ 54%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-2.5-scalers0-aggregators0-16-32] PASSED [ 54%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-2.5-scalers0-aggregators0-32-16] PASSED [ 54%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-2.5-scalers0-aggregators0-32-32] PASSED [ 54%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-2.5-scalers0-aggregators1-16-16] PASSED [ 54%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-2.5-scalers0-aggregators1-16-32] PASSED [ 54%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-2.5-scalers0-aggregators1-32-16] PASSED [ 54%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-2.5-scalers0-aggregators1-32-32] PASSED [ 54%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-2.5-scalers0-aggregators2-16-16] PASSED [ 54%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-2.5-scalers0-aggregators2-16-32] PASSED [ 54%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-2.5-scalers0-aggregators2-32-16] PASSED [ 54%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-2.5-scalers0-aggregators2-32-32] PASSED [ 54%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-2.5-scalers1-aggregators0-16-16] PASSED [ 54%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-2.5-scalers1-aggregators0-16-32] PASSED [ 54%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-2.5-scalers1-aggregators0-32-16] PASSED [ 54%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-2.5-scalers1-aggregators0-32-32] PASSED [ 54%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-2.5-scalers1-aggregators1-16-16] PASSED [ 54%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-2.5-scalers1-aggregators1-16-32] PASSED [ 54%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-2.5-scalers1-aggregators1-32-16] PASSED [ 54%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-2.5-scalers1-aggregators1-32-32] PASSED [ 54%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-2.5-scalers1-aggregators2-16-16] PASSED [ 54%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-2.5-scalers1-aggregators2-16-32] PASSED [ 54%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-2.5-scalers1-aggregators2-32-16] PASSED [ 54%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-2.5-scalers1-aggregators2-32-32] PASSED [ 54%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-7.4-scalers0-aggregators0-16-16] PASSED [ 54%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-7.4-scalers0-aggregators0-16-32] PASSED [ 54%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-7.4-scalers0-aggregators0-32-16] PASSED [ 54%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-7.4-scalers0-aggregators0-32-32] PASSED [ 54%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-7.4-scalers0-aggregators1-16-16] PASSED [ 54%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-7.4-scalers0-aggregators1-16-32] Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/libxsmm'...
PASSED [ 54%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-7.4-scalers0-aggregators1-32-16] PASSED [ 54%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-7.4-scalers0-aggregators1-32-32] PASSED [ 54%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-7.4-scalers0-aggregators2-16-16] PASSED [ 54%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-7.4-scalers0-aggregators2-16-32] PASSED [ 54%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-7.4-scalers0-aggregators2-32-16] PASSED [ 54%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-7.4-scalers0-aggregators2-32-32] PASSED [ 54%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-7.4-scalers1-aggregators0-16-16] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-7.4-scalers1-aggregators0-16-32] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-7.4-scalers1-aggregators0-32-16] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-7.4-scalers1-aggregators0-32-32] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-7.4-scalers1-aggregators1-16-16] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-7.4-scalers1-aggregators1-16-32] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-7.4-scalers1-aggregators1-32-16] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-7.4-scalers1-aggregators1-32-32] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-7.4-scalers1-aggregators2-16-16] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-7.4-scalers1-aggregators2-16-32] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-7.4-scalers1-aggregators2-32-16] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-16-4-0.1-7.4-scalers1-aggregators2-32-32] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-2.5-scalers0-aggregators0-16-16] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-2.5-scalers0-aggregators0-16-32] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-2.5-scalers0-aggregators0-32-16] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-2.5-scalers0-aggregators0-32-32] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-2.5-scalers0-aggregators1-16-16] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-2.5-scalers0-aggregators1-16-32] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-2.5-scalers0-aggregators1-32-16] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-2.5-scalers0-aggregators1-32-32] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-2.5-scalers0-aggregators2-16-16] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-2.5-scalers0-aggregators2-16-32] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-2.5-scalers0-aggregators2-32-16] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-2.5-scalers0-aggregators2-32-32] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-2.5-scalers1-aggregators0-16-16] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-2.5-scalers1-aggregators0-16-32] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-2.5-scalers1-aggregators0-32-16] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-2.5-scalers1-aggregators0-32-32] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-2.5-scalers1-aggregators1-16-16] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-2.5-scalers1-aggregators1-16-32] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-2.5-scalers1-aggregators1-32-16] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-2.5-scalers1-aggregators1-32-32] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-2.5-scalers1-aggregators2-16-16] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-2.5-scalers1-aggregators2-16-32] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-2.5-scalers1-aggregators2-32-16] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-2.5-scalers1-aggregators2-32-32] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-7.4-scalers0-aggregators0-16-16] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-7.4-scalers0-aggregators0-16-32] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-7.4-scalers0-aggregators0-32-16] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-7.4-scalers0-aggregators0-32-32] PASSED [ 55%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-7.4-scalers0-aggregators1-16-16] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-7.4-scalers0-aggregators1-16-32] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-7.4-scalers0-aggregators1-32-16] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-7.4-scalers0-aggregators1-32-32] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-7.4-scalers0-aggregators2-16-16] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-7.4-scalers0-aggregators2-16-32] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-7.4-scalers0-aggregators2-32-16] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-7.4-scalers0-aggregators2-32-32] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-7.4-scalers1-aggregators0-16-16] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-7.4-scalers1-aggregators0-16-32] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-7.4-scalers1-aggregators0-32-16] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-7.4-scalers1-aggregators0-32-32] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-7.4-scalers1-aggregators1-16-16] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-7.4-scalers1-aggregators1-16-32] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-7.4-scalers1-aggregators1-32-16] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-7.4-scalers1-aggregators1-32-32] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-7.4-scalers1-aggregators2-16-16] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-7.4-scalers1-aggregators2-16-32] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-7.4-scalers1-aggregators2-32-16] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.0-7.4-scalers1-aggregators2-32-32] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-2.5-scalers0-aggregators0-16-16] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-2.5-scalers0-aggregators0-16-32] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-2.5-scalers0-aggregators0-32-16] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-2.5-scalers0-aggregators0-32-32] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-2.5-scalers0-aggregators1-16-16] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-2.5-scalers0-aggregators1-16-32] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-2.5-scalers0-aggregators1-32-16] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-2.5-scalers0-aggregators1-32-32] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-2.5-scalers0-aggregators2-16-16] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-2.5-scalers0-aggregators2-16-32] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-2.5-scalers0-aggregators2-32-16] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-2.5-scalers0-aggregators2-32-32] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-2.5-scalers1-aggregators0-16-16] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-2.5-scalers1-aggregators0-16-32] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-2.5-scalers1-aggregators0-32-16] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-2.5-scalers1-aggregators0-32-32] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-2.5-scalers1-aggregators1-16-16] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-2.5-scalers1-aggregators1-16-32] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-2.5-scalers1-aggregators1-32-16] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-2.5-scalers1-aggregators1-32-32] PASSED [ 56%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-2.5-scalers1-aggregators2-16-16] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-2.5-scalers1-aggregators2-16-32] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-2.5-scalers1-aggregators2-32-16] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-2.5-scalers1-aggregators2-32-32] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-7.4-scalers0-aggregators0-16-16] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-7.4-scalers0-aggregators0-16-32] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-7.4-scalers0-aggregators0-32-16] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-7.4-scalers0-aggregators0-32-32] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-7.4-scalers0-aggregators1-16-16] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-7.4-scalers0-aggregators1-16-32] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-7.4-scalers0-aggregators1-32-16] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-7.4-scalers0-aggregators1-32-32] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-7.4-scalers0-aggregators2-16-16] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-7.4-scalers0-aggregators2-16-32] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-7.4-scalers0-aggregators2-32-16] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-7.4-scalers0-aggregators2-32-32] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-7.4-scalers1-aggregators0-16-16] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-7.4-scalers1-aggregators0-16-32] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-7.4-scalers1-aggregators0-32-16] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-7.4-scalers1-aggregators0-32-32] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-7.4-scalers1-aggregators1-16-16] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-7.4-scalers1-aggregators1-16-32] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-7.4-scalers1-aggregators1-32-16] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-7.4-scalers1-aggregators1-32-32] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-7.4-scalers1-aggregators2-16-16] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-7.4-scalers1-aggregators2-16-32] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-7.4-scalers1-aggregators2-32-16] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-1-0.1-7.4-scalers1-aggregators2-32-32] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-2.5-scalers0-aggregators0-16-16] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-2.5-scalers0-aggregators0-16-32] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-2.5-scalers0-aggregators0-32-16] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-2.5-scalers0-aggregators0-32-32] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-2.5-scalers0-aggregators1-16-16] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-2.5-scalers0-aggregators1-16-32] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-2.5-scalers0-aggregators1-32-16] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-2.5-scalers0-aggregators1-32-32] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-2.5-scalers0-aggregators2-16-16] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-2.5-scalers0-aggregators2-16-32] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-2.5-scalers0-aggregators2-32-16] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-2.5-scalers0-aggregators2-32-32] PASSED [ 57%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-2.5-scalers1-aggregators0-16-16] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-2.5-scalers1-aggregators0-16-32] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-2.5-scalers1-aggregators0-32-16] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-2.5-scalers1-aggregators0-32-32] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-2.5-scalers1-aggregators1-16-16] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-2.5-scalers1-aggregators1-16-32] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-2.5-scalers1-aggregators1-32-16] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-2.5-scalers1-aggregators1-32-32] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-2.5-scalers1-aggregators2-16-16] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-2.5-scalers1-aggregators2-16-32] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-2.5-scalers1-aggregators2-32-16] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-2.5-scalers1-aggregators2-32-32] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-7.4-scalers0-aggregators0-16-16] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-7.4-scalers0-aggregators0-16-32] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-7.4-scalers0-aggregators0-32-16] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-7.4-scalers0-aggregators0-32-32] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-7.4-scalers0-aggregators1-16-16] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-7.4-scalers0-aggregators1-16-32] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-7.4-scalers0-aggregators1-32-16] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-7.4-scalers0-aggregators1-32-32] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-7.4-scalers0-aggregators2-16-16] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-7.4-scalers0-aggregators2-16-32] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-7.4-scalers0-aggregators2-32-16] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-7.4-scalers0-aggregators2-32-32] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-7.4-scalers1-aggregators0-16-16] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-7.4-scalers1-aggregators0-16-32] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-7.4-scalers1-aggregators0-32-16] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-7.4-scalers1-aggregators0-32-32] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-7.4-scalers1-aggregators1-16-16] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-7.4-scalers1-aggregators1-16-32] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-7.4-scalers1-aggregators1-32-16] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-7.4-scalers1-aggregators1-32-32] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-7.4-scalers1-aggregators2-16-16] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-7.4-scalers1-aggregators2-16-32] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-7.4-scalers1-aggregators2-32-16] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.0-7.4-scalers1-aggregators2-32-32] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-2.5-scalers0-aggregators0-16-16] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-2.5-scalers0-aggregators0-16-32] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-2.5-scalers0-aggregators0-32-16] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-2.5-scalers0-aggregators0-32-32] PASSED [ 58%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-2.5-scalers0-aggregators1-16-16] PASSED [ 59%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-2.5-scalers0-aggregators1-16-32] PASSED [ 59%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-2.5-scalers0-aggregators1-32-16] PASSED [ 59%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-2.5-scalers0-aggregators1-32-32] PASSED [ 59%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-2.5-scalers0-aggregators2-16-16] PASSED [ 59%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-2.5-scalers0-aggregators2-16-32] PASSED [ 59%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-2.5-scalers0-aggregators2-32-16] PASSED [ 59%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-2.5-scalers0-aggregators2-32-32] PASSED [ 59%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-2.5-scalers1-aggregators0-16-16] PASSED [ 59%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-2.5-scalers1-aggregators0-16-32] PASSED [ 59%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-2.5-scalers1-aggregators0-32-16] PASSED [ 59%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-2.5-scalers1-aggregators0-32-32] PASSED [ 59%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-2.5-scalers1-aggregators1-16-16] PASSED [ 59%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-2.5-scalers1-aggregators1-16-32] PASSED [ 59%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-2.5-scalers1-aggregators1-32-16] PASSED [ 59%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-2.5-scalers1-aggregators1-32-32] PASSED [ 59%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-2.5-scalers1-aggregators2-16-16] PASSED [ 59%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-2.5-scalers1-aggregators2-16-32] PASSED [ 59%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-2.5-scalers1-aggregators2-32-16] PASSED [ 59%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-2.5-scalers1-aggregators2-32-32] PASSED [ 59%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-7.4-scalers0-aggregators0-16-16] PASSED [ 59%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-7.4-scalers0-aggregators0-16-32] PASSED [ 59%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-7.4-scalers0-aggregators0-32-16] PASSED [ 59%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-7.4-scalers0-aggregators0-32-32] PASSED [ 59%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-7.4-scalers0-aggregators1-16-16] PASSED [ 59%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-7.4-scalers0-aggregators1-16-32] PASSED [ 59%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-7.4-scalers0-aggregators1-32-16] PASSED [ 59%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-7.4-scalers0-aggregators1-32-32] PASSED [ 59%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-7.4-scalers0-aggregators2-16-16] PASSED [ 59%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-7.4-scalers0-aggregators2-16-32] PASSED [ 59%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-7.4-scalers0-aggregators2-32-16] PASSED [ 59%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-7.4-scalers0-aggregators2-32-32] PASSED [ 59%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-7.4-scalers1-aggregators0-16-16] PASSED [ 59%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-7.4-scalers1-aggregators0-16-32] PASSED [ 59%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-7.4-scalers1-aggregators0-32-16] PASSED [ 59%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-7.4-scalers1-aggregators0-32-32] PASSED [ 59%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-7.4-scalers1-aggregators1-16-16] PASSED [ 59%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-7.4-scalers1-aggregators1-16-32] PASSED [ 59%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-7.4-scalers1-aggregators1-32-16] PASSED [ 59%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-7.4-scalers1-aggregators1-32-32] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-7.4-scalers1-aggregators2-16-16] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-7.4-scalers1-aggregators2-16-32] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-7.4-scalers1-aggregators2-32-16] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[True-0-4-0.1-7.4-scalers1-aggregators2-32-32] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-2.5-scalers0-aggregators0-16-16] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-2.5-scalers0-aggregators0-16-32] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-2.5-scalers0-aggregators0-32-16] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-2.5-scalers0-aggregators0-32-32] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-2.5-scalers0-aggregators1-16-16] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-2.5-scalers0-aggregators1-16-32] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-2.5-scalers0-aggregators1-32-16] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-2.5-scalers0-aggregators1-32-32] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-2.5-scalers0-aggregators2-16-16] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-2.5-scalers0-aggregators2-16-32] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-2.5-scalers0-aggregators2-32-16] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-2.5-scalers0-aggregators2-32-32] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-2.5-scalers1-aggregators0-16-16] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-2.5-scalers1-aggregators0-16-32] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-2.5-scalers1-aggregators0-32-16] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-2.5-scalers1-aggregators0-32-32] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-2.5-scalers1-aggregators1-16-16] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-2.5-scalers1-aggregators1-16-32] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-2.5-scalers1-aggregators1-32-16] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-2.5-scalers1-aggregators1-32-32] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-2.5-scalers1-aggregators2-16-16] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-2.5-scalers1-aggregators2-16-32] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-2.5-scalers1-aggregators2-32-16] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-2.5-scalers1-aggregators2-32-32] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-7.4-scalers0-aggregators0-16-16] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-7.4-scalers0-aggregators0-16-32] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-7.4-scalers0-aggregators0-32-16] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-7.4-scalers0-aggregators0-32-32] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-7.4-scalers0-aggregators1-16-16] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-7.4-scalers0-aggregators1-16-32] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-7.4-scalers0-aggregators1-32-16] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-7.4-scalers0-aggregators1-32-32] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-7.4-scalers0-aggregators2-16-16] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-7.4-scalers0-aggregators2-16-32] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-7.4-scalers0-aggregators2-32-16] PASSED [ 60%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-7.4-scalers0-aggregators2-32-32] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-7.4-scalers1-aggregators0-16-16] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-7.4-scalers1-aggregators0-16-32] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-7.4-scalers1-aggregators0-32-16] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-7.4-scalers1-aggregators0-32-32] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-7.4-scalers1-aggregators1-16-16] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-7.4-scalers1-aggregators1-16-32] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-7.4-scalers1-aggregators1-32-16] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-7.4-scalers1-aggregators1-32-32] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-7.4-scalers1-aggregators2-16-16] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-7.4-scalers1-aggregators2-16-32] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-7.4-scalers1-aggregators2-32-16] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.0-7.4-scalers1-aggregators2-32-32] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-2.5-scalers0-aggregators0-16-16] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-2.5-scalers0-aggregators0-16-32] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-2.5-scalers0-aggregators0-32-16] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-2.5-scalers0-aggregators0-32-32] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-2.5-scalers0-aggregators1-16-16] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-2.5-scalers0-aggregators1-16-32] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-2.5-scalers0-aggregators1-32-16] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-2.5-scalers0-aggregators1-32-32] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-2.5-scalers0-aggregators2-16-16] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-2.5-scalers0-aggregators2-16-32] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-2.5-scalers0-aggregators2-32-16] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-2.5-scalers0-aggregators2-32-32] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-2.5-scalers1-aggregators0-16-16] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-2.5-scalers1-aggregators0-16-32] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-2.5-scalers1-aggregators0-32-16] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-2.5-scalers1-aggregators0-32-32] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-2.5-scalers1-aggregators1-16-16] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-2.5-scalers1-aggregators1-16-32] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-2.5-scalers1-aggregators1-32-16] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-2.5-scalers1-aggregators1-32-32] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-2.5-scalers1-aggregators2-16-16] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-2.5-scalers1-aggregators2-16-32] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-2.5-scalers1-aggregators2-32-16] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-2.5-scalers1-aggregators2-32-32] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-7.4-scalers0-aggregators0-16-16] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-7.4-scalers0-aggregators0-16-32] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-7.4-scalers0-aggregators0-32-16] PASSED [ 61%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-7.4-scalers0-aggregators0-32-32] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-7.4-scalers0-aggregators1-16-16] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-7.4-scalers0-aggregators1-16-32] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-7.4-scalers0-aggregators1-32-16] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-7.4-scalers0-aggregators1-32-32] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-7.4-scalers0-aggregators2-16-16] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-7.4-scalers0-aggregators2-16-32] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-7.4-scalers0-aggregators2-32-16] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-7.4-scalers0-aggregators2-32-32] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-7.4-scalers1-aggregators0-16-16] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-7.4-scalers1-aggregators0-16-32] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-7.4-scalers1-aggregators0-32-16] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-7.4-scalers1-aggregators0-32-32] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-7.4-scalers1-aggregators1-16-16] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-7.4-scalers1-aggregators1-16-32] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-7.4-scalers1-aggregators1-32-16] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-7.4-scalers1-aggregators1-32-32] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-7.4-scalers1-aggregators2-16-16] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-7.4-scalers1-aggregators2-16-32] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-7.4-scalers1-aggregators2-32-16] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-1-0.1-7.4-scalers1-aggregators2-32-32] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-2.5-scalers0-aggregators0-16-16] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-2.5-scalers0-aggregators0-16-32] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-2.5-scalers0-aggregators0-32-16] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-2.5-scalers0-aggregators0-32-32] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-2.5-scalers0-aggregators1-16-16] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-2.5-scalers0-aggregators1-16-32] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-2.5-scalers0-aggregators1-32-16] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-2.5-scalers0-aggregators1-32-32] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-2.5-scalers0-aggregators2-16-16] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-2.5-scalers0-aggregators2-16-32] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-2.5-scalers0-aggregators2-32-16] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-2.5-scalers0-aggregators2-32-32] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-2.5-scalers1-aggregators0-16-16] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-2.5-scalers1-aggregators0-16-32] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-2.5-scalers1-aggregators0-32-16] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-2.5-scalers1-aggregators0-32-32] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-2.5-scalers1-aggregators1-16-16] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-2.5-scalers1-aggregators1-16-32] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-2.5-scalers1-aggregators1-32-16] PASSED [ 62%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-2.5-scalers1-aggregators1-32-32] PASSED [ 63%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-2.5-scalers1-aggregators2-16-16] PASSED [ 63%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-2.5-scalers1-aggregators2-16-32] PASSED [ 63%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-2.5-scalers1-aggregators2-32-16] PASSED [ 63%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-2.5-scalers1-aggregators2-32-32] PASSED [ 63%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-7.4-scalers0-aggregators0-16-16] PASSED [ 63%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-7.4-scalers0-aggregators0-16-32] PASSED [ 63%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-7.4-scalers0-aggregators0-32-16] PASSED [ 63%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-7.4-scalers0-aggregators0-32-32] PASSED [ 63%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-7.4-scalers0-aggregators1-16-16] PASSED [ 63%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-7.4-scalers0-aggregators1-16-32] PASSED [ 63%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-7.4-scalers0-aggregators1-32-16] PASSED [ 63%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-7.4-scalers0-aggregators1-32-32] PASSED [ 63%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-7.4-scalers0-aggregators2-16-16] PASSED [ 63%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-7.4-scalers0-aggregators2-16-32] PASSED [ 63%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-7.4-scalers0-aggregators2-32-16] PASSED [ 63%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-7.4-scalers0-aggregators2-32-32] PASSED [ 63%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-7.4-scalers1-aggregators0-16-16] PASSED [ 63%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-7.4-scalers1-aggregators0-16-32] PASSED [ 63%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-7.4-scalers1-aggregators0-32-16] PASSED [ 63%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-7.4-scalers1-aggregators0-32-32] PASSED [ 63%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-7.4-scalers1-aggregators1-16-16] PASSED [ 63%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-7.4-scalers1-aggregators1-16-32] PASSED [ 63%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-7.4-scalers1-aggregators1-32-16] PASSED [ 63%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-7.4-scalers1-aggregators1-32-32] PASSED [ 63%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-7.4-scalers1-aggregators2-16-16] PASSED [ 63%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-7.4-scalers1-aggregators2-16-32] PASSED [ 63%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-7.4-scalers1-aggregators2-32-16] PASSED [ 63%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.0-7.4-scalers1-aggregators2-32-32] PASSED [ 63%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-2.5-scalers0-aggregators0-16-16] PASSED [ 63%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-2.5-scalers0-aggregators0-16-32] PASSED [ 63%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-2.5-scalers0-aggregators0-32-16] PASSED [ 63%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-2.5-scalers0-aggregators0-32-32] PASSED [ 63%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-2.5-scalers0-aggregators1-16-16] PASSED [ 63%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-2.5-scalers0-aggregators1-16-32] PASSED [ 63%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-2.5-scalers0-aggregators1-32-16] PASSED [ 63%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-2.5-scalers0-aggregators1-32-32] PASSED [ 63%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-2.5-scalers0-aggregators2-16-16] PASSED [ 63%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-2.5-scalers0-aggregators2-16-32] PASSED [ 63%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-2.5-scalers0-aggregators2-32-16] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-2.5-scalers0-aggregators2-32-32] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-2.5-scalers1-aggregators0-16-16] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-2.5-scalers1-aggregators0-16-32] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-2.5-scalers1-aggregators0-32-16] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-2.5-scalers1-aggregators0-32-32] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-2.5-scalers1-aggregators1-16-16] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-2.5-scalers1-aggregators1-16-32] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-2.5-scalers1-aggregators1-32-16] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-2.5-scalers1-aggregators1-32-32] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-2.5-scalers1-aggregators2-16-16] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-2.5-scalers1-aggregators2-16-32] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-2.5-scalers1-aggregators2-32-16] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-2.5-scalers1-aggregators2-32-32] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-7.4-scalers0-aggregators0-16-16] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-7.4-scalers0-aggregators0-16-32] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-7.4-scalers0-aggregators0-32-16] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-7.4-scalers0-aggregators0-32-32] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-7.4-scalers0-aggregators1-16-16] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-7.4-scalers0-aggregators1-16-32] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-7.4-scalers0-aggregators1-32-16] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-7.4-scalers0-aggregators1-32-32] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-7.4-scalers0-aggregators2-16-16] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-7.4-scalers0-aggregators2-16-32] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-7.4-scalers0-aggregators2-32-16] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-7.4-scalers0-aggregators2-32-32] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-7.4-scalers1-aggregators0-16-16] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-7.4-scalers1-aggregators0-16-32] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-7.4-scalers1-aggregators0-32-16] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-7.4-scalers1-aggregators0-32-32] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-7.4-scalers1-aggregators1-16-16] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-7.4-scalers1-aggregators1-16-32] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-7.4-scalers1-aggregators1-32-16] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-7.4-scalers1-aggregators1-32-32] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-7.4-scalers1-aggregators2-16-16] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-7.4-scalers1-aggregators2-16-32] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-7.4-scalers1-aggregators2-32-16] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-16-4-0.1-7.4-scalers1-aggregators2-32-32] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-2.5-scalers0-aggregators0-16-16] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-2.5-scalers0-aggregators0-16-32] PASSED [ 64%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-2.5-scalers0-aggregators0-32-16] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-2.5-scalers0-aggregators0-32-32] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-2.5-scalers0-aggregators1-16-16] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-2.5-scalers0-aggregators1-16-32] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-2.5-scalers0-aggregators1-32-16] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-2.5-scalers0-aggregators1-32-32] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-2.5-scalers0-aggregators2-16-16] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-2.5-scalers0-aggregators2-16-32] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-2.5-scalers0-aggregators2-32-16] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-2.5-scalers0-aggregators2-32-32] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-2.5-scalers1-aggregators0-16-16] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-2.5-scalers1-aggregators0-16-32] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-2.5-scalers1-aggregators0-32-16] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-2.5-scalers1-aggregators0-32-32] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-2.5-scalers1-aggregators1-16-16] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-2.5-scalers1-aggregators1-16-32] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-2.5-scalers1-aggregators1-32-16] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-2.5-scalers1-aggregators1-32-32] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-2.5-scalers1-aggregators2-16-16] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-2.5-scalers1-aggregators2-16-32] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-2.5-scalers1-aggregators2-32-16] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-2.5-scalers1-aggregators2-32-32] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-7.4-scalers0-aggregators0-16-16] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-7.4-scalers0-aggregators0-16-32] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-7.4-scalers0-aggregators0-32-16] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-7.4-scalers0-aggregators0-32-32] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-7.4-scalers0-aggregators1-16-16] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-7.4-scalers0-aggregators1-16-32] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-7.4-scalers0-aggregators1-32-16] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-7.4-scalers0-aggregators1-32-32] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-7.4-scalers0-aggregators2-16-16] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-7.4-scalers0-aggregators2-16-32] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-7.4-scalers0-aggregators2-32-16] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-7.4-scalers0-aggregators2-32-32] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-7.4-scalers1-aggregators0-16-16] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-7.4-scalers1-aggregators0-16-32] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-7.4-scalers1-aggregators0-32-16] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-7.4-scalers1-aggregators0-32-32] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-7.4-scalers1-aggregators1-16-16] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-7.4-scalers1-aggregators1-16-32] PASSED [ 65%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-7.4-scalers1-aggregators1-32-16] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-7.4-scalers1-aggregators1-32-32] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-7.4-scalers1-aggregators2-16-16] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-7.4-scalers1-aggregators2-16-32] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-7.4-scalers1-aggregators2-32-16] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.0-7.4-scalers1-aggregators2-32-32] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-2.5-scalers0-aggregators0-16-16] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-2.5-scalers0-aggregators0-16-32] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-2.5-scalers0-aggregators0-32-16] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-2.5-scalers0-aggregators0-32-32] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-2.5-scalers0-aggregators1-16-16] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-2.5-scalers0-aggregators1-16-32] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-2.5-scalers0-aggregators1-32-16] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-2.5-scalers0-aggregators1-32-32] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-2.5-scalers0-aggregators2-16-16] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-2.5-scalers0-aggregators2-16-32] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-2.5-scalers0-aggregators2-32-16] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-2.5-scalers0-aggregators2-32-32] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-2.5-scalers1-aggregators0-16-16] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-2.5-scalers1-aggregators0-16-32] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-2.5-scalers1-aggregators0-32-16] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-2.5-scalers1-aggregators0-32-32] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-2.5-scalers1-aggregators1-16-16] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-2.5-scalers1-aggregators1-16-32] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-2.5-scalers1-aggregators1-32-16] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-2.5-scalers1-aggregators1-32-32] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-2.5-scalers1-aggregators2-16-16] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-2.5-scalers1-aggregators2-16-32] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-2.5-scalers1-aggregators2-32-16] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-2.5-scalers1-aggregators2-32-32] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-7.4-scalers0-aggregators0-16-16] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-7.4-scalers0-aggregators0-16-32] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-7.4-scalers0-aggregators0-32-16] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-7.4-scalers0-aggregators0-32-32] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-7.4-scalers0-aggregators1-16-16] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-7.4-scalers0-aggregators1-16-32] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-7.4-scalers0-aggregators1-32-16] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-7.4-scalers0-aggregators1-32-32] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-7.4-scalers0-aggregators2-16-16] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-7.4-scalers0-aggregators2-16-32] PASSED [ 66%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-7.4-scalers0-aggregators2-32-16] PASSED [ 67%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-7.4-scalers0-aggregators2-32-32] PASSED [ 67%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-7.4-scalers1-aggregators0-16-16] PASSED [ 67%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-7.4-scalers1-aggregators0-16-32] PASSED [ 67%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-7.4-scalers1-aggregators0-32-16] PASSED [ 67%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-7.4-scalers1-aggregators0-32-32] PASSED [ 67%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-7.4-scalers1-aggregators1-16-16] PASSED [ 67%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-7.4-scalers1-aggregators1-16-32] PASSED [ 67%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-7.4-scalers1-aggregators1-32-16] PASSED [ 67%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-7.4-scalers1-aggregators1-32-32] PASSED [ 67%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-7.4-scalers1-aggregators2-16-16] PASSED [ 67%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-7.4-scalers1-aggregators2-16-32] PASSED [ 67%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-7.4-scalers1-aggregators2-32-16] PASSED [ 67%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-1-0.1-7.4-scalers1-aggregators2-32-32] PASSED [ 67%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-2.5-scalers0-aggregators0-16-16] PASSED [ 67%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-2.5-scalers0-aggregators0-16-32] PASSED [ 67%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-2.5-scalers0-aggregators0-32-16] PASSED [ 67%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-2.5-scalers0-aggregators0-32-32] PASSED [ 67%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-2.5-scalers0-aggregators1-16-16] PASSED [ 67%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-2.5-scalers0-aggregators1-16-32] PASSED [ 67%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-2.5-scalers0-aggregators1-32-16] PASSED [ 67%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-2.5-scalers0-aggregators1-32-32] PASSED [ 67%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-2.5-scalers0-aggregators2-16-16] PASSED [ 67%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-2.5-scalers0-aggregators2-16-32] PASSED [ 67%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-2.5-scalers0-aggregators2-32-16] PASSED [ 67%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-2.5-scalers0-aggregators2-32-32] PASSED [ 67%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-2.5-scalers1-aggregators0-16-16] PASSED [ 67%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-2.5-scalers1-aggregators0-16-32] PASSED [ 67%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-2.5-scalers1-aggregators0-32-16] PASSED [ 67%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-2.5-scalers1-aggregators0-32-32] PASSED [ 67%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-2.5-scalers1-aggregators1-16-16] PASSED [ 67%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-2.5-scalers1-aggregators1-16-32] PASSED [ 67%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-2.5-scalers1-aggregators1-32-16] PASSED [ 67%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-2.5-scalers1-aggregators1-32-32] PASSED [ 67%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-2.5-scalers1-aggregators2-16-16] PASSED [ 67%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-2.5-scalers1-aggregators2-16-32] PASSED [ 67%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-2.5-scalers1-aggregators2-32-16] PASSED [ 67%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-2.5-scalers1-aggregators2-32-32] PASSED [ 67%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-7.4-scalers0-aggregators0-16-16] PASSED [ 67%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-7.4-scalers0-aggregators0-16-32] PASSED [ 67%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-7.4-scalers0-aggregators0-32-16] PASSED [ 68%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-7.4-scalers0-aggregators0-32-32] PASSED [ 68%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-7.4-scalers0-aggregators1-16-16] PASSED [ 68%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-7.4-scalers0-aggregators1-16-32] PASSED [ 68%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-7.4-scalers0-aggregators1-32-16] PASSED [ 68%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-7.4-scalers0-aggregators1-32-32] PASSED [ 68%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-7.4-scalers0-aggregators2-16-16] PASSED [ 68%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-7.4-scalers0-aggregators2-16-32] PASSED [ 68%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-7.4-scalers0-aggregators2-32-16] PASSED [ 68%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-7.4-scalers0-aggregators2-32-32] PASSED [ 68%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-7.4-scalers1-aggregators0-16-16] PASSED [ 68%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-7.4-scalers1-aggregators0-16-32] PASSED [ 68%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-7.4-scalers1-aggregators0-32-16] PASSED [ 68%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-7.4-scalers1-aggregators0-32-32] PASSED [ 68%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-7.4-scalers1-aggregators1-16-16] PASSED [ 68%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-7.4-scalers1-aggregators1-16-32] PASSED [ 68%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-7.4-scalers1-aggregators1-32-16] PASSED [ 68%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-7.4-scalers1-aggregators1-32-32] PASSED [ 68%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-7.4-scalers1-aggregators2-16-16] PASSED [ 68%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-7.4-scalers1-aggregators2-16-32] PASSED [ 68%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-7.4-scalers1-aggregators2-32-16] PASSED [ 68%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.0-7.4-scalers1-aggregators2-32-32] PASSED [ 68%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-2.5-scalers0-aggregators0-16-16] PASSED [ 68%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-2.5-scalers0-aggregators0-16-32] PASSED [ 68%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-2.5-scalers0-aggregators0-32-16] PASSED [ 68%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-2.5-scalers0-aggregators0-32-32] PASSED [ 68%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-2.5-scalers0-aggregators1-16-16] PASSED [ 68%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-2.5-scalers0-aggregators1-16-32] PASSED [ 68%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-2.5-scalers0-aggregators1-32-16] PASSED [ 68%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-2.5-scalers0-aggregators1-32-32] PASSED [ 68%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-2.5-scalers0-aggregators2-16-16] PASSED [ 68%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-2.5-scalers0-aggregators2-16-32] PASSED [ 68%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-2.5-scalers0-aggregators2-32-16] PASSED [ 68%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-2.5-scalers0-aggregators2-32-32] PASSED [ 68%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-2.5-scalers1-aggregators0-16-16] PASSED [ 68%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-2.5-scalers1-aggregators0-16-32] PASSED [ 68%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-2.5-scalers1-aggregators0-32-16] PASSED [ 68%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-2.5-scalers1-aggregators0-32-32] PASSED [ 68%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-2.5-scalers1-aggregators1-16-16] PASSED [ 68%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-2.5-scalers1-aggregators1-16-32] PASSED [ 69%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-2.5-scalers1-aggregators1-32-16] PASSED [ 69%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-2.5-scalers1-aggregators1-32-32] PASSED [ 69%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-2.5-scalers1-aggregators2-16-16] PASSED [ 69%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-2.5-scalers1-aggregators2-16-32] PASSED [ 69%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-2.5-scalers1-aggregators2-32-16] PASSED [ 69%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-2.5-scalers1-aggregators2-32-32] PASSED [ 69%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-7.4-scalers0-aggregators0-16-16] PASSED [ 69%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-7.4-scalers0-aggregators0-16-32] PASSED [ 69%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-7.4-scalers0-aggregators0-32-16] PASSED [ 69%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-7.4-scalers0-aggregators0-32-32] PASSED [ 69%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-7.4-scalers0-aggregators1-16-16] PASSED [ 69%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-7.4-scalers0-aggregators1-16-32] PASSED [ 69%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-7.4-scalers0-aggregators1-32-16] PASSED [ 69%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-7.4-scalers0-aggregators1-32-32] PASSED [ 69%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-7.4-scalers0-aggregators2-16-16] PASSED [ 69%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-7.4-scalers0-aggregators2-16-32] PASSED [ 69%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-7.4-scalers0-aggregators2-32-16] PASSED [ 69%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-7.4-scalers0-aggregators2-32-32] PASSED [ 69%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-7.4-scalers1-aggregators0-16-16] PASSED [ 69%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-7.4-scalers1-aggregators0-16-32] PASSED [ 69%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-7.4-scalers1-aggregators0-32-16] PASSED [ 69%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-7.4-scalers1-aggregators0-32-32] PASSED [ 69%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-7.4-scalers1-aggregators1-16-16] PASSED [ 69%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-7.4-scalers1-aggregators1-16-32] PASSED [ 69%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-7.4-scalers1-aggregators1-32-16] PASSED [ 69%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-7.4-scalers1-aggregators1-32-32] PASSED [ 69%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-7.4-scalers1-aggregators2-16-16] PASSED [ 69%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-7.4-scalers1-aggregators2-16-32] PASSED [ 69%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-7.4-scalers1-aggregators2-32-16] PASSED [ 69%]
tests/pytorch/test_nn.py::test_pna_conv[False-0-4-0.1-7.4-scalers1-aggregators2-32-32] PASSED [ 69%]
tests/pytorch/test_nn.py::test_label_prop[True-True-True-sym-0.0-3] PASSED [ 69%]
tests/pytorch/test_nn.py::test_label_prop[True-True-True-sym-0.0-5] PASSED [ 69%]
tests/pytorch/test_nn.py::test_label_prop[True-True-True-sym-0.5-3] PASSED [ 69%]
tests/pytorch/test_nn.py::test_label_prop[True-True-True-sym-0.5-5] PASSED [ 69%]
tests/pytorch/test_nn.py::test_label_prop[True-True-True-sym-1.0-3] PASSED [ 69%]
tests/pytorch/test_nn.py::test_label_prop[True-True-True-sym-1.0-5] PASSED [ 69%]
tests/pytorch/test_nn.py::test_label_prop[True-True-True-row-0.0-3] PASSED [ 69%]
tests/pytorch/test_nn.py::test_label_prop[True-True-True-row-0.0-5] PASSED [ 69%]
tests/pytorch/test_nn.py::test_label_prop[True-True-True-row-0.5-3] PASSED [ 69%]
tests/pytorch/test_nn.py::test_label_prop[True-True-True-row-0.5-5] PASSED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[True-True-True-row-1.0-3] PASSED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[True-True-True-row-1.0-5] PASSED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[True-True-False-sym-0.0-3] PASSED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[True-True-False-sym-0.0-5] PASSED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[True-True-False-sym-0.5-3] PASSED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[True-True-False-sym-0.5-5] PASSED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[True-True-False-sym-1.0-3] PASSED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[True-True-False-sym-1.0-5] PASSED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[True-True-False-row-0.0-3] PASSED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[True-True-False-row-0.0-5] PASSED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[True-True-False-row-0.5-3] PASSED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[True-True-False-row-0.5-5] PASSED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[True-True-False-row-1.0-3] PASSED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[True-True-False-row-1.0-5] PASSED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[True-False-True-sym-0.0-3] PASSED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[True-False-True-sym-0.0-5] PASSED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[True-False-True-sym-0.5-3] PASSED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[True-False-True-sym-0.5-5] PASSED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[True-False-True-sym-1.0-3] PASSED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[True-False-True-sym-1.0-5] PASSED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[True-False-True-row-0.0-3] PASSED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[True-False-True-row-0.0-5] PASSED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[True-False-True-row-0.5-3] PASSED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[True-False-True-row-0.5-5] PASSED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[True-False-True-row-1.0-3] PASSED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[True-False-True-row-1.0-5] PASSED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[True-False-False-sym-0.0-3] PASSED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[True-False-False-sym-0.0-5] PASSED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[True-False-False-sym-0.5-3] PASSED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[True-False-False-sym-0.5-5] PASSED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[True-False-False-sym-1.0-3] PASSED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[True-False-False-sym-1.0-5] PASSED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[True-False-False-row-0.0-3] PASSED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[True-False-False-row-0.0-5] PASSED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[True-False-False-row-0.5-3] PASSED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[True-False-False-row-0.5-5] PASSED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[True-False-False-row-1.0-3] PASSED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[True-False-False-row-1.0-5] PASSED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[False-True-True-sym-0.0-3] PASSED [ 70%]
tests/pytorch/test_nn.py::test_label_prop[False-True-True-sym-0.0-5] PASSED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-True-True-sym-0.5-3] PASSED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-True-True-sym-0.5-5] PASSED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-True-True-sym-1.0-3] PASSED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-True-True-sym-1.0-5] PASSED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-True-True-row-0.0-3] PASSED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-True-True-row-0.0-5] PASSED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-True-True-row-0.5-3] PASSED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-True-True-row-0.5-5] PASSED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-True-True-row-1.0-3] PASSED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-True-True-row-1.0-5] PASSED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-True-False-sym-0.0-3] PASSED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-True-False-sym-0.0-5] PASSED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-True-False-sym-0.5-3] PASSED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-True-False-sym-0.5-5] PASSED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-True-False-sym-1.0-3] PASSED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-True-False-sym-1.0-5] PASSED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-True-False-row-0.0-3] PASSED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-True-False-row-0.0-5] PASSED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-True-False-row-0.5-3] PASSED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-True-False-row-0.5-5] PASSED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-True-False-row-1.0-3] PASSED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-True-False-row-1.0-5] PASSED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-False-True-sym-0.0-3] PASSED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-False-True-sym-0.0-5] PASSED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-False-True-sym-0.5-3] PASSED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-False-True-sym-0.5-5] PASSED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-False-True-sym-1.0-3] PASSED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-False-True-sym-1.0-5] PASSED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-False-True-row-0.0-3] PASSED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-False-True-row-0.0-5] PASSED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-False-True-row-0.5-3] PASSED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-False-True-row-0.5-5] PASSED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-False-True-row-1.0-3] PASSED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-False-True-row-1.0-5] PASSED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-False-False-sym-0.0-3] PASSED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-False-False-sym-0.0-5] PASSED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-False-False-sym-0.5-3] PASSED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-False-False-sym-0.5-5] PASSED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-False-False-sym-1.0-3] PASSED [ 71%]
tests/pytorch/test_nn.py::test_label_prop[False-False-False-sym-1.0-5] PASSED [ 72%]
tests/pytorch/test_nn.py::test_label_prop[False-False-False-row-0.0-3] PASSED [ 72%]
tests/pytorch/test_nn.py::test_label_prop[False-False-False-row-0.0-5] PASSED [ 72%]
tests/pytorch/test_nn.py::test_label_prop[False-False-False-row-0.5-3] PASSED [ 72%]
tests/pytorch/test_nn.py::test_label_prop[False-False-False-row-0.5-5] PASSED [ 72%]
tests/pytorch/test_nn.py::test_label_prop[False-False-False-row-1.0-3] PASSED [ 72%]
tests/pytorch/test_nn.py::test_label_prop[False-False-False-row-1.0-5] PASSED [ 72%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-2.5-scalers0-aggregators0-16-16] PASSED [ 72%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-2.5-scalers0-aggregators0-16-32] PASSED [ 72%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-2.5-scalers0-aggregators0-32-16] PASSED [ 72%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-2.5-scalers0-aggregators0-32-32] PASSED [ 72%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-2.5-scalers0-aggregators1-16-16] PASSED [ 72%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-2.5-scalers0-aggregators1-16-32] PASSED [ 72%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-2.5-scalers0-aggregators1-32-16] PASSED [ 72%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-2.5-scalers0-aggregators1-32-32] PASSED [ 72%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-2.5-scalers0-aggregators2-16-16] PASSED [ 72%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-2.5-scalers0-aggregators2-16-32] PASSED [ 72%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-2.5-scalers0-aggregators2-32-16] PASSED [ 72%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-2.5-scalers0-aggregators2-32-32] PASSED [ 72%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-2.5-scalers1-aggregators0-16-16] PASSED [ 72%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-2.5-scalers1-aggregators0-16-32] PASSED [ 72%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-2.5-scalers1-aggregators0-32-16] PASSED [ 72%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-2.5-scalers1-aggregators0-32-32] PASSED [ 72%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-2.5-scalers1-aggregators1-16-16] PASSED [ 72%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-2.5-scalers1-aggregators1-16-32] PASSED [ 72%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-2.5-scalers1-aggregators1-32-16] PASSED [ 72%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-2.5-scalers1-aggregators1-32-32] PASSED [ 72%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-2.5-scalers1-aggregators2-16-16] PASSED [ 72%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-2.5-scalers1-aggregators2-16-32] PASSED [ 72%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-2.5-scalers1-aggregators2-32-16] PASSED [ 72%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-2.5-scalers1-aggregators2-32-32] PASSED [ 72%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-7.4-scalers0-aggregators0-16-16] PASSED [ 72%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-7.4-scalers0-aggregators0-16-32] PASSED [ 72%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-7.4-scalers0-aggregators0-32-16] PASSED [ 72%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-7.4-scalers0-aggregators0-32-32] PASSED [ 72%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-7.4-scalers0-aggregators1-16-16] PASSED [ 72%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-7.4-scalers0-aggregators1-16-32] PASSED [ 72%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-7.4-scalers0-aggregators1-32-16] PASSED [ 72%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-7.4-scalers0-aggregators1-32-32] PASSED [ 72%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-7.4-scalers0-aggregators2-16-16] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-7.4-scalers0-aggregators2-16-32] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-7.4-scalers0-aggregators2-32-16] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-7.4-scalers0-aggregators2-32-32] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-7.4-scalers1-aggregators0-16-16] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-7.4-scalers1-aggregators0-16-32] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-7.4-scalers1-aggregators0-32-16] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-7.4-scalers1-aggregators0-32-32] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-7.4-scalers1-aggregators1-16-16] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-7.4-scalers1-aggregators1-16-32] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-7.4-scalers1-aggregators1-32-16] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-7.4-scalers1-aggregators1-32-32] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-7.4-scalers1-aggregators2-16-16] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-7.4-scalers1-aggregators2-16-32] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-7.4-scalers1-aggregators2-32-16] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.0-7.4-scalers1-aggregators2-32-32] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-2.5-scalers0-aggregators0-16-16] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-2.5-scalers0-aggregators0-16-32] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-2.5-scalers0-aggregators0-32-16] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-2.5-scalers0-aggregators0-32-32] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-2.5-scalers0-aggregators1-16-16] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-2.5-scalers0-aggregators1-16-32] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-2.5-scalers0-aggregators1-32-16] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-2.5-scalers0-aggregators1-32-32] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-2.5-scalers0-aggregators2-16-16] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-2.5-scalers0-aggregators2-16-32] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-2.5-scalers0-aggregators2-32-16] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-2.5-scalers0-aggregators2-32-32] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-2.5-scalers1-aggregators0-16-16] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-2.5-scalers1-aggregators0-16-32] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-2.5-scalers1-aggregators0-32-16] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-2.5-scalers1-aggregators0-32-32] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-2.5-scalers1-aggregators1-16-16] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-2.5-scalers1-aggregators1-16-32] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-2.5-scalers1-aggregators1-32-16] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-2.5-scalers1-aggregators1-32-32] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-2.5-scalers1-aggregators2-16-16] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-2.5-scalers1-aggregators2-16-32] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-2.5-scalers1-aggregators2-32-16] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-2.5-scalers1-aggregators2-32-32] PASSED [ 73%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-7.4-scalers0-aggregators0-16-16] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-7.4-scalers0-aggregators0-16-32] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-7.4-scalers0-aggregators0-32-16] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-7.4-scalers0-aggregators0-32-32] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-7.4-scalers0-aggregators1-16-16] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-7.4-scalers0-aggregators1-16-32] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-7.4-scalers0-aggregators1-32-16] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-7.4-scalers0-aggregators1-32-32] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-7.4-scalers0-aggregators2-16-16] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-7.4-scalers0-aggregators2-16-32] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-7.4-scalers0-aggregators2-32-16] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-7.4-scalers0-aggregators2-32-32] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-7.4-scalers1-aggregators0-16-16] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-7.4-scalers1-aggregators0-16-32] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-7.4-scalers1-aggregators0-32-16] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-7.4-scalers1-aggregators0-32-32] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-7.4-scalers1-aggregators1-16-16] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-7.4-scalers1-aggregators1-16-32] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-7.4-scalers1-aggregators1-32-16] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-7.4-scalers1-aggregators1-32-32] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-7.4-scalers1-aggregators2-16-16] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-7.4-scalers1-aggregators2-16-32] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-7.4-scalers1-aggregators2-32-16] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-1-0.1-7.4-scalers1-aggregators2-32-32] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-2.5-scalers0-aggregators0-16-16] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-2.5-scalers0-aggregators0-16-32] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-2.5-scalers0-aggregators0-32-16] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-2.5-scalers0-aggregators0-32-32] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-2.5-scalers0-aggregators1-16-16] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-2.5-scalers0-aggregators1-16-32] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-2.5-scalers0-aggregators1-32-16] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-2.5-scalers0-aggregators1-32-32] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-2.5-scalers0-aggregators2-16-16] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-2.5-scalers0-aggregators2-16-32] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-2.5-scalers0-aggregators2-32-16] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-2.5-scalers0-aggregators2-32-32] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-2.5-scalers1-aggregators0-16-16] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-2.5-scalers1-aggregators0-16-32] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-2.5-scalers1-aggregators0-32-16] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-2.5-scalers1-aggregators0-32-32] PASSED [ 74%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-2.5-scalers1-aggregators1-16-16] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-2.5-scalers1-aggregators1-16-32] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-2.5-scalers1-aggregators1-32-16] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-2.5-scalers1-aggregators1-32-32] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-2.5-scalers1-aggregators2-16-16] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-2.5-scalers1-aggregators2-16-32] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-2.5-scalers1-aggregators2-32-16] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-2.5-scalers1-aggregators2-32-32] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-7.4-scalers0-aggregators0-16-16] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-7.4-scalers0-aggregators0-16-32] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-7.4-scalers0-aggregators0-32-16] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-7.4-scalers0-aggregators0-32-32] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-7.4-scalers0-aggregators1-16-16] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-7.4-scalers0-aggregators1-16-32] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-7.4-scalers0-aggregators1-32-16] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-7.4-scalers0-aggregators1-32-32] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-7.4-scalers0-aggregators2-16-16] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-7.4-scalers0-aggregators2-16-32] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-7.4-scalers0-aggregators2-32-16] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-7.4-scalers0-aggregators2-32-32] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-7.4-scalers1-aggregators0-16-16] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-7.4-scalers1-aggregators0-16-32] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-7.4-scalers1-aggregators0-32-16] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-7.4-scalers1-aggregators0-32-32] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-7.4-scalers1-aggregators1-16-16] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-7.4-scalers1-aggregators1-16-32] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-7.4-scalers1-aggregators1-32-16] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-7.4-scalers1-aggregators1-32-32] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-7.4-scalers1-aggregators2-16-16] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-7.4-scalers1-aggregators2-16-32] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-7.4-scalers1-aggregators2-32-16] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.0-7.4-scalers1-aggregators2-32-32] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-2.5-scalers0-aggregators0-16-16] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-2.5-scalers0-aggregators0-16-32] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-2.5-scalers0-aggregators0-32-16] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-2.5-scalers0-aggregators0-32-32] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-2.5-scalers0-aggregators1-16-16] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-2.5-scalers0-aggregators1-16-32] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-2.5-scalers0-aggregators1-32-16] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-2.5-scalers0-aggregators1-32-32] PASSED [ 75%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-2.5-scalers0-aggregators2-16-16] PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-2.5-scalers0-aggregators2-16-32] PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-2.5-scalers0-aggregators2-32-16] PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-2.5-scalers0-aggregators2-32-32] PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-2.5-scalers1-aggregators0-16-16] PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-2.5-scalers1-aggregators0-16-32] PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-2.5-scalers1-aggregators0-32-16] PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-2.5-scalers1-aggregators0-32-32] PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-2.5-scalers1-aggregators1-16-16] PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-2.5-scalers1-aggregators1-16-32] PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-2.5-scalers1-aggregators1-32-16] PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-2.5-scalers1-aggregators1-32-32] PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-2.5-scalers1-aggregators2-16-16] PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-2.5-scalers1-aggregators2-16-32] PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-2.5-scalers1-aggregators2-32-16] PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-2.5-scalers1-aggregators2-32-32] PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-7.4-scalers0-aggregators0-16-16] PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-7.4-scalers0-aggregators0-16-32] PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-7.4-scalers0-aggregators0-32-16] PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-7.4-scalers0-aggregators0-32-32] PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-7.4-scalers0-aggregators1-16-16] PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-7.4-scalers0-aggregators1-16-32] PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-7.4-scalers0-aggregators1-32-16] PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-7.4-scalers0-aggregators1-32-32] PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-7.4-scalers0-aggregators2-16-16] PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-7.4-scalers0-aggregators2-16-32] PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-7.4-scalers0-aggregators2-32-16] PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-7.4-scalers0-aggregators2-32-32] PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-7.4-scalers1-aggregators0-16-16] PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-7.4-scalers1-aggregators0-16-32] PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-7.4-scalers1-aggregators0-32-16] PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-7.4-scalers1-aggregators0-32-32] PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-7.4-scalers1-aggregators1-16-16] PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-7.4-scalers1-aggregators1-16-32] PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-7.4-scalers1-aggregators1-32-16] PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-7.4-scalers1-aggregators1-32-32] PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-7.4-scalers1-aggregators2-16-16] PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-7.4-scalers1-aggregators2-16-32] PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-7.4-scalers1-aggregators2-32-16] PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-16-4-0.1-7.4-scalers1-aggregators2-32-32] PASSED [ 76%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-2.5-scalers0-aggregators0-16-16] PASSED [ 77%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-2.5-scalers0-aggregators0-16-32] PASSED [ 77%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-2.5-scalers0-aggregators0-32-16] PASSED [ 77%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-2.5-scalers0-aggregators0-32-32] PASSED [ 77%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-2.5-scalers0-aggregators1-16-16] PASSED [ 77%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-2.5-scalers0-aggregators1-16-32] PASSED [ 77%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-2.5-scalers0-aggregators1-32-16] PASSED [ 77%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-2.5-scalers0-aggregators1-32-32] PASSED [ 77%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-2.5-scalers0-aggregators2-16-16] PASSED [ 77%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-2.5-scalers0-aggregators2-16-32] PASSED [ 77%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-2.5-scalers0-aggregators2-32-16] PASSED [ 77%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-2.5-scalers0-aggregators2-32-32] PASSED [ 77%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-2.5-scalers1-aggregators0-16-16] PASSED [ 77%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-2.5-scalers1-aggregators0-16-32] PASSED [ 77%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-2.5-scalers1-aggregators0-32-16] PASSED [ 77%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-2.5-scalers1-aggregators0-32-32] PASSED [ 77%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-2.5-scalers1-aggregators1-16-16] PASSED [ 77%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-2.5-scalers1-aggregators1-16-32] PASSED [ 77%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-2.5-scalers1-aggregators1-32-16] PASSED [ 77%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-2.5-scalers1-aggregators1-32-32] PASSED [ 77%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-2.5-scalers1-aggregators2-16-16] PASSED [ 77%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-2.5-scalers1-aggregators2-16-32] PASSED [ 77%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-2.5-scalers1-aggregators2-32-16] PASSED [ 77%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-2.5-scalers1-aggregators2-32-32] PASSED [ 77%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-7.4-scalers0-aggregators0-16-16] PASSED [ 77%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-7.4-scalers0-aggregators0-16-32] PASSED [ 77%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-7.4-scalers0-aggregators0-32-16] PASSED [ 77%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-7.4-scalers0-aggregators0-32-32] PASSED [ 77%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-7.4-scalers0-aggregators1-16-16] PASSED [ 77%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-7.4-scalers0-aggregators1-16-32] PASSED [ 77%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-7.4-scalers0-aggregators1-32-16] PASSED [ 77%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-7.4-scalers0-aggregators1-32-32] PASSED [ 77%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-7.4-scalers0-aggregators2-16-16] PASSED [ 77%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-7.4-scalers0-aggregators2-16-32] PASSED [ 77%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-7.4-scalers0-aggregators2-32-16] PASSED [ 77%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-7.4-scalers0-aggregators2-32-32] PASSED [ 77%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-7.4-scalers1-aggregators0-16-16] PASSED [ 77%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-7.4-scalers1-aggregators0-16-32] PASSED [ 77%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-7.4-scalers1-aggregators0-32-16] PASSED [ 77%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-7.4-scalers1-aggregators0-32-32] PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-7.4-scalers1-aggregators1-16-16] PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-7.4-scalers1-aggregators1-16-32] PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-7.4-scalers1-aggregators1-32-16] PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-7.4-scalers1-aggregators1-32-32] PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-7.4-scalers1-aggregators2-16-16] PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-7.4-scalers1-aggregators2-16-32] PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-7.4-scalers1-aggregators2-32-16] PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.0-7.4-scalers1-aggregators2-32-32] PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-2.5-scalers0-aggregators0-16-16] PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-2.5-scalers0-aggregators0-16-32] PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-2.5-scalers0-aggregators0-32-16] PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-2.5-scalers0-aggregators0-32-32] PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-2.5-scalers0-aggregators1-16-16] PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-2.5-scalers0-aggregators1-16-32] PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-2.5-scalers0-aggregators1-32-16] PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-2.5-scalers0-aggregators1-32-32] PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-2.5-scalers0-aggregators2-16-16] PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-2.5-scalers0-aggregators2-16-32] PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-2.5-scalers0-aggregators2-32-16] PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-2.5-scalers0-aggregators2-32-32] PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-2.5-scalers1-aggregators0-16-16] PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-2.5-scalers1-aggregators0-16-32] PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-2.5-scalers1-aggregators0-32-16] PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-2.5-scalers1-aggregators0-32-32] PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-2.5-scalers1-aggregators1-16-16] PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-2.5-scalers1-aggregators1-16-32] PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-2.5-scalers1-aggregators1-32-16] PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-2.5-scalers1-aggregators1-32-32] PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-2.5-scalers1-aggregators2-16-16] PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-2.5-scalers1-aggregators2-16-32] PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-2.5-scalers1-aggregators2-32-16] PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-2.5-scalers1-aggregators2-32-32] PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-7.4-scalers0-aggregators0-16-16] PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-7.4-scalers0-aggregators0-16-32] PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-7.4-scalers0-aggregators0-32-16] PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-7.4-scalers0-aggregators0-32-32] PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-7.4-scalers0-aggregators1-16-16] PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-7.4-scalers0-aggregators1-16-32] PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-7.4-scalers0-aggregators1-32-16] PASSED [ 78%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-7.4-scalers0-aggregators1-32-32] PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-7.4-scalers0-aggregators2-16-16] PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-7.4-scalers0-aggregators2-16-32] PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-7.4-scalers0-aggregators2-32-16] PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-7.4-scalers0-aggregators2-32-32] PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-7.4-scalers1-aggregators0-16-16] PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-7.4-scalers1-aggregators0-16-32] PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-7.4-scalers1-aggregators0-32-16] PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-7.4-scalers1-aggregators0-32-32] PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-7.4-scalers1-aggregators1-16-16] PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-7.4-scalers1-aggregators1-16-32] PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-7.4-scalers1-aggregators1-32-16] PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-7.4-scalers1-aggregators1-32-32] PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-7.4-scalers1-aggregators2-16-16] PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-7.4-scalers1-aggregators2-16-32] PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-7.4-scalers1-aggregators2-32-16] PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-1-0.1-7.4-scalers1-aggregators2-32-32] PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-2.5-scalers0-aggregators0-16-16] PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-2.5-scalers0-aggregators0-16-32] PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-2.5-scalers0-aggregators0-32-16] PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-2.5-scalers0-aggregators0-32-32] PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-2.5-scalers0-aggregators1-16-16] PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-2.5-scalers0-aggregators1-16-32] PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-2.5-scalers0-aggregators1-32-16] PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-2.5-scalers0-aggregators1-32-32] PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-2.5-scalers0-aggregators2-16-16] PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-2.5-scalers0-aggregators2-16-32] PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-2.5-scalers0-aggregators2-32-16] PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-2.5-scalers0-aggregators2-32-32] PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-2.5-scalers1-aggregators0-16-16] PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-2.5-scalers1-aggregators0-16-32] PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-2.5-scalers1-aggregators0-32-16] PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-2.5-scalers1-aggregators0-32-32] PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-2.5-scalers1-aggregators1-16-16] PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-2.5-scalers1-aggregators1-16-32] PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-2.5-scalers1-aggregators1-32-16] PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-2.5-scalers1-aggregators1-32-32] PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-2.5-scalers1-aggregators2-16-16] PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-2.5-scalers1-aggregators2-16-32] PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-2.5-scalers1-aggregators2-32-16] PASSED [ 79%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-2.5-scalers1-aggregators2-32-32] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-7.4-scalers0-aggregators0-16-16] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-7.4-scalers0-aggregators0-16-32] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-7.4-scalers0-aggregators0-32-16] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-7.4-scalers0-aggregators0-32-32] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-7.4-scalers0-aggregators1-16-16] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-7.4-scalers0-aggregators1-16-32] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-7.4-scalers0-aggregators1-32-16] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-7.4-scalers0-aggregators1-32-32] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-7.4-scalers0-aggregators2-16-16] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-7.4-scalers0-aggregators2-16-32] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-7.4-scalers0-aggregators2-32-16] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-7.4-scalers0-aggregators2-32-32] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-7.4-scalers1-aggregators0-16-16] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-7.4-scalers1-aggregators0-16-32] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-7.4-scalers1-aggregators0-32-16] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-7.4-scalers1-aggregators0-32-32] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-7.4-scalers1-aggregators1-16-16] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-7.4-scalers1-aggregators1-16-32] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-7.4-scalers1-aggregators1-32-16] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-7.4-scalers1-aggregators1-32-32] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-7.4-scalers1-aggregators2-16-16] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-7.4-scalers1-aggregators2-16-32] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-7.4-scalers1-aggregators2-32-16] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.0-7.4-scalers1-aggregators2-32-32] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-2.5-scalers0-aggregators0-16-16] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-2.5-scalers0-aggregators0-16-32] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-2.5-scalers0-aggregators0-32-16] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-2.5-scalers0-aggregators0-32-32] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-2.5-scalers0-aggregators1-16-16] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-2.5-scalers0-aggregators1-16-32] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-2.5-scalers0-aggregators1-32-16] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-2.5-scalers0-aggregators1-32-32] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-2.5-scalers0-aggregators2-16-16] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-2.5-scalers0-aggregators2-16-32] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-2.5-scalers0-aggregators2-32-16] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-2.5-scalers0-aggregators2-32-32] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-2.5-scalers1-aggregators0-16-16] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-2.5-scalers1-aggregators0-16-32] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-2.5-scalers1-aggregators0-32-16] PASSED [ 80%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-2.5-scalers1-aggregators0-32-32] PASSED [ 81%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-2.5-scalers1-aggregators1-16-16] PASSED [ 81%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-2.5-scalers1-aggregators1-16-32] PASSED [ 81%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-2.5-scalers1-aggregators1-32-16] PASSED [ 81%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-2.5-scalers1-aggregators1-32-32] PASSED [ 81%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-2.5-scalers1-aggregators2-16-16] PASSED [ 81%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-2.5-scalers1-aggregators2-16-32] PASSED [ 81%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-2.5-scalers1-aggregators2-32-16] PASSED [ 81%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-2.5-scalers1-aggregators2-32-32] PASSED [ 81%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-7.4-scalers0-aggregators0-16-16] PASSED [ 81%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-7.4-scalers0-aggregators0-16-32] PASSED [ 81%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-7.4-scalers0-aggregators0-32-16] PASSED [ 81%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-7.4-scalers0-aggregators0-32-32] PASSED [ 81%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-7.4-scalers0-aggregators1-16-16] PASSED [ 81%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-7.4-scalers0-aggregators1-16-32] PASSED [ 81%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-7.4-scalers0-aggregators1-32-16] PASSED [ 81%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-7.4-scalers0-aggregators1-32-32] PASSED [ 81%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-7.4-scalers0-aggregators2-16-16] PASSED [ 81%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-7.4-scalers0-aggregators2-16-32] PASSED [ 81%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-7.4-scalers0-aggregators2-32-16] PASSED [ 81%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-7.4-scalers0-aggregators2-32-32] PASSED [ 81%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-7.4-scalers1-aggregators0-16-16] PASSED [ 81%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-7.4-scalers1-aggregators0-16-32] PASSED [ 81%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-7.4-scalers1-aggregators0-32-16] PASSED [ 81%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-7.4-scalers1-aggregators0-32-32] PASSED [ 81%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-7.4-scalers1-aggregators1-16-16] PASSED [ 81%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-7.4-scalers1-aggregators1-16-32] PASSED [ 81%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-7.4-scalers1-aggregators1-32-16] PASSED [ 81%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-7.4-scalers1-aggregators1-32-32] PASSED [ 81%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-7.4-scalers1-aggregators2-16-16] PASSED [ 81%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-7.4-scalers1-aggregators2-16-32] PASSED [ 81%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-7.4-scalers1-aggregators2-32-16] PASSED [ 81%]
tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-7.4-scalers1-aggregators2-32-32] PASSED [ 81%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-2.5-scalers0-aggregators0-16-16] PASSED [ 81%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-2.5-scalers0-aggregators0-16-32] PASSED [ 81%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-2.5-scalers0-aggregators0-32-16] PASSED [ 81%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-2.5-scalers0-aggregators0-32-32] PASSED [ 81%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-2.5-scalers0-aggregators1-16-16] PASSED [ 81%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-2.5-scalers0-aggregators1-16-32] PASSED [ 81%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-2.5-scalers0-aggregators1-32-16] PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-2.5-scalers0-aggregators1-32-32] PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-2.5-scalers0-aggregators2-16-16] PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-2.5-scalers0-aggregators2-16-32] PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-2.5-scalers0-aggregators2-32-16] PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-2.5-scalers0-aggregators2-32-32] PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-2.5-scalers1-aggregators0-16-16] PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-2.5-scalers1-aggregators0-16-32] PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-2.5-scalers1-aggregators0-32-16] PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-2.5-scalers1-aggregators0-32-32] PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-2.5-scalers1-aggregators1-16-16] PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-2.5-scalers1-aggregators1-16-32] PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-2.5-scalers1-aggregators1-32-16] PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-2.5-scalers1-aggregators1-32-32] PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-2.5-scalers1-aggregators2-16-16] PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-2.5-scalers1-aggregators2-16-32] PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-2.5-scalers1-aggregators2-32-16] PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-2.5-scalers1-aggregators2-32-32] PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-7.4-scalers0-aggregators0-16-16] PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-7.4-scalers0-aggregators0-16-32] PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-7.4-scalers0-aggregators0-32-16] PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-7.4-scalers0-aggregators0-32-32] PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-7.4-scalers0-aggregators1-16-16] PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-7.4-scalers0-aggregators1-16-32] PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-7.4-scalers0-aggregators1-32-16] PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-7.4-scalers0-aggregators1-32-32] PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-7.4-scalers0-aggregators2-16-16] PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-7.4-scalers0-aggregators2-16-32] PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-7.4-scalers0-aggregators2-32-16] PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-7.4-scalers0-aggregators2-32-32] PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-7.4-scalers1-aggregators0-16-16] PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-7.4-scalers1-aggregators0-16-32] PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-7.4-scalers1-aggregators0-32-16] PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-7.4-scalers1-aggregators0-32-32] PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-7.4-scalers1-aggregators1-16-16] PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-7.4-scalers1-aggregators1-16-32] PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-7.4-scalers1-aggregators1-32-16] PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-7.4-scalers1-aggregators1-32-32] PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-7.4-scalers1-aggregators2-16-16] PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-7.4-scalers1-aggregators2-16-32] PASSED [ 82%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-7.4-scalers1-aggregators2-32-16] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.0-7.4-scalers1-aggregators2-32-32] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-2.5-scalers0-aggregators0-16-16] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-2.5-scalers0-aggregators0-16-32] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-2.5-scalers0-aggregators0-32-16] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-2.5-scalers0-aggregators0-32-32] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-2.5-scalers0-aggregators1-16-16] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-2.5-scalers0-aggregators1-16-32] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-2.5-scalers0-aggregators1-32-16] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-2.5-scalers0-aggregators1-32-32] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-2.5-scalers0-aggregators2-16-16] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-2.5-scalers0-aggregators2-16-32] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-2.5-scalers0-aggregators2-32-16] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-2.5-scalers0-aggregators2-32-32] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-2.5-scalers1-aggregators0-16-16] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-2.5-scalers1-aggregators0-16-32] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-2.5-scalers1-aggregators0-32-16] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-2.5-scalers1-aggregators0-32-32] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-2.5-scalers1-aggregators1-16-16] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-2.5-scalers1-aggregators1-16-32] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-2.5-scalers1-aggregators1-32-16] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-2.5-scalers1-aggregators1-32-32] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-2.5-scalers1-aggregators2-16-16] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-2.5-scalers1-aggregators2-16-32] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-2.5-scalers1-aggregators2-32-16] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-2.5-scalers1-aggregators2-32-32] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-7.4-scalers0-aggregators0-16-16] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-7.4-scalers0-aggregators0-16-32] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-7.4-scalers0-aggregators0-32-16] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-7.4-scalers0-aggregators0-32-32] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-7.4-scalers0-aggregators1-16-16] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-7.4-scalers0-aggregators1-16-32] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-7.4-scalers0-aggregators1-32-16] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-7.4-scalers0-aggregators1-32-32] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-7.4-scalers0-aggregators2-16-16] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-7.4-scalers0-aggregators2-16-32] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-7.4-scalers0-aggregators2-32-16] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-7.4-scalers0-aggregators2-32-32] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-7.4-scalers1-aggregators0-16-16] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-7.4-scalers1-aggregators0-16-32] PASSED [ 83%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-7.4-scalers1-aggregators0-32-16] PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-7.4-scalers1-aggregators0-32-32] PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-7.4-scalers1-aggregators1-16-16] PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-7.4-scalers1-aggregators1-16-32] PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-7.4-scalers1-aggregators1-32-16] PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-7.4-scalers1-aggregators1-32-32] PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-7.4-scalers1-aggregators2-16-16] PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-7.4-scalers1-aggregators2-16-32] PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-7.4-scalers1-aggregators2-32-16] PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-1-0.1-7.4-scalers1-aggregators2-32-32] PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-2.5-scalers0-aggregators0-16-16] PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-2.5-scalers0-aggregators0-16-32] PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-2.5-scalers0-aggregators0-32-16] PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-2.5-scalers0-aggregators0-32-32] PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-2.5-scalers0-aggregators1-16-16] PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-2.5-scalers0-aggregators1-16-32] PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-2.5-scalers0-aggregators1-32-16] PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-2.5-scalers0-aggregators1-32-32] PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-2.5-scalers0-aggregators2-16-16] PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-2.5-scalers0-aggregators2-16-32] PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-2.5-scalers0-aggregators2-32-16] PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-2.5-scalers0-aggregators2-32-32] PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-2.5-scalers1-aggregators0-16-16] PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-2.5-scalers1-aggregators0-16-32] PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-2.5-scalers1-aggregators0-32-16] PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-2.5-scalers1-aggregators0-32-32] PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-2.5-scalers1-aggregators1-16-16] PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-2.5-scalers1-aggregators1-16-32] PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-2.5-scalers1-aggregators1-32-16] PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-2.5-scalers1-aggregators1-32-32] PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-2.5-scalers1-aggregators2-16-16] PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-2.5-scalers1-aggregators2-16-32] PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-2.5-scalers1-aggregators2-32-16] PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-2.5-scalers1-aggregators2-32-32] PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-7.4-scalers0-aggregators0-16-16] PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-7.4-scalers0-aggregators0-16-32] PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-7.4-scalers0-aggregators0-32-16] PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-7.4-scalers0-aggregators0-32-32] PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-7.4-scalers0-aggregators1-16-16] PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-7.4-scalers0-aggregators1-16-32] PASSED [ 84%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-7.4-scalers0-aggregators1-32-16] PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-7.4-scalers0-aggregators1-32-32] PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-7.4-scalers0-aggregators2-16-16] PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-7.4-scalers0-aggregators2-16-32] PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-7.4-scalers0-aggregators2-32-16] PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-7.4-scalers0-aggregators2-32-32] PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-7.4-scalers1-aggregators0-16-16] PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-7.4-scalers1-aggregators0-16-32] PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-7.4-scalers1-aggregators0-32-16] Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/nanoflann'...
PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-7.4-scalers1-aggregators0-32-32] PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-7.4-scalers1-aggregators1-16-16] PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-7.4-scalers1-aggregators1-16-32] PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-7.4-scalers1-aggregators1-32-16] PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-7.4-scalers1-aggregators1-32-32] PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-7.4-scalers1-aggregators2-16-16] Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/nccl'...
PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-7.4-scalers1-aggregators2-16-32] PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-7.4-scalers1-aggregators2-32-16] PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.0-7.4-scalers1-aggregators2-32-32] PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-2.5-scalers0-aggregators0-16-16] PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-2.5-scalers0-aggregators0-16-32] PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-2.5-scalers0-aggregators0-32-16] PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-2.5-scalers0-aggregators0-32-32] PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-2.5-scalers0-aggregators1-16-16] PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-2.5-scalers0-aggregators1-16-32] PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-2.5-scalers0-aggregators1-32-16] PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-2.5-scalers0-aggregators1-32-32] PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-2.5-scalers0-aggregators2-16-16] Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/phmap'...
PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-2.5-scalers0-aggregators2-16-32] PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-2.5-scalers0-aggregators2-32-16] PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-2.5-scalers0-aggregators2-32-32] PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-2.5-scalers1-aggregators0-16-16] PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-2.5-scalers1-aggregators0-16-32] PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-2.5-scalers1-aggregators0-32-16] PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-2.5-scalers1-aggregators0-32-32] PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-2.5-scalers1-aggregators1-16-16] PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-2.5-scalers1-aggregators1-16-32] PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-2.5-scalers1-aggregators1-32-16] PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-2.5-scalers1-aggregators1-32-32] PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-2.5-scalers1-aggregators2-16-16] PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-2.5-scalers1-aggregators2-16-32] PASSED [ 85%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-2.5-scalers1-aggregators2-32-16] PASSED [ 86%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-2.5-scalers1-aggregators2-32-32] PASSED [ 86%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-7.4-scalers0-aggregators0-16-16] PASSED [ 86%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-7.4-scalers0-aggregators0-16-32] PASSED [ 86%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-7.4-scalers0-aggregators0-32-16] PASSED [ 86%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-7.4-scalers0-aggregators0-32-32] PASSED [ 86%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-7.4-scalers0-aggregators1-16-16] PASSED [ 86%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-7.4-scalers0-aggregators1-16-32] PASSED [ 86%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-7.4-scalers0-aggregators1-32-16] PASSED [ 86%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-7.4-scalers0-aggregators1-32-32] PASSED [ 86%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-7.4-scalers0-aggregators2-16-16] PASSED [ 86%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-7.4-scalers0-aggregators2-16-32] Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tensorpipe'...
PASSED [ 86%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-7.4-scalers0-aggregators2-32-16] PASSED [ 86%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-7.4-scalers0-aggregators2-32-32] PASSED [ 86%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-7.4-scalers1-aggregators0-16-16] PASSED [ 86%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-7.4-scalers1-aggregators0-16-32] PASSED [ 86%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-7.4-scalers1-aggregators0-32-16] PASSED [ 86%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-7.4-scalers1-aggregators0-32-32] PASSED [ 86%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-7.4-scalers1-aggregators1-16-16] PASSED [ 86%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-7.4-scalers1-aggregators1-16-32] PASSED [ 86%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-7.4-scalers1-aggregators1-32-16] PASSED [ 86%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-7.4-scalers1-aggregators1-32-32] PASSED [ 86%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-7.4-scalers1-aggregators2-16-16] PASSED [ 86%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-7.4-scalers1-aggregators2-16-32] Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/thrust'...
PASSED [ 86%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-7.4-scalers1-aggregators2-32-16] PASSED [ 86%]
tests/pytorch/test_nn.py::test_dgn_conv[False-16-4-0.1-7.4-scalers1-aggregators2-32-32] PASSED [ 86%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-2.5-scalers0-aggregators0-16-16] PASSED [ 86%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-2.5-scalers0-aggregators0-16-32] PASSED [ 86%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-2.5-scalers0-aggregators0-32-16] PASSED [ 86%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-2.5-scalers0-aggregators0-32-32] PASSED [ 86%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-2.5-scalers0-aggregators1-16-16] PASSED [ 86%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-2.5-scalers0-aggregators1-16-32] PASSED [ 86%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-2.5-scalers0-aggregators1-32-16] PASSED [ 86%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-2.5-scalers0-aggregators1-32-32] PASSED [ 86%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-2.5-scalers0-aggregators2-16-16] PASSED [ 86%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-2.5-scalers0-aggregators2-16-32] PASSED [ 86%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-2.5-scalers0-aggregators2-32-16] PASSED [ 86%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-2.5-scalers0-aggregators2-32-32] PASSED [ 86%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-2.5-scalers1-aggregators0-16-16] PASSED [ 86%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-2.5-scalers1-aggregators0-16-32] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-2.5-scalers1-aggregators0-32-16] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-2.5-scalers1-aggregators0-32-32] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-2.5-scalers1-aggregators1-16-16] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-2.5-scalers1-aggregators1-16-32] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-2.5-scalers1-aggregators1-32-16] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-2.5-scalers1-aggregators1-32-32] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-2.5-scalers1-aggregators2-16-16] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-2.5-scalers1-aggregators2-16-32] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-2.5-scalers1-aggregators2-32-16] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-2.5-scalers1-aggregators2-32-32] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-7.4-scalers0-aggregators0-16-16] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-7.4-scalers0-aggregators0-16-32] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-7.4-scalers0-aggregators0-32-16] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-7.4-scalers0-aggregators0-32-32] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-7.4-scalers0-aggregators1-16-16] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-7.4-scalers0-aggregators1-16-32] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-7.4-scalers0-aggregators1-32-16] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-7.4-scalers0-aggregators1-32-32] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-7.4-scalers0-aggregators2-16-16] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-7.4-scalers0-aggregators2-16-32] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-7.4-scalers0-aggregators2-32-16] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-7.4-scalers0-aggregators2-32-32] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-7.4-scalers1-aggregators0-16-16] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-7.4-scalers1-aggregators0-16-32] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-7.4-scalers1-aggregators0-32-16] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-7.4-scalers1-aggregators0-32-32] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-7.4-scalers1-aggregators1-16-16] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-7.4-scalers1-aggregators1-16-32] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-7.4-scalers1-aggregators1-32-16] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-7.4-scalers1-aggregators1-32-32] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-7.4-scalers1-aggregators2-16-16] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-7.4-scalers1-aggregators2-16-32] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-7.4-scalers1-aggregators2-32-16] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.0-7.4-scalers1-aggregators2-32-32] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-2.5-scalers0-aggregators0-16-16] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-2.5-scalers0-aggregators0-16-32] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-2.5-scalers0-aggregators0-32-16] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-2.5-scalers0-aggregators0-32-32] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-2.5-scalers0-aggregators1-16-16] PASSED [ 87%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-2.5-scalers0-aggregators1-16-32] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-2.5-scalers0-aggregators1-32-16] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-2.5-scalers0-aggregators1-32-32] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-2.5-scalers0-aggregators2-16-16] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-2.5-scalers0-aggregators2-16-32] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-2.5-scalers0-aggregators2-32-16] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-2.5-scalers0-aggregators2-32-32] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-2.5-scalers1-aggregators0-16-16] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-2.5-scalers1-aggregators0-16-32] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-2.5-scalers1-aggregators0-32-16] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-2.5-scalers1-aggregators0-32-32] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-2.5-scalers1-aggregators1-16-16] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-2.5-scalers1-aggregators1-16-32] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-2.5-scalers1-aggregators1-32-16] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-2.5-scalers1-aggregators1-32-32] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-2.5-scalers1-aggregators2-16-16] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-2.5-scalers1-aggregators2-16-32] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-2.5-scalers1-aggregators2-32-16] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-2.5-scalers1-aggregators2-32-32] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-7.4-scalers0-aggregators0-16-16] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-7.4-scalers0-aggregators0-16-32] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-7.4-scalers0-aggregators0-32-16] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-7.4-scalers0-aggregators0-32-32] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-7.4-scalers0-aggregators1-16-16] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-7.4-scalers0-aggregators1-16-32] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-7.4-scalers0-aggregators1-32-16] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-7.4-scalers0-aggregators1-32-32] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-7.4-scalers0-aggregators2-16-16] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-7.4-scalers0-aggregators2-16-32] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-7.4-scalers0-aggregators2-32-16] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-7.4-scalers0-aggregators2-32-32] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-7.4-scalers1-aggregators0-16-16] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-7.4-scalers1-aggregators0-16-32] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-7.4-scalers1-aggregators0-32-16] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-7.4-scalers1-aggregators0-32-32] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-7.4-scalers1-aggregators1-16-16] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-7.4-scalers1-aggregators1-16-32] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-7.4-scalers1-aggregators1-32-16] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-7.4-scalers1-aggregators1-32-32] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-7.4-scalers1-aggregators2-16-16] PASSED [ 88%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-7.4-scalers1-aggregators2-16-32] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-7.4-scalers1-aggregators2-32-16] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-1-0.1-7.4-scalers1-aggregators2-32-32] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-2.5-scalers0-aggregators0-16-16] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-2.5-scalers0-aggregators0-16-32] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-2.5-scalers0-aggregators0-32-16] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-2.5-scalers0-aggregators0-32-32] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-2.5-scalers0-aggregators1-16-16] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-2.5-scalers0-aggregators1-16-32] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-2.5-scalers0-aggregators1-32-16] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-2.5-scalers0-aggregators1-32-32] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-2.5-scalers0-aggregators2-16-16] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-2.5-scalers0-aggregators2-16-32] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-2.5-scalers0-aggregators2-32-16] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-2.5-scalers0-aggregators2-32-32] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-2.5-scalers1-aggregators0-16-16] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-2.5-scalers1-aggregators0-16-32] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-2.5-scalers1-aggregators0-32-16] Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tvm'...
PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-2.5-scalers1-aggregators0-32-32] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-2.5-scalers1-aggregators1-16-16] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-2.5-scalers1-aggregators1-16-32] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-2.5-scalers1-aggregators1-32-16] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-2.5-scalers1-aggregators1-32-32] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-2.5-scalers1-aggregators2-16-16] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-2.5-scalers1-aggregators2-16-32] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-2.5-scalers1-aggregators2-32-16] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-2.5-scalers1-aggregators2-32-32] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-7.4-scalers0-aggregators0-16-16] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-7.4-scalers0-aggregators0-16-32] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-7.4-scalers0-aggregators0-32-16] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-7.4-scalers0-aggregators0-32-32] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-7.4-scalers0-aggregators1-16-16] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-7.4-scalers0-aggregators1-16-32] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-7.4-scalers0-aggregators1-32-16] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-7.4-scalers0-aggregators1-32-32] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-7.4-scalers0-aggregators2-16-16] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-7.4-scalers0-aggregators2-16-32] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-7.4-scalers0-aggregators2-32-16] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-7.4-scalers0-aggregators2-32-32] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-7.4-scalers1-aggregators0-16-16] PASSED [ 89%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-7.4-scalers1-aggregators0-16-32] PASSED [ 90%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-7.4-scalers1-aggregators0-32-16] PASSED [ 90%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-7.4-scalers1-aggregators0-32-32] PASSED [ 90%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-7.4-scalers1-aggregators1-16-16] PASSED [ 90%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-7.4-scalers1-aggregators1-16-32] PASSED [ 90%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-7.4-scalers1-aggregators1-32-16] PASSED [ 90%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-7.4-scalers1-aggregators1-32-32] PASSED [ 90%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-7.4-scalers1-aggregators2-16-16] PASSED [ 90%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-7.4-scalers1-aggregators2-16-32] PASSED [ 90%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-7.4-scalers1-aggregators2-32-16] PASSED [ 90%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.0-7.4-scalers1-aggregators2-32-32] PASSED [ 90%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-2.5-scalers0-aggregators0-16-16] PASSED [ 90%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-2.5-scalers0-aggregators0-16-32] PASSED [ 90%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-2.5-scalers0-aggregators0-32-16] PASSED [ 90%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-2.5-scalers0-aggregators0-32-32] PASSED [ 90%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-2.5-scalers0-aggregators1-16-16] PASSED [ 90%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-2.5-scalers0-aggregators1-16-32] PASSED [ 90%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-2.5-scalers0-aggregators1-32-16] PASSED [ 90%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-2.5-scalers0-aggregators1-32-32] PASSED [ 90%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-2.5-scalers0-aggregators2-16-16] PASSED [ 90%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-2.5-scalers0-aggregators2-16-32] PASSED [ 90%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-2.5-scalers0-aggregators2-32-16] PASSED [ 90%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-2.5-scalers0-aggregators2-32-32] PASSED [ 90%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-2.5-scalers1-aggregators0-16-16] PASSED [ 90%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-2.5-scalers1-aggregators0-16-32] PASSED [ 90%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-2.5-scalers1-aggregators0-32-16] PASSED [ 90%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-2.5-scalers1-aggregators0-32-32] PASSED [ 90%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-2.5-scalers1-aggregators1-16-16] PASSED [ 90%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-2.5-scalers1-aggregators1-16-32] PASSED [ 90%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-2.5-scalers1-aggregators1-32-16] PASSED [ 90%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-2.5-scalers1-aggregators1-32-32] PASSED [ 90%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-2.5-scalers1-aggregators2-16-16] PASSED [ 90%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-2.5-scalers1-aggregators2-16-32] PASSED [ 90%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-2.5-scalers1-aggregators2-32-16] PASSED [ 90%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-2.5-scalers1-aggregators2-32-32] PASSED [ 90%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-7.4-scalers0-aggregators0-16-16] PASSED [ 90%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-7.4-scalers0-aggregators0-16-32] PASSED [ 90%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-7.4-scalers0-aggregators0-32-16] PASSED [ 90%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-7.4-scalers0-aggregators0-32-32] PASSED [ 90%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-7.4-scalers0-aggregators1-16-16] PASSED [ 91%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-7.4-scalers0-aggregators1-16-32] PASSED [ 91%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-7.4-scalers0-aggregators1-32-16] PASSED [ 91%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-7.4-scalers0-aggregators1-32-32] PASSED [ 91%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-7.4-scalers0-aggregators2-16-16] PASSED [ 91%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-7.4-scalers0-aggregators2-16-32] PASSED [ 91%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-7.4-scalers0-aggregators2-32-16] PASSED [ 91%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-7.4-scalers0-aggregators2-32-32] PASSED [ 91%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-7.4-scalers1-aggregators0-16-16] PASSED [ 91%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-7.4-scalers1-aggregators0-16-32] PASSED [ 91%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-7.4-scalers1-aggregators0-32-16] PASSED [ 91%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-7.4-scalers1-aggregators0-32-32] PASSED [ 91%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-7.4-scalers1-aggregators1-16-16] PASSED [ 91%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-7.4-scalers1-aggregators1-16-32] PASSED [ 91%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-7.4-scalers1-aggregators1-32-16] PASSED [ 91%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-7.4-scalers1-aggregators1-32-32] PASSED [ 91%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-7.4-scalers1-aggregators2-16-16] PASSED [ 91%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-7.4-scalers1-aggregators2-16-32] PASSED [ 91%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-7.4-scalers1-aggregators2-32-16] PASSED [ 91%]
tests/pytorch/test_nn.py::test_dgn_conv[False-0-4-0.1-7.4-scalers1-aggregators2-32-32] PASSED [ 91%]
tests/pytorch/test_optim.py::test_sparse_adam[1] PASSED                  [ 91%]
tests/pytorch/test_optim.py::test_sparse_adam[4] PASSED                  [ 91%]
tests/pytorch/test_optim.py::test_sparse_adam[101] PASSED                [ 91%]
tests/pytorch/test_optim.py::test_sparse_adam[1024] PASSED               [ 91%]
tests/pytorch/test_optim.py::test_sparse_adam_uva[1-False] PASSED        [ 91%]
tests/pytorch/test_optim.py::test_sparse_adam_uva[1-True] PASSED         [ 91%]
tests/pytorch/test_optim.py::test_sparse_adam_uva[1-None] PASSED         [ 91%]
tests/pytorch/test_optim.py::test_sparse_adam_uva[4-False] PASSED        [ 91%]
tests/pytorch/test_optim.py::test_sparse_adam_uva[4-True] PASSED         [ 91%]
tests/pytorch/test_optim.py::test_sparse_adam_uva[4-None] PASSED         [ 91%]
tests/pytorch/test_optim.py::test_sparse_adam_uva[101-False] PASSED      [ 91%]
tests/pytorch/test_optim.py::test_sparse_adam_uva[101-True] PASSED       [ 91%]
tests/pytorch/test_optim.py::test_sparse_adam_uva[101-None] PASSED       [ 91%]
tests/pytorch/test_optim.py::test_sparse_adam_uva[1024-False] PASSED     [ 91%]
tests/pytorch/test_optim.py::test_sparse_adam_uva[1024-True] PASSED      [ 91%]
tests/pytorch/test_optim.py::test_sparse_adam_uva[1024-None] PASSED      [ 91%]
tests/pytorch/test_optim.py::test_sparse_adam_dtype[1-dtype0] PASSED     [ 91%]
tests/pytorch/test_optim.py::test_sparse_adam_dtype[1-dtype1] PASSED     [ 91%]
tests/pytorch/test_optim.py::test_sparse_adam_dtype[4-dtype0] PASSED     [ 91%]
tests/pytorch/test_optim.py::test_sparse_adam_dtype[4-dtype1] PASSED     [ 91%]
tests/pytorch/test_optim.py::test_sparse_adam_dtype[101-dtype0] PASSED   [ 92%]
tests/pytorch/test_optim.py::test_sparse_adam_dtype[101-dtype1] PASSED   [ 92%]
tests/pytorch/test_optim.py::test_sparse_adam_dtype[1024-dtype0] PASSED  [ 92%]
tests/pytorch/test_optim.py::test_sparse_adam_dtype[1024-dtype1] PASSED  [ 92%]
tests/pytorch/test_optim.py::test_sparse_adam_zero_step PASSED           [ 92%]
tests/pytorch/test_optim.py::test_multiprocess_cpu_sparse_adam[2] SKIPPED [ 92%]
tests/pytorch/test_optim.py::test_multiprocess_cpu_sparse_adam[4] SKIPPED [ 92%]
tests/pytorch/test_optim.py::test_multiprocess_sparse_adam[nccl-2] SKIPPED [ 92%]
tests/pytorch/test_optim.py::test_multiprocess_sparse_adam[nccl-4] SKIPPED [ 92%]
tests/pytorch/test_optim.py::test_multiprocess_sparse_adam[nccl-8] SKIPPED [ 92%]
tests/pytorch/test_optim.py::test_multiprocess_sparse_adam[gloo-2] SKIPPED [ 92%]
tests/pytorch/test_optim.py::test_multiprocess_sparse_adam[gloo-4] SKIPPED [ 92%]
tests/pytorch/test_optim.py::test_multiprocess_sparse_adam[gloo-8] SKIPPED [ 92%]
tests/pytorch/test_optim.py::test_multiprocess_sparse_adam_cuda_tensor[2] SKIPPED [ 92%]
tests/pytorch/test_optim.py::test_multiprocess_sparse_adam_cuda_tensor[4] SKIPPED [ 92%]
tests/pytorch/test_optim.py::test_multiprocess_sparse_adam_cuda_tensor[8] SKIPPED [ 92%]
tests/pytorch/test_optim.py::test_multiprocess_sparse_adam_cpu_zero_step[2] SKIPPED [ 92%]
tests/pytorch/test_optim.py::test_multiprocess_sparse_adam_cpu_zero_step[4] SKIPPED [ 92%]
tests/pytorch/test_optim.py::test_multiprocess_sparse_adam_zero_step[nccl-2] SKIPPED [ 92%]
tests/pytorch/test_optim.py::test_multiprocess_sparse_adam_zero_step[nccl-4] SKIPPED [ 92%]
tests/pytorch/test_optim.py::test_multiprocess_sparse_adam_zero_step[nccl-8] SKIPPED [ 92%]
tests/pytorch/test_optim.py::test_multiprocess_sparse_adam_zero_step[gloo-2] SKIPPED [ 92%]
tests/pytorch/test_optim.py::test_multiprocess_sparse_adam_zero_step[gloo-4] SKIPPED [ 92%]
tests/pytorch/test_optim.py::test_multiprocess_sparse_adam_zero_step[gloo-8] SKIPPED [ 92%]
tests/pytorch/test_optim.py::test_multiprocess_sparse_adam_zero_step_cuda_tensor[2] SKIPPED [ 92%]
tests/pytorch/test_optim.py::test_multiprocess_sparse_adam_zero_step_cuda_tensor[4] SKIPPED [ 92%]
tests/pytorch/test_optim.py::test_multiprocess_sparse_adam_zero_step_cuda_tensor[8] SKIPPED [ 92%]
tests/pytorch/test_pickle.py::test_pickling_batched_graph PASSED         [ 92%]
tests/pytorch/test_pin_memory.py::test_pin_noncontiguous PASSED          [ 92%]
tests/pytorch/test_pin_memory.py::test_pin_view PASSED                   [ 92%]
tests/pytorch/test_pin_memory.py::test_unpin_automatically PASSED        [ 92%]
tests/pytorch/test_pin_memory.py::test_pin_unpin_column PASSED           [ 92%]
tests/pytorch/test_pin_memory.py::test_pin_empty PASSED                  [ 92%]
tests/pytorch/test_sparse_emb.py::test_multiprocess_sparse_emb_get_set[1] Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/xbyak'...
Submodule path 'third_party/METIS': checked out '10603482cc055626ff99d2ac3ab7e73d1119439d'
Submodule 'GKlib' (https://github.com/KarypisLab/GKlib.git) registered for path 'third_party/METIS/GKlib'
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/METIS/GKlib'...
Submodule path 'third_party/METIS/GKlib': checked out '62de20c7f97c719abdc96d7c3e266a1f7cb52bc5'
Submodule path 'third_party/dlpack': checked out 'e2bdd3bee8cb6501558042633fa59144cc8b7f5f'
Submodule path 'third_party/dmlc-core': checked out 'bfad207b448480783a1f428ae3d93d87032d8349'
Submodule path 'third_party/googletest': checked out 'f71fb4f9a912ec945401cc49a287a759b6131026'
Submodule path 'third_party/libxsmm': checked out 'fa687556130b6298430f1c0555a14cf79ab6101c'
Submodule path 'third_party/nanoflann': checked out '4c47ca200209550c5628c89803591f8a753c8181'
Submodule path 'third_party/nccl': checked out 'e11238b3029795d33f958b5868d47c90c4f22628'
Submodule path 'third_party/phmap': checked out '25293cefd8b85491b45600c03fe8edf07647553f'
Submodule path 'third_party/tensorpipe': checked out '6042f1a4cbce8eef997f11ed0012de137b317361'
Submodule 'third_party/googletest' (https://github.com/google/googletest.git) registered for path 'third_party/tensorpipe/third_party/googletest'
Submodule 'third_party/libnop' (https://github.com/google/libnop.git) registered for path 'third_party/tensorpipe/third_party/libnop'
Submodule 'third_party/libuv' (https://github.com/libuv/libuv.git) registered for path 'third_party/tensorpipe/third_party/libuv'
Submodule 'third_party/pybind11' (https://github.com/pybind/pybind11.git) registered for path 'third_party/tensorpipe/third_party/pybind11'
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tensorpipe/third_party/googletest'...
PASSED [ 92%]
tests/pytorch/test_sparse_emb.py::test_multiprocess_sparse_emb_get_set[2] SKIPPED [ 92%]
tests/pytorch/test_sparse_emb.py::test_multiprocess_sparse_emb_get_set[3] SKIPPED [ 92%]
tests/pytorch/test_stream.py::test_basics PASSED                         [ 92%]
tests/pytorch/test_stream.py::test_set_get_stream PASSED                 [ 92%]
tests/pytorch/test_stream.py::test_record_stream_ndarray PASSED          [ 92%]
tests/pytorch/test_stream.py::test_record_stream_graph_positive Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tensorpipe/third_party/libnop'...
PASSED   [ 92%]
tests/pytorch/test_stream.py::test_record_stream_graph_negative PASSED   [ 93%]
tests/pytorch/test_unified_tensor.py::test_unified_tensor PASSED         [ 93%]
tests/pytorch/test_unified_tensor.py::test_multi_gpu_unified_tensor[1] Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tensorpipe/third_party/libuv'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tensorpipe/third_party/pybind11'...
Submodule path 'third_party/tensorpipe/third_party/googletest': checked out 'aee0f9d9b5b87796ee8a0ab26b7587ec30e8858e'
Submodule path 'third_party/tensorpipe/third_party/libnop': checked out 'aa95422ea8c409e3f078d2ee7708a5f59a8b9fa2'
Submodule path 'third_party/tensorpipe/third_party/libuv': checked out '1dff88e5161cba5c59276d2070d2e304e4dcb242'
Submodule path 'third_party/tensorpipe/third_party/pybind11': checked out 'a23996fce38ff6ccfbcdc09f1e63f2c4be5ea2ef'
Submodule 'tools/clang' (https://github.com/wjakob/clang-cindex-python3) registered for path 'third_party/tensorpipe/third_party/pybind11/tools/clang'
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tensorpipe/third_party/pybind11/tools/clang'...
Submodule path 'third_party/tensorpipe/third_party/pybind11/tools/clang': checked out '6a00cbc4a9b8e68b71caf7f774b3f9c753ae84d5'
Submodule path 'third_party/thrust': checked out '6a3078c64cab0e2f276340fa5dcafa0d758ed890'
Submodule 'cub' (https://github.com/NVIDIA/cub.git) registered for path 'third_party/thrust/dependencies/cub'
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/thrust/dependencies/cub'...
PASSED [ 93%]
tests/pytorch/test_unified_tensor.py::test_multi_gpu_unified_tensor[2] SKIPPED [ 93%]
tests/pytorch/mock_sparse/test_diag.py::test_diag[None-val_shape0] PASSED [ 93%]
tests/pytorch/mock_sparse/test_diag.py::test_diag[None-val_shape1] PASSED [ 93%]
tests/pytorch/mock_sparse/test_diag.py::test_diag[mat_shape1-val_shape0] PASSED [ 93%]
tests/pytorch/mock_sparse/test_diag.py::test_diag[mat_shape1-val_shape1] PASSED [ 93%]
tests/pytorch/mock_sparse/test_diag.py::test_diag[mat_shape2-val_shape0] PASSED [ 93%]
tests/pytorch/mock_sparse/test_diag.py::test_diag[mat_shape2-val_shape1] PASSED [ 93%]
tests/pytorch/mock_sparse/test_diag.py::test_identity[None-shape0] PASSED [ 93%]
tests/pytorch/mock_sparse/test_diag.py::test_identity[None-shape1] PASSED [ 93%]
tests/pytorch/mock_sparse/test_diag.py::test_identity[None-shape2] PASSED [ 93%]
tests/pytorch/mock_sparse/test_diag.py::test_identity[2-shape0] PASSED   [ 93%]
tests/pytorch/mock_sparse/test_diag.py::test_identity[2-shape1] PASSED   [ 93%]
tests/pytorch/mock_sparse/test_diag.py::test_identity[2-shape2] PASSED   [ 93%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_sparse[add-dtype0-idtype0] PASSED [ 93%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_sparse[add-dtype0-idtype1] PASSED [ 93%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_sparse[add-dtype1-idtype0] PASSED [ 93%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_sparse[add-dtype1-idtype1] PASSED [ 93%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_sparse[sub-dtype0-idtype0] PASSED [ 93%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_sparse[sub-dtype0-idtype1] PASSED [ 93%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_sparse[sub-dtype1-idtype0] PASSED [ 93%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_sparse[sub-dtype1-idtype1] PASSED [ 93%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_sparse[mul-dtype0-idtype0] PASSED [ 93%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_sparse[mul-dtype0-idtype1] PASSED [ 93%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_sparse[mul-dtype1-idtype0] PASSED [ 93%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_sparse[mul-dtype1-idtype1] PASSED [ 93%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_sparse[truediv-dtype0-idtype0] PASSED [ 93%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_sparse[truediv-dtype0-idtype1] PASSED [ 93%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_sparse[truediv-dtype1-idtype0] PASSED [ 93%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_sparse[truediv-dtype1-idtype1] PASSED [ 93%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_diag[add-dtype0-idtype0] PASSED [ 93%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_diag[add-dtype0-idtype1] PASSED [ 93%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_diag[add-dtype1-idtype0] PASSED [ 93%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_diag[add-dtype1-idtype1] PASSED [ 93%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_diag[sub-dtype0-idtype0] PASSED [ 93%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_diag[sub-dtype0-idtype1] PASSED [ 93%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_diag[sub-dtype1-idtype0] PASSED [ 93%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_diag[sub-dtype1-idtype1] PASSED [ 93%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_diag[mul-dtype0-idtype0] PASSED [ 94%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_diag[mul-dtype0-idtype1] PASSED [ 94%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_diag[mul-dtype1-idtype0] PASSED [ 94%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_diag[mul-dtype1-idtype1] PASSED [ 94%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_diag[truediv-dtype0-idtype0] PASSED [ 94%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_diag[truediv-dtype0-idtype1] PASSED [ 94%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_diag[truediv-dtype1-idtype0] PASSED [ 94%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_diag[truediv-dtype1-idtype1] PASSED [ 94%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_scalar[2-dtype0-idtype0] PASSED [ 94%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_scalar[2-dtype0-idtype1] PASSED [ 94%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_scalar[2-dtype1-idtype0] PASSED [ 94%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_scalar[2-dtype1-idtype1] PASSED [ 94%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_scalar[2.5-dtype0-idtype0] PASSED [ 94%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_scalar[2.5-dtype0-idtype1] PASSED [ 94%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_scalar[2.5-dtype1-idtype0] PASSED [ 94%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_sparse_op_scalar[2.5-dtype1-idtype1] PASSED [ 94%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_scalar_op_sparse[2-dtype0-idtype0] PASSED [ 94%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_scalar_op_sparse[2-dtype0-idtype1] PASSED [ 94%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_scalar_op_sparse[2-dtype1-idtype0] PASSED [ 94%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_scalar_op_sparse[2-dtype1-idtype1] PASSED [ 94%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_scalar_op_sparse[2.5-dtype0-idtype0] PASSED [ 94%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_scalar_op_sparse[2.5-dtype0-idtype1] PASSED [ 94%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_scalar_op_sparse[2.5-dtype1-idtype0] PASSED [ 94%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_scalar_op_sparse[2.5-dtype1-idtype1] PASSED [ 94%]
tests/pytorch/mock_sparse/test_elementwise_op_sp.py::test_expose_op PASSED [ 94%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-sum-extra_shape0-col0-row0-None] SKIPPED [ 94%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-sum-extra_shape0-col0-row0-2] SKIPPED [ 94%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-sum-extra_shape0-col0-row1-None] SKIPPED [ 94%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-sum-extra_shape0-col0-row1-2] SKIPPED [ 94%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-sum-extra_shape0-col1-row0-None] SKIPPED [ 94%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-sum-extra_shape0-col1-row0-2] SKIPPED [ 94%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-sum-extra_shape0-col1-row1-None] SKIPPED [ 94%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-sum-extra_shape0-col1-row1-2] SKIPPED [ 94%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-sum-extra_shape1-col0-row0-None] SKIPPED [ 94%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-sum-extra_shape1-col0-row0-2] SKIPPED [ 94%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-sum-extra_shape1-col0-row1-None] SKIPPED [ 94%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-sum-extra_shape1-col0-row1-2] SKIPPED [ 94%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-sum-extra_shape1-col1-row0-None] SKIPPED [ 94%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-sum-extra_shape1-col1-row0-2] SKIPPED [ 94%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-sum-extra_shape1-col1-row1-None] SKIPPED [ 94%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-sum-extra_shape1-col1-row1-2] SKIPPED [ 95%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smax-extra_shape0-col0-row0-None] SKIPPED [ 95%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smax-extra_shape0-col0-row0-2] SKIPPED [ 95%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smax-extra_shape0-col0-row1-None] SKIPPED [ 95%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smax-extra_shape0-col0-row1-2] SKIPPED [ 95%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smax-extra_shape0-col1-row0-None] SKIPPED [ 95%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smax-extra_shape0-col1-row0-2] SKIPPED [ 95%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smax-extra_shape0-col1-row1-None] SKIPPED [ 95%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smax-extra_shape0-col1-row1-2] SKIPPED [ 95%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smax-extra_shape1-col0-row0-None] SKIPPED [ 95%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smax-extra_shape1-col0-row0-2] SKIPPED [ 95%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smax-extra_shape1-col0-row1-None] SKIPPED [ 95%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smax-extra_shape1-col0-row1-2] SKIPPED [ 95%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smax-extra_shape1-col1-row0-None] SKIPPED [ 95%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smax-extra_shape1-col1-row0-2] SKIPPED [ 95%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smax-extra_shape1-col1-row1-None] SKIPPED [ 95%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smax-extra_shape1-col1-row1-2] SKIPPED [ 95%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smin-extra_shape0-col0-row0-None] SKIPPED [ 95%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smin-extra_shape0-col0-row0-2] SKIPPED [ 95%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smin-extra_shape0-col0-row1-None] SKIPPED [ 95%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smin-extra_shape0-col0-row1-2] SKIPPED [ 95%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smin-extra_shape0-col1-row0-None] SKIPPED [ 95%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smin-extra_shape0-col1-row0-2] SKIPPED [ 95%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smin-extra_shape0-col1-row1-None] SKIPPED [ 95%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smin-extra_shape0-col1-row1-2] SKIPPED [ 95%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smin-extra_shape1-col0-row0-None] SKIPPED [ 95%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smin-extra_shape1-col0-row0-2] SKIPPED [ 95%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smin-extra_shape1-col0-row1-None] SKIPPED [ 95%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smin-extra_shape1-col0-row1-2] SKIPPED [ 95%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smin-extra_shape1-col1-row0-None] SKIPPED [ 95%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smin-extra_shape1-col1-row0-2] SKIPPED [ 95%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smin-extra_shape1-col1-row1-None] SKIPPED [ 95%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smin-extra_shape1-col1-row1-2] SKIPPED [ 95%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smean-extra_shape0-col0-row0-None] SKIPPED [ 95%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smean-extra_shape0-col0-row0-2] SKIPPED [ 95%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smean-extra_shape0-col0-row1-None] SKIPPED [ 95%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smean-extra_shape0-col0-row1-2] SKIPPED [ 95%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smean-extra_shape0-col1-row0-None] SKIPPED [ 95%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smean-extra_shape0-col1-row0-2] SKIPPED [ 95%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smean-extra_shape0-col1-row1-None] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smean-extra_shape0-col1-row1-2] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smean-extra_shape1-col0-row0-None] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smean-extra_shape1-col0-row0-2] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smean-extra_shape1-col0-row1-None] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smean-extra_shape1-col0-row1-2] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smean-extra_shape1-col1-row0-None] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smean-extra_shape1-col1-row0-2] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smean-extra_shape1-col1-row1-None] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[None-smean-extra_shape1-col1-row1-2] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-sum-extra_shape0-col0-row0-None] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-sum-extra_shape0-col0-row0-2] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-sum-extra_shape0-col0-row1-None] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-sum-extra_shape0-col0-row1-2] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-sum-extra_shape0-col1-row0-None] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-sum-extra_shape0-col1-row0-2] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-sum-extra_shape0-col1-row1-None] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-sum-extra_shape0-col1-row1-2] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-sum-extra_shape1-col0-row0-None] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-sum-extra_shape1-col0-row0-2] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-sum-extra_shape1-col0-row1-None] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-sum-extra_shape1-col0-row1-2] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-sum-extra_shape1-col1-row0-None] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-sum-extra_shape1-col1-row0-2] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-sum-extra_shape1-col1-row1-None] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-sum-extra_shape1-col1-row1-2] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smax-extra_shape0-col0-row0-None] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smax-extra_shape0-col0-row0-2] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smax-extra_shape0-col0-row1-None] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smax-extra_shape0-col0-row1-2] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smax-extra_shape0-col1-row0-None] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smax-extra_shape0-col1-row0-2] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smax-extra_shape0-col1-row1-None] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smax-extra_shape0-col1-row1-2] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smax-extra_shape1-col0-row0-None] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smax-extra_shape1-col0-row0-2] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smax-extra_shape1-col0-row1-None] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smax-extra_shape1-col0-row1-2] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smax-extra_shape1-col1-row0-None] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smax-extra_shape1-col1-row0-2] SKIPPED [ 96%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smax-extra_shape1-col1-row1-None] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smax-extra_shape1-col1-row1-2] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smin-extra_shape0-col0-row0-None] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smin-extra_shape0-col0-row0-2] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smin-extra_shape0-col0-row1-None] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smin-extra_shape0-col0-row1-2] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smin-extra_shape0-col1-row0-None] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smin-extra_shape0-col1-row0-2] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smin-extra_shape0-col1-row1-None] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smin-extra_shape0-col1-row1-2] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smin-extra_shape1-col0-row0-None] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smin-extra_shape1-col0-row0-2] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smin-extra_shape1-col0-row1-None] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smin-extra_shape1-col0-row1-2] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smin-extra_shape1-col1-row0-None] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smin-extra_shape1-col1-row0-2] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smin-extra_shape1-col1-row1-None] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smin-extra_shape1-col1-row1-2] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smean-extra_shape0-col0-row0-None] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smean-extra_shape0-col0-row0-2] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smean-extra_shape0-col0-row1-None] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smean-extra_shape0-col0-row1-2] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smean-extra_shape0-col1-row0-None] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smean-extra_shape0-col1-row0-2] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smean-extra_shape0-col1-row1-None] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smean-extra_shape0-col1-row1-2] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smean-extra_shape1-col0-row0-None] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smean-extra_shape1-col0-row0-2] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smean-extra_shape1-col0-row1-None] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smean-extra_shape1-col0-row1-2] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smean-extra_shape1-col1-row0-None] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smean-extra_shape1-col1-row0-2] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smean-extra_shape1-col1-row1-None] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[0-smean-extra_shape1-col1-row1-2] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-sum-extra_shape0-col0-row0-None] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-sum-extra_shape0-col0-row0-2] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-sum-extra_shape0-col0-row1-None] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-sum-extra_shape0-col0-row1-2] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-sum-extra_shape0-col1-row0-None] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-sum-extra_shape0-col1-row0-2] SKIPPED [ 97%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-sum-extra_shape0-col1-row1-None] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-sum-extra_shape0-col1-row1-2] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-sum-extra_shape1-col0-row0-None] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-sum-extra_shape1-col0-row0-2] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-sum-extra_shape1-col0-row1-None] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-sum-extra_shape1-col0-row1-2] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-sum-extra_shape1-col1-row0-None] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-sum-extra_shape1-col1-row0-2] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-sum-extra_shape1-col1-row1-None] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-sum-extra_shape1-col1-row1-2] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smax-extra_shape0-col0-row0-None] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smax-extra_shape0-col0-row0-2] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smax-extra_shape0-col0-row1-None] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smax-extra_shape0-col0-row1-2] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smax-extra_shape0-col1-row0-None] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smax-extra_shape0-col1-row0-2] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smax-extra_shape0-col1-row1-None] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smax-extra_shape0-col1-row1-2] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smax-extra_shape1-col0-row0-None] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smax-extra_shape1-col0-row0-2] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smax-extra_shape1-col0-row1-None] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smax-extra_shape1-col0-row1-2] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smax-extra_shape1-col1-row0-None] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smax-extra_shape1-col1-row0-2] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smax-extra_shape1-col1-row1-None] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smax-extra_shape1-col1-row1-2] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smin-extra_shape0-col0-row0-None] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smin-extra_shape0-col0-row0-2] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smin-extra_shape0-col0-row1-None] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smin-extra_shape0-col0-row1-2] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smin-extra_shape0-col1-row0-None] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smin-extra_shape0-col1-row0-2] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smin-extra_shape0-col1-row1-None] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smin-extra_shape0-col1-row1-2] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smin-extra_shape1-col0-row0-None] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smin-extra_shape1-col0-row0-2] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smin-extra_shape1-col0-row1-None] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smin-extra_shape1-col0-row1-2] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smin-extra_shape1-col1-row0-None] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smin-extra_shape1-col1-row0-2] SKIPPED [ 98%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smin-extra_shape1-col1-row1-None] SKIPPED [ 99%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smin-extra_shape1-col1-row1-2] SKIPPED [ 99%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smean-extra_shape0-col0-row0-None] SKIPPED [ 99%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smean-extra_shape0-col0-row0-2] SKIPPED [ 99%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smean-extra_shape0-col0-row1-None] SKIPPED [ 99%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smean-extra_shape0-col0-row1-2] SKIPPED [ 99%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smean-extra_shape0-col1-row0-None] SKIPPED [ 99%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smean-extra_shape0-col1-row0-2] SKIPPED [ 99%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smean-extra_shape0-col1-row1-None] SKIPPED [ 99%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smean-extra_shape0-col1-row1-2] SKIPPED [ 99%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smean-extra_shape1-col0-row0-None] SKIPPED [ 99%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smean-extra_shape1-col0-row0-2] SKIPPED [ 99%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smean-extra_shape1-col0-row1-None] SKIPPED [ 99%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smean-extra_shape1-col0-row1-2] SKIPPED [ 99%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smean-extra_shape1-col1-row0-None] SKIPPED [ 99%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smean-extra_shape1-col1-row0-2] SKIPPED [ 99%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smean-extra_shape1-col1-row1-None] SKIPPED [ 99%]
tests/pytorch/mock_sparse/test_reduction.py::test_reduction[1-smean-extra_shape1-col1-row1-2] SKIPPED [ 99%]
tests/pytorch/mock_sparse/test_transpose.py::test_diag_matrix_transpose[None-val_shape0] PASSED [ 99%]
tests/pytorch/mock_sparse/test_transpose.py::test_diag_matrix_transpose[None-val_shape1] PASSED [ 99%]
tests/pytorch/mock_sparse/test_transpose.py::test_diag_matrix_transpose[mat_shape1-val_shape0] PASSED [ 99%]
tests/pytorch/mock_sparse/test_transpose.py::test_diag_matrix_transpose[mat_shape1-val_shape1] PASSED [ 99%]
tests/pytorch/mock_sparse/test_transpose.py::test_diag_matrix_transpose[mat_shape2-val_shape0] PASSED [ 99%]
tests/pytorch/mock_sparse/test_transpose.py::test_diag_matrix_transpose[mat_shape2-val_shape1] PASSED [ 99%]
tests/pytorch/mock_sparse/test_transpose.py::test_sparse_matrix_transpose[extra_shape0-col0-row0-None] PASSED [ 99%]
tests/pytorch/mock_sparse/test_transpose.py::test_sparse_matrix_transpose[extra_shape0-col0-row0-2] PASSED [ 99%]
tests/pytorch/mock_sparse/test_transpose.py::test_sparse_matrix_transpose[extra_shape0-col0-row1-None] PASSED [ 99%]
tests/pytorch/mock_sparse/test_transpose.py::test_sparse_matrix_transpose[extra_shape0-col0-row1-2] PASSED [ 99%]
tests/pytorch/mock_sparse/test_transpose.py::test_sparse_matrix_transpose[extra_shape0-col1-row0-None] PASSED [ 99%]
tests/pytorch/mock_sparse/test_transpose.py::test_sparse_matrix_transpose[extra_shape0-col1-row0-2] PASSED [ 99%]
tests/pytorch/mock_sparse/test_transpose.py::test_sparse_matrix_transpose[extra_shape0-col1-row1-None] PASSED [ 99%]
tests/pytorch/mock_sparse/test_transpose.py::test_sparse_matrix_transpose[extra_shape0-col1-row1-2] PASSED [ 99%]
tests/pytorch/mock_sparse/test_transpose.py::test_sparse_matrix_transpose[extra_shape1-col0-row0-None] PASSED [ 99%]
tests/pytorch/mock_sparse/test_transpose.py::test_sparse_matrix_transpose[extra_shape1-col0-row0-2] PASSED [ 99%]
tests/pytorch/mock_sparse/test_transpose.py::test_sparse_matrix_transpose[extra_shape1-col0-row1-None] PASSED [ 99%]
tests/pytorch/mock_sparse/test_transpose.py::test_sparse_matrix_transpose[extra_shape1-col0-row1-2] PASSED [ 99%]
tests/pytorch/mock_sparse/test_transpose.py::test_sparse_matrix_transpose[extra_shape1-col1-row0-None] PASSED [ 99%]
tests/pytorch/mock_sparse/test_transpose.py::test_sparse_matrix_transpose[extra_shape1-col1-row0-2] PASSED [ 99%]
tests/pytorch/mock_sparse/test_transpose.py::test_sparse_matrix_transpose[extra_shape1-col1-row1-None] PASSED [ 99%]
tests/pytorch/mock_sparse/test_transpose.py::test_sparse_matrix_transpose[extra_shape1-col1-row1-2] PASSED [100%]

=============================== warnings summary ===============================
python/dgl/backend/pytorch/tensor.py:16
python/dgl/backend/pytorch/tensor.py:16
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/backend/pytorch/tensor.py:16: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    if LooseVersion(th.__version__) < LooseVersion("1.9.0"):

python/dgl/backend/pytorch/tensor.py:340
python/dgl/backend/pytorch/tensor.py:340
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/backend/pytorch/tensor.py:340: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    if LooseVersion(th.__version__) >= LooseVersion("1.10.0"):

python/dgl/dataloading/dataloader.py:33
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/dataloading/dataloader.py:33: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    PYTORCH_VER = LooseVersion(torch.__version__)

python/dgl/_dataloading/pytorch/dataloader.py:23
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/_dataloading/pytorch/dataloader.py:23: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    PYTORCH_VER = LooseVersion(th.__version__)

python/dgl/_dataloading/pytorch/dataloader.py:24
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/_dataloading/pytorch/dataloader.py:24: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    PYTORCH_16 = PYTORCH_VER >= LooseVersion("1.6.0")

python/dgl/_dataloading/pytorch/dataloader.py:25
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/_dataloading/pytorch/dataloader.py:25: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    PYTORCH_17 = PYTORCH_VER >= LooseVersion("1.7.0")

tests/pytorch/test_dataloader.py::test_graph_dataloader
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/data/minigc.py:159: DeprecationWarning: `np.int` is a deprecated alias for the builtin `int`. To silence this warning, use `int` by itself. Doing this will not modify any behavior and is safe. When replacing `np.int`, you may wish to use e.g. `np.int64` or `np.int32` to specify the precision. If you wish to review your current use, check the release note link for additional information.
  Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations
    self.labels = F.tensor(np.array(self.labels).astype(np.int))

tests/pytorch/test_dataloader.py: 10 warnings
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/dataloading/dataloader.py:863: DGLWarning: Dataloader CPU affinity opt is not enabled, consider switching it on (see enable_cpu_affinity() or CPU best practices for DGL [https://docs.dgl.ai/tutorials/cpu/cpu_best_practises.html])
    dgl_warning(f'Dataloader CPU affinity opt is not enabled, consider switching it on '

tests/pytorch/test_dataloader.py: 700 warnings
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/dataloading/dataloader.py:78: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    if PYTORCH_VER >= LooseVersion("1.10.0"):

tests/pytorch/test_dataloader.py: 128 warnings
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/dataloading/dataloader.py:83: DGLWarning: The current output_nodes are out of order even if set shuffle to False in Dataloader, the reason is that the current version of torch dose not support stable sort. Please update torch to 1.10.0 or higher to fix it.
    'The current output_nodes are out of order even if set shuffle '

tests/pytorch/test_nn.py: 11 warnings
tests/pytorch/test_pickle.py: 1 warning
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/heterograph.py:72: DGLWarning: Recommend creating graphs by `dgl.graph(data)` instead of `dgl.DGLGraph(data)`.
    dgl_warning('Recommend creating graphs by `dgl.graph(data)`'

tests/pytorch/test_nn.py::test_set_trans
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/nn/pytorch/glob.py:700: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).
    lengths_x = th.tensor(lengths_x, dtype=th.int64, device=device)

tests/pytorch/test_nn.py::test_set_trans
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/nn/pytorch/glob.py:701: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).
    lengths_mem = th.tensor(lengths_mem, dtype=th.int64, device=device)

tests/pytorch/test_nn.py::test_dense_cheb_conv[1]
tests/pytorch/test_nn.py::test_dense_cheb_conv[2]
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/heterograph.py:84: DGLWarning: Keyword arguments ['readonly'] are deprecated in v0.5, and can be safely removed in all cases.
    ' removed in all cases.'.format(list(deprecate_kwargs.keys())))

-- Docs: https://docs.pytest.org/en/stable/how-to/capture-warnings.html
-- generated xml file: /root/jenkins/workspace/dgl_PR-4648/pytest_backend.xml --
============================ slowest 100 durations =============================
25.71s call     tests/pytorch/test_dist_optim.py::test_sparse_opt
5.17s call     tests/pytorch/test_unified_tensor.py::test_multi_gpu_unified_tensor[1]
5.11s call     tests/pytorch/test_sparse_emb.py::test_multiprocess_sparse_emb_get_set[1]
2.52s call     tests/pytorch/test_dataloader.py::test_cluster_gcn[0]
1.75s call     tests/pytorch/test_dataloader.py::test_edge_dataloader_excludes[1-sampler1-None-False]
1.45s call     tests/pytorch/test_dataloader.py::test_edge_dataloader_excludes[1-sampler0-None-False]
1.31s call     tests/pytorch/test_dataloader.py::test_node_dataloader[True-cpu-neighbor-idtype0]
1.30s call     tests/pytorch/test_dataloader.py::test_node_dataloader[True-cpu-full-idtype0]
1.28s call     tests/pytorch/test_dataloader.py::test_node_dataloader[True-cpu-full-idtype1]
1.25s call     tests/pytorch/test_dataloader.py::test_node_dataloader[True-cpu-neighbor2-idtype1]
1.21s call     tests/pytorch/test_dataloader.py::test_node_dataloader[True-cpu-neighbor2-idtype0]
1.20s call     tests/pytorch/test_dataloader.py::test_node_dataloader[True-cpu-neighbor-idtype1]
1.18s call     tests/pytorch/test_dataloader.py::test_node_dataloader[False-cpu-full-idtype0]
1.14s call     tests/pytorch/test_dataloader.py::test_node_dataloader[False-cpu-neighbor-idtype1]
1.12s call     tests/pytorch/test_dataloader.py::test_node_dataloader[False-cpu-neighbor2-idtype1]
1.10s call     tests/pytorch/test_dataloader.py::test_node_dataloader[False-cpu-full-idtype1]
1.07s call     tests/pytorch/test_dataloader.py::test_node_dataloader[False-cpu-neighbor2-idtype0]
1.07s call     tests/pytorch/test_ipc.py::test_torch_ipc
1.04s call     tests/pytorch/test_dataloader.py::test_node_dataloader[False-cpu-neighbor-idtype0]
0.98s call     tests/pytorch/test_dataloader.py::test_neighbor_nonuniform[False-cpu-idtype0]
0.82s call     tests/pytorch/test_dataloader.py::test_saint[edge-4]
0.81s call     tests/pytorch/test_dataloader.py::test_saint[walk-4]
0.80s call     tests/pytorch/test_dataloader.py::test_saint[node-4]
0.77s call     tests/pytorch/test_dataloader.py::test_neighbor_nonuniform[True-cpu-idtype0]
0.74s call     tests/pytorch/test_dataloader.py::test_neighbor_nonuniform[True-cpu-idtype1]
0.70s call     tests/pytorch/test_dataloader.py::test_cluster_gcn[4]
0.69s call     tests/pytorch/test_dataloader.py::test_shadow[4]
0.63s call     tests/pytorch/test_dataloader.py::test_neighbor_nonuniform[False-cpu-idtype1]
0.53s call     tests/pytorch/test_dataloader.py::test_saint[edge-0]
0.49s call     tests/pytorch/test_nn.py::test_gnnexplainer[1-g0-idtype0]
0.48s call     tests/pytorch/test_nn.py::test_gnnexplainer[2-g1-idtype1]
0.48s call     tests/pytorch/test_nn.py::test_gnnexplainer[2-g2-idtype1]
0.48s call     tests/pytorch/test_nn.py::test_gnnexplainer[2-g2-idtype0]
0.48s call     tests/pytorch/test_nn.py::test_gnnexplainer[2-g5-idtype1]
0.48s call     tests/pytorch/test_nn.py::test_gnnexplainer[2-g1-idtype0]
0.48s call     tests/pytorch/test_nn.py::test_gnnexplainer[2-g3-idtype1]
0.48s call     tests/pytorch/test_nn.py::test_gnnexplainer[2-g0-idtype1]
0.48s call     tests/pytorch/test_nn.py::test_gnnexplainer[2-g4-idtype1]
0.48s call     tests/pytorch/test_nn.py::test_gnnexplainer[1-g1-idtype0]
0.48s call     tests/pytorch/test_nn.py::test_gnnexplainer[2-g4-idtype0]
0.48s call     tests/pytorch/test_nn.py::test_gnnexplainer[2-g5-idtype0]
0.48s call     tests/pytorch/test_nn.py::test_gnnexplainer[1-g2-idtype0]
0.48s call     tests/pytorch/test_nn.py::test_gnnexplainer[2-g0-idtype0]
0.48s call     tests/pytorch/test_nn.py::test_gnnexplainer[2-g3-idtype0]
0.48s call     tests/pytorch/test_dataloader.py::test_saint[walk-0]
0.48s call     tests/pytorch/test_nn.py::test_gnnexplainer[1-g5-idtype0]
0.48s call     tests/pytorch/test_nn.py::test_gnnexplainer[1-g4-idtype0]
0.48s call     tests/pytorch/test_nn.py::test_gnnexplainer[1-g3-idtype0]
0.47s call     tests/pytorch/test_nn.py::test_gnnexplainer[1-g0-idtype1]
0.47s call     tests/pytorch/test_nn.py::test_gnnexplainer[1-g2-idtype1]
0.47s call     tests/pytorch/test_nn.py::test_gnnexplainer[1-g1-idtype1]
0.47s call     tests/pytorch/test_dataloader.py::test_shadow[0]
0.47s call     tests/pytorch/test_nn.py::test_gnnexplainer[1-g5-idtype1]
0.47s call     tests/pytorch/test_nn.py::test_gnnexplainer[1-g4-idtype1]
0.47s call     tests/pytorch/test_nn.py::test_gnnexplainer[1-g3-idtype1]
0.45s call     tests/pytorch/test_dataloader.py::test_saint[node-0]
0.32s call     tests/pytorch/test_dataloader.py::test_neighbor_nonuniform[False-pure_gpu-idtype0]
0.27s call     tests/pytorch/test_dataloader.py::test_edge_dataloader[True-cpu-neg_sampler1-full-idtype0]
0.27s call     tests/pytorch/test_dataloader.py::test_edge_dataloader[True-cpu-neg_sampler2-full-idtype0]
0.26s call     tests/pytorch/test_dataloader.py::test_edge_dataloader[True-cpu-neg_sampler0-full-idtype0]
0.26s call     tests/pytorch/test_dataloader.py::test_edge_dataloader[True-cpu-neg_sampler0-full-idtype1]
0.26s call     tests/pytorch/test_dataloader.py::test_edge_dataloader[True-cpu-neg_sampler1-full-idtype1]
0.24s call     tests/pytorch/test_dataloader.py::test_edge_dataloader[False-cpu-neg_sampler0-full-idtype0]
0.24s call     tests/pytorch/test_dataloader.py::test_edge_dataloader[False-cpu-neg_sampler1-full-idtype0]
0.24s call     tests/pytorch/test_dataloader.py::test_edge_dataloader[True-cpu-neg_sampler2-full-idtype1]
0.23s call     tests/pytorch/test_dataloader.py::test_edge_dataloader[False-cpu-neg_sampler1-full-idtype1]
0.22s call     tests/pytorch/test_nn.py::test_graph_conv[1-True-True-none-g0-idtype0]
0.22s call     tests/pytorch/test_dataloader.py::test_edge_dataloader[False-cpu-neg_sampler2-full-idtype1]
0.22s call     tests/pytorch/test_dataloader.py::test_edge_dataloader[False-cpu-neg_sampler2-full-idtype0]
0.21s call     tests/pytorch/test_dataloader.py::test_edge_dataloader[False-cpu-neg_sampler0-full-idtype1]
0.19s call     tests/pytorch/test_dataloader.py::test_edge_dataloader[True-cpu-neg_sampler1-neighbor-idtype0]
0.19s call     tests/pytorch/test_dataloader.py::test_edge_dataloader[True-cpu-neg_sampler1-neighbor-idtype1]
0.18s call     tests/pytorch/test_dataloader.py::test_edge_dataloader[True-cpu-neg_sampler2-neighbor-idtype0]
0.18s call     tests/pytorch/test_dataloader.py::test_edge_dataloader[False-cpu-neg_sampler1-neighbor-idtype0]
0.17s call     tests/pytorch/test_dataloader.py::test_edge_dataloader[True-cpu-neg_sampler2-neighbor-idtype1]
0.17s call     tests/pytorch/test_dataloader.py::test_edge_dataloader[True-cpu-neg_sampler0-neighbor-idtype1]
0.17s call     tests/pytorch/test_dataloader.py::test_edge_dataloader[True-cpu-neg_sampler0-neighbor-idtype0]
0.16s call     tests/pytorch/test_dataloader.py::test_edge_dataloader[False-cpu-neg_sampler2-neighbor-idtype0]
0.15s call     tests/pytorch/test_dataloader.py::test_edge_dataloader[False-cpu-neg_sampler2-neighbor-idtype1]
0.15s call     tests/pytorch/test_dataloader.py::test_edge_dataloader[False-cpu-neg_sampler1-neighbor-idtype1]
0.14s call     tests/pytorch/test_dataloader.py::test_edge_dataloader[False-cpu-neg_sampler0-neighbor-idtype1]
0.14s call     tests/pytorch/test_dataloader.py::test_edge_dataloader[False-cpu-neg_sampler0-neighbor-idtype0]
0.14s call     tests/pytorch/test_dataloader.py::test_edge_dataloader_excludes[50-sampler1-reverse_types-True]
0.12s call     tests/pytorch/test_dataloader.py::test_edge_dataloader_excludes[50-sampler1-reverse_types-False]
0.12s call     tests/pytorch/test_dataloader.py::test_edge_dataloader_excludes[1-sampler1-reverse_types-True]
0.12s call     tests/pytorch/test_dataloader.py::test_edge_dataloader[True-uva-neg_sampler0-full-idtype0]
0.11s call     tests/pytorch/test_dataloader.py::test_edge_dataloader_excludes[1-sampler1-reverse_types-False]
0.10s call     tests/pytorch/test_unified_tensor.py::test_unified_tensor
0.10s call     tests/pytorch/test_dataloader.py::test_edge_dataloader[True-uva-neg_sampler0-full-idtype1]
0.10s call     tests/pytorch/test_nn.py::test_dgn_conv[True-0-4-0.1-2.5-scalers0-aggregators0-16-16]
0.09s call     tests/pytorch/test_dataloader.py::test_edge_dataloader[True-pure_gpu-neg_sampler2-neighbor-idtype0]
0.09s call     tests/pytorch/test_dataloader.py::test_edge_dataloader[True-uva-neg_sampler0-neighbor-idtype1]
0.09s call     tests/pytorch/test_dataloader.py::test_edge_dataloader[True-uva-neg_sampler0-neighbor-idtype0]
0.09s call     tests/pytorch/test_dataloader.py::test_edge_dataloader[True-pure_gpu-neg_sampler0-neighbor-idtype1]
0.09s call     tests/pytorch/test_dataloader.py::test_edge_dataloader[True-pure_gpu-neg_sampler2-full-idtype0]
0.09s call     tests/pytorch/test_dataloader.py::test_edge_dataloader[True-pure_gpu-neg_sampler1-full-idtype1]
0.09s call     tests/pytorch/test_dataloader.py::test_edge_dataloader[True-pure_gpu-neg_sampler1-full-idtype0]
0.09s call     tests/pytorch/test_dataloader.py::test_edge_dataloader[True-pure_gpu-neg_sampler1-neighbor-idtype0]
0.09s call     tests/pytorch/test_dataloader.py::test_edge_dataloader[True-pure_gpu-neg_sampler2-full-idtype1]
0.09s call     tests/pytorch/test_dataloader.py::test_edge_dataloader[True-pure_gpu-neg_sampler1-neighbor-idtype1]
========= 3745 passed, 233 skipped, 863 warnings in 134.40s (0:02:14) ==========
[Pipeline] }
[Pipeline] // timeout
[Pipeline] }
[Pipeline] // stage
[Pipeline] stage
[Pipeline] { (Torch GPU Example test)
[Pipeline] sh
Submodule path 'third_party/thrust/dependencies/cub': checked out 'cdaa9558a85e45d849016e5fe7b6e4ee79113f95'
+ rm -rf CMakeLists.txt CONTRIBUTORS.md Jenkinsfile LICENSE NEWS.md README.md apps benchmarks build cluster_gcn.pkl cmake conda cora_full cora_full.zip dataset dglgo docker docs examples featgraph include minigc ogbg-molhiv-as-graphpred ogbl-collab-as-linkpred ogbn-arxiv-as-nodepred optim_ip_config.txt pyproject.toml pytest_backend.xml pytest_compute.xml python readthedocs.yml src tensoradapter tests third_party tools tutorials
Submodule path 'third_party/tvm': checked out 'b2e418cb109df4cd1f17a2cf2894a1b396a6b838'
Submodule 'dlpack' (https://github.com/dmlc/dlpack) registered for path 'third_party/tvm/3rdparty/dlpack'
Submodule 'dmlc-core' (https://github.com/dmlc/dmlc-core) registered for path 'third_party/tvm/3rdparty/dmlc-core'
Submodule '3rdparty/rang' (https://github.com/agauniyal/rang) registered for path 'third_party/tvm/3rdparty/rang'
Submodule '3rdparty/vta-hw' (https://github.com/apache/incubator-tvm-vta) registered for path 'third_party/tvm/3rdparty/vta-hw'
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tvm/3rdparty/dlpack'...
[Pipeline] checkout
The recommended git tool is: git
using credential 150de63f-189c-4717-bcaf-010460d2f51a
Warning: JENKINS-30600: special launcher org.jenkinsci.plugins.docker.workflow.WithContainerStep$Decorator$1@61a0ca14; decorates RemoteLauncher[hudson.remoting.Channel@41a64267:dglci-manual-gpu-worker] will be ignored (a typical symptom is the Git executable not being run inside a designated container)
Fetching changes from the remote Git repository
Cleaning workspace
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tvm/3rdparty/dmlc-core'...
Fetching without tags
Merging remotes/origin/master commit d78a3a4baf611b90871a849f58647160c7cd9ab4 into PR head commit 37ed78035f406940ed88a4b60cadd2b7cfde5fea
Merge succeeded, producing 37ed78035f406940ed88a4b60cadd2b7cfde5fea
Checking out Revision 37ed78035f406940ed88a4b60cadd2b7cfde5fea (PR-4648)
Commit message: "fix for pytorch < 1.12"
Cleaning workspace
 > git rev-parse --resolve-git-dir /root/jenkins/workspace/dgl_PR-4648/.git # timeout=10
 > git config remote.origin.url https://github.com/dmlc/dgl.git # timeout=10
 > git rev-parse --verify HEAD # timeout=10
Resetting working tree
 > git reset --hard # timeout=10
 > git clean -fdx # timeout=10
Fetching upstream changes from https://github.com/dmlc/dgl.git
 > git --version # timeout=10
 > git --version # 'git version 2.17.1'
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git fetch --no-tags --progress -- https://github.com/dmlc/dgl.git +refs/pull/4648/head:refs/remotes/origin/PR-4648 +refs/heads/master:refs/remotes/origin/master # timeout=10
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git remote # timeout=10
 > git config --get remote.origin.url # timeout=10
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git merge d78a3a4baf611b90871a849f58647160c7cd9ab4 # timeout=10
 > git rev-parse HEAD^{commit} # timeout=10
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git branch -a -v --no-abbrev # timeout=10
 > git branch -D PR-4648 # timeout=10
 > git checkout -b PR-4648 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git rev-parse --verify HEAD # timeout=10
Resetting working tree
 > git reset --hard # timeout=10
 > git clean -fdx # timeout=10
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tvm/3rdparty/rang'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tvm/3rdparty/vta-hw'...
[Pipeline] sh
+ git submodule update --recursive --init
Submodule path 'third_party/tvm/3rdparty/dlpack': checked out '3ec04430e89a6834e5a1b99471f415fa939bf642'
Submodule path 'third_party/METIS': checked out '10603482cc055626ff99d2ac3ab7e73d1119439d'
Submodule path 'third_party/METIS/GKlib': checked out '62de20c7f97c719abdc96d7c3e266a1f7cb52bc5'
Submodule path 'third_party/dlpack': checked out 'e2bdd3bee8cb6501558042633fa59144cc8b7f5f'
Submodule path 'third_party/dmlc-core': checked out 'bfad207b448480783a1f428ae3d93d87032d8349'
Submodule path 'third_party/googletest': checked out 'f71fb4f9a912ec945401cc49a287a759b6131026'
Submodule path 'third_party/tvm/3rdparty/dmlc-core': checked out '6c401e242c59a1f4c913918246591bb13fd714e7'
Submodule path 'third_party/tvm/3rdparty/rang': checked out 'cabe04d6d6b05356fa8f9741704924788f0dd762'
Submodule path 'third_party/tvm/3rdparty/vta-hw': checked out '87ce9acfae550d1a487746e9d06c2e250076e54c'
Submodule path 'third_party/xbyak': checked out '757e4063f6464740b8ff4a2cae9136d2f8458020'
[Pipeline] unstash
[Pipeline] echo
Unpacked build/libdgl.so, build/runUnitTests, python/dgl/_ffi/_cy3/core.cpython-*-x86_64-linux-gnu.so, build/tensoradapter/pytorch/*.so from dgl-cpu-linux
[Pipeline] timeout
Timeout set to expire in 30 min
[Pipeline] {
[Pipeline] sh
Submodule path 'third_party/libxsmm': checked out 'fa687556130b6298430f1c0555a14cf79ab6101c'
Submodule path 'third_party/nanoflann': checked out '4c47ca200209550c5628c89803591f8a753c8181'
Submodule path 'third_party/nccl': checked out 'e11238b3029795d33f958b5868d47c90c4f22628'
Submodule path 'third_party/phmap': checked out '25293cefd8b85491b45600c03fe8edf07647553f'
+ bash tests/scripts/task_distributed_test.sh pytorch cpu
Submodule path 'third_party/tensorpipe': checked out '6042f1a4cbce8eef997f11ed0012de137b317361'
Submodule path 'third_party/tensorpipe/third_party/googletest': checked out 'aee0f9d9b5b87796ee8a0ab26b7587ec30e8858e'
Submodule path 'third_party/tensorpipe/third_party/libnop': checked out 'aa95422ea8c409e3f078d2ee7708a5f59a8b9fa2'
Submodule path 'third_party/tensorpipe/third_party/libuv': checked out '1dff88e5161cba5c59276d2070d2e304e4dcb242'
Submodule path 'third_party/tensorpipe/third_party/pybind11': checked out 'a23996fce38ff6ccfbcdc09f1e63f2c4be5ea2ef'
Submodule path 'third_party/tensorpipe/third_party/pybind11/tools/clang': checked out '6a00cbc4a9b8e68b71caf7f774b3f9c753ae84d5'
Submodule path 'third_party/thrust': checked out '6a3078c64cab0e2f276340fa5dcafa0d758ed890'
Submodule path 'third_party/thrust/dependencies/cub': checked out 'cdaa9558a85e45d849016e5fe7b6e4ee79113f95'
Requirement already satisfied: pytest in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (7.1.2)
Collecting psutil
  Downloading psutil-5.9.2-cp37-cp37m-manylinux_2_12_x86_64.manylinux2010_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl (281 kB)
Submodule path 'third_party/tvm': checked out 'b2e418cb109df4cd1f17a2cf2894a1b396a6b838'
Submodule path 'third_party/tvm/3rdparty/dlpack': checked out '3ec04430e89a6834e5a1b99471f415fa939bf642'
Submodule path 'third_party/tvm/3rdparty/dmlc-core': checked out '6c401e242c59a1f4c913918246591bb13fd714e7'
Submodule path 'third_party/tvm/3rdparty/rang': checked out 'cabe04d6d6b05356fa8f9741704924788f0dd762'
Submodule path 'third_party/tvm/3rdparty/vta-hw': checked out '87ce9acfae550d1a487746e9d06c2e250076e54c'
Submodule path 'third_party/xbyak': checked out '757e4063f6464740b8ff4a2cae9136d2f8458020'
[Pipeline] unstash
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 281.3/281.3 kB 8.5 MB/s eta 0:00:00
Collecting pyyaml
  Downloading PyYAML-6.0-cp37-cp37m-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_12_x86_64.manylinux2010_x86_64.whl (596 kB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 596.3/596.3 kB 32.7 MB/s eta 0:00:00
Collecting pydantic
  Downloading pydantic-1.10.2-cp37-cp37m-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (11.8 MB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 11.8/11.8 MB 84.0 MB/s eta 0:00:00
Requirement already satisfied: pandas in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (1.1.5)
Collecting rdflib
  Downloading rdflib-6.2.0-py3-none-any.whl (500 kB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 500.3/500.3 kB 48.0 MB/s eta 0:00:00
Requirement already satisfied: ogb in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (1.3.3)
Collecting filelock
  Downloading filelock-3.8.0-py3-none-any.whl (10 kB)
Collecting pyarrow
  Downloading pyarrow-9.0.0-cp37-cp37m-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (35.3 MB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 35.3/35.3 MB 53.0 MB/s eta 0:00:00
Requirement already satisfied: pluggy<2.0,>=0.12 in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from pytest) (1.0.0)
Requirement already satisfied: attrs>=19.2.0 in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from pytest) (22.1.0)
Requirement already satisfied: iniconfig in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from pytest) (1.1.1)
Requirement already satisfied: packaging in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from pytest) (21.3)
Requirement already satisfied: tomli>=1.0.0 in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from pytest) (2.0.1)
Requirement already satisfied: py>=1.8.2 in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from pytest) (1.11.0)
Requirement already satisfied: importlib-metadata>=0.12 in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from pytest) (4.12.0)
Requirement already satisfied: typing-extensions>=4.1.0 in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from pydantic) (4.3.0)
Requirement already satisfied: numpy>=1.15.4 in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from pandas) (1.21.6)
Requirement already satisfied: pytz>=2017.2 in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from pandas) (2022.2.1)
Requirement already satisfied: python-dateutil>=2.7.3 in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from pandas) (2.8.2)
Requirement already satisfied: pyparsing in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from rdflib) (3.0.9)
Requirement already satisfied: setuptools in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from rdflib) (61.2.0)
Collecting isodate
  Downloading isodate-0.6.1-py2.py3-none-any.whl (41 kB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 41.7/41.7 kB 7.1 MB/s eta 0:00:00
Requirement already satisfied: torch>=1.6.0 in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from ogb) (1.9.0+cpu)
Requirement already satisfied: tqdm>=4.29.0 in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from ogb) (4.64.0)
Requirement already satisfied: six>=1.12.0 in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from ogb) (1.16.0)
Requirement already satisfied: scikit-learn>=0.20.0 in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from ogb) (1.0.2)
Requirement already satisfied: urllib3>=1.24.0 in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from ogb) (1.26.11)
Requirement already satisfied: outdated>=0.2.0 in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from ogb) (0.2.1)
Requirement already satisfied: zipp>=0.5 in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from importlib-metadata>=0.12->pytest) (3.8.1)
Requirement already satisfied: requests in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from outdated>=0.2.0->ogb) (2.28.1)
Requirement already satisfied: littleutils in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from outdated>=0.2.0->ogb) (0.2.2)
Requirement already satisfied: threadpoolctl>=2.0.0 in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from scikit-learn>=0.20.0->ogb) (3.1.0)
Requirement already satisfied: scipy>=1.1.0 in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from scikit-learn>=0.20.0->ogb) (1.7.3)
Requirement already satisfied: joblib>=0.11 in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from scikit-learn>=0.20.0->ogb) (1.1.0)
Requirement already satisfied: certifi>=2017.4.17 in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from requests->outdated>=0.2.0->ogb) (2022.6.15)
Requirement already satisfied: charset-normalizer<3,>=2 in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from requests->outdated>=0.2.0->ogb) (2.1.0)
Requirement already satisfied: idna<4,>=2.5 in /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages (from requests->outdated>=0.2.0->ogb) (3.3)
Installing collected packages: pyyaml, pydantic, pyarrow, psutil, isodate, filelock, rdflib
Successfully installed filelock-3.8.0 isodate-0.6.1 psutil-5.9.2 pyarrow-9.0.0 pydantic-1.10.2 pyyaml-6.0 rdflib-6.2.0
WARNING: Running pip as the 'root' user can result in broken permissions and conflicting behaviour with the system package manager. It is recommended to use a virtual environment instead: https://pip.pypa.io/warnings/venv
============================= test session starts ==============================
platform linux -- Python 3.7.0, pytest-7.1.2, pluggy-1.0.0 -- /opt/conda/envs/pytorch-ci/bin/python3
cachedir: .pytest_cache
rootdir: /root/jenkins/workspace/dgl_PR-4648
collecting ... [Pipeline] echo
Unpacked build/libdgl.so, build/runUnitTests, python/dgl/_ffi/_cy3/core.cpython-*-x86_64-linux-gnu.so, build/tensoradapter/pytorch/*.so from dgl-gpu-linux
[Pipeline] timeout
Timeout set to expire in 20 min
[Pipeline] {
[Pipeline] sh
+ bash tests/scripts/task_example_test.sh gpu
run graph store with port 7815
collected 55 items

tests/distributed/test_dist_graph_store.py::test_server_client /root/jenkins/workspace/dgl_PR-4648/python/dgl/heterograph.py:72: DGLWarning: Recommend creating graphs by `dgl.graph(data)` instead of `dgl.DGLGraph(data)`.
  dgl_warning('Recommend creating graphs by `dgl.graph(data)`'
tensor([0.0036, 0.0078, 0.0176, 0.0096, 0.0114, 0.0081, 0.0141, 0.0080, 0.0070,
        0.0140, 0.0134, 0.0092, 0.0082, 0.0113, 0.0089, 0.0142, 0.0063, 0.0083,
        0.0083, 0.0097, 0.0065, 0.0048, 0.0065, 0.0069, 0.0061, 0.0092, 0.0079,
        0.0112, 0.0069, 0.0067, 0.0122, 0.0094, 0.0090, 0.0130, 0.0107, 0.0128,
        0.0052, 0.0114, 0.0078, 0.0101, 0.0037, 0.0127, 0.0129, 0.0141, 0.0095,
        0.0049, 0.0080, 0.0133, 0.0204, 0.0128, 0.0053, 0.0130, 0.0142, 0.0123,
        0.0062, 0.0124, 0.0082, 0.0124, 0.0124, 0.0101, 0.0052, 0.0143, 0.0048,
        0.0156, 0.0129, 0.0142, 0.0078, 0.0097, 0.0063, 0.0046, 0.0085, 0.0120,
        0.0063, 0.0108, 0.0099, 0.0128, 0.0095, 0.0120, 0.0051, 0.0080, 0.0069,
        0.0195, 0.0102, 0.0080, 0.0079, 0.0015, 0.0098, 0.0151, 0.0092, 0.0137,
        0.0225, 0.0066, 0.0141, 0.0100, 0.0079, 0.0083, 0.0098, 0.0098, 0.0082,
        0.0101])
Converting to homogeneous graph takes 0.004s, peak mem: 1.356 GB
Convert a graph into a bidirected graph: 0.005 seconds, peak memory: 1.356 GB
Construct multi-constraint weights: 0.000 seconds, peak memory: 1.356 GB
[05:28:40] /root/jenkins/workspace/dgl_PR-4648/src/graph/transform/metis_partition_hetero.cc:87: Partition a graph with 10000 nodes and 199872 edges into 4 parts and get 56541 edge cuts
Metis partitioning: 0.049 seconds, peak memory: 1.356 GB
Assigning nodes to METIS partitions takes 0.055s, peak mem: 1.356 GB
Save partitions: 0.004 seconds, peak memory: 1.356 GB
There are 100000 edges in the graph and 0 edge cuts for 1 partitions.
start client 0
start client 1
[05:28:41] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:28:41] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load dist_graph_test_2
start server 0
start graph service on server 0 for part 0
[05:28:41] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:28:41] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:13571]...
start client 2
[05:28:42] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:28:42] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
start client 3
[05:28:42] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:28:42] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
[05:28:43] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:28:43] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Training with DGL built-in GraphConv module.
Downloading /root/jenkins/workspace/dgl_PR-4648/cora_v2.zip from https://data.dgl.ai/dataset/cora_v2.zip...
Extracting file to /root/jenkins/workspace/dgl_PR-4648/cora_v2
Finished data loading and preprocessing.
  NumNodes: 2708
  NumEdges: 10556
  NumFeats: 1433
  NumClasses: 7
  NumTrainingSamples: 140
  NumValidationSamples: 500
  NumTestSamples: 1000
Done saving data into cached files.
Training...
Epoch 00000 | Loss 1.9462 | Accuracy 0.2580 
Epoch 00001 | Loss 1.9409 | Accuracy 0.3240 
Epoch 00002 | Loss 1.9337 | Accuracy 0.3020 
Epoch 00003 | Loss 1.9265 | Accuracy 0.3420 
Epoch 00004 | Loss 1.9188 | Accuracy 0.3700 
Epoch 00005 | Loss 1.9101 | Accuracy 0.4140 
Epoch 00006 | Loss 1.9004 | Accuracy 0.4980 
Epoch 00007 | Loss 1.8909 | Accuracy 0.5740 
Epoch 00008 | Loss 1.8814 | Accuracy 0.6220 
Epoch 00009 | Loss 1.8704 | Accuracy 0.6320 
Epoch 00010 | Loss 1.8582 | Accuracy 0.6320 
Epoch 00011 | Loss 1.8513 | Accuracy 0.6280 
Epoch 00012 | Loss 1.8343 | Accuracy 0.6480 
Epoch 00013 | Loss 1.8235 | Accuracy 0.6680 
Epoch 00014 | Loss 1.8096 | Accuracy 0.6700 
Epoch 00015 | Loss 1.7921 | Accuracy 0.6900 
Epoch 00016 | Loss 1.7764 | Accuracy 0.7080 
Epoch 00017 | Loss 1.7533 | Accuracy 0.7020 
Epoch 00018 | Loss 1.7434 | Accuracy 0.6940 
Epoch 00019 | Loss 1.7264 | Accuracy 0.6920 
Epoch 00020 | Loss 1.7177 | Accuracy 0.6980 
Epoch 00021 | Loss 1.6990 | Accuracy 0.7080 
Epoch 00022 | Loss 1.6810 | Accuracy 0.7060 
Epoch 00023 | Loss 1.6548 | Accuracy 0.7120 
Epoch 00024 | Loss 1.6515 | Accuracy 0.7160 
Epoch 00025 | Loss 1.6183 | Accuracy 0.7220 
Epoch 00026 | Loss 1.6064 | Accuracy 0.7300 
Epoch 00027 | Loss 1.5987 | Accuracy 0.7420 
Epoch 00028 | Loss 1.5722 | Accuracy 0.7480 
Epoch 00029 | Loss 1.5536 | Accuracy 0.7560 
Epoch 00030 | Loss 1.5210 | Accuracy 0.7600 
Epoch 00031 | Loss 1.5227 | Accuracy 0.7740 
Epoch 00032 | Loss 1.4687 | Accuracy 0.7740 
Epoch 00033 | Loss 1.4591 | Accuracy 0.7760 
Epoch 00034 | Loss 1.4282 | Accuracy 0.7780 
Epoch 00035 | Loss 1.3976 | Accuracy 0.7780 
Epoch 00036 | Loss 1.3886 | Accuracy 0.7800 
Epoch 00037 | Loss 1.3805 | Accuracy 0.7800 
Epoch 00038 | Loss 1.3435 | Accuracy 0.7820 
Epoch 00039 | Loss 1.3721 | Accuracy 0.7820 
Epoch 00040 | Loss 1.3254 | Accuracy 0.7820 
Epoch 00041 | Loss 1.2808 | Accuracy 0.7800 
Epoch 00042 | Loss 1.2840 | Accuracy 0.7740 
Epoch 00043 | Loss 1.2243 | Accuracy 0.7780 
Epoch 00044 | Loss 1.2208 | Accuracy 0.7760 
Epoch 00045 | Loss 1.2212 | Accuracy 0.7760 
Epoch 00046 | Loss 1.1585 | Accuracy 0.7780 
Epoch 00047 | Loss 1.1558 | Accuracy 0.7800 
Epoch 00048 | Loss 1.1347 | Accuracy 0.7820 
Epoch 00049 | Loss 1.1403 | Accuracy 0.7880 
Epoch 00050 | Loss 1.1074 | Accuracy 0.7920 
Epoch 00051 | Loss 1.0424 | Accuracy 0.7920 
Epoch 00052 | Loss 1.0537 | Accuracy 0.7940 
Epoch 00053 | Loss 1.0229 | Accuracy 0.7960 
Epoch 00054 | Loss 1.0028 | Accuracy 0.7940 
Epoch 00055 | Loss 1.0399 | Accuracy 0.7940 
Epoch 00056 | Loss 0.9750 | Accuracy 0.7960 
Epoch 00057 | Loss 0.9453 | Accuracy 0.7920 
Epoch 00058 | Loss 0.9648 | Accuracy 0.7880 
Epoch 00059 | Loss 0.9406 | Accuracy 0.7880 
Epoch 00060 | Loss 0.9175 | Accuracy 0.7880 
Epoch 00061 | Loss 0.9537 | Accuracy 0.7880 
Epoch 00062 | Loss 0.9176 | Accuracy 0.7820 
Epoch 00063 | Loss 0.8880 | Accuracy 0.7840 
Epoch 00064 | Loss 0.8529 | Accuracy 0.7840 
Epoch 00065 | Loss 0.8151 | Accuracy 0.7840 
Epoch 00066 | Loss 0.8164 | Accuracy 0.7880 
Epoch 00067 | Loss 0.8098 | Accuracy 0.7900 
Epoch 00068 | Loss 0.8055 | Accuracy 0.7920 
Epoch 00069 | Loss 0.7819 | Accuracy 0.7940 
Epoch 00070 | Loss 0.7831 | Accuracy 0.8000 
Epoch 00071 | Loss 0.7502 | Accuracy 0.7980 
Epoch 00072 | Loss 0.7618 | Accuracy 0.8000 
Epoch 00073 | Loss 0.7383 | Accuracy 0.7980 
Epoch 00074 | Loss 0.6967 | Accuracy 0.7980 
Epoch 00075 | Loss 0.7508 | Accuracy 0.7920 
Epoch 00076 | Loss 0.7197 | Accuracy 0.7840 
Epoch 00077 | Loss 0.7183 | Accuracy 0.7840 
Epoch 00078 | Loss 0.7003 | Accuracy 0.7840 
Epoch 00079 | Loss 0.7260 | Accuracy 0.7840 
Epoch 00080 | Loss 0.6990 | Accuracy 0.7840 
Epoch 00081 | Loss 0.7018 | Accuracy 0.7860 
Epoch 00082 | Loss 0.6802 | Accuracy 0.7900 
Epoch 00083 | Loss 0.6447 | Accuracy 0.7900 
Epoch 00084 | Loss 0.6260 | Accuracy 0.7880 
Epoch 00085 | Loss 0.6349 | Accuracy 0.7900 
Epoch 00086 | Loss 0.6400 | Accuracy 0.7900 
Epoch 00087 | Loss 0.6182 | Accuracy 0.7920 
Epoch 00088 | Loss 0.5869 | Accuracy 0.7920 
Epoch 00089 | Loss 0.6108 | Accuracy 0.7900 
Epoch 00090 | Loss 0.6057 | Accuracy 0.7900 
Epoch 00091 | Loss 0.5550 | Accuracy 0.7900 
Epoch 00092 | Loss 0.5728 | Accuracy 0.7860 
Epoch 00093 | Loss 0.6150 | Accuracy 0.7860 
Epoch 00094 | Loss 0.5653 | Accuracy 0.7860 
Epoch 00095 | Loss 0.5723 | Accuracy 0.7880 
Epoch 00096 | Loss 0.5682 | Accuracy 0.7880 
Epoch 00097 | Loss 0.5510 | Accuracy 0.7880 
Epoch 00098 | Loss 0.5710 | Accuracy 0.7880 
Epoch 00099 | Loss 0.5857 | Accuracy 0.7860 
Epoch 00100 | Loss 0.5164 | Accuracy 0.7860 
Epoch 00101 | Loss 0.5594 | Accuracy 0.7940 
Epoch 00102 | Loss 0.5639 | Accuracy 0.7920 
Epoch 00103 | Loss 0.4940 | Accuracy 0.7940 
Epoch 00104 | Loss 0.5070 | Accuracy 0.7940 
Epoch 00105 | Loss 0.5081 | Accuracy 0.7940 
Epoch 00106 | Loss 0.5046 | Accuracy 0.7940 
Epoch 00107 | Loss 0.5082 | Accuracy 0.7940 
Epoch 00108 | Loss 0.4807 | Accuracy 0.7920 
Epoch 00109 | Loss 0.4636 | Accuracy 0.7920 
Epoch 00110 | Loss 0.5204 | Accuracy 0.7880 
Epoch 00111 | Loss 0.4859 | Accuracy 0.7900 
Epoch 00112 | Loss 0.4774 | Accuracy 0.7900 
Epoch 00113 | Loss 0.4575 | Accuracy 0.7880 
Epoch 00114 | Loss 0.4771 | Accuracy 0.7960 
Epoch 00115 | Loss 0.5051 | Accuracy 0.7960 
Epoch 00116 | Loss 0.4516 | Accuracy 0.7960 
Epoch 00117 | Loss 0.4474 | Accuracy 0.7920 
Epoch 00118 | Loss 0.4601 | Accuracy 0.7920 
Epoch 00119 | Loss 0.4451 | Accuracy 0.7900 
Epoch 00120 | Loss 0.4848 | Accuracy 0.7900 
Epoch 00121 | Loss 0.4628 | Accuracy 0.7880 
Epoch 00122 | Loss 0.4522 | Accuracy 0.7880 
Epoch 00123 | Loss 0.4658 | Accuracy 0.7860 
Epoch 00124 | Loss 0.4926 | Accuracy 0.7920 
Epoch 00125 | Loss 0.4521 | Accuracy 0.7920 
Epoch 00126 | Loss 0.4403 | Accuracy 0.7920 
Epoch 00127 | Loss 0.4325 | Accuracy 0.7960 
Epoch 00128 | Loss 0.4528 | Accuracy 0.7920 
Epoch 00129 | Loss 0.4277 | Accuracy 0.7940 
Epoch 00130 | Loss 0.4162 | Accuracy 0.7980 
Epoch 00131 | Loss 0.4363 | Accuracy 0.7980 
Epoch 00132 | Loss 0.4738 | Accuracy 0.7980 
Epoch 00133 | Loss 0.4386 | Accuracy 0.7980 
Epoch 00134 | Loss 0.4705 | Accuracy 0.7940 
Epoch 00135 | Loss 0.4417 | Accuracy 0.7960 
Epoch 00136 | Loss 0.3999 | Accuracy 0.7940 
Epoch 00137 | Loss 0.4179 | Accuracy 0.7960 
Epoch 00138 | Loss 0.4020 | Accuracy 0.7920 
Epoch 00139 | Loss 0.4014 | Accuracy 0.7840 
Epoch 00140 | Loss 0.4077 | Accuracy 0.7860 
Epoch 00141 | Loss 0.3857 | Accuracy 0.7900 
Epoch 00142 | Loss 0.4134 | Accuracy 0.7920 
Epoch 00143 | Loss 0.4346 | Accuracy 0.7960 
Epoch 00144 | Loss 0.4352 | Accuracy 0.7960 
Epoch 00145 | Loss 0.4007 | Accuracy 0.8020 
Epoch 00146 | Loss 0.4016 | Accuracy 0.7980 
Epoch 00147 | Loss 0.3762 | Accuracy 0.7920 
Epoch 00148 | Loss 0.3756 | Accuracy 0.7880 
Epoch 00149 | Loss 0.3848 | Accuracy 0.7860 
Epoch 00150 | Loss 0.4018 | Accuracy 0.7820 
Epoch 00151 | Loss 0.3802 | Accuracy 0.7920 
Epoch 00152 | Loss 0.4017 | Accuracy 0.7900 
Epoch 00153 | Loss 0.3435 | Accuracy 0.7940 
Epoch 00154 | Loss 0.3867 | Accuracy 0.7980 
Epoch 00155 | Loss 0.3685 | Accuracy 0.8040 
Epoch 00156 | Loss 0.4117 | Accuracy 0.8000 
Epoch 00157 | Loss 0.3740 | Accuracy 0.8020 
Epoch 00158 | Loss 0.4028 | Accuracy 0.7980 
Epoch 00159 | Loss 0.3579 | Accuracy 0.7920 
Epoch 00160 | Loss 0.3868 | Accuracy 0.7960 
Epoch 00161 | Loss 0.3549 | Accuracy 0.7940 
Epoch 00162 | Loss 0.3999 | Accuracy 0.7940 
Epoch 00163 | Loss 0.3725 | Accuracy 0.7940 
Epoch 00164 | Loss 0.3903 | Accuracy 0.7900 
Epoch 00165 | Loss 0.3571 | Accuracy 0.7940 
Epoch 00166 | Loss 0.3791 | Accuracy 0.7960 
Epoch 00167 | Loss 0.3439 | Accuracy 0.8020 
Epoch 00168 | Loss 0.3250 | Accuracy 0.8040 
Epoch 00169 | Loss 0.3488 | Accuracy 0.8020 
Epoch 00170 | Loss 0.3246 | Accuracy 0.8020 
Epoch 00171 | Loss 0.3719 | Accuracy 0.8020 
Epoch 00172 | Loss 0.3256 | Accuracy 0.7940 
Epoch 00173 | Loss 0.3185 | Accuracy 0.7940 
Epoch 00174 | Loss 0.3587 | Accuracy 0.7920 
Epoch 00175 | Loss 0.3587 | Accuracy 0.7900 
Epoch 00176 | Loss 0.3303 | Accuracy 0.7880 
Epoch 00177 | Loss 0.3268 | Accuracy 0.7880 
Epoch 00178 | Loss 0.3563 | Accuracy 0.7880 
Epoch 00179 | Loss 0.3828 | Accuracy 0.7920 
Epoch 00180 | Loss 0.3369 | Accuracy 0.7980 
Epoch 00181 | Loss 0.3357 | Accuracy 0.8020 
Epoch 00182 | Loss 0.3519 | Accuracy 0.8020 
Epoch 00183 | Loss 0.3699 | Accuracy 0.8040 
Epoch 00184 | Loss 0.3319 | Accuracy 0.8020 
Epoch 00185 | Loss 0.3564 | Accuracy 0.7980 
Epoch 00186 | Loss 0.3564 | Accuracy 0.7960 
Epoch 00187 | Loss 0.3439 | Accuracy 0.7900 
Epoch 00188 | Loss 0.3303 | Accuracy 0.7920 
Epoch 00189 | Loss 0.3270 | Accuracy 0.7940 
Epoch 00190 | Loss 0.3114 | Accuracy 0.7960 
Epoch 00191 | Loss 0.3439 | Accuracy 0.7960 
Epoch 00192 | Loss 0.3258 | Accuracy 0.7980 
Epoch 00193 | Loss 0.3129 | Accuracy 0.8000 
Epoch 00194 | Loss 0.3424 | Accuracy 0.7980 
Epoch 00195 | Loss 0.3218 | Accuracy 0.7940 
Epoch 00196 | Loss 0.3207 | Accuracy 0.8000 
Epoch 00197 | Loss 0.3627 | Accuracy 0.7980 
Epoch 00198 | Loss 0.3112 | Accuracy 0.8000 
Epoch 00199 | Loss 0.3258 | Accuracy 0.7960 
Testing...
Test accuracy 0.8190
lda/lda_model.py:29: UserWarning: cached_property not found - using property instead
  warnings.warn("cached_property not found - using property instead")
Testing LatentDirichletAllocation ...
{'prior': {'doc': 0.1, 'word': 0.1}, 'rho': 1, 'mult': {'doc': 1, 'word': 1}, 'init': {'doc': (100.0, 100.0), 'word': (100.0, 100.0)}, 'word_data': [tensor([[0.0000e+00, 0.0000e+00, 0.0000e+00, 4.5631e-05, 0.0000e+00],
        [0.0000e+00, 0.0000e+00, 0.0000e+00, 6.1491e-06, 0.0000e+00],
        [0.0000e+00, 0.0000e+00, 0.0000e+00, 9.9963e-01, 0.0000e+00],
        [0.0000e+00, 0.0000e+00, 0.0000e+00, 4.5618e-05, 0.0000e+00],
        [0.0000e+00, 0.0000e+00, 0.0000e+00, 4.5618e-05, 0.0000e+00],
        [0.0000e+00, 0.0000e+00, 0.0000e+00, 4.5618e-05, 0.0000e+00],
        [0.0000e+00, 0.0000e+00, 0.0000e+00, 4.5618e-05, 0.0000e+00],
        [0.0000e+00, 0.0000e+00, 0.0000e+00, 4.5618e-05, 0.0000e+00],
        [0.0000e+00, 0.0000e+00, 0.0000e+00, 4.5620e-05, 0.0000e+00],
        [0.0000e+00, 0.0000e+00, 0.0000e+00, 4.5618e-05, 0.0000e+00]])]}
Testing LatentDirichletAllocation passed!
[Pipeline] }
[Pipeline] // timeout
[Pipeline] }
[Pipeline] // stage
Post stage
[Pipeline] cleanWs
[WS-CLEANUP] Deleting project workspace...
[WS-CLEANUP] Deferred wipeout is disabled by the job configuration...
[WS-CLEANUP] done
[Pipeline] }
$ docker stop --time=1 f9b6859df1d65887fd5c1562bda6c9300bf7673dd8efba82320500053c49173b
$ docker rm -f f9b6859df1d65887fd5c1562bda6c9300bf7673dd8efba82320500053c49173b
[Pipeline] // withDockerContainer
[Pipeline] }
[Pipeline] // withEnv
[Pipeline] }
Running on dglci-manual-gpu-worker in /root/jenkins/workspace/dgl_PR-4648
[Pipeline] // node
[Pipeline] {
[Pipeline] }
[Pipeline] // stage
[Pipeline] }
[Pipeline] checkout
The recommended git tool is: git
using credential 150de63f-189c-4717-bcaf-010460d2f51a
Cloning the remote Git repository
Cloning with configured refspecs honoured and without tags
Client [1104] waits on 172.17.0.3:56245Client [1112] waits on 172.17.0.3:53595Client [1122] waits on 172.17.0.3:40013

Client [1117] waits on 172.17.0.3:49365

Machine (0) group (0) client (0) connect to server successfuly!
Machine (0) group (0) client (1) connect to server successfuly!
Machine (0) group (0) client (2) connect to server successfuly!
Machine (0) group (0) client (3) connect to server successfuly!
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
Client[1] in group[0] is exiting...
Client[2] in group[0] is exiting...
Client[3] in group[0] is exiting...
Client[0] in group[0] is exiting...
Server (0) shutdown.
Server is exiting...
clients have terminated
Cloning repository https://github.com/dmlc/dgl.git
 > git init /root/jenkins/workspace/dgl_PR-4648 # timeout=10
Fetching upstream changes from https://github.com/dmlc/dgl.git
 > git --version # timeout=10
 > git --version # 'git version 2.17.1'
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git fetch --no-tags --progress -- https://github.com/dmlc/dgl.git +refs/pull/4648/head:refs/remotes/origin/PR-4648 +refs/heads/master:refs/remotes/origin/master # timeout=10
Cleaning workspace
Fetching without tags
 > git config remote.origin.url https://github.com/dmlc/dgl.git # timeout=10
 > git config --add remote.origin.fetch +refs/pull/4648/head:refs/remotes/origin/PR-4648 # timeout=10
 > git config --add remote.origin.fetch +refs/heads/master:refs/remotes/origin/master # timeout=10
 > git config remote.origin.url https://github.com/dmlc/dgl.git # timeout=10
 > git rev-parse --verify HEAD # timeout=10
No valid HEAD. Skipping the resetting
 > git clean -fdx # timeout=10
Fetching upstream changes from https://github.com/dmlc/dgl.git
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git fetch --no-tags --progress -- https://github.com/dmlc/dgl.git +refs/pull/4648/head:refs/remotes/origin/PR-4648 +refs/heads/master:refs/remotes/origin/master # timeout=10
Merging remotes/origin/master commit d78a3a4baf611b90871a849f58647160c7cd9ab4 into PR head commit 37ed78035f406940ed88a4b60cadd2b7cfde5fea
Merge succeeded, producing 37ed78035f406940ed88a4b60cadd2b7cfde5fea
Checking out Revision 37ed78035f406940ed88a4b60cadd2b7cfde5fea (PR-4648)
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git remote # timeout=10
 > git config --get remote.origin.url # timeout=10
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git merge d78a3a4baf611b90871a849f58647160c7cd9ab4 # timeout=10
 > git rev-parse HEAD^{commit} # timeout=10
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
Commit message: "fix for pytorch < 1.12"
Cleaning workspace
 > git branch -a -v --no-abbrev # timeout=10
 > git checkout -b PR-4648 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git rev-parse --verify HEAD # timeout=10
Resetting working tree
 > git reset --hard # timeout=10
 > git clean -fdx # timeout=10
[Pipeline] withEnv
[Pipeline] {
[Pipeline] isUnix
[Pipeline] withEnv
[Pipeline] {
[Pipeline] sh
Converting to homogeneous graph takes 0.003s, peak mem: 1.369 GB
Save partitions: 0.007 seconds, peak memory: 1.369 GB
There are 100000 edges in the graph and 0 edge cuts for 1 partitions.
start client 0
+ docker pull rapidsai/cugraph_nightly_torch-cuda:11.5-base-ubuntu18.04-py3.9-pytorch1.12.0-rapids22.10
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load dist_graph_test_1
start server 0
start graph service on server 0 for part 0
[05:28:52] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:28:52] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:24053]...
[05:28:52] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:28:52] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
11.5-base-ubuntu18.04-py3.9-pytorch1.12.0-rapids22.10: Pulling from rapidsai/cugraph_nightly_torch-cuda
Digest: sha256:72e8cb2632449beac4895f24b55018dea21f79da47fe39be05ef83a4fd3ddb67
Status: Image is up to date for rapidsai/cugraph_nightly_torch-cuda:11.5-base-ubuntu18.04-py3.9-pytorch1.12.0-rapids22.10
docker.io/rapidsai/cugraph_nightly_torch-cuda:11.5-base-ubuntu18.04-py3.9-pytorch1.12.0-rapids22.10
[Pipeline] }
[Pipeline] // withEnv
[Pipeline] isUnix
[Pipeline] withEnv
[Pipeline] {
[Pipeline] sh
+ docker inspect -f . rapidsai/cugraph_nightly_torch-cuda:11.5-base-ubuntu18.04-py3.9-pytorch1.12.0-rapids22.10
.
[Pipeline] }
[Pipeline] // withEnv
[Pipeline] withDockerContainer
dglci-manual-gpu-worker does not seem to be running inside a container
$ docker run -t -d -u 0:0 --runtime nvidia --shm-size=8gb -w /root/jenkins/workspace/dgl_PR-4648 -v /root/jenkins/workspace/dgl_PR-4648:/root/jenkins/workspace/dgl_PR-4648:rw,z -v /root/jenkins/workspace/dgl_PR-4648@tmp:/root/jenkins/workspace/dgl_PR-4648@tmp:rw,z -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** rapidsai/cugraph_nightly_torch-cuda:11.5-base-ubuntu18.04-py3.9-pytorch1.12.0-rapids22.10 cat
$ docker top b7adb4db5f4cf1d886ff23a8dd7a1caea99af2b94c6076d915adcd91d0c69ad9 -eo pid,comm
[Pipeline] {
[Pipeline] stage
[Pipeline] { (PyTorch Cugraph GPU Unit test)
[Pipeline] sh
+ nvidia-smi
Tue Sep 27 05:28:53 2022       
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 510.47.03    Driver Version: 510.47.03    CUDA Version: 11.6     |
|-------------------------------+----------------------+----------------------+
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|                               |                      |               MIG M. |
|===============================+======================+======================|
|   0  Tesla T4            On   | 00000000:00:1E.0 Off |                    0 |
| N/A   44C    P8    16W /  70W |      0MiB / 15360MiB |      0%      Default |
|                               |                      |                  N/A |
+-------------------------------+----------------------+----------------------+
                                                                               
+-----------------------------------------------------------------------------+
| Processes:                                                                  |
|  GPU   GI   CI        PID   Type   Process name                  GPU Memory |
|        ID   ID                                                   Usage      |
|=============================================================================|
|  No running processes found                                                 |
+-----------------------------------------------------------------------------+
[Pipeline] sh
+ rm -rf CMakeLists.txt CONTRIBUTORS.md Jenkinsfile LICENSE NEWS.md README.md apps benchmarks cmake conda dglgo docker docs examples featgraph include pyproject.toml python readthedocs.yml src tensoradapter tests third_party tools tutorials
[Pipeline] checkout
The recommended git tool is: git
using credential 150de63f-189c-4717-bcaf-010460d2f51a
Warning: JENKINS-30600: special launcher org.jenkinsci.plugins.docker.workflow.WithContainerStep$Decorator$1@b0ad4e; decorates RemoteLauncher[hudson.remoting.Channel@41a64267:dglci-manual-gpu-worker] will be ignored (a typical symptom is the Git executable not being run inside a designated container)
Fetching changes from the remote Git repository
Cleaning workspace
Fetching without tags
Merging remotes/origin/master commit d78a3a4baf611b90871a849f58647160c7cd9ab4 into PR head commit 37ed78035f406940ed88a4b60cadd2b7cfde5fea
Merge succeeded, producing 37ed78035f406940ed88a4b60cadd2b7cfde5fea
Checking out Revision 37ed78035f406940ed88a4b60cadd2b7cfde5fea (PR-4648)
Commit message: "fix for pytorch < 1.12"
Cleaning workspace
 > git rev-parse --resolve-git-dir /root/jenkins/workspace/dgl_PR-4648/.git # timeout=10
 > git config remote.origin.url https://github.com/dmlc/dgl.git # timeout=10
 > git rev-parse --verify HEAD # timeout=10
Resetting working tree
 > git reset --hard # timeout=10
 > git clean -fdx # timeout=10
Fetching upstream changes from https://github.com/dmlc/dgl.git
 > git --version # timeout=10
 > git --version # 'git version 2.17.1'
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git fetch --no-tags --progress -- https://github.com/dmlc/dgl.git +refs/pull/4648/head:refs/remotes/origin/PR-4648 +refs/heads/master:refs/remotes/origin/master # timeout=10
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git remote # timeout=10
 > git config --get remote.origin.url # timeout=10
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git merge d78a3a4baf611b90871a849f58647160c7cd9ab4 # timeout=10
 > git rev-parse HEAD^{commit} # timeout=10
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git branch -a -v --no-abbrev # timeout=10
 > git branch -D PR-4648 # timeout=10
 > git checkout -b PR-4648 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git rev-parse --verify HEAD # timeout=10
Resetting working tree
 > git reset --hard # timeout=10
 > git clean -fdx # timeout=10
Client [1163] waits on 172.17.0.3:32903
Machine (0) group (0) client (0) connect to server successfuly!
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
end
Client[0] in group[0] is exiting...
Server (0) shutdown.
Server is exiting...
clients have terminated
[Pipeline] sh
+ git submodule update --recursive --init
Submodule 'third_party/METIS' (https://github.com/KarypisLab/METIS.git) registered for path 'third_party/METIS'
Submodule 'third_party/dlpack' (https://github.com/dmlc/dlpack.git) registered for path 'third_party/dlpack'
Submodule 'third_party/dmlc-core' (https://github.com/dmlc/dmlc-core.git) registered for path 'third_party/dmlc-core'
Submodule 'third_party/googletest' (https://github.com/google/googletest.git) registered for path 'third_party/googletest'
Submodule 'third_party/libxsmm' (https://github.com/hfp/libxsmm.git) registered for path 'third_party/libxsmm'
Submodule 'third_party/nanoflann' (https://github.com/jlblancoc/nanoflann) registered for path 'third_party/nanoflann'
Submodule 'third_party/nccl' (https://github.com/nvidia/nccl) registered for path 'third_party/nccl'
Submodule 'third_party/phmap' (https://github.com/greg7mdp/parallel-hashmap.git) registered for path 'third_party/phmap'
Submodule 'third_party/tensorpipe' (https://github.com/pytorch/tensorpipe) registered for path 'third_party/tensorpipe'
Submodule 'third_party/thrust' (https://github.com/NVIDIA/thrust.git) registered for path 'third_party/thrust'
Submodule 'third_party/tvm' (https://github.com/apache/incubator-tvm) registered for path 'third_party/tvm'
Submodule 'third_party/xbyak' (https://github.com/herumi/xbyak) registered for path 'third_party/xbyak'
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/METIS'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/dlpack'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/dmlc-core'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/googletest'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/libxsmm'...
Converting to homogeneous graph takes 0.011s, peak mem: 1.371 GB
Save partitions: 0.029 seconds, peak memory: 1.371 GB
There are 300600 edges in the graph and 0 edge cuts for 1 partitions.
start client 0
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load dist_graph_test_3
[05:29:09] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:29:09] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
start server 0
start graph service on server 0 for part 0
[05:29:09] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:29:09] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:11907]...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/nanoflann'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/nccl'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/phmap'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tensorpipe'...
Client [1185] waits on 172.17.0.3:53831
Machine (0) group (0) client (0) connect to server successfuly!
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
end
Client[0] in group[0] is exiting...
Server (0) shutdown.
Server is exiting...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/thrust'...
clients have terminated
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tvm'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/xbyak'...
Submodule path 'third_party/METIS': checked out '10603482cc055626ff99d2ac3ab7e73d1119439d'
Submodule 'GKlib' (https://github.com/KarypisLab/GKlib.git) registered for path 'third_party/METIS/GKlib'
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/METIS/GKlib'...
Submodule path 'third_party/METIS/GKlib': checked out '62de20c7f97c719abdc96d7c3e266a1f7cb52bc5'
Submodule path 'third_party/dlpack': checked out 'e2bdd3bee8cb6501558042633fa59144cc8b7f5f'
Submodule path 'third_party/dmlc-core': checked out 'bfad207b448480783a1f428ae3d93d87032d8349'
Submodule path 'third_party/googletest': checked out 'f71fb4f9a912ec945401cc49a287a759b6131026'
Submodule path 'third_party/libxsmm': checked out 'fa687556130b6298430f1c0555a14cf79ab6101c'
Submodule path 'third_party/nanoflann': checked out '4c47ca200209550c5628c89803591f8a753c8181'
Submodule path 'third_party/nccl': checked out 'e11238b3029795d33f958b5868d47c90c4f22628'
Submodule path 'third_party/phmap': checked out '25293cefd8b85491b45600c03fe8edf07647553f'
Submodule path 'third_party/tensorpipe': checked out '6042f1a4cbce8eef997f11ed0012de137b317361'
Submodule 'third_party/googletest' (https://github.com/google/googletest.git) registered for path 'third_party/tensorpipe/third_party/googletest'
Submodule 'third_party/libnop' (https://github.com/google/libnop.git) registered for path 'third_party/tensorpipe/third_party/libnop'
Submodule 'third_party/libuv' (https://github.com/libuv/libuv.git) registered for path 'third_party/tensorpipe/third_party/libuv'
Submodule 'third_party/pybind11' (https://github.com/pybind/pybind11.git) registered for path 'third_party/tensorpipe/third_party/pybind11'
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tensorpipe/third_party/googletest'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tensorpipe/third_party/libnop'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tensorpipe/third_party/libuv'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tensorpipe/third_party/pybind11'...
Submodule path 'third_party/tensorpipe/third_party/googletest': checked out 'aee0f9d9b5b87796ee8a0ab26b7587ec30e8858e'
Submodule path 'third_party/tensorpipe/third_party/libnop': checked out 'aa95422ea8c409e3f078d2ee7708a5f59a8b9fa2'
Submodule path 'third_party/tensorpipe/third_party/libuv': checked out '1dff88e5161cba5c59276d2070d2e304e4dcb242'
Submodule path 'third_party/tensorpipe/third_party/pybind11': checked out 'a23996fce38ff6ccfbcdc09f1e63f2c4be5ea2ef'
Submodule 'tools/clang' (https://github.com/wjakob/clang-cindex-python3) registered for path 'third_party/tensorpipe/third_party/pybind11/tools/clang'
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tensorpipe/third_party/pybind11/tools/clang'...
Submodule path 'third_party/tensorpipe/third_party/pybind11/tools/clang': checked out '6a00cbc4a9b8e68b71caf7f774b3f9c753ae84d5'
Submodule path 'third_party/thrust': checked out '6a3078c64cab0e2f276340fa5dcafa0d758ed890'
Submodule 'cub' (https://github.com/NVIDIA/cub.git) registered for path 'third_party/thrust/dependencies/cub'
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/thrust/dependencies/cub'...
Submodule path 'third_party/thrust/dependencies/cub': checked out 'cdaa9558a85e45d849016e5fe7b6e4ee79113f95'
Submodule path 'third_party/tvm': checked out 'b2e418cb109df4cd1f17a2cf2894a1b396a6b838'
Submodule 'dlpack' (https://github.com/dmlc/dlpack) registered for path 'third_party/tvm/3rdparty/dlpack'
Submodule 'dmlc-core' (https://github.com/dmlc/dmlc-core) registered for path 'third_party/tvm/3rdparty/dmlc-core'
Submodule '3rdparty/rang' (https://github.com/agauniyal/rang) registered for path 'third_party/tvm/3rdparty/rang'
Submodule '3rdparty/vta-hw' (https://github.com/apache/incubator-tvm-vta) registered for path 'third_party/tvm/3rdparty/vta-hw'
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tvm/3rdparty/dlpack'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tvm/3rdparty/dmlc-core'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tvm/3rdparty/rang'...
Converting to homogeneous graph takes 0.019s, peak mem: 1.390 GB
Save partitions: 0.028 seconds, peak memory: 1.390 GB
There are 300600 edges in the graph and 0 edge cuts for 1 partitions.
start client 0
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tvm/3rdparty/vta-hw'...
[05:29:34] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:29:34] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load dist_graph_test_3
start server 0
start graph service on server 0 for part 0
[05:29:34] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:29:34] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:20710]...
Submodule path 'third_party/tvm/3rdparty/dlpack': checked out '3ec04430e89a6834e5a1b99471f415fa939bf642'
Submodule path 'third_party/tvm/3rdparty/dmlc-core': checked out '6c401e242c59a1f4c913918246591bb13fd714e7'
Submodule path 'third_party/tvm/3rdparty/rang': checked out 'cabe04d6d6b05356fa8f9741704924788f0dd762'
Submodule path 'third_party/tvm/3rdparty/vta-hw': checked out '87ce9acfae550d1a487746e9d06c2e250076e54c'
Submodule path 'third_party/xbyak': checked out '757e4063f6464740b8ff4a2cae9136d2f8458020'
[Pipeline] unstash
Client [1211] waits on 172.17.0.3:42831
Machine (0) group (0) client (0) connect to server successfuly!
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
end
Client[0] in group[0] is exiting...
Server (0) shutdown.
Server is exiting...
clients have terminated
[Pipeline] echo
Unpacked build/libdgl.so, build/runUnitTests, python/dgl/_ffi/_cy3/core.cpython-*-x86_64-linux-gnu.so, build/tensoradapter/pytorch/*.so from dgl-cugraph-linux
[Pipeline] timeout
Timeout set to expire in 15 min
[Pipeline] {
[Pipeline] sh
+ bash tests/scripts/cugraph_unit_test.sh pytorch
Collecting pytest
  Downloading pytest-7.1.3-py3-none-any.whl (298 kB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 298.2/298.2 kB 8.6 MB/s eta 0:00:00
Requirement already satisfied: psutil in /opt/conda/lib/python3.9/site-packages (5.9.2)
Requirement already satisfied: pyyaml in /opt/conda/lib/python3.9/site-packages (6.0)
Collecting pydantic
  Downloading pydantic-1.10.2-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (13.2 MB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 13.2/13.2 MB 98.0 MB/s eta 0:00:00
Requirement already satisfied: pandas in /opt/conda/lib/python3.9/site-packages (1.4.4)
Collecting rdflib
  Downloading rdflib-6.2.0-py3-none-any.whl (500 kB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 500.3/500.3 kB 67.0 MB/s eta 0:00:00
Collecting ogb
  Downloading ogb-1.3.4-py3-none-any.whl (78 kB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 78.6/78.6 kB 21.3 MB/s eta 0:00:00
Collecting pluggy<2.0,>=0.12
  Downloading pluggy-1.0.0-py2.py3-none-any.whl (13 kB)
Collecting iniconfig
  Downloading iniconfig-1.1.1-py2.py3-none-any.whl (5.0 kB)
Collecting tomli>=1.0.0
  Downloading tomli-2.0.1-py3-none-any.whl (12 kB)
Collecting attrs>=19.2.0
  Downloading attrs-22.1.0-py2.py3-none-any.whl (58 kB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 58.8/58.8 kB 12.6 MB/s eta 0:00:00
Collecting py>=1.8.2
  Downloading py-1.11.0-py2.py3-none-any.whl (98 kB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 98.7/98.7 kB 21.9 MB/s eta 0:00:00
Requirement already satisfied: packaging in /opt/conda/lib/python3.9/site-packages (from pytest) (21.3)
Requirement already satisfied: typing-extensions>=4.1.0 in /opt/conda/lib/python3.9/site-packages (from pydantic) (4.3.0)
Requirement already satisfied: python-dateutil>=2.8.1 in /opt/conda/lib/python3.9/site-packages (from pandas) (2.8.2)
Requirement already satisfied: pytz>=2020.1 in /opt/conda/lib/python3.9/site-packages (from pandas) (2022.2.1)
Requirement already satisfied: numpy>=1.18.5 in /opt/conda/lib/python3.9/site-packages (from pandas) (1.22.4)
Requirement already satisfied: pyparsing in /opt/conda/lib/python3.9/site-packages (from rdflib) (3.0.9)
Requirement already satisfied: setuptools in /opt/conda/lib/python3.9/site-packages (from rdflib) (65.3.0)
Collecting isodate
  Downloading isodate-0.6.1-py2.py3-none-any.whl (41 kB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 41.7/41.7 kB 11.3 MB/s eta 0:00:00
Requirement already satisfied: urllib3>=1.24.0 in /opt/conda/lib/python3.9/site-packages (from ogb) (1.26.11)
Requirement already satisfied: tqdm>=4.29.0 in /opt/conda/lib/python3.9/site-packages (from ogb) (4.64.1)
Requirement already satisfied: torch>=1.6.0 in /opt/conda/lib/python3.9/site-packages (from ogb) (1.12.0)
Collecting outdated>=0.2.0
  Downloading outdated-0.2.1-py3-none-any.whl (7.5 kB)
Collecting scikit-learn>=0.20.0
  Downloading scikit_learn-1.1.2-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (30.8 MB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 30.8/30.8 MB 68.6 MB/s eta 0:00:00
Requirement already satisfied: six>=1.12.0 in /opt/conda/lib/python3.9/site-packages (from ogb) (1.16.0)
Collecting littleutils
  Downloading littleutils-0.2.2.tar.gz (6.6 kB)
  Preparing metadata (setup.py): started
  Preparing metadata (setup.py): finished with status 'done'
Requirement already satisfied: requests in /opt/conda/lib/python3.9/site-packages (from outdated>=0.2.0->ogb) (2.28.1)
Requirement already satisfied: scipy>=1.3.2 in /opt/conda/lib/python3.9/site-packages (from scikit-learn>=0.20.0->ogb) (1.9.1)
Collecting threadpoolctl>=2.0.0
  Downloading threadpoolctl-3.1.0-py3-none-any.whl (14 kB)
Requirement already satisfied: joblib>=1.0.0 in /opt/conda/lib/python3.9/site-packages (from scikit-learn>=0.20.0->ogb) (1.2.0)
Requirement already satisfied: certifi>=2017.4.17 in /opt/conda/lib/python3.9/site-packages (from requests->outdated>=0.2.0->ogb) (2022.9.14)
Requirement already satisfied: charset-normalizer<3,>=2 in /opt/conda/lib/python3.9/site-packages (from requests->outdated>=0.2.0->ogb) (2.1.1)
Requirement already satisfied: idna<4,>=2.5 in /opt/conda/lib/python3.9/site-packages (from requests->outdated>=0.2.0->ogb) (3.3)
Building wheels for collected packages: littleutils
  Building wheel for littleutils (setup.py): started
  Building wheel for littleutils (setup.py): finished with status 'done'
  Created wheel for littleutils: filename=littleutils-0.2.2-py3-none-any.whl size=7028 sha256=e9f7281664cc0f6a5d41f79ccf3d793de57dc8c78c3d4405d9eb041e027b9f59
  Stored in directory: /root/.cache/pip/wheels/04/bb/0d/2d02ec45f29c48d6192476bfb59c5a0e64b605e7212374dd15
Successfully built littleutils
Installing collected packages: littleutils, iniconfig, tomli, threadpoolctl, pydantic, py, pluggy, isodate, attrs, scikit-learn, rdflib, pytest, outdated, ogb
Running on dgl-manual-large-cpu in /root/jenkins/workspace/dgl_PR-4648@4
[Pipeline] {
[Pipeline] checkout
The recommended git tool is: git
using credential 150de63f-189c-4717-bcaf-010460d2f51a
Cloning the remote Git repository
Cloning with configured refspecs honoured and without tags
Converting to homogeneous graph takes 0.005s, peak mem: 1.390 GB
Save partitions: 0.008 seconds, peak memory: 1.390 GB
There are 100000 edges in the graph and 0 edge cuts for 1 partitions.
start client[0] for group[0]
Cloning repository https://github.com/dmlc/dgl.git
 > git init /root/jenkins/workspace/dgl_PR-4648@4 # timeout=10
Fetching upstream changes from https://github.com/dmlc/dgl.git
 > git --version # timeout=10
 > git --version # 'git version 2.17.1'
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git fetch --no-tags --progress -- https://github.com/dmlc/dgl.git +refs/pull/4648/head:refs/remotes/origin/PR-4648 +refs/heads/master:refs/remotes/origin/master # timeout=10
[05:29:47] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:29:47] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load check_server_client_True_1_1_1
start server 0
start graph service on server 0 for part 0
[05:29:47] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:29:47] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:12813]...
Successfully installed attrs-22.1.0 iniconfig-1.1.1 isodate-0.6.1 littleutils-0.2.2 ogb-1.3.4 outdated-0.2.1 pluggy-1.0.0 py-1.11.0 pydantic-1.10.2 pytest-7.1.3 rdflib-6.2.0 scikit-learn-1.1.2 threadpoolctl-3.1.0 tomli-2.0.1
WARNING: Running pip as the 'root' user can result in broken permissions and conflicting behaviour with the system package manager. It is recommended to use a virtual environment instead: https://pip.pypa.io/warnings/venv
============================= test session starts ==============================
platform linux -- Python 3.9.13, pytest-7.1.3, pluggy-1.0.0 -- /opt/conda/bin/python3
cachedir: .pytest_cache
rootdir: /root/jenkins/workspace/dgl_PR-4648
collecting ... Cleaning workspace
Fetching without tags
PASSED                           [  5%]
tests/compute/test_data.py::test_explain_syn Merging remotes/origin/master commit d78a3a4baf611b90871a849f58647160c7cd9ab4 into PR head commit 37ed78035f406940ed88a4b60cadd2b7cfde5fea
Merge succeeded, producing 37ed78035f406940ed88a4b60cadd2b7cfde5fea
Checking out Revision 37ed78035f406940ed88a4b60cadd2b7cfde5fea (PR-4648)
 > git config remote.origin.url https://github.com/dmlc/dgl.git # timeout=10
 > git config --add remote.origin.fetch +refs/pull/4648/head:refs/remotes/origin/PR-4648 # timeout=10
 > git config --add remote.origin.fetch +refs/heads/master:refs/remotes/origin/master # timeout=10
 > git config remote.origin.url https://github.com/dmlc/dgl.git # timeout=10
 > git rev-parse --verify HEAD # timeout=10
No valid HEAD. Skipping the resetting
 > git clean -fdx # timeout=10
Fetching upstream changes from https://github.com/dmlc/dgl.git
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git fetch --no-tags --progress -- https://github.com/dmlc/dgl.git +refs/pull/4648/head:refs/remotes/origin/PR-4648 +refs/heads/master:refs/remotes/origin/master # timeout=10
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git remote # timeout=10
 > git config --get remote.origin.url # timeout=10
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git merge d78a3a4baf611b90871a849f58647160c7cd9ab4 # timeout=10
 > git rev-parse HEAD^{commit} # timeout=10
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
Commit message: "fix for pytorch < 1.12"
Cleaning workspace
[Pipeline] withEnv
[Pipeline] {
[Pipeline] isUnix
[Pipeline] withEnv
[Pipeline] {
[Pipeline] sh
PASSED                      [  5%]
tests/compute/test_data.py::test_wiki_cs + docker pull dgllib/dgl-ci-cpu:cu101_v220629
 > git branch -a -v --no-abbrev # timeout=10
 > git checkout -b PR-4648 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git rev-parse --verify HEAD # timeout=10
Resetting working tree
 > git reset --hard # timeout=10
 > git clean -fdx # timeout=10
cu101_v220629: Pulling from dgllib/dgl-ci-cpu
Digest: sha256:1f698bdb17257817c14ca7339a211328f6dc310f344a4a7fbf4cc7e53854427c
Status: Image is up to date for dgllib/dgl-ci-cpu:cu101_v220629
docker.io/dgllib/dgl-ci-cpu:cu101_v220629
[Pipeline] }
[Pipeline] // withEnv
[Pipeline] isUnix
[Pipeline] withEnv
[Pipeline] {
[Pipeline] sh
+ docker inspect -f . dgllib/dgl-ci-cpu:cu101_v220629
.
[Pipeline] }
[Pipeline] // withEnv
[Pipeline] withDockerContainer
dgl-manual-large-cpu does not seem to be running inside a container
$ docker run -t -d -u 0:0 -w /root/jenkins/workspace/dgl_PR-4648@4 -v /root/jenkins/workspace/dgl_PR-4648@4:/root/jenkins/workspace/dgl_PR-4648@4:rw,z -v /root/jenkins/workspace/dgl_PR-4648@4@tmp:/root/jenkins/workspace/dgl_PR-4648@4@tmp:rw,z -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** dgllib/dgl-ci-cpu:cu101_v220629 cat
$ docker top 6db97419e6b6b2f1427e0e55271c5f8554e9cd989f9b44da6a3861a1be7f5a89 -eo pid,comm
[Pipeline] {
[Pipeline] stage
[Pipeline] { (MXNet CPU Unit test)
[Pipeline] sh
+ rm -rf CMakeLists.txt CONTRIBUTORS.md Jenkinsfile LICENSE NEWS.md README.md apps benchmarks cmake conda dglgo docker docs examples featgraph include pyproject.toml python readthedocs.yml src tensoradapter tests third_party tools tutorials
[Pipeline] checkout
The recommended git tool is: git
using credential 150de63f-189c-4717-bcaf-010460d2f51a
Warning: JENKINS-30600: special launcher org.jenkinsci.plugins.docker.workflow.WithContainerStep$Decorator$1@45e8259d; decorates RemoteLauncher[hudson.remoting.Channel@1d9b9638:dgl-manual-large-cpu] will be ignored (a typical symptom is the Git executable not being run inside a designated container)
Fetching changes from the remote Git repository
Cleaning workspace
 > git rev-parse --resolve-git-dir /root/jenkins/workspace/dgl_PR-4648@4/.git # timeout=10
 > git config remote.origin.url https://github.com/dmlc/dgl.git # timeout=10
 > git rev-parse --verify HEAD # timeout=10
Resetting working tree
 > git reset --hard # timeout=10
Fetching without tags
Merging remotes/origin/master commit d78a3a4baf611b90871a849f58647160c7cd9ab4 into PR head commit 37ed78035f406940ed88a4b60cadd2b7cfde5fea
Merge succeeded, producing 37ed78035f406940ed88a4b60cadd2b7cfde5fea
Checking out Revision 37ed78035f406940ed88a4b60cadd2b7cfde5fea (PR-4648)
Client [1229] waits on 172.17.0.3:49273
Machine (0) group (0) client (0) connect to server successfuly!
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
end
Client[0] in group[0] is exiting...
Commit message: "fix for pytorch < 1.12"
Cleaning workspace
 > git clean -fdx # timeout=10
Fetching upstream changes from https://github.com/dmlc/dgl.git
 > git --version # timeout=10
 > git --version # 'git version 2.17.1'
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git fetch --no-tags --progress -- https://github.com/dmlc/dgl.git +refs/pull/4648/head:refs/remotes/origin/PR-4648 +refs/heads/master:refs/remotes/origin/master # timeout=10
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git remote # timeout=10
 > git config --get remote.origin.url # timeout=10
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git merge d78a3a4baf611b90871a849f58647160c7cd9ab4 # timeout=10
 > git rev-parse HEAD^{commit} # timeout=10
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git branch -a -v --no-abbrev # timeout=10
 > git branch -D PR-4648 # timeout=10
 > git checkout -b PR-4648 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git rev-parse --verify HEAD # timeout=10
Resetting working tree
 > git reset --hard # timeout=10
 > git clean -fdx # timeout=10
Server (0) shutdown.
Server is exiting...
clients have terminated
[Pipeline] sh
+ git submodule update --recursive --init
Submodule 'third_party/METIS' (https://github.com/KarypisLab/METIS.git) registered for path 'third_party/METIS'
Submodule 'third_party/dlpack' (https://github.com/dmlc/dlpack.git) registered for path 'third_party/dlpack'
Submodule 'third_party/dmlc-core' (https://github.com/dmlc/dmlc-core.git) registered for path 'third_party/dmlc-core'
Submodule 'third_party/googletest' (https://github.com/google/googletest.git) registered for path 'third_party/googletest'
Submodule 'third_party/libxsmm' (https://github.com/hfp/libxsmm.git) registered for path 'third_party/libxsmm'
Submodule 'third_party/nanoflann' (https://github.com/jlblancoc/nanoflann) registered for path 'third_party/nanoflann'
PASSED                           [  5%]
tests/compute/test_data.py::test_explain_syn PASSED                      [  5%]
tests/compute/test_data.py::test_wiki_cs Submodule 'third_party/nccl' (https://github.com/nvidia/nccl) registered for path 'third_party/nccl'
Submodule 'third_party/phmap' (https://github.com/greg7mdp/parallel-hashmap.git) registered for path 'third_party/phmap'
Submodule 'third_party/tensorpipe' (https://github.com/pytorch/tensorpipe) registered for path 'third_party/tensorpipe'
Submodule 'third_party/thrust' (https://github.com/NVIDIA/thrust.git) registered for path 'third_party/thrust'
Submodule 'third_party/tvm' (https://github.com/apache/incubator-tvm) registered for path 'third_party/tvm'
Submodule 'third_party/xbyak' (https://github.com/herumi/xbyak) registered for path 'third_party/xbyak'
Cloning into '/root/jenkins/workspace/dgl_PR-4648@4/third_party/METIS'...
PASSED                          [  5%]
tests/compute/test_data.py::test_yelp SKIPPED (Dataset too large to ...) [  5%]
tests/compute/test_data.py::test_flickr PASSED                          [  5%]
tests/compute/test_data.py::test_yelp SKIPPED (Dataset too large to ...) [  5%]
tests/compute/test_data.py::test_flickr Cloning into '/root/jenkins/workspace/dgl_PR-4648@4/third_party/dlpack'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648@4/third_party/dmlc-core'...
collected 3 items

tests/cugraph/test_basics.py::test_dummy PASSED                          [ 33%]
tests/cugraph/test_basics.py::test_to_cugraph_conversion Cloning into '/root/jenkins/workspace/dgl_PR-4648@4/third_party/googletest'...
PASSED                           [  5%]
tests/compute/test_data.py::test_extract_archive PASSED                  [  5%]
tests/compute/test_data.py::test_csvdataset Converting to homogeneous graph takes 0.003s, peak mem: 1.390 GB
Save partitions: 0.005 seconds, peak memory: 1.390 GB
There are 100000 edges in the graph and 0 edge cuts for 1 partitions.
start client[0] for group[0]
PASSED                       [  5%]
tests/compute/test_data.py::test_add_nodepred_split PASSED                           [  5%]
tests/compute/test_data.py::test_extract_archive PASSED                  [  5%]
tests/compute/test_data.py::test_csvdataset Cloning into '/root/jenkins/workspace/dgl_PR-4648@4/third_party/libxsmm'...
[05:29:59] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:29:59] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load check_server_client_False_1_1_1
start server 0
start graph service on server 0 for part 0
[05:29:59] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:29:59] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:10759]...
PASSED                       [  5%]
tests/compute/test_data.py::test_add_nodepred_split PASSED               [  5%]
tests/compute/test_data.py::test_as_nodepred1 PASSED                     [  5%]
tests/compute/test_data.py::test_as_nodepred2 PASSED                     [  5%]
tests/compute/test_data.py::test_as_nodepred_ogb PASSED               [  5%]
tests/compute/test_data.py::test_as_nodepred1 PASSED                     [  5%]
tests/compute/test_data.py::test_as_nodepred2 PASSED                     [  5%]
tests/compute/test_data.py::test_as_nodepred_ogb SKIPPED (ogb only s...) [  5%]
tests/compute/test_data.py::test_as_linkpred PASSED                      [  5%]
tests/compute/test_data.py::test_as_linkpred_ogb SKIPPED (ogb only s...) [  5%]
tests/compute/test_data.py::test_as_nodepred_csvdataset PASSED           [  5%]
tests/compute/test_data.py::test_as_graphpred Client [1247] waits on 172.17.0.3:52321
Machine (0) group (0) client (0) connect to server successfuly!
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
end
Client[0] in group[0] is exiting...
Server (0) shutdown.
Server is exiting...
clients have terminated
PASSED    [  1%]
tests/distributed/test_dist_graph_store.py::test_dist_emb_server_client PASSED          [ 66%]
tests/cugraph/test_basics.py::test_from_cugraph_conversion PASSED        [100%]

=============================== warnings summary ===============================
../../../../opt/conda/lib/python3.9/site-packages/dask_cudf/core.py:33
  /opt/conda/lib/python3.9/site-packages/dask_cudf/core.py:33: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    DASK_VERSION = LooseVersion(dask.__version__)

../../../../opt/conda/lib/python3.9/site-packages/setuptools/_distutils/version.py:346: 34 warnings
  /opt/conda/lib/python3.9/site-packages/setuptools/_distutils/version.py:346: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    other = LooseVersion(other)

python/dgl/backend/backend.py:1717
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/backend/backend.py:1717: DeprecationWarning: invalid escape sequence \P
    """Segment reduction operator.

python/dgl/backend/pytorch/tensor.py:16
python/dgl/backend/pytorch/tensor.py:16
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/backend/pytorch/tensor.py:16: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    if LooseVersion(th.__version__) < LooseVersion("1.9.0"):

python/dgl/backend/pytorch/tensor.py:340
python/dgl/backend/pytorch/tensor.py:340
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/backend/pytorch/tensor.py:340: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    if LooseVersion(th.__version__) >= LooseVersion("1.10.0"):

python/dgl/dataloading/dataloader.py:33
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/dataloading/dataloader.py:33: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    PYTORCH_VER = LooseVersion(torch.__version__)

python/dgl/_dataloading/pytorch/dataloader.py:23
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/_dataloading/pytorch/dataloader.py:23: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    PYTORCH_VER = LooseVersion(th.__version__)

python/dgl/_dataloading/pytorch/dataloader.py:24
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/_dataloading/pytorch/dataloader.py:24: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    PYTORCH_16 = PYTORCH_VER >= LooseVersion("1.6.0")

python/dgl/_dataloading/pytorch/dataloader.py:25
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/_dataloading/pytorch/dataloader.py:25: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    PYTORCH_17 = PYTORCH_VER >= LooseVersion("1.7.0")

-- Docs: https://docs.pytest.org/en/stable/how-to/capture-warnings.html
-- generated xml file: /root/jenkins/workspace/dgl_PR-4648/pytest_cugraph.xml --
============================= slowest 20 durations =============================
12.17s call     tests/cugraph/test_basics.py::test_to_cugraph_conversion
0.05s call     tests/cugraph/test_basics.py::test_from_cugraph_conversion

(7 durations < 0.005s hidden.  Use -vv to show these durations.)
======================= 3 passed, 44 warnings in 19.04s ========================
[Pipeline] }
[Pipeline] // timeout
[Pipeline] }
[Pipeline] // stage
Post stage
[Pipeline] cleanWs
[WS-CLEANUP] Deleting project workspace...
[WS-CLEANUP] Deferred wipeout is disabled by the job configuration...
[WS-CLEANUP] done
[Pipeline] }
$ docker stop --time=1 b7adb4db5f4cf1d886ff23a8dd7a1caea99af2b94c6076d915adcd91d0c69ad9
$ docker rm -f b7adb4db5f4cf1d886ff23a8dd7a1caea99af2b94c6076d915adcd91d0c69ad9
[Pipeline] // withDockerContainer
[Pipeline] }
[Pipeline] // withEnv
[Pipeline] }
Running on dglci-manual-gpu-worker in /root/jenkins/workspace/dgl_PR-4648
[Pipeline] // node
[Pipeline] {
PASSED                  [  5%]
tests/compute/test_data.py::test_as_linkpred PASSED                      [  5%]
tests/compute/test_data.py::test_as_linkpred_ogb [Pipeline] }
[Pipeline] // stage
[Pipeline] }
[Pipeline] checkout
The recommended git tool is: git
using credential 150de63f-189c-4717-bcaf-010460d2f51a
Cloning the remote Git repository
Cloning with configured refspecs honoured and without tags
Converting to homogeneous graph takes 0.003s, peak mem: 1.390 GB
Save partitions: 0.008 seconds, peak memory: 1.390 GB
There are 100000 edges in the graph and 0 edge cuts for 1 partitions.
start client[0] for group[0]
Cloning repository https://github.com/dmlc/dgl.git
 > git init /root/jenkins/workspace/dgl_PR-4648 # timeout=10
Fetching upstream changes from https://github.com/dmlc/dgl.git
 > git --version # timeout=10
 > git --version # 'git version 2.17.1'
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git fetch --no-tags --progress -- https://github.com/dmlc/dgl.git +refs/pull/4648/head:refs/remotes/origin/PR-4648 +refs/heads/master:refs/remotes/origin/master # timeout=10
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load check_dist_emb_True_1_1_1
start server 0
start graph service on server 0 for part 0
[05:30:11] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:30:11] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:18465]...
[05:30:11] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:30:11] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Cleaning workspace
Fetching without tags
Merging remotes/origin/master commit d78a3a4baf611b90871a849f58647160c7cd9ab4 into PR head commit 37ed78035f406940ed88a4b60cadd2b7cfde5fea
 > git config remote.origin.url https://github.com/dmlc/dgl.git # timeout=10
 > git config --add remote.origin.fetch +refs/pull/4648/head:refs/remotes/origin/PR-4648 # timeout=10
 > git config --add remote.origin.fetch +refs/heads/master:refs/remotes/origin/master # timeout=10
 > git config remote.origin.url https://github.com/dmlc/dgl.git # timeout=10
 > git rev-parse --verify HEAD # timeout=10
No valid HEAD. Skipping the resetting
 > git clean -fdx # timeout=10
Fetching upstream changes from https://github.com/dmlc/dgl.git
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git fetch --no-tags --progress -- https://github.com/dmlc/dgl.git +refs/pull/4648/head:refs/remotes/origin/PR-4648 +refs/heads/master:refs/remotes/origin/master # timeout=10
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
Merge succeeded, producing 37ed78035f406940ed88a4b60cadd2b7cfde5fea
Checking out Revision 37ed78035f406940ed88a4b60cadd2b7cfde5fea (PR-4648)
Commit message: "fix for pytorch < 1.12"
Cleaning workspace
[Pipeline] withEnv
[Pipeline] {
[Pipeline] isUnix
[Pipeline] withEnv
[Pipeline] {
[Pipeline] sh
 > git remote # timeout=10
 > git config --get remote.origin.url # timeout=10
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git merge d78a3a4baf611b90871a849f58647160c7cd9ab4 # timeout=10
 > git rev-parse HEAD^{commit} # timeout=10
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git branch -a -v --no-abbrev # timeout=10
 > git checkout -b PR-4648 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git rev-parse --verify HEAD # timeout=10
Resetting working tree
 > git reset --hard # timeout=10
 > git clean -fdx # timeout=10
+ docker pull dgllib/dgl-ci-gpu:cu101_v220816
cu101_v220816: Pulling from dgllib/dgl-ci-gpu
Digest: sha256:ca40fc52876a2563a4e904d0c271d658c1acc8e6a4f8611b578bb49f8c7fd925
Status: Image is up to date for dgllib/dgl-ci-gpu:cu101_v220816
docker.io/dgllib/dgl-ci-gpu:cu101_v220816
[Pipeline] }
[Pipeline] // withEnv
[Pipeline] isUnix
[Pipeline] withEnv
[Pipeline] {
[Pipeline] sh
+ docker inspect -f . dgllib/dgl-ci-gpu:cu101_v220816
.
[Pipeline] }
[Pipeline] // withEnv
[Pipeline] withDockerContainer
dglci-manual-gpu-worker does not seem to be running inside a container
$ docker run -t -d -u 0:0 --runtime nvidia -w /root/jenkins/workspace/dgl_PR-4648 -v /root/jenkins/workspace/dgl_PR-4648:/root/jenkins/workspace/dgl_PR-4648:rw,z -v /root/jenkins/workspace/dgl_PR-4648@tmp:/root/jenkins/workspace/dgl_PR-4648@tmp:rw,z -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** dgllib/dgl-ci-gpu:cu101_v220816 cat
Client [1265] waits on 172.17.0.3:51429
Machine (0) group (0) client (0) connect to server successfuly!
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
Client[0] in group[0] is exiting...
$ docker top 93431508da8c7bd1ffa88cc07c573b36fd6622338dac6808937ce4f0b46901f0 -eo pid,comm
[Pipeline] {
[Pipeline] stage
[Pipeline] { (MXNet GPU Unit test)
[Pipeline] sh
+ nvidia-smi
Tue Sep 27 05:30:16 2022       
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 510.47.03    Driver Version: 510.47.03    CUDA Version: 11.6     |
|-------------------------------+----------------------+----------------------+
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|                               |                      |               MIG M. |
|===============================+======================+======================|
|   0  Tesla T4            On   | 00000000:00:1E.0 Off |                    0 |
| N/A   42C    P8    16W /  70W |      0MiB / 15360MiB |      0%      Default |
|                               |                      |                  N/A |
+-------------------------------+----------------------+----------------------+
                                                                               
+-----------------------------------------------------------------------------+
| Processes:                                                                  |
|  GPU   GI   CI        PID   Type   Process name                  GPU Memory |
|        ID   ID                                                   Usage      |
|=============================================================================|
|  No running processes found                                                 |
+-----------------------------------------------------------------------------+
[Pipeline] sh
Server (0) shutdown.
Server is exiting...
+ rm -rf CMakeLists.txt CONTRIBUTORS.md Jenkinsfile LICENSE NEWS.md README.md apps benchmarks cmake conda dglgo docker docs examples featgraph include pyproject.toml python readthedocs.yml src tensoradapter tests third_party tools tutorials
clients have terminated
[Pipeline] checkout
The recommended git tool is: git
using credential 150de63f-189c-4717-bcaf-010460d2f51a
Warning: JENKINS-30600: special launcher org.jenkinsci.plugins.docker.workflow.WithContainerStep$Decorator$1@276ec34c; decorates RemoteLauncher[hudson.remoting.Channel@41a64267:dglci-manual-gpu-worker] will be ignored (a typical symptom is the Git executable not being run inside a designated container)
Fetching changes from the remote Git repository
Cleaning workspace
Fetching without tags
Merging remotes/origin/master commit d78a3a4baf611b90871a849f58647160c7cd9ab4 into PR head commit 37ed78035f406940ed88a4b60cadd2b7cfde5fea
Merge succeeded, producing 37ed78035f406940ed88a4b60cadd2b7cfde5fea
Checking out Revision 37ed78035f406940ed88a4b60cadd2b7cfde5fea (PR-4648)
Commit message: "fix for pytorch < 1.12"
Cleaning workspace
 > git rev-parse --resolve-git-dir /root/jenkins/workspace/dgl_PR-4648/.git # timeout=10
 > git config remote.origin.url https://github.com/dmlc/dgl.git # timeout=10
 > git rev-parse --verify HEAD # timeout=10
Resetting working tree
 > git reset --hard # timeout=10
 > git clean -fdx # timeout=10
Fetching upstream changes from https://github.com/dmlc/dgl.git
 > git --version # timeout=10
 > git --version # 'git version 2.17.1'
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git fetch --no-tags --progress -- https://github.com/dmlc/dgl.git +refs/pull/4648/head:refs/remotes/origin/PR-4648 +refs/heads/master:refs/remotes/origin/master # timeout=10
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git remote # timeout=10
 > git config --get remote.origin.url # timeout=10
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git merge d78a3a4baf611b90871a849f58647160c7cd9ab4 # timeout=10
 > git rev-parse HEAD^{commit} # timeout=10
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git branch -a -v --no-abbrev # timeout=10
 > git branch -D PR-4648 # timeout=10
 > git checkout -b PR-4648 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git rev-parse --verify HEAD # timeout=10
Resetting working tree
 > git reset --hard # timeout=10
 > git clean -fdx # timeout=10
[Pipeline] sh
+ git submodule update --recursive --init
Submodule 'third_party/METIS' (https://github.com/KarypisLab/METIS.git) registered for path 'third_party/METIS'
Submodule 'third_party/dlpack' (https://github.com/dmlc/dlpack.git) registered for path 'third_party/dlpack'
Submodule 'third_party/dmlc-core' (https://github.com/dmlc/dmlc-core.git) registered for path 'third_party/dmlc-core'
Submodule 'third_party/googletest' (https://github.com/google/googletest.git) registered for path 'third_party/googletest'
Submodule 'third_party/libxsmm' (https://github.com/hfp/libxsmm.git) registered for path 'third_party/libxsmm'
Submodule 'third_party/nanoflann' (https://github.com/jlblancoc/nanoflann) registered for path 'third_party/nanoflann'
Submodule 'third_party/nccl' (https://github.com/nvidia/nccl) registered for path 'third_party/nccl'
Submodule 'third_party/phmap' (https://github.com/greg7mdp/parallel-hashmap.git) registered for path 'third_party/phmap'
Submodule 'third_party/tensorpipe' (https://github.com/pytorch/tensorpipe) registered for path 'third_party/tensorpipe'
Submodule 'third_party/thrust' (https://github.com/NVIDIA/thrust.git) registered for path 'third_party/thrust'
Submodule 'third_party/tvm' (https://github.com/apache/incubator-tvm) registered for path 'third_party/tvm'
Submodule 'third_party/xbyak' (https://github.com/herumi/xbyak) registered for path 'third_party/xbyak'
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/METIS'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/dlpack'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/dmlc-core'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/googletest'...
Converting to homogeneous graph takes 0.003s, peak mem: 1.390 GB
Save partitions: 0.008 seconds, peak memory: 1.390 GB
There are 100000 edges in the graph and 0 edge cuts for 1 partitions.
start client[0] for group[0]
Cloning into '/root/jenkins/workspace/dgl_PR-4648@4/third_party/nanoflann'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/libxsmm'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648@4/third_party/nccl'...
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load check_dist_emb_False_1_1_1
start server 0
start graph service on server 0 for part 0
[05:30:21] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:30:21] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:11787]...
[05:30:21] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:30:21] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Cloning into '/root/jenkins/workspace/dgl_PR-4648@4/third_party/phmap'...
PASSED                  [  5%]
tests/compute/test_data.py::test_as_nodepred_csvdataset PASSED           [  5%]
tests/compute/test_data.py::test_as_graphpred Cloning into '/root/jenkins/workspace/dgl_PR-4648@4/third_party/tensorpipe'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648@4/third_party/thrust'...
Client [1284] waits on 172.17.0.3:34707
Machine (0) group (0) client (0) connect to server successfuly!
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
Client[0] in group[0] is exiting...
Cloning into '/root/jenkins/workspace/dgl_PR-4648@4/third_party/tvm'...
Server (0) shutdown.
Server is exiting...
clients have terminated
PASSED [  3%]
tests/distributed/test_dist_graph_store.py::test_standalone Converting to homogeneous graph takes 0.003s, peak mem: 1.390 GB
Save partitions: 0.005 seconds, peak memory: 1.390 GB
There are 100000 edges in the graph and 0 edge cuts for 1 partitions.
end
Client[-1] in group[-1] is exiting...
PASSED       [  5%]
tests/distributed/test_dist_graph_store.py::test_standalone_node_emb Cloning into '/root/jenkins/workspace/dgl_PR-4648@4/third_party/xbyak'...
Submodule path 'third_party/METIS': checked out '10603482cc055626ff99d2ac3ab7e73d1119439d'
Submodule 'GKlib' (https://github.com/KarypisLab/GKlib.git) registered for path 'third_party/METIS/GKlib'
Cloning into '/root/jenkins/workspace/dgl_PR-4648@4/third_party/METIS/GKlib'...
Submodule path 'third_party/METIS/GKlib': checked out '62de20c7f97c719abdc96d7c3e266a1f7cb52bc5'
Submodule path 'third_party/dlpack': checked out 'e2bdd3bee8cb6501558042633fa59144cc8b7f5f'
Submodule path 'third_party/dmlc-core': checked out 'bfad207b448480783a1f428ae3d93d87032d8349'
Submodule path 'third_party/googletest': checked out 'f71fb4f9a912ec945401cc49a287a759b6131026'
Submodule path 'third_party/libxsmm': checked out 'fa687556130b6298430f1c0555a14cf79ab6101c'
Submodule path 'third_party/nanoflann': checked out '4c47ca200209550c5628c89803591f8a753c8181'
Submodule path 'third_party/nccl': checked out 'e11238b3029795d33f958b5868d47c90c4f22628'
Submodule path 'third_party/phmap': checked out '25293cefd8b85491b45600c03fe8edf07647553f'
Submodule path 'third_party/tensorpipe': checked out '6042f1a4cbce8eef997f11ed0012de137b317361'
Submodule 'third_party/googletest' (https://github.com/google/googletest.git) registered for path 'third_party/tensorpipe/third_party/googletest'
Submodule 'third_party/libnop' (https://github.com/google/libnop.git) registered for path 'third_party/tensorpipe/third_party/libnop'
Submodule 'third_party/libuv' (https://github.com/libuv/libuv.git) registered for path 'third_party/tensorpipe/third_party/libuv'
Submodule 'third_party/pybind11' (https://github.com/pybind/pybind11.git) registered for path 'third_party/tensorpipe/third_party/pybind11'
Cloning into '/root/jenkins/workspace/dgl_PR-4648@4/third_party/tensorpipe/third_party/googletest'...
Converting to homogeneous graph takes 0.003s, peak mem: 1.390 GB
Save partitions: 0.005 seconds, peak memory: 1.390 GB
There are 100000 edges in the graph and 0 edge cuts for 1 partitions.
Client[-1] in group[-1] is exiting...
PASSED [  7%]
tests/distributed/test_dist_graph_store.py::test_split[True] Cloning into '/root/jenkins/workspace/dgl_PR-4648@4/third_party/tensorpipe/third_party/libnop'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648@4/third_party/tensorpipe/third_party/libuv'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648@4/third_party/tensorpipe/third_party/pybind11'...
Submodule path 'third_party/tensorpipe/third_party/googletest': checked out 'aee0f9d9b5b87796ee8a0ab26b7587ec30e8858e'
Submodule path 'third_party/tensorpipe/third_party/libnop': checked out 'aa95422ea8c409e3f078d2ee7708a5f59a8b9fa2'
Submodule path 'third_party/tensorpipe/third_party/libuv': checked out '1dff88e5161cba5c59276d2070d2e304e4dcb242'
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/nanoflann'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/nccl'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/phmap'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tensorpipe'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/thrust'...
Submodule path 'third_party/tensorpipe/third_party/pybind11': checked out 'a23996fce38ff6ccfbcdc09f1e63f2c4be5ea2ef'
Submodule 'tools/clang' (https://github.com/wjakob/clang-cindex-python3) registered for path 'third_party/tensorpipe/third_party/pybind11/tools/clang'
Cloning into '/root/jenkins/workspace/dgl_PR-4648@4/third_party/tensorpipe/third_party/pybind11/tools/clang'...
Submodule path 'third_party/tensorpipe/third_party/pybind11/tools/clang': checked out '6a00cbc4a9b8e68b71caf7f774b3f9c753ae84d5'
Submodule path 'third_party/thrust': checked out '6a3078c64cab0e2f276340fa5dcafa0d758ed890'
Submodule 'cub' (https://github.com/NVIDIA/cub.git) registered for path 'third_party/thrust/dependencies/cub'
Cloning into '/root/jenkins/workspace/dgl_PR-4648@4/third_party/thrust/dependencies/cub'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tvm'...
Submodule path 'third_party/thrust/dependencies/cub': checked out 'cdaa9558a85e45d849016e5fe7b6e4ee79113f95'
Submodule path 'third_party/tvm': checked out 'b2e418cb109df4cd1f17a2cf2894a1b396a6b838'
Submodule 'dlpack' (https://github.com/dmlc/dlpack) registered for path 'third_party/tvm/3rdparty/dlpack'
Submodule 'dmlc-core' (https://github.com/dmlc/dmlc-core) registered for path 'third_party/tvm/3rdparty/dmlc-core'
Submodule '3rdparty/rang' (https://github.com/agauniyal/rang) registered for path 'third_party/tvm/3rdparty/rang'
Submodule '3rdparty/vta-hw' (https://github.com/apache/incubator-tvm-vta) registered for path 'third_party/tvm/3rdparty/vta-hw'
Cloning into '/root/jenkins/workspace/dgl_PR-4648@4/third_party/tvm/3rdparty/dlpack'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648@4/third_party/tvm/3rdparty/dmlc-core'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648@4/third_party/tvm/3rdparty/rang'...
PASSED                     [  5%]
tests/compute/test_data.py::test_as_graphpred_reprocess Cloning into '/root/jenkins/workspace/dgl_PR-4648@4/third_party/tvm/3rdparty/vta-hw'...
Converting to homogeneous graph takes 0.008s, peak mem: 1.451 GB
Convert a graph into a bidirected graph: 0.016 seconds, peak memory: 1.451 GB
Construct multi-constraint weights: 0.001 seconds, peak memory: 1.451 GB
[05:30:43] /root/jenkins/workspace/dgl_PR-4648/src/graph/transform/metis_partition_hetero.cc:87: Partition a graph with 30030 nodes and 601200 edges into 4 parts and get 175264 edge cuts
Metis partitioning: 0.187 seconds, peak memory: 1.451 GB
Assigning nodes to METIS partitions takes 0.205s, peak mem: 1.451 GB
Reshuffle nodes and edges: 0.020 seconds
Split the graph: 0.069 seconds
Construct subgraphs: 0.077 seconds
Splitting the graph into partitions takes 0.166s, peak mem: 1.451 GB
part 0 has 10000 nodes of type n1 and 2574 are inside the partition
part 0 has 9873 nodes of type n2 and 2576 are inside the partition
part 0 has 9901 nodes of type n3 and 2580 are inside the partition
part 0 has 92426 edges of type r1 and 25798 are inside the partition
part 0 has 40959 edges of type r2 and 25834 are inside the partition
part 0 has 40713 edges of type r3 and 25619 are inside the partition
part 1 has 10000 nodes of type n1 and 2574 are inside the partition
part 1 has 9904 nodes of type n2 and 2576 are inside the partition
part 1 has 9906 nodes of type n3 and 2580 are inside the partition
part 1 has 92768 edges of type r1 and 25800 are inside the partition
part 1 has 40705 edges of type r2 and 25798 are inside the partition
part 1 has 41105 edges of type r3 and 25978 are inside the partition
part 2 has 9999 nodes of type n1 and 2281 are inside the partition
part 2 has 9794 nodes of type n2 and 2281 are inside the partition
part 2 has 9801 nodes of type n3 and 2280 are inside the partition
part 2 has 89496 edges of type r1 and 22774 are inside the partition
part 2 has 36244 edges of type r2 and 22739 are inside the partition
part 2 has 36291 edges of type r3 and 22826 are inside the partition
part 3 has 10000 nodes of type n1 and 2571 are inside the partition
part 3 has 9883 nodes of type n2 and 2577 are inside the partition
part 3 has 9906 nodes of type n3 and 2580 are inside the partition
part 3 has 92623 edges of type r1 and 25728 are inside the partition
part 3 has 40948 edges of type r2 and 25829 are inside the partition
part 3 has 40613 edges of type r3 and 25877 are inside the partition
Save partitions: 0.043 seconds, peak memory: 1.451 GB
There are 300600 edges in the graph and 0 edge cuts for 4 partitions.
Submodule path 'third_party/tvm/3rdparty/dlpack': checked out '3ec04430e89a6834e5a1b99471f415fa939bf642'
Submodule path 'third_party/tvm/3rdparty/dmlc-core': checked out '6c401e242c59a1f4c913918246591bb13fd714e7'
Submodule path 'third_party/tvm/3rdparty/rang': checked out 'cabe04d6d6b05356fa8f9741704924788f0dd762'
Submodule path 'third_party/tvm/3rdparty/vta-hw': checked out '87ce9acfae550d1a487746e9d06c2e250076e54c'
Submodule path 'third_party/xbyak': checked out '757e4063f6464740b8ff4a2cae9136d2f8458020'
[Pipeline] unstash
[Pipeline] echo
Unpacked build/libdgl.so, build/runUnitTests, python/dgl/_ffi/_cy3/core.cpython-*-x86_64-linux-gnu.so, build/tensoradapter/pytorch/*.so from dgl-cpu-linux
[Pipeline] timeout
Timeout set to expire in 30 min
[Pipeline] {
PASSED      [  9%]
tests/distributed/test_dist_graph_store.py::test_split[False] [Pipeline] sh
+ bash tests/scripts/task_unit_test.sh mxnet cpu
Requirement already satisfied: pytest in /opt/conda/envs/mxnet-ci/lib/python3.6/site-packages (7.0.1)
Collecting psutil
  Downloading psutil-5.9.2-cp36-cp36m-manylinux_2_12_x86_64.manylinux2010_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl (280 kB)
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/xbyak'...
Collecting pyyaml
  Downloading PyYAML-6.0-cp36-cp36m-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_12_x86_64.manylinux2010_x86_64.whl (603 kB)
Collecting pydantic
  Downloading pydantic-1.9.2-cp36-cp36m-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (11.2 MB)
Submodule path 'third_party/METIS': checked out '10603482cc055626ff99d2ac3ab7e73d1119439d'
Submodule 'GKlib' (https://github.com/KarypisLab/GKlib.git) registered for path 'third_party/METIS/GKlib'
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/METIS/GKlib'...
Collecting pandas
  Using cached pandas-1.1.5-cp36-cp36m-manylinux1_x86_64.whl (9.5 MB)
Collecting rdflib
Submodule path 'third_party/METIS/GKlib': checked out '62de20c7f97c719abdc96d7c3e266a1f7cb52bc5'
Submodule path 'third_party/dlpack': checked out 'e2bdd3bee8cb6501558042633fa59144cc8b7f5f'
  Downloading rdflib-5.0.0-py3-none-any.whl (231 kB)
Collecting ogb
  Downloading ogb-1.3.4-py3-none-any.whl (78 kB)
Requirement already satisfied: py>=1.8.2 in /opt/conda/envs/mxnet-ci/lib/python3.6/site-packages (from pytest) (1.11.0)
Requirement already satisfied: iniconfig in /opt/conda/envs/mxnet-ci/lib/python3.6/site-packages (from pytest) (1.1.1)
Requirement already satisfied: importlib-metadata>=0.12 in /opt/conda/envs/mxnet-ci/lib/python3.6/site-packages (from pytest) (4.8.3)
Requirement already satisfied: pluggy<2.0,>=0.12 in /opt/conda/envs/mxnet-ci/lib/python3.6/site-packages (from pytest) (1.0.0)
Requirement already satisfied: packaging in /opt/conda/envs/mxnet-ci/lib/python3.6/site-packages (from pytest) (21.3)
Requirement already satisfied: attrs>=19.2.0 in /opt/conda/envs/mxnet-ci/lib/python3.6/site-packages (from pytest) (21.4.0)
Requirement already satisfied: tomli>=1.0.0 in /opt/conda/envs/mxnet-ci/lib/python3.6/site-packages (from pytest) (1.2.3)
Requirement already satisfied: typing-extensions>=3.7.4.3 in /opt/conda/envs/mxnet-ci/lib/python3.6/site-packages (from pydantic) (4.1.1)
Submodule path 'third_party/dmlc-core': checked out 'bfad207b448480783a1f428ae3d93d87032d8349'
Submodule path 'third_party/googletest': checked out 'f71fb4f9a912ec945401cc49a287a759b6131026'
Collecting dataclasses>=0.6
  Using cached dataclasses-0.8-py3-none-any.whl (19 kB)
Collecting pytz>=2017.2
  Downloading pytz-2022.2.1-py2.py3-none-any.whl (500 kB)
Converting to homogeneous graph takes 0.002s, peak mem: 1.529 GB
Convert a graph into a bidirected graph: 0.004 seconds, peak memory: 1.529 GB
Construct multi-constraint weights: 0.000 seconds, peak memory: 1.529 GB
[05:30:47] /root/jenkins/workspace/dgl_PR-4648/src/graph/transform/metis_partition_hetero.cc:87: Partition a graph with 10000 nodes and 199872 edges into 4 parts and get 56578 edge cuts
Metis partitioning: 0.046 seconds, peak memory: 1.529 GB
Assigning nodes to METIS partitions takes 0.051s, peak mem: 1.529 GB
Reshuffle nodes and edges: 0.005 seconds
Split the graph: 0.024 seconds
Construct subgraphs: 0.034 seconds
Splitting the graph into partitions takes 0.063s, peak mem: 1.529 GB
part 0 has 10000 nodes and 2500 are inside the partition
part 0 has 91023 edges and 24712 are inside the partition
part 1 has 10000 nodes and 2500 are inside the partition
part 1 has 91347 edges and 25164 are inside the partition
part 2 has 10000 nodes and 2500 are inside the partition
part 2 has 91578 edges and 25151 are inside the partition
part 3 has 10000 nodes and 2500 are inside the partition
part 3 has 91266 edges and 24973 are inside the partition
Save partitions: 0.015 seconds, peak memory: 1.529 GB
There are 100000 edges in the graph and 0 edge cuts for 4 partitions.
Requirement already satisfied: python-dateutil>=2.7.3 in /opt/conda/envs/mxnet-ci/lib/python3.6/site-packages (from pandas) (2.8.2)
Requirement already satisfied: numpy>=1.15.4 in /opt/conda/envs/mxnet-ci/lib/python3.6/site-packages (from pandas) (1.19.5)
Requirement already satisfied: pyparsing in /opt/conda/envs/mxnet-ci/lib/python3.6/site-packages (from rdflib) (3.0.9)
Requirement already satisfied: six in /opt/conda/envs/mxnet-ci/lib/python3.6/site-packages (from rdflib) (1.16.0)
Collecting isodate
  Downloading isodate-0.6.1-py2.py3-none-any.whl (41 kB)
Requirement already satisfied: urllib3>=1.24.0 in /opt/conda/envs/mxnet-ci/lib/python3.6/site-packages (from ogb) (1.26.9)
Requirement already satisfied: tqdm>=4.29.0 in /opt/conda/envs/mxnet-ci/lib/python3.6/site-packages (from ogb) (4.64.0)
Collecting torch>=1.6.0
  Downloading torch-1.10.2-cp36-cp36m-manylinux1_x86_64.whl (881.9 MB)
PASSED     [ 10%]
tests/distributed/test_dist_graph_store.py::test_split_even Submodule path 'third_party/libxsmm': checked out 'fa687556130b6298430f1c0555a14cf79ab6101c'
Submodule path 'third_party/nanoflann': checked out '4c47ca200209550c5628c89803591f8a753c8181'
Submodule path 'third_party/nccl': checked out 'e11238b3029795d33f958b5868d47c90c4f22628'
Submodule path 'third_party/phmap': checked out '25293cefd8b85491b45600c03fe8edf07647553f'
Submodule path 'third_party/tensorpipe': checked out '6042f1a4cbce8eef997f11ed0012de137b317361'
Submodule 'third_party/googletest' (https://github.com/google/googletest.git) registered for path 'third_party/tensorpipe/third_party/googletest'
Submodule 'third_party/libnop' (https://github.com/google/libnop.git) registered for path 'third_party/tensorpipe/third_party/libnop'
Submodule 'third_party/libuv' (https://github.com/libuv/libuv.git) registered for path 'third_party/tensorpipe/third_party/libuv'
Submodule 'third_party/pybind11' (https://github.com/pybind/pybind11.git) registered for path 'third_party/tensorpipe/third_party/pybind11'
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tensorpipe/third_party/googletest'...
PASSED                     [  5%]
tests/compute/test_data.py::test_as_graphpred_reprocess PASSED           [  5%]
tests/compute/test_data.py::test_as_graphpred_ogb SKIPPED (ogb only ...) [  5%]
tests/compute/test_dataloader.py::test_edge_prediction_sampler[idtype0] PASSED [  5%]
tests/compute/test_dataloader.py::test_edge_prediction_sampler[idtype1] PASSED [  5%]
tests/compute/test_edge_softmax_hetero.py::test_edge_softmax_unidirectional SKIPPED [  6%]
tests/compute/test_edge_softmax_hetero.py::test_edge_softmax[idtype0-src-g0] SKIPPED [  6%]
tests/compute/test_edge_softmax_hetero.py::test_edge_softmax[idtype0-dst-g0] SKIPPED [  6%]
tests/compute/test_edge_softmax_hetero.py::test_edge_softmax[idtype1-src-g0] SKIPPED [  6%]
tests/compute/test_edge_softmax_hetero.py::test_edge_softmax[idtype1-dst-g0] SKIPPED [  6%]
tests/compute/test_ffi.py::test_cython PASSED                            [  6%]
tests/compute/test_ffi.py::test_callback[1] PASSED                       [  6%]
tests/compute/test_ffi.py::test_callback[2.3] PASSED                     [  6%]
tests/compute/test_ffi.py::test_callback_thread[1] PASSED                [  6%]
tests/compute/test_ffi.py::test_callback_thread[2.3] PASSED              [  6%]
tests/compute/test_filter.py::test_graph_filter PASSED                   [  6%]
tests/compute/test_filter.py::test_array_filter[idtype0] SKIPPED (CP...) [  6%]
tests/compute/test_filter.py::test_array_filter[idtype1] SKIPPED (CP...) [  6%]
tests/compute/test_frame.py::test_column_subcolumn PASSED                [  6%]
tests/compute/test_frame.py::test_serialize_deserialize_plain PASSED     [  6%]
tests/compute/test_frame.py::test_serialize_deserialize_subcolumn PASSED [  6%]
tests/compute/test_frame.py::test_serialize_deserialize_dtype PASSED     [  6%]
tests/compute/test_generators.py::test_rand_graph PASSED                 [  6%]
tests/compute/test_graph.py::test_query Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tensorpipe/third_party/libnop'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tensorpipe/third_party/libuv'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tensorpipe/third_party/pybind11'...
Submodule path 'third_party/tensorpipe/third_party/googletest': checked out 'aee0f9d9b5b87796ee8a0ab26b7587ec30e8858e'
Submodule path 'third_party/tensorpipe/third_party/libnop': checked out 'aa95422ea8c409e3f078d2ee7708a5f59a8b9fa2'
Submodule path 'third_party/tensorpipe/third_party/libuv': checked out '1dff88e5161cba5c59276d2070d2e304e4dcb242'
Submodule path 'third_party/tensorpipe/third_party/pybind11': checked out 'a23996fce38ff6ccfbcdc09f1e63f2c4be5ea2ef'
Submodule 'tools/clang' (https://github.com/wjakob/clang-cindex-python3) registered for path 'third_party/tensorpipe/third_party/pybind11/tools/clang'
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tensorpipe/third_party/pybind11/tools/clang'...
Converting to homogeneous graph takes 0.003s, peak mem: 1.529 GB
Convert a graph into a bidirected graph: 0.005 seconds, peak memory: 1.529 GB
Construct multi-constraint weights: 0.000 seconds, peak memory: 1.529 GB
[05:30:53] /root/jenkins/workspace/dgl_PR-4648/src/graph/transform/metis_partition_hetero.cc:87: Partition a graph with 10000 nodes and 199872 edges into 4 parts and get 56461 edge cuts
Metis partitioning: 0.049 seconds, peak memory: 1.529 GB
Assigning nodes to METIS partitions takes 0.054s, peak mem: 1.529 GB
Reshuffle nodes and edges: 0.005 seconds
Split the graph: 0.028 seconds
Construct subgraphs: 0.038 seconds
Splitting the graph into partitions takes 0.072s, peak mem: 1.529 GB
part 0 has 10000 nodes and 2501 are inside the partition
part 0 has 91468 edges and 25337 are inside the partition
part 1 has 10000 nodes and 2499 are inside the partition
part 1 has 91662 edges and 24912 are inside the partition
part 2 has 9999 nodes and 2500 are inside the partition
part 2 has 91219 edges and 24847 are inside the partition
part 3 has 9999 nodes and 2500 are inside the partition
part 3 has 90856 edges and 24904 are inside the partition
Save partitions: 0.023 seconds, peak memory: 1.529 GB
There are 100000 edges in the graph and 0 edge cuts for 4 partitions.
part 0 get 1706 nodes and 1706 are in the partition
intersection has 1706
part 0 get 17268 edges and 17268 are in the partition
intersection has 17268
part 1 get 1706 nodes and 1689 are in the partition
intersection has 1706
part 1 get 17267 edges and 16989 are in the partition
intersection has 17267
part 2 get 1705 nodes and 1705 are in the partition
intersection has 1705
part 2 get 17267 edges and 16964 are in the partition
intersection has 17267
part 3 get 1705 nodes and 1703 are in the partition
intersection has 1705
part 3 get 17267 edges and 17070 are in the partition
intersection has 17267
PASSED       [ 12%]
tests/distributed/test_distributed_sampling.py::test_rpc_find_edges_shuffle[1] Converting to homogeneous graph takes 0.001s, peak mem: 1.529 GB
Save partitions: 0.002 seconds, peak memory: 1.529 GB
There are 3060 edges in the graph and 0 edge cuts for 1 partitions.
Submodule path 'third_party/tensorpipe/third_party/pybind11/tools/clang': checked out '6a00cbc4a9b8e68b71caf7f774b3f9c753ae84d5'
Submodule path 'third_party/thrust': checked out '6a3078c64cab0e2f276340fa5dcafa0d758ed890'
Submodule 'cub' (https://github.com/NVIDIA/cub.git) registered for path 'third_party/thrust/dependencies/cub'
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/thrust/dependencies/cub'...
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_find_edges
start graph service on server 0 for part 0
[05:30:54] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:30:54] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:19883]...
[05:30:54] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:30:54] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
PASSED                           [  6%]
tests/compute/test_graph.py::test_mutation PASSED                        [  6%]
tests/compute/test_graph.py::test_scipy_adjmat PASSED                    [  6%]
tests/compute/test_graph.py::test_incmat PASSED                          [  6%]
tests/compute/test_graph.py::test_find_edges PASSED                      [  6%]
tests/compute/test_graph.py::test_ismultigraph PASSED                    [  6%]
tests/compute/test_graph.py::test_hypersparse_query PASSED               [  6%]
tests/compute/test_graph.py::test_empty_data_initialized PASSED          [  7%]
tests/compute/test_graph.py::test_is_sorted PASSED                       [  7%]
tests/compute/test_graph.py::test_default_types PASSED                   [  7%]
tests/compute/test_graph.py::test_formats PASSED                         [  7%]
tests/compute/test_heterograph.py::test_create[idtype0] PASSED           [  7%]
tests/compute/test_heterograph.py::test_create[idtype1] PASSED           [  7%]
tests/compute/test_heterograph.py::test_create2 PASSED                   [  7%]
tests/compute/test_heterograph.py::test_query[idtype0] Submodule path 'third_party/thrust/dependencies/cub': checked out 'cdaa9558a85e45d849016e5fe7b6e4ee79113f95'
Submodule path 'third_party/tvm': checked out 'b2e418cb109df4cd1f17a2cf2894a1b396a6b838'
Submodule 'dlpack' (https://github.com/dmlc/dlpack) registered for path 'third_party/tvm/3rdparty/dlpack'
Submodule 'dmlc-core' (https://github.com/dmlc/dmlc-core) registered for path 'third_party/tvm/3rdparty/dmlc-core'
Submodule '3rdparty/rang' (https://github.com/agauniyal/rang) registered for path 'third_party/tvm/3rdparty/rang'
Submodule '3rdparty/vta-hw' (https://github.com/apache/incubator-tvm-vta) registered for path 'third_party/tvm/3rdparty/vta-hw'
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tvm/3rdparty/dlpack'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tvm/3rdparty/dmlc-core'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tvm/3rdparty/rang'...
PASSED           [  5%]
tests/compute/test_data.py::test_as_graphpred_ogb Cloning into '/root/jenkins/workspace/dgl_PR-4648/third_party/tvm/3rdparty/vta-hw'...
Collecting scikit-learn>=0.20.0
  Using cached scikit_learn-0.24.2-cp36-cp36m-manylinux2010_x86_64.whl (22.2 MB)
Collecting outdated>=0.2.0
  Using cached outdated-0.2.1-py3-none-any.whl (7.5 kB)
Requirement already satisfied: zipp>=0.5 in /opt/conda/envs/mxnet-ci/lib/python3.6/site-packages (from importlib-metadata>=0.12->pytest) (3.6.0)
Collecting littleutils
  Using cached littleutils-0.2.2-py3-none-any.whl
Requirement already satisfied: requests in /opt/conda/envs/mxnet-ci/lib/python3.6/site-packages (from outdated>=0.2.0->ogb) (2.27.1)
Collecting threadpoolctl>=2.0.0
  Using cached threadpoolctl-3.1.0-py3-none-any.whl (14 kB)
Requirement already satisfied: scipy>=0.19.1 in /opt/conda/envs/mxnet-ci/lib/python3.6/site-packages (from scikit-learn>=0.20.0->ogb) (1.5.4)
Requirement already satisfied: joblib>=0.11 in /opt/conda/envs/mxnet-ci/lib/python3.6/site-packages (from scikit-learn>=0.20.0->ogb) (1.1.0)
Requirement already satisfied: importlib-resources in /opt/conda/envs/mxnet-ci/lib/python3.6/site-packages (from tqdm>=4.29.0->ogb) (5.4.0)
Requirement already satisfied: idna<4,>=2.5 in /opt/conda/envs/mxnet-ci/lib/python3.6/site-packages (from requests->outdated>=0.2.0->ogb) (3.3)
Requirement already satisfied: charset-normalizer~=2.0.0 in /opt/conda/envs/mxnet-ci/lib/python3.6/site-packages (from requests->outdated>=0.2.0->ogb) (2.0.12)
Requirement already satisfied: certifi>=2017.4.17 in /opt/conda/envs/mxnet-ci/lib/python3.6/site-packages (from requests->outdated>=0.2.0->ogb) (2021.5.30)
Client [1090] waits on 172.17.0.3:45449
Machine (0) group (0) client (0) connect to server successfuly!
Client[0] in group[0] is exiting...
Server (0) shutdown.
Server is exiting...
Downloading /root/jenkins/workspace/dgl_PR-4648/cora_v2.zip from https://data.dgl.ai/dataset/cora_v2.zip...
Extracting file to /root/jenkins/workspace/dgl_PR-4648/cora_v2
Submodule path 'third_party/tvm/3rdparty/dlpack': checked out '3ec04430e89a6834e5a1b99471f415fa939bf642'
Submodule path 'third_party/tvm/3rdparty/dmlc-core': checked out '6c401e242c59a1f4c913918246591bb13fd714e7'
Submodule path 'third_party/tvm/3rdparty/rang': checked out 'cabe04d6d6b05356fa8f9741704924788f0dd762'
Submodule path 'third_party/tvm/3rdparty/vta-hw': checked out '87ce9acfae550d1a487746e9d06c2e250076e54c'
Submodule path 'third_party/xbyak': checked out '757e4063f6464740b8ff4a2cae9136d2f8458020'
[Pipeline] unstash
Installing collected packages: threadpoolctl, pytz, littleutils, dataclasses, torch, scikit-learn, pandas, outdated, isodate, rdflib, pyyaml, pydantic, psutil, ogb
Finished data loading and preprocessing.
  NumNodes: 2708
  NumEdges: 10556
  NumFeats: 1433
  NumClasses: 7
  NumTrainingSamples: 140
  NumValidationSamples: 500
  NumTestSamples: 1000
Done saving data into cached files.
Converting to homogeneous graph takes 0.001s, peak mem: 1.529 GB
Save partitions: 0.029 seconds, peak memory: 1.529 GB
There are 10556 edges in the graph and 0 edge cuts for 1 partitions.
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_find_edges
start graph service on server 0 for part 0
[05:30:58] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:30:58] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:13648]...
[05:30:59] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:30:59] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
PASSED            [  7%]
tests/compute/test_heterograph.py::test_query[idtype1] Client [1090] waits on 172.17.0.3:60937
Machine (0) group (0) client (0) connect to server successfuly!
Client[0] in group[0] is exiting...
Server (0) shutdown.
Server is exiting...
PASSED [ 14%]
tests/distributed/test_distributed_sampling.py::test_rpc_find_edges_shuffle[2] Converting to homogeneous graph takes 0.001s, peak mem: 1.529 GB
Convert a graph into a bidirected graph: 0.000 seconds, peak memory: 1.529 GB
Construct multi-constraint weights: 0.000 seconds, peak memory: 1.529 GB
[05:31:02] /root/jenkins/workspace/dgl_PR-4648/src/graph/transform/metis_partition_hetero.cc:87: Partition a graph with 3030 nodes and 6120 edges into 2 parts and get 204 edge cuts
Metis partitioning: 0.003 seconds, peak memory: 1.529 GB
Assigning nodes to METIS partitions takes 0.004s, peak mem: 1.529 GB
Reshuffle nodes and edges: 0.001 seconds
Split the graph: 0.002 seconds
Construct subgraphs: 0.002 seconds
Splitting the graph into partitions takes 0.005s, peak mem: 1.529 GB
part 0 has 587 nodes of type n1 and 519 are inside the partition
part 0 has 574 nodes of type n2 and 511 are inside the partition
part 0 has 585 nodes of type n3 and 525 are inside the partition
part 0 has 565 edges of type r12 and 531 are inside the partition
part 0 has 571 edges of type r13 and 535 are inside the partition
part 0 has 554 edges of type r23 and 527 are inside the partition
part 1 has 556 nodes of type n1 and 491 are inside the partition
part 1 has 549 nodes of type n2 and 489 are inside the partition
part 1 has 559 nodes of type n3 and 495 are inside the partition
part 1 has 515 edges of type r12 and 479 are inside the partition
part 1 has 530 edges of type r13 and 495 are inside the partition
part 1 has 529 edges of type r23 and 493 are inside the partition
Save partitions: 0.003 seconds, peak memory: 1.529 GB
There are 3060 edges in the graph and 0 edge cuts for 2 partitions.
[Pipeline] echo
Unpacked build/libdgl.so, build/runUnitTests, python/dgl/_ffi/_cy3/core.cpython-*-x86_64-linux-gnu.so, build/tensoradapter/pytorch/*.so from dgl-gpu-linux
[Pipeline] timeout
Timeout set to expire in 30 min
[Pipeline] {
[Pipeline] sh
+ bash tests/scripts/task_unit_test.sh mxnet gpu
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_find_edges
start graph service on server 0 for part 0
[05:31:03] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:31:03] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:10819]...
Requirement already satisfied: pytest in /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages (7.1.2)
Collecting psutil
PASSED            [  7%]
tests/compute/test_heterograph.py::test_empty_query[idtype0] PASSED      [  7%]
tests/compute/test_heterograph.py::test_empty_query[idtype1]   Downloading psutil-5.9.2-cp37-cp37m-manylinux_2_12_x86_64.manylinux2010_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl (281 kB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 281.3/281.3 kB 10.0 MB/s eta 0:00:00
Collecting pyyaml
  Downloading PyYAML-6.0-cp37-cp37m-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_12_x86_64.manylinux2010_x86_64.whl (596 kB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 596.3/596.3 kB 40.6 MB/s eta 0:00:00
PASSED      [  7%]
tests/compute/test_heterograph.py::test_adj[idtype0] PASSED              [  7%]
tests/compute/test_heterograph.py::test_adj[idtype1] Collecting pydantic
  Downloading pydantic-1.10.2-cp37-cp37m-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (11.8 MB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 11.8/11.8 MB 115.3 MB/s eta 0:00:00
PASSED              [  7%]
tests/compute/test_heterograph.py::test_inc[idtype0] PASSED              [  7%]
tests/compute/test_heterograph.py::test_inc[idtype1] PASSED              [  7%]
tests/compute/test_heterograph.py::test_view[idtype0] PASSED             [  7%]
tests/compute/test_heterograph.py::test_view[idtype1] PASSED             [  7%]
tests/compute/test_heterograph.py::test_view1[idtype0] /root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_find_edges
start graph service on server 1 for part 1
[05:31:04] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:31:04] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:10821]...
Collecting pandas
  Using cached pandas-1.1.5-cp37-cp37m-manylinux1_x86_64.whl (9.5 MB)
Collecting rdflib
[05:31:04] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:31:04] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
  Downloading rdflib-6.2.0-py3-none-any.whl (500 kB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 500.3/500.3 kB 68.7 MB/s eta 0:00:00
Collecting ogb
  Downloading ogb-1.3.4-py3-none-any.whl (78 kB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 78.6/78.6 kB 16.1 MB/s eta 0:00:00
Requirement already satisfied: importlib-metadata>=0.12 in /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages (from pytest) (4.12.0)
Requirement already satisfied: packaging in /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages (from pytest) (21.3)
Requirement already satisfied: pluggy<2.0,>=0.12 in /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages (from pytest) (1.0.0)
Requirement already satisfied: tomli>=1.0.0 in /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages (from pytest) (2.0.1)
Requirement already satisfied: attrs>=19.2.0 in /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages (from pytest) (22.1.0)
Requirement already satisfied: iniconfig in /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages (from pytest) (1.1.1)
Requirement already satisfied: py>=1.8.2 in /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages (from pytest) (1.11.0)
Requirement already satisfied: typing-extensions>=4.1.0 in /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages (from pydantic) (4.3.0)
Requirement already satisfied: numpy>=1.15.4 in /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages (from pandas) (1.21.6)
Collecting pytz>=2017.2
  Using cached pytz-2022.2.1-py2.py3-none-any.whl (500 kB)
Requirement already satisfied: python-dateutil>=2.7.3 in /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages (from pandas) (2.8.2)
Requirement already satisfied: pyparsing in /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages (from rdflib) (3.0.9)
Requirement already satisfied: setuptools in /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages (from rdflib) (61.2.0)
Collecting isodate
  Downloading isodate-0.6.1-py2.py3-none-any.whl (41 kB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 41.7/41.7 kB 9.7 MB/s eta 0:00:00
Requirement already satisfied: urllib3>=1.24.0 in /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages (from ogb) (1.26.11)
Collecting scikit-learn>=0.20.0
  Using cached scikit_learn-1.0.2-cp37-cp37m-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (24.8 MB)
Collecting outdated>=0.2.0
  Using cached outdated-0.2.1-py3-none-any.whl (7.5 kB)
Collecting torch>=1.6.0
  Downloading torch-1.12.1-cp37-cp37m-manylinux1_x86_64.whl (776.3 MB)
PASSED            [  7%]
tests/compute/test_heterograph.py::test_view1[idtype1] PASSED            [  7%]
tests/compute/test_heterograph.py::test_flatten[idtype0] PASSED          [  7%]
tests/compute/test_heterograph.py::test_flatten[idtype1] PASSED          [  7%]
tests/compute/test_heterograph.py::test_to_device[idtype0] SKIPPED (...) [  7%]
tests/compute/test_heterograph.py::test_to_device[idtype1] SKIPPED (...) [  7%]
tests/compute/test_heterograph.py::test_to_device2[g0-idtype0] SKIPPED   [  7%]
tests/compute/test_heterograph.py::test_to_device2[g0-idtype1] SKIPPED   [  8%]
tests/compute/test_heterograph.py::test_to_device2[g1-idtype0] SKIPPED   [  8%]
tests/compute/test_heterograph.py::test_to_device2[g1-idtype1] SKIPPED   [  8%]
tests/compute/test_heterograph.py::test_pin_memory_[idtype0] SKIPPED     [  8%]
tests/compute/test_heterograph.py::test_pin_memory_[idtype1] SKIPPED     [  8%]
tests/compute/test_heterograph.py::test_convert_bound[idtype0] PASSED    [  8%]
tests/compute/test_heterograph.py::test_convert_bound[idtype1] PASSED    [  8%]
tests/compute/test_heterograph.py::test_convert[idtype0] PASSED          [  8%]
tests/compute/test_heterograph.py::test_convert[idtype1] PASSED          [  8%]
tests/compute/test_heterograph.py::test_to_homo_zero_nodes[idtype0] PASSED [  8%]
tests/compute/test_heterograph.py::test_to_homo_zero_nodes[idtype1] PASSED [  8%]
tests/compute/test_heterograph.py::test_to_homo2[idtype0] PASSED         [  8%]
tests/compute/test_heterograph.py::test_to_homo2[idtype1] PASSED         [  8%]
tests/compute/test_heterograph.py::test_invertible_conversion[idtype0] PASSED [  8%]
tests/compute/test_heterograph.py::test_invertible_conversion[idtype1] PASSED [  8%]
tests/compute/test_heterograph.py::test_metagraph_reachable[idtype0] PASSED [  8%]
tests/compute/test_heterograph.py::test_metagraph_reachable[idtype1] PASSED [  8%]
tests/compute/test_heterograph.py::test_subgraph_mask[idtype0] PASSED    [  8%]
tests/compute/test_heterograph.py::test_subgraph_mask[idtype1] PASSED    [  8%]
tests/compute/test_heterograph.py::test_subgraph[idtype0] PASSED         [  8%]
tests/compute/test_heterograph.py::test_subgraph[idtype1] Client [1090] waits on 172.17.0.3:36223
Machine (0) group (0) client (0) connect to server successfuly!
Client[0] in group[0] is exiting...
Server (0) shutdown.
Server is exiting...
Server (1) shutdown.
Server is exiting...
  NumNodes: 2708
  NumEdges: 10556
  NumFeats: 1433
  NumClasses: 7
  NumTrainingSamples: 140
  NumValidationSamples: 500
  NumTestSamples: 1000
Done loading data from cached files.
Converting to homogeneous graph takes 0.001s, peak mem: 1.529 GB
Convert a graph into a bidirected graph: 0.001 seconds, peak memory: 1.529 GB
Construct multi-constraint weights: 0.000 seconds, peak memory: 1.529 GB
[05:31:07] /root/jenkins/workspace/dgl_PR-4648/src/graph/transform/metis_partition_hetero.cc:87: Partition a graph with 2708 nodes and 10556 edges into 2 parts and get 260 edge cuts
Metis partitioning: 0.002 seconds, peak memory: 1.529 GB
Assigning nodes to METIS partitions takes 0.004s, peak mem: 1.529 GB
Reshuffle nodes and edges: 0.001 seconds
Split the graph: 0.002 seconds
Construct subgraphs: 0.003 seconds
Splitting the graph into partitions takes 0.007s, peak mem: 1.529 GB
part 0 has 1512 nodes and 1354 are inside the partition
part 0 has 5630 edges and 5370 are inside the partition
part 1 has 1552 nodes and 1354 are inside the partition
part 1 has 5446 edges and 5186 are inside the partition
Save partitions: 0.032 seconds, peak memory: 1.529 GB
There are 10556 edges in the graph and 0 edge cuts for 2 partitions.
PASSED         [  8%]
tests/compute/test_heterograph.py::test_apply[idtype0] PASSED            [  8%]
tests/compute/test_heterograph.py::test_apply[idtype1] PASSED            [  8%]
tests/compute/test_heterograph.py::test_level2[idtype0] /root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_find_edges
start graph service on server 0 for part 0
[05:31:08] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:31:08] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:18024]...
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_find_edges
start graph service on server 1 for part 1
[05:31:09] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:31:09] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:18026]...
[05:31:09] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:31:09] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
PASSED           [  8%]
tests/compute/test_heterograph.py::test_level2[idtype1]      ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 776.3/776.3 MB 4.2 MB/s eta 0:00:00
PASSED           [  8%]
tests/compute/test_heterograph.py::test_more_nnz[idtype0] SKIPPED (N...) [  9%]
tests/compute/test_heterograph.py::test_more_nnz[idtype1] SKIPPED (N...) [  9%]
tests/compute/test_heterograph.py::test_updates[idtype0] Client [1090] waits on 172.17.0.3:42195
Machine (0) group (0) client (0) connect to server successfuly!
Client[0] in group[0] is exiting...
Server (1) shutdown.
Server is exiting...
Server (0) shutdown.
Server is exiting...
PASSED [ 16%]
tests/distributed/test_distributed_sampling.py::test_rpc_get_degree_shuffle[1]   NumNodes: 2708
  NumEdges: 10556
  NumFeats: 1433
  NumClasses: 7
  NumTrainingSamples: 140
  NumValidationSamples: 500
  NumTestSamples: 1000
Done loading data from cached files.
Converting to homogeneous graph takes 0.001s, peak mem: 1.529 GB
Save partitions: 0.023 seconds, peak memory: 1.529 GB
There are 10556 edges in the graph and 0 edge cuts for 1 partitions.
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_get_degrees
start graph service on server 0 for part 0
[05:31:13] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:31:13] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:27870]...
[05:31:13] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:31:13] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Requirement already satisfied: six>=1.12.0 in /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages (from ogb) (1.16.0)
Requirement already satisfied: tqdm>=4.29.0 in /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages (from ogb) (4.64.0)
Requirement already satisfied: zipp>=0.5 in /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages (from importlib-metadata>=0.12->pytest) (3.8.1)
Requirement already satisfied: requests in /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages (from outdated>=0.2.0->ogb) (2.28.1)
Collecting littleutils
  Using cached littleutils-0.2.2-py3-none-any.whl
Requirement already satisfied: scipy>=1.1.0 in /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages (from scikit-learn>=0.20.0->ogb) (1.7.3)
Collecting threadpoolctl>=2.0.0
  Using cached threadpoolctl-3.1.0-py3-none-any.whl (14 kB)
Requirement already satisfied: joblib>=0.11 in /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages (from scikit-learn>=0.20.0->ogb) (1.1.0)
Requirement already satisfied: idna<4,>=2.5 in /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages (from requests->outdated>=0.2.0->ogb) (3.3)
Requirement already satisfied: certifi>=2017.4.17 in /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages (from requests->outdated>=0.2.0->ogb) (2022.6.15)
Requirement already satisfied: charset-normalizer<3,>=2 in /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages (from requests->outdated>=0.2.0->ogb) (2.1.0)
Installing collected packages: pytz, littleutils, torch, threadpoolctl, pyyaml, pydantic, psutil, isodate, scikit-learn, rdflib, pandas, outdated, ogb
PASSED          [  9%]
tests/compute/test_heterograph.py::test_updates[idtype1] Client [1090] waits on 172.17.0.3:36663
Machine (0) group (0) client (0) connect to server successfuly!
Client[0] in group[0] is exiting...
Done get_degreeServer (0) shutdown.

Server is exiting...
PASSED          [  9%]
tests/compute/test_heterograph.py::test_backward[idtype0] PASSED         [  9%]
tests/compute/test_heterograph.py::test_backward[idtype1] PASSED         [  9%]
tests/compute/test_heterograph.py::test_empty_heterograph[idtype0] PASSED [  9%]
tests/compute/test_heterograph.py::test_empty_heterograph[idtype1] PASSED [  9%]
tests/compute/test_heterograph.py::test_types_in_function[idtype0] check results
PASSED [ 18%]
tests/distributed/test_distributed_sampling.py::test_rpc_get_degree_shuffle[2]   NumNodes: 2708
  NumEdges: 10556
  NumFeats: 1433
  NumClasses: 7
  NumTrainingSamples: 140
  NumValidationSamples: 500
  NumTestSamples: 1000
Done loading data from cached files.
Converting to homogeneous graph takes 0.001s, peak mem: 1.529 GB
Convert a graph into a bidirected graph: 0.001 seconds, peak memory: 1.529 GB
Construct multi-constraint weights: 0.000 seconds, peak memory: 1.529 GB
[05:31:17] /root/jenkins/workspace/dgl_PR-4648/src/graph/transform/metis_partition_hetero.cc:87: Partition a graph with 2708 nodes and 10556 edges into 2 parts and get 260 edge cuts
Metis partitioning: 0.003 seconds, peak memory: 1.529 GB
Assigning nodes to METIS partitions takes 0.004s, peak mem: 1.529 GB
Reshuffle nodes and edges: 0.001 seconds
Split the graph: 0.002 seconds
Construct subgraphs: 0.003 seconds
Splitting the graph into partitions takes 0.007s, peak mem: 1.529 GB
part 0 has 1512 nodes and 1354 are inside the partition
part 0 has 5630 edges and 5370 are inside the partition
part 1 has 1552 nodes and 1354 are inside the partition
part 1 has 5446 edges and 5186 are inside the partition
Save partitions: 0.021 seconds, peak memory: 1.529 GB
There are 10556 edges in the graph and 0 edge cuts for 2 partitions.
PASSED [  9%]
tests/compute/test_heterograph.py::test_types_in_function[idtype1] /root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_get_degrees
start graph service on server 0 for part 0
[05:31:18] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:31:18] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:12848]...
PASSED [  9%]
tests/compute/test_heterograph.py::test_stack_reduce[idtype0] PASSED     [  9%]
tests/compute/test_heterograph.py::test_stack_reduce[idtype1] /root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_get_degrees
start graph service on server 1 for part 1
[05:31:19] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:31:19] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:12850]...
PASSED     [  9%]
tests/compute/test_heterograph.py::test_isolated_ntype[idtype0] PASSED   [  9%]
tests/compute/test_heterograph.py::test_isolated_ntype[idtype1] PASSED   [  9%]
tests/compute/test_heterograph.py::test_ismultigraph[idtype0] [05:31:19] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:31:19] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Successfully installed dataclasses-0.8 isodate-0.6.1 littleutils-0.2.2 ogb-1.3.4 outdated-0.2.1 pandas-1.1.5 psutil-5.9.2 pydantic-1.9.2 pytz-2022.2.1 pyyaml-6.0 rdflib-5.0.0 scikit-learn-0.24.2 threadpoolctl-3.1.0 torch-1.10.2
WARNING: Running pip as the 'root' user can result in broken permissions and conflicting behaviour with the system package manager. It is recommended to use a virtual environment instead: https://pip.pypa.io/warnings/venv
============================= test session starts ==============================
platform linux -- Python 3.6.9, pytest-7.0.1, pluggy-1.0.0 -- /opt/conda/envs/mxnet-ci/bin/python3
cachedir: .pytest_cache
rootdir: /root/jenkins/workspace/dgl_PR-4648@4
collecting ... PASSED     [  9%]
tests/compute/test_heterograph.py::test_ismultigraph[idtype1] PASSED     [  9%]
tests/compute/test_heterograph.py::test_graph_index_is_unibipartite[idtype0] PASSED [  9%]
tests/compute/test_heterograph.py::test_graph_index_is_unibipartite[idtype1] PASSED [  9%]
tests/compute/test_heterograph.py::test_bipartite[idtype0] PASSED        [  9%]
tests/compute/test_heterograph.py::test_bipartite[idtype1] PASSED        [  9%]
tests/compute/test_heterograph.py::test_dtype_cast[idtype0] PASSED       [  9%]
tests/compute/test_heterograph.py::test_dtype_cast[idtype1] PASSED       [  9%]
tests/compute/test_heterograph.py::test_float_cast PASSED                [  9%]
tests/compute/test_heterograph.py::test_format[idtype0] PASSED           [  9%]
tests/compute/test_heterograph.py::test_format[idtype1] Client [1090] waits on 172.17.0.3:34321
Machine (0) group (0) client (0) connect to server successfuly!
Client[0] in group[0] is exiting...
Server (0) shutdown.
Server is exiting...
Server (1) shutdown.
Server is exiting...
Done get_degree
check results
PASSED [ 20%]
tests/distributed/test_distributed_sampling.py::test_rpc_sampling SKIPPED [ 21%]
tests/distributed/test_distributed_sampling.py::test_rpc_sampling_shuffle[1]   NumNodes: 2708
  NumEdges: 10556
  NumFeats: 1433
  NumClasses: 7
  NumTrainingSamples: 140
  NumValidationSamples: 500
  NumTestSamples: 1000
Done loading data from cached files.
Converting to homogeneous graph takes 0.001s, peak mem: 1.529 GB
Save partitions: 0.026 seconds, peak memory: 1.529 GB
There are 10556 edges in the graph and 0 edge cuts for 1 partitions.
PASSED           [ 10%]
tests/compute/test_heterograph.py::test_edges_order[idtype0] PASSED      [ 10%]
tests/compute/test_heterograph.py::test_edges_order[idtype1] PASSED      [ 10%]
tests/compute/test_heterograph.py::test_reverse[idtype0] PASSED          [ 10%]
tests/compute/test_heterograph.py::test_reverse[idtype1] PASSED                 [  5%]
tests/compute/test_dataloader.py::test_edge_prediction_sampler[idtype0] PASSED [  5%]
tests/compute/test_dataloader.py::test_edge_prediction_sampler[idtype1] PASSED [  5%]
tests/compute/test_edge_softmax_hetero.py::test_edge_softmax_unidirectional FAILED [  6%]
tests/compute/test_edge_softmax_hetero.py::test_edge_softmax[idtype0-src-g0] FAILED [  6%]
tests/compute/test_edge_softmax_hetero.py::test_edge_softmax[idtype0-dst-g0] FAILED [  6%]
tests/compute/test_edge_softmax_hetero.py::test_edge_softmax[idtype1-src-g0] FAILED [  6%]
tests/compute/test_edge_softmax_hetero.py::test_edge_softmax[idtype1-dst-g0] FAILED [  6%]
tests/compute/test_ffi.py::test_cython PASSED                            [  6%]
tests/compute/test_ffi.py::test_callback[1] PASSED                       [  6%]
tests/compute/test_ffi.py::test_callback[2.3] PASSED                     [  6%]
tests/compute/test_ffi.py::test_callback_thread[1] PASSED                [  6%]
tests/compute/test_ffi.py::test_callback_thread[2.3] PASSED              [  6%]
tests/compute/test_filter.py::test_graph_filter PASSED                   [  6%]
tests/compute/test_filter.py::test_array_filter[idtype0] SKIPPED (CP...) [  6%]
tests/compute/test_filter.py::test_array_filter[idtype1] SKIPPED (CP...) [  6%]
tests/compute/test_frame.py::test_column_subcolumn PASSED                [  6%]
tests/compute/test_frame.py::test_serialize_deserialize_plain PASSED     [  6%]
tests/compute/test_frame.py::test_serialize_deserialize_subcolumn PASSED [  6%]
tests/compute/test_frame.py::test_serialize_deserialize_dtype PASSED     [  6%]
tests/compute/test_generators.py::test_rand_graph PASSED                 [  6%]
tests/compute/test_graph.py::test_query /root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_sampling
start graph service on server 0 for part 0
[05:31:23] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:31:23] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:17223]...
PASSED          [ 10%]
tests/compute/test_heterograph.py::test_clone[idtype0] PASSED            [ 10%]
tests/compute/test_heterograph.py::test_clone[idtype1] PASSED            [ 10%]
tests/compute/test_heterograph.py::test_add_edges[idtype0] PASSED        [ 10%]
tests/compute/test_heterograph.py::test_add_edges[idtype1] PASSED        [ 10%]
tests/compute/test_heterograph.py::test_add_nodes[idtype0] PASSED        [ 10%]
tests/compute/test_heterograph.py::test_add_nodes[idtype1] PASSED        [ 10%]
tests/compute/test_heterograph.py::test_remove_edges[idtype0] PASSED     [ 10%]
tests/compute/test_heterograph.py::test_remove_edges[idtype1] PASSED     [ 10%]
tests/compute/test_heterograph.py::test_remove_nodes[idtype0] PASSED     [ 10%]
tests/compute/test_heterograph.py::test_remove_nodes[idtype1] [05:31:24] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:31:24] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
PASSED     [ 10%]
tests/compute/test_heterograph.py::test_frame[idtype0] PASSED            [ 10%]
tests/compute/test_heterograph.py::test_frame[idtype1] PASSED            [ 10%]
tests/compute/test_heterograph.py::test_frame_device[idtype0] SKIPPED    [ 10%]
tests/compute/test_heterograph.py::test_frame_device[idtype1] SKIPPED    [ 10%]
tests/compute/test_heterograph.py::test_create_block[idtype0] PASSED     [ 10%]
tests/compute/test_heterograph.py::test_create_block[idtype1] PASSED     [ 10%]
tests/compute/test_heterograph.py::test_adj_sparse[coo-idtype0] PASSED   [ 10%]
tests/compute/test_heterograph.py::test_adj_sparse[coo-idtype1] PASSED   [ 10%]
tests/compute/test_heterograph.py::test_adj_sparse[csr-idtype0] PASSED   [ 10%]
tests/compute/test_heterograph.py::test_adj_sparse[csr-idtype1] PASSED   [ 10%]
tests/compute/test_heterograph.py::test_adj_sparse[csc-idtype0] PASSED   [ 11%]
tests/compute/test_heterograph.py::test_adj_sparse[csc-idtype1] PASSED   [ 11%]
tests/compute/test_heterograph.py::test_forking_pickler collected 2453 items

tests/compute/test_apply_edges_hetero.py::test_unary_copy_u[int32] SKIPPED [  0%]
tests/compute/test_apply_edges_hetero.py::test_unary_copy_u[int64] SKIPPED [  0%]
tests/compute/test_apply_edges_hetero.py::test_unary_copy_e[int32] SKIPPED [  0%]
tests/compute/test_apply_edges_hetero.py::test_unary_copy_e[int64] SKIPPED [  0%]
tests/compute/test_apply_edges_hetero.py::test_binary_op[int32] SKIPPED  [  0%]
tests/compute/test_apply_edges_hetero.py::test_binary_op[int64] SKIPPED  [  0%]
tests/compute/test_backend.py::test_set_default_backend PASSED           [  0%]
tests/compute/test_basics.py::test_compatible PASSED                     [  0%]
tests/compute/test_basics.py::test_batch_setter_getter[int32] PASSED     [  0%]
tests/compute/test_basics.py::test_batch_setter_getter[int64] PASSED     [  0%]
tests/compute/test_basics.py::test_batch_setter_autograd[int32] PASSED   [  0%]
tests/compute/test_basics.py::test_batch_setter_autograd[int64] PASSED   [  0%]
tests/compute/test_basics.py::test_apply_nodes[int32] PASSED             [  0%]
tests/compute/test_basics.py::test_apply_nodes[int64] PASSED             [  0%]
tests/compute/test_basics.py::test_apply_edges[int32] PASSED             [  0%]
tests/compute/test_basics.py::test_apply_edges[int64] PASSED             [  0%]
tests/compute/test_basics.py::test_update_routines[int32] PASSED         [  0%]
tests/compute/test_basics.py::test_update_routines[int64] PASSED                           [  6%]
tests/compute/test_graph.py::test_mutation PASSED                        [  6%]
tests/compute/test_graph.py::test_scipy_adjmat PASSED                    [  6%]
tests/compute/test_graph.py::test_incmat PASSED                          [  6%]
tests/compute/test_graph.py::test_find_edges PASSED                      [  6%]
tests/compute/test_graph.py::test_ismultigraph PASSED                    [  6%]
tests/compute/test_graph.py::test_hypersparse_query PASSED               [  6%]
tests/compute/test_graph.py::test_empty_data_initialized PASSED          [  7%]
tests/compute/test_graph.py::test_is_sorted PASSED                       [  7%]
tests/compute/test_graph.py::test_default_types PASSED                   [  7%]
tests/compute/test_graph.py::test_formats PASSED                         [  7%]
tests/compute/test_heterograph.py::test_create[idtype0] PASSED           [  7%]
tests/compute/test_heterograph.py::test_create[idtype1] PASSED           [  7%]
tests/compute/test_heterograph.py::test_create2 PASSED                   [  7%]
tests/compute/test_heterograph.py::test_query[idtype0] PASSED         [  0%]
tests/compute/test_basics.py::test_update_all_0deg[int32] PASSED         [  0%]
tests/compute/test_basics.py::test_update_all_0deg[int64] PASSED         [  0%]
tests/compute/test_basics.py::test_pull_0deg[int32] PASSED               [  0%]
tests/compute/test_basics.py::test_pull_0deg[int64] PASSED               [  0%]
tests/compute/test_basics.py::test_dynamic_addition PASSED               [  0%]
tests/compute/test_basics.py::test_repr[int32] PASSED                    [  0%]
tests/compute/test_basics.py::test_repr[int64] PASSED                    [  1%]
tests/compute/test_basics.py::test_local_var[int32] PASSED               [  1%]
tests/compute/test_basics.py::test_local_var[int64] PASSED               [  1%]
tests/compute/test_basics.py::test_local_scope[int32] PASSED             [  1%]
tests/compute/test_basics.py::test_local_scope[int64] PASSED             [  1%]
tests/compute/test_basics.py::test_isolated_nodes[int32] PASSED          [  1%]
tests/compute/test_basics.py::test_isolated_nodes[int64] PASSED          [  1%]
tests/compute/test_basics.py::test_send_multigraph[int32] Client [1403] waits on 172.17.0.3:44143
Machine (0) group (0) client (0) connect to server successfuly!
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
Client[0] in group[0] is exiting...
Server (0) shutdown.
Server is exiting...
Converting to homogeneous graph takes 0.001s, peak mem: 1.529 GB
Save partitions: 0.004 seconds, peak memory: 1.529 GB
There are 3060 edges in the graph and 0 edge cuts for 1 partitions.
PASSED         [  1%]
tests/compute/test_basics.py::test_send_multigraph[int64] PASSED         [  1%]
tests/compute/test_basics.py::test_issue_1088[int32] PASSED              [  1%]
tests/compute/test_basics.py::test_issue_1088[int64] PASSED              [  1%]
tests/compute/test_basics.py::test_degree_bucket_edge_ordering[int32] PASSED            [  7%]
tests/compute/test_heterograph.py::test_query[idtype1] PASSED [  1%]
tests/compute/test_basics.py::test_degree_bucket_edge_ordering[int64] PASSED [  1%]
tests/compute/test_basics.py::test_issue_2484[int32] PASSED           [ 11%]
tests/compute/test_index.py::test_dlpack SKIPPED (TF doesn't support...) [ 11%]
tests/compute/test_kernel.py::test_copy_src_reduce /root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_sampling
start graph service on server 0 for part 0
[05:31:29] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:31:29] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:24719]...
PASSED              [  1%]
tests/compute/test_basics.py::test_issue_2484[int64] PASSED              [  1%]
tests/compute/test_batched_graph.py::test_batch_unbatch[int32] PASSED    [  1%]
tests/compute/test_batched_graph.py::test_batch_unbatch[int64] [05:31:29] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:31:29] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
PASSED    [  1%]
tests/compute/test_batched_graph.py::test_batch_unbatch1[int32] PASSED   [  1%]
tests/compute/test_batched_graph.py::test_batch_unbatch1[int64] PASSED   [  1%]
tests/compute/test_batched_graph.py::test_batch_unbatch_frame[int32] PASSED [  1%]
tests/compute/test_batched_graph.py::test_batch_unbatch_frame[int64] PASSED [  1%]
tests/compute/test_batched_graph.py::test_batch_unbatch2[int32] PASSED   [  1%]
tests/compute/test_batched_graph.py::test_batch_unbatch2[int64] PASSED   [  1%]
tests/compute/test_batched_graph.py::test_batch_send_and_recv[int32] PASSED [  1%]
tests/compute/test_batched_graph.py::test_batch_send_and_recv[int64] PASSED [  1%]
tests/compute/test_batched_graph.py::test_batch_propagate[int32] PASSED  [  2%]
tests/compute/test_batched_graph.py::test_batch_propagate[int64] PASSED            [  7%]
tests/compute/test_heterograph.py::test_empty_query[idtype0] PASSED      [  7%]
tests/compute/test_heterograph.py::test_empty_query[idtype1] PASSED      [  7%]
tests/compute/test_heterograph.py::test_adj[idtype0] PASSED              [  7%]
tests/compute/test_heterograph.py::test_adj[idtype1] PASSED              [  7%]
tests/compute/test_heterograph.py::test_inc[idtype0] PASSED              [  7%]
tests/compute/test_heterograph.py::test_inc[idtype1] PASSED              [  7%]
tests/compute/test_heterograph.py::test_view[idtype0] PASSED             [  7%]
tests/compute/test_heterograph.py::test_view[idtype1] PASSED             [  7%]
tests/compute/test_heterograph.py::test_view1[idtype0] PASSED  [  2%]
tests/compute/test_batched_graph.py::test_batched_edge_ordering[int32] PASSED [  2%]
tests/compute/test_batched_graph.py::test_batched_edge_ordering[int64] PASSED [  2%]
tests/compute/test_batched_graph.py::test_batch_no_edge[int32] PASSED    [  2%]
tests/compute/test_batched_graph.py::test_batch_no_edge[int64] Client [1090] waits on 172.17.0.3:50765
Machine (0) group (0) client (0) connect to server successfuly!
Client[0] in group[0] is exiting...
Server (0) shutdown.
Done samplingServer is exiting...

Converting to homogeneous graph takes 0.001s, peak mem: 1.529 GB
Save partitions: 0.001 seconds, peak memory: 1.529 GB
There are 3000 edges in the graph and 0 edge cuts for 1 partitions.
PASSED            [  7%]
tests/compute/test_heterograph.py::test_view1[idtype1] PASSED    [  2%]
tests/compute/test_batched_graph.py::test_batch_keeps_empty_data[int32] PASSED [  2%]
tests/compute/test_batched_graph.py::test_batch_keeps_empty_data[int64] PASSED [  2%]
tests/compute/test_batched_graph.py::test_set_batch_info[int32] PASSED   [  2%]
tests/compute/test_batched_graph.py::test_set_batch_info[int64] /root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_sampling
start graph service on server 0 for part 0
[05:31:33] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:31:33] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:13648]...
[05:31:33] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:31:33] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
PASSED            [  7%]
tests/compute/test_heterograph.py::test_flatten[idtype0] PASSED          [  7%]
tests/compute/test_heterograph.py::test_flatten[idtype1] PASSED          [  7%]
tests/compute/test_heterograph.py::test_to_device[idtype0] SKIPPED (...) [  7%]
tests/compute/test_heterograph.py::test_to_device[idtype1] SKIPPED (...) [  7%]
tests/compute/test_heterograph.py::test_to_device2[g0-idtype0] SKIPPED   [  7%]
tests/compute/test_heterograph.py::test_to_device2[g0-idtype1] SKIPPED   [  8%]
tests/compute/test_heterograph.py::test_to_device2[g1-idtype0] SKIPPED   [  8%]
tests/compute/test_heterograph.py::test_to_device2[g1-idtype1] SKIPPED   [  8%]
tests/compute/test_heterograph.py::test_pin_memory_[idtype0] SKIPPED     [  8%]
tests/compute/test_heterograph.py::test_pin_memory_[idtype1] SKIPPED     [  8%]
tests/compute/test_heterograph.py::test_convert_bound[idtype0] PASSED    [  8%]
tests/compute/test_heterograph.py::test_convert_bound[idtype1] PASSED    [  8%]
tests/compute/test_heterograph.py::test_convert[idtype0] PASSED   [  2%]
tests/compute/test_batched_heterograph.py::test_topology[int32-gs0] PASSED [  2%]
tests/compute/test_batched_heterograph.py::test_topology[int32-gs1] PASSED          [  8%]
tests/compute/test_heterograph.py::test_convert[idtype1] PASSED          [  8%]
tests/compute/test_heterograph.py::test_to_homo_zero_nodes[idtype0] PASSED [  8%]
tests/compute/test_heterograph.py::test_to_homo_zero_nodes[idtype1] PASSED [  8%]
tests/compute/test_heterograph.py::test_to_homo2[idtype0] PASSED         [  8%]
tests/compute/test_heterograph.py::test_to_homo2[idtype1] PASSED         [  8%]
tests/compute/test_heterograph.py::test_invertible_conversion[idtype0] PASSED [  8%]
tests/compute/test_heterograph.py::test_invertible_conversion[idtype1] PASSED [  8%]
tests/compute/test_heterograph.py::test_metagraph_reachable[idtype0] PASSED [  2%]
tests/compute/test_batched_heterograph.py::test_topology[int64-gs0] PASSED [  2%]
tests/compute/test_batched_heterograph.py::test_topology[int64-gs1] PASSED [  2%]
tests/compute/test_batched_heterograph.py::test_batching_batched[int32] PASSED [  8%]
tests/compute/test_heterograph.py::test_metagraph_reachable[idtype1] PASSED [  8%]
tests/compute/test_heterograph.py::test_subgraph_mask[idtype0] PASSED [  2%]
tests/compute/test_batched_heterograph.py::test_batching_batched[int64] PASSED [  2%]
tests/compute/test_batched_heterograph.py::test_features[int32] PASSED    [  8%]
tests/compute/test_heterograph.py::test_subgraph_mask[idtype1] PASSED    [  8%]
tests/compute/test_heterograph.py::test_subgraph[idtype0] PASSED         [  8%]
tests/compute/test_heterograph.py::test_subgraph[idtype1] PASSED         [  8%]
tests/compute/test_heterograph.py::test_apply[idtype0] PASSED            [  8%]
tests/compute/test_heterograph.py::test_apply[idtype1] PASSED            [  8%]
tests/compute/test_heterograph.py::test_level2[idtype0] PASSED   [  2%]
tests/compute/test_batched_heterograph.py::test_features[int64] PASSED   [  2%]
tests/compute/test_batched_heterograph.py::test_empty_relation[int32] SKIPPED [  2%]
tests/compute/test_batched_heterograph.py::test_empty_relation[int64] SKIPPED [  2%]
tests/compute/test_batched_heterograph.py::test_unbatch2[int32] PASSED   [  2%]
tests/compute/test_batched_heterograph.py::test_unbatch2[int64] PASSED   [  2%]
tests/compute/test_batched_heterograph.py::test_slice_batch[int32] PASSED [  2%]
tests/compute/test_batched_heterograph.py::test_slice_batch[int64] PASSED [  2%]
tests/compute/test_batched_heterograph.py::test_batch_keeps_empty_data[int32] PASSED [  3%]
tests/compute/test_batched_heterograph.py::test_batch_keeps_empty_data[int64] PASSED [  3%]
tests/compute/test_batched_heterograph.py::test_batch_netypes PASSED     [  3%]
tests/compute/test_csrmm.py::test_csrmm[float32-int32] PASSED            [  3%]
tests/compute/test_csrmm.py::test_csrmm[float32-int64] PASSED           [  8%]
tests/compute/test_heterograph.py::test_level2[idtype1] PASSED            [  3%]
tests/compute/test_csrmm.py::test_csrmm[float64-int32] PASSED            [  3%]
tests/compute/test_csrmm.py::test_csrmm[float64-int64] Successfully installed isodate-0.6.1 littleutils-0.2.2 ogb-1.3.4 outdated-0.2.1 pandas-1.1.5 psutil-5.9.2 pydantic-1.10.2 pytz-2022.2.1 pyyaml-6.0 rdflib-6.2.0 scikit-learn-1.0.2 threadpoolctl-3.1.0 torch-1.12.1
WARNING: Running pip as the 'root' user can result in broken permissions and conflicting behaviour with the system package manager. It is recommended to use a virtual environment instead: https://pip.pypa.io/warnings/venv
============================= test session starts ==============================
platform linux -- Python 3.7.0, pytest-7.1.2, pluggy-1.0.0 -- /opt/conda/envs/mxnet-ci/bin/python3
cachedir: .pytest_cache
rootdir: /root/jenkins/workspace/dgl_PR-4648
collecting ... PASSED            [  3%]
tests/compute/test_csrmm.py::test_csrmm_backward[1-float32-int32] PASSED [  3%]
tests/compute/test_csrmm.py::test_csrmm_backward[1-float32-int64] PASSED [  3%]
tests/compute/test_csrmm.py::test_csrmm_backward[1-float64-int32] PASSED [  3%]
tests/compute/test_csrmm.py::test_csrmm_backward[1-float64-int64] Client [1090] waits on 172.17.0.3:43441
Machine (0) group (0) client (0) connect to server successfuly!
Client[0] in group[0] is exiting...
Done sampling
Server (0) shutdown.
Server is exiting...
Converting to homogeneous graph takes 0.001s, peak mem: 1.529 GB
Save partitions: 0.002 seconds, peak memory: 1.529 GB
There are 13220 edges in the graph and 0 edge cuts for 1 partitions.
PASSED [  3%]
tests/compute/test_csrmm.py::test_csrmm_backward[2-float32-int32] PASSED [  3%]
tests/compute/test_csrmm.py::test_csrmm_backward[2-float32-int64] PASSED [  3%]
tests/compute/test_csrmm.py::test_csrmm_backward[2-float64-int32] PASSED [  3%]
tests/compute/test_csrmm.py::test_csrmm_backward[2-float64-int64] PASSED           [  8%]
tests/compute/test_heterograph.py::test_more_nnz[idtype0] SKIPPED (N...) [  9%]
tests/compute/test_heterograph.py::test_more_nnz[idtype1] SKIPPED (N...) [  9%]
tests/compute/test_heterograph.py::test_updates[idtype0] FAILED          [  9%]
tests/compute/test_heterograph.py::test_updates[idtype1] PASSED [  3%]
tests/compute/test_csrmm.py::test_csrsum[float32-int32] PASSED           [  3%]
tests/compute/test_csrmm.py::test_csrsum[float32-int64] PASSED           [  3%]
tests/compute/test_csrmm.py::test_csrsum[float64-int32] PASSED           [  3%]
tests/compute/test_csrmm.py::test_csrsum[float64-int64] PASSED           [  3%]
tests/compute/test_csrmm.py::test_csrsum_backward[1-float32-int32] PASSED [  3%]
tests/compute/test_csrmm.py::test_csrsum_backward[1-float32-int64] FAILED          [  9%]
tests/compute/test_heterograph.py::test_backward[idtype0] FAILED         [  9%]
tests/compute/test_heterograph.py::test_backward[idtype1] PASSED [  3%]
tests/compute/test_csrmm.py::test_csrsum_backward[1-float64-int32] PASSED [  3%]
tests/compute/test_csrmm.py::test_csrsum_backward[1-float64-int64] FAILED         [  9%]
tests/compute/test_heterograph.py::test_empty_heterograph[idtype0] PASSED [  9%]
tests/compute/test_heterograph.py::test_empty_heterograph[idtype1] PASSED [  9%]
tests/compute/test_heterograph.py::test_types_in_function[idtype0] /root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_sampling
start graph service on server 0 for part 0
[05:31:37] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:31:37] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:13648]...
[05:31:37] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:31:37] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
PASSED [  3%]
tests/compute/test_csrmm.py::test_csrsum_backward[2-float32-int32] PASSED [  3%]
tests/compute/test_csrmm.py::test_csrsum_backward[2-float32-int64] PASSED [  3%]
tests/compute/test_csrmm.py::test_csrsum_backward[2-float64-int32] PASSED [  4%]
tests/compute/test_csrmm.py::test_csrsum_backward[2-float64-int64] PASSED [  4%]
tests/compute/test_csrmm.py::test_csrmask[9000-9000-float32-int32] PASSED [  4%]
tests/compute/test_csrmm.py::test_csrmask[9000-9000-float32-int64] PASSED [  4%]
tests/compute/test_csrmm.py::test_csrmask[9000-9000-float64-int32] PASSED [  4%]
tests/compute/test_csrmm.py::test_csrmask[9000-9000-float64-int64] PASSED [  9%]
tests/compute/test_heterograph.py::test_types_in_function[idtype1] PASSED [  9%]
tests/compute/test_heterograph.py::test_stack_reduce[idtype0] PASSED     [  9%]
tests/compute/test_heterograph.py::test_stack_reduce[idtype1] PASSED [  4%]
tests/compute/test_csrmm.py::test_csrmask[9000-0-float32-int32] PASSED   [  4%]
tests/compute/test_csrmm.py::test_csrmask[9000-0-float32-int64] PASSED   [  4%]
tests/compute/test_csrmm.py::test_csrmask[9000-0-float64-int32] PASSED   [  4%]
tests/compute/test_csrmm.py::test_csrmask[9000-0-float64-int64] PASSED   [  4%]
tests/compute/test_csrmm.py::test_csrmask[0-9000-float32-int32] PASSED   [  4%]
tests/compute/test_csrmm.py::test_csrmask[0-9000-float32-int64] PASSED   [  4%]
tests/compute/test_csrmm.py::test_csrmask[0-9000-float64-int32] PASSED   [  4%]
tests/compute/test_csrmm.py::test_csrmask[0-9000-float64-int64] PASSED   [  4%]
tests/compute/test_csrmm.py::test_csrmask[0-0-float32-int32] PASSED      [  4%]
tests/compute/test_csrmm.py::test_csrmask[0-0-float32-int64] PASSED      [  4%]
tests/compute/test_csrmm.py::test_csrmask[0-0-float64-int32] PASSED      [  4%]
tests/compute/test_csrmm.py::test_csrmask[0-0-float64-int64] PASSED      [  4%]
tests/compute/test_csrmm.py::test_csrmask_backward[float32-int32] PASSED     [  9%]
tests/compute/test_heterograph.py::test_isolated_ntype[idtype0] PASSED   [  9%]
tests/compute/test_heterograph.py::test_isolated_ntype[idtype1] PASSED   [  9%]
tests/compute/test_heterograph.py::test_ismultigraph[idtype0] PASSED     [  9%]
tests/compute/test_heterograph.py::test_ismultigraph[idtype1] PASSED     [  9%]
tests/compute/test_heterograph.py::test_graph_index_is_unibipartite[idtype0] PASSED [  9%]
tests/compute/test_heterograph.py::test_graph_index_is_unibipartite[idtype1] PASSED [  9%]
tests/compute/test_heterograph.py::test_bipartite[idtype0] PASSED        [  9%]
tests/compute/test_heterograph.py::test_bipartite[idtype1] PASSED        [  9%]
tests/compute/test_heterograph.py::test_dtype_cast[idtype0] PASSED       [  9%]
tests/compute/test_heterograph.py::test_dtype_cast[idtype1] PASSED       [  9%]
tests/compute/test_heterograph.py::test_float_cast PASSED                [  9%]
tests/compute/test_heterograph.py::test_format[idtype0] PASSED [  4%]
tests/compute/test_csrmm.py::test_csrmask_backward[float32-int64] PASSED [  4%]
tests/compute/test_csrmm.py::test_csrmask_backward[float64-int32] PASSED           [  9%]
tests/compute/test_heterograph.py::test_format[idtype1] PASSED           [ 10%]
tests/compute/test_heterograph.py::test_edges_order[idtype0] PASSED      [ 10%]
tests/compute/test_heterograph.py::test_edges_order[idtype1] PASSED      [ 10%]
tests/compute/test_heterograph.py::test_reverse[idtype0] PASSED          [ 10%]
tests/compute/test_heterograph.py::test_reverse[idtype1] PASSED          [ 10%]
tests/compute/test_heterograph.py::test_clone[idtype0] PASSED            [ 10%]
tests/compute/test_heterograph.py::test_clone[idtype1] PASSED            [ 10%]
tests/compute/test_heterograph.py::test_add_edges[idtype0] PASSED        [ 10%]
tests/compute/test_heterograph.py::test_add_edges[idtype1] PASSED        [ 10%]
tests/compute/test_heterograph.py::test_add_nodes[idtype0] PASSED [  4%]
tests/compute/test_csrmm.py::test_csrmask_backward[float64-int64] PASSED [  4%]
tests/compute/test_data.py::test_minigc SKIPPED (Skip MXNet)             [  4%]
tests/compute/test_data.py::test_gin SKIPPED (Skip MXNet)                [  4%]
tests/compute/test_data.py::test_fraud SKIPPED (Skip MXNet)              [  5%]
tests/compute/test_data.py::test_fakenews SKIPPED (Skip MXNet)           [  5%]
tests/compute/test_data.py::test_tudataset_regression SKIPPED (Skip ...) [  5%]
tests/compute/test_data.py::test_data_hash SKIPPED (Skip MXNet)          [  5%]
tests/compute/test_data.py::test_citation_graph SKIPPED (Skip MXNet)     [  5%]
tests/compute/test_data.py::test_gnn_benchmark SKIPPED (Skip MXNet)      [  5%]
tests/compute/test_data.py::test_reddit SKIPPED (Skip MXNet)             [  5%]
tests/compute/test_data.py::test_explain_syn SKIPPED (Skip MXNet)        [  5%]
tests/compute/test_data.py::test_wiki_cs SKIPPED (Skip MXNet)            [  5%]
tests/compute/test_data.py::test_yelp SKIPPED (Dataset too large to ...) [  5%]
tests/compute/test_data.py::test_flickr SKIPPED (Skip MXNet)             [  5%]
tests/compute/test_data.py::test_extract_archive SKIPPED (Skip MXNet)    [  5%]
tests/compute/test_data.py::test_csvdataset SKIPPED (Skip MXNet)         [  5%]
tests/compute/test_data.py::test_add_nodepred_split SKIPPED (Skip MXNet) [  5%]
tests/compute/test_data.py::test_as_nodepred1 SKIPPED (Skip MXNet)       [  5%]
tests/compute/test_data.py::test_as_nodepred2 SKIPPED (Skip MXNet)       [  5%]
tests/compute/test_data.py::test_as_nodepred_ogb SKIPPED (ogb only s...) [  5%]
tests/compute/test_data.py::test_as_linkpred SKIPPED (Skip MXNet)        [  5%]
tests/compute/test_data.py::test_as_linkpred_ogb SKIPPED (ogb only s...) [  5%]
tests/compute/test_data.py::test_as_nodepred_csvdataset SKIPPED (Ski...) [  5%]
tests/compute/test_data.py::test_as_graphpred SKIPPED (Skip MXNet)       [  5%]
tests/compute/test_data.py::test_as_graphpred_reprocess SKIPPED (Ski...) [  5%]
tests/compute/test_data.py::test_as_graphpred_ogb SKIPPED (ogb only ...) [  5%]
tests/compute/test_dataloader.py::test_edge_prediction_sampler[int32] PASSED        [ 10%]
tests/compute/test_heterograph.py::test_add_nodes[idtype1] PASSED        [ 10%]
tests/compute/test_heterograph.py::test_remove_edges[idtype0] PASSED     [ 10%]
tests/compute/test_heterograph.py::test_remove_edges[idtype1] PASSED     [ 10%]
tests/compute/test_heterograph.py::test_remove_nodes[idtype0] PASSED     [ 10%]
tests/compute/test_heterograph.py::test_remove_nodes[idtype1] PASSED     [ 10%]
tests/compute/test_heterograph.py::test_frame[idtype0] PASSED            [ 10%]
tests/compute/test_heterograph.py::test_frame[idtype1] PASSED            [ 10%]
tests/compute/test_heterograph.py::test_frame_device[idtype0] SKIPPED    [ 10%]
tests/compute/test_heterograph.py::test_frame_device[idtype1] SKIPPED    [ 10%]
tests/compute/test_heterograph.py::test_create_block[idtype0] PASSED     [ 10%]
tests/compute/test_heterograph.py::test_create_block[idtype1] PASSED     [ 10%]
tests/compute/test_heterograph.py::test_adj_sparse[coo-idtype0] PASSED   [ 10%]
tests/compute/test_heterograph.py::test_adj_sparse[coo-idtype1] PASSED   [ 10%]
tests/compute/test_heterograph.py::test_adj_sparse[csr-idtype0] PASSED   [ 10%]
tests/compute/test_heterograph.py::test_adj_sparse[csr-idtype1] PASSED   [ 10%]
tests/compute/test_heterograph.py::test_adj_sparse[csc-idtype0] PASSED   [ 11%]
tests/compute/test_heterograph.py::test_adj_sparse[csc-idtype1] PASSED   [ 11%]
tests/compute/test_heterograph.py::test_forking_pickler PASSED                [ 11%]
tests/compute/test_kernel.py::test_copy_edge_reduce PASSED [  5%]
tests/compute/test_dataloader.py::test_edge_prediction_sampler[int64] PASSED [  5%]
tests/compute/test_edge_softmax_hetero.py::test_edge_softmax_unidirectional SKIPPED [  6%]
tests/compute/test_edge_softmax_hetero.py::test_edge_softmax[int32-src-g0] SKIPPED [  6%]
tests/compute/test_edge_softmax_hetero.py::test_edge_softmax[int32-dst-g0] SKIPPED [  6%]
tests/compute/test_edge_softmax_hetero.py::test_edge_softmax[int64-src-g0] SKIPPED [  6%]
tests/compute/test_edge_softmax_hetero.py::test_edge_softmax[int64-dst-g0] SKIPPED [  6%]
tests/compute/test_filter.py::test_graph_filter PASSED                   [  6%]
tests/compute/test_filter.py::test_array_filter[int32] SKIPPED (CPU ...) [  6%]
tests/compute/test_filter.py::test_array_filter[int64] SKIPPED (CPU ...) [  6%]
tests/compute/test_frame.py::test_column_subcolumn PASSED                [  6%]
tests/compute/test_frame.py::test_serialize_deserialize_plain PASSED     [  6%]
tests/compute/test_frame.py::test_serialize_deserialize_subcolumn PASSED [  6%]
tests/compute/test_frame.py::test_serialize_deserialize_dtype PASSED     [  6%]
tests/compute/test_generators.py::test_rand_graph PASSED                 [  6%]
tests/compute/test_graph.py::test_query PASSED           [ 11%]
tests/compute/test_index.py::test_dlpack PASSED                          [ 11%]
tests/compute/test_kernel.py::test_copy_src_reduce FAILED                [ 11%]
tests/compute/test_kernel.py::test_copy_edge_reduce FAILED               [ 11%]
tests/compute/test_kernel.py::test_all_binary_builtins FAILED            [ 11%]
tests/compute/test_kernel.py::test_mean_zero_degree[g0-idtype0] FAILED   [ 11%]
tests/compute/test_kernel.py::test_mean_zero_degree[g0-idtype1] FAILED   [ 11%]
tests/compute/test_merge.py::test_heterograph_merge[idtype0] PASSED      [ 11%]
tests/compute/test_merge.py::test_heterograph_merge[idtype1] PASSED      [ 11%]
tests/compute/test_nccl.py::test_nccl_id SKIPPED (NCCL only runs on ...) [ 11%]
tests/compute/test_nccl.py::test_nccl_sparse_push_single_remainder SKIPPED [ 11%]
tests/compute/test_nccl.py::test_nccl_sparse_pull_single_remainder SKIPPED [ 11%]
tests/compute/test_nccl.py::test_nccl_sparse_push_single_range SKIPPED   [ 11%]
tests/compute/test_nccl.py::test_nccl_sparse_pull_single_range SKIPPED   [ 11%]
tests/compute/test_nccl.py::test_nccl_support SKIPPED (NCCL only run...) [ 11%]
tests/compute/test_new_update_all_hetero.py::test_unary_copy_u[idtype0] FAILED [ 11%]
tests/compute/test_new_update_all_hetero.py::test_unary_copy_u[idtype1] FAILED [ 11%]
tests/compute/test_new_update_all_hetero.py::test_unary_copy_e[idtype0] Client [1090] waits on 172.17.0.3:58957
Machine (0) group (0) client (0) connect to server successfuly!
Client[0] in group[0] is exiting...
FAILED [ 11%]
tests/compute/test_new_update_all_hetero.py::test_unary_copy_e[idtype1] FAILED [ 11%]
tests/compute/test_new_update_all_hetero.py::test_binary_op[idtype0] FAILED [ 11%]
tests/compute/test_new_update_all_hetero.py::test_binary_op[idtype1] FAILED [ 11%]
tests/compute/test_partition.py::test_get_node_partition_from_book[idtype0] SKIPPED [ 11%]
tests/compute/test_partition.py::test_get_node_partition_from_book[idtype1] SKIPPED [ 12%]
tests/compute/test_pickle.py::test_pickling_index PASSED                 [ 12%]
tests/compute/test_pickle.py::test_pickling_graph_index PASSED           [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g0-idtype0] PASSED     [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g0-idtype1] PASSED     [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g1-idtype0] PASSED     [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g1-idtype1] PASSED     [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g2-idtype0] PASSED     [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g2-idtype1] PASSED     [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g3-idtype0] PASSED     [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g3-idtype1] PASSED     [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g4-idtype0] PASSED     [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g4-idtype1] PASSED     [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g5-idtype0] PASSED     [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g5-idtype1] PASSED     [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g6-idtype0] PASSED     [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g6-idtype1] PASSED     [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g7-idtype0] PASSED     [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g7-idtype1] PASSED     [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g8-idtype0] PASSED     [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g8-idtype1] PASSED     [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g9-idtype0] PASSED     [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g9-idtype1] PASSED     [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g10-idtype0] PASSED    [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g10-idtype1] PASSED    [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g11-idtype0] PASSED    [ 13%]
tests/compute/test_pickle.py::test_pickling_graph[g11-idtype1] PASSED    [ 13%]
tests/compute/test_pickle.py::test_pickling_graph[g12-idtype0] PASSED    [ 13%]
tests/compute/test_pickle.py::test_pickling_graph[g12-idtype1] PASSED    [ 13%]
tests/compute/test_pickle.py::test_pickling_batched_heterograph PASSED   [ 13%]
tests/compute/test_pickle.py::test_pickling_subgraph PASSED              [ 13%]
tests/compute/test_pickle.py::test_pickling_is_pinned[idtype0] SKIPPED   [ 13%]
tests/compute/test_pickle.py::test_pickling_is_pinned[idtype1] SKIPPED   [ 13%]
tests/compute/test_pin_memory.py::test_pin_unpin SKIPPED (Need gpu f...) [ 13%]
tests/compute/test_propagate.py::test_prop_nodes_bfs[idtype0] PASSED     [ 13%]
tests/compute/test_propagate.py::test_prop_nodes_bfs[idtype1] PASSED     [ 13%]
tests/compute/test_propagate.py::test_prop_edges_dfs[idtype0] Done sampling
Server (0) shutdown.
Server is exiting...
Converting to homogeneous graph takes 0.001s, peak mem: 1.529 GB
Save partitions: 0.002 seconds, peak memory: 1.529 GB
There are 13220 edges in the graph and 0 edge cuts for 1 partitions.
PASSED     [ 13%]
tests/compute/test_propagate.py::test_prop_edges_dfs[idtype1] PASSED     [ 13%]
tests/compute/test_propagate.py::test_prop_nodes_topo[idtype0] PASSED    [ 13%]
tests/compute/test_propagate.py::test_prop_nodes_topo[idtype1] /root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_sampling
start graph service on server 0 for part 0
[05:31:42] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:31:42] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:13648]...
[05:31:42] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:31:42] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
PASSED    [ 13%]
tests/compute/test_random.py::test_random_choice PASSED                  [ 13%]
tests/compute/test_readout.py::test_sum_case1[idtype0] FAILED            [ 13%]
tests/compute/test_readout.py::test_sum_case1[idtype1] FAILED            [ 13%]
tests/compute/test_readout.py::test_reduce_readout[sum-g0-idtype0] FAILED [ 13%]
tests/compute/test_readout.py::test_reduce_readout[sum-g0-idtype1] FAILED [ 13%]
tests/compute/test_readout.py::test_reduce_readout[sum-g1-idtype0] FAILED [ 13%]
tests/compute/test_readout.py::test_reduce_readout[sum-g1-idtype1] PASSED                           [  6%]
tests/compute/test_graph.py::test_mutation PASSED                        [  6%]
tests/compute/test_graph.py::test_scipy_adjmat PASSED                    [  6%]
tests/compute/test_graph.py::test_incmat PASSED                          [  6%]FAILED [ 13%]
tests/compute/test_readout.py::test_reduce_readout[sum-g2-idtype0] FAILED [ 13%]
tests/compute/test_readout.py::test_reduce_readout[sum-g2-idtype1] FAILED [ 13%]
tests/compute/test_readout.py::test_reduce_readout[sum-g3-idtype0] FAILED [ 13%]
tests/compute/test_readout.py::test_reduce_readout[sum-g3-idtype1] FAILED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[sum-g4-idtype0] FAILED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[sum-g4-idtype1] FAILED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[sum-g5-idtype0] 
tests/compute/test_graph.py::test_find_edges PASSED                      [  6%]
tests/compute/test_graph.py::test_ismultigraph PASSED                    [  6%]
tests/compute/test_graph.py::test_hypersparse_query PASSED               [  6%]
tests/compute/test_graph.py::test_empty_data_initialized PASSED          [  6%]
tests/compute/test_graph.py::test_is_sorted PASSED                       [  6%]
tests/compute/test_graph.py::test_default_types PASSED                   [  6%]
tests/compute/test_graph.py::test_formats PASSED                         [  6%]
tests/compute/test_heterograph.py::test_create[int32] FAILED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[sum-g5-idtype1] FAILED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[sum-g6-idtype0] FAILED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[sum-g6-idtype1] FAILED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g0-idtype0] FAILED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g0-idtype1] FAILED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g1-idtype0] PASSED             [  7%]
tests/compute/test_heterograph.py::test_create[int64] PASSED             [  7%]
tests/compute/test_heterograph.py::test_create2 PASSED                   [  7%]
tests/compute/test_heterograph.py::test_query[int32] FAILED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g1-idtype1] FAILED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g2-idtype0] FAILED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g2-idtype1] FAILED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g3-idtype0] FAILED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g3-idtype1] FAILED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g4-idtype0] FAILED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g4-idtype1] PASSED               [ 11%]
tests/compute/test_kernel.py::test_all_binary_builtins FAILED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g5-idtype0] FAILED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g5-idtype1] FAILED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g6-idtype0] FAILED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g6-idtype1] FAILED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[mean-g0-idtype0] FAILED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[mean-g0-idtype1] FAILED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[mean-g1-idtype0] collected 2453 items

tests/compute/test_apply_edges_hetero.py::test_unary_copy_u[int32] SKIPPED [  0%]
tests/compute/test_apply_edges_hetero.py::test_unary_copy_u[int64] SKIPPED [  0%]
tests/compute/test_apply_edges_hetero.py::test_unary_copy_e[int32] SKIPPED [  0%]
tests/compute/test_apply_edges_hetero.py::test_unary_copy_e[int64] SKIPPED [  0%]
tests/compute/test_apply_edges_hetero.py::test_binary_op[int32] SKIPPED  [  0%]
tests/compute/test_apply_edges_hetero.py::test_binary_op[int64] SKIPPED  [  0%]
tests/compute/test_backend.py::test_set_default_backend PASSED           [  0%]
tests/compute/test_basics.py::test_compatible FAILED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[mean-g1-idtype1] FAILED [ 15%]
tests/compute/test_readout.py::test_reduce_readout[mean-g2-idtype0] FAILED [ 15%]
tests/compute/test_readout.py::test_reduce_readout[mean-g2-idtype1] FAILED [ 15%]
tests/compute/test_readout.py::test_reduce_readout[mean-g3-idtype0] FAILED [ 15%]
tests/compute/test_readout.py::test_reduce_readout[mean-g3-idtype1] FAILED [ 15%]
tests/compute/test_readout.py::test_reduce_readout[mean-g4-idtype0] FAILED [ 15%]
tests/compute/test_readout.py::test_reduce_readout[mean-g4-idtype1] FAILED [ 15%]
tests/compute/test_readout.py::test_reduce_readout[mean-g5-idtype0] FAILED [ 15%]
tests/compute/test_readout.py::test_reduce_readout[mean-g5-idtype1] FAILED [ 15%]
tests/compute/test_readout.py::test_reduce_readout[mean-g6-idtype0] FAILED [ 15%]
tests/compute/test_readout.py::test_reduce_readout[mean-g6-idtype1] FAILED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g0-idtype0] FAILED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g0-idtype1] PASSED                     [  0%]
tests/compute/test_basics.py::test_batch_setter_getter[int32] FAILED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g1-idtype0] FAILED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g1-idtype1] FAILED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g2-idtype0] FAILED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g2-idtype1] FAILED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g3-idtype0] FAILED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g3-idtype1] FAILED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g4-idtype0] PASSED     [  0%]
tests/compute/test_basics.py::test_batch_setter_getter[int64] PASSED     [  0%]
tests/compute/test_basics.py::test_batch_setter_autograd[int32] PASSED   [  0%]
tests/compute/test_basics.py::test_batch_setter_autograd[int64] PASSED   [  0%]
tests/compute/test_basics.py::test_apply_nodes[int32] PASSED             [  0%]
tests/compute/test_basics.py::test_apply_nodes[int64] PASSED             [  0%]
tests/compute/test_basics.py::test_apply_edges[int32] PASSED             [  0%]
tests/compute/test_basics.py::test_apply_edges[int64] PASSED             [  0%]
tests/compute/test_basics.py::test_update_routines[int32] PASSED         [  0%]
tests/compute/test_basics.py::test_update_routines[int64] PASSED         [  0%]
tests/compute/test_basics.py::test_update_all_0deg[int32] PASSED         [  0%]
tests/compute/test_basics.py::test_update_all_0deg[int64] PASSED         [  0%]
tests/compute/test_basics.py::test_pull_0deg[int32] PASSED               [  0%]
tests/compute/test_basics.py::test_pull_0deg[int64] PASSED               [  0%]
tests/compute/test_basics.py::test_dynamic_addition PASSED               [  0%]
tests/compute/test_basics.py::test_repr[int32] PASSED                    [  0%]
tests/compute/test_basics.py::test_repr[int64] PASSED                    [  1%]
tests/compute/test_basics.py::test_local_var[int32] PASSED               [  1%]
tests/compute/test_basics.py::test_local_var[int64] PASSED               [  1%]
tests/compute/test_basics.py::test_local_scope[int32] FAILED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g4-idtype1] FAILED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g5-idtype0] FAILED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g5-idtype1] FAILED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g6-idtype0] FAILED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g6-idtype1] FAILED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g0-idtype0] PASSED             [  1%]
tests/compute/test_basics.py::test_local_scope[int64] PASSED             [  1%]
tests/compute/test_basics.py::test_isolated_nodes[int32] PASSED          [  1%]
tests/compute/test_basics.py::test_isolated_nodes[int64] PASSED          [  1%]
tests/compute/test_basics.py::test_send_multigraph[int32] PASSED         [  1%]
tests/compute/test_basics.py::test_send_multigraph[int64] PASSED         [  1%]
tests/compute/test_basics.py::test_issue_1088[int32] PASSED              [  1%]
tests/compute/test_basics.py::test_issue_1088[int64] PASSED              [  1%]
tests/compute/test_basics.py::test_degree_bucket_edge_ordering[int32] PASSED [  1%]
tests/compute/test_basics.py::test_degree_bucket_edge_ordering[int64] PASSED [  1%]
tests/compute/test_basics.py::test_issue_2484[int32] PASSED              [  1%]
tests/compute/test_basics.py::test_issue_2484[int64] PASSED              [  1%]
tests/compute/test_batched_graph.py::test_batch_unbatch[int32] PASSED    [  1%]
tests/compute/test_batched_graph.py::test_batch_unbatch[int64] PASSED    [  1%]
tests/compute/test_batched_graph.py::test_batch_unbatch1[int32] PASSED   [  1%]
tests/compute/test_batched_graph.py::test_batch_unbatch1[int64] PASSED   [  1%]
tests/compute/test_batched_graph.py::test_batch_unbatch_frame[int32] PASSED [  1%]
tests/compute/test_batched_graph.py::test_batch_unbatch_frame[int64] PASSED [  1%]
tests/compute/test_batched_graph.py::test_batch_unbatch2[int32] PASSED   [  1%]
tests/compute/test_batched_graph.py::test_batch_unbatch2[int64] FAILED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g0-idtype1] FAILED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g1-idtype0] FAILED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g1-idtype1] FAILED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g2-idtype0] FAILED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g2-idtype1] FAILED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g3-idtype0] FAILED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g3-idtype1] FAILED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g4-idtype0] PASSED   [  1%]
tests/compute/test_batched_graph.py::test_batch_send_and_recv[int32] PASSED [  1%]
tests/compute/test_batched_graph.py::test_batch_send_and_recv[int64] PASSED [  1%]
tests/compute/test_batched_graph.py::test_batch_propagate[int32] PASSED  [  2%]
tests/compute/test_batched_graph.py::test_batch_propagate[int64] PASSED  [  2%]
tests/compute/test_batched_graph.py::test_batched_edge_ordering[int32] PASSED [  2%]
tests/compute/test_batched_graph.py::test_batched_edge_ordering[int64] PASSED [  2%]
tests/compute/test_batched_graph.py::test_batch_no_edge[int32] PASSED    [  2%]
tests/compute/test_batched_graph.py::test_batch_no_edge[int64] PASSED    [  2%]
tests/compute/test_batched_graph.py::test_batch_keeps_empty_data[int32] PASSED [  2%]
tests/compute/test_batched_graph.py::test_batch_keeps_empty_data[int64] PASSED [  2%]
tests/compute/test_batched_graph.py::test_set_batch_info[int32] PASSED   [  2%]
tests/compute/test_batched_graph.py::test_set_batch_info[int64] PASSED   [  2%]
tests/compute/test_batched_heterograph.py::test_topology[int32-gs0] PASSED [  2%]
tests/compute/test_batched_heterograph.py::test_topology[int32-gs1] FAILED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g4-idtype1] FAILED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g5-idtype0] FAILED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g5-idtype1] FAILED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g6-idtype0] FAILED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g6-idtype1] FAILED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g0-idtype0] PASSED [  2%]
tests/compute/test_batched_heterograph.py::test_topology[int64-gs0] PASSED [  2%]
tests/compute/test_batched_heterograph.py::test_topology[int64-gs1] PASSED [  2%]
tests/compute/test_batched_heterograph.py::test_batching_batched[int32] PASSED [  2%]
tests/compute/test_batched_heterograph.py::test_batching_batched[int64] PASSED [  2%]
tests/compute/test_batched_heterograph.py::test_features[int32] PASSED   [  2%]
tests/compute/test_batched_heterograph.py::test_features[int64] PASSED   [  2%]
tests/compute/test_batched_heterograph.py::test_empty_relation[int32] SKIPPED [  2%]
tests/compute/test_batched_heterograph.py::test_empty_relation[int64] SKIPPED [  2%]
tests/compute/test_batched_heterograph.py::test_unbatch2[int32] PASSED   [  2%]
tests/compute/test_batched_heterograph.py::test_unbatch2[int64] PASSED   [  2%]
tests/compute/test_batched_heterograph.py::test_slice_batch[int32] PASSED [  2%]
tests/compute/test_batched_heterograph.py::test_slice_batch[int64] FAILED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g0-idtype1] FAILED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g1-idtype0] FAILED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g1-idtype1] FAILED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g2-idtype0] FAILED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g2-idtype1] FAILED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g3-idtype0] PASSED [  2%]
tests/compute/test_batched_heterograph.py::test_batch_keeps_empty_data[int32] PASSED [  3%]
tests/compute/test_batched_heterograph.py::test_batch_keeps_empty_data[int64] PASSED [  3%]
tests/compute/test_batched_heterograph.py::test_batch_netypes SKIPPED    [  3%]
tests/compute/test_csrmm.py::test_csrmm[float32-int32] PASSED            [  3%]
tests/compute/test_csrmm.py::test_csrmm[float32-int64] PASSED            [  3%]
tests/compute/test_csrmm.py::test_csrmm[float64-int32] PASSED            [  3%]
tests/compute/test_csrmm.py::test_csrmm[float64-int64] PASSED            [  3%]
tests/compute/test_csrmm.py::test_csrmm_backward[1-float32-int32] PASSED [  3%]
tests/compute/test_csrmm.py::test_csrmm_backward[1-float32-int64] PASSED [  3%]
tests/compute/test_csrmm.py::test_csrmm_backward[1-float64-int32] PASSED [  3%]
tests/compute/test_csrmm.py::test_csrmm_backward[1-float64-int64] PASSED [  3%]
tests/compute/test_csrmm.py::test_csrmm_backward[2-float32-int32] PASSED [  3%]
tests/compute/test_csrmm.py::test_csrmm_backward[2-float32-int64] PASSED [  3%]
tests/compute/test_csrmm.py::test_csrmm_backward[2-float64-int32] PASSED [  3%]
tests/compute/test_csrmm.py::test_csrmm_backward[2-float64-int64] PASSED [  3%]
tests/compute/test_csrmm.py::test_csrsum[float32-int32] PASSED           [  3%]
tests/compute/test_csrmm.py::test_csrsum[float32-int64] PASSED           [  3%]
tests/compute/test_csrmm.py::test_csrsum[float64-int32] PASSED              [  7%]
tests/compute/test_heterograph.py::test_query[int64] FAILED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g3-idtype1] FAILED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g4-idtype0] FAILED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g4-idtype1] FAILED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g5-idtype0] FAILED [ 17%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g5-idtype1] FAILED [ 17%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g6-idtype0] PASSED           [  3%]
tests/compute/test_csrmm.py::test_csrsum[float64-int64] PASSED           [  3%]
tests/compute/test_csrmm.py::test_csrsum_backward[1-float32-int32] PASSED [  3%]
tests/compute/test_csrmm.py::test_csrsum_backward[1-float32-int64] PASSED [  3%]
tests/compute/test_csrmm.py::test_csrsum_backward[1-float64-int32] PASSED [  3%]
tests/compute/test_csrmm.py::test_csrsum_backward[1-float64-int64] PASSED [  3%]
tests/compute/test_csrmm.py::test_csrsum_backward[2-float32-int32] PASSED [  3%]
tests/compute/test_csrmm.py::test_csrsum_backward[2-float32-int64] PASSED [  3%]
tests/compute/test_csrmm.py::test_csrsum_backward[2-float64-int32] PASSED [  4%]
tests/compute/test_csrmm.py::test_csrsum_backward[2-float64-int64] PASSED [  4%]
tests/compute/test_csrmm.py::test_csrmask[9000-9000-float32-int32] PASSED [  4%]
tests/compute/test_csrmm.py::test_csrmask[9000-9000-float32-int64] PASSED [  4%]
tests/compute/test_csrmm.py::test_csrmask[9000-9000-float64-int32] PASSED [  4%]
tests/compute/test_csrmm.py::test_csrmask[9000-9000-float64-int64] PASSED [  4%]
tests/compute/test_csrmm.py::test_csrmask[9000-0-float32-int32] PASSED   [  4%]
tests/compute/test_csrmm.py::test_csrmask[9000-0-float32-int64] PASSED   [  4%]
tests/compute/test_csrmm.py::test_csrmask[9000-0-float64-int32] PASSED   [  4%]
tests/compute/test_csrmm.py::test_csrmask[9000-0-float64-int64] PASSED   [  4%]
tests/compute/test_csrmm.py::test_csrmask[0-9000-float32-int32] PASSED   [  4%]
tests/compute/test_csrmm.py::test_csrmask[0-9000-float32-int64] PASSED   [  4%]
tests/compute/test_csrmm.py::test_csrmask[0-9000-float64-int32] PASSED   [  4%]
tests/compute/test_csrmm.py::test_csrmask[0-9000-float64-int64] PASSED   [  4%]
tests/compute/test_csrmm.py::test_csrmask[0-0-float32-int32] PASSED      [  4%]
tests/compute/test_csrmm.py::test_csrmask[0-0-float32-int64] PASSED      [  4%]
tests/compute/test_csrmm.py::test_csrmask[0-0-float64-int32] PASSED      [  4%]
tests/compute/test_csrmm.py::test_csrmask[0-0-float64-int64] PASSED      [  4%]
tests/compute/test_csrmm.py::test_csrmask_backward[float32-int32] PASSED [  4%]
tests/compute/test_csrmm.py::test_csrmask_backward[float32-int64] PASSED [  4%]
tests/compute/test_csrmm.py::test_csrmask_backward[float64-int32] PASSED [  4%]
tests/compute/test_csrmm.py::test_csrmask_backward[float64-int64] Client [1090] waits on 172.17.0.3:37923
Machine (0) group (0) client (0) connect to server successfuly!
Client[0] in group[0] is exiting...
Server (0) shutdown.
Server is exiting...
Done sampling
Converting to homogeneous graph takes 0.001s, peak mem: 1.529 GB
Save partitions: 0.002 seconds, peak memory: 1.529 GB
There are 11990 edges in the graph and 0 edge cuts for 1 partitions.
FAILED [ 17%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g6-idtype1] FAILED [ 17%]
tests/compute/test_readout.py::test_topk[True-g0-idtype0] PASSED         [ 17%]
tests/compute/test_readout.py::test_topk[True-g0-idtype1] PASSED         [ 17%]
tests/compute/test_readout.py::test_topk[True-g1-idtype0] PASSED         [ 17%]
tests/compute/test_readout.py::test_topk[True-g1-idtype1] PASSED         [ 17%]
tests/compute/test_readout.py::test_topk[True-g2-idtype0] PASSED         [ 17%]
tests/compute/test_readout.py::test_topk[True-g2-idtype1] PASSED         [ 17%]
tests/compute/test_readout.py::test_topk[True-g3-idtype0] PASSED         [ 17%]
tests/compute/test_readout.py::test_topk[True-g3-idtype1] PASSED         [ 17%]
tests/compute/test_readout.py::test_topk[True-g4-idtype0] PASSED         [ 17%]
tests/compute/test_readout.py::test_topk[True-g4-idtype1] PASSED         [ 17%]
tests/compute/test_readout.py::test_topk[True-g5-idtype0] PASSED         [ 17%]
tests/compute/test_readout.py::test_topk[True-g5-idtype1] PASSED         [ 17%]
tests/compute/test_readout.py::test_topk[True-g6-idtype0] PASSED [  4%]
tests/compute/test_data.py::test_minigc SKIPPED (Datasets don't need...) [  4%]
tests/compute/test_data.py::test_gin SKIPPED (Datasets don't need to...) [  4%]
tests/compute/test_data.py::test_fraud SKIPPED (Datasets don't need ...) [  5%]
tests/compute/test_data.py::test_fakenews SKIPPED (Datasets don't ne...) [  5%]
tests/compute/test_data.py::test_tudataset_regression SKIPPED (Datas...) [  5%]
tests/compute/test_data.py::test_data_hash SKIPPED (Datasets don't n...) [  5%]
tests/compute/test_data.py::test_citation_graph SKIPPED (Datasets do...) [  5%]
tests/compute/test_data.py::test_gnn_benchmark SKIPPED (Datasets don...) [  5%]
tests/compute/test_data.py::test_reddit SKIPPED (Datasets don't need...) [  5%]
tests/compute/test_data.py::test_explain_syn SKIPPED (Datasets don't...) [  5%]
tests/compute/test_data.py::test_wiki_cs SKIPPED (Datasets don't nee...) [  5%]
tests/compute/test_data.py::test_yelp SKIPPED (Dataset too large to ...) [  5%]
tests/compute/test_data.py::test_flickr SKIPPED (Datasets don't need...) [  5%]
tests/compute/test_data.py::test_extract_archive SKIPPED (Datasets d...) [  5%]
tests/compute/test_data.py::test_csvdataset SKIPPED (Datasets don't ...) [  5%]
tests/compute/test_data.py::test_add_nodepred_split SKIPPED (Dataset...) [  5%]
tests/compute/test_data.py::test_as_nodepred1 SKIPPED (Datasets don'...) [  5%]
tests/compute/test_data.py::test_as_nodepred2 SKIPPED (Datasets don'...) [  5%]
tests/compute/test_data.py::test_as_nodepred_ogb SKIPPED (ogb only s...) [  5%]
tests/compute/test_data.py::test_as_linkpred SKIPPED (Datasets don't...) [  5%]
tests/compute/test_data.py::test_as_linkpred_ogb SKIPPED (ogb only s...) [  5%]
tests/compute/test_data.py::test_as_nodepred_csvdataset SKIPPED (Dat...) [  5%]
tests/compute/test_data.py::test_as_graphpred SKIPPED (Datasets don'...) [  5%]
tests/compute/test_data.py::test_as_graphpred_reprocess SKIPPED (Dat...) [  5%]
tests/compute/test_data.py::test_as_graphpred_ogb SKIPPED (ogb only ...) [  5%]
tests/compute/test_dataloader.py::test_edge_prediction_sampler[int32] PASSED [  5%]
tests/compute/test_dataloader.py::test_edge_prediction_sampler[int64] PASSED [  5%]
tests/compute/test_edge_softmax_hetero.py::test_edge_softmax_unidirectional SKIPPED [  6%]
tests/compute/test_edge_softmax_hetero.py::test_edge_softmax[int32-src-g0] SKIPPED [  6%]
tests/compute/test_edge_softmax_hetero.py::test_edge_softmax[int32-dst-g0] SKIPPED [  6%]
tests/compute/test_edge_softmax_hetero.py::test_edge_softmax[int64-src-g0] SKIPPED [  6%]
tests/compute/test_edge_softmax_hetero.py::test_edge_softmax[int64-dst-g0] SKIPPED [  6%]
tests/compute/test_filter.py::test_graph_filter PASSED                   [  6%]
tests/compute/test_filter.py::test_array_filter[int32] PASSED            [  6%]
tests/compute/test_filter.py::test_array_filter[int64] PASSED            [  6%]
tests/compute/test_frame.py::test_column_subcolumn PASSED                [  6%]
tests/compute/test_frame.py::test_serialize_deserialize_plain PASSED     [  6%]
tests/compute/test_frame.py::test_serialize_deserialize_subcolumn PASSED [  6%]
tests/compute/test_frame.py::test_serialize_deserialize_dtype PASSED     [  6%]
tests/compute/test_generators.py::test_rand_graph SKIPPED (GPU rando...) [  6%]
tests/compute/test_graph.py::test_query PASSED         [ 17%]
tests/compute/test_readout.py::test_topk[True-g6-idtype1] PASSED         [ 17%]
tests/compute/test_readout.py::test_topk[False-g0-idtype0] PASSED        [ 17%]
tests/compute/test_readout.py::test_topk[False-g0-idtype1] PASSED        [ 17%]
tests/compute/test_readout.py::test_topk[False-g1-idtype0] PASSED        [ 17%]
tests/compute/test_readout.py::test_topk[False-g1-idtype1] PASSED        [ 17%]
tests/compute/test_readout.py::test_topk[False-g2-idtype0] PASSED        [ 17%]
tests/compute/test_readout.py::test_topk[False-g2-idtype1] PASSED        [ 17%]
tests/compute/test_readout.py::test_topk[False-g3-idtype0] PASSED        [ 17%]
tests/compute/test_readout.py::test_topk[False-g3-idtype1] PASSED        [ 18%]
tests/compute/test_readout.py::test_topk[False-g4-idtype0] PASSED        [ 18%]
tests/compute/test_readout.py::test_topk[False-g4-idtype1] PASSED        [ 18%]
tests/compute/test_readout.py::test_topk[False-g5-idtype0] PASSED        [ 18%]
tests/compute/test_readout.py::test_topk[False-g5-idtype1] PASSED        [ 18%]
tests/compute/test_readout.py::test_topk[False-g6-idtype0] PASSED        [ 18%]
tests/compute/test_readout.py::test_topk[False-g6-idtype1] PASSED        [ 18%]
tests/compute/test_readout.py::test_softmax[g0-idtype0] FAILED           [ 18%]
tests/compute/test_readout.py::test_softmax[g0-idtype1] FAILED           [ 18%]
tests/compute/test_readout.py::test_softmax[g1-idtype0] FAILED           [ 18%]
tests/compute/test_readout.py::test_softmax[g1-idtype1] FAILED           [ 18%]
tests/compute/test_readout.py::test_softmax[g2-idtype0] FAILED           [ 18%]
tests/compute/test_readout.py::test_softmax[g2-idtype1] FAILED           [ 18%]
tests/compute/test_readout.py::test_softmax[g3-idtype0] FAILED           [ 18%]
tests/compute/test_readout.py::test_softmax[g3-idtype1] FAILED           [ 18%]
tests/compute/test_readout.py::test_softmax[g4-idtype0] FAILED           [ 18%]
tests/compute/test_readout.py::test_softmax[g4-idtype1] FAILED           [ 18%]
tests/compute/test_readout.py::test_softmax[g5-idtype0] FAILED           [ 18%]
tests/compute/test_readout.py::test_softmax[g5-idtype1] FAILED           [ 18%]
tests/compute/test_readout.py::test_softmax[g6-idtype0] FAILED           [ 18%]
tests/compute/test_readout.py::test_softmax[g6-idtype1] FAILED           [ 18%]
tests/compute/test_readout.py::test_broadcast[g0-idtype0] PASSED         [ 18%]
tests/compute/test_readout.py::test_broadcast[g0-idtype1] PASSED         [ 18%]
tests/compute/test_readout.py::test_broadcast[g1-idtype0] PASSED         [ 18%]
tests/compute/test_readout.py::test_broadcast[g1-idtype1] PASSED         [ 18%]
tests/compute/test_readout.py::test_broadcast[g2-idtype0] PASSED         [ 19%]
tests/compute/test_readout.py::test_broadcast[g2-idtype1] PASSED         [ 19%]
tests/compute/test_readout.py::test_broadcast[g3-idtype0] PASSED         [ 19%]
tests/compute/test_readout.py::test_broadcast[g3-idtype1] PASSED         [ 19%]
tests/compute/test_readout.py::test_broadcast[g4-idtype0] PASSED         [ 19%]
tests/compute/test_readout.py::test_broadcast[g4-idtype1] PASSED         [ 19%]
tests/compute/test_readout.py::test_broadcast[g5-idtype0] PASSED         [ 19%]
tests/compute/test_readout.py::test_broadcast[g5-idtype1] PASSED         [ 19%]
tests/compute/test_readout.py::test_broadcast[g6-idtype0] PASSED         [ 19%]
tests/compute/test_readout.py::test_broadcast[g6-idtype1] PASSED         [ 19%]
tests/compute/test_removal.py::test_node_removal[idtype0] PASSED         [ 19%]
tests/compute/test_removal.py::test_node_removal[idtype1] PASSED         [ 19%]
tests/compute/test_removal.py::test_multigraph_node_removal[idtype0] PASSED [ 19%]
tests/compute/test_removal.py::test_multigraph_node_removal[idtype1] PASSED [ 19%]
tests/compute/test_removal.py::test_multigraph_edge_removal[idtype0] PASSED [ 19%]
tests/compute/test_removal.py::test_multigraph_edge_removal[idtype1] PASSED [ 19%]
tests/compute/test_removal.py::test_edge_removal[idtype0] PASSED         [ 19%]
tests/compute/test_removal.py::test_edge_removal[idtype1] PASSED         [ 19%]
tests/compute/test_removal.py::test_node_and_edge_removal[idtype0] /root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_sampling
start graph service on server 0 for part 0
[05:31:47] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:31:47] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:13648]...
PASSED [ 19%]
tests/compute/test_removal.py::test_node_and_edge_removal[idtype1] PASSED [ 19%]
tests/compute/test_removal.py::test_node_frame[idtype0] PASSED           [ 19%]
tests/compute/test_removal.py::test_node_frame[idtype1] PASSED           [ 19%]
tests/compute/test_removal.py::test_edge_frame[idtype0] PASSED           [ 19%]
tests/compute/test_removal.py::test_edge_frame[idtype1] PASSED           [ 19%]
tests/compute/test_removal.py::test_issue1287[idtype0] PASSED            [ 20%]
tests/compute/test_removal.py::test_issue1287[idtype1] PASSED            [ 20%]
tests/compute/test_sampler.py::test_create_full PASSED                   [ 20%]
tests/compute/test_sampler.py::test_1neighbor_sampler_all PASSED         [ 20%]
tests/compute/test_sampler.py::test_1neighbor_sampler [05:31:47] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:31:47] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
PASSED             [ 20%]
tests/compute/test_sampler.py::test_prefetch_neighbor_sampler PASSED     [ 20%]
tests/compute/test_sampler.py::test_10neighbor_sampler_all PASSED                           [  6%]
tests/compute/test_graph.py::test_mutation PASSED                        [  6%]
tests/compute/test_graph.py::test_scipy_adjmat PASSED                    [  6%]
tests/compute/test_graph.py::test_incmat PASSED                          [  6%]
tests/compute/test_graph.py::test_find_edges PASSED                      [  6%]
tests/compute/test_graph.py::test_ismultigraph PASSED                    [  6%]
tests/compute/test_graph.py::test_hypersparse_query PASSED               [  6%]
tests/compute/test_graph.py::test_empty_data_initialized PASSED          [  6%]
tests/compute/test_graph.py::test_is_sorted PASSED                       [  6%]
tests/compute/test_graph.py::test_default_types PASSED                   [  6%]
tests/compute/test_graph.py::test_formats PASSED                         [  6%]
tests/compute/test_heterograph.py::test_create[int32] PASSED             [  7%]
tests/compute/test_heterograph.py::test_create[int64] PASSED             [  7%]
tests/compute/test_heterograph.py::test_create2 PASSED                   [  7%]
tests/compute/test_heterograph.py::test_query[int32] PASSED        [ 20%]
tests/compute/test_sampler.py::test_10neighbor_sampler PASSED            [ 20%]
tests/compute/test_sampler.py::test_layer_sampler PASSED                 [ 20%]
tests/compute/test_sampler.py::test_nonuniform_neighbor_sampler PASSED   [ 20%]
tests/compute/test_sampler.py::test_setseed PASSED                       [ 20%]
tests/compute/test_sampler.py::test_negative_sampler PASSED              [  7%]
tests/compute/test_heterograph.py::test_query[int64] PASSED              [  7%]
tests/compute/test_heterograph.py::test_empty_query[int32] PASSED        [  7%]
tests/compute/test_heterograph.py::test_empty_query[int64] PASSED        [  7%]
tests/compute/test_heterograph.py::test_adj[int32] PASSED                [  7%]
tests/compute/test_heterograph.py::test_adj[int64] PASSED                [  7%]
tests/compute/test_heterograph.py::test_inc[int32] PASSED                [  7%]
tests/compute/test_heterograph.py::test_inc[int64] PASSED                [  7%]
tests/compute/test_heterograph.py::test_view[int32] PASSED               [  7%]
tests/compute/test_heterograph.py::test_view[int64] PASSED               [  7%]
tests/compute/test_heterograph.py::test_view1[int32] PASSED              [  7%]
tests/compute/test_heterograph.py::test_view1[int64] PASSED              [  7%]
tests/compute/test_heterograph.py::test_flatten[int32] PASSED            [  7%]
tests/compute/test_heterograph.py::test_flatten[int64] PASSED            [  7%]
tests/compute/test_heterograph.py::test_to_device[int32] PASSED          [  7%]
tests/compute/test_heterograph.py::test_to_device[int64] PASSED          [  7%]
tests/compute/test_heterograph.py::test_to_device2[g0-int32] PASSED      [  7%]
tests/compute/test_heterograph.py::test_to_device2[g0-int64] PASSED      [  7%]
tests/compute/test_heterograph.py::test_to_device2[g1-int32] PASSED      [  7%]
tests/compute/test_heterograph.py::test_to_device2[g1-int64] PASSED      [  7%]
tests/compute/test_heterograph.py::test_pin_memory_[int32] SKIPPED (...) [  7%]
tests/compute/test_heterograph.py::test_pin_memory_[int64] SKIPPED (...) [  7%]
tests/compute/test_heterograph.py::test_convert_bound[int32] PASSED      [  8%]
tests/compute/test_heterograph.py::test_convert_bound[int64] PASSED      [  8%]
tests/compute/test_heterograph.py::test_convert[int32] PASSED            [  8%]
tests/compute/test_heterograph.py::test_convert[int64] PASSED            [  8%]
tests/compute/test_heterograph.py::test_to_homo_zero_nodes[int32] SKIPPED [  8%]
tests/compute/test_heterograph.py::test_to_homo_zero_nodes[int64] SKIPPED [  8%]
tests/compute/test_heterograph.py::test_to_homo2[int32] PASSED           [  8%]
tests/compute/test_heterograph.py::test_to_homo2[int64] PASSED           [  8%]
tests/compute/test_heterograph.py::test_invertible_conversion[int32] PASSED [  8%]
tests/compute/test_heterograph.py::test_invertible_conversion[int64] PASSED [  8%]
tests/compute/test_heterograph.py::test_metagraph_reachable[int32] PASSED [  8%]
tests/compute/test_heterograph.py::test_metagraph_reachable[int64] PASSED [  8%]
tests/compute/test_heterograph.py::test_subgraph_mask[int32] SKIPPED     [  8%]
tests/compute/test_heterograph.py::test_subgraph_mask[int64] SKIPPED     [  8%]
tests/compute/test_heterograph.py::test_subgraph[int32] PASSED           [  8%]
tests/compute/test_heterograph.py::test_subgraph[int64] PASSED           [  8%]
tests/compute/test_heterograph.py::test_apply[int32] PASSED              [  8%]
tests/compute/test_heterograph.py::test_apply[int64] PASSED              [  8%]
tests/compute/test_heterograph.py::test_level2[int32] PASSED             [  8%]
tests/compute/test_heterograph.py::test_level2[int64] Client [1090] waits on 172.17.0.3:34041
Machine (0) group (0) client (0) connect to server successfuly!
Client[0] in group[0] is exiting...
Server (0) shutdown.
Server is exiting...
Done sampling
Converting to homogeneous graph takes 0.001s, peak mem: 1.529 GB
Save partitions: 0.001 seconds, peak memory: 1.529 GB
There are 1000 edges in the graph and 0 edge cuts for 1 partitions.
PASSED             [  8%]
tests/compute/test_heterograph.py::test_more_nnz[int32] PASSED           [  8%]
tests/compute/test_heterograph.py::test_more_nnz[int64] PASSED           [  8%]
tests/compute/test_heterograph.py::test_updates[int32] PASSED            [  8%]
tests/compute/test_heterograph.py::test_updates[int64] PASSED            [  8%]
tests/compute/test_heterograph.py::test_backward[int32] PASSED           [  9%]
tests/compute/test_heterograph.py::test_backward[int64] PASSED           [  9%]
tests/compute/test_heterograph.py::test_empty_heterograph[int32] PASSED  [  9%]
tests/compute/test_heterograph.py::test_empty_heterograph[int64] PASSED  [  9%]
tests/compute/test_heterograph.py::test_types_in_function[int32] PASSED  [  9%]
tests/compute/test_heterograph.py::test_types_in_function[int64] PASSED  [  9%]
tests/compute/test_heterograph.py::test_stack_reduce[int32] PASSED       [  9%]
tests/compute/test_heterograph.py::test_stack_reduce[int64] PASSED       [  9%]
tests/compute/test_heterograph.py::test_isolated_ntype[int32] PASSED     [  9%]
tests/compute/test_heterograph.py::test_isolated_ntype[int64] PASSED     [  9%]
tests/compute/test_heterograph.py::test_ismultigraph[int32] PASSED       [  9%]
tests/compute/test_heterograph.py::test_ismultigraph[int64] PASSED       [  9%]
tests/compute/test_heterograph.py::test_graph_index_is_unibipartite[int32] PASSED [  9%]
tests/compute/test_heterograph.py::test_graph_index_is_unibipartite[int64] PASSED [  9%]
tests/compute/test_heterograph.py::test_bipartite[int32] PASSED          [  9%]
tests/compute/test_heterograph.py::test_bipartite[int64] PASSED          [  9%]
tests/compute/test_heterograph.py::test_dtype_cast[int32] PASSED         [  9%]
tests/compute/test_heterograph.py::test_dtype_cast[int64] PASSED         [  9%]
tests/compute/test_heterograph.py::test_float_cast PASSED                [  9%]
tests/compute/test_heterograph.py::test_format[int32] PASSED             [  9%]
tests/compute/test_heterograph.py::test_format[int64] /root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_sampling
start graph service on server 0 for part 0
[05:31:51] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:31:51] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:13648]...
[05:31:51] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:31:51] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
PASSED             [  9%]
tests/compute/test_heterograph.py::test_edges_order[int32] PASSED        [  9%]
tests/compute/test_heterograph.py::test_edges_order[int64] PASSED        [  9%]
tests/compute/test_heterograph.py::test_reverse[int32] PASSED            [  9%]
tests/compute/test_heterograph.py::test_reverse[int64] PASSED            [  9%]
tests/compute/test_heterograph.py::test_clone[int32] PASSED              [ 10%]
tests/compute/test_heterograph.py::test_clone[int64] PASSED              [ 10%]
tests/compute/test_heterograph.py::test_add_edges[int32] PASSED          [ 10%]
tests/compute/test_heterograph.py::test_add_edges[int64] PASSED          [ 10%]
tests/compute/test_heterograph.py::test_add_nodes[int32] PASSED          [ 10%]
tests/compute/test_heterograph.py::test_add_nodes[int64] PASSED          [ 10%]
tests/compute/test_heterograph.py::test_remove_edges[int32] SKIPPED      [ 10%]
tests/compute/test_heterograph.py::test_remove_edges[int64] SKIPPED      [ 10%]
tests/compute/test_heterograph.py::test_remove_nodes[int32] PASSED       [ 10%]
tests/compute/test_heterograph.py::test_remove_nodes[int64] PASSED       [ 10%]
tests/compute/test_heterograph.py::test_frame[int32] PASSED              [ 10%]
tests/compute/test_heterograph.py::test_frame[int64] PASSED              [ 10%]
tests/compute/test_heterograph.py::test_frame_device[int32] PASSED       [ 10%]
tests/compute/test_heterograph.py::test_frame_device[int64] PASSED       [ 10%]
tests/compute/test_heterograph.py::test_create_block[int32] PASSED       [ 10%]
tests/compute/test_heterograph.py::test_create_block[int64] PASSED       [ 10%]
tests/compute/test_heterograph.py::test_adj_sparse[coo-int32] PASSED     [ 10%]
tests/compute/test_heterograph.py::test_adj_sparse[coo-int64] PASSED     [ 10%]
tests/compute/test_heterograph.py::test_adj_sparse[csr-int32] PASSED     [ 10%]
tests/compute/test_heterograph.py::test_adj_sparse[csr-int64] PASSED     [ 10%]
tests/compute/test_heterograph.py::test_adj_sparse[csc-int32] PASSED     [ 10%]
tests/compute/test_heterograph.py::test_adj_sparse[csc-int64] PASSED     [ 10%]
tests/compute/test_heterograph.py::test_forking_pickler SKIPPED (MXN...) [ 10%]
tests/compute/test_index.py::test_dlpack PASSED                          [ 10%]
tests/compute/test_kernel.py::test_copy_src_reduce PASSED                [ 11%]
tests/compute/test_kernel.py::test_copy_edge_reduce PASSED               [ 11%]
tests/compute/test_kernel.py::test_all_binary_builtins PASSED              [  7%]
tests/compute/test_heterograph.py::test_empty_query[int32] PASSED        [  7%]
tests/compute/test_heterograph.py::test_empty_query[int64] PASSED        [  7%]
tests/compute/test_heterograph.py::test_adj[int32] PASSED                [  7%]
tests/compute/test_heterograph.py::test_adj[int64] PASSED                [  7%]
tests/compute/test_heterograph.py::test_inc[int32] PASSED                [  7%]
tests/compute/test_heterograph.py::test_inc[int64] PASSED                [  7%]
tests/compute/test_heterograph.py::test_view[int32] PASSED               [  7%]
tests/compute/test_heterograph.py::test_view[int64] PASSED               [  7%]
tests/compute/test_heterograph.py::test_view1[int32] Client [1090] waits on 172.17.0.3:58207
Machine (0) group (0) client (0) connect to server successfuly!
Client[0] in group[0] is exiting...
Done sampling
Server (0) shutdown.
Server is exiting...
Converting to homogeneous graph takes 0.001s, peak mem: 1.529 GB
Save partitions: 0.001 seconds, peak memory: 1.529 GB
There are 1000 edges in the graph and 0 edge cuts for 1 partitions.
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_sampling
start graph service on server 0 for part 0
[05:31:55] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:31:55] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:10819]...
[05:31:55] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:31:55] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
PASSED              [  7%]
tests/compute/test_heterograph.py::test_view1[int64] PASSED              [  7%]
tests/compute/test_heterograph.py::test_flatten[int32] PASSED            [  7%]
tests/compute/test_heterograph.py::test_flatten[int64] Client [1090] waits on 172.17.0.3:45951
Machine (0) group (0) client (0) connect to server successfuly!
Client[0] in group[0] is exiting...
Server (0) shutdown.
Server is exiting...
Done sampling
Converting to homogeneous graph takes 0.001s, peak mem: 1.529 GB
Save partitions: 0.001 seconds, peak memory: 1.529 GB
There are 1000 edges in the graph and 0 edge cuts for 1 partitions.
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_sampling
start graph service on server 0 for part 0
[05:31:59] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:31:59] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:19012]...
[05:31:59] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:31:59] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
PASSED            [  7%]
tests/compute/test_heterograph.py::test_to_device[int32] SKIPPED (Ne...) [  7%]
tests/compute/test_heterograph.py::test_to_device[int64] SKIPPED (Ne...) [  7%]
tests/compute/test_heterograph.py::test_to_device2[g0-int32] SKIPPED     [  7%]
tests/compute/test_heterograph.py::test_to_device2[g0-int64] SKIPPED     [  7%]
tests/compute/test_heterograph.py::test_to_device2[g1-int32] SKIPPED     [  7%]
tests/compute/test_heterograph.py::test_to_device2[g1-int64] SKIPPED     [  7%]
tests/compute/test_heterograph.py::test_pin_memory_[int32] SKIPPED (...) [  7%]
tests/compute/test_heterograph.py::test_pin_memory_[int64] SKIPPED (...) [  7%]
tests/compute/test_heterograph.py::test_convert_bound[int32] PASSED      [  8%]
tests/compute/test_heterograph.py::test_convert_bound[int64] PASSED      [  8%]
tests/compute/test_heterograph.py::test_convert[int32] PASSED            [  8%]
tests/compute/test_heterograph.py::test_convert[int64] PASSED            [  8%]
tests/compute/test_heterograph.py::test_to_homo_zero_nodes[int32] PASSED [  8%]
tests/compute/test_heterograph.py::test_to_homo_zero_nodes[int64] PASSED [  8%]
tests/compute/test_heterograph.py::test_to_homo2[int32] PASSED           [  8%]
tests/compute/test_heterograph.py::test_to_homo2[int64] PASSED           [  8%]
tests/compute/test_heterograph.py::test_invertible_conversion[int32] PASSED [  8%]
tests/compute/test_heterograph.py::test_invertible_conversion[int64] PASSED [  8%]
tests/compute/test_heterograph.py::test_metagraph_reachable[int32] PASSED [  8%]
tests/compute/test_heterograph.py::test_metagraph_reachable[int64] PASSED [  8%]
tests/compute/test_heterograph.py::test_subgraph_mask[int32] SKIPPED     [  8%]
tests/compute/test_heterograph.py::test_subgraph_mask[int64] SKIPPED     [  8%]
tests/compute/test_heterograph.py::test_subgraph[int32] PASSED            [ 11%]
tests/compute/test_kernel.py::test_mean_zero_degree[g0-int32] PASSED     [ 11%]
tests/compute/test_kernel.py::test_mean_zero_degree[g0-int64] PASSED     [ 11%]
tests/compute/test_merge.py::test_heterograph_merge[int32] PASSED        [ 11%]
tests/compute/test_merge.py::test_heterograph_merge[int64] PASSED        [ 11%]
tests/compute/test_nccl.py::test_nccl_id PASSED                          [ 11%]
tests/compute/test_nccl.py::test_nccl_sparse_push_single_remainder PASSED [ 11%]
tests/compute/test_nccl.py::test_nccl_sparse_pull_single_remainder PASSED [ 11%]
tests/compute/test_nccl.py::test_nccl_sparse_push_single_range PASSED    [ 11%]
tests/compute/test_nccl.py::test_nccl_sparse_pull_single_range PASSED    [ 11%]
tests/compute/test_nccl.py::test_nccl_support PASSED                     [ 11%]
tests/compute/test_new_update_all_hetero.py::test_unary_copy_u[int32] SKIPPED [ 11%]
tests/compute/test_new_update_all_hetero.py::test_unary_copy_u[int64] SKIPPED [ 11%]
tests/compute/test_new_update_all_hetero.py::test_unary_copy_e[int32] SKIPPED [ 11%]
tests/compute/test_new_update_all_hetero.py::test_unary_copy_e[int64] SKIPPED [ 11%]
tests/compute/test_new_update_all_hetero.py::test_binary_op[int32] SKIPPED [ 11%]
tests/compute/test_new_update_all_hetero.py::test_binary_op[int64] SKIPPED [ 11%]
tests/compute/test_partition.py::test_get_node_partition_from_book[int32] PASSED [ 11%]
tests/compute/test_partition.py::test_get_node_partition_from_book[int64] PASSED [ 11%]
tests/compute/test_pickle.py::test_pickling_index PASSED                 [ 11%]
tests/compute/test_pickle.py::test_pickling_graph_index PASSED           [ 11%]
tests/compute/test_pickle.py::test_pickling_graph[g0-int32] SKIPPED      [ 11%]
tests/compute/test_pickle.py::test_pickling_graph[g0-int64] SKIPPED      [ 11%]
tests/compute/test_pickle.py::test_pickling_graph[g1-int32] SKIPPED      [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g1-int64] SKIPPED      [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g2-int32] SKIPPED      [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g2-int64] SKIPPED      [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g3-int32] SKIPPED      [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g3-int64] SKIPPED      [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g4-int32] SKIPPED      [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g4-int64] SKIPPED      [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g5-int32] SKIPPED      [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g5-int64] SKIPPED      [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g6-int32] SKIPPED      [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g6-int64] SKIPPED      [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g7-int32] SKIPPED      [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g7-int64] SKIPPED      [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g8-int32] SKIPPED      [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g8-int64] SKIPPED      [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g9-int32] SKIPPED      [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g9-int64] SKIPPED      [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g10-int32] SKIPPED     [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g10-int64] SKIPPED     [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g11-int32] SKIPPED     [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g11-int64] SKIPPED     [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g12-int32] SKIPPED     [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g12-int64] SKIPPED     [ 12%]
tests/compute/test_pickle.py::test_pickling_batched_heterograph SKIPPED  [ 13%]
tests/compute/test_pickle.py::test_pickling_subgraph SKIPPED (GPU ed...) [ 13%]
tests/compute/test_pickle.py::test_pickling_is_pinned[int32] PASSED      [ 13%]
tests/compute/test_pickle.py::test_pickling_is_pinned[int64] PASSED      [ 13%]
tests/compute/test_pin_memory.py::test_pin_unpin PASSED                  [ 13%]
tests/compute/test_propagate.py::test_prop_nodes_bfs[int32] SKIPPED      [ 13%]
tests/compute/test_propagate.py::test_prop_nodes_bfs[int64] SKIPPED      [ 13%]
tests/compute/test_propagate.py::test_prop_edges_dfs[int32] SKIPPED      [ 13%]
tests/compute/test_propagate.py::test_prop_edges_dfs[int64] SKIPPED      [ 13%]
tests/compute/test_propagate.py::test_prop_nodes_topo[int32] SKIPPED     [ 13%]
tests/compute/test_propagate.py::test_prop_nodes_topo[int64] SKIPPED     [ 13%]
tests/compute/test_random.py::test_random_choice SKIPPED (GPU random...) [ 13%]
tests/compute/test_readout.py::test_sum_case1[int32] PASSED              [ 13%]
tests/compute/test_readout.py::test_sum_case1[int64] PASSED              [ 13%]
tests/compute/test_readout.py::test_reduce_readout[sum-g0-int32] PASSED  [ 13%]
tests/compute/test_readout.py::test_reduce_readout[sum-g0-int64] PASSED  [ 13%]
tests/compute/test_readout.py::test_reduce_readout[sum-g1-int32] PASSED  [ 13%]
tests/compute/test_readout.py::test_reduce_readout[sum-g1-int64] PASSED  [ 13%]
tests/compute/test_readout.py::test_reduce_readout[sum-g2-int32] PASSED  [ 13%]
tests/compute/test_readout.py::test_reduce_readout[sum-g2-int64] PASSED  [ 13%]
tests/compute/test_readout.py::test_reduce_readout[sum-g3-int32] PASSED  [ 13%]
tests/compute/test_readout.py::test_reduce_readout[sum-g3-int64] PASSED  [ 13%]
tests/compute/test_readout.py::test_reduce_readout[sum-g4-int32] PASSED  [ 13%]
tests/compute/test_readout.py::test_reduce_readout[sum-g4-int64] PASSED  [ 13%]
tests/compute/test_readout.py::test_reduce_readout[sum-g5-int32] PASSED  [ 13%]
tests/compute/test_readout.py::test_reduce_readout[sum-g5-int64] PASSED  [ 14%]
tests/compute/test_readout.py::test_reduce_readout[sum-g6-int32] PASSED  [ 14%]
tests/compute/test_readout.py::test_reduce_readout[sum-g6-int64] PASSED  [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g0-int32] PASSED  [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g0-int64] PASSED  [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g1-int32] PASSED  [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g1-int64] PASSED  [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g2-int32] PASSED  [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g2-int64] PASSED  [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g3-int32] PASSED  [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g3-int64] PASSED  [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g4-int32] PASSED  [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g4-int64] PASSED  [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g5-int32] PASSED  [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g5-int64] PASSED  [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g6-int32] PASSED  [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g6-int64] PASSED  [ 14%]
tests/compute/test_readout.py::test_reduce_readout[mean-g0-int32] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[mean-g0-int64] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[mean-g1-int32] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[mean-g1-int64] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[mean-g2-int32] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[mean-g2-int64] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[mean-g3-int32] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[mean-g3-int64] PASSED [ 15%]
tests/compute/test_readout.py::test_reduce_readout[mean-g4-int32] PASSED [ 15%]
tests/compute/test_readout.py::test_reduce_readout[mean-g4-int64] PASSED [ 15%]
tests/compute/test_readout.py::test_reduce_readout[mean-g5-int32] PASSED [ 15%]
tests/compute/test_readout.py::test_reduce_readout[mean-g5-int64] PASSED [ 15%]
tests/compute/test_readout.py::test_reduce_readout[mean-g6-int32] PASSED [ 15%]
tests/compute/test_readout.py::test_reduce_readout[mean-g6-int64] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g0-int32] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g0-int64] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g1-int32] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g1-int64] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g2-int32] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g2-int64] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g3-int32] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g3-int64] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g4-int32] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g4-int64] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g5-int32] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g5-int64] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g6-int32] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g6-int64] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g0-int32] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g0-int64] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g1-int32] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g1-int64] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g2-int32] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g2-int64] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g3-int32] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g3-int64] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g4-int32] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g4-int64] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g5-int32] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g5-int64] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g6-int32] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g6-int64] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g0-int32] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g0-int64] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g1-int32] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g1-int64] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g2-int32] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g2-int64] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g3-int32] PASSED           [  8%]
tests/compute/test_heterograph.py::test_subgraph[int64] PASSED           [  8%]
tests/compute/test_heterograph.py::test_apply[int32] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g3-int64] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g4-int32] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g4-int64] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g5-int32] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g5-int64] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g6-int32] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g6-int64] PASSED [ 16%]
tests/compute/test_readout.py::test_topk[True-g0-int32] PASSED           [ 16%]
tests/compute/test_readout.py::test_topk[True-g0-int64] PASSED           [ 17%]
tests/compute/test_readout.py::test_topk[True-g1-int32] PASSED           [ 17%]
tests/compute/test_readout.py::test_topk[True-g1-int64] PASSED           [ 17%]
tests/compute/test_readout.py::test_topk[True-g2-int32] PASSED           [ 17%]
tests/compute/test_readout.py::test_topk[True-g2-int64] Client [1090] waits on 172.17.0.3:60369
Machine (0) group (0) client (0) connect to server successfuly!
Client[0] in group[0] is exiting...
Server (0) shutdown.Done sampling

Server is exiting...
PASSED              [  8%]
tests/compute/test_heterograph.py::test_apply[int64] PASSED              [  8%]
tests/compute/test_heterograph.py::test_level2[int32] PASSED           [ 17%]
tests/compute/test_readout.py::test_topk[True-g3-int32] PASSED           [ 17%]
tests/compute/test_readout.py::test_topk[True-g3-int64] PASSED           [ 17%]
tests/compute/test_readout.py::test_topk[True-g4-int32] PASSED           [ 17%]
tests/compute/test_readout.py::test_topk[True-g4-int64] PASSED           [ 17%]
tests/compute/test_readout.py::test_topk[True-g5-int32] PASSED           [ 17%]
tests/compute/test_readout.py::test_topk[True-g5-int64] PASSED           [ 17%]
tests/compute/test_readout.py::test_topk[True-g6-int32] PASSED           [ 17%]
tests/compute/test_readout.py::test_topk[True-g6-int64] PASSED           [ 17%]
tests/compute/test_readout.py::test_topk[False-g0-int32] PASSED          [ 17%]
tests/compute/test_readout.py::test_topk[False-g0-int64] PASSED          [ 17%]
tests/compute/test_readout.py::test_topk[False-g1-int32] PASSED          [ 17%]
tests/compute/test_readout.py::test_topk[False-g1-int64] PASSED          [ 17%]
tests/compute/test_readout.py::test_topk[False-g2-int32] Converting to homogeneous graph takes 0.001s, peak mem: 1.529 GB
Save partitions: 0.001 seconds, peak memory: 1.529 GB
There are 1000 edges in the graph and 0 edge cuts for 1 partitions.
PASSED          [ 17%]
tests/compute/test_readout.py::test_topk[False-g2-int64] PASSED          [ 17%]
tests/compute/test_readout.py::test_topk[False-g3-int32] PASSED          [ 17%]
tests/compute/test_readout.py::test_topk[False-g3-int64] PASSED          [ 17%]
tests/compute/test_readout.py::test_topk[False-g4-int32] PASSED          [ 17%]
tests/compute/test_readout.py::test_topk[False-g4-int64] PASSED          [ 17%]
tests/compute/test_readout.py::test_topk[False-g5-int32] PASSED          [ 17%]
tests/compute/test_readout.py::test_topk[False-g5-int64] PASSED          [ 18%]
tests/compute/test_readout.py::test_topk[False-g6-int32] PASSED          [ 18%]
tests/compute/test_readout.py::test_topk[False-g6-int64] PASSED          [ 18%]
tests/compute/test_readout.py::test_softmax[g0-int32] PASSED             [ 18%]
tests/compute/test_readout.py::test_softmax[g0-int64] PASSED             [ 18%]
tests/compute/test_readout.py::test_softmax[g1-int32] PASSED             [ 18%]
tests/compute/test_readout.py::test_softmax[g1-int64] PASSED             [ 18%]
tests/compute/test_readout.py::test_softmax[g2-int32] PASSED             [ 18%]
tests/compute/test_readout.py::test_softmax[g2-int64] PASSED             [ 18%]
tests/compute/test_readout.py::test_softmax[g3-int32] PASSED             [ 18%]
tests/compute/test_readout.py::test_softmax[g3-int64] PASSED             [ 18%]
tests/compute/test_readout.py::test_softmax[g4-int32] PASSED             [ 18%]
tests/compute/test_readout.py::test_softmax[g4-int64] PASSED             [ 18%]
tests/compute/test_readout.py::test_softmax[g5-int32] PASSED             [ 18%]
tests/compute/test_readout.py::test_softmax[g5-int64] PASSED             [ 18%]
tests/compute/test_readout.py::test_softmax[g6-int32] PASSED             [ 18%]
tests/compute/test_readout.py::test_softmax[g6-int64] PASSED             [ 18%]
tests/compute/test_readout.py::test_broadcast[g0-int32] PASSED           [ 18%]
tests/compute/test_readout.py::test_broadcast[g0-int64] PASSED           [ 18%]
tests/compute/test_readout.py::test_broadcast[g1-int32] PASSED           [ 18%]
tests/compute/test_readout.py::test_broadcast[g1-int64] PASSED           [ 18%]
tests/compute/test_readout.py::test_broadcast[g2-int32] PASSED           [ 18%]
tests/compute/test_readout.py::test_broadcast[g2-int64] PASSED           [ 18%]
tests/compute/test_readout.py::test_broadcast[g3-int32] PASSED           [ 18%]
tests/compute/test_readout.py::test_broadcast[g3-int64] PASSED           [ 18%]
tests/compute/test_readout.py::test_broadcast[g4-int32] PASSED           [ 19%]
tests/compute/test_readout.py::test_broadcast[g4-int64] PASSED           [ 19%]
tests/compute/test_readout.py::test_broadcast[g5-int32] PASSED           [ 19%]
tests/compute/test_readout.py::test_broadcast[g5-int64] PASSED           [ 19%]
tests/compute/test_readout.py::test_broadcast[g6-int32] PASSED           [ 19%]
tests/compute/test_readout.py::test_broadcast[g6-int64] PASSED           [ 19%]
tests/compute/test_removal.py::test_node_removal[int32] PASSED           [ 19%]
tests/compute/test_removal.py::test_node_removal[int64] PASSED           [ 19%]
tests/compute/test_removal.py::test_multigraph_node_removal[int32] PASSED [ 19%]
tests/compute/test_removal.py::test_multigraph_node_removal[int64] PASSED [ 19%]
tests/compute/test_removal.py::test_multigraph_edge_removal[int32] PASSED [ 19%]
tests/compute/test_removal.py::test_multigraph_edge_removal[int64] PASSED [ 19%]
tests/compute/test_removal.py::test_edge_removal[int32] PASSED           [ 19%]
tests/compute/test_removal.py::test_edge_removal[int64] PASSED           [ 19%]
tests/compute/test_removal.py::test_node_and_edge_removal[int32] /root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_sampling
start graph service on server 0 for part 0
[05:32:04] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:32:04] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:18024]...
[05:32:04] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:32:04] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
PASSED  [ 19%]
tests/compute/test_removal.py::test_node_and_edge_removal[int64] PASSED  [ 19%]
tests/compute/test_removal.py::test_node_frame[int32] PASSED             [ 19%]
tests/compute/test_removal.py::test_node_frame[int64] PASSED             [ 19%]
tests/compute/test_removal.py::test_edge_frame[int32] PASSED             [ 19%]
tests/compute/test_removal.py::test_edge_frame[int64] PASSED             [ 19%]
tests/compute/test_removal.py::test_issue1287[int32] PASSED              [ 19%]
tests/compute/test_removal.py::test_issue1287[int64] PASSED              [ 19%]
tests/compute/test_sampler.py::test_create_full PASSED                   [ 19%]
tests/compute/test_sampler.py::test_1neighbor_sampler_all PASSED         [ 19%]
tests/compute/test_sampler.py::test_1neighbor_sampler PASSED             [ 20%]
tests/compute/test_sampler.py::test_prefetch_neighbor_sampler PASSED     [ 20%]
tests/compute/test_sampler.py::test_10neighbor_sampler_all PASSED        [ 20%]
tests/compute/test_sampler.py::test_10neighbor_sampler PASSED            [ 20%]
tests/compute/test_sampler.py::test_layer_sampler PASSED                 [ 20%]
tests/compute/test_sampler.py::test_nonuniform_neighbor_sampler PASSED   [ 20%]
tests/compute/test_sampler.py::test_setseed PASSED                       [ 20%]
tests/compute/test_sampler.py::test_negative_sampler PASSED             [  8%]
tests/compute/test_heterograph.py::test_level2[int64] Client [1090] waits on 172.17.0.3:56707
Machine (0) group (0) client (0) connect to server successfuly!
Client[0] in group[0] is exiting...
Server (0) shutdown.
Server is exiting...
Done sampling
PASSED [ 23%]
tests/distributed/test_distributed_sampling.py::test_rpc_sampling_shuffle[2]   NumNodes: 2708
  NumEdges: 10556
  NumFeats: 1433
  NumClasses: 7
  NumTrainingSamples: 140
  NumValidationSamples: 500
  NumTestSamples: 1000
Done loading data from cached files.
Converting to homogeneous graph takes 0.001s, peak mem: 1.529 GB
Convert a graph into a bidirected graph: 0.001 seconds, peak memory: 1.529 GB
Construct multi-constraint weights: 0.000 seconds, peak memory: 1.529 GB
[05:32:07] /root/jenkins/workspace/dgl_PR-4648/src/graph/transform/metis_partition_hetero.cc:87: Partition a graph with 2708 nodes and 10556 edges into 2 parts and get 260 edge cuts
Metis partitioning: 0.002 seconds, peak memory: 1.529 GB
Assigning nodes to METIS partitions takes 0.003s, peak mem: 1.529 GB
Reshuffle nodes and edges: 0.001 seconds
Split the graph: 0.002 seconds
Construct subgraphs: 0.003 seconds
Splitting the graph into partitions takes 0.006s, peak mem: 1.529 GB
part 0 has 1512 nodes and 1354 are inside the partition
part 0 has 5630 edges and 5370 are inside the partition
part 1 has 1552 nodes and 1354 are inside the partition
part 1 has 5446 edges and 5186 are inside the partition
Save partitions: 0.020 seconds, peak memory: 1.529 GB
There are 10556 edges in the graph and 0 edge cuts for 2 partitions.
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_sampling
start graph service on server 0 for part 0
[05:32:08] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:32:08] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:17314]...
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_sampling
start graph service on server 1 for part 1
[05:32:09] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:32:09] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:17316]...
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
[05:32:10] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:32:10] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
PASSED             [  8%]
tests/compute/test_heterograph.py::test_more_nnz[int32] SKIPPED (Nee...) [  8%]
tests/compute/test_heterograph.py::test_more_nnz[int64] SKIPPED (Nee...) [  8%]
tests/compute/test_heterograph.py::test_updates[int32] PASSED            [  8%]
tests/compute/test_heterograph.py::test_updates[int64] Client [1509] waits on 172.17.0.3:59519
Machine (0) group (0) client (0) connect to server successfuly!
Client[0] in group[0] is exiting...
Server (0) shutdown.
Server is exiting...
PASSED            [  8%]
tests/compute/test_heterograph.py::test_backward[int32] PASSED           [  9%]
tests/compute/test_heterograph.py::test_backward[int64] PASSED           [  9%]
tests/compute/test_heterograph.py::test_empty_heterograph[int32] PASSED  [  9%]
tests/compute/test_heterograph.py::test_empty_heterograph[int64] PASSED  [  9%]
tests/compute/test_heterograph.py::test_types_in_function[int32] Server (1) shutdown.
Server is exiting...
PASSED              [ 20%]
tests/compute/test_sampling.py::test_non_uniform_random_walk[True] SKIPPED [ 20%]
tests/compute/test_sampling.py::test_non_uniform_random_walk[False] PASSED [ 20%]
tests/compute/test_sampling.py::test_uniform_random_walk[True] SKIPPED   [ 20%]
tests/compute/test_sampling.py::test_uniform_random_walk[False] PASSED   [ 20%]
tests/compute/test_sampling.py::test_node2vec PASSED                     [ 20%]
tests/compute/test_sampling.py::test_pack_traces PASSED                  [ 20%]
tests/compute/test_sampling.py::test_pinsage_sampling[True] SKIPPED      [ 20%]
tests/compute/test_sampling.py::test_pinsage_sampling[False] PASSED      [ 20%]
tests/compute/test_sampling.py::test_sample_neighbors_noprob PASSED      [ 20%]
tests/compute/test_sampling.py::test_sample_neighbors_prob PASSED        [ 20%]
tests/compute/test_sampling.py::test_sample_neighbors_outedge Converting to homogeneous graph takes 0.001s, peak mem: 1.529 GB
Convert a graph into a bidirected graph: 0.000 seconds, peak memory: 1.529 GB
Construct multi-constraint weights: 0.000 seconds, peak memory: 1.529 GB
[05:32:14] /root/jenkins/workspace/dgl_PR-4648/src/graph/transform/metis_partition_hetero.cc:87: Partition a graph with 3030 nodes and 6120 edges into 2 parts and get 204 edge cuts
Metis partitioning: 0.003 seconds, peak memory: 1.529 GB
Assigning nodes to METIS partitions takes 0.004s, peak mem: 1.529 GB
Reshuffle nodes and edges: 0.001 seconds
Split the graph: 0.002 seconds
Construct subgraphs: 0.002 seconds
Splitting the graph into partitions takes 0.005s, peak mem: 1.529 GB
part 0 has 587 nodes of type n1 and 519 are inside the partition
part 0 has 574 nodes of type n2 and 511 are inside the partition
part 0 has 585 nodes of type n3 and 525 are inside the partition
part 0 has 565 edges of type r12 and 531 are inside the partition
part 0 has 571 edges of type r13 and 535 are inside the partition
part 0 has 554 edges of type r23 and 527 are inside the partition
part 1 has 556 nodes of type n1 and 491 are inside the partition
part 1 has 549 nodes of type n2 and 489 are inside the partition
part 1 has 559 nodes of type n3 and 495 are inside the partition
part 1 has 515 edges of type r12 and 479 are inside the partition
part 1 has 530 edges of type r13 and 495 are inside the partition
part 1 has 529 edges of type r23 and 493 are inside the partition
Save partitions: 0.005 seconds, peak memory: 1.529 GB
There are 3060 edges in the graph and 0 edge cuts for 2 partitions.
PASSED  [  9%]
tests/compute/test_heterograph.py::test_types_in_function[int64] PASSED  [  9%]
tests/compute/test_heterograph.py::test_stack_reduce[int32] PASSED       [  9%]
tests/compute/test_heterograph.py::test_stack_reduce[int64] PASSED       [  9%]
tests/compute/test_heterograph.py::test_isolated_ntype[int32] /root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_sampling
start graph service on server 0 for part 0
[05:32:15] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:32:15] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:14572]...
PASSED     [  9%]
tests/compute/test_heterograph.py::test_isolated_ntype[int64] PASSED     [  9%]
tests/compute/test_heterograph.py::test_ismultigraph[int32] PASSED       [  9%]
tests/compute/test_heterograph.py::test_ismultigraph[int64] PASSED       [  9%]
tests/compute/test_heterograph.py::test_graph_index_is_unibipartite[int32] PASSED [  9%]
tests/compute/test_heterograph.py::test_graph_index_is_unibipartite[int64] PASSED [  9%]
tests/compute/test_heterograph.py::test_bipartite[int32] PASSED          [  9%]
tests/compute/test_heterograph.py::test_bipartite[int64] PASSED          [  9%]
tests/compute/test_heterograph.py::test_dtype_cast[int32] PASSED         [  9%]
tests/compute/test_heterograph.py::test_dtype_cast[int64] PASSED         [  9%]
tests/compute/test_heterograph.py::test_float_cast /root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_sampling
start graph service on server 1 for part 1
[05:32:16] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:32:16] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:14574]...
[05:32:17] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:32:17] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
PASSED                [  9%]
tests/compute/test_heterograph.py::test_format[int32] PASSED             [  9%]
tests/compute/test_heterograph.py::test_format[int64] PASSED             [  9%]
tests/compute/test_heterograph.py::test_edges_order[int32] PASSED        [  9%]
tests/compute/test_heterograph.py::test_edges_order[int64] PASSED        [  9%]
tests/compute/test_heterograph.py::test_reverse[int32] PASSED            [  9%]
tests/compute/test_heterograph.py::test_reverse[int64] PASSED     [ 20%]
tests/compute/test_sampling.py::test_sample_neighbors_topk PASSED            [  9%]
tests/compute/test_heterograph.py::test_clone[int32] PASSED              [ 10%]
tests/compute/test_heterograph.py::test_clone[int64] PASSED              [ 10%]
tests/compute/test_heterograph.py::test_add_edges[int32] PASSED        [ 20%]
tests/compute/test_sampling.py::test_sample_neighbors_topk_outedge PASSED [ 20%]
tests/compute/test_sampling.py::test_sample_neighbors_with_0deg PASSED   [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_biased_homogeneous PASSED          [ 10%]
tests/compute/test_heterograph.py::test_add_edges[int64] PASSED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_biased_bipartite PASSED          [ 10%]
tests/compute/test_heterograph.py::test_add_nodes[int32] PASSED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_homogeneous[False-in-coo] PASSED          [ 10%]
tests/compute/test_heterograph.py::test_add_nodes[int64] PASSED          [ 10%]
tests/compute/test_heterograph.py::test_remove_edges[int32] SKIPPED      [ 10%]
tests/compute/test_heterograph.py::test_remove_edges[int64] SKIPPED      [ 10%]
tests/compute/test_heterograph.py::test_remove_nodes[int32] PASSED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_homogeneous[False-in-csr] PASSED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_homogeneous[False-in-csc] PASSED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_homogeneous[False-out-coo] PASSED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_homogeneous[False-out-csr] PASSED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_homogeneous[False-out-csc] PASSED       [ 10%]
tests/compute/test_heterograph.py::test_remove_nodes[int64] PASSED       [ 10%]
tests/compute/test_heterograph.py::test_frame[int32] PASSED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_homogeneous[True-in-coo] PASSED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_homogeneous[True-in-csr] PASSED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_homogeneous[True-in-csc] PASSED              [ 10%]
tests/compute/test_heterograph.py::test_frame[int64] PASSED              [ 10%]
tests/compute/test_heterograph.py::test_frame_device[int32] SKIPPED      [ 10%]
tests/compute/test_heterograph.py::test_frame_device[int64] SKIPPED      [ 10%]
tests/compute/test_heterograph.py::test_create_block[int32] PASSED       [ 10%]
tests/compute/test_heterograph.py::test_create_block[int64] PASSED       [ 10%]
tests/compute/test_heterograph.py::test_adj_sparse[coo-int32] PASSED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_homogeneous[True-out-coo] PASSED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_homogeneous[True-out-csr] PASSED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_homogeneous[True-out-csc] PASSED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_sorted_homogeneous[in-csr] PASSED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_sorted_homogeneous[in-csc] PASSED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_sorted_homogeneous[out-csr] PASSED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_sorted_homogeneous[out-csc] PASSED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_exclude_edges_heteroG[int32] PASSED     [ 10%]
tests/compute/test_heterograph.py::test_adj_sparse[coo-int64] PASSED     [ 10%]
tests/compute/test_heterograph.py::test_adj_sparse[csr-int32] PASSED     [ 10%]
tests/compute/test_heterograph.py::test_adj_sparse[csr-int64] PASSED     [ 10%]
tests/compute/test_heterograph.py::test_adj_sparse[csc-int32] PASSED     [ 10%]
tests/compute/test_heterograph.py::test_adj_sparse[csc-int64] PASSED     [ 10%]
tests/compute/test_heterograph.py::test_forking_pickler SKIPPED (MXN...) [ 10%]
tests/compute/test_index.py::test_dlpack PASSED                          [ 10%]
tests/compute/test_kernel.py::test_copy_src_reduce PASSED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_exclude_edges_heteroG[int64] PASSED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_exclude_edges_homoG[int32] PASSED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_exclude_edges_homoG[int64] PASSED [ 21%]
tests/compute/test_sampling.py::test_global_uniform_negative_sampling[int32] PASSED [ 21%]
tests/compute/test_sampling.py::test_global_uniform_negative_sampling[int64] PASSED [ 22%]
tests/compute/test_serialize.py::test_graph_serialize_with_feature[True] PASSED [ 22%]
tests/compute/test_serialize.py::test_graph_serialize_with_feature[False] Client [1090] waits on 172.17.0.3:39975
Machine (0) group (0) client (0) connect to server successfuly!
Client[0] in group[0] is exiting...
PASSED [ 22%]
tests/compute/test_serialize.py::test_graph_serialize_without_feature[True] PASSED [ 22%]
tests/compute/test_serialize.py::test_graph_serialize_without_feature[False] PASSED [ 22%]
tests/compute/test_serialize.py::test_graph_serialize_with_labels[True] PASSED [ 22%]
tests/compute/test_serialize.py::test_graph_serialize_with_labels[False] PASSED [ 22%]
tests/compute/test_serialize.py::test_serialize_tensors PASSED           [ 22%]
tests/compute/test_serialize.py::test_serialize_empty_dict PASSED        [ 22%]
tests/compute/test_serialize.py::test_load_old_files1 PASSED             [ 22%]
tests/compute/test_serialize.py::test_load_old_files2 PASSED             [ 22%]
tests/compute/test_serialize.py::test_deserialize_old_heterograph_file PASSED [ 22%]
tests/compute/test_serialize.py::test_serialize_heterograph PASSED       [ 22%]
tests/compute/test_serialize.py::test_serialize_heterograph_s3 SKIPPED   [ 22%]
tests/compute/test_shared_mem.py::test_single_process[idtype0] PASSED    [ 22%]
tests/compute/test_shared_mem.py::test_single_process[idtype1] PASSED    [ 22%]
tests/compute/test_shared_mem.py::test_multi_process[idtype0] PASSED     [ 22%]
tests/compute/test_shared_mem.py::test_multi_process[idtype1] PASSED     [ 22%]
tests/compute/test_shared_mem.py::test_copy_from_gpu SKIPPED (Need g...) [ 22%]
tests/compute/test_sort.py::test_sort_with_tag[idtype0] PASSED           [ 22%]
tests/compute/test_sort.py::test_sort_with_tag[idtype1] Server (1) shutdown.Server (0) shutdown.

Server is exiting...
Server is exiting...
Done sampling
PASSED           [ 22%]
tests/compute/test_sort.py::test_sort_with_tag_bipartite[idtype0] PASSED [ 22%]
tests/compute/test_sort.py::test_sort_with_tag_bipartite[idtype1] PASSED [ 22%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-add-shp0-g0] Converting to homogeneous graph takes 0.001s, peak mem: 1.529 GB
Convert a graph into a bidirected graph: 0.000 seconds, peak memory: 1.529 GB
Construct multi-constraint weights: 0.000 seconds, peak memory: 1.529 GB
[05:32:21] /root/jenkins/workspace/dgl_PR-4648/src/graph/transform/metis_partition_hetero.cc:87: Partition a graph with 3030 nodes and 6000 edges into 2 parts and get 199 edge cuts
Metis partitioning: 0.003 seconds, peak memory: 1.529 GB
Assigning nodes to METIS partitions takes 0.003s, peak mem: 1.529 GB
Reshuffle nodes and edges: 0.001 seconds
Split the graph: 0.002 seconds
Construct subgraphs: 0.002 seconds
Splitting the graph into partitions takes 0.005s, peak mem: 1.529 GB
part 0 has 555 nodes of type n1 and 493 are inside the partition
part 0 has 550 nodes of type n2 and 486 are inside the partition
part 0 has 556 nodes of type n3 and 496 are inside the partition
part 0 has 508 edges of type r12 and 473 are inside the partition
part 0 has 517 edges of type r13 and 486 are inside the partition
part 0 has 494 edges of type r23 and 462 are inside the partition
part 1 has 577 nodes of type n1 and 517 are inside the partition
part 1 has 567 nodes of type n2 and 514 are inside the partition
part 1 has 594 nodes of type n3 and 524 are inside the partition
part 1 has 541 edges of type r12 and 517 are inside the partition
part 1 has 565 edges of type r13 and 524 are inside the partition
part 1 has 574 edges of type r23 and 538 are inside the partition
Save partitions: 0.002 seconds, peak memory: 1.529 GB
There are 3000 edges in the graph and 0 edge cuts for 2 partitions.
FAILED  [ 22%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-add-shp0-g1] FAILED  [ 22%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-add-shp1-g0] FAILED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-add-shp1-g1] FAILED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-add-shp2-g0] FAILED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-add-shp2-g1] FAILED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-add-shp3-g0] FAILED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-add-shp3-g1] FAILED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-add-shp4-g0] FAILED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-add-shp4-g1] FAILED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-add-shp5-g0] FAILED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-add-shp5-g1] FAILED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-sub-shp0-g0] FAILED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-sub-shp0-g1] FAILED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-sub-shp1-g0] FAILED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-sub-shp1-g1] FAILED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-sub-shp2-g0] FAILED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-sub-shp2-g1] FAILED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-sub-shp3-g0] FAILED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-sub-shp3-g1] FAILED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-sub-shp4-g0] FAILED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-sub-shp4-g1] FAILED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-sub-shp5-g0] /root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_sampling
start graph service on server 0 for part 0
[05:32:22] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:32:22] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:19012]...
FAILED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-sub-shp5-g1] FAILED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-mul-shp0-g0] FAILED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-mul-shp0-g1] FAILED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-mul-shp1-g0] FAILED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-mul-shp1-g1] FAILED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-mul-shp2-g0] FAILED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-mul-shp2-g1] FAILED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-mul-shp3-g0] FAILED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-mul-shp3-g1] FAILED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-mul-shp4-g0] FAILED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-mul-shp4-g1] FAILED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-mul-shp5-g0] FAILED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-mul-shp5-g1] FAILED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-div-shp0-g0] FAILED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-div-shp0-g1] FAILED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-div-shp1-g0] FAILED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-div-shp1-g1] FAILED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-div-shp2-g0] FAILED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-div-shp2-g1] FAILED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-div-shp3-g0] /root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_sampling
start graph service on server 1 for part 1
[05:32:23] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:32:23] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:19014]...
[05:32:23] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:32:23] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
FAILED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-div-shp3-g1] FAILED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-div-shp4-g0] FAILED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-div-shp4-g1] FAILED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-div-shp5-g0] FAILED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-div-shp5-g1] FAILED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_lhs-shp0-g0] FAILED [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_lhs-shp0-g1] FAILED [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_lhs-shp1-g0] FAILED [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_lhs-shp1-g1] FAILED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_lhs-shp2-g0] FAILED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_lhs-shp2-g1] FAILED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_lhs-shp3-g0] FAILED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_lhs-shp3-g1] FAILED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_lhs-shp4-g0] FAILED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_lhs-shp4-g1] FAILED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_lhs-shp5-g0] FAILED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_lhs-shp5-g1] FAILED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_rhs-shp0-g0] FAILED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_rhs-shp0-g1] FAILED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_rhs-shp1-g0] FAILED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_rhs-shp1-g1] FAILED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_rhs-shp2-g0] FAILED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_rhs-shp2-g1] FAILED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_rhs-shp3-g0] FAILED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_rhs-shp3-g1] FAILED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_rhs-shp4-g0] FAILED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_rhs-shp4-g1] FAILED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_rhs-shp5-g0] FAILED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_rhs-shp5-g1] FAILED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-add-shp0-g0] FAILED  [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-add-shp0-g1] FAILED  [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-add-shp1-g0] FAILED  [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-add-shp1-g1] FAILED  [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-add-shp2-g0] FAILED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-add-shp2-g1] PASSED                [ 11%]
tests/compute/test_kernel.py::test_copy_edge_reduce FAILED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-add-shp3-g0] FAILED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-add-shp3-g1] FAILED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-add-shp4-g0] FAILED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-add-shp4-g1] FAILED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-add-shp5-g0] FAILED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-add-shp5-g1] FAILED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-sub-shp0-g0] FAILED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-sub-shp0-g1] FAILED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-sub-shp1-g0] FAILED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-sub-shp1-g1] FAILED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-sub-shp2-g0] FAILED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-sub-shp2-g1] FAILED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-sub-shp3-g0] FAILED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-sub-shp3-g1] FAILED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-sub-shp4-g0] FAILED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-sub-shp4-g1] FAILED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-sub-shp5-g0] FAILED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-sub-shp5-g1] FAILED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-mul-shp0-g0] FAILED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-mul-shp0-g1] FAILED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-mul-shp1-g0] FAILED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-mul-shp1-g1] FAILED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-mul-shp2-g0] FAILED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-mul-shp2-g1] FAILED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-mul-shp3-g0] FAILED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-mul-shp3-g1] FAILED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-mul-shp4-g0] FAILED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-mul-shp4-g1] FAILED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-mul-shp5-g0] FAILED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-mul-shp5-g1] FAILED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-div-shp0-g0] FAILED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-div-shp0-g1] FAILED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-div-shp1-g0] FAILED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-div-shp1-g1] FAILED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-div-shp2-g0] FAILED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-div-shp2-g1] Client [1090] waits on 172.17.0.3:56167
Machine (0) group (0) client (0) connect to server successfuly!
Client[0] in group[0] is exiting...
Server (1) shutdown.Done sampling

Server is exiting...Server (0) shutdown.

Server is exiting...
Converting to homogeneous graph takes 0.001s, peak mem: 1.529 GB
Convert a graph into a bidirected graph: 0.001 seconds, peak memory: 1.529 GB
Construct multi-constraint weights: 0.000 seconds, peak memory: 1.529 GB
[05:32:26] /root/jenkins/workspace/dgl_PR-4648/src/graph/transform/metis_partition_hetero.cc:87: Partition a graph with 630 nodes and 26440 edges into 2 parts and get 5478 edge cuts
Metis partitioning: 0.005 seconds, peak memory: 1.529 GB
Assigning nodes to METIS partitions takes 0.006s, peak mem: 1.529 GB
Reshuffle nodes and edges: 0.001 seconds
Split the graph: 0.001 seconds
Construct subgraphs: 0.003 seconds
Splitting the graph into partitions takes 0.006s, peak mem: 1.529 GB
part 0 has 210 nodes of type n1 and 108 are inside the partition
part 0 has 200 nodes of type n2 and 100 are inside the partition
part 0 has 220 nodes of type n3 and 113 are inside the partition
part 0 has 2989 edges of type r12 and 2109 are inside the partition
part 0 has 3328 edges of type r13 and 2342 are inside the partition
part 0 has 3166 edges of type r23 and 2217 are inside the partition
part 1 has 210 nodes of type n1 and 102 are inside the partition
part 1 has 200 nodes of type n2 and 100 are inside the partition
part 1 has 220 nodes of type n3 and 107 are inside the partition
part 1 has 2942 edges of type r12 and 2091 are inside the partition
part 1 has 3191 edges of type r13 and 2278 are inside the partition
part 1 has 3082 edges of type r23 and 2183 are inside the partition
Save partitions: 0.003 seconds, peak memory: 1.529 GB
There are 13220 edges in the graph and 0 edge cuts for 2 partitions.
FAILED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-div-shp3-g0] FAILED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-div-shp3-g1] FAILED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-div-shp4-g0] FAILED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-div-shp4-g1] FAILED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-div-shp5-g0] FAILED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-div-shp5-g1] FAILED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_lhs-shp0-g0] FAILED [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_lhs-shp0-g1] FAILED [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_lhs-shp1-g0] FAILED [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_lhs-shp1-g1] FAILED [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_lhs-shp2-g0] PASSED              [ 20%]
tests/compute/test_sampling.py::test_non_uniform_random_walk[True] SKIPPED [ 20%]
tests/compute/test_sampling.py::test_non_uniform_random_walk[False] PASSED [ 20%]
tests/compute/test_sampling.py::test_uniform_random_walk[True] PASSED    [ 20%]
tests/compute/test_sampling.py::test_uniform_random_walk[False] PASSED   [ 20%]
tests/compute/test_sampling.py::test_node2vec SKIPPED (GPU random wa...) [ 20%]
tests/compute/test_sampling.py::test_pack_traces SKIPPED (GPU pack t...) [ 20%]
tests/compute/test_sampling.py::test_pinsage_sampling[True] PASSED       [ 20%]
tests/compute/test_sampling.py::test_pinsage_sampling[False] PASSED      [ 20%]
tests/compute/test_sampling.py::test_sample_neighbors_noprob FAILED [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_lhs-shp2-g1] FAILED [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_lhs-shp3-g0] FAILED [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_lhs-shp3-g1] FAILED [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_lhs-shp4-g0] FAILED [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_lhs-shp4-g1] FAILED [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_lhs-shp5-g0] PASSED      [ 20%]
tests/compute/test_sampling.py::test_sample_neighbors_prob FAILED [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_lhs-shp5-g1] FAILED [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_rhs-shp0-g0] FAILED [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_rhs-shp0-g1] FAILED [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_rhs-shp1-g0] FAILED [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_rhs-shp1-g1] /root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_sampling
start graph service on server 0 for part 0
[05:32:27] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:32:27] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:18024]...
PASSED        [ 20%]
tests/compute/test_sampling.py::test_sample_neighbors_outedge FAILED [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_rhs-shp2-g0] FAILED [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_rhs-shp2-g1] FAILED [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_rhs-shp3-g0] FAILED [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_rhs-shp3-g1] FAILED [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_rhs-shp4-g0] FAILED [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_rhs-shp4-g1] FAILED [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_rhs-shp5-g0] FAILED [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_rhs-shp5-g1] FAILED [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-add-shp0-g0] FAILED  [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-add-shp0-g1] FAILED  [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-add-shp1-g0] PASSED     [ 20%]
tests/compute/test_sampling.py::test_sample_neighbors_topk SKIPPED (...) [ 20%]
tests/compute/test_sampling.py::test_sample_neighbors_topk_outedge SKIPPED [ 20%]
tests/compute/test_sampling.py::test_sample_neighbors_with_0deg PASSED   [ 20%]
tests/compute/test_sampling.py::test_sample_neighbors_biased_homogeneous SKIPPED [ 20%]
tests/compute/test_sampling.py::test_sample_neighbors_biased_bipartite SKIPPED [ 20%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_homogeneous[False-in-coo] SKIPPED [ 20%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_homogeneous[False-in-csr] SKIPPED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_homogeneous[False-in-csc] SKIPPED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_homogeneous[False-out-coo] SKIPPED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_homogeneous[False-out-csr] SKIPPED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_homogeneous[False-out-csc] SKIPPED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_homogeneous[True-in-coo] SKIPPED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_homogeneous[True-in-csr] SKIPPED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_homogeneous[True-in-csc] SKIPPED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_homogeneous[True-out-coo] SKIPPED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_homogeneous[True-out-csr] SKIPPED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_homogeneous[True-out-csc] SKIPPED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_sorted_homogeneous[in-csr] SKIPPED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_sorted_homogeneous[in-csc] SKIPPED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_sorted_homogeneous[out-csr] SKIPPED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_sorted_homogeneous[out-csc] SKIPPED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_exclude_edges_heteroG[int32] PASSED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_exclude_edges_heteroG[int64] PASSED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_exclude_edges_homoG[int32] PASSED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_exclude_edges_homoG[int64] PASSED [ 21%]
tests/compute/test_sampling.py::test_global_uniform_negative_sampling[int32] FAILED  [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-add-shp1-g1] FAILED  [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-add-shp2-g0] FAILED  [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-add-shp2-g1] FAILED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-add-shp3-g0] FAILED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-add-shp3-g1] FAILED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-add-shp4-g0] PASSED [ 21%]
tests/compute/test_sampling.py::test_global_uniform_negative_sampling[int64] PASSED [ 21%]
tests/compute/test_serialize.py::test_graph_serialize_with_feature[True] SKIPPED [ 21%]
tests/compute/test_serialize.py::test_graph_serialize_with_feature[False] SKIPPED [ 21%]
tests/compute/test_serialize.py::test_graph_serialize_without_feature[True] SKIPPED [ 21%]
tests/compute/test_serialize.py::test_graph_serialize_without_feature[False] SKIPPED [ 22%]
tests/compute/test_serialize.py::test_graph_serialize_with_labels[True] SKIPPED [ 22%]
tests/compute/test_serialize.py::test_graph_serialize_with_labels[False] SKIPPED [ 22%]
tests/compute/test_serialize.py::test_serialize_tensors PASSED           [ 22%]
tests/compute/test_serialize.py::test_serialize_empty_dict PASSED        [ 22%]
tests/compute/test_serialize.py::test_load_old_files1 PASSED             [ 22%]
tests/compute/test_serialize.py::test_load_old_files2 PASSED             [ 22%]
tests/compute/test_serialize.py::test_deserialize_old_heterograph_file PASSED [ 22%]
tests/compute/test_serialize.py::test_serialize_heterograph SKIPPED      [ 22%]
tests/compute/test_serialize.py::test_serialize_heterograph_s3 SKIPPED   [ 22%]
tests/compute/test_shared_mem.py::test_single_process[int32] PASSED      [ 22%]
tests/compute/test_shared_mem.py::test_single_process[int64] PASSED      [ 22%]
tests/compute/test_shared_mem.py::test_multi_process[int32] PASSED       [ 22%]
tests/compute/test_shared_mem.py::test_multi_process[int64] FAILED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-add-shp4-g1] FAILED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-add-shp5-g0] FAILED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-add-shp5-g1] FAILED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-sub-shp0-g0] FAILED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-sub-shp0-g1] FAILED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-sub-shp1-g0] FAILED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-sub-shp1-g1] PASSED       [ 22%]
tests/compute/test_shared_mem.py::test_copy_from_gpu PASSED              [ 22%]
tests/compute/test_sort.py::test_sort_with_tag[int32] SKIPPED (GPU s...) [ 22%]
tests/compute/test_sort.py::test_sort_with_tag[int64] SKIPPED (GPU s...) [ 22%]
tests/compute/test_sort.py::test_sort_with_tag_bipartite[int32] SKIPPED  [ 22%]
tests/compute/test_sort.py::test_sort_with_tag_bipartite[int64] SKIPPED  [ 22%]
tests/compute/test_sparse.py::test_spmm[int32-sum-add-shp0-g0] PASSED    [ 22%]
tests/compute/test_sparse.py::test_spmm[int32-sum-add-shp0-g1] PASSED    [ 22%]
tests/compute/test_sparse.py::test_spmm[int32-sum-add-shp1-g0] PASSED    [ 22%]
tests/compute/test_sparse.py::test_spmm[int32-sum-add-shp1-g1] PASSED    [ 22%]
tests/compute/test_sparse.py::test_spmm[int32-sum-add-shp2-g0] PASSED    [ 22%]
tests/compute/test_sparse.py::test_spmm[int32-sum-add-shp2-g1] FAILED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-sub-shp2-g0] FAILED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-sub-shp2-g1] FAILED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-sub-shp3-g0] FAILED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-sub-shp3-g1] /root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_sampling
start graph service on server 1 for part 1
[05:32:28] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:32:28] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:18026]...
[05:32:28] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:32:28] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
PASSED    [ 22%]
tests/compute/test_sparse.py::test_spmm[int32-sum-add-shp3-g0] PASSED    [ 23%]
tests/compute/test_sparse.py::test_spmm[int32-sum-add-shp3-g1] PASSED    [ 23%]
tests/compute/test_sparse.py::test_spmm[int32-sum-add-shp4-g0] PASSED    [ 23%]
tests/compute/test_sparse.py::test_spmm[int32-sum-add-shp4-g1] PASSED    [ 23%]
tests/compute/test_sparse.py::test_spmm[int32-sum-add-shp5-g0] PASSED    [ 23%]
tests/compute/test_sparse.py::test_spmm[int32-sum-add-shp5-g1] PASSED    [ 23%]
tests/compute/test_sparse.py::test_spmm[int32-sum-sub-shp0-g0] PASSED    [ 23%]
tests/compute/test_sparse.py::test_spmm[int32-sum-sub-shp0-g1] PASSED    [ 23%]
tests/compute/test_sparse.py::test_spmm[int32-sum-sub-shp1-g0] PASSED    [ 23%]
tests/compute/test_sparse.py::test_spmm[int32-sum-sub-shp1-g1] PASSED    [ 23%]
tests/compute/test_sparse.py::test_spmm[int32-sum-sub-shp2-g0] PASSED    [ 23%]
tests/compute/test_sparse.py::test_spmm[int32-sum-sub-shp2-g1] PASSED    [ 23%]
tests/compute/test_sparse.py::test_spmm[int32-sum-sub-shp3-g0] PASSED    [ 23%]
tests/compute/test_sparse.py::test_spmm[int32-sum-sub-shp3-g1] FAILED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-sub-shp4-g0] FAILED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-sub-shp4-g1] FAILED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-sub-shp5-g0] FAILED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-sub-shp5-g1] FAILED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-mul-shp0-g0] PASSED    [ 23%]
tests/compute/test_sparse.py::test_spmm[int32-sum-sub-shp4-g0] PASSED    [ 23%]
tests/compute/test_sparse.py::test_spmm[int32-sum-sub-shp4-g1] PASSED    [ 23%]
tests/compute/test_sparse.py::test_spmm[int32-sum-sub-shp5-g0] PASSED    [ 23%]
tests/compute/test_sparse.py::test_spmm[int32-sum-sub-shp5-g1] PASSED    [ 23%]
tests/compute/test_sparse.py::test_spmm[int32-sum-mul-shp0-g0] PASSED    [ 23%]
tests/compute/test_sparse.py::test_spmm[int32-sum-mul-shp0-g1] PASSED    [ 23%]
tests/compute/test_sparse.py::test_spmm[int32-sum-mul-shp1-g0] PASSED    [ 23%]
tests/compute/test_sparse.py::test_spmm[int32-sum-mul-shp1-g1] PASSED    [ 23%]
tests/compute/test_sparse.py::test_spmm[int32-sum-mul-shp2-g0] PASSED    [ 23%]
tests/compute/test_sparse.py::test_spmm[int32-sum-mul-shp2-g1] PASSED    [ 23%]
tests/compute/test_sparse.py::test_spmm[int32-sum-mul-shp3-g0] PASSED    [ 24%]
tests/compute/test_sparse.py::test_spmm[int32-sum-mul-shp3-g1] PASSED    [ 24%]
tests/compute/test_sparse.py::test_spmm[int32-sum-mul-shp4-g0] PASSED    [ 24%]
tests/compute/test_sparse.py::test_spmm[int32-sum-mul-shp4-g1] FAILED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-mul-shp0-g1] FAILED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-mul-shp1-g0] FAILED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-mul-shp1-g1] FAILED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-mul-shp2-g0] PASSED               [ 11%]
tests/compute/test_kernel.py::test_all_binary_builtins PASSED    [ 24%]
tests/compute/test_sparse.py::test_spmm[int32-sum-mul-shp5-g0] PASSED    [ 24%]
tests/compute/test_sparse.py::test_spmm[int32-sum-mul-shp5-g1] PASSED    [ 24%]
tests/compute/test_sparse.py::test_spmm[int32-sum-div-shp0-g0] PASSED    [ 24%]
tests/compute/test_sparse.py::test_spmm[int32-sum-div-shp0-g1] PASSED    [ 24%]
tests/compute/test_sparse.py::test_spmm[int32-sum-div-shp1-g0] PASSED    [ 24%]
tests/compute/test_sparse.py::test_spmm[int32-sum-div-shp1-g1] PASSED    [ 24%]
tests/compute/test_sparse.py::test_spmm[int32-sum-div-shp2-g0] PASSED    [ 24%]
tests/compute/test_sparse.py::test_spmm[int32-sum-div-shp2-g1] PASSED    [ 24%]
tests/compute/test_sparse.py::test_spmm[int32-sum-div-shp3-g0] PASSED    [ 24%]
tests/compute/test_sparse.py::test_spmm[int32-sum-div-shp3-g1] PASSED    [ 24%]
tests/compute/test_sparse.py::test_spmm[int32-sum-div-shp4-g0] PASSED    [ 24%]
tests/compute/test_sparse.py::test_spmm[int32-sum-div-shp4-g1] PASSED    [ 24%]
tests/compute/test_sparse.py::test_spmm[int32-sum-div-shp5-g0] FAILED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-mul-shp2-g1] FAILED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-mul-shp3-g0] FAILED  [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-mul-shp3-g1] FAILED  [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-mul-shp4-g0] FAILED  [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-mul-shp4-g1] FAILED  [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-mul-shp5-g0] PASSED    [ 24%]
tests/compute/test_sparse.py::test_spmm[int32-sum-div-shp5-g1] PASSED    [ 24%]
tests/compute/test_sparse.py::test_spmm[int32-sum-copy_lhs-shp0-g0] PASSED [ 24%]
tests/compute/test_sparse.py::test_spmm[int32-sum-copy_lhs-shp0-g1] PASSED [ 24%]
tests/compute/test_sparse.py::test_spmm[int32-sum-copy_lhs-shp1-g0] PASSED [ 24%]
tests/compute/test_sparse.py::test_spmm[int32-sum-copy_lhs-shp1-g1] PASSED [ 24%]
tests/compute/test_sparse.py::test_spmm[int32-sum-copy_lhs-shp2-g0] PASSED [ 24%]
tests/compute/test_sparse.py::test_spmm[int32-sum-copy_lhs-shp2-g1] PASSED [ 24%]
tests/compute/test_sparse.py::test_spmm[int32-sum-copy_lhs-shp3-g0] PASSED [ 24%]
tests/compute/test_sparse.py::test_spmm[int32-sum-copy_lhs-shp3-g1] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[int32-sum-copy_lhs-shp4-g0] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[int32-sum-copy_lhs-shp4-g1] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[int32-sum-copy_lhs-shp5-g0] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[int32-sum-copy_lhs-shp5-g1] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[int32-sum-copy_rhs-shp0-g0] FAILED  [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-mul-shp5-g1] FAILED  [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-div-shp0-g0] FAILED  [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-div-shp0-g1] FAILED  [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-div-shp1-g0] FAILED  [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-div-shp1-g1] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[int32-sum-copy_rhs-shp0-g1] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[int32-sum-copy_rhs-shp1-g0] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[int32-sum-copy_rhs-shp1-g1] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[int32-sum-copy_rhs-shp2-g0] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[int32-sum-copy_rhs-shp2-g1] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[int32-sum-copy_rhs-shp3-g0] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[int32-sum-copy_rhs-shp3-g1] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[int32-sum-copy_rhs-shp4-g0] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[int32-sum-copy_rhs-shp4-g1] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[int32-sum-copy_rhs-shp5-g0] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[int32-sum-copy_rhs-shp5-g1] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[int32-min-add-shp0-g0] PASSED    [ 25%]
tests/compute/test_sparse.py::test_spmm[int32-min-add-shp0-g1] PASSED    [ 25%]
tests/compute/test_sparse.py::test_spmm[int32-min-add-shp1-g0] FAILED  [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-div-shp2-g0] FAILED  [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-div-shp2-g1] FAILED  [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-div-shp3-g0] FAILED  [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-div-shp3-g1] FAILED  [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-div-shp4-g0] PASSED    [ 25%]
tests/compute/test_sparse.py::test_spmm[int32-min-add-shp1-g1] PASSED    [ 25%]
tests/compute/test_sparse.py::test_spmm[int32-min-add-shp2-g0] PASSED    [ 25%]
tests/compute/test_sparse.py::test_spmm[int32-min-add-shp2-g1] PASSED    [ 25%]
tests/compute/test_sparse.py::test_spmm[int32-min-add-shp3-g0] PASSED    [ 25%]
tests/compute/test_sparse.py::test_spmm[int32-min-add-shp3-g1] PASSED    [ 26%]
tests/compute/test_sparse.py::test_spmm[int32-min-add-shp4-g0] PASSED    [ 26%]
tests/compute/test_sparse.py::test_spmm[int32-min-add-shp4-g1] PASSED    [ 26%]
tests/compute/test_sparse.py::test_spmm[int32-min-add-shp5-g0] PASSED    [ 26%]
tests/compute/test_sparse.py::test_spmm[int32-min-add-shp5-g1] PASSED    [ 26%]
tests/compute/test_sparse.py::test_spmm[int32-min-sub-shp0-g0] PASSED    [ 26%]
tests/compute/test_sparse.py::test_spmm[int32-min-sub-shp0-g1] PASSED    [ 26%]
tests/compute/test_sparse.py::test_spmm[int32-min-sub-shp1-g0] PASSED    [ 26%]
tests/compute/test_sparse.py::test_spmm[int32-min-sub-shp1-g1] FAILED  [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-div-shp4-g1] FAILED  [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-div-shp5-g0] FAILED  [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-div-shp5-g1] FAILED  [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_lhs-shp0-g0] FAILED [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_lhs-shp0-g1] PASSED    [ 26%]
tests/compute/test_sparse.py::test_spmm[int32-min-sub-shp2-g0] PASSED    [ 26%]
tests/compute/test_sparse.py::test_spmm[int32-min-sub-shp2-g1] PASSED    [ 26%]
tests/compute/test_sparse.py::test_spmm[int32-min-sub-shp3-g0] PASSED    [ 26%]
tests/compute/test_sparse.py::test_spmm[int32-min-sub-shp3-g1] PASSED    [ 26%]
tests/compute/test_sparse.py::test_spmm[int32-min-sub-shp4-g0] PASSED    [ 26%]
tests/compute/test_sparse.py::test_spmm[int32-min-sub-shp4-g1] PASSED    [ 26%]
tests/compute/test_sparse.py::test_spmm[int32-min-sub-shp5-g0] PASSED    [ 26%]
tests/compute/test_sparse.py::test_spmm[int32-min-sub-shp5-g1] PASSED    [ 26%]
tests/compute/test_sparse.py::test_spmm[int32-min-mul-shp0-g0] PASSED    [ 26%]
tests/compute/test_sparse.py::test_spmm[int32-min-mul-shp0-g1] PASSED    [ 26%]
tests/compute/test_sparse.py::test_spmm[int32-min-mul-shp1-g0] PASSED    [ 26%]
tests/compute/test_sparse.py::test_spmm[int32-min-mul-shp1-g1] FAILED [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_lhs-shp1-g0] FAILED [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_lhs-shp1-g1] FAILED [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_lhs-shp2-g0] FAILED [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_lhs-shp2-g1] FAILED [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_lhs-shp3-g0] PASSED    [ 26%]
tests/compute/test_sparse.py::test_spmm[int32-min-mul-shp2-g0] PASSED    [ 26%]
tests/compute/test_sparse.py::test_spmm[int32-min-mul-shp2-g1] PASSED    [ 26%]
tests/compute/test_sparse.py::test_spmm[int32-min-mul-shp3-g0] PASSED    [ 26%]
tests/compute/test_sparse.py::test_spmm[int32-min-mul-shp3-g1] PASSED    [ 26%]
tests/compute/test_sparse.py::test_spmm[int32-min-mul-shp4-g0] PASSED    [ 27%]
tests/compute/test_sparse.py::test_spmm[int32-min-mul-shp4-g1] PASSED    [ 27%]
tests/compute/test_sparse.py::test_spmm[int32-min-mul-shp5-g0] PASSED    [ 27%]
tests/compute/test_sparse.py::test_spmm[int32-min-mul-shp5-g1] PASSED    [ 27%]
tests/compute/test_sparse.py::test_spmm[int32-min-div-shp0-g0] PASSED    [ 27%]
tests/compute/test_sparse.py::test_spmm[int32-min-div-shp0-g1] PASSED    [ 27%]
tests/compute/test_sparse.py::test_spmm[int32-min-div-shp1-g0] PASSED    [ 27%]
tests/compute/test_sparse.py::test_spmm[int32-min-div-shp1-g1] FAILED [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_lhs-shp3-g1] FAILED [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_lhs-shp4-g0] FAILED [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_lhs-shp4-g1] FAILED [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_lhs-shp5-g0] FAILED [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_lhs-shp5-g1] PASSED    [ 27%]
tests/compute/test_sparse.py::test_spmm[int32-min-div-shp2-g0] PASSED    [ 27%]
tests/compute/test_sparse.py::test_spmm[int32-min-div-shp2-g1] PASSED    [ 27%]
tests/compute/test_sparse.py::test_spmm[int32-min-div-shp3-g0] PASSED    [ 27%]
tests/compute/test_sparse.py::test_spmm[int32-min-div-shp3-g1] PASSED    [ 27%]
tests/compute/test_sparse.py::test_spmm[int32-min-div-shp4-g0] PASSED    [ 27%]
tests/compute/test_sparse.py::test_spmm[int32-min-div-shp4-g1] PASSED    [ 27%]
tests/compute/test_sparse.py::test_spmm[int32-min-div-shp5-g0] PASSED    [ 27%]
tests/compute/test_sparse.py::test_spmm[int32-min-div-shp5-g1] PASSED    [ 27%]
tests/compute/test_sparse.py::test_spmm[int32-min-copy_lhs-shp0-g0] PASSED [ 27%]
tests/compute/test_sparse.py::test_spmm[int32-min-copy_lhs-shp0-g1] PASSED [ 27%]
tests/compute/test_sparse.py::test_spmm[int32-min-copy_lhs-shp1-g0] PASSED [ 27%]
tests/compute/test_sparse.py::test_spmm[int32-min-copy_lhs-shp1-g1] PASSED [ 27%]
tests/compute/test_sparse.py::test_spmm[int32-min-copy_lhs-shp2-g0] FAILED [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_rhs-shp0-g0] FAILED [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_rhs-shp0-g1] FAILED [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_rhs-shp1-g0] FAILED [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_rhs-shp1-g1] PASSED [ 27%]
tests/compute/test_sparse.py::test_spmm[int32-min-copy_lhs-shp2-g1] PASSED [ 27%]
tests/compute/test_sparse.py::test_spmm[int32-min-copy_lhs-shp3-g0] PASSED [ 27%]
tests/compute/test_sparse.py::test_spmm[int32-min-copy_lhs-shp3-g1] PASSED [ 27%]
tests/compute/test_sparse.py::test_spmm[int32-min-copy_lhs-shp4-g0] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[int32-min-copy_lhs-shp4-g1] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[int32-min-copy_lhs-shp5-g0] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[int32-min-copy_lhs-shp5-g1] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[int32-min-copy_rhs-shp0-g0] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[int32-min-copy_rhs-shp0-g1] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[int32-min-copy_rhs-shp1-g0] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[int32-min-copy_rhs-shp1-g1] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[int32-min-copy_rhs-shp2-g0] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[int32-min-copy_rhs-shp2-g1] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[int32-min-copy_rhs-shp3-g0] FAILED [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_rhs-shp2-g0] FAILED [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_rhs-shp2-g1] FAILED [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_rhs-shp3-g0] FAILED [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_rhs-shp3-g1] FAILED [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_rhs-shp4-g0] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[int32-min-copy_rhs-shp3-g1] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[int32-min-copy_rhs-shp4-g0] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[int32-min-copy_rhs-shp4-g1] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[int32-min-copy_rhs-shp5-g0] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[int32-min-copy_rhs-shp5-g1] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[int32-max-add-shp0-g0] PASSED    [ 28%]
tests/compute/test_sparse.py::test_spmm[int32-max-add-shp0-g1] PASSED    [ 28%]
tests/compute/test_sparse.py::test_spmm[int32-max-add-shp1-g0] PASSED    [ 28%]
tests/compute/test_sparse.py::test_spmm[int32-max-add-shp1-g1] PASSED    [ 28%]
tests/compute/test_sparse.py::test_spmm[int32-max-add-shp2-g0] PASSED    [ 28%]
tests/compute/test_sparse.py::test_spmm[int32-max-add-shp2-g1] PASSED    [ 28%]
tests/compute/test_sparse.py::test_spmm[int32-max-add-shp3-g0] PASSED    [ 28%]
tests/compute/test_sparse.py::test_spmm[int32-max-add-shp3-g1] FAILED [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_rhs-shp4-g1] FAILED [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_rhs-shp5-g0] FAILED [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_rhs-shp5-g1] FAILED [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-add-shp0-g0] FAILED  [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-add-shp0-g1] PASSED    [ 28%]
tests/compute/test_sparse.py::test_spmm[int32-max-add-shp4-g0] PASSED    [ 28%]
tests/compute/test_sparse.py::test_spmm[int32-max-add-shp4-g1] PASSED    [ 29%]
tests/compute/test_sparse.py::test_spmm[int32-max-add-shp5-g0] PASSED    [ 29%]
tests/compute/test_sparse.py::test_spmm[int32-max-add-shp5-g1] PASSED    [ 29%]
tests/compute/test_sparse.py::test_spmm[int32-max-sub-shp0-g0] PASSED    [ 29%]
tests/compute/test_sparse.py::test_spmm[int32-max-sub-shp0-g1] PASSED    [ 29%]
tests/compute/test_sparse.py::test_spmm[int32-max-sub-shp1-g0] PASSED    [ 29%]
tests/compute/test_sparse.py::test_spmm[int32-max-sub-shp1-g1] PASSED    [ 29%]
tests/compute/test_sparse.py::test_spmm[int32-max-sub-shp2-g0] PASSED    [ 29%]
tests/compute/test_sparse.py::test_spmm[int32-max-sub-shp2-g1] PASSED    [ 29%]
tests/compute/test_sparse.py::test_spmm[int32-max-sub-shp3-g0] PASSED    [ 29%]
tests/compute/test_sparse.py::test_spmm[int32-max-sub-shp3-g1] PASSED    [ 29%]
tests/compute/test_sparse.py::test_spmm[int32-max-sub-shp4-g0] FAILED  [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-add-shp1-g0] FAILED  [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-add-shp1-g1] FAILED  [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-add-shp2-g0] FAILED  [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-add-shp2-g1] FAILED  [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-add-shp3-g0] PASSED    [ 29%]
tests/compute/test_sparse.py::test_spmm[int32-max-sub-shp4-g1] PASSED    [ 29%]
tests/compute/test_sparse.py::test_spmm[int32-max-sub-shp5-g0] PASSED    [ 29%]
tests/compute/test_sparse.py::test_spmm[int32-max-sub-shp5-g1] PASSED    [ 29%]
tests/compute/test_sparse.py::test_spmm[int32-max-mul-shp0-g0] PASSED    [ 29%]
tests/compute/test_sparse.py::test_spmm[int32-max-mul-shp0-g1] PASSED    [ 29%]
tests/compute/test_sparse.py::test_spmm[int32-max-mul-shp1-g0] PASSED    [ 29%]
tests/compute/test_sparse.py::test_spmm[int32-max-mul-shp1-g1] PASSED    [ 29%]
tests/compute/test_sparse.py::test_spmm[int32-max-mul-shp2-g0] PASSED    [ 29%]
tests/compute/test_sparse.py::test_spmm[int32-max-mul-shp2-g1] PASSED    [ 29%]
tests/compute/test_sparse.py::test_spmm[int32-max-mul-shp3-g0] PASSED    [ 29%]
tests/compute/test_sparse.py::test_spmm[int32-max-mul-shp3-g1] PASSED    [ 29%]
tests/compute/test_sparse.py::test_spmm[int32-max-mul-shp4-g0] FAILED  [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-add-shp3-g1] FAILED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-add-shp4-g0] FAILED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-add-shp4-g1] FAILED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-add-shp5-g0] FAILED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-add-shp5-g1] Client [1090] waits on 172.17.0.3:45547
Machine (0) group (0) client (0) connect to server successfuly!
Client[0] in group[0] is exiting...
Server (0) shutdown.
Server is exiting...
Done sampling
Server (1) shutdown.
Server is exiting...
Converting to homogeneous graph takes 0.001s, peak mem: 1.529 GB
Convert a graph into a bidirected graph: 0.000 seconds, peak memory: 1.529 GB
Construct multi-constraint weights: 0.000 seconds, peak memory: 1.529 GB
[05:32:32] /root/jenkins/workspace/dgl_PR-4648/src/graph/transform/metis_partition_hetero.cc:87: Partition a graph with 630 nodes and 26440 edges into 2 parts and get 5478 edge cuts
Metis partitioning: 0.005 seconds, peak memory: 1.529 GB
Assigning nodes to METIS partitions takes 0.006s, peak mem: 1.529 GB
Reshuffle nodes and edges: 0.001 seconds
Split the graph: 0.001 seconds
Construct subgraphs: 0.003 seconds
Splitting the graph into partitions takes 0.005s, peak mem: 1.529 GB
part 0 has 210 nodes of type n1 and 108 are inside the partition
part 0 has 200 nodes of type n2 and 100 are inside the partition
part 0 has 220 nodes of type n3 and 113 are inside the partition
part 0 has 2989 edges of type r12 and 2109 are inside the partition
part 0 has 3328 edges of type r13 and 2342 are inside the partition
part 0 has 3166 edges of type r23 and 2217 are inside the partition
part 1 has 210 nodes of type n1 and 102 are inside the partition
part 1 has 200 nodes of type n2 and 100 are inside the partition
part 1 has 220 nodes of type n3 and 107 are inside the partition
part 1 has 2942 edges of type r12 and 2091 are inside the partition
part 1 has 3191 edges of type r13 and 2278 are inside the partition
part 1 has 3082 edges of type r23 and 2183 are inside the partition
Save partitions: 0.003 seconds, peak memory: 1.529 GB
There are 13220 edges in the graph and 0 edge cuts for 2 partitions.
PASSED    [ 29%]
tests/compute/test_sparse.py::test_spmm[int32-max-mul-shp4-g1] PASSED    [ 30%]
tests/compute/test_sparse.py::test_spmm[int32-max-mul-shp5-g0] PASSED    [ 30%]
tests/compute/test_sparse.py::test_spmm[int32-max-mul-shp5-g1] PASSED    [ 30%]
tests/compute/test_sparse.py::test_spmm[int32-max-div-shp0-g0] PASSED    [ 30%]
tests/compute/test_sparse.py::test_spmm[int32-max-div-shp0-g1] PASSED    [ 30%]
tests/compute/test_sparse.py::test_spmm[int32-max-div-shp1-g0] PASSED    [ 30%]
tests/compute/test_sparse.py::test_spmm[int32-max-div-shp1-g1] PASSED    [ 30%]
tests/compute/test_sparse.py::test_spmm[int32-max-div-shp2-g0] PASSED    [ 30%]
tests/compute/test_sparse.py::test_spmm[int32-max-div-shp2-g1] PASSED    [ 30%]
tests/compute/test_sparse.py::test_spmm[int32-max-div-shp3-g0] PASSED    [ 30%]
tests/compute/test_sparse.py::test_spmm[int32-max-div-shp3-g1] PASSED    [ 30%]
tests/compute/test_sparse.py::test_spmm[int32-max-div-shp4-g0] FAILED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-sub-shp0-g0] FAILED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-sub-shp0-g1] FAILED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-sub-shp1-g0] FAILED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-sub-shp1-g1] FAILED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-sub-shp2-g0] PASSED    [ 30%]
tests/compute/test_sparse.py::test_spmm[int32-max-div-shp4-g1] PASSED    [ 30%]
tests/compute/test_sparse.py::test_spmm[int32-max-div-shp5-g0] PASSED    [ 30%]
tests/compute/test_sparse.py::test_spmm[int32-max-div-shp5-g1] PASSED    [ 30%]
tests/compute/test_sparse.py::test_spmm[int32-max-copy_lhs-shp0-g0] PASSED [ 30%]
tests/compute/test_sparse.py::test_spmm[int32-max-copy_lhs-shp0-g1] PASSED [ 30%]
tests/compute/test_sparse.py::test_spmm[int32-max-copy_lhs-shp1-g0] PASSED [ 30%]
tests/compute/test_sparse.py::test_spmm[int32-max-copy_lhs-shp1-g1] PASSED [ 30%]
tests/compute/test_sparse.py::test_spmm[int32-max-copy_lhs-shp2-g0] PASSED [ 30%]
tests/compute/test_sparse.py::test_spmm[int32-max-copy_lhs-shp2-g1] PASSED [ 30%]
tests/compute/test_sparse.py::test_spmm[int32-max-copy_lhs-shp3-g0] PASSED [ 30%]
tests/compute/test_sparse.py::test_spmm[int32-max-copy_lhs-shp3-g1] PASSED [ 30%]
tests/compute/test_sparse.py::test_spmm[int32-max-copy_lhs-shp4-g0] PASSED [ 30%]
tests/compute/test_sparse.py::test_spmm[int32-max-copy_lhs-shp4-g1] FAILED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-sub-shp2-g1] FAILED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-sub-shp3-g0] FAILED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-sub-shp3-g1] FAILED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-sub-shp4-g0] FAILED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-sub-shp4-g1] PASSED [ 30%]
tests/compute/test_sparse.py::test_spmm[int32-max-copy_lhs-shp5-g0] PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[int32-max-copy_lhs-shp5-g1] PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[int32-max-copy_rhs-shp0-g0] PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[int32-max-copy_rhs-shp0-g1] PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[int32-max-copy_rhs-shp1-g0] PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[int32-max-copy_rhs-shp1-g1] PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[int32-max-copy_rhs-shp2-g0] PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[int32-max-copy_rhs-shp2-g1] PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[int32-max-copy_rhs-shp3-g0] PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[int32-max-copy_rhs-shp3-g1] PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[int32-max-copy_rhs-shp4-g0] PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[int32-max-copy_rhs-shp4-g1] PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[int32-max-copy_rhs-shp5-g0] PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[int32-max-copy_rhs-shp5-g1] FAILED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-sub-shp5-g0] FAILED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-sub-shp5-g1] FAILED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-mul-shp0-g0] FAILED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-mul-shp0-g1] PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[int64-sum-add-shp0-g0] PASSED    [ 31%]
tests/compute/test_sparse.py::test_spmm[int64-sum-add-shp0-g1] PASSED    [ 31%]
tests/compute/test_sparse.py::test_spmm[int64-sum-add-shp1-g0] PASSED    [ 31%]
tests/compute/test_sparse.py::test_spmm[int64-sum-add-shp1-g1] PASSED    [ 31%]
tests/compute/test_sparse.py::test_spmm[int64-sum-add-shp2-g0] PASSED    [ 31%]
tests/compute/test_sparse.py::test_spmm[int64-sum-add-shp2-g1] PASSED    [ 31%]
tests/compute/test_sparse.py::test_spmm[int64-sum-add-shp3-g0] PASSED    [ 31%]
tests/compute/test_sparse.py::test_spmm[int64-sum-add-shp3-g1] PASSED    [ 31%]
tests/compute/test_sparse.py::test_spmm[int64-sum-add-shp4-g0] PASSED    [ 31%]
tests/compute/test_sparse.py::test_spmm[int64-sum-add-shp4-g1] PASSED    [ 31%]
tests/compute/test_sparse.py::test_spmm[int64-sum-add-shp5-g0] PASSED    [ 32%]
tests/compute/test_sparse.py::test_spmm[int64-sum-add-shp5-g1] PASSED    [ 32%]
tests/compute/test_sparse.py::test_spmm[int64-sum-sub-shp0-g0] FAILED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-mul-shp1-g0] FAILED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-mul-shp1-g1] FAILED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-mul-shp2-g0] FAILED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-mul-shp2-g1] FAILED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-mul-shp3-g0] /root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_sampling
start graph service on server 0 for part 0
[05:32:33] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:32:33] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:18024]...
PASSED    [ 32%]
tests/compute/test_sparse.py::test_spmm[int64-sum-sub-shp0-g1] PASSED    [ 32%]
tests/compute/test_sparse.py::test_spmm[int64-sum-sub-shp1-g0] PASSED    [ 32%]
tests/compute/test_sparse.py::test_spmm[int64-sum-sub-shp1-g1] PASSED    [ 32%]
tests/compute/test_sparse.py::test_spmm[int64-sum-sub-shp2-g0] PASSED    [ 32%]
tests/compute/test_sparse.py::test_spmm[int64-sum-sub-shp2-g1] PASSED    [ 32%]
tests/compute/test_sparse.py::test_spmm[int64-sum-sub-shp3-g0] PASSED    [ 32%]
tests/compute/test_sparse.py::test_spmm[int64-sum-sub-shp3-g1] PASSED    [ 32%]
tests/compute/test_sparse.py::test_spmm[int64-sum-sub-shp4-g0] PASSED    [ 32%]
tests/compute/test_sparse.py::test_spmm[int64-sum-sub-shp4-g1] PASSED    [ 32%]
tests/compute/test_sparse.py::test_spmm[int64-sum-sub-shp5-g0] PASSED    [ 32%]
tests/compute/test_sparse.py::test_spmm[int64-sum-sub-shp5-g1] PASSED    [ 32%]
tests/compute/test_sparse.py::test_spmm[int64-sum-mul-shp0-g0] PASSED    [ 32%]
tests/compute/test_sparse.py::test_spmm[int64-sum-mul-shp0-g1] PASSED    [ 32%]
tests/compute/test_sparse.py::test_spmm[int64-sum-mul-shp1-g0] FAILED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-mul-shp3-g1] FAILED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-mul-shp4-g0] FAILED  [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-mul-shp4-g1] FAILED  [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-mul-shp5-g0] FAILED  [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-mul-shp5-g1] PASSED    [ 32%]
tests/compute/test_sparse.py::test_spmm[int64-sum-mul-shp1-g1] PASSED    [ 32%]
tests/compute/test_sparse.py::test_spmm[int64-sum-mul-shp2-g0] PASSED    [ 32%]
tests/compute/test_sparse.py::test_spmm[int64-sum-mul-shp2-g1] PASSED    [ 32%]
tests/compute/test_sparse.py::test_spmm[int64-sum-mul-shp3-g0] PASSED    [ 32%]
tests/compute/test_sparse.py::test_spmm[int64-sum-mul-shp3-g1] PASSED    [ 32%]
tests/compute/test_sparse.py::test_spmm[int64-sum-mul-shp4-g0] PASSED    [ 32%]
tests/compute/test_sparse.py::test_spmm[int64-sum-mul-shp4-g1] PASSED    [ 32%]
tests/compute/test_sparse.py::test_spmm[int64-sum-mul-shp5-g0] PASSED    [ 32%]
tests/compute/test_sparse.py::test_spmm[int64-sum-mul-shp5-g1] PASSED    [ 33%]
tests/compute/test_sparse.py::test_spmm[int64-sum-div-shp0-g0] PASSED    [ 33%]
tests/compute/test_sparse.py::test_spmm[int64-sum-div-shp0-g1] PASSED    [ 33%]
tests/compute/test_sparse.py::test_spmm[int64-sum-div-shp1-g0] PASSED    [ 33%]
tests/compute/test_sparse.py::test_spmm[int64-sum-div-shp1-g1] FAILED  [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-div-shp0-g0] FAILED  [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-div-shp0-g1] FAILED  [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-div-shp1-g0] FAILED  [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-div-shp1-g1] FAILED  [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-div-shp2-g0] PASSED    [ 33%]
tests/compute/test_sparse.py::test_spmm[int64-sum-div-shp2-g0] PASSED    [ 33%]
tests/compute/test_sparse.py::test_spmm[int64-sum-div-shp2-g1] PASSED    [ 33%]
tests/compute/test_sparse.py::test_spmm[int64-sum-div-shp3-g0] PASSED    [ 33%]
tests/compute/test_sparse.py::test_spmm[int64-sum-div-shp3-g1] PASSED    [ 33%]
tests/compute/test_sparse.py::test_spmm[int64-sum-div-shp4-g0] PASSED    [ 33%]
tests/compute/test_sparse.py::test_spmm[int64-sum-div-shp4-g1] PASSED    [ 33%]
tests/compute/test_sparse.py::test_spmm[int64-sum-div-shp5-g0] PASSED    [ 33%]
tests/compute/test_sparse.py::test_spmm[int64-sum-div-shp5-g1] PASSED    [ 33%]
tests/compute/test_sparse.py::test_spmm[int64-sum-copy_lhs-shp0-g0] PASSED [ 33%]
tests/compute/test_sparse.py::test_spmm[int64-sum-copy_lhs-shp0-g1] PASSED [ 33%]
tests/compute/test_sparse.py::test_spmm[int64-sum-copy_lhs-shp1-g0] PASSED [ 33%]
tests/compute/test_sparse.py::test_spmm[int64-sum-copy_lhs-shp1-g1] PASSED [ 33%]
tests/compute/test_sparse.py::test_spmm[int64-sum-copy_lhs-shp2-g0] PASSED [ 33%]
tests/compute/test_sparse.py::test_spmm[int64-sum-copy_lhs-shp2-g1] FAILED  [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-div-shp2-g1] FAILED  [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-div-shp3-g0] FAILED  [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-div-shp3-g1] FAILED  [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-div-shp4-g0] FAILED  [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-div-shp4-g1] PASSED [ 33%]
tests/compute/test_sparse.py::test_spmm[int64-sum-copy_lhs-shp3-g0] PASSED [ 33%]
tests/compute/test_sparse.py::test_spmm[int64-sum-copy_lhs-shp3-g1] PASSED [ 33%]
tests/compute/test_sparse.py::test_spmm[int64-sum-copy_lhs-shp4-g0] PASSED [ 33%]
tests/compute/test_sparse.py::test_spmm[int64-sum-copy_lhs-shp4-g1] PASSED [ 33%]
tests/compute/test_sparse.py::test_spmm[int64-sum-copy_lhs-shp5-g0] PASSED [ 33%]
tests/compute/test_sparse.py::test_spmm[int64-sum-copy_lhs-shp5-g1] PASSED [ 33%]
tests/compute/test_sparse.py::test_spmm[int64-sum-copy_rhs-shp0-g0] PASSED [ 34%]
tests/compute/test_sparse.py::test_spmm[int64-sum-copy_rhs-shp0-g1] PASSED [ 34%]
tests/compute/test_sparse.py::test_spmm[int64-sum-copy_rhs-shp1-g0] PASSED [ 34%]
tests/compute/test_sparse.py::test_spmm[int64-sum-copy_rhs-shp1-g1] PASSED [ 34%]
tests/compute/test_sparse.py::test_spmm[int64-sum-copy_rhs-shp2-g0] PASSED [ 34%]
tests/compute/test_sparse.py::test_spmm[int64-sum-copy_rhs-shp2-g1] PASSED [ 34%]
tests/compute/test_sparse.py::test_spmm[int64-sum-copy_rhs-shp3-g0] PASSED [ 34%]
tests/compute/test_sparse.py::test_spmm[int64-sum-copy_rhs-shp3-g1] [05:32:34] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:32:34] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
FAILED  [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-div-shp5-g0] FAILED  [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-div-shp5-g1] FAILED  [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_lhs-shp0-g0] FAILED [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_lhs-shp0-g1] FAILED [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_lhs-shp1-g0] PASSED [ 34%]
tests/compute/test_sparse.py::test_spmm[int64-sum-copy_rhs-shp4-g0] PASSED [ 34%]
tests/compute/test_sparse.py::test_spmm[int64-sum-copy_rhs-shp4-g1] PASSED [ 34%]
tests/compute/test_sparse.py::test_spmm[int64-sum-copy_rhs-shp5-g0] PASSED [ 34%]
tests/compute/test_sparse.py::test_spmm[int64-sum-copy_rhs-shp5-g1] PASSED [ 34%]
tests/compute/test_sparse.py::test_spmm[int64-min-add-shp0-g0] PASSED    [ 34%]
tests/compute/test_sparse.py::test_spmm[int64-min-add-shp0-g1] PASSED    [ 34%]
tests/compute/test_sparse.py::test_spmm[int64-min-add-shp1-g0] PASSED    [ 34%]
tests/compute/test_sparse.py::test_spmm[int64-min-add-shp1-g1] PASSED    [ 34%]
tests/compute/test_sparse.py::test_spmm[int64-min-add-shp2-g0] PASSED    [ 34%]
tests/compute/test_sparse.py::test_spmm[int64-min-add-shp2-g1] PASSED    [ 34%]
tests/compute/test_sparse.py::test_spmm[int64-min-add-shp3-g0] PASSED    [ 34%]
tests/compute/test_sparse.py::test_spmm[int64-min-add-shp3-g1] PASSED    [ 34%]
tests/compute/test_sparse.py::test_spmm[int64-min-add-shp4-g0] /root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_sampling
start graph service on server 1 for part 1
[05:32:34] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:32:34] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:18026]...
FAILED [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_lhs-shp1-g1] FAILED [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_lhs-shp2-g0] FAILED [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_lhs-shp2-g1] FAILED [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_lhs-shp3-g0] PASSED    [ 34%]
tests/compute/test_sparse.py::test_spmm[int64-min-add-shp4-g1] PASSED    [ 34%]
tests/compute/test_sparse.py::test_spmm[int64-min-add-shp5-g0] PASSED    [ 34%]
tests/compute/test_sparse.py::test_spmm[int64-min-add-shp5-g1] PASSED    [ 34%]
tests/compute/test_sparse.py::test_spmm[int64-min-sub-shp0-g0] PASSED    [ 35%]
tests/compute/test_sparse.py::test_spmm[int64-min-sub-shp0-g1] PASSED    [ 35%]
tests/compute/test_sparse.py::test_spmm[int64-min-sub-shp1-g0] PASSED    [ 35%]
tests/compute/test_sparse.py::test_spmm[int64-min-sub-shp1-g1] PASSED    [ 35%]
tests/compute/test_sparse.py::test_spmm[int64-min-sub-shp2-g0] PASSED    [ 35%]
tests/compute/test_sparse.py::test_spmm[int64-min-sub-shp2-g1] PASSED    [ 35%]
tests/compute/test_sparse.py::test_spmm[int64-min-sub-shp3-g0] PASSED    [ 35%]
tests/compute/test_sparse.py::test_spmm[int64-min-sub-shp3-g1] PASSED    [ 35%]
tests/compute/test_sparse.py::test_spmm[int64-min-sub-shp4-g0] PASSED    [ 35%]
tests/compute/test_sparse.py::test_spmm[int64-min-sub-shp4-g1] FAILED [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_lhs-shp3-g1] FAILED [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_lhs-shp4-g0] FAILED [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_lhs-shp4-g1] FAILED [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_lhs-shp5-g0] FAILED [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_lhs-shp5-g1] FAILED [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_rhs-shp0-g0] PASSED    [ 35%]
tests/compute/test_sparse.py::test_spmm[int64-min-sub-shp5-g0] PASSED    [ 35%]
tests/compute/test_sparse.py::test_spmm[int64-min-sub-shp5-g1] PASSED    [ 35%]
tests/compute/test_sparse.py::test_spmm[int64-min-mul-shp0-g0] PASSED    [ 35%]
tests/compute/test_sparse.py::test_spmm[int64-min-mul-shp0-g1] PASSED    [ 35%]
tests/compute/test_sparse.py::test_spmm[int64-min-mul-shp1-g0] PASSED    [ 35%]
tests/compute/test_sparse.py::test_spmm[int64-min-mul-shp1-g1] PASSED    [ 35%]
tests/compute/test_sparse.py::test_spmm[int64-min-mul-shp2-g0] PASSED    [ 35%]
tests/compute/test_sparse.py::test_spmm[int64-min-mul-shp2-g1] PASSED    [ 35%]
tests/compute/test_sparse.py::test_spmm[int64-min-mul-shp3-g0] PASSED    [ 35%]
tests/compute/test_sparse.py::test_spmm[int64-min-mul-shp3-g1] PASSED    [ 35%]
tests/compute/test_sparse.py::test_spmm[int64-min-mul-shp4-g0] PASSED    [ 35%]
tests/compute/test_sparse.py::test_spmm[int64-min-mul-shp4-g1] FAILED [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_rhs-shp0-g1] FAILED [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_rhs-shp1-g0] FAILED [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_rhs-shp1-g1] FAILED [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_rhs-shp2-g0] PASSED    [ 35%]
tests/compute/test_sparse.py::test_spmm[int64-min-mul-shp5-g0] PASSED    [ 35%]
tests/compute/test_sparse.py::test_spmm[int64-min-mul-shp5-g1] PASSED    [ 35%]
tests/compute/test_sparse.py::test_spmm[int64-min-div-shp0-g0] PASSED    [ 35%]
tests/compute/test_sparse.py::test_spmm[int64-min-div-shp0-g1] PASSED    [ 36%]
tests/compute/test_sparse.py::test_spmm[int64-min-div-shp1-g0] PASSED    [ 36%]
tests/compute/test_sparse.py::test_spmm[int64-min-div-shp1-g1] PASSED    [ 36%]
tests/compute/test_sparse.py::test_spmm[int64-min-div-shp2-g0] PASSED    [ 36%]
tests/compute/test_sparse.py::test_spmm[int64-min-div-shp2-g1] PASSED    [ 36%]
tests/compute/test_sparse.py::test_spmm[int64-min-div-shp3-g0] PASSED    [ 36%]
tests/compute/test_sparse.py::test_spmm[int64-min-div-shp3-g1] PASSED    [ 36%]
tests/compute/test_sparse.py::test_spmm[int64-min-div-shp4-g0] PASSED    [ 36%]
tests/compute/test_sparse.py::test_spmm[int64-min-div-shp4-g1] FAILED [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_rhs-shp2-g1] FAILED [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_rhs-shp3-g0] FAILED [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_rhs-shp3-g1] FAILED [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_rhs-shp4-g0] FAILED [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_rhs-shp4-g1] PASSED    [ 36%]
tests/compute/test_sparse.py::test_spmm[int64-min-div-shp5-g0] PASSED    [ 36%]
tests/compute/test_sparse.py::test_spmm[int64-min-div-shp5-g1] PASSED    [ 36%]
tests/compute/test_sparse.py::test_spmm[int64-min-copy_lhs-shp0-g0] PASSED [ 36%]
tests/compute/test_sparse.py::test_spmm[int64-min-copy_lhs-shp0-g1] PASSED [ 36%]
tests/compute/test_sparse.py::test_spmm[int64-min-copy_lhs-shp1-g0] PASSED [ 36%]
tests/compute/test_sparse.py::test_spmm[int64-min-copy_lhs-shp1-g1] PASSED [ 36%]
tests/compute/test_sparse.py::test_spmm[int64-min-copy_lhs-shp2-g0] PASSED [ 36%]
tests/compute/test_sparse.py::test_spmm[int64-min-copy_lhs-shp2-g1] PASSED [ 36%]
tests/compute/test_sparse.py::test_spmm[int64-min-copy_lhs-shp3-g0] PASSED [ 36%]
tests/compute/test_sparse.py::test_spmm[int64-min-copy_lhs-shp3-g1] PASSED [ 36%]
tests/compute/test_sparse.py::test_spmm[int64-min-copy_lhs-shp4-g0] PASSED [ 36%]
tests/compute/test_sparse.py::test_spmm[int64-min-copy_lhs-shp4-g1] PASSED [ 36%]
tests/compute/test_sparse.py::test_spmm[int64-min-copy_lhs-shp5-g0] FAILED [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_rhs-shp5-g0] FAILED [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_rhs-shp5-g1] FAILED [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-add-shp0-g0] FAILED  [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-add-shp0-g1] FAILED  [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-add-shp1-g0] PASSED [ 36%]
tests/compute/test_sparse.py::test_spmm[int64-min-copy_lhs-shp5-g1] PASSED [ 36%]
tests/compute/test_sparse.py::test_spmm[int64-min-copy_rhs-shp0-g0] PASSED [ 36%]
tests/compute/test_sparse.py::test_spmm[int64-min-copy_rhs-shp0-g1] PASSED [ 37%]
tests/compute/test_sparse.py::test_spmm[int64-min-copy_rhs-shp1-g0] PASSED [ 37%]
tests/compute/test_sparse.py::test_spmm[int64-min-copy_rhs-shp1-g1] PASSED [ 37%]
tests/compute/test_sparse.py::test_spmm[int64-min-copy_rhs-shp2-g0] PASSED [ 37%]
tests/compute/test_sparse.py::test_spmm[int64-min-copy_rhs-shp2-g1] PASSED [ 37%]
tests/compute/test_sparse.py::test_spmm[int64-min-copy_rhs-shp3-g0] PASSED [ 37%]
tests/compute/test_sparse.py::test_spmm[int64-min-copy_rhs-shp3-g1] PASSED [ 37%]
tests/compute/test_sparse.py::test_spmm[int64-min-copy_rhs-shp4-g0] PASSED [ 37%]
tests/compute/test_sparse.py::test_spmm[int64-min-copy_rhs-shp4-g1] PASSED [ 37%]
tests/compute/test_sparse.py::test_spmm[int64-min-copy_rhs-shp5-g0] PASSED [ 37%]
tests/compute/test_sparse.py::test_spmm[int64-min-copy_rhs-shp5-g1] PASSED [ 37%]
tests/compute/test_sparse.py::test_spmm[int64-max-add-shp0-g0] FAILED  [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-add-shp1-g1] FAILED  [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-add-shp2-g0] FAILED  [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-add-shp2-g1] FAILED  [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-add-shp3-g0] FAILED  [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-add-shp3-g1] PASSED    [ 37%]
tests/compute/test_sparse.py::test_spmm[int64-max-add-shp0-g1] PASSED    [ 37%]
tests/compute/test_sparse.py::test_spmm[int64-max-add-shp1-g0] PASSED    [ 37%]
tests/compute/test_sparse.py::test_spmm[int64-max-add-shp1-g1] PASSED    [ 37%]
tests/compute/test_sparse.py::test_spmm[int64-max-add-shp2-g0] PASSED    [ 37%]
tests/compute/test_sparse.py::test_spmm[int64-max-add-shp2-g1] PASSED    [ 37%]
tests/compute/test_sparse.py::test_spmm[int64-max-add-shp3-g0] PASSED    [ 37%]
tests/compute/test_sparse.py::test_spmm[int64-max-add-shp3-g1] PASSED    [ 37%]
tests/compute/test_sparse.py::test_spmm[int64-max-add-shp4-g0] PASSED    [ 37%]
tests/compute/test_sparse.py::test_spmm[int64-max-add-shp4-g1] PASSED    [ 37%]
tests/compute/test_sparse.py::test_spmm[int64-max-add-shp5-g0] PASSED    [ 37%]
tests/compute/test_sparse.py::test_spmm[int64-max-add-shp5-g1] PASSED    [ 37%]
tests/compute/test_sparse.py::test_spmm[int64-max-sub-shp0-g0] FAILED  [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-add-shp4-g0] FAILED  [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-add-shp4-g1] FAILED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-add-shp5-g0] FAILED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-add-shp5-g1] FAILED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-sub-shp0-g0] PASSED    [ 37%]
tests/compute/test_sparse.py::test_spmm[int64-max-sub-shp0-g1] PASSED    [ 37%]
tests/compute/test_sparse.py::test_spmm[int64-max-sub-shp1-g0] PASSED    [ 38%]
tests/compute/test_sparse.py::test_spmm[int64-max-sub-shp1-g1] PASSED    [ 38%]
tests/compute/test_sparse.py::test_spmm[int64-max-sub-shp2-g0] PASSED    [ 38%]
tests/compute/test_sparse.py::test_spmm[int64-max-sub-shp2-g1] PASSED    [ 38%]
tests/compute/test_sparse.py::test_spmm[int64-max-sub-shp3-g0] PASSED    [ 38%]
tests/compute/test_sparse.py::test_spmm[int64-max-sub-shp3-g1] PASSED    [ 38%]
tests/compute/test_sparse.py::test_spmm[int64-max-sub-shp4-g0] PASSED    [ 38%]
tests/compute/test_sparse.py::test_spmm[int64-max-sub-shp4-g1] PASSED    [ 38%]
tests/compute/test_sparse.py::test_spmm[int64-max-sub-shp5-g0] PASSED    [ 38%]
tests/compute/test_sparse.py::test_spmm[int64-max-sub-shp5-g1] PASSED    [ 38%]
tests/compute/test_sparse.py::test_spmm[int64-max-mul-shp0-g0] PASSED    [ 38%]
tests/compute/test_sparse.py::test_spmm[int64-max-mul-shp0-g1] FAILED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-sub-shp0-g1] FAILED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-sub-shp1-g0] FAILED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-sub-shp1-g1] FAILED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-sub-shp2-g0] FAILED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-sub-shp2-g1] PASSED    [ 38%]
tests/compute/test_sparse.py::test_spmm[int64-max-mul-shp1-g0] PASSED    [ 38%]
tests/compute/test_sparse.py::test_spmm[int64-max-mul-shp1-g1] PASSED    [ 38%]
tests/compute/test_sparse.py::test_spmm[int64-max-mul-shp2-g0] PASSED    [ 38%]
tests/compute/test_sparse.py::test_spmm[int64-max-mul-shp2-g1] PASSED    [ 38%]
tests/compute/test_sparse.py::test_spmm[int64-max-mul-shp3-g0] PASSED    [ 38%]
tests/compute/test_sparse.py::test_spmm[int64-max-mul-shp3-g1] PASSED    [ 38%]
tests/compute/test_sparse.py::test_spmm[int64-max-mul-shp4-g0] PASSED    [ 38%]
tests/compute/test_sparse.py::test_spmm[int64-max-mul-shp4-g1] PASSED    [ 38%]
tests/compute/test_sparse.py::test_spmm[int64-max-mul-shp5-g0] PASSED    [ 38%]
tests/compute/test_sparse.py::test_spmm[int64-max-mul-shp5-g1] PASSED    [ 38%]
tests/compute/test_sparse.py::test_spmm[int64-max-div-shp0-g0] FAILED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-sub-shp3-g0] FAILED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-sub-shp3-g1] FAILED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-sub-shp4-g0] FAILED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-sub-shp4-g1] FAILED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-sub-shp5-g0] PASSED    [ 38%]
tests/compute/test_sparse.py::test_spmm[int64-max-div-shp0-g1] PASSED    [ 38%]
tests/compute/test_sparse.py::test_spmm[int64-max-div-shp1-g0] PASSED    [ 39%]
tests/compute/test_sparse.py::test_spmm[int64-max-div-shp1-g1] PASSED    [ 39%]
tests/compute/test_sparse.py::test_spmm[int64-max-div-shp2-g0] PASSED    [ 39%]
tests/compute/test_sparse.py::test_spmm[int64-max-div-shp2-g1] PASSED    [ 39%]
tests/compute/test_sparse.py::test_spmm[int64-max-div-shp3-g0] PASSED    [ 39%]
tests/compute/test_sparse.py::test_spmm[int64-max-div-shp3-g1] PASSED    [ 39%]
tests/compute/test_sparse.py::test_spmm[int64-max-div-shp4-g0] PASSED    [ 39%]
tests/compute/test_sparse.py::test_spmm[int64-max-div-shp4-g1] PASSED    [ 39%]
tests/compute/test_sparse.py::test_spmm[int64-max-div-shp5-g0] PASSED    [ 39%]
tests/compute/test_sparse.py::test_spmm[int64-max-div-shp5-g1] PASSED    [ 39%]
tests/compute/test_sparse.py::test_spmm[int64-max-copy_lhs-shp0-g0] FAILED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-sub-shp5-g1] FAILED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-mul-shp0-g0] FAILED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-mul-shp0-g1] FAILED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-mul-shp1-g0] PASSED [ 39%]
tests/compute/test_sparse.py::test_spmm[int64-max-copy_lhs-shp0-g1] PASSED [ 39%]
tests/compute/test_sparse.py::test_spmm[int64-max-copy_lhs-shp1-g0] PASSED [ 39%]
tests/compute/test_sparse.py::test_spmm[int64-max-copy_lhs-shp1-g1] PASSED [ 39%]
tests/compute/test_sparse.py::test_spmm[int64-max-copy_lhs-shp2-g0] PASSED [ 39%]
tests/compute/test_sparse.py::test_spmm[int64-max-copy_lhs-shp2-g1] PASSED [ 39%]
tests/compute/test_sparse.py::test_spmm[int64-max-copy_lhs-shp3-g0] PASSED [ 39%]
tests/compute/test_sparse.py::test_spmm[int64-max-copy_lhs-shp3-g1] PASSED [ 39%]
tests/compute/test_sparse.py::test_spmm[int64-max-copy_lhs-shp4-g0] PASSED [ 39%]
tests/compute/test_sparse.py::test_spmm[int64-max-copy_lhs-shp4-g1] PASSED [ 39%]
tests/compute/test_sparse.py::test_spmm[int64-max-copy_lhs-shp5-g0] PASSED [ 39%]
tests/compute/test_sparse.py::test_spmm[int64-max-copy_lhs-shp5-g1] PASSED [ 39%]
tests/compute/test_sparse.py::test_spmm[int64-max-copy_rhs-shp0-g0] PASSED [ 39%]
tests/compute/test_sparse.py::test_spmm[int64-max-copy_rhs-shp0-g1] FAILED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-mul-shp1-g1] FAILED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-mul-shp2-g0] FAILED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-mul-shp2-g1] FAILED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-mul-shp3-g0] FAILED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-mul-shp3-g1] PASSED [ 39%]
tests/compute/test_sparse.py::test_spmm[int64-max-copy_rhs-shp1-g0] PASSED [ 39%]
tests/compute/test_sparse.py::test_spmm[int64-max-copy_rhs-shp1-g1] PASSED [ 40%]
tests/compute/test_sparse.py::test_spmm[int64-max-copy_rhs-shp2-g0] PASSED [ 40%]
tests/compute/test_sparse.py::test_spmm[int64-max-copy_rhs-shp2-g1] PASSED [ 40%]
tests/compute/test_sparse.py::test_spmm[int64-max-copy_rhs-shp3-g0] PASSED [ 40%]
tests/compute/test_sparse.py::test_spmm[int64-max-copy_rhs-shp3-g1] PASSED [ 40%]
tests/compute/test_sparse.py::test_spmm[int64-max-copy_rhs-shp4-g0] PASSED [ 40%]
tests/compute/test_sparse.py::test_spmm[int64-max-copy_rhs-shp4-g1] PASSED [ 40%]
tests/compute/test_sparse.py::test_spmm[int64-max-copy_rhs-shp5-g0] PASSED [ 40%]
tests/compute/test_sparse.py::test_spmm[int64-max-copy_rhs-shp5-g1] PASSED [ 40%]
tests/compute/test_sparse.py::test_sddmm[int32-add-u-u-shp0-g0] PASSED   [ 40%]
tests/compute/test_sparse.py::test_sddmm[int32-add-u-u-shp0-g1] PASSED   [ 40%]
tests/compute/test_sparse.py::test_sddmm[int32-add-u-u-shp1-g0] PASSED   [ 40%]
tests/compute/test_sparse.py::test_sddmm[int32-add-u-u-shp1-g1] PASSED   [ 40%]
tests/compute/test_sparse.py::test_sddmm[int32-add-u-u-shp2-g0] PASSED   [ 40%]
tests/compute/test_sparse.py::test_sddmm[int32-add-u-u-shp2-g1] PASSED   [ 40%]
tests/compute/test_sparse.py::test_sddmm[int32-add-u-u-shp3-g0] PASSED   [ 40%]
tests/compute/test_sparse.py::test_sddmm[int32-add-u-u-shp3-g1] PASSED   [ 40%]
tests/compute/test_sparse.py::test_sddmm[int32-add-u-u-shp4-g0] PASSED   [ 40%]
tests/compute/test_sparse.py::test_sddmm[int32-add-u-u-shp4-g1] PASSED   [ 40%]
tests/compute/test_sparse.py::test_sddmm[int32-add-u-v-shp0-g0] PASSED   [ 40%]
tests/compute/test_sparse.py::test_sddmm[int32-add-u-v-shp0-g1] PASSED   [ 40%]
tests/compute/test_sparse.py::test_sddmm[int32-add-u-v-shp1-g0] PASSED   [ 40%]
tests/compute/test_sparse.py::test_sddmm[int32-add-u-v-shp1-g1] PASSED   [ 40%]
tests/compute/test_sparse.py::test_sddmm[int32-add-u-v-shp2-g0] PASSED   [ 40%]
tests/compute/test_sparse.py::test_sddmm[int32-add-u-v-shp2-g1] PASSED   [ 41%]
tests/compute/test_sparse.py::test_sddmm[int32-add-u-v-shp3-g0] PASSED   [ 41%]
tests/compute/test_sparse.py::test_sddmm[int32-add-u-v-shp3-g1] FAILED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-mul-shp4-g0] FAILED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-mul-shp4-g1] FAILED  [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-mul-shp5-g0] FAILED  [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-mul-shp5-g1] FAILED  [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-div-shp0-g0] PASSED   [ 41%]
tests/compute/test_sparse.py::test_sddmm[int32-add-u-v-shp4-g0] PASSED   [ 41%]
tests/compute/test_sparse.py::test_sddmm[int32-add-u-v-shp4-g1] PASSED   [ 41%]
tests/compute/test_sparse.py::test_sddmm[int32-add-u-e-shp0-g0] PASSED   [ 41%]
tests/compute/test_sparse.py::test_sddmm[int32-add-u-e-shp0-g1] PASSED   [ 41%]
tests/compute/test_sparse.py::test_sddmm[int32-add-u-e-shp1-g0] PASSED   [ 41%]
tests/compute/test_sparse.py::test_sddmm[int32-add-u-e-shp1-g1] PASSED   [ 41%]
tests/compute/test_sparse.py::test_sddmm[int32-add-u-e-shp2-g0] PASSED   [ 41%]
tests/compute/test_sparse.py::test_sddmm[int32-add-u-e-shp2-g1] PASSED   [ 41%]
tests/compute/test_sparse.py::test_sddmm[int32-add-u-e-shp3-g0] PASSED   [ 41%]
tests/compute/test_sparse.py::test_sddmm[int32-add-u-e-shp3-g1] PASSED   [ 41%]
tests/compute/test_sparse.py::test_sddmm[int32-add-u-e-shp4-g0] PASSED   [ 41%]
tests/compute/test_sparse.py::test_sddmm[int32-add-u-e-shp4-g1] PASSED   [ 41%]
tests/compute/test_sparse.py::test_sddmm[int32-add-v-u-shp0-g0] PASSED   [ 41%]
tests/compute/test_sparse.py::test_sddmm[int32-add-v-u-shp0-g1] PASSED   [ 41%]
tests/compute/test_sparse.py::test_sddmm[int32-add-v-u-shp1-g0] PASSED   [ 41%]
tests/compute/test_sparse.py::test_sddmm[int32-add-v-u-shp1-g1] PASSED   [ 41%]
tests/compute/test_sparse.py::test_sddmm[int32-add-v-u-shp2-g0] PASSED   [ 41%]
tests/compute/test_sparse.py::test_sddmm[int32-add-v-u-shp2-g1] PASSED   [ 41%]
tests/compute/test_sparse.py::test_sddmm[int32-add-v-u-shp3-g0] PASSED   [ 41%]
tests/compute/test_sparse.py::test_sddmm[int32-add-v-u-shp3-g1] PASSED   [ 41%]
tests/compute/test_sparse.py::test_sddmm[int32-add-v-u-shp4-g0] PASSED   [ 41%]
tests/compute/test_sparse.py::test_sddmm[int32-add-v-u-shp4-g1] PASSED   [ 41%]
tests/compute/test_sparse.py::test_sddmm[int32-add-v-v-shp0-g0] PASSED   [ 42%]
tests/compute/test_sparse.py::test_sddmm[int32-add-v-v-shp0-g1] PASSED   [ 42%]
tests/compute/test_sparse.py::test_sddmm[int32-add-v-v-shp1-g0] PASSED   [ 42%]
tests/compute/test_sparse.py::test_sddmm[int32-add-v-v-shp1-g1] PASSED   [ 42%]
tests/compute/test_sparse.py::test_sddmm[int32-add-v-v-shp2-g0] PASSED   [ 42%]
tests/compute/test_sparse.py::test_sddmm[int32-add-v-v-shp2-g1] PASSED   [ 42%]
tests/compute/test_sparse.py::test_sddmm[int32-add-v-v-shp3-g0] PASSED   [ 42%]
tests/compute/test_sparse.py::test_sddmm[int32-add-v-v-shp3-g1] PASSED   [ 42%]
tests/compute/test_sparse.py::test_sddmm[int32-add-v-v-shp4-g0] PASSED   [ 42%]
tests/compute/test_sparse.py::test_sddmm[int32-add-v-v-shp4-g1] PASSED   [ 42%]
tests/compute/test_sparse.py::test_sddmm[int32-add-v-e-shp0-g0] PASSED   [ 42%]
tests/compute/test_sparse.py::test_sddmm[int32-add-v-e-shp0-g1] PASSED   [ 42%]
tests/compute/test_sparse.py::test_sddmm[int32-add-v-e-shp1-g0] PASSED   [ 42%]
tests/compute/test_sparse.py::test_sddmm[int32-add-v-e-shp1-g1] PASSED   [ 42%]
tests/compute/test_sparse.py::test_sddmm[int32-add-v-e-shp2-g0] PASSED   [ 42%]
tests/compute/test_sparse.py::test_sddmm[int32-add-v-e-shp2-g1] PASSED   [ 42%]
tests/compute/test_sparse.py::test_sddmm[int32-add-v-e-shp3-g0] PASSED   [ 42%]
tests/compute/test_sparse.py::test_sddmm[int32-add-v-e-shp3-g1] PASSED   [ 42%]
tests/compute/test_sparse.py::test_sddmm[int32-add-v-e-shp4-g0] PASSED   [ 42%]
tests/compute/test_sparse.py::test_sddmm[int32-add-v-e-shp4-g1] PASSED   [ 42%]
tests/compute/test_sparse.py::test_sddmm[int32-add-e-u-shp0-g0] PASSED   [ 42%]
tests/compute/test_sparse.py::test_sddmm[int32-add-e-u-shp0-g1] PASSED   [ 42%]
tests/compute/test_sparse.py::test_sddmm[int32-add-e-u-shp1-g0] PASSED   [ 42%]
tests/compute/test_sparse.py::test_sddmm[int32-add-e-u-shp1-g1] FAILED  [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-div-shp0-g1] FAILED  [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-div-shp1-g0] FAILED  [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-div-shp1-g1] FAILED  [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-div-shp2-g0] FAILED  [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-div-shp2-g1] PASSED   [ 42%]
tests/compute/test_sparse.py::test_sddmm[int32-add-e-u-shp2-g0] PASSED   [ 43%]
tests/compute/test_sparse.py::test_sddmm[int32-add-e-u-shp2-g1] PASSED   [ 43%]
tests/compute/test_sparse.py::test_sddmm[int32-add-e-u-shp3-g0] PASSED   [ 43%]
tests/compute/test_sparse.py::test_sddmm[int32-add-e-u-shp3-g1] PASSED   [ 43%]
tests/compute/test_sparse.py::test_sddmm[int32-add-e-u-shp4-g0] PASSED   [ 43%]
tests/compute/test_sparse.py::test_sddmm[int32-add-e-u-shp4-g1] PASSED   [ 43%]
tests/compute/test_sparse.py::test_sddmm[int32-add-e-v-shp0-g0] PASSED   [ 43%]
tests/compute/test_sparse.py::test_sddmm[int32-add-e-v-shp0-g1] PASSED   [ 43%]
tests/compute/test_sparse.py::test_sddmm[int32-add-e-v-shp1-g0] PASSED   [ 43%]
tests/compute/test_sparse.py::test_sddmm[int32-add-e-v-shp1-g1] PASSED   [ 43%]
tests/compute/test_sparse.py::test_sddmm[int32-add-e-v-shp2-g0] PASSED   [ 43%]
tests/compute/test_sparse.py::test_sddmm[int32-add-e-v-shp2-g1] PASSED   [ 43%]
tests/compute/test_sparse.py::test_sddmm[int32-add-e-v-shp3-g0] PASSED   [ 43%]
tests/compute/test_sparse.py::test_sddmm[int32-add-e-v-shp3-g1] PASSED   [ 43%]
tests/compute/test_sparse.py::test_sddmm[int32-add-e-v-shp4-g0] PASSED   [ 43%]
tests/compute/test_sparse.py::test_sddmm[int32-add-e-v-shp4-g1] PASSED   [ 43%]
tests/compute/test_sparse.py::test_sddmm[int32-add-e-e-shp0-g0] PASSED   [ 43%]
tests/compute/test_sparse.py::test_sddmm[int32-add-e-e-shp0-g1] PASSED   [ 43%]
tests/compute/test_sparse.py::test_sddmm[int32-add-e-e-shp1-g0] PASSED   [ 43%]
tests/compute/test_sparse.py::test_sddmm[int32-add-e-e-shp1-g1] PASSED   [ 43%]
tests/compute/test_sparse.py::test_sddmm[int32-add-e-e-shp2-g0] PASSED   [ 43%]
tests/compute/test_sparse.py::test_sddmm[int32-add-e-e-shp2-g1] PASSED   [ 43%]
tests/compute/test_sparse.py::test_sddmm[int32-add-e-e-shp3-g0] PASSED   [ 43%]
tests/compute/test_sparse.py::test_sddmm[int32-add-e-e-shp3-g1] PASSED   [ 43%]
tests/compute/test_sparse.py::test_sddmm[int32-add-e-e-shp4-g0] PASSED   [ 43%]
tests/compute/test_sparse.py::test_sddmm[int32-add-e-e-shp4-g1] PASSED   [ 44%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-u-u-shp0-g0] PASSED   [ 44%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-u-u-shp0-g1] PASSED   [ 44%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-u-u-shp1-g0] PASSED   [ 44%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-u-u-shp1-g1] PASSED   [ 44%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-u-u-shp2-g0] PASSED   [ 44%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-u-u-shp2-g1] PASSED   [ 44%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-u-u-shp3-g0] PASSED   [ 44%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-u-u-shp3-g1] PASSED   [ 44%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-u-u-shp4-g0] PASSED   [ 44%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-u-u-shp4-g1] PASSED   [ 44%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-u-v-shp0-g0] PASSED   [ 44%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-u-v-shp0-g1] PASSED   [ 44%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-u-v-shp1-g0] PASSED   [ 44%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-u-v-shp1-g1] PASSED   [ 44%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-u-v-shp2-g0] PASSED   [ 44%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-u-v-shp2-g1] PASSED   [ 44%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-u-v-shp3-g0] PASSED   [ 44%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-u-v-shp3-g1] PASSED   [ 44%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-u-v-shp4-g0] PASSED   [ 44%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-u-v-shp4-g1] PASSED   [ 44%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-u-e-shp0-g0] PASSED   [ 44%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-u-e-shp0-g1] PASSED   [ 44%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-u-e-shp1-g0] PASSED   [ 44%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-u-e-shp1-g1] PASSED   [ 45%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-u-e-shp2-g0] PASSED   [ 45%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-u-e-shp2-g1] PASSED   [ 45%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-u-e-shp3-g0] PASSED   [ 45%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-u-e-shp3-g1] PASSED   [ 45%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-u-e-shp4-g0] PASSED   [ 45%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-u-e-shp4-g1] FAILED  [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-div-shp3-g0] FAILED  [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-div-shp3-g1] FAILED  [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-div-shp4-g0] FAILED  [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-div-shp4-g1] FAILED  [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-div-shp5-g0] PASSED   [ 45%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-v-u-shp0-g0] PASSED   [ 45%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-v-u-shp0-g1] PASSED   [ 45%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-v-u-shp1-g0] PASSED   [ 45%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-v-u-shp1-g1] PASSED   [ 45%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-v-u-shp2-g0] PASSED   [ 45%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-v-u-shp2-g1] PASSED   [ 45%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-v-u-shp3-g0] PASSED   [ 45%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-v-u-shp3-g1] PASSED   [ 45%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-v-u-shp4-g0] PASSED   [ 45%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-v-u-shp4-g1] PASSED   [ 45%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-v-v-shp0-g0] PASSED   [ 45%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-v-v-shp0-g1] PASSED   [ 45%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-v-v-shp1-g0] PASSED   [ 45%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-v-v-shp1-g1] PASSED   [ 45%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-v-v-shp2-g0] PASSED   [ 45%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-v-v-shp2-g1] PASSED   [ 45%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-v-v-shp3-g0] PASSED   [ 45%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-v-v-shp3-g1] PASSED   [ 45%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-v-v-shp4-g0] PASSED   [ 46%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-v-v-shp4-g1] PASSED   [ 46%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-v-e-shp0-g0] PASSED   [ 46%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-v-e-shp0-g1] PASSED   [ 46%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-v-e-shp1-g0] PASSED   [ 46%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-v-e-shp1-g1] PASSED   [ 46%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-v-e-shp2-g0] PASSED   [ 46%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-v-e-shp2-g1] PASSED   [ 46%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-v-e-shp3-g0] PASSED   [ 46%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-v-e-shp3-g1] PASSED   [ 46%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-v-e-shp4-g0] PASSED   [ 46%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-v-e-shp4-g1] PASSED   [ 46%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-e-u-shp0-g0] PASSED   [ 46%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-e-u-shp0-g1] PASSED   [ 46%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-e-u-shp1-g0] PASSED   [ 46%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-e-u-shp1-g1] PASSED   [ 46%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-e-u-shp2-g0] PASSED   [ 46%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-e-u-shp2-g1] PASSED   [ 46%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-e-u-shp3-g0] PASSED   [ 46%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-e-u-shp3-g1] PASSED   [ 46%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-e-u-shp4-g0] PASSED   [ 46%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-e-u-shp4-g1] PASSED   [ 46%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-e-v-shp0-g0] PASSED   [ 46%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-e-v-shp0-g1] PASSED   [ 46%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-e-v-shp1-g0] PASSED   [ 47%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-e-v-shp1-g1] PASSED   [ 47%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-e-v-shp2-g0] PASSED   [ 47%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-e-v-shp2-g1] FAILED  [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-div-shp5-g1] FAILED  [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_lhs-shp0-g0] FAILED [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_lhs-shp0-g1] FAILED [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_lhs-shp1-g0] FAILED [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_lhs-shp1-g1] PASSED   [ 47%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-e-v-shp3-g0] PASSED   [ 47%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-e-v-shp3-g1] PASSED   [ 47%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-e-v-shp4-g0] PASSED   [ 47%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-e-v-shp4-g1] PASSED   [ 47%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-e-e-shp0-g0] PASSED   [ 47%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-e-e-shp0-g1] PASSED   [ 47%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-e-e-shp1-g0] PASSED   [ 47%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-e-e-shp1-g1] PASSED   [ 47%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-e-e-shp2-g0] PASSED   [ 47%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-e-e-shp2-g1] PASSED   [ 47%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-e-e-shp3-g0] PASSED   [ 47%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-e-e-shp3-g1] PASSED   [ 47%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-e-e-shp4-g0] PASSED   [ 47%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-e-e-shp4-g1] PASSED   [ 47%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-u-u-shp0-g0] PASSED   [ 47%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-u-u-shp0-g1] PASSED   [ 47%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-u-u-shp1-g0] PASSED   [ 47%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-u-u-shp1-g1] PASSED   [ 47%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-u-u-shp2-g0] PASSED   [ 47%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-u-u-shp2-g1] PASSED   [ 47%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-u-u-shp3-g0] PASSED   [ 47%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-u-u-shp3-g1] PASSED   [ 48%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-u-u-shp4-g0] PASSED   [ 48%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-u-u-shp4-g1] PASSED   [ 48%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-u-v-shp0-g0] PASSED   [ 48%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-u-v-shp0-g1] PASSED   [ 48%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-u-v-shp1-g0] PASSED   [ 48%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-u-v-shp1-g1] PASSED   [ 48%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-u-v-shp2-g0] PASSED   [ 48%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-u-v-shp2-g1] PASSED   [ 48%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-u-v-shp3-g0] PASSED   [ 48%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-u-v-shp3-g1] PASSED   [ 48%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-u-v-shp4-g0] PASSED   [ 48%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-u-v-shp4-g1] PASSED   [ 48%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-u-e-shp0-g0] PASSED   [ 48%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-u-e-shp0-g1] PASSED   [ 48%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-u-e-shp1-g0] PASSED   [ 48%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-u-e-shp1-g1] PASSED   [ 48%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-u-e-shp2-g0] PASSED   [ 48%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-u-e-shp2-g1] PASSED   [ 48%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-u-e-shp3-g0] PASSED   [ 48%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-u-e-shp3-g1] PASSED   [ 48%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-u-e-shp4-g0] PASSED   [ 48%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-u-e-shp4-g1] PASSED   [ 48%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-v-u-shp0-g0] PASSED   [ 48%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-v-u-shp0-g1] PASSED   [ 49%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-v-u-shp1-g0] PASSED   [ 49%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-v-u-shp1-g1] PASSED   [ 49%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-v-u-shp2-g0] PASSED   [ 49%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-v-u-shp2-g1] PASSED   [ 49%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-v-u-shp3-g0] PASSED   [ 49%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-v-u-shp3-g1] PASSED   [ 49%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-v-u-shp4-g0] PASSED   [ 49%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-v-u-shp4-g1] PASSED   [ 49%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-v-v-shp0-g0] PASSED   [ 49%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-v-v-shp0-g1] PASSED   [ 49%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-v-v-shp1-g0] FAILED [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_lhs-shp2-g0] FAILED [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_lhs-shp2-g1] FAILED [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_lhs-shp3-g0] FAILED [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_lhs-shp3-g1] FAILED [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_lhs-shp4-g0] PASSED   [ 49%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-v-v-shp1-g1] PASSED   [ 49%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-v-v-shp2-g0] PASSED   [ 49%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-v-v-shp2-g1] PASSED   [ 49%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-v-v-shp3-g0] PASSED   [ 49%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-v-v-shp3-g1] PASSED   [ 49%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-v-v-shp4-g0] PASSED   [ 49%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-v-v-shp4-g1] PASSED   [ 49%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-v-e-shp0-g0] PASSED   [ 49%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-v-e-shp0-g1] PASSED   [ 49%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-v-e-shp1-g0] PASSED   [ 49%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-v-e-shp1-g1] PASSED   [ 49%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-v-e-shp2-g0] PASSED   [ 49%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-v-e-shp2-g1] PASSED   [ 49%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-v-e-shp3-g0] PASSED   [ 50%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-v-e-shp3-g1] PASSED   [ 50%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-v-e-shp4-g0] PASSED   [ 50%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-v-e-shp4-g1] PASSED   [ 50%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-e-u-shp0-g0] PASSED   [ 50%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-e-u-shp0-g1] PASSED   [ 50%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-e-u-shp1-g0] PASSED   [ 50%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-e-u-shp1-g1] PASSED   [ 50%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-e-u-shp2-g0] PASSED   [ 50%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-e-u-shp2-g1] PASSED   [ 50%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-e-u-shp3-g0] PASSED   [ 50%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-e-u-shp3-g1] PASSED   [ 50%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-e-u-shp4-g0] PASSED   [ 50%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-e-u-shp4-g1] PASSED   [ 50%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-e-v-shp0-g0] PASSED   [ 50%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-e-v-shp0-g1] PASSED   [ 50%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-e-v-shp1-g0] PASSED   [ 50%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-e-v-shp1-g1] PASSED   [ 50%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-e-v-shp2-g0] PASSED   [ 50%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-e-v-shp2-g1] PASSED   [ 50%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-e-v-shp3-g0] PASSED   [ 50%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-e-v-shp3-g1] PASSED   [ 50%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-e-v-shp4-g0] PASSED   [ 50%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-e-v-shp4-g1] PASSED   [ 50%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-e-e-shp0-g0] PASSED   [ 50%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-e-e-shp0-g1] PASSED   [ 51%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-e-e-shp1-g0] PASSED   [ 51%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-e-e-shp1-g1] PASSED   [ 51%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-e-e-shp2-g0] PASSED   [ 51%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-e-e-shp2-g1] PASSED   [ 51%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-e-e-shp3-g0] PASSED   [ 51%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-e-e-shp3-g1] PASSED   [ 51%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-e-e-shp4-g0] PASSED   [ 51%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-e-e-shp4-g1] PASSED   [ 51%]
tests/compute/test_sparse.py::test_sddmm[int32-div-u-u-shp0-g0] PASSED   [ 51%]
tests/compute/test_sparse.py::test_sddmm[int32-div-u-u-shp0-g1] PASSED   [ 51%]
tests/compute/test_sparse.py::test_sddmm[int32-div-u-u-shp1-g0] PASSED   [ 51%]
tests/compute/test_sparse.py::test_sddmm[int32-div-u-u-shp1-g1] PASSED   [ 51%]
tests/compute/test_sparse.py::test_sddmm[int32-div-u-u-shp2-g0] PASSED   [ 51%]
tests/compute/test_sparse.py::test_sddmm[int32-div-u-u-shp2-g1] PASSED   [ 51%]
tests/compute/test_sparse.py::test_sddmm[int32-div-u-u-shp3-g0] PASSED   [ 51%]
tests/compute/test_sparse.py::test_sddmm[int32-div-u-u-shp3-g1] PASSED   [ 51%]
tests/compute/test_sparse.py::test_sddmm[int32-div-u-u-shp4-g0] PASSED   [ 51%]
tests/compute/test_sparse.py::test_sddmm[int32-div-u-u-shp4-g1] PASSED   [ 51%]
tests/compute/test_sparse.py::test_sddmm[int32-div-u-v-shp0-g0] PASSED   [ 51%]
tests/compute/test_sparse.py::test_sddmm[int32-div-u-v-shp0-g1] PASSED   [ 51%]
tests/compute/test_sparse.py::test_sddmm[int32-div-u-v-shp1-g0] PASSED   [ 51%]
tests/compute/test_sparse.py::test_sddmm[int32-div-u-v-shp1-g1] FAILED [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_lhs-shp4-g1] FAILED [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_lhs-shp5-g0] FAILED [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_lhs-shp5-g1] FAILED [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_rhs-shp0-g0] FAILED [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_rhs-shp0-g1] PASSED   [ 51%]
tests/compute/test_sparse.py::test_sddmm[int32-div-u-v-shp2-g0] PASSED   [ 51%]
tests/compute/test_sparse.py::test_sddmm[int32-div-u-v-shp2-g1] PASSED   [ 52%]
tests/compute/test_sparse.py::test_sddmm[int32-div-u-v-shp3-g0] PASSED   [ 52%]
tests/compute/test_sparse.py::test_sddmm[int32-div-u-v-shp3-g1] PASSED   [ 52%]
tests/compute/test_sparse.py::test_sddmm[int32-div-u-v-shp4-g0] PASSED   [ 52%]
tests/compute/test_sparse.py::test_sddmm[int32-div-u-v-shp4-g1] PASSED   [ 52%]
tests/compute/test_sparse.py::test_sddmm[int32-div-u-e-shp0-g0] PASSED   [ 52%]
tests/compute/test_sparse.py::test_sddmm[int32-div-u-e-shp0-g1] PASSED   [ 52%]
tests/compute/test_sparse.py::test_sddmm[int32-div-u-e-shp1-g0] PASSED   [ 52%]
tests/compute/test_sparse.py::test_sddmm[int32-div-u-e-shp1-g1] PASSED   [ 52%]
tests/compute/test_sparse.py::test_sddmm[int32-div-u-e-shp2-g0] PASSED   [ 52%]
tests/compute/test_sparse.py::test_sddmm[int32-div-u-e-shp2-g1] PASSED   [ 52%]
tests/compute/test_sparse.py::test_sddmm[int32-div-u-e-shp3-g0] PASSED   [ 52%]
tests/compute/test_sparse.py::test_sddmm[int32-div-u-e-shp3-g1] PASSED   [ 52%]
tests/compute/test_sparse.py::test_sddmm[int32-div-u-e-shp4-g0] PASSED   [ 52%]
tests/compute/test_sparse.py::test_sddmm[int32-div-u-e-shp4-g1] PASSED   [ 52%]
tests/compute/test_sparse.py::test_sddmm[int32-div-v-u-shp0-g0] PASSED   [ 52%]
tests/compute/test_sparse.py::test_sddmm[int32-div-v-u-shp0-g1] PASSED   [ 52%]
tests/compute/test_sparse.py::test_sddmm[int32-div-v-u-shp1-g0] PASSED   [ 52%]
tests/compute/test_sparse.py::test_sddmm[int32-div-v-u-shp1-g1] PASSED   [ 52%]
tests/compute/test_sparse.py::test_sddmm[int32-div-v-u-shp2-g0] PASSED   [ 52%]
tests/compute/test_sparse.py::test_sddmm[int32-div-v-u-shp2-g1] PASSED   [ 52%]
tests/compute/test_sparse.py::test_sddmm[int32-div-v-u-shp3-g0] PASSED   [ 52%]
tests/compute/test_sparse.py::test_sddmm[int32-div-v-u-shp3-g1] PASSED   [ 52%]
tests/compute/test_sparse.py::test_sddmm[int32-div-v-u-shp4-g0] PASSED   [ 52%]
tests/compute/test_sparse.py::test_sddmm[int32-div-v-u-shp4-g1] PASSED   [ 52%]
tests/compute/test_sparse.py::test_sddmm[int32-div-v-v-shp0-g0] PASSED   [ 53%]
tests/compute/test_sparse.py::test_sddmm[int32-div-v-v-shp0-g1] PASSED   [ 53%]
tests/compute/test_sparse.py::test_sddmm[int32-div-v-v-shp1-g0] PASSED   [ 53%]
tests/compute/test_sparse.py::test_sddmm[int32-div-v-v-shp1-g1] PASSED   [ 53%]
tests/compute/test_sparse.py::test_sddmm[int32-div-v-v-shp2-g0] PASSED   [ 53%]
tests/compute/test_sparse.py::test_sddmm[int32-div-v-v-shp2-g1] PASSED   [ 53%]
tests/compute/test_sparse.py::test_sddmm[int32-div-v-v-shp3-g0] PASSED   [ 53%]
tests/compute/test_sparse.py::test_sddmm[int32-div-v-v-shp3-g1] PASSED   [ 53%]
tests/compute/test_sparse.py::test_sddmm[int32-div-v-v-shp4-g0] PASSED   [ 53%]
tests/compute/test_sparse.py::test_sddmm[int32-div-v-v-shp4-g1] PASSED   [ 53%]
tests/compute/test_sparse.py::test_sddmm[int32-div-v-e-shp0-g0] PASSED   [ 53%]
tests/compute/test_sparse.py::test_sddmm[int32-div-v-e-shp0-g1] PASSED   [ 53%]
tests/compute/test_sparse.py::test_sddmm[int32-div-v-e-shp1-g0] PASSED   [ 53%]
tests/compute/test_sparse.py::test_sddmm[int32-div-v-e-shp1-g1] PASSED   [ 53%]
tests/compute/test_sparse.py::test_sddmm[int32-div-v-e-shp2-g0] PASSED   [ 53%]
tests/compute/test_sparse.py::test_sddmm[int32-div-v-e-shp2-g1] PASSED   [ 53%]
tests/compute/test_sparse.py::test_sddmm[int32-div-v-e-shp3-g0] PASSED   [ 53%]
tests/compute/test_sparse.py::test_sddmm[int32-div-v-e-shp3-g1] PASSED   [ 53%]
tests/compute/test_sparse.py::test_sddmm[int32-div-v-e-shp4-g0] FAILED [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_rhs-shp1-g0] FAILED [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_rhs-shp1-g1] FAILED [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_rhs-shp2-g0] FAILED [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_rhs-shp2-g1] FAILED [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_rhs-shp3-g0] PASSED   [ 53%]
tests/compute/test_sparse.py::test_sddmm[int32-div-v-e-shp4-g1] PASSED   [ 53%]
tests/compute/test_sparse.py::test_sddmm[int32-div-e-u-shp0-g0] PASSED   [ 53%]
tests/compute/test_sparse.py::test_sddmm[int32-div-e-u-shp0-g1] PASSED   [ 53%]
tests/compute/test_sparse.py::test_sddmm[int32-div-e-u-shp1-g0] PASSED   [ 53%]
tests/compute/test_sparse.py::test_sddmm[int32-div-e-u-shp1-g1] PASSED   [ 53%]
tests/compute/test_sparse.py::test_sddmm[int32-div-e-u-shp2-g0] PASSED   [ 54%]
tests/compute/test_sparse.py::test_sddmm[int32-div-e-u-shp2-g1] PASSED   [ 54%]
tests/compute/test_sparse.py::test_sddmm[int32-div-e-u-shp3-g0] PASSED   [ 54%]
tests/compute/test_sparse.py::test_sddmm[int32-div-e-u-shp3-g1] PASSED   [ 54%]
tests/compute/test_sparse.py::test_sddmm[int32-div-e-u-shp4-g0] PASSED   [ 54%]
tests/compute/test_sparse.py::test_sddmm[int32-div-e-u-shp4-g1] PASSED   [ 54%]
tests/compute/test_sparse.py::test_sddmm[int32-div-e-v-shp0-g0] PASSED   [ 54%]
tests/compute/test_sparse.py::test_sddmm[int32-div-e-v-shp0-g1] PASSED   [ 54%]
tests/compute/test_sparse.py::test_sddmm[int32-div-e-v-shp1-g0] PASSED   [ 54%]
tests/compute/test_sparse.py::test_sddmm[int32-div-e-v-shp1-g1] PASSED   [ 54%]
tests/compute/test_sparse.py::test_sddmm[int32-div-e-v-shp2-g0] PASSED   [ 54%]
tests/compute/test_sparse.py::test_sddmm[int32-div-e-v-shp2-g1] PASSED   [ 54%]
tests/compute/test_sparse.py::test_sddmm[int32-div-e-v-shp3-g0] PASSED   [ 54%]
tests/compute/test_sparse.py::test_sddmm[int32-div-e-v-shp3-g1] PASSED   [ 54%]
tests/compute/test_sparse.py::test_sddmm[int32-div-e-v-shp4-g0] PASSED   [ 54%]
tests/compute/test_sparse.py::test_sddmm[int32-div-e-v-shp4-g1] PASSED   [ 54%]
tests/compute/test_sparse.py::test_sddmm[int32-div-e-e-shp0-g0] PASSED   [ 54%]
tests/compute/test_sparse.py::test_sddmm[int32-div-e-e-shp0-g1] PASSED   [ 54%]
tests/compute/test_sparse.py::test_sddmm[int32-div-e-e-shp1-g0] PASSED   [ 54%]
tests/compute/test_sparse.py::test_sddmm[int32-div-e-e-shp1-g1] PASSED   [ 54%]
tests/compute/test_sparse.py::test_sddmm[int32-div-e-e-shp2-g0] PASSED   [ 54%]
tests/compute/test_sparse.py::test_sddmm[int32-div-e-e-shp2-g1] PASSED   [ 54%]
tests/compute/test_sparse.py::test_sddmm[int32-div-e-e-shp3-g0] PASSED   [ 54%]
tests/compute/test_sparse.py::test_sddmm[int32-div-e-e-shp3-g1] PASSED   [ 54%]
tests/compute/test_sparse.py::test_sddmm[int32-div-e-e-shp4-g0] PASSED   [ 54%]
tests/compute/test_sparse.py::test_sddmm[int32-div-e-e-shp4-g1] PASSED   [ 55%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-u-u-shp0-g0] PASSED   [ 55%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-u-u-shp0-g1] PASSED   [ 55%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-u-u-shp1-g0] PASSED   [ 55%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-u-u-shp1-g1] PASSED   [ 55%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-u-u-shp2-g0] PASSED   [ 55%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-u-u-shp2-g1] PASSED   [ 55%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-u-u-shp3-g0] PASSED   [ 55%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-u-u-shp3-g1] PASSED   [ 55%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-u-u-shp4-g0] PASSED   [ 55%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-u-u-shp4-g1] PASSED   [ 55%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-u-v-shp0-g0] PASSED   [ 55%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-u-v-shp0-g1] PASSED   [ 55%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-u-v-shp1-g0] PASSED   [ 55%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-u-v-shp1-g1] PASSED   [ 55%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-u-v-shp2-g0] PASSED   [ 55%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-u-v-shp2-g1] PASSED   [ 55%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-u-v-shp3-g0] PASSED   [ 55%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-u-v-shp3-g1] PASSED   [ 55%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-u-v-shp4-g0] PASSED   [ 55%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-u-v-shp4-g1] PASSED   [ 55%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-u-e-shp0-g0] PASSED   [ 55%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-u-e-shp0-g1] PASSED   [ 55%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-u-e-shp1-g0] PASSED   [ 55%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-u-e-shp1-g1] PASSED   [ 56%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-u-e-shp2-g0] PASSED   [ 56%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-u-e-shp2-g1] PASSED   [ 56%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-u-e-shp3-g0] PASSED   [ 56%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-u-e-shp3-g1] PASSED   [ 56%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-u-e-shp4-g0] PASSED   [ 56%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-u-e-shp4-g1] PASSED   [ 56%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-v-u-shp0-g0] PASSED   [ 56%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-v-u-shp0-g1] PASSED   [ 56%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-v-u-shp1-g0] PASSED   [ 56%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-v-u-shp1-g1] PASSED   [ 56%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-v-u-shp2-g0] PASSED   [ 56%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-v-u-shp2-g1] PASSED   [ 56%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-v-u-shp3-g0] PASSED   [ 56%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-v-u-shp3-g1] PASSED   [ 56%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-v-u-shp4-g0] PASSED   [ 56%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-v-u-shp4-g1] PASSED   [ 56%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-v-v-shp0-g0] PASSED   [ 56%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-v-v-shp0-g1] PASSED   [ 56%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-v-v-shp1-g0] PASSED   [ 56%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-v-v-shp1-g1] PASSED   [ 56%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-v-v-shp2-g0] PASSED   [ 56%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-v-v-shp2-g1] PASSED   [ 56%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-v-v-shp3-g0] PASSED   [ 56%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-v-v-shp3-g1] PASSED   [ 56%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-v-v-shp4-g0] PASSED   [ 57%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-v-v-shp4-g1] PASSED   [ 57%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-v-e-shp0-g0] PASSED   [ 57%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-v-e-shp0-g1] PASSED   [ 57%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-v-e-shp1-g0] PASSED   [ 57%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-v-e-shp1-g1] PASSED   [ 57%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-v-e-shp2-g0] PASSED   [ 57%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-v-e-shp2-g1] PASSED   [ 57%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-v-e-shp3-g0] PASSED   [ 57%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-v-e-shp3-g1] PASSED   [ 57%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-v-e-shp4-g0] PASSED   [ 57%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-v-e-shp4-g1] PASSED   [ 57%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-e-u-shp0-g0] PASSED   [ 57%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-e-u-shp0-g1] PASSED   [ 57%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-e-u-shp1-g0] PASSED   [ 57%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-e-u-shp1-g1] PASSED   [ 57%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-e-u-shp2-g0] PASSED   [ 57%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-e-u-shp2-g1] PASSED   [ 57%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-e-u-shp3-g0] PASSED   [ 57%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-e-u-shp3-g1] PASSED   [ 57%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-e-u-shp4-g0] PASSED   [ 57%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-e-u-shp4-g1] PASSED   [ 57%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-e-v-shp0-g0] PASSED   [ 57%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-e-v-shp0-g1] FAILED [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_rhs-shp3-g1] FAILED [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_rhs-shp4-g0] FAILED [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_rhs-shp4-g1] FAILED [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_rhs-shp5-g0] FAILED [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_rhs-shp5-g1] PASSED   [ 57%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-e-v-shp1-g0] PASSED   [ 58%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-e-v-shp1-g1] PASSED   [ 58%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-e-v-shp2-g0] PASSED   [ 58%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-e-v-shp2-g1] PASSED   [ 58%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-e-v-shp3-g0] PASSED   [ 58%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-e-v-shp3-g1] PASSED   [ 58%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-e-v-shp4-g0] PASSED   [ 58%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-e-v-shp4-g1] PASSED   [ 58%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-e-e-shp0-g0] PASSED   [ 58%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-e-e-shp0-g1] PASSED   [ 58%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-e-e-shp1-g0] PASSED   [ 58%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-e-e-shp1-g1] PASSED   [ 58%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-e-e-shp2-g0] PASSED   [ 58%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-e-e-shp2-g1] PASSED   [ 58%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-e-e-shp3-g0] PASSED   [ 58%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-e-e-shp3-g1] PASSED   [ 58%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-e-e-shp4-g0] PASSED   [ 58%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-e-e-shp4-g1] PASSED   [ 58%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-u-u-shp0-g0] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-u-u-shp0-g1] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-u-u-shp1-g0] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-u-u-shp1-g1] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-u-u-shp2-g0] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-u-u-shp2-g1] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-u-u-shp3-g0] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-u-u-shp3-g1] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-u-u-shp4-g0] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-u-u-shp4-g1] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-u-v-shp0-g0] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-u-v-shp0-g1] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-u-v-shp1-g0] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-u-v-shp1-g1] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-u-v-shp2-g0] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-u-v-shp2-g1] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-u-v-shp3-g0] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-u-v-shp3-g1] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-u-v-shp4-g0] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-u-v-shp4-g1] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-u-e-shp0-g0] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-u-e-shp0-g1] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-u-e-shp1-g0] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-u-e-shp1-g1] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-u-e-shp2-g0] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-u-e-shp2-g1] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-u-e-shp3-g0] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-u-e-shp3-g1] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-u-e-shp4-g0] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-u-e-shp4-g1] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-v-u-shp0-g0] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-v-u-shp0-g1] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-v-u-shp1-g0] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-v-u-shp1-g1] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-v-u-shp2-g0] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-v-u-shp2-g1] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-v-u-shp3-g0] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-v-u-shp3-g1] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-v-u-shp4-g0] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-v-u-shp4-g1] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-v-v-shp0-g0] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-v-v-shp0-g1] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-v-v-shp1-g0] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-v-v-shp1-g1] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-v-v-shp2-g0] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-v-v-shp2-g1] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-v-v-shp3-g0] PASSED [ 60%]FAILED [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-add-shp0-g0] FAILED  [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-add-shp0-g1] FAILED  [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-add-shp1-g0] FAILED  [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-add-shp1-g1] FAILED  [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-add-shp2-g0] 
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-v-v-shp3-g1] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-v-v-shp4-g0] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-v-v-shp4-g1] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-v-e-shp0-g0] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-v-e-shp0-g1] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-v-e-shp1-g0] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-v-e-shp1-g1] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-v-e-shp2-g0] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-v-e-shp2-g1] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-v-e-shp3-g0] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-v-e-shp3-g1] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-v-e-shp4-g0] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-v-e-shp4-g1] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-e-u-shp0-g0] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-e-u-shp0-g1] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-e-u-shp1-g0] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-e-u-shp1-g1] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-e-u-shp2-g0] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-e-u-shp2-g1] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-e-u-shp3-g0] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-e-u-shp3-g1] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-e-u-shp4-g0] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-e-u-shp4-g1] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-e-v-shp0-g0] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-e-v-shp0-g1] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-e-v-shp1-g0] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-e-v-shp1-g1] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-e-v-shp2-g0] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-e-v-shp2-g1] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-e-v-shp3-g0] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-e-v-shp3-g1] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-e-v-shp4-g0] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-e-v-shp4-g1] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-e-e-shp0-g0] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-e-e-shp0-g1] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-e-e-shp1-g0] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-e-e-shp1-g1] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-e-e-shp2-g0] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-e-e-shp2-g1] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-e-e-shp3-g0] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-e-e-shp3-g1] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-e-e-shp4-g0] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-e-e-shp4-g1] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-u-u-shp0-g0] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-u-u-shp0-g1] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-u-u-shp1-g0] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-u-u-shp1-g1] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-u-u-shp2-g0] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-u-u-shp2-g1] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-u-u-shp3-g0] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-u-u-shp3-g1] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-u-u-shp4-g0] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-u-u-shp4-g1] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-u-v-shp0-g0] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-u-v-shp0-g1] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-u-v-shp1-g0] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-u-v-shp1-g1] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-u-v-shp2-g0] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-u-v-shp2-g1] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-u-v-shp3-g0] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-u-v-shp3-g1] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-u-v-shp4-g0] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-u-v-shp4-g1] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-u-e-shp0-g0] FAILED  [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-add-shp2-g1] FAILED  [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-add-shp3-g0] FAILED  [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-add-shp3-g1] FAILED  [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-add-shp4-g0] FAILED  [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-add-shp4-g1] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-u-e-shp0-g1] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-u-e-shp1-g0] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-u-e-shp1-g1] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-u-e-shp2-g0] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-u-e-shp2-g1] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-u-e-shp3-g0] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-u-e-shp3-g1] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-u-e-shp4-g0] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-u-e-shp4-g1] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-v-u-shp0-g0] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-v-u-shp0-g1] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-v-u-shp1-g0] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-v-u-shp1-g1] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-v-u-shp2-g0] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-v-u-shp2-g1] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-v-u-shp3-g0] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-v-u-shp3-g1] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-v-u-shp4-g0] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-v-u-shp4-g1] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-v-v-shp0-g0] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-v-v-shp0-g1] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-v-v-shp1-g0] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-v-v-shp1-g1] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-v-v-shp2-g0] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-v-v-shp2-g1] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-v-v-shp3-g0] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-v-v-shp3-g1] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-v-v-shp4-g0] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-v-v-shp4-g1] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-v-e-shp0-g0] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-v-e-shp0-g1] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-v-e-shp1-g0] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-v-e-shp1-g1] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-v-e-shp2-g0] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-v-e-shp2-g1] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-v-e-shp3-g0] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-v-e-shp3-g1] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-v-e-shp4-g0] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-v-e-shp4-g1] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-e-u-shp0-g0] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-e-u-shp0-g1] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-e-u-shp1-g0] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-e-u-shp1-g1] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-e-u-shp2-g0] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-e-u-shp2-g1] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-e-u-shp3-g0] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-e-u-shp3-g1] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-e-u-shp4-g0] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-e-u-shp4-g1] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-e-v-shp0-g0] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-e-v-shp0-g1] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-e-v-shp1-g0] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-e-v-shp1-g1] PASSED [ 65%]FAILED  [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-add-shp5-g0] FAILED  [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-add-shp5-g1] FAILED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-sub-shp0-g0] FAILED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-sub-shp0-g1] FAILED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-sub-shp1-g0] 
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-e-v-shp2-g0] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-e-v-shp2-g1] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-e-v-shp3-g0] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-e-v-shp3-g1] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-e-v-shp4-g0] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-e-v-shp4-g1] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-e-e-shp0-g0] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-e-e-shp0-g1] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-e-e-shp1-g0] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-e-e-shp1-g1] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-e-e-shp2-g0] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-e-e-shp2-g1] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-e-e-shp3-g0] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-e-e-shp3-g1] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-e-e-shp4-g0] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-e-e-shp4-g1] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[int64-add-u-u-shp0-g0] PASSED   [ 66%]
tests/compute/test_sparse.py::test_sddmm[int64-add-u-u-shp0-g1] PASSED   [ 66%]
tests/compute/test_sparse.py::test_sddmm[int64-add-u-u-shp1-g0] PASSED   [ 66%]
tests/compute/test_sparse.py::test_sddmm[int64-add-u-u-shp1-g1] PASSED   [ 66%]
tests/compute/test_sparse.py::test_sddmm[int64-add-u-u-shp2-g0] PASSED   [ 66%]
tests/compute/test_sparse.py::test_sddmm[int64-add-u-u-shp2-g1] PASSED   [ 66%]
tests/compute/test_sparse.py::test_sddmm[int64-add-u-u-shp3-g0] PASSED   [ 66%]
tests/compute/test_sparse.py::test_sddmm[int64-add-u-u-shp3-g1] PASSED   [ 66%]
tests/compute/test_sparse.py::test_sddmm[int64-add-u-u-shp4-g0] PASSED   [ 66%]
tests/compute/test_sparse.py::test_sddmm[int64-add-u-u-shp4-g1] PASSED   [ 66%]
tests/compute/test_sparse.py::test_sddmm[int64-add-u-v-shp0-g0] PASSED   [ 66%]
tests/compute/test_sparse.py::test_sddmm[int64-add-u-v-shp0-g1] PASSED   [ 66%]
tests/compute/test_sparse.py::test_sddmm[int64-add-u-v-shp1-g0] PASSED   [ 66%]
tests/compute/test_sparse.py::test_sddmm[int64-add-u-v-shp1-g1] PASSED   [ 66%]
tests/compute/test_sparse.py::test_sddmm[int64-add-u-v-shp2-g0] PASSED   [ 66%]
tests/compute/test_sparse.py::test_sddmm[int64-add-u-v-shp2-g1] PASSED   [ 66%]
tests/compute/test_sparse.py::test_sddmm[int64-add-u-v-shp3-g0] PASSED   [ 66%]
tests/compute/test_sparse.py::test_sddmm[int64-add-u-v-shp3-g1] PASSED   [ 66%]FAILED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-sub-shp1-g1] FAILED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-sub-shp2-g0] FAILED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-sub-shp2-g1] FAILED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-sub-shp3-g0] FAILED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-sub-shp3-g1] 
tests/compute/test_sparse.py::test_sddmm[int64-add-u-v-shp4-g0] PASSED   [ 66%]
tests/compute/test_sparse.py::test_sddmm[int64-add-u-v-shp4-g1] PASSED   [ 66%]
tests/compute/test_sparse.py::test_sddmm[int64-add-u-e-shp0-g0] PASSED   [ 66%]
tests/compute/test_sparse.py::test_sddmm[int64-add-u-e-shp0-g1] PASSED   [ 66%]
tests/compute/test_sparse.py::test_sddmm[int64-add-u-e-shp1-g0] PASSED   [ 66%]
tests/compute/test_sparse.py::test_sddmm[int64-add-u-e-shp1-g1] PASSED   [ 67%]
tests/compute/test_sparse.py::test_sddmm[int64-add-u-e-shp2-g0] PASSED   [ 67%]
tests/compute/test_sparse.py::test_sddmm[int64-add-u-e-shp2-g1] PASSED   [ 67%]
tests/compute/test_sparse.py::test_sddmm[int64-add-u-e-shp3-g0] PASSED   [ 67%]
tests/compute/test_sparse.py::test_sddmm[int64-add-u-e-shp3-g1] PASSED   [ 67%]
tests/compute/test_sparse.py::test_sddmm[int64-add-u-e-shp4-g0] PASSED   [ 67%]
tests/compute/test_sparse.py::test_sddmm[int64-add-u-e-shp4-g1] PASSED   [ 67%]
tests/compute/test_sparse.py::test_sddmm[int64-add-v-u-shp0-g0] PASSED   [ 67%]
tests/compute/test_sparse.py::test_sddmm[int64-add-v-u-shp0-g1] PASSED   [ 67%]
tests/compute/test_sparse.py::test_sddmm[int64-add-v-u-shp1-g0] PASSED   [ 67%]
tests/compute/test_sparse.py::test_sddmm[int64-add-v-u-shp1-g1] PASSED   [ 67%]
tests/compute/test_sparse.py::test_sddmm[int64-add-v-u-shp2-g0] PASSED   [ 67%]
tests/compute/test_sparse.py::test_sddmm[int64-add-v-u-shp2-g1] PASSED   [ 67%]
tests/compute/test_sparse.py::test_sddmm[int64-add-v-u-shp3-g0] PASSED   [ 67%]
tests/compute/test_sparse.py::test_sddmm[int64-add-v-u-shp3-g1] PASSED   [ 67%]
tests/compute/test_sparse.py::test_sddmm[int64-add-v-u-shp4-g0] PASSED   [ 67%]
tests/compute/test_sparse.py::test_sddmm[int64-add-v-u-shp4-g1] PASSED   [ 67%]
tests/compute/test_sparse.py::test_sddmm[int64-add-v-v-shp0-g0] PASSED   [ 67%]
tests/compute/test_sparse.py::test_sddmm[int64-add-v-v-shp0-g1] PASSED   [ 67%]
tests/compute/test_sparse.py::test_sddmm[int64-add-v-v-shp1-g0] PASSED   [ 67%]
tests/compute/test_sparse.py::test_sddmm[int64-add-v-v-shp1-g1] PASSED   [ 67%]
tests/compute/test_sparse.py::test_sddmm[int64-add-v-v-shp2-g0] PASSED   [ 67%]
tests/compute/test_sparse.py::test_sddmm[int64-add-v-v-shp2-g1] PASSED   [ 67%]
tests/compute/test_sparse.py::test_sddmm[int64-add-v-v-shp3-g0] PASSED   [ 67%]
tests/compute/test_sparse.py::test_sddmm[int64-add-v-v-shp3-g1] PASSED   [ 67%]
tests/compute/test_sparse.py::test_sddmm[int64-add-v-v-shp4-g0] PASSED   [ 68%]
tests/compute/test_sparse.py::test_sddmm[int64-add-v-v-shp4-g1] PASSED   [ 68%]
tests/compute/test_sparse.py::test_sddmm[int64-add-v-e-shp0-g0] PASSED   [ 68%]
tests/compute/test_sparse.py::test_sddmm[int64-add-v-e-shp0-g1] PASSED   [ 68%]
tests/compute/test_sparse.py::test_sddmm[int64-add-v-e-shp1-g0] PASSED   [ 68%]
tests/compute/test_sparse.py::test_sddmm[int64-add-v-e-shp1-g1] PASSED   [ 68%]
tests/compute/test_sparse.py::test_sddmm[int64-add-v-e-shp2-g0] PASSED   [ 68%]
tests/compute/test_sparse.py::test_sddmm[int64-add-v-e-shp2-g1] PASSED   [ 68%]
tests/compute/test_sparse.py::test_sddmm[int64-add-v-e-shp3-g0] PASSED   [ 68%]
tests/compute/test_sparse.py::test_sddmm[int64-add-v-e-shp3-g1] PASSED   [ 68%]
tests/compute/test_sparse.py::test_sddmm[int64-add-v-e-shp4-g0] PASSED   [ 68%]
tests/compute/test_sparse.py::test_sddmm[int64-add-v-e-shp4-g1] PASSED   [ 68%]
tests/compute/test_sparse.py::test_sddmm[int64-add-e-u-shp0-g0] PASSED   [ 68%]
tests/compute/test_sparse.py::test_sddmm[int64-add-e-u-shp0-g1] PASSED   [ 68%]
tests/compute/test_sparse.py::test_sddmm[int64-add-e-u-shp1-g0] PASSED   [ 68%]
tests/compute/test_sparse.py::test_sddmm[int64-add-e-u-shp1-g1] PASSED   [ 68%]
tests/compute/test_sparse.py::test_sddmm[int64-add-e-u-shp2-g0] PASSED   [ 68%]
tests/compute/test_sparse.py::test_sddmm[int64-add-e-u-shp2-g1] FAILED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-sub-shp4-g0] FAILED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-sub-shp4-g1] FAILED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-sub-shp5-g0] FAILED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-sub-shp5-g1] PASSED   [ 68%]
tests/compute/test_sparse.py::test_sddmm[int64-add-e-u-shp3-g0] PASSED   [ 68%]
tests/compute/test_sparse.py::test_sddmm[int64-add-e-u-shp3-g1] PASSED   [ 68%]
tests/compute/test_sparse.py::test_sddmm[int64-add-e-u-shp4-g0] PASSED   [ 68%]
tests/compute/test_sparse.py::test_sddmm[int64-add-e-u-shp4-g1] PASSED   [ 68%]
tests/compute/test_sparse.py::test_sddmm[int64-add-e-v-shp0-g0] PASSED   [ 68%]
tests/compute/test_sparse.py::test_sddmm[int64-add-e-v-shp0-g1] PASSED   [ 68%]
tests/compute/test_sparse.py::test_sddmm[int64-add-e-v-shp1-g0] PASSED   [ 69%]
tests/compute/test_sparse.py::test_sddmm[int64-add-e-v-shp1-g1] PASSED   [ 69%]
tests/compute/test_sparse.py::test_sddmm[int64-add-e-v-shp2-g0] PASSED   [ 69%]
tests/compute/test_sparse.py::test_sddmm[int64-add-e-v-shp2-g1] PASSED   [ 69%]
tests/compute/test_sparse.py::test_sddmm[int64-add-e-v-shp3-g0] PASSED   [ 69%]
tests/compute/test_sparse.py::test_sddmm[int64-add-e-v-shp3-g1] PASSED   [ 69%]
tests/compute/test_sparse.py::test_sddmm[int64-add-e-v-shp4-g0] PASSED   [ 69%]
tests/compute/test_sparse.py::test_sddmm[int64-add-e-v-shp4-g1] PASSED   [ 69%]
tests/compute/test_sparse.py::test_sddmm[int64-add-e-e-shp0-g0] PASSED   [ 69%]
tests/compute/test_sparse.py::test_sddmm[int64-add-e-e-shp0-g1] PASSED   [ 69%]
tests/compute/test_sparse.py::test_sddmm[int64-add-e-e-shp1-g0] PASSED   [ 69%]
tests/compute/test_sparse.py::test_sddmm[int64-add-e-e-shp1-g1] PASSED   [ 69%]
tests/compute/test_sparse.py::test_sddmm[int64-add-e-e-shp2-g0] PASSED   [ 69%]
tests/compute/test_sparse.py::test_sddmm[int64-add-e-e-shp2-g1] PASSED   [ 69%]
tests/compute/test_sparse.py::test_sddmm[int64-add-e-e-shp3-g0] PASSED   [ 69%]
tests/compute/test_sparse.py::test_sddmm[int64-add-e-e-shp3-g1] PASSED   [ 69%]
tests/compute/test_sparse.py::test_sddmm[int64-add-e-e-shp4-g0] PASSED   [ 69%]
tests/compute/test_sparse.py::test_sddmm[int64-add-e-e-shp4-g1] PASSED   [ 69%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-u-u-shp0-g0] PASSED   [ 69%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-u-u-shp0-g1] PASSED   [ 69%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-u-u-shp1-g0] PASSED   [ 69%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-u-u-shp1-g1] PASSED   [ 69%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-u-u-shp2-g0] PASSED   [ 69%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-u-u-shp2-g1] PASSED   [ 69%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-u-u-shp3-g0] PASSED   [ 69%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-u-u-shp3-g1] PASSED   [ 70%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-u-u-shp4-g0] PASSED   [ 70%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-u-u-shp4-g1] PASSED   [ 70%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-u-v-shp0-g0] PASSED   [ 70%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-u-v-shp0-g1] PASSED   [ 70%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-u-v-shp1-g0] PASSED   [ 70%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-u-v-shp1-g1] PASSED   [ 70%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-u-v-shp2-g0] PASSED   [ 70%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-u-v-shp2-g1] PASSED   [ 70%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-u-v-shp3-g0] PASSED   [ 70%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-u-v-shp3-g1] PASSED   [ 70%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-u-v-shp4-g0] PASSED   [ 70%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-u-v-shp4-g1] PASSED   [ 70%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-u-e-shp0-g0] PASSED   [ 70%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-u-e-shp0-g1] PASSED   [ 70%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-u-e-shp1-g0] PASSED   [ 70%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-u-e-shp1-g1] PASSED   [ 70%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-u-e-shp2-g0] PASSED   [ 70%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-u-e-shp2-g1] PASSED   [ 70%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-u-e-shp3-g0] PASSED   [ 70%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-u-e-shp3-g1] PASSED   [ 70%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-u-e-shp4-g0] PASSED   [ 70%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-u-e-shp4-g1] PASSED   [ 70%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-v-u-shp0-g0] PASSED   [ 70%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-v-u-shp0-g1] FAILED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-mul-shp0-g0] FAILED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-mul-shp0-g1] FAILED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-mul-shp1-g0] FAILED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-mul-shp1-g1] FAILED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-mul-shp2-g0] PASSED   [ 71%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-v-u-shp1-g0] PASSED   [ 71%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-v-u-shp1-g1] PASSED   [ 71%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-v-u-shp2-g0] PASSED   [ 71%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-v-u-shp2-g1] PASSED   [ 71%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-v-u-shp3-g0] PASSED   [ 71%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-v-u-shp3-g1] PASSED   [ 71%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-v-u-shp4-g0] PASSED   [ 71%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-v-u-shp4-g1] PASSED   [ 71%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-v-v-shp0-g0] PASSED   [ 71%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-v-v-shp0-g1] PASSED   [ 71%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-v-v-shp1-g0] PASSED   [ 71%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-v-v-shp1-g1] PASSED   [ 71%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-v-v-shp2-g0] PASSED   [ 71%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-v-v-shp2-g1] PASSED   [ 71%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-v-v-shp3-g0] PASSED   [ 71%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-v-v-shp3-g1] PASSED   [ 71%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-v-v-shp4-g0] PASSED   [ 71%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-v-v-shp4-g1] PASSED   [ 71%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-v-e-shp0-g0] PASSED   [ 71%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-v-e-shp0-g1] PASSED   [ 71%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-v-e-shp1-g0] PASSED   [ 71%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-v-e-shp1-g1] PASSED   [ 71%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-v-e-shp2-g0] PASSED   [ 71%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-v-e-shp2-g1] PASSED   [ 71%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-v-e-shp3-g0] PASSED   [ 72%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-v-e-shp3-g1] PASSED   [ 72%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-v-e-shp4-g0] PASSED   [ 72%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-v-e-shp4-g1] PASSED   [ 72%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-e-u-shp0-g0] PASSED   [ 72%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-e-u-shp0-g1] PASSED   [ 72%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-e-u-shp1-g0] PASSED   [ 72%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-e-u-shp1-g1] PASSED   [ 72%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-e-u-shp2-g0] PASSED   [ 72%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-e-u-shp2-g1] PASSED   [ 72%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-e-u-shp3-g0] PASSED   [ 72%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-e-u-shp3-g1] PASSED   [ 72%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-e-u-shp4-g0] PASSED   [ 72%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-e-u-shp4-g1] PASSED   [ 72%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-e-v-shp0-g0] PASSED   [ 72%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-e-v-shp0-g1] PASSED   [ 72%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-e-v-shp1-g0] PASSED   [ 72%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-e-v-shp1-g1] PASSED   [ 72%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-e-v-shp2-g0] PASSED   [ 72%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-e-v-shp2-g1] PASSED   [ 72%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-e-v-shp3-g0] PASSED   [ 72%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-e-v-shp3-g1] PASSED   [ 72%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-e-v-shp4-g0] FAILED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-mul-shp2-g1] FAILED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-mul-shp3-g0] FAILED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-mul-shp3-g1] FAILED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-mul-shp4-g0] FAILED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-mul-shp4-g1] Client [1090] waits on 172.17.0.3:57707
Machine (0) group (0) client (0) connect to server successfuly!
Client[0] in group[0] is exiting...
Server (0) shutdown.
Server is exiting...
PASSED   [ 72%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-e-v-shp4-g1] PASSED   [ 72%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-e-e-shp0-g0] PASSED   [ 73%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-e-e-shp0-g1] PASSED   [ 73%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-e-e-shp1-g0] PASSED   [ 73%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-e-e-shp1-g1] PASSED   [ 73%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-e-e-shp2-g0] PASSED   [ 73%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-e-e-shp2-g1] PASSED   [ 73%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-e-e-shp3-g0] PASSED   [ 73%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-e-e-shp3-g1] PASSED   [ 73%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-e-e-shp4-g0] PASSED   [ 73%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-e-e-shp4-g1] PASSED   [ 73%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-u-u-shp0-g0] PASSED   [ 73%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-u-u-shp0-g1] PASSED   [ 73%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-u-u-shp1-g0] PASSED   [ 73%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-u-u-shp1-g1] PASSED   [ 73%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-u-u-shp2-g0] PASSED   [ 73%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-u-u-shp2-g1] PASSED   [ 73%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-u-u-shp3-g0] PASSED   [ 73%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-u-u-shp3-g1] PASSED   [ 73%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-u-u-shp4-g0] PASSED   [ 73%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-u-u-shp4-g1] PASSED   [ 73%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-u-v-shp0-g0] PASSED   [ 73%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-u-v-shp0-g1] PASSED   [ 73%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-u-v-shp1-g0] PASSED   [ 73%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-u-v-shp1-g1] PASSED   [ 73%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-u-v-shp2-g0] PASSED   [ 73%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-u-v-shp2-g1] PASSED   [ 74%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-u-v-shp3-g0] PASSED   [ 74%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-u-v-shp3-g1] PASSED   [ 74%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-u-v-shp4-g0] PASSED   [ 74%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-u-v-shp4-g1] PASSED   [ 74%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-u-e-shp0-g0] PASSED   [ 74%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-u-e-shp0-g1] PASSED   [ 74%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-u-e-shp1-g0] PASSED   [ 74%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-u-e-shp1-g1] PASSED   [ 74%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-u-e-shp2-g0] PASSED   [ 74%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-u-e-shp2-g1] PASSED   [ 74%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-u-e-shp3-g0] PASSED   [ 74%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-u-e-shp3-g1] PASSED   [ 74%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-u-e-shp4-g0] PASSED   [ 74%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-u-e-shp4-g1] PASSED   [ 74%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-v-u-shp0-g0] PASSED   [ 74%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-v-u-shp0-g1] PASSED   [ 74%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-v-u-shp1-g0] PASSED   [ 74%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-v-u-shp1-g1] PASSED   [ 74%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-v-u-shp2-g0] PASSED   [ 74%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-v-u-shp2-g1] PASSED   [ 74%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-v-u-shp3-g0] PASSED   [ 74%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-v-u-shp3-g1] PASSED   [ 74%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-v-u-shp4-g0] PASSED   [ 74%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-v-u-shp4-g1] PASSED   [ 75%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-v-v-shp0-g0] PASSED   [ 75%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-v-v-shp0-g1] PASSED   [ 75%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-v-v-shp1-g0] PASSED   [ 75%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-v-v-shp1-g1] PASSED   [ 75%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-v-v-shp2-g0] PASSED   [ 75%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-v-v-shp2-g1] PASSED   [ 75%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-v-v-shp3-g0] PASSED   [ 75%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-v-v-shp3-g1] PASSED   [ 75%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-v-v-shp4-g0] PASSED   [ 75%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-v-v-shp4-g1] PASSED   [ 75%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-v-e-shp0-g0] PASSED   [ 75%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-v-e-shp0-g1] PASSED   [ 75%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-v-e-shp1-g0] PASSED   [ 75%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-v-e-shp1-g1] FAILED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-mul-shp5-g0] FAILED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-mul-shp5-g1] FAILED  [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-div-shp0-g0] FAILED  [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-div-shp0-g1] FAILED  [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-div-shp1-g0] PASSED   [ 75%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-v-e-shp2-g0] PASSED   [ 75%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-v-e-shp2-g1] PASSED   [ 75%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-v-e-shp3-g0] PASSED   [ 75%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-v-e-shp3-g1] PASSED   [ 75%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-v-e-shp4-g0] PASSED   [ 75%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-v-e-shp4-g1] PASSED   [ 75%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-e-u-shp0-g0] PASSED   [ 75%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-e-u-shp0-g1] PASSED   [ 75%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-e-u-shp1-g0] PASSED   [ 75%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-e-u-shp1-g1] PASSED   [ 75%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-e-u-shp2-g0] PASSED   [ 76%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-e-u-shp2-g1] PASSED   [ 76%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-e-u-shp3-g0] PASSED   [ 76%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-e-u-shp3-g1] PASSED   [ 76%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-e-u-shp4-g0] PASSED   [ 76%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-e-u-shp4-g1] PASSED   [ 76%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-e-v-shp0-g0] PASSED   [ 76%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-e-v-shp0-g1] PASSED   [ 76%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-e-v-shp1-g0] PASSED   [ 76%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-e-v-shp1-g1] PASSED   [ 76%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-e-v-shp2-g0] PASSED   [ 76%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-e-v-shp2-g1] PASSED   [ 76%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-e-v-shp3-g0] PASSED   [ 76%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-e-v-shp3-g1] PASSED   [ 76%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-e-v-shp4-g0] PASSED   [ 76%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-e-v-shp4-g1] PASSED   [ 76%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-e-e-shp0-g0] PASSED   [ 76%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-e-e-shp0-g1] PASSED   [ 76%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-e-e-shp1-g0] PASSED   [ 76%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-e-e-shp1-g1] PASSED   [ 76%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-e-e-shp2-g0] PASSED   [ 76%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-e-e-shp2-g1] PASSED   [ 76%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-e-e-shp3-g0] PASSED   [ 76%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-e-e-shp3-g1] PASSED   [ 76%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-e-e-shp4-g0] PASSED   [ 77%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-e-e-shp4-g1] PASSED   [ 77%]
tests/compute/test_sparse.py::test_sddmm[int64-div-u-u-shp0-g0] PASSED   [ 77%]
tests/compute/test_sparse.py::test_sddmm[int64-div-u-u-shp0-g1] PASSED   [ 77%]
tests/compute/test_sparse.py::test_sddmm[int64-div-u-u-shp1-g0] PASSED   [ 77%]
tests/compute/test_sparse.py::test_sddmm[int64-div-u-u-shp1-g1] PASSED   [ 77%]
tests/compute/test_sparse.py::test_sddmm[int64-div-u-u-shp2-g0] PASSED   [ 77%]
tests/compute/test_sparse.py::test_sddmm[int64-div-u-u-shp2-g1] PASSED   [ 77%]
tests/compute/test_sparse.py::test_sddmm[int64-div-u-u-shp3-g0] PASSED   [ 77%]
tests/compute/test_sparse.py::test_sddmm[int64-div-u-u-shp3-g1] PASSED   [ 77%]
tests/compute/test_sparse.py::test_sddmm[int64-div-u-u-shp4-g0] PASSED   [ 77%]
tests/compute/test_sparse.py::test_sddmm[int64-div-u-u-shp4-g1] PASSED   [ 77%]
tests/compute/test_sparse.py::test_sddmm[int64-div-u-v-shp0-g0] PASSED   [ 77%]
tests/compute/test_sparse.py::test_sddmm[int64-div-u-v-shp0-g1] PASSED   [ 77%]
tests/compute/test_sparse.py::test_sddmm[int64-div-u-v-shp1-g0] PASSED   [ 77%]
tests/compute/test_sparse.py::test_sddmm[int64-div-u-v-shp1-g1] PASSED   [ 77%]
tests/compute/test_sparse.py::test_sddmm[int64-div-u-v-shp2-g0] PASSED   [ 77%]
tests/compute/test_sparse.py::test_sddmm[int64-div-u-v-shp2-g1] PASSED   [ 77%]
tests/compute/test_sparse.py::test_sddmm[int64-div-u-v-shp3-g0] PASSED   [ 77%]
tests/compute/test_sparse.py::test_sddmm[int64-div-u-v-shp3-g1] PASSED   [ 77%]
tests/compute/test_sparse.py::test_sddmm[int64-div-u-v-shp4-g0] FAILED  [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-div-shp1-g1] FAILED  [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-div-shp2-g0] FAILED  [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-div-shp2-g1] FAILED  [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-div-shp3-g0] FAILED  [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-div-shp3-g1] Done sampling
Server (1) shutdown.
Server is exiting...
Converting to homogeneous graph takes 0.001s, peak mem: 1.529 GB
Convert a graph into a bidirected graph: 0.001 seconds, peak memory: 1.529 GB
Construct multi-constraint weights: 0.000 seconds, peak memory: 1.529 GB
[05:32:41] /root/jenkins/workspace/dgl_PR-4648/src/graph/transform/metis_partition_hetero.cc:87: Partition a graph with 630 nodes and 23980 edges into 2 parts and get 4875 edge cuts
Metis partitioning: 0.006 seconds, peak memory: 1.529 GB
Assigning nodes to METIS partitions takes 0.007s, peak mem: 1.529 GB
Reshuffle nodes and edges: 0.001 seconds
Split the graph: 0.001 seconds
Construct subgraphs: 0.003 seconds
Splitting the graph into partitions takes 0.005s, peak mem: 1.529 GB
part 0 has 200 nodes of type n1 and 108 are inside the partition
part 0 has 190 nodes of type n2 and 101 are inside the partition
part 0 has 212 nodes of type n3 and 113 are inside the partition
part 0 has 2824 edges of type r12 and 2054 are inside the partition
part 0 has 3108 edges of type r13 and 2219 are inside the partition
part 0 has 2907 edges of type r23 and 2078 are inside the partition
part 1 has 210 nodes of type n1 and 102 are inside the partition
part 1 has 200 nodes of type n2 and 99 are inside the partition
part 1 has 218 nodes of type n3 and 107 are inside the partition
part 1 has 2530 edges of type r12 and 1746 are inside the partition
part 1 has 2809 edges of type r13 and 1981 are inside the partition
part 1 has 2687 edges of type r23 and 1912 are inside the partition
Save partitions: 0.003 seconds, peak memory: 1.529 GB
There are 11990 edges in the graph and 0 edge cuts for 2 partitions.
PASSED   [ 77%]
tests/compute/test_sparse.py::test_sddmm[int64-div-u-v-shp4-g1] PASSED   [ 77%]
tests/compute/test_sparse.py::test_sddmm[int64-div-u-e-shp0-g0] PASSED   [ 77%]
tests/compute/test_sparse.py::test_sddmm[int64-div-u-e-shp0-g1] PASSED   [ 77%]
tests/compute/test_sparse.py::test_sddmm[int64-div-u-e-shp1-g0] PASSED   [ 77%]
tests/compute/test_sparse.py::test_sddmm[int64-div-u-e-shp1-g1] PASSED   [ 78%]
tests/compute/test_sparse.py::test_sddmm[int64-div-u-e-shp2-g0] PASSED   [ 78%]
tests/compute/test_sparse.py::test_sddmm[int64-div-u-e-shp2-g1] PASSED   [ 78%]
tests/compute/test_sparse.py::test_sddmm[int64-div-u-e-shp3-g0] PASSED   [ 78%]
tests/compute/test_sparse.py::test_sddmm[int64-div-u-e-shp3-g1] PASSED   [ 78%]
tests/compute/test_sparse.py::test_sddmm[int64-div-u-e-shp4-g0] PASSED   [ 78%]
tests/compute/test_sparse.py::test_sddmm[int64-div-u-e-shp4-g1] PASSED   [ 78%]
tests/compute/test_sparse.py::test_sddmm[int64-div-v-u-shp0-g0] PASSED   [ 78%]
tests/compute/test_sparse.py::test_sddmm[int64-div-v-u-shp0-g1] PASSED   [ 78%]
tests/compute/test_sparse.py::test_sddmm[int64-div-v-u-shp1-g0] PASSED   [ 78%]
tests/compute/test_sparse.py::test_sddmm[int64-div-v-u-shp1-g1] PASSED   [ 78%]
tests/compute/test_sparse.py::test_sddmm[int64-div-v-u-shp2-g0] PASSED   [ 78%]
tests/compute/test_sparse.py::test_sddmm[int64-div-v-u-shp2-g1] PASSED   [ 78%]
tests/compute/test_sparse.py::test_sddmm[int64-div-v-u-shp3-g0] PASSED   [ 78%]
tests/compute/test_sparse.py::test_sddmm[int64-div-v-u-shp3-g1] PASSED   [ 78%]
tests/compute/test_sparse.py::test_sddmm[int64-div-v-u-shp4-g0] PASSED   [ 78%]
tests/compute/test_sparse.py::test_sddmm[int64-div-v-u-shp4-g1] PASSED   [ 78%]
tests/compute/test_sparse.py::test_sddmm[int64-div-v-v-shp0-g0] PASSED   [ 78%]
tests/compute/test_sparse.py::test_sddmm[int64-div-v-v-shp0-g1] PASSED   [ 78%]
tests/compute/test_sparse.py::test_sddmm[int64-div-v-v-shp1-g0] PASSED   [ 78%]
tests/compute/test_sparse.py::test_sddmm[int64-div-v-v-shp1-g1] PASSED   [ 78%]
tests/compute/test_sparse.py::test_sddmm[int64-div-v-v-shp2-g0] PASSED   [ 78%]
tests/compute/test_sparse.py::test_sddmm[int64-div-v-v-shp2-g1] PASSED   [ 78%]
tests/compute/test_sparse.py::test_sddmm[int64-div-v-v-shp3-g0] PASSED   [ 78%]
tests/compute/test_sparse.py::test_sddmm[int64-div-v-v-shp3-g1] PASSED   [ 79%]
tests/compute/test_sparse.py::test_sddmm[int64-div-v-v-shp4-g0] PASSED   [ 79%]
tests/compute/test_sparse.py::test_sddmm[int64-div-v-v-shp4-g1] PASSED   [ 79%]
tests/compute/test_sparse.py::test_sddmm[int64-div-v-e-shp0-g0] PASSED   [ 79%]
tests/compute/test_sparse.py::test_sddmm[int64-div-v-e-shp0-g1] PASSED   [ 79%]
tests/compute/test_sparse.py::test_sddmm[int64-div-v-e-shp1-g0] PASSED   [ 79%]
tests/compute/test_sparse.py::test_sddmm[int64-div-v-e-shp1-g1] PASSED   [ 79%]
tests/compute/test_sparse.py::test_sddmm[int64-div-v-e-shp2-g0] PASSED   [ 79%]
tests/compute/test_sparse.py::test_sddmm[int64-div-v-e-shp2-g1] PASSED   [ 79%]
tests/compute/test_sparse.py::test_sddmm[int64-div-v-e-shp3-g0] PASSED   [ 79%]
tests/compute/test_sparse.py::test_sddmm[int64-div-v-e-shp3-g1] PASSED   [ 79%]
tests/compute/test_sparse.py::test_sddmm[int64-div-v-e-shp4-g0] PASSED   [ 79%]
tests/compute/test_sparse.py::test_sddmm[int64-div-v-e-shp4-g1] PASSED   [ 79%]
tests/compute/test_sparse.py::test_sddmm[int64-div-e-u-shp0-g0] PASSED   [ 79%]
tests/compute/test_sparse.py::test_sddmm[int64-div-e-u-shp0-g1] PASSED   [ 79%]
tests/compute/test_sparse.py::test_sddmm[int64-div-e-u-shp1-g0] PASSED   [ 79%]
tests/compute/test_sparse.py::test_sddmm[int64-div-e-u-shp1-g1] FAILED  [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-div-shp4-g0] FAILED  [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-div-shp4-g1] FAILED  [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-div-shp5-g0] FAILED  [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-div-shp5-g1] FAILED  [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_lhs-shp0-g0] PASSED   [ 79%]
tests/compute/test_sparse.py::test_sddmm[int64-div-e-u-shp2-g0] PASSED   [ 79%]
tests/compute/test_sparse.py::test_sddmm[int64-div-e-u-shp2-g1] PASSED   [ 79%]
tests/compute/test_sparse.py::test_sddmm[int64-div-e-u-shp3-g0] PASSED   [ 79%]
tests/compute/test_sparse.py::test_sddmm[int64-div-e-u-shp3-g1] PASSED   [ 79%]
tests/compute/test_sparse.py::test_sddmm[int64-div-e-u-shp4-g0] PASSED   [ 79%]
tests/compute/test_sparse.py::test_sddmm[int64-div-e-u-shp4-g1] PASSED   [ 79%]
tests/compute/test_sparse.py::test_sddmm[int64-div-e-v-shp0-g0] PASSED   [ 79%]
tests/compute/test_sparse.py::test_sddmm[int64-div-e-v-shp0-g1] PASSED   [ 79%]
tests/compute/test_sparse.py::test_sddmm[int64-div-e-v-shp1-g0] PASSED   [ 80%]
tests/compute/test_sparse.py::test_sddmm[int64-div-e-v-shp1-g1] PASSED   [ 80%]
tests/compute/test_sparse.py::test_sddmm[int64-div-e-v-shp2-g0] PASSED   [ 80%]
tests/compute/test_sparse.py::test_sddmm[int64-div-e-v-shp2-g1] PASSED   [ 80%]
tests/compute/test_sparse.py::test_sddmm[int64-div-e-v-shp3-g0] PASSED   [ 80%]
tests/compute/test_sparse.py::test_sddmm[int64-div-e-v-shp3-g1] PASSED   [ 80%]
tests/compute/test_sparse.py::test_sddmm[int64-div-e-v-shp4-g0] PASSED   [ 80%]
tests/compute/test_sparse.py::test_sddmm[int64-div-e-v-shp4-g1] PASSED   [ 80%]
tests/compute/test_sparse.py::test_sddmm[int64-div-e-e-shp0-g0] PASSED   [ 80%]
tests/compute/test_sparse.py::test_sddmm[int64-div-e-e-shp0-g1] PASSED   [ 80%]
tests/compute/test_sparse.py::test_sddmm[int64-div-e-e-shp1-g0] PASSED   [ 80%]
tests/compute/test_sparse.py::test_sddmm[int64-div-e-e-shp1-g1] PASSED   [ 80%]
tests/compute/test_sparse.py::test_sddmm[int64-div-e-e-shp2-g0] PASSED   [ 80%]
tests/compute/test_sparse.py::test_sddmm[int64-div-e-e-shp2-g1] PASSED   [ 80%]
tests/compute/test_sparse.py::test_sddmm[int64-div-e-e-shp3-g0] PASSED   [ 80%]
tests/compute/test_sparse.py::test_sddmm[int64-div-e-e-shp3-g1] PASSED   [ 80%]
tests/compute/test_sparse.py::test_sddmm[int64-div-e-e-shp4-g0] PASSED   [ 80%]
tests/compute/test_sparse.py::test_sddmm[int64-div-e-e-shp4-g1] PASSED   [ 80%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-u-u-shp0-g0] PASSED   [ 80%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-u-u-shp0-g1] PASSED   [ 80%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-u-u-shp1-g0] PASSED   [ 80%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-u-u-shp1-g1] PASSED   [ 80%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-u-u-shp2-g0] PASSED   [ 80%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-u-u-shp2-g1] PASSED   [ 80%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-u-u-shp3-g0] PASSED   [ 81%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-u-u-shp3-g1] PASSED   [ 81%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-u-u-shp4-g0] PASSED   [ 81%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-u-u-shp4-g1] PASSED   [ 81%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-u-v-shp0-g0] PASSED   [ 81%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-u-v-shp0-g1] PASSED   [ 81%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-u-v-shp1-g0] PASSED   [ 81%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-u-v-shp1-g1] PASSED   [ 81%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-u-v-shp2-g0] PASSED   [ 81%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-u-v-shp2-g1] PASSED   [ 81%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-u-v-shp3-g0] PASSED   [ 81%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-u-v-shp3-g1] PASSED   [ 81%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-u-v-shp4-g0] PASSED   [ 81%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-u-v-shp4-g1] PASSED   [ 81%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-u-e-shp0-g0] PASSED   [ 81%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-u-e-shp0-g1] PASSED   [ 81%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-u-e-shp1-g0] PASSED   [ 81%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-u-e-shp1-g1] PASSED   [ 81%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-u-e-shp2-g0] PASSED   [ 81%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-u-e-shp2-g1] PASSED   [ 81%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-u-e-shp3-g0] PASSED   [ 81%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-u-e-shp3-g1] PASSED   [ 81%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-u-e-shp4-g0] PASSED   [ 81%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-u-e-shp4-g1] FAILED [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_lhs-shp0-g1] FAILED [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_lhs-shp1-g0] FAILED [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_lhs-shp1-g1] FAILED [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_lhs-shp2-g0] FAILED [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_lhs-shp2-g1] PASSED   [ 81%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-v-u-shp0-g0] PASSED   [ 81%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-v-u-shp0-g1] PASSED   [ 82%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-v-u-shp1-g0] PASSED   [ 82%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-v-u-shp1-g1] PASSED   [ 82%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-v-u-shp2-g0] PASSED   [ 82%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-v-u-shp2-g1] PASSED   [ 82%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-v-u-shp3-g0] PASSED   [ 82%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-v-u-shp3-g1] PASSED   [ 82%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-v-u-shp4-g0] PASSED   [ 82%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-v-u-shp4-g1] PASSED   [ 82%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-v-v-shp0-g0] PASSED   [ 82%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-v-v-shp0-g1] PASSED   [ 82%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-v-v-shp1-g0] PASSED   [ 82%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-v-v-shp1-g1] PASSED   [ 82%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-v-v-shp2-g0] PASSED   [ 82%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-v-v-shp2-g1] PASSED   [ 82%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-v-v-shp3-g0] PASSED   [ 82%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-v-v-shp3-g1] PASSED   [ 82%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-v-v-shp4-g0] PASSED   [ 82%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-v-v-shp4-g1] PASSED   [ 82%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-v-e-shp0-g0] PASSED   [ 82%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-v-e-shp0-g1] PASSED   [ 82%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-v-e-shp1-g0] PASSED   [ 82%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-v-e-shp1-g1] PASSED   [ 82%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-v-e-shp2-g0] PASSED   [ 82%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-v-e-shp2-g1] PASSED   [ 83%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-v-e-shp3-g0] PASSED   [ 83%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-v-e-shp3-g1] PASSED   [ 83%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-v-e-shp4-g0] PASSED   [ 83%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-v-e-shp4-g1] PASSED   [ 83%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-e-u-shp0-g0] PASSED   [ 83%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-e-u-shp0-g1] PASSED   [ 83%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-e-u-shp1-g0] PASSED   [ 83%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-e-u-shp1-g1] PASSED   [ 83%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-e-u-shp2-g0] PASSED   [ 83%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-e-u-shp2-g1] PASSED   [ 83%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-e-u-shp3-g0] PASSED   [ 83%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-e-u-shp3-g1] PASSED   [ 83%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-e-u-shp4-g0] PASSED   [ 83%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-e-u-shp4-g1] PASSED   [ 83%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-e-v-shp0-g0] PASSED   [ 83%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-e-v-shp0-g1] PASSED   [ 83%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-e-v-shp1-g0] PASSED   [ 83%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-e-v-shp1-g1] PASSED   [ 83%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-e-v-shp2-g0] PASSED   [ 83%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-e-v-shp2-g1] PASSED   [ 83%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-e-v-shp3-g0] PASSED   [ 83%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-e-v-shp3-g1] FAILED [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_lhs-shp3-g0] FAILED [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_lhs-shp3-g1] FAILED [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_lhs-shp4-g0] FAILED [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_lhs-shp4-g1] PASSED   [ 83%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-e-v-shp4-g0] PASSED   [ 83%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-e-v-shp4-g1] PASSED   [ 83%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-e-e-shp0-g0] PASSED   [ 84%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-e-e-shp0-g1] PASSED   [ 84%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-e-e-shp1-g0] PASSED   [ 84%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-e-e-shp1-g1] PASSED   [ 84%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-e-e-shp2-g0] PASSED   [ 84%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-e-e-shp2-g1] PASSED   [ 84%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-e-e-shp3-g0] PASSED   [ 84%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-e-e-shp3-g1] PASSED   [ 84%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-e-e-shp4-g0] PASSED   [ 84%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-e-e-shp4-g1] PASSED   [ 84%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-u-u-shp0-g0] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-u-u-shp0-g1] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-u-u-shp1-g0] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-u-u-shp1-g1] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-u-u-shp2-g0] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-u-u-shp2-g1] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-u-u-shp3-g0] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-u-u-shp3-g1] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-u-u-shp4-g0] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-u-u-shp4-g1] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-u-v-shp0-g0] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-u-v-shp0-g1] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-u-v-shp1-g0] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-u-v-shp1-g1] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-u-v-shp2-g0] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-u-v-shp2-g1] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-u-v-shp3-g0] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-u-v-shp3-g1] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-u-v-shp4-g0] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-u-v-shp4-g1] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-u-e-shp0-g0] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-u-e-shp0-g1] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-u-e-shp1-g0] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-u-e-shp1-g1] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-u-e-shp2-g0] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-u-e-shp2-g1] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-u-e-shp3-g0] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-u-e-shp3-g1] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-u-e-shp4-g0] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-u-e-shp4-g1] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-v-u-shp0-g0] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-v-u-shp0-g1] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-v-u-shp1-g0] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-v-u-shp1-g1] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-v-u-shp2-g0] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-v-u-shp2-g1] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-v-u-shp3-g0] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-v-u-shp3-g1] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-v-u-shp4-g0] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-v-u-shp4-g1] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-v-v-shp0-g0] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-v-v-shp0-g1] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-v-v-shp1-g0] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-v-v-shp1-g1] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-v-v-shp2-g0] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-v-v-shp2-g1] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-v-v-shp3-g0] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-v-v-shp3-g1] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-v-v-shp4-g0] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-v-v-shp4-g1] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-v-e-shp0-g0] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-v-e-shp0-g1] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-v-e-shp1-g0] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-v-e-shp1-g1] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-v-e-shp2-g0] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-v-e-shp2-g1] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-v-e-shp3-g0] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-v-e-shp3-g1] /root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_sampling
start graph service on server 0 for part 0
[05:32:42] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:32:42] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:18024]...
FAILED [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_lhs-shp5-g0] FAILED [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_lhs-shp5-g1] FAILED [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_rhs-shp0-g0] FAILED [ 40%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_rhs-shp0-g1] FAILED [ 40%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_rhs-shp1-g0] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-v-e-shp4-g0] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-v-e-shp4-g1] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-e-u-shp0-g0] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-e-u-shp0-g1] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-e-u-shp1-g0] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-e-u-shp1-g1] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-e-u-shp2-g0] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-e-u-shp2-g1] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-e-u-shp3-g0] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-e-u-shp3-g1] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-e-u-shp4-g0] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-e-u-shp4-g1] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-e-v-shp0-g0] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-e-v-shp0-g1] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-e-v-shp1-g0] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-e-v-shp1-g1] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-e-v-shp2-g0] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-e-v-shp2-g1] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-e-v-shp3-g0] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-e-v-shp3-g1] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-e-v-shp4-g0] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-e-v-shp4-g1] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-e-e-shp0-g0] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-e-e-shp0-g1] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-e-e-shp1-g0] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-e-e-shp1-g1] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-e-e-shp2-g0] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-e-e-shp2-g1] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-e-e-shp3-g0] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-e-e-shp3-g1] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-e-e-shp4-g0] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-e-e-shp4-g1] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-u-u-shp0-g0] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-u-u-shp0-g1] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-u-u-shp1-g0] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-u-u-shp1-g1] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-u-u-shp2-g0] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-u-u-shp2-g1] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-u-u-shp3-g0] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-u-u-shp3-g1] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-u-u-shp4-g0] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-u-u-shp4-g1] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-u-v-shp0-g0] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-u-v-shp0-g1] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-u-v-shp1-g0] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-u-v-shp1-g1] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-u-v-shp2-g0] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-u-v-shp2-g1] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-u-v-shp3-g0] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-u-v-shp3-g1] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-u-v-shp4-g0] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-u-v-shp4-g1] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-u-e-shp0-g0] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-u-e-shp0-g1] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-u-e-shp1-g0] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-u-e-shp1-g1] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-u-e-shp2-g0] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-u-e-shp2-g1] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-u-e-shp3-g0] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-u-e-shp3-g1] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-u-e-shp4-g0] FAILED [ 40%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_rhs-shp1-g1] FAILED [ 40%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_rhs-shp2-g0] FAILED [ 40%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_rhs-shp2-g1] FAILED [ 40%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_rhs-shp3-g0] FAILED [ 40%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_rhs-shp3-g1] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-u-e-shp4-g1] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-v-u-shp0-g0] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-v-u-shp0-g1] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-v-u-shp1-g0] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-v-u-shp1-g1] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-v-u-shp2-g0] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-v-u-shp2-g1] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-v-u-shp3-g0] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-v-u-shp3-g1] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-v-u-shp4-g0] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-v-u-shp4-g1] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-v-v-shp0-g0] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-v-v-shp0-g1] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-v-v-shp1-g0] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-v-v-shp1-g1] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-v-v-shp2-g0] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-v-v-shp2-g1] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-v-v-shp3-g0] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-v-v-shp3-g1] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-v-v-shp4-g0] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-v-v-shp4-g1] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-v-e-shp0-g0] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-v-e-shp0-g1] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-v-e-shp1-g0] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-v-e-shp1-g1] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-v-e-shp2-g0] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-v-e-shp2-g1] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-v-e-shp3-g0] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-v-e-shp3-g1] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-v-e-shp4-g0] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-v-e-shp4-g1] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-e-u-shp0-g0] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-e-u-shp0-g1] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-e-u-shp1-g0] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-e-u-shp1-g1] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-e-u-shp2-g0] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-e-u-shp2-g1] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-e-u-shp3-g0] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-e-u-shp3-g1] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-e-u-shp4-g0] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-e-u-shp4-g1] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-e-v-shp0-g0] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-e-v-shp0-g1] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-e-v-shp1-g0] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-e-v-shp1-g1] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-e-v-shp2-g0] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-e-v-shp2-g1] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-e-v-shp3-g0] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-e-v-shp3-g1] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-e-v-shp4-g0] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-e-v-shp4-g1] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-e-e-shp0-g0] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-e-e-shp0-g1] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-e-e-shp1-g0] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-e-e-shp1-g1] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-e-e-shp2-g0] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-e-e-shp2-g1] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-e-e-shp3-g0] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-e-e-shp3-g1] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-e-e-shp4-g0] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-e-e-shp4-g1] PASSED [ 91%]FAILED [ 40%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_rhs-shp4-g0] FAILED [ 40%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_rhs-shp4-g1] FAILED [ 40%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_rhs-shp5-g0] FAILED [ 40%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_rhs-shp5-g1] FAILED [ 40%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-u-shp0-g0] PASSED [ 40%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-u-shp0-g1] PASSED [ 40%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-u-shp1-g0] PASSED [ 40%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-u-shp1-g1] PASSED [ 40%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-u-shp2-g0] PASSED [ 40%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-u-shp2-g1] PASSED [ 40%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-u-shp3-g0] PASSED [ 40%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-u-shp3-g1] PASSED [ 40%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-u-shp4-g0] PASSED [ 40%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-u-shp4-g1] PASSED [ 40%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-v-shp0-g0] 
tests/compute/test_sparse.py::test_edge_softmax[int32-shp0-src-g0] PASSED [ 91%]
tests/compute/test_sparse.py::test_edge_softmax[int32-shp0-dst-g0] PASSED [ 91%]
tests/compute/test_sparse.py::test_edge_softmax[int32-shp1-src-g0] PASSED [ 91%]
tests/compute/test_sparse.py::test_edge_softmax[int32-shp1-dst-g0] PASSED [ 91%]
tests/compute/test_sparse.py::test_edge_softmax[int32-shp2-src-g0] PASSED [ 91%]
tests/compute/test_sparse.py::test_edge_softmax[int32-shp2-dst-g0] PASSED [ 91%]
tests/compute/test_sparse.py::test_edge_softmax[int64-shp0-src-g0] PASSED [ 92%]
tests/compute/test_sparse.py::test_edge_softmax[int64-shp0-dst-g0] PASSED [ 92%]
tests/compute/test_sparse.py::test_edge_softmax[int64-shp1-src-g0] PASSED [ 92%]
tests/compute/test_sparse.py::test_edge_softmax[int64-shp1-dst-g0] PASSED [ 92%]
tests/compute/test_sparse.py::test_edge_softmax[int64-shp2-src-g0] PASSED [ 92%]
tests/compute/test_sparse.py::test_edge_softmax[int64-shp2-dst-g0] PASSED [ 92%]
tests/compute/test_sparse.py::test_segment_reduce[sum] PASSED            [ 92%]
tests/compute/test_sparse.py::test_segment_reduce[max] PASSED            [ 92%]
tests/compute/test_sparse.py::test_segment_reduce[min] PASSED            [ 92%]
tests/compute/test_sparse.py::test_segment_reduce[mean] PASSED           [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype0-0.01-1-int32] SKIPPED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype0-0.01-1-int64] SKIPPED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype0-0.01-8-int32] SKIPPED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype0-0.01-8-int64] SKIPPED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype0-0.01-16-int32] SKIPPED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype0-0.01-16-int64] SKIPPED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype0-0.01-64-int32] SKIPPED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype0-0.01-64-int64] SKIPPED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype0-0.01-256-int32] SKIPPED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype0-0.01-256-int64] SKIPPED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype1-0.01-1-int32] SKIPPED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype1-0.01-1-int64] SKIPPED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype1-0.01-8-int32] SKIPPED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype1-0.01-8-int64] SKIPPED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype1-0.01-16-int32] SKIPPED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype1-0.01-16-int64] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype1-0.01-64-int32] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype1-0.01-64-int64] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype1-0.01-256-int32] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype1-0.01-256-int64] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype2-0.003-1-int32] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype2-0.003-1-int64] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype2-0.003-8-int32] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype2-0.003-8-int64] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype2-0.003-16-int32] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype2-0.003-16-int64] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype2-0.003-64-int32] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype2-0.003-64-int64] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype2-0.003-256-int32] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype2-0.003-256-int64] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype3-0.0001-1-int32] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype3-0.0001-1-int64] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype3-0.0001-8-int32] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype3-0.0001-8-int64] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype3-0.0001-16-int32] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype3-0.0001-16-int64] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype3-0.0001-64-int32] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype3-0.0001-64-int64] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype3-0.0001-256-int32] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype3-0.0001-256-int64] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype0-0.01-1] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype0-0.01-8] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype0-0.01-16] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype0-0.01-64] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype0-0.01-256] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype1-0.02-1] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype1-0.02-8] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype1-0.02-16] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype1-0.02-64] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype1-0.02-256] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype2-0.003-1] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype2-0.003-8] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype2-0.003-16] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype2-0.003-64] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype2-0.003-256] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype3-0.0001-1] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype3-0.0001-8] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype3-0.0001-16] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype3-0.0001-64] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype3-0.0001-256] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_use_libxsmm_switch SKIPPED (Only ...) [ 94%]
tests/compute/test_specialization.py::test_v2v_update_all[int32] PASSED  [ 94%]
tests/compute/test_specialization.py::test_v2v_update_all[int64] FAILED [ 40%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-v-shp0-g1] FAILED [ 40%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-v-shp1-g0] FAILED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-v-shp1-g1] FAILED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-v-shp2-g0] PASSED  [ 94%]
tests/compute/test_specialization.py::test_v2v_snr[int32] PASSED         [ 94%]
tests/compute/test_specialization.py::test_v2v_snr[int64] FAILED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-v-shp2-g1] FAILED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-v-shp3-g0] FAILED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-v-shp3-g1] FAILED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-v-shp4-g0] FAILED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-v-shp4-g1] PASSED         [ 95%]
tests/compute/test_specialization.py::test_v2v_pull[int32] PASSED        [ 95%]
tests/compute/test_specialization.py::test_v2v_pull[int64] PASSED        [ 95%]
tests/compute/test_specialization.py::test_update_all_multi_fallback[int32] PASSED [ 95%]
tests/compute/test_specialization.py::test_update_all_multi_fallback[int64] FAILED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-e-shp0-g0] FAILED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-e-shp0-g1] FAILED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-e-shp1-g0] FAILED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-e-shp1-g1] FAILED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-e-shp2-g0] /root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_sampling
start graph service on server 1 for part 1
[05:32:43] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:32:43] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:18026]...
[05:32:43] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:32:43] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
PASSED [ 95%]
tests/compute/test_specialization.py::test_pull_multi_fallback[int32] PASSED [ 95%]
tests/compute/test_specialization.py::test_pull_multi_fallback[int64] PASSED [ 95%]
tests/compute/test_specialization.py::test_spmv_3d_feat[int32] PASSED    [ 95%]
tests/compute/test_specialization.py::test_spmv_3d_feat[int64] FAILED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-e-shp2-g1] FAILED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-e-shp3-g0] FAILED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-e-shp3-g1] FAILED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-e-shp4-g0] PASSED    [ 95%]
tests/compute/test_subgraph.py::test_edge_subgraph PASSED                [ 95%]
tests/compute/test_subgraph.py::test_subgraph PASSED                     [ 95%]
tests/compute/test_subgraph.py::test_subgraph_mask[int32] SKIPPED (M...) [ 95%]
tests/compute/test_subgraph.py::test_subgraph_mask[int64] SKIPPED (M...) [ 95%]
tests/compute/test_subgraph.py::test_subgraph1[int32] PASSED             [ 95%]
tests/compute/test_subgraph.py::test_subgraph1[int64] PASSED             [ 95%]
tests/compute/test_subgraph.py::test_in_subgraph[int32] PASSED           [ 95%]FAILED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-e-shp4-g1] FAILED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-u-shp0-g0] FAILED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-u-shp0-g1] FAILED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-u-shp1-g0] FAILED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-u-shp1-g1] 
tests/compute/test_subgraph.py::test_in_subgraph[int64] PASSED           [ 95%]
tests/compute/test_subgraph.py::test_out_subgraph[int32] PASSED          [ 95%]
tests/compute/test_subgraph.py::test_out_subgraph[int64] PASSED          [ 95%]
tests/compute/test_subgraph.py::test_subgraph_message_passing PASSED     [ 95%]
tests/compute/test_subgraph.py::test_khop_in_subgraph[int32] PASSED      [ 95%]
tests/compute/test_subgraph.py::test_khop_in_subgraph[int64] PASSED      [ 95%]
tests/compute/test_subgraph.py::test_khop_out_subgraph[int32] PASSED     [ 95%]
tests/compute/test_subgraph.py::test_khop_out_subgraph[int64] PASSED     [ 95%]
tests/compute/test_subgraph.py::test_subframes[child_device0-parent_idx_device0] PASSED [ 96%]
tests/compute/test_subgraph.py::test_subframes[child_device0-parent_idx_device1] PASSED [ 96%]
tests/compute/test_subgraph.py::test_subframes[child_device0-parent_idx_device2] SKIPPED [ 96%]
tests/compute/test_subgraph.py::test_subframes[child_device0-parent_idx_device3] FAILED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-u-shp2-g0] FAILED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-u-shp2-g1] FAILED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-u-shp3-g0] FAILED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-u-shp3-g1] FAILED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-u-shp4-g0] SKIPPED [ 96%]
tests/compute/test_subgraph.py::test_subframes[child_device1-parent_idx_device0] PASSED [ 96%]
tests/compute/test_subgraph.py::test_subframes[child_device1-parent_idx_device1] PASSED [ 96%]
tests/compute/test_subgraph.py::test_subframes[child_device1-parent_idx_device2] SKIPPED [ 96%]
tests/compute/test_subgraph.py::test_subframes[child_device1-parent_idx_device3] SKIPPED [ 96%]
tests/compute/test_subgraph.py::test_uva_subgraph[int32-device0] SKIPPED [ 96%]
tests/compute/test_subgraph.py::test_uva_subgraph[int32-device1] SKIPPED [ 96%]
tests/compute/test_subgraph.py::test_uva_subgraph[int64-device0] SKIPPED [ 96%]
tests/compute/test_subgraph.py::test_uva_subgraph[int64-device1] SKIPPED [ 96%]
tests/compute/test_transform.py::test_line_graph1 PASSED                 [ 96%]
tests/compute/test_transform.py::test_line_graph2[int32] PASSED          [ 96%]
tests/compute/test_transform.py::test_line_graph2[int64] PASSED          [ 96%]
tests/compute/test_transform.py::test_no_backtracking PASSED             [ 96%]
tests/compute/test_transform.py::test_reverse[int32] PASSED              [ 96%]
tests/compute/test_transform.py::test_reverse[int64] PASSED              [ 96%]
tests/compute/test_transform.py::test_reverse_shared_frames[int32] PASSED [ 96%]
tests/compute/test_transform.py::test_reverse_shared_frames[int64] PASSED [ 96%]
tests/compute/test_transform.py::test_to_bidirected SKIPPED (GPU not...) [ 96%]
tests/compute/test_transform.py::test_add_reverse_edges PASSED           [ 96%]
tests/compute/test_transform.py::test_simple_graph SKIPPED (GPU not ...) [ 96%]
tests/compute/test_transform.py::test_khop_graph SKIPPED (GPU not im...) [ 96%]
tests/compute/test_transform.py::test_khop_adj SKIPPED (GPU not impl...) [ 96%]
tests/compute/test_transform.py::test_laplacian_lambda_max SKIPPED (...) [ 97%]
tests/compute/test_transform.py::test_partition_with_halo SKIPPED (G...) [ 97%]
tests/compute/test_transform.py::test_metis_partition[int32] SKIPPED     [ 97%]
tests/compute/test_transform.py::test_metis_partition[int64] SKIPPED     [ 97%]
tests/compute/test_transform.py::test_reorder_nodes SKIPPED (It does...) [ 97%]
tests/compute/test_transform.py::test_compact[int32] PASSED              [ 97%]
tests/compute/test_transform.py::test_compact[int64] FAILED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-u-shp4-g1] FAILED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-v-shp0-g0] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-v-shp0-g1] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-v-shp1-g0] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-v-shp1-g1] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-v-shp2-g0] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-v-shp2-g1] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-v-shp3-g0] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-v-shp3-g1] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-v-shp4-g0] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-v-shp4-g1] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-e-shp0-g0] FAILED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-e-shp0-g1] FAILED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-e-shp1-g0] PASSED              [ 97%]
tests/compute/test_transform.py::test_to_simple[int32] SKIPPED (GPU ...) [ 97%]
tests/compute/test_transform.py::test_to_simple[int64] SKIPPED (GPU ...) [ 97%]
tests/compute/test_transform.py::test_to_block[int32] PASSED             [ 97%]
tests/compute/test_transform.py::test_to_block[int64] PASSED             [ 97%]
tests/compute/test_transform.py::test_remove_edges[int32] PASSED         [ 97%]
tests/compute/test_transform.py::test_remove_edges[int64] FAILED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-e-shp1-g1] FAILED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-e-shp2-g0] FAILED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-e-shp2-g1] PASSED         [ 97%]
tests/compute/test_transform.py::test_add_edges[int32] PASSED            [ 97%]
tests/compute/test_transform.py::test_add_edges[int64] PASSED            [ 97%]
tests/compute/test_transform.py::test_add_nodes[int32] PASSED            [ 97%]
tests/compute/test_transform.py::test_add_nodes[int64] PASSED            [ 97%]
tests/compute/test_transform.py::test_remove_nodes[int32] FAILED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-e-shp3-g0] FAILED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-e-shp3-g1] FAILED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-e-shp4-g0] FAILED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-e-shp4-g1] FAILED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-u-shp0-g0] FAILED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-u-shp0-g1] PASSED         [ 97%]
tests/compute/test_transform.py::test_remove_nodes[int64] PASSED         [ 97%]
tests/compute/test_transform.py::test_add_selfloop[int32] PASSED         [ 97%]
tests/compute/test_transform.py::test_add_selfloop[int64] FAILED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-u-shp1-g0] FAILED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-u-shp1-g1] FAILED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-u-shp2-g0] FAILED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-u-shp2-g1] PASSED         [ 97%]
tests/compute/test_transform.py::test_remove_selfloop[int32] PASSED      [ 97%]
tests/compute/test_transform.py::test_remove_selfloop[int64] PASSED      [ 97%]
tests/compute/test_transform.py::test_reorder_graph[int32] PASSED        [ 97%]
tests/compute/test_transform.py::test_reorder_graph[int64] PASSED        [ 98%]
tests/compute/test_transform.py::test_norm_by_dst[int32] PASSED          [ 98%]
tests/compute/test_transform.py::test_norm_by_dst[int64] PASSED          [ 98%]
tests/compute/test_transform.py::test_module_add_self_loop[int32] PASSED [ 98%]
tests/compute/test_transform.py::test_module_add_self_loop[int64] PASSED [ 98%]
tests/compute/test_transform.py::test_module_remove_self_loop[int32] PASSED [ 98%]
tests/compute/test_transform.py::test_module_remove_self_loop[int64] PASSED [ 98%]
tests/compute/test_transform.py::test_module_add_reverse[int32] FAILED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-u-shp3-g0] FAILED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-u-shp3-g1] FAILED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-u-shp4-g0] FAILED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-u-shp4-g1] FAILED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-v-shp0-g0] PASSED   [ 98%]
tests/compute/test_transform.py::test_module_add_reverse[int64] PASSED   [ 98%]
tests/compute/test_transform.py::test_module_to_simple[int32] SKIPPED    [ 98%]
tests/compute/test_transform.py::test_module_to_simple[int64] SKIPPED    [ 98%]
tests/compute/test_transform.py::test_module_line_graph[int32] PASSED    [ 98%]
tests/compute/test_transform.py::test_module_line_graph[int64] PASSED    [ 98%]
tests/compute/test_transform.py::test_module_khop_graph[int32] PASSED    [ 98%]
tests/compute/test_transform.py::test_module_khop_graph[int64] PASSED    [ 98%]
tests/compute/test_transform.py::test_module_add_metapaths[int32] PASSED [ 98%]
tests/compute/test_transform.py::test_module_add_metapaths[int64] PASSED [ 98%]
tests/compute/test_transform.py::test_module_compose[int32] PASSED       [ 98%]
tests/compute/test_transform.py::test_module_compose[int64] PASSED       [ 98%]
tests/compute/test_transform.py::test_module_gcnnorm[int32] PASSED       [ 98%]
tests/compute/test_transform.py::test_module_gcnnorm[int64] PASSED       [ 98%]
tests/compute/test_transform.py::test_module_ppr[int32] SKIPPED (Onl...) [ 98%]
tests/compute/test_transform.py::test_module_ppr[int64] SKIPPED (Onl...) [ 98%]
tests/compute/test_transform.py::test_module_heat_kernel[int32] SKIPPED  [ 98%]
tests/compute/test_transform.py::test_module_heat_kernel[int64] SKIPPED  [ 98%]
tests/compute/test_transform.py::test_module_gdc[int32] SKIPPED (Onl...) [ 99%]
tests/compute/test_transform.py::test_module_gdc[int64] SKIPPED (Onl...) [ 99%]
tests/compute/test_transform.py::test_module_node_shuffle[int32] PASSED  [ 99%]
tests/compute/test_transform.py::test_module_node_shuffle[int64] PASSED  [ 99%]
tests/compute/test_transform.py::test_module_drop_node[int32] SKIPPED    [ 99%]
tests/compute/test_transform.py::test_module_drop_node[int64] SKIPPED    [ 99%]
tests/compute/test_transform.py::test_module_drop_edge[int32] SKIPPED    [ 99%]
tests/compute/test_transform.py::test_module_drop_edge[int64] SKIPPED    [ 99%]
tests/compute/test_transform.py::test_module_add_edge[int32] PASSED      [ 99%]
tests/compute/test_transform.py::test_module_add_edge[int64] PASSED      [ 99%]
tests/compute/test_transform.py::test_module_random_walk_pe[int32] PASSED [ 99%]
tests/compute/test_transform.py::test_module_random_walk_pe[int64] PASSED [ 99%]
tests/compute/test_transform.py::test_module_laplacian_pe[int32] PASSED  [ 99%]
tests/compute/test_transform.py::test_module_laplacian_pe[int64] PASSED  [ 99%]
tests/compute/test_transform.py::test_module_sign[g0] SKIPPED (Only ...) [ 99%]
tests/compute/test_transform.py::test_module_row_feat_normalizer[int32] SKIPPED [ 99%]
tests/compute/test_transform.py::test_module_row_feat_normalizer[int64] SKIPPED [ 99%]
tests/compute/test_transform.py::test_module_feat_mask[int32] SKIPPED    [ 99%]
tests/compute/test_transform.py::test_module_feat_mask[int64] SKIPPED    [ 99%]
tests/compute/test_traversal.py::test_bfs[int32] FAILED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-v-shp0-g1] FAILED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-v-shp1-g0] FAILED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-v-shp1-g1] FAILED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-v-shp2-g0] FAILED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-v-shp2-g1] FAILED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-v-shp3-g0] FAILED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-v-shp3-g1] FAILED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-v-shp4-g0] FAILED [ 43%]PASSED                  [ 99%]
tests/compute/test_traversal.py::test_bfs[int64] 
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-v-shp4-g1] FAILED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-e-shp0-g0] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-e-shp0-g1] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-e-shp1-g0] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-e-shp1-g1] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-e-shp2-g0] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-e-shp2-g1] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-e-shp3-g0] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-e-shp3-g1] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-e-shp4-g0] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-e-shp4-g1] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-u-shp0-g0] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-u-shp0-g1] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-u-shp1-g0] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-u-shp1-g1] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-u-shp2-g0] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-u-shp2-g1] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-u-shp3-g0] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-u-shp3-g1] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-u-shp4-g0] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-u-shp4-g1] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-v-shp0-g0] FAILED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-v-shp0-g1] FAILED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-v-shp1-g0] PASSED                  [ 99%]
tests/compute/test_traversal.py::test_topological_nodes[int32] PASSED    [ 99%]
tests/compute/test_traversal.py::test_topological_nodes[int64] PASSED    [ 99%]
tests/compute/test_traversal.py::test_dfs_labeled_edges[int32] PASSED    [ 99%]
tests/compute/test_traversal.py::test_dfs_labeled_edges[int64] PASSED    [100%]

=============================== warnings summary ===============================
python/dgl/backend/backend.py:1741
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/backend/backend.py:1741: DeprecationWarning: invalid escape sequence \P
    """

../../../../opt/conda/envs/mxnet-ci/lib/python3.7/site-packages/mxnet/numpy/utils.py:37
  /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages/mxnet/numpy/utils.py:37: DeprecationWarning: `np.bool` is a deprecated alias for the builtin `bool`. To silence this warning, use `bool` by itself. Doing this will not modify any behavior and is safe. If you specifically wanted the numpy scalar type, use `np.bool_` here.
  Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations
    bool = onp.bool

../../../../opt/conda/envs/mxnet-ci/lib/python3.7/site-packages/mxnet/numpy/fallback.py:143
  /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages/mxnet/numpy/fallback.py:143: DeprecationWarning: In accordance with NEP 32, the function mirr was removed from NumPy version 1.20.  A replacement for this function is available in the numpy_financial library: https://pypi.org/project/numpy-financial
    mirr = onp.mirr

../../../../opt/conda/envs/mxnet-ci/lib/python3.7/site-packages/mxnet/numpy/fallback.py:160
  /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages/mxnet/numpy/fallback.py:160: DeprecationWarning: In accordance with NEP 32, the function npv was removed from NumPy version 1.20.  A replacement for this function is available in the numpy_financial library: https://pypi.org/project/numpy-financial
    npv = onp.npv

../../../../opt/conda/envs/mxnet-ci/lib/python3.7/site-packages/mxnet/numpy/fallback.py:164
  /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages/mxnet/numpy/fallback.py:164: DeprecationWarning: In accordance with NEP 32, the function pmt was removed from NumPy version 1.20.  A replacement for this function is available in the numpy_financial library: https://pypi.org/project/numpy-financial
    pmt = onp.pmt

../../../../opt/conda/envs/mxnet-ci/lib/python3.7/site-packages/mxnet/numpy/fallback.py:173
  /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages/mxnet/numpy/fallback.py:173: DeprecationWarning: In accordance with NEP 32, the function ppmt was removed from NumPy version 1.20.  A replacement for this function is available in the numpy_financial library: https://pypi.org/project/numpy-financial
    ppmt = onp.ppmt

../../../../opt/conda/envs/mxnet-ci/lib/python3.7/site-packages/mxnet/numpy/fallback.py:176
  /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages/mxnet/numpy/fallback.py:176: DeprecationWarning: In accordance with NEP 32, the function pv was removed from NumPy version 1.20.  A replacement for this function is available in the numpy_financial library: https://pypi.org/project/numpy-financial
    pv = onp.pv

../../../../opt/conda/envs/mxnet-ci/lib/python3.7/site-packages/mxnet/numpy/fallback.py:177
  /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages/mxnet/numpy/fallback.py:177: DeprecationWarning: In accordance with NEP 32, the function rate was removed from NumPy version 1.20.  A replacement for this function is available in the numpy_financial library: https://pypi.org/project/numpy-financial
    rate = onp.rate

../../../../opt/conda/envs/mxnet-ci/lib/python3.7/site-packages/mxnet/numpy_dispatch_protocol.py:48
  /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages/mxnet/numpy_dispatch_protocol.py:48: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    cur_np_ver = LooseVersion(_np.__version__)

../../../../opt/conda/envs/mxnet-ci/lib/python3.7/site-packages/mxnet/numpy_dispatch_protocol.py:49
  /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages/mxnet/numpy_dispatch_protocol.py:49: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    np_1_17_ver = LooseVersion('1.17')

../../../../opt/conda/envs/mxnet-ci/lib/python3.7/site-packages/mxnet/numpy_dispatch_protocol.py:68
  /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages/mxnet/numpy_dispatch_protocol.py:68: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    cur_np_ver = LooseVersion(_np.__version__)

../../../../opt/conda/envs/mxnet-ci/lib/python3.7/site-packages/mxnet/numpy_dispatch_protocol.py:69
  /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages/mxnet/numpy_dispatch_protocol.py:69: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    np_1_15_ver = LooseVersion('1.15')

python/dgl/backend/mxnet/tensor.py:15
python/dgl/backend/mxnet/tensor.py:15
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/backend/mxnet/tensor.py:15: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    if LooseVersion(mx.__version__) < LooseVersion("1.6.0"):

python/dgl/backend/mxnet/tensor.py:31
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/backend/mxnet/tensor.py:31: DeprecationWarning: `np.bool` is a deprecated alias for the builtin `bool`. To silence this warning, use `bool` by itself. Doing this will not modify any behavior and is safe. If you specifically wanted the numpy scalar type, use `np.bool_` here.
  Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations
    'bool'    : np.bool}  # mxnet does not support bool

python/dgl/backend/mxnet/tensor.py:37: 202 warnings
tests/compute/test_basics.py: 229 warnings
tests/compute/test_batched_graph.py: 432 warnings
tests/compute/test_batched_heterograph.py: 310 warnings
tests/compute/test_csrmm.py: 332 warnings
tests/compute/test_dataloader.py: 50 warnings
tests/compute/test_filter.py: 22 warnings
tests/compute/test_frame.py: 8 warnings
tests/compute/test_graph.py: 1626 warnings
tests/compute/test_heterograph.py: 4432 warnings
tests/compute/test_index.py: 1 warning
tests/compute/test_kernel.py: 2511 warnings
tests/compute/test_merge.py: 12 warnings
tests/compute/test_nccl.py: 2 warnings
tests/compute/test_new_update_all_hetero.py: 56 warnings
tests/compute/test_partition.py: 22 warnings
tests/compute/test_pickle.py: 16 warnings
tests/compute/test_readout.py: 844 warnings
tests/compute/test_removal.py: 838 warnings
tests/compute/test_sampler.py: 167 warnings
tests/compute/test_sampling.py: 2737 warnings
tests/compute/test_serialize.py: 2 warnings
tests/compute/test_shared_mem.py: 40 warnings
tests/compute/test_sparse.py: 2564 warnings
tests/compute/test_specialization.py: 598 warnings
tests/compute/test_subgraph.py: 666 warnings
tests/compute/test_transform.py: 1818 warnings
tests/compute/test_traversal.py: 792 warnings
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/backend/mxnet/tensor.py:37: DeprecationWarning: `np.bool` is a deprecated alias for the builtin `bool`. To silence this warning, use `bool` by itself. Doing this will not modify any behavior and is safe. If you specifically wanted the numpy scalar type, use `np.bool_` here.
  Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations
    if dtype == np.bool:

python/dgl/backend/mxnet/tensor.py:136: 174 warnings
tests/compute/test_basics.py: 225 warnings
tests/compute/test_batched_graph.py: 396 warnings
tests/compute/test_batched_heterograph.py: 204 warnings
tests/compute/test_dataloader.py: 16 warnings
tests/compute/test_filter.py: 8 warnings
tests/compute/test_frame.py: 1 warning
tests/compute/test_graph.py: 2377 warnings
tests/compute/test_heterograph.py: 5224 warnings
tests/compute/test_index.py: 1 warning
tests/compute/test_kernel.py: 2050 warnings
tests/compute/test_merge.py: 16 warnings
tests/compute/test_new_update_all_hetero.py: 56 warnings
tests/compute/test_pickle.py: 24 warnings
tests/compute/test_readout.py: 1416 warnings
tests/compute/test_removal.py: 662 warnings
tests/compute/test_sampling.py: 5106 warnings
tests/compute/test_shared_mem.py: 40 warnings
tests/compute/test_sparse.py: 2 warnings
tests/compute/test_specialization.py: 588 warnings
tests/compute/test_subgraph.py: 354 warnings
tests/compute/test_transform.py: 902 warnings
tests/compute/test_traversal.py: 1563 warnings
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/backend/mxnet/tensor.py:136: DeprecationWarning: `np.bool` is a deprecated alias for the builtin `bool`. To silence this warning, use `bool` by itself. Doing this will not modify any behavior and is safe. If you specifically wanted the numpy scalar type, use `np.bool_` here.
  Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations
    if ty == np.bool:

../../../../opt/conda/envs/mxnet-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:10
  /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:10: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    _nlv = LooseVersion(_np_version)

../../../../opt/conda/envs/mxnet-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:11
  /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:11: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    _np_version_under1p16 = _nlv < LooseVersion("1.16")

../../../../opt/conda/envs/mxnet-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:12
  /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:12: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    _np_version_under1p17 = _nlv < LooseVersion("1.17")

../../../../opt/conda/envs/mxnet-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:13
  /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:13: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    _np_version_under1p18 = _nlv < LooseVersion("1.18")

../../../../opt/conda/envs/mxnet-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:14
  /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:14: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    _np_version_under1p19 = _nlv < LooseVersion("1.19")

../../../../opt/conda/envs/mxnet-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:15
  /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:15: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    _np_version_under1p20 = _nlv < LooseVersion("1.20")

../../../../opt/conda/envs/mxnet-ci/lib/python3.7/site-packages/setuptools/_distutils/version.py:351
  /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages/setuptools/_distutils/version.py:351: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    other = LooseVersion(other)

../../../../opt/conda/envs/mxnet-ci/lib/python3.7/site-packages/pandas/compat/numpy/function.py:125
../../../../opt/conda/envs/mxnet-ci/lib/python3.7/site-packages/pandas/compat/numpy/function.py:125
  /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages/pandas/compat/numpy/function.py:125: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    if LooseVersion(_np_version) >= LooseVersion("1.17.0"):

tests/compute/test_basics.py: 2 warnings
tests/compute/test_filter.py: 1 warning
tests/compute/test_graph.py: 9 warnings
tests/compute/test_kernel.py: 3 warnings
tests/compute/test_removal.py: 16 warnings
tests/compute/test_specialization.py: 12 warnings
tests/compute/test_subgraph.py: 2 warnings
tests/compute/test_transform.py: 6 warnings
tests/compute/test_traversal.py: 2 warnings
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/heterograph.py:72: DGLWarning: Recommend creating graphs by `dgl.graph(data)` instead of `dgl.DGLGraph(data)`.
    dgl_warning('Recommend creating graphs by `dgl.graph(data)`'

tests/compute/test_basics.py: 2 warnings
tests/compute/test_batched_graph.py: 10 warnings
tests/compute/test_graph.py: 2 warnings
tests/compute/test_kernel.py: 1 warning
tests/compute/test_removal.py: 10 warnings
tests/compute/test_specialization.py: 10 warnings
tests/compute/test_subgraph.py: 2 warnings
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/heterograph.py:354: DGLWarning: DGLGraph.add_edge is deprecated. Please use DGLGraph.add_edges
    dgl_warning("DGLGraph.add_edge is deprecated. Please use DGLGraph.add_edges")

tests/compute/test_basics.py::test_update_all_0deg[int32]
tests/compute/test_basics.py::test_update_all_0deg[int64]
tests/compute/test_basics.py::test_pull_0deg[int32]
tests/compute/test_basics.py::test_pull_0deg[int64]
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/core.py:79: DGLWarning: The input graph for the user-defined edge function does not contain valid edges
    dgl_warning('The input graph for the user-defined edge function ' \

tests/compute/test_batched_graph.py::test_batched_edge_ordering[int32]
tests/compute/test_batched_graph.py::test_batched_edge_ordering[int64]
tests/compute/test_graph.py::test_query
tests/compute/test_graph.py::test_hypersparse_query
tests/compute/test_transform.py::test_no_backtracking
tests/compute/test_transform.py::test_reverse[int32]
tests/compute/test_transform.py::test_reverse[int64]
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/heterograph.py:2978: DGLWarning: DGLGraph.edge_id is deprecated. Please use DGLGraph.edge_ids.
    dgl_warning("DGLGraph.edge_id is deprecated. Please use DGLGraph.edge_ids.")

tests/compute/test_batched_heterograph.py::test_features[int32]
tests/compute/test_batched_heterograph.py::test_features[int64]
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/batch.py:159: DGLWarning: Arguments edge_attrs has been deprecated. Please use edata instead.
    dgl_warning('Arguments edge_attrs has been deprecated. Please use'

tests/compute/test_csrmm.py::test_csrmm[float32-int32]
tests/compute/test_csrmm.py::test_csrmm[float32-int64]
tests/compute/test_csrmm.py::test_csrmm[float64-int32]
tests/compute/test_csrmm.py::test_csrmm[float64-int64]
tests/compute/test_csrmm.py::test_csrsum[float32-int32]
tests/compute/test_csrmm.py::test_csrsum[float32-int64]
tests/compute/test_csrmm.py::test_csrsum[float64-int32]
tests/compute/test_csrmm.py::test_csrsum[float64-int64]
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/heterograph_index.py:797: FutureWarning: Adjacency matrix by default currently returns edge IDs.  As a result there is one 0 entry which is not eliminated.  In the next release it will return 1s by default, and 0 will be eliminated otherwise.
    FutureWarning)

tests/compute/test_graph.py::test_query
tests/compute/test_graph.py::test_hypersparse_query
tests/compute/test_heterograph.py::test_query[int32]
tests/compute/test_heterograph.py::test_query[int64]
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/heterograph.py:2753: DGLWarning: DGLGraph.has_node is deprecated. Please use DGLGraph.has_nodes
    dgl_warning("DGLGraph.has_node is deprecated. Please use DGLGraph.has_nodes")

tests/compute/test_graph.py::test_query
tests/compute/test_graph.py::test_hypersparse_query
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/heterograph.py:2687: DGLWarning: DGLGraph.__contains__ is deprecated. Please directly call has_nodes.
    dgl_warning('DGLGraph.__contains__ is deprecated.'

tests/compute/test_graph.py::test_query
tests/compute/test_sampling.py::test_non_uniform_random_walk[False]
tests/compute/test_sampling.py::test_uniform_random_walk[True]
tests/compute/test_sampling.py::test_uniform_random_walk[False]
tests/compute/test_transform.py::test_no_backtracking
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/heterograph.py:2851: DGLWarning: DGLGraph.has_edge_between is deprecated. Please use DGLGraph.has_edges_between
    dgl_warning("DGLGraph.has_edge_between is deprecated. "

tests/compute/test_graph.py::test_query
tests/compute/test_graph.py::test_hypersparse_query
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/heterograph.py:3432: DGLWarning: DGLGraph.in_degree is deprecated. Please use DGLGraph.in_degrees
    dgl_warning("DGLGraph.in_degree is deprecated. Please use DGLGraph.in_degrees")

tests/compute/test_graph.py::test_query
tests/compute/test_graph.py::test_hypersparse_query
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/heterograph.py:3516: DGLWarning: DGLGraph.out_degree is deprecated. Please use DGLGraph.out_degrees
    dgl_warning("DGLGraph.out_degree is deprecated. Please use DGLGraph.out_degrees")

tests/compute/test_graph.py::test_query
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/heterograph.py:84: DGLWarning: Keyword arguments ['readonly', 'sort_csr'] are deprecated in v0.5, and can be safely removed in all cases.
    ' removed in all cases.'.format(list(deprecate_kwargs.keys())))

tests/compute/test_heterograph.py: 352 warnings
tests/compute/test_kernel.py: 1766 warnings
tests/compute/test_new_update_all_hetero.py: 56 warnings
tests/compute/test_readout.py: 144 warnings
tests/compute/test_removal.py: 8 warnings
tests/compute/test_sampler.py: 22 warnings
tests/compute/test_sampling.py: 12 warnings
tests/compute/test_sparse.py: 2560 warnings
tests/compute/test_subgraph.py: 10 warnings
tests/compute/test_transform.py: 52 warnings
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/backend/mxnet/tensor.py:50: DeprecationWarning: `np.bool` is a deprecated alias for the builtin `bool`. To silence this warning, use `bool` by itself. Doing this will not modify any behavior and is safe. If you specifically wanted the numpy scalar type, use `np.bool_` here.
  Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations
    dtype = np.int32 if data.dtype == np.bool else data.dtype

tests/compute/test_heterograph.py: 20 warnings
  /root/jenkins/workspace/dgl_PR-4648/tests/compute/test_heterograph.py:1128: DeprecationWarning: np.asscalar(a) is deprecated since NumPy v1.16, use a.item() instead
    assert np.asscalar(F.asnumpy(src_i)) == nid[src[i]]

tests/compute/test_heterograph.py: 20 warnings
  /root/jenkins/workspace/dgl_PR-4648/tests/compute/test_heterograph.py:1129: DeprecationWarning: np.asscalar(a) is deprecated since NumPy v1.16, use a.item() instead
    assert np.asscalar(F.asnumpy(dst_i)) == nid[dst[i]]

tests/compute/test_heterograph.py::test_invertible_conversion[int32]
tests/compute/test_heterograph.py::test_invertible_conversion[int64]
tests/compute/test_shared_mem.py::test_single_process[int32]
tests/compute/test_shared_mem.py::test_single_process[int64]
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/heterograph.py:2635: DGLWarning: DGLGraph.is_readonly is deprecated in v0.5.
  DGLGraph now always supports mutable operations like add_nodes and add_edges.
    dgl_warning('DGLGraph.is_readonly is deprecated in v0.5.\n'

tests/compute/test_kernel.py: 3 warnings
tests/compute/test_sparse.py: 290 warnings
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/backend/mxnet/sparse.py:17: DGLWarning: MXNet do not support scatter_add, fallback to numpy.
    dgl_warning("MXNet do not support scatter_add, fallback to numpy.")

tests/compute/test_partition.py::test_get_node_partition_from_book[int32]
tests/compute/test_partition.py::test_get_node_partition_from_book[int64]
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
    "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")

tests/compute/test_sampler.py::test_create_full
tests/compute/test_sampler.py::test_1neighbor_sampler_all
tests/compute/test_sampler.py::test_1neighbor_sampler
tests/compute/test_sampler.py::test_prefetch_neighbor_sampler
tests/compute/test_sampler.py::test_10neighbor_sampler_all
tests/compute/test_sampler.py::test_10neighbor_sampler
tests/compute/test_sampler.py::test_nonuniform_neighbor_sampler
tests/compute/test_sampler.py::test_setseed
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/contrib/sampling/sampler.py:317: DGLWarning: dgl.contrib.sampling.NeighborSampler is deprecated starting from v0.5. Please read our guide<link> for how to use the new sampling APIs.
    dgl_warning('dgl.contrib.sampling.NeighborSampler is deprecated starting from v0.5.'

tests/compute/test_sampler.py::test_create_full
tests/compute/test_sampler.py::test_1neighbor_sampler_all
tests/compute/test_sampler.py::test_1neighbor_sampler
tests/compute/test_sampler.py::test_prefetch_neighbor_sampler
tests/compute/test_sampler.py::test_10neighbor_sampler_all
tests/compute/test_sampler.py::test_10neighbor_sampler
tests/compute/test_sampler.py::test_layer_sampler
tests/compute/test_sampler.py::test_nonuniform_neighbor_sampler
tests/compute/test_sampler.py::test_setseed
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/_deprecate/nodeflow.py:99: DGLWarning: NodeFlow APIs are deprecated starting from v0.5. Please read our guide<link> for how to use the new sampling APIs.
    dgl_warning('NodeFlow APIs are deprecated starting from v0.5. Please read our'

tests/compute/test_sampling.py::test_global_uniform_negative_sampling[int32]
tests/compute/test_sampling.py::test_global_uniform_negative_sampling[int64]
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/sampling/negative.py:102: ComplexWarning: Casting complex values to real discards the imaginary part
    g._graph, etype_id, num_samples, 3, exclude_self_loops, replace, redundancy)

tests/compute/test_serialize.py::test_load_old_files1
tests/compute/test_serialize.py::test_load_old_files2
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/data/graph_serialize.py:179: DGLWarning: You are loading a graph file saved by old version of dgl.              Please consider saving it again with the current format.
    Please consider saving it again with the current format.")

tests/compute/test_transform.py::test_reverse_shared_frames[int32]
tests/compute/test_transform.py::test_reverse_shared_frames[int64]
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/transforms/functional.py:1267: DGLWarning: share_ndata argument has been renamed to copy_ndata.
    dgl_warning('share_ndata argument has been renamed to copy_ndata.')

tests/compute/test_transform.py::test_reverse_shared_frames[int32]
tests/compute/test_transform.py::test_reverse_shared_frames[int64]
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/transforms/functional.py:1270: DGLWarning: share_edata argument has been renamed to copy_edata.
    dgl_warning('share_edata argument has been renamed to copy_edata.')

-- Docs: https://docs.pytest.org/en/stable/how-to/capture-warnings.html
-- generated xml file: /root/jenkins/workspace/dgl_PR-4648/pytest_compute.xml --
============================ slowest 100 durations =============================
19.59s call     tests/compute/test_sampler.py::test_negative_sampler
8.74s call     tests/compute/test_kernel.py::test_all_binary_builtins
1.20s call     tests/compute/test_graph.py::test_query
1.02s call     tests/compute/test_sampling.py::test_non_uniform_random_walk[False]
0.75s call     tests/compute/test_heterograph.py::test_query[int64]
0.69s call     tests/compute/test_heterograph.py::test_query[int32]
0.65s call     tests/compute/test_sampling.py::test_uniform_random_walk[False]
0.50s call     tests/compute/test_sampling.py::test_uniform_random_walk[True]
0.48s call     tests/compute/test_sampling.py::test_sample_neighbors_outedge
0.34s call     tests/compute/test_traversal.py::test_bfs[int32]
0.34s call     tests/compute/test_basics.py::test_compatible
0.32s call     tests/compute/test_basics.py::test_batch_setter_getter[int32]
0.31s call     tests/compute/test_traversal.py::test_bfs[int64]
0.29s call     tests/compute/test_heterograph.py::test_view1[int32]
0.27s call     tests/compute/test_heterograph.py::test_view1[int64]
0.26s call     tests/compute/test_sampling.py::test_sample_neighbors_prob
0.26s call     tests/compute/test_sampling.py::test_sample_neighbors_noprob
0.23s call     tests/compute/test_kernel.py::test_copy_edge_reduce
0.22s call     tests/compute/test_kernel.py::test_copy_src_reduce
0.17s call     tests/compute/test_heterograph.py::test_updates[int32]
0.17s call     tests/compute/test_heterograph.py::test_updates[int64]
0.17s call     tests/compute/test_removal.py::test_node_and_edge_removal[int32]
0.16s call     tests/compute/test_removal.py::test_node_and_edge_removal[int64]
0.15s call     tests/compute/test_sparse.py::test_sddmm[int64-add-u-v-shp0-g0]
0.15s call     tests/compute/test_sampler.py::test_prefetch_neighbor_sampler
0.13s call     tests/compute/test_sampler.py::test_1neighbor_sampler
0.12s call     tests/compute/test_sampler.py::test_10neighbor_sampler
0.12s call     tests/compute/test_transform.py::test_remove_nodes[int32]
0.12s call     tests/compute/test_transform.py::test_remove_nodes[int64]
0.12s call     tests/compute/test_heterograph.py::test_format[int64]
0.12s call     tests/compute/test_heterograph.py::test_format[int32]
0.11s call     tests/compute/test_heterograph.py::test_level2[int32]
0.11s call     tests/compute/test_heterograph.py::test_level2[int64]
0.11s call     tests/compute/test_transform.py::test_remove_edges[int32]
0.10s call     tests/compute/test_specialization.py::test_v2v_snr[int32]
0.10s call     tests/compute/test_sampler.py::test_1neighbor_sampler_all
0.10s call     tests/compute/test_specialization.py::test_v2v_snr[int64]
0.10s call     tests/compute/test_transform.py::test_remove_edges[int64]
0.10s call     tests/compute/test_shared_mem.py::test_multi_process[int32]
0.09s call     tests/compute/test_specialization.py::test_v2v_pull[int32]
0.09s call     tests/compute/test_shared_mem.py::test_copy_from_gpu
0.09s call     tests/compute/test_specialization.py::test_v2v_pull[int64]
0.09s call     tests/compute/test_shared_mem.py::test_multi_process[int64]
0.08s call     tests/compute/test_specialization.py::test_pull_multi_fallback[int32]
0.08s call     tests/compute/test_specialization.py::test_pull_multi_fallback[int64]
0.07s call     tests/compute/test_specialization.py::test_v2v_update_all[int32]
0.07s call     tests/compute/test_heterograph.py::test_convert[int64]
0.07s call     tests/compute/test_specialization.py::test_v2v_update_all[int64]
0.07s call     tests/compute/test_heterograph.py::test_flatten[int32]
0.07s call     tests/compute/test_heterograph.py::test_convert[int32]
0.07s call     tests/compute/test_heterograph.py::test_flatten[int64]
0.06s call     tests/compute/test_subgraph.py::test_subgraph1[int32]
0.06s call     tests/compute/test_subgraph.py::test_subgraph1[int64]
0.06s call     tests/compute/test_heterograph.py::test_float_cast
0.06s call     tests/compute/test_transform.py::test_add_edges[int32]
0.06s call     tests/compute/test_transform.py::test_add_edges[int64]
0.05s call     tests/compute/test_sampler.py::test_layer_sampler
0.05s call     tests/compute/test_heterograph.py::test_add_edges[int32]
0.05s call     tests/compute/test_heterograph.py::test_add_edges[int64]
0.05s call     tests/compute/test_specialization.py::test_spmv_3d_feat[int64]
0.05s call     tests/compute/test_heterograph.py::test_remove_nodes[int32]
0.05s call     tests/compute/test_removal.py::test_edge_removal[int32]
0.05s call     tests/compute/test_nccl.py::test_nccl_sparse_push_single_remainder
0.05s call     tests/compute/test_heterograph.py::test_remove_nodes[int64]
0.05s call     tests/compute/test_transform.py::test_add_selfloop[int32]
0.05s call     tests/compute/test_transform.py::test_add_selfloop[int64]
0.05s call     tests/compute/test_transform.py::test_to_block[int32]
0.05s call     tests/compute/test_transform.py::test_to_block[int64]
0.04s call     tests/compute/test_removal.py::test_edge_removal[int64]
0.04s call     tests/compute/test_specialization.py::test_spmv_3d_feat[int32]
0.04s call     tests/compute/test_heterograph.py::test_types_in_function[int32]
0.04s call     tests/compute/test_heterograph.py::test_types_in_function[int64]
0.04s call     tests/compute/test_nccl.py::test_nccl_sparse_pull_single_remainder
0.04s call     tests/compute/test_subgraph.py::test_khop_out_subgraph[int64]
0.04s call     tests/compute/test_subgraph.py::test_khop_out_subgraph[int32]
0.04s call     tests/compute/test_transform.py::test_add_reverse_edges
0.04s call     tests/compute/test_heterograph.py::test_create[int32]
0.04s call     tests/compute/test_specialization.py::test_update_all_multi_fallback[int32]
0.04s call     tests/compute/test_specialization.py::test_update_all_multi_fallback[int64]
0.04s call     tests/compute/test_batched_heterograph.py::test_slice_batch[int32]
0.04s call     tests/compute/test_batched_heterograph.py::test_slice_batch[int64]
0.04s call     tests/compute/test_heterograph.py::test_create[int64]
0.04s call     tests/compute/test_nccl.py::test_nccl_sparse_pull_single_range
0.04s call     tests/compute/test_transform.py::test_no_backtracking
0.04s call     tests/compute/test_sparse.py::test_spmm[int32-sum-add-shp0-g0]
0.03s call     tests/compute/test_readout.py::test_weighted_reduce_readout[max-g1-int32]
0.03s call     tests/compute/test_batched_graph.py::test_batch_propagate[int32]
0.03s call     tests/compute/test_readout.py::test_weighted_reduce_readout[max-g1-int64]
0.03s call     tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g1-int64]
0.03s call     tests/compute/test_batched_graph.py::test_batch_propagate[int64]
0.03s call     tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g1-int32]
0.03s call     tests/compute/test_transform.py::test_reorder_graph[int32]
0.03s call     tests/compute/test_readout.py::test_reduce_readout[max-g1-int32]
0.03s call     tests/compute/test_readout.py::test_reduce_readout[max-g1-int64]
0.03s call     tests/compute/test_readout.py::test_reduce_readout[mean-g1-int64]
0.03s call     tests/compute/test_subgraph.py::test_khop_in_subgraph[int32]
0.03s call     tests/compute/test_readout.py::test_reduce_readout[mean-g1-int32]
0.03s call     tests/compute/test_transform.py::test_reorder_graph[int64]
0.03s call     tests/compute/test_nccl.py::test_nccl_sparse_push_single_range
0.03s call     tests/compute/test_sampler.py::test_nonuniform_neighbor_sampler
========= 2229 passed, 224 skipped, 48231 warnings in 72.02s (0:01:12) =========
FAILED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-v-shp1-g1] FAILED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-v-shp2-g0] FAILED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-v-shp2-g1] FAILED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-v-shp3-g0] Client [1090] waits on 172.17.0.3:49037
Machine (0) group (0) client (0) connect to server successfuly!
Client[0] in group[0] is exiting...
Server (0) shutdown.
Server is exiting...
Server (1) shutdown.
Server is exiting...
Done sampling
Converting to homogeneous graph takes 0.001s, peak mem: 1.529 GB
Convert a graph into a bidirected graph: 0.000 seconds, peak memory: 1.529 GB
Construct multi-constraint weights: 0.000 seconds, peak memory: 1.529 GB
[05:32:46] /root/jenkins/workspace/dgl_PR-4648/src/graph/transform/metis_partition_hetero.cc:87: Partition a graph with 1500 nodes and 2000 edges into 2 parts and get 24 edge cuts
Metis partitioning: 0.001 seconds, peak memory: 1.529 GB
Assigning nodes to METIS partitions takes 0.002s, peak mem: 1.529 GB
Reshuffle nodes and edges: 0.001 seconds
Split the graph: 0.001 seconds
Construct subgraphs: 0.002 seconds
Splitting the graph into partitions takes 0.003s, peak mem: 1.529 GB
part 0 has 519 nodes of type game and 511 are inside the partition
part 0 has 266 nodes of type user and 252 are inside the partition
part 0 has 497 edges of type buys and 489 are inside the partition
part 1 has 505 nodes of type game and 489 are inside the partition
part 1 has 256 nodes of type user and 248 are inside the partition
part 1 has 527 edges of type buys and 511 are inside the partition
Save partitions: 0.002 seconds, peak memory: 1.529 GB
There are 1000 edges in the graph and 0 edge cuts for 2 partitions.
FAILED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-v-shp3-g1] FAILED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-v-shp4-g0] FAILED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-v-shp4-g1] FAILED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-e-shp0-g0] Error in sys.excepthook:

Original exception was:
Error in sys.excepthook:

Original exception was:
Error in sys.excepthook:

Original exception was:
Error in sys.excepthook:

Original exception was:
Error in sys.excepthook:

Original exception was:
Error in sys.excepthook:

Original exception was:
Error in sys.excepthook:

Original exception was:
Error in sys.excepthook:

Original exception was:
Error in sys.excepthook:

Original exception was:
Error in sys.excepthook:

Original exception was:
Error in sys.excepthook:

Original exception was:
Error in sys.excepthook:

Original exception was:
Error in sys.excepthook:

Original exception was:
Error in sys.excepthook:

Original exception was:
Error in sys.excepthook:

Original exception was:
Error in sys.excepthook:

Original exception was:
FAILED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-e-shp0-g1] FAILED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-e-shp1-g0] FAILED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-e-shp1-g1] FAILED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-e-shp2-g0] FAILED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-e-shp2-g1] FAILED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-e-shp3-g0] FAILED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-e-shp3-g1] FAILED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-e-shp4-g0] FAILED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-e-shp4-g1] ============================= test session starts ==============================
platform linux -- Python 3.7.0, pytest-7.1.2, pluggy-1.0.0 -- /opt/conda/envs/mxnet-ci/bin/python3
cachedir: .pytest_cache
rootdir: /root/jenkins/workspace/dgl_PR-4648
collecting ... FAILED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-u-shp0-g0] FAILED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-u-shp0-g1] FAILED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-u-shp1-g0] FAILED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-u-shp1-g1] /root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_sampling
start graph service on server 0 for part 0
[05:32:47] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:32:47] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:18024]...
FAILED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-u-shp2-g0] FAILED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-u-shp2-g1] FAILED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-u-shp3-g0] FAILED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-u-shp3-g1] FAILED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-u-shp4-g0] FAILED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-u-shp4-g1] FAILED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-v-shp0-g0] PASSED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-v-shp0-g1] PASSED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-v-shp1-g0] PASSED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-v-shp1-g1] PASSED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-v-shp2-g0] PASSED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-v-shp2-g1] PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-v-shp3-g0] PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-v-shp3-g1] PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-v-shp4-g0] PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-v-shp4-g1] PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-e-shp0-g0] FAILED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-e-shp0-g1] FAILED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-e-shp1-g0] FAILED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-e-shp1-g1] FAILED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-e-shp2-g0] FAILED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-e-shp2-g1] FAILED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-e-shp3-g0] FAILED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-e-shp3-g1] FAILED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-e-shp4-g0] FAILED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-e-shp4-g1] FAILED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-u-shp0-g0] FAILED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-u-shp0-g1] /root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_sampling
start graph service on server 1 for part 1
[05:32:48] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:32:48] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:18026]...
[05:32:48] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:32:48] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
FAILED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-u-shp1-g0] FAILED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-u-shp1-g1] FAILED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-u-shp2-g0] FAILED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-u-shp2-g1] FAILED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-u-shp3-g0] FAILED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-u-shp3-g1] FAILED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-u-shp4-g0] FAILED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-u-shp4-g1] FAILED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-v-shp0-g0] FAILED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-v-shp0-g1] FAILED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-v-shp1-g0] FAILED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-v-shp1-g1] FAILED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-v-shp2-g0] FAILED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-v-shp2-g1] FAILED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-v-shp3-g0] FAILED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-v-shp3-g1] FAILED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-v-shp4-g0] FAILED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-v-shp4-g1] FAILED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-e-shp0-g0] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-e-shp0-g1] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-e-shp1-g0] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-e-shp1-g1] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-e-shp2-g0] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-e-shp2-g1] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-e-shp3-g0] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-e-shp3-g1] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-e-shp4-g0] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-e-shp4-g1] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-u-shp0-g0] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-u-shp0-g1] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-u-shp1-g0] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-u-shp1-g1] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-u-shp2-g0] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-u-shp2-g1] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-u-shp3-g0] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-u-shp3-g1] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-u-shp4-g0] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-u-shp4-g1] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-v-shp0-g0] FAILED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-v-shp0-g1] FAILED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-v-shp1-g0] FAILED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-v-shp1-g1] FAILED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-v-shp2-g0] FAILED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-v-shp2-g1] FAILED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-v-shp3-g0] FAILED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-v-shp3-g1] FAILED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-v-shp4-g0] FAILED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-v-shp4-g1] FAILED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-e-shp0-g0] FAILED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-e-shp0-g1] FAILED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-e-shp1-g0] FAILED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-e-shp1-g1] FAILED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-e-shp2-g0] FAILED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-e-shp2-g1] FAILED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-e-shp3-g0] FAILED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-e-shp3-g1] FAILED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-e-shp4-g0] FAILED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-e-shp4-g1] FAILED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-u-shp0-g0] FAILED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-u-shp0-g1] FAILED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-u-shp1-g0] FAILED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-u-shp1-g1] FAILED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-u-shp2-g0] FAILED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-u-shp2-g1] FAILED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-u-shp3-g0] FAILED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-u-shp3-g1] FAILED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-u-shp4-g0] FAILED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-u-shp4-g1] FAILED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-v-shp0-g0] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-v-shp0-g1] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-v-shp1-g0] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-v-shp1-g1] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-v-shp2-g0] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-v-shp2-g1] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-v-shp3-g0] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-v-shp3-g1] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-v-shp4-g0] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-v-shp4-g1] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-e-shp0-g0] FAILED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-e-shp0-g1] FAILED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-e-shp1-g0] FAILED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-e-shp1-g1] FAILED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-e-shp2-g0] Client [1090] waits on 172.17.0.3:36771
Machine (0) group (0) client (0) connect to server successfuly!
Client[0] in group[0] is exiting...
FAILED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-e-shp2-g1] FAILED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-e-shp3-g0] FAILED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-e-shp3-g1] FAILED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-e-shp4-g0] FAILED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-e-shp4-g1] FAILED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-u-shp0-g0] FAILED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-u-shp0-g1] FAILED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-u-shp1-g0] FAILED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-u-shp1-g1] FAILED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-u-shp2-g0] FAILED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-u-shp2-g1] FAILED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-u-shp3-g0] FAILED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-u-shp3-g1] FAILED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-u-shp4-g0] FAILED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-u-shp4-g1] FAILED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-v-shp0-g0] FAILED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-v-shp0-g1] Server (1) shutdown.
Server is exiting...
Done sampling
Server (0) shutdown.
Server is exiting...
Converting to homogeneous graph takes 0.001s, peak mem: 1.529 GB
Convert a graph into a bidirected graph: 0.000 seconds, peak memory: 1.529 GB
Construct multi-constraint weights: 0.000 seconds, peak memory: 1.529 GB
[05:32:53] /root/jenkins/workspace/dgl_PR-4648/src/graph/transform/metis_partition_hetero.cc:87: Partition a graph with 1500 nodes and 2000 edges into 2 parts and get 23 edge cuts
Metis partitioning: 0.001 seconds, peak memory: 1.529 GB
Assigning nodes to METIS partitions takes 0.002s, peak mem: 1.529 GB
Reshuffle nodes and edges: 0.001 seconds
Split the graph: 0.001 seconds
Construct subgraphs: 0.002 seconds
Splitting the graph into partitions takes 0.003s, peak mem: 1.529 GB
part 0 has 506 nodes of type game and 496 are inside the partition
part 0 has 256 nodes of type user and 244 are inside the partition
part 0 has 533 edges of type buys and 522 are inside the partition
part 1 has 516 nodes of type game and 504 are inside the partition
part 1 has 267 nodes of type user and 256 are inside the partition
part 1 has 490 edges of type buys and 478 are inside the partition
Save partitions: 0.002 seconds, peak memory: 1.529 GB
There are 1000 edges in the graph and 0 edge cuts for 2 partitions.
FAILED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-v-shp1-g0] FAILED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-v-shp1-g1] FAILED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-v-shp2-g0] FAILED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-v-shp2-g1] FAILED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-v-shp3-g0] FAILED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-v-shp3-g1] FAILED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-v-shp4-g0] FAILED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-v-shp4-g1] FAILED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-e-shp0-g0] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-e-shp0-g1] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-e-shp1-g0] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-e-shp1-g1] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-e-shp2-g0] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-e-shp2-g1] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-e-shp3-g0] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-e-shp3-g1] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-e-shp4-g0] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-e-shp4-g1] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-u-shp0-g0] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-u-shp0-g1] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-u-shp1-g0] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-u-shp1-g1] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-u-shp2-g0] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-u-shp2-g1] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-u-shp3-g0] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-u-shp3-g1] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-u-shp4-g0] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-u-shp4-g1] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-v-shp0-g0] FAILED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-v-shp0-g1] FAILED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-v-shp1-g0] FAILED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-v-shp1-g1] FAILED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-v-shp2-g0] FAILED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-v-shp2-g1] FAILED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-v-shp3-g0] FAILED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-v-shp3-g1] FAILED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-v-shp4-g0] FAILED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-v-shp4-g1] FAILED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-e-shp0-g0] /root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_sampling
start graph service on server 0 for part 0
[05:32:54] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:32:54] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:27870]...
FAILED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-e-shp0-g1] FAILED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-e-shp1-g0] FAILED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-e-shp1-g1] FAILED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-e-shp2-g0] collected 806 items

tests/mxnet/test_geometry.py::test_fps PASSED                            [  0%]
tests/mxnet/test_nn.py::test_graph_conv[1-int32] PASSED                  [  0%]
tests/mxnet/test_nn.py::test_graph_conv[1-int64] PASSED                  [  0%]
tests/mxnet/test_nn.py::test_graph_conv[2-int32] PASSED                  [  0%]
tests/mxnet/test_nn.py::test_graph_conv[2-int64] FAILED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-e-shp2-g1] FAILED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-e-shp3-g0] FAILED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-e-shp3-g1] FAILED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-e-shp4-g0] FAILED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-e-shp4-g1] PASSED                  [  0%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-none-g0-int32] PASSED [  0%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-none-g0-int64] PASSED [  0%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-none-g1-int32] PASSED [  0%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-none-g1-int64] PASSED [  1%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-none-g2-int32] PASSED [  1%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-none-g2-int64] PASSED [  1%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-none-g3-int32] PASSED [  1%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-none-g3-int64] PASSED [  1%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-none-g4-int32] PASSED [  1%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-none-g4-int64] PASSED [  1%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-none-g5-int32] PASSED [  1%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-none-g5-int64] PASSED [  2%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-none-g6-int32] PASSED [  2%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-none-g6-int64] PASSED [  2%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-both-g0-int32] PASSED [  2%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-both-g0-int64] PASSED [  2%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-both-g1-int32] PASSED [  2%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-both-g1-int64] PASSED [  2%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-both-g2-int32] PASSED [  2%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-both-g2-int64] PASSED [  3%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-both-g3-int32] PASSED [  3%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-both-g3-int64] PASSED [  3%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-both-g4-int32] PASSED [  3%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-both-g4-int64] PASSED [  3%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-both-g5-int32] PASSED [  3%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-both-g5-int64] PASSED [  3%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-both-g6-int32] PASSED [  3%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-both-g6-int64] PASSED [  4%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-right-g0-int32] PASSED [  4%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-right-g0-int64] PASSED [  4%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-right-g1-int32] PASSED [  4%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-right-g1-int64] PASSED [  4%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-right-g2-int32] PASSED [  4%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-right-g2-int64] PASSED [  4%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-right-g3-int32] PASSED [  4%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-right-g3-int64] PASSED [  5%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-right-g4-int32] PASSED [  5%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-right-g4-int64] PASSED [  5%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-right-g5-int32] PASSED [  5%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-right-g5-int64] PASSED [  5%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-right-g6-int32] PASSED [  5%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-right-g6-int64] PASSED [  5%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-left-g0-int32] PASSED [  5%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-left-g0-int64] PASSED [  6%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-left-g1-int32] PASSED [  6%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-left-g1-int64] PASSED [  6%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-left-g2-int32] PASSED [  6%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-left-g2-int64] PASSED [  6%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-left-g3-int32] PASSED [  6%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-left-g3-int64] PASSED [  6%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-left-g4-int32] PASSED [  6%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-left-g4-int64] PASSED [  7%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-left-g5-int32] PASSED [  7%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-left-g5-int64] PASSED [  7%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-left-g6-int32] PASSED [  7%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-left-g6-int64] PASSED [  7%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-none-g0-int32] PASSED [  7%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-none-g0-int64] PASSED [  7%]FAILED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-u-shp0-g0] FAILED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-u-shp0-g1] FAILED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-u-shp1-g0] FAILED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-u-shp1-g1] 
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-none-g1-int32] PASSED [  7%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-none-g1-int64] PASSED [  8%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-none-g2-int32] PASSED [  8%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-none-g2-int64] PASSED [  8%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-none-g3-int32] PASSED [  8%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-none-g3-int64] PASSED [  8%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-none-g4-int32] PASSED [  8%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-none-g4-int64] PASSED [  8%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-none-g5-int32] PASSED [  8%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-none-g5-int64] PASSED [  9%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-none-g6-int32] PASSED [  9%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-none-g6-int64] PASSED [  9%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-both-g0-int32] PASSED [  9%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-both-g0-int64] PASSED [  9%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-both-g1-int32] PASSED [  9%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-both-g1-int64] PASSED [  9%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-both-g2-int32] PASSED [  9%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-both-g2-int64] PASSED [ 10%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-both-g3-int32] PASSED [ 10%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-both-g3-int64] PASSED [ 10%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-both-g4-int32] PASSED [ 10%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-both-g4-int64] PASSED [ 10%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-both-g5-int32] PASSED [ 10%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-both-g5-int64] PASSED [ 10%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-both-g6-int32] PASSED [ 10%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-both-g6-int64] PASSED [ 11%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-right-g0-int32] PASSED [ 11%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-right-g0-int64] PASSED [ 11%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-right-g1-int32] PASSED [ 11%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-right-g1-int64] PASSED [ 11%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-right-g2-int32] PASSED [ 11%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-right-g2-int64] PASSED [ 11%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-right-g3-int32] PASSED [ 11%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-right-g3-int64] PASSED [ 12%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-right-g4-int32] PASSED [ 12%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-right-g4-int64] PASSED [ 12%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-right-g5-int32] PASSED [ 12%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-right-g5-int64] PASSED [ 12%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-right-g6-int32] PASSED [ 12%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-right-g6-int64] PASSED [ 12%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-left-g0-int32] PASSED [ 12%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-left-g0-int64] PASSED [ 13%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-left-g1-int32] PASSED [ 13%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-left-g1-int64] PASSED [ 13%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-left-g2-int32] PASSED [ 13%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-left-g2-int64] PASSED [ 13%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-left-g3-int32] PASSED [ 13%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-left-g3-int64] PASSED [ 13%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-left-g4-int32] PASSED [ 13%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-left-g4-int64] PASSED [ 14%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-left-g5-int32] PASSED [ 14%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-left-g5-int64] PASSED [ 14%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-left-g6-int32] PASSED [ 14%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-left-g6-int64] PASSED [ 14%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-none-g0-int32] PASSED [ 14%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-none-g0-int64] PASSED [ 14%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-none-g1-int32] PASSED [ 14%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-none-g1-int64] PASSED [ 15%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-none-g2-int32] PASSED [ 15%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-none-g2-int64] PASSED [ 15%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-none-g3-int32] PASSED [ 15%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-none-g3-int64] PASSED [ 15%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-none-g4-int32] PASSED [ 15%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-none-g4-int64] PASSED [ 15%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-none-g5-int32] PASSED [ 15%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-none-g5-int64] FAILED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-u-shp2-g0] FAILED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-u-shp2-g1] FAILED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-u-shp3-g0] FAILED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-u-shp3-g1] FAILED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-u-shp4-g0] PASSED [ 16%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-none-g6-int32] PASSED [ 16%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-none-g6-int64] PASSED [ 16%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-both-g0-int32] PASSED [ 16%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-both-g0-int64] PASSED [ 16%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-both-g1-int32] PASSED [ 16%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-both-g1-int64] PASSED [ 16%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-both-g2-int32] PASSED [ 16%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-both-g2-int64] PASSED [ 16%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-both-g3-int32] PASSED [ 17%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-both-g3-int64] PASSED [ 17%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-both-g4-int32] PASSED [ 17%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-both-g4-int64] PASSED [ 17%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-both-g5-int32] PASSED [ 17%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-both-g5-int64] PASSED [ 17%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-both-g6-int32] PASSED [ 17%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-both-g6-int64] PASSED [ 17%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-right-g0-int32] PASSED [ 18%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-right-g0-int64] PASSED [ 18%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-right-g1-int32] PASSED [ 18%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-right-g1-int64] PASSED [ 18%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-right-g2-int32] PASSED [ 18%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-right-g2-int64] PASSED [ 18%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-right-g3-int32] PASSED [ 18%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-right-g3-int64] PASSED [ 18%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-right-g4-int32] PASSED [ 19%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-right-g4-int64] PASSED [ 19%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-right-g5-int32] PASSED [ 19%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-right-g5-int64] PASSED [ 19%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-right-g6-int32] PASSED [ 19%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-right-g6-int64] PASSED [ 19%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-left-g0-int32] PASSED [ 19%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-left-g0-int64] PASSED [ 19%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-left-g1-int32] PASSED [ 20%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-left-g1-int64] PASSED [ 20%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-left-g2-int32] PASSED [ 20%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-left-g2-int64] PASSED [ 20%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-left-g3-int32] PASSED [ 20%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-left-g3-int64] PASSED [ 20%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-left-g4-int32] PASSED [ 20%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-left-g4-int64] PASSED [ 20%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-left-g5-int32] PASSED [ 21%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-left-g5-int64] PASSED [ 21%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-left-g6-int32] PASSED [ 21%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-left-g6-int64] PASSED [ 21%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-none-g0-int32] PASSED [ 21%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-none-g0-int64] PASSED [ 21%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-none-g1-int32] PASSED [ 21%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-none-g1-int64] PASSED [ 21%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-none-g2-int32] PASSED [ 22%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-none-g2-int64] PASSED [ 22%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-none-g3-int32] PASSED [ 22%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-none-g3-int64] PASSED [ 22%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-none-g4-int32] PASSED [ 22%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-none-g4-int64] PASSED [ 22%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-none-g5-int32] PASSED [ 22%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-none-g5-int64] PASSED [ 22%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-none-g6-int32] PASSED [ 23%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-none-g6-int64] PASSED [ 23%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-both-g0-int32] PASSED [ 23%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-both-g0-int64] FAILED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-u-shp4-g1] FAILED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-v-shp0-g0] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-v-shp0-g1] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-v-shp1-g0] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-v-shp1-g1] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-v-shp2-g0] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-v-shp2-g1] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-v-shp3-g0] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-v-shp3-g1] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-v-shp4-g0] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-v-shp4-g1] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-e-shp0-g0] FAILED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-e-shp0-g1] FAILED [ 53%]PASSED [ 23%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-both-g1-int32] PASSED [ 23%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-both-g1-int64] PASSED [ 23%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-both-g2-int32] PASSED [ 23%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-both-g2-int64] PASSED [ 23%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-both-g3-int32] PASSED [ 24%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-both-g3-int64] PASSED [ 24%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-both-g4-int32] PASSED [ 24%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-both-g4-int64] PASSED [ 24%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-both-g5-int32] PASSED [ 24%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-both-g5-int64] PASSED [ 24%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-both-g6-int32] PASSED [ 24%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-both-g6-int64] PASSED [ 24%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-right-g0-int32] PASSED [ 25%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-right-g0-int64] PASSED [ 25%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-right-g1-int32] PASSED [ 25%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-right-g1-int64] PASSED [ 25%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-right-g2-int32] PASSED [ 25%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-right-g2-int64] PASSED [ 25%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-right-g3-int32] PASSED [ 25%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-right-g3-int64] PASSED [ 25%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-right-g4-int32] PASSED [ 26%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-right-g4-int64] PASSED [ 26%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-right-g5-int32] PASSED [ 26%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-right-g5-int64] PASSED [ 26%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-right-g6-int32] PASSED [ 26%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-right-g6-int64] PASSED [ 26%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-left-g0-int32] PASSED [ 26%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-left-g0-int64] PASSED [ 26%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-left-g1-int32] PASSED [ 27%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-left-g1-int64] PASSED [ 27%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-left-g2-int32] PASSED [ 27%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-left-g2-int64] PASSED [ 27%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-left-g3-int32] PASSED [ 27%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-left-g3-int64] PASSED [ 27%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-left-g4-int32] PASSED [ 27%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-left-g4-int64] PASSED [ 27%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-left-g5-int32] PASSED [ 28%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-left-g5-int64] PASSED [ 28%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-left-g6-int32] PASSED [ 28%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-left-g6-int64] PASSED [ 28%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[1-False-True-none-g0-int32] PASSED [ 28%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[1-False-True-none-g0-int64] PASSED [ 28%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[1-False-True-none-g1-int32] PASSED [ 28%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[1-False-True-none-g1-int64] PASSED [ 28%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[1-False-True-both-g0-int32] PASSED [ 29%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[1-False-True-both-g0-int64] PASSED [ 29%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[1-False-True-both-g1-int32] PASSED [ 29%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[1-False-True-both-g1-int64] PASSED [ 29%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[1-False-True-right-g0-int32] PASSED [ 29%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[1-False-True-right-g0-int64] PASSED [ 29%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[1-False-True-right-g1-int32] PASSED [ 29%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[1-False-True-right-g1-int64] PASSED [ 29%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[1-False-False-none-g0-int32] PASSED [ 30%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[1-False-False-none-g0-int64] PASSED [ 30%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[1-False-False-none-g1-int32] PASSED [ 30%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[1-False-False-none-g1-int64] PASSED [ 30%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[1-False-False-both-g0-int32] PASSED [ 30%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[1-False-False-both-g0-int64] PASSED [ 30%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[1-False-False-both-g1-int32] PASSED [ 30%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[1-False-False-both-g1-int64] PASSED [ 30%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[1-False-False-right-g0-int32] PASSED [ 31%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[1-False-False-right-g0-int64] /root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_sampling
start graph service on server 1 for part 1
[05:32:55] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:32:55] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:27872]...
[05:32:55] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:32:55] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
PASSED [ 31%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[1-False-False-right-g1-int32] PASSED [ 31%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[1-False-False-right-g1-int64] PASSED [ 31%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[2-False-True-none-g0-int32] PASSED [ 31%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[2-False-True-none-g0-int64] PASSED [ 31%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[2-False-True-none-g1-int32] PASSED [ 31%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[2-False-True-none-g1-int64] PASSED [ 31%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[2-False-True-both-g0-int32] PASSED [ 32%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[2-False-True-both-g0-int64] PASSED [ 32%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[2-False-True-both-g1-int32] PASSED [ 32%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[2-False-True-both-g1-int64] PASSED [ 32%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[2-False-True-right-g0-int32] PASSED [ 32%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[2-False-True-right-g0-int64] PASSED [ 32%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[2-False-True-right-g1-int32] PASSED [ 32%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[2-False-True-right-g1-int64] PASSED [ 32%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[2-False-False-none-g0-int32] PASSED [ 33%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[2-False-False-none-g0-int64] PASSED [ 33%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[2-False-False-none-g1-int32] PASSED [ 33%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[2-False-False-none-g1-int64] PASSED [ 33%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[2-False-False-both-g0-int32] PASSED [ 33%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[2-False-False-both-g0-int64] PASSED [ 33%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[2-False-False-both-g1-int32] PASSED [ 33%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[2-False-False-both-g1-int64] PASSED [ 33%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[2-False-False-right-g0-int32] PASSED [ 33%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[2-False-False-right-g0-int64] PASSED [ 34%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[2-False-False-right-g1-int32] PASSED [ 34%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[2-False-False-right-g1-int64] PASSED [ 34%]
tests/mxnet/test_nn.py::test_tagconv[1] PASSED                           [ 34%]
tests/mxnet/test_nn.py::test_tagconv[2] PASSED                           [ 34%]
tests/mxnet/test_nn.py::test_gat_conv[1-1-g0-int32] PASSED               [ 34%]
tests/mxnet/test_nn.py::test_gat_conv[1-1-g0-int64] PASSED               [ 34%]
tests/mxnet/test_nn.py::test_gat_conv[1-1-g1-int32] PASSED               [ 34%]
tests/mxnet/test_nn.py::test_gat_conv[1-1-g1-int64] PASSED               [ 35%]
tests/mxnet/test_nn.py::test_gat_conv[1-1-g2-int32] PASSED               [ 35%]
tests/mxnet/test_nn.py::test_gat_conv[1-1-g2-int64] PASSED               [ 35%]
tests/mxnet/test_nn.py::test_gat_conv[1-1-g3-int32] PASSED               [ 35%]
tests/mxnet/test_nn.py::test_gat_conv[1-1-g3-int64] PASSED               [ 35%]
tests/mxnet/test_nn.py::test_gat_conv[1-1-g4-int32] PASSED               [ 35%]
tests/mxnet/test_nn.py::test_gat_conv[1-1-g4-int64] 
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-e-shp1-g0] FAILED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-e-shp1-g1] FAILED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-e-shp2-g0] FAILED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-e-shp2-g1] FAILED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-e-shp3-g0] PASSED               [ 35%]
tests/mxnet/test_nn.py::test_gat_conv[1-1-g5-int32] PASSED               [ 35%]
tests/mxnet/test_nn.py::test_gat_conv[1-1-g5-int64] PASSED               [ 36%]
tests/mxnet/test_nn.py::test_gat_conv[1-1-g6-int32] PASSED               [ 36%]
tests/mxnet/test_nn.py::test_gat_conv[1-1-g6-int64] PASSED               [ 36%]
tests/mxnet/test_nn.py::test_gat_conv[1-20-g0-int32] PASSED              [ 36%]
tests/mxnet/test_nn.py::test_gat_conv[1-20-g0-int64] PASSED              [ 36%]
tests/mxnet/test_nn.py::test_gat_conv[1-20-g1-int32] PASSED              [ 36%]
tests/mxnet/test_nn.py::test_gat_conv[1-20-g1-int64] PASSED              [ 36%]
tests/mxnet/test_nn.py::test_gat_conv[1-20-g2-int32] PASSED              [ 36%]
tests/mxnet/test_nn.py::test_gat_conv[1-20-g2-int64] PASSED              [ 37%]
tests/mxnet/test_nn.py::test_gat_conv[1-20-g3-int32] PASSED              [ 37%]
tests/mxnet/test_nn.py::test_gat_conv[1-20-g3-int64] PASSED              [ 37%]
tests/mxnet/test_nn.py::test_gat_conv[1-20-g4-int32] PASSED              [ 37%]
tests/mxnet/test_nn.py::test_gat_conv[1-20-g4-int64] PASSED              [ 37%]
tests/mxnet/test_nn.py::test_gat_conv[1-20-g5-int32] PASSED              [ 37%]
tests/mxnet/test_nn.py::test_gat_conv[1-20-g5-int64] PASSED              [ 37%]
tests/mxnet/test_nn.py::test_gat_conv[1-20-g6-int32] PASSED              [ 37%]
tests/mxnet/test_nn.py::test_gat_conv[1-20-g6-int64] FAILED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-e-shp3-g1] FAILED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-e-shp4-g0] FAILED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-e-shp4-g1] FAILED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-u-shp0-g0] FAILED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-u-shp0-g1] PASSED              [ 38%]
tests/mxnet/test_nn.py::test_gat_conv[5-1-g0-int32] PASSED               [ 38%]
tests/mxnet/test_nn.py::test_gat_conv[5-1-g0-int64] PASSED               [ 38%]
tests/mxnet/test_nn.py::test_gat_conv[5-1-g1-int32] PASSED               [ 38%]
tests/mxnet/test_nn.py::test_gat_conv[5-1-g1-int64] PASSED               [ 38%]
tests/mxnet/test_nn.py::test_gat_conv[5-1-g2-int32] PASSED               [ 38%]
tests/mxnet/test_nn.py::test_gat_conv[5-1-g2-int64] PASSED               [ 38%]
tests/mxnet/test_nn.py::test_gat_conv[5-1-g3-int32] PASSED               [ 38%]
tests/mxnet/test_nn.py::test_gat_conv[5-1-g3-int64] PASSED               [ 39%]
tests/mxnet/test_nn.py::test_gat_conv[5-1-g4-int32] PASSED               [ 39%]
tests/mxnet/test_nn.py::test_gat_conv[5-1-g4-int64] PASSED               [ 39%]
tests/mxnet/test_nn.py::test_gat_conv[5-1-g5-int32] PASSED               [ 39%]
tests/mxnet/test_nn.py::test_gat_conv[5-1-g5-int64] PASSED               [ 39%]
tests/mxnet/test_nn.py::test_gat_conv[5-1-g6-int32] PASSED               [ 39%]
tests/mxnet/test_nn.py::test_gat_conv[5-1-g6-int64] PASSED               [ 39%]
tests/mxnet/test_nn.py::test_gat_conv[5-20-g0-int32] PASSED              [ 39%]
tests/mxnet/test_nn.py::test_gat_conv[5-20-g0-int64] PASSED              [ 40%]
tests/mxnet/test_nn.py::test_gat_conv[5-20-g1-int32] PASSED              [ 40%]
tests/mxnet/test_nn.py::test_gat_conv[5-20-g1-int64] FAILED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-u-shp1-g0] FAILED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-u-shp1-g1] FAILED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-u-shp2-g0] FAILED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-u-shp2-g1] FAILED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-u-shp3-g0] PASSED              [ 40%]
tests/mxnet/test_nn.py::test_gat_conv[5-20-g2-int32] PASSED              [ 40%]
tests/mxnet/test_nn.py::test_gat_conv[5-20-g2-int64] PASSED              [ 40%]
tests/mxnet/test_nn.py::test_gat_conv[5-20-g3-int32] PASSED              [ 40%]
tests/mxnet/test_nn.py::test_gat_conv[5-20-g3-int64] PASSED              [ 40%]
tests/mxnet/test_nn.py::test_gat_conv[5-20-g4-int32] PASSED              [ 40%]
tests/mxnet/test_nn.py::test_gat_conv[5-20-g4-int64] PASSED              [ 41%]
tests/mxnet/test_nn.py::test_gat_conv[5-20-g5-int32] PASSED              [ 41%]
tests/mxnet/test_nn.py::test_gat_conv[5-20-g5-int64] PASSED              [ 41%]
tests/mxnet/test_nn.py::test_gat_conv[5-20-g6-int32] PASSED              [ 41%]
tests/mxnet/test_nn.py::test_gat_conv[5-20-g6-int64] PASSED              [ 41%]
tests/mxnet/test_nn.py::test_gat_conv_bi[1-1-g0-int32] PASSED            [ 41%]
tests/mxnet/test_nn.py::test_gat_conv_bi[1-1-g0-int64] PASSED            [ 41%]
tests/mxnet/test_nn.py::test_gat_conv_bi[1-1-g1-int32] PASSED            [ 41%]
tests/mxnet/test_nn.py::test_gat_conv_bi[1-1-g1-int64] PASSED            [ 42%]
tests/mxnet/test_nn.py::test_gat_conv_bi[1-2-g0-int32] PASSED            [ 42%]
tests/mxnet/test_nn.py::test_gat_conv_bi[1-2-g0-int64] PASSED            [ 42%]
tests/mxnet/test_nn.py::test_gat_conv_bi[1-2-g1-int32] FAILED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-u-shp3-g1] FAILED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-u-shp4-g0] FAILED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-u-shp4-g1] FAILED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-v-shp0-g0] FAILED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-v-shp0-g1] PASSED            [ 42%]
tests/mxnet/test_nn.py::test_gat_conv_bi[1-2-g1-int64] PASSED            [ 42%]
tests/mxnet/test_nn.py::test_gat_conv_bi[4-1-g0-int32] PASSED            [ 42%]
tests/mxnet/test_nn.py::test_gat_conv_bi[4-1-g0-int64] PASSED            [ 42%]
tests/mxnet/test_nn.py::test_gat_conv_bi[4-1-g1-int32] PASSED            [ 42%]
tests/mxnet/test_nn.py::test_gat_conv_bi[4-1-g1-int64] PASSED            [ 43%]
tests/mxnet/test_nn.py::test_gat_conv_bi[4-2-g0-int32] PASSED            [ 43%]
tests/mxnet/test_nn.py::test_gat_conv_bi[4-2-g0-int64] PASSED            [ 43%]
tests/mxnet/test_nn.py::test_gat_conv_bi[4-2-g1-int32] PASSED            [ 43%]
tests/mxnet/test_nn.py::test_gat_conv_bi[4-2-g1-int64] PASSED            [ 43%]
tests/mxnet/test_nn.py::test_sage_conv[1-mean-g0-int32] PASSED           [ 43%]
tests/mxnet/test_nn.py::test_sage_conv[1-mean-g0-int64] PASSED           [ 43%]
tests/mxnet/test_nn.py::test_sage_conv[1-mean-g1-int32] PASSED           [ 43%]
tests/mxnet/test_nn.py::test_sage_conv[1-mean-g1-int64] PASSED           [ 44%]
tests/mxnet/test_nn.py::test_sage_conv[1-mean-g2-int32] PASSED           [ 44%]
tests/mxnet/test_nn.py::test_sage_conv[1-mean-g2-int64] PASSED           [ 44%]
tests/mxnet/test_nn.py::test_sage_conv[1-mean-g3-int32] PASSED           [ 44%]
tests/mxnet/test_nn.py::test_sage_conv[1-mean-g3-int64] PASSED           [ 44%]
tests/mxnet/test_nn.py::test_sage_conv[1-mean-g4-int32] PASSED           [ 44%]
tests/mxnet/test_nn.py::test_sage_conv[1-mean-g4-int64] PASSED           [ 44%]
tests/mxnet/test_nn.py::test_sage_conv[1-mean-g5-int32] PASSED           [ 44%]
tests/mxnet/test_nn.py::test_sage_conv[1-mean-g5-int64] PASSED           [ 45%]
tests/mxnet/test_nn.py::test_sage_conv[1-mean-g6-int32] PASSED           [ 45%]
tests/mxnet/test_nn.py::test_sage_conv[1-mean-g6-int64] PASSED           [ 45%]
tests/mxnet/test_nn.py::test_sage_conv[1-mean-g7-int32] PASSED           [ 45%]
tests/mxnet/test_nn.py::test_sage_conv[1-mean-g7-int64] PASSED           [ 45%]
tests/mxnet/test_nn.py::test_sage_conv[1-pool-g0-int32] PASSED           [ 45%]
tests/mxnet/test_nn.py::test_sage_conv[1-pool-g0-int64] PASSED           [ 45%]
tests/mxnet/test_nn.py::test_sage_conv[1-pool-g1-int32] PASSED           [ 45%]
tests/mxnet/test_nn.py::test_sage_conv[1-pool-g1-int64] PASSED           [ 46%]
tests/mxnet/test_nn.py::test_sage_conv[1-pool-g2-int32] PASSED           [ 46%]
tests/mxnet/test_nn.py::test_sage_conv[1-pool-g2-int64] PASSED           [ 46%]
tests/mxnet/test_nn.py::test_sage_conv[1-pool-g3-int32] PASSED           [ 46%]
tests/mxnet/test_nn.py::test_sage_conv[1-pool-g3-int64] PASSED           [ 46%]
tests/mxnet/test_nn.py::test_sage_conv[1-pool-g4-int32] PASSED           [ 46%]
tests/mxnet/test_nn.py::test_sage_conv[1-pool-g4-int64] PASSED           [ 46%]
tests/mxnet/test_nn.py::test_sage_conv[1-pool-g5-int32] PASSED           [ 46%]
tests/mxnet/test_nn.py::test_sage_conv[1-pool-g5-int64] PASSED           [ 47%]
tests/mxnet/test_nn.py::test_sage_conv[1-pool-g6-int32] PASSED           [ 47%]
tests/mxnet/test_nn.py::test_sage_conv[1-pool-g6-int64] PASSED           [ 47%]
tests/mxnet/test_nn.py::test_sage_conv[1-pool-g7-int32] PASSED           [ 47%]
tests/mxnet/test_nn.py::test_sage_conv[1-pool-g7-int64] FAILED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-v-shp1-g0] FAILED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-v-shp1-g1] FAILED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-v-shp2-g0] FAILED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-v-shp2-g1] FAILED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-v-shp3-g0] PASSED           [ 47%]
tests/mxnet/test_nn.py::test_sage_conv[1-gcn-g0-int32] PASSED            [ 47%]
tests/mxnet/test_nn.py::test_sage_conv[1-gcn-g0-int64] PASSED            [ 47%]
tests/mxnet/test_nn.py::test_sage_conv[1-gcn-g1-int32] PASSED            [ 47%]
tests/mxnet/test_nn.py::test_sage_conv[1-gcn-g1-int64] PASSED            [ 48%]
tests/mxnet/test_nn.py::test_sage_conv[1-gcn-g2-int32] PASSED            [ 48%]
tests/mxnet/test_nn.py::test_sage_conv[1-gcn-g2-int64] PASSED            [ 48%]
tests/mxnet/test_nn.py::test_sage_conv[1-gcn-g3-int32] PASSED            [ 48%]
tests/mxnet/test_nn.py::test_sage_conv[1-gcn-g3-int64] PASSED            [ 48%]
tests/mxnet/test_nn.py::test_sage_conv[1-gcn-g4-int32] PASSED            [ 48%]
tests/mxnet/test_nn.py::test_sage_conv[1-gcn-g4-int64] PASSED            [ 48%]
tests/mxnet/test_nn.py::test_sage_conv[1-gcn-g5-int32] PASSED            [ 48%]
tests/mxnet/test_nn.py::test_sage_conv[1-gcn-g5-int64] PASSED            [ 49%]
tests/mxnet/test_nn.py::test_sage_conv[1-gcn-g6-int32] PASSED            [ 49%]
tests/mxnet/test_nn.py::test_sage_conv[1-gcn-g6-int64] PASSED            [ 49%]
tests/mxnet/test_nn.py::test_sage_conv[1-gcn-g7-int32] PASSED            [ 49%]
tests/mxnet/test_nn.py::test_sage_conv[1-gcn-g7-int64] PASSED            [ 49%]
tests/mxnet/test_nn.py::test_sage_conv[10-mean-g0-int32] PASSED          [ 49%]
tests/mxnet/test_nn.py::test_sage_conv[10-mean-g0-int64] PASSED          [ 49%]
tests/mxnet/test_nn.py::test_sage_conv[10-mean-g1-int32] PASSED          [ 49%]
tests/mxnet/test_nn.py::test_sage_conv[10-mean-g1-int64] PASSED          [ 50%]
tests/mxnet/test_nn.py::test_sage_conv[10-mean-g2-int32] PASSED          [ 50%]
tests/mxnet/test_nn.py::test_sage_conv[10-mean-g2-int64] PASSED          [ 50%]
tests/mxnet/test_nn.py::test_sage_conv[10-mean-g3-int32] PASSED          [ 50%]
tests/mxnet/test_nn.py::test_sage_conv[10-mean-g3-int64] PASSED          [ 50%]
tests/mxnet/test_nn.py::test_sage_conv[10-mean-g4-int32] PASSED          [ 50%]
tests/mxnet/test_nn.py::test_sage_conv[10-mean-g4-int64] PASSED          [ 50%]
tests/mxnet/test_nn.py::test_sage_conv[10-mean-g5-int32] PASSED          [ 50%]
tests/mxnet/test_nn.py::test_sage_conv[10-mean-g5-int64] PASSED          [ 50%]
tests/mxnet/test_nn.py::test_sage_conv[10-mean-g6-int32] PASSED          [ 51%]
tests/mxnet/test_nn.py::test_sage_conv[10-mean-g6-int64] PASSED          [ 51%]
tests/mxnet/test_nn.py::test_sage_conv[10-mean-g7-int32] PASSED          [ 51%]
tests/mxnet/test_nn.py::test_sage_conv[10-mean-g7-int64] PASSED          [ 51%]
tests/mxnet/test_nn.py::test_sage_conv[10-pool-g0-int32] PASSED          [ 51%]
tests/mxnet/test_nn.py::test_sage_conv[10-pool-g0-int64] PASSED          [ 51%]
tests/mxnet/test_nn.py::test_sage_conv[10-pool-g1-int32] PASSED          [ 51%]
tests/mxnet/test_nn.py::test_sage_conv[10-pool-g1-int64] PASSED          [ 51%]
tests/mxnet/test_nn.py::test_sage_conv[10-pool-g2-int32] PASSED          [ 52%]
tests/mxnet/test_nn.py::test_sage_conv[10-pool-g2-int64] PASSED          [ 52%]
tests/mxnet/test_nn.py::test_sage_conv[10-pool-g3-int32] PASSED          [ 52%]
tests/mxnet/test_nn.py::test_sage_conv[10-pool-g3-int64] PASSED          [ 52%]
tests/mxnet/test_nn.py::test_sage_conv[10-pool-g4-int32] PASSED          [ 52%]
tests/mxnet/test_nn.py::test_sage_conv[10-pool-g4-int64] PASSED          [ 52%]
tests/mxnet/test_nn.py::test_sage_conv[10-pool-g5-int32] PASSED          [ 52%]
tests/mxnet/test_nn.py::test_sage_conv[10-pool-g5-int64] PASSED          [ 52%]
tests/mxnet/test_nn.py::test_sage_conv[10-pool-g6-int32] PASSED          [ 53%]
tests/mxnet/test_nn.py::test_sage_conv[10-pool-g6-int64] PASSED          [ 53%]
tests/mxnet/test_nn.py::test_sage_conv[10-pool-g7-int32] PASSED          [ 53%]
tests/mxnet/test_nn.py::test_sage_conv[10-pool-g7-int64] PASSED          [ 53%]
tests/mxnet/test_nn.py::test_sage_conv[10-gcn-g0-int32] PASSED           [ 53%]
tests/mxnet/test_nn.py::test_sage_conv[10-gcn-g0-int64] PASSED           [ 53%]
tests/mxnet/test_nn.py::test_sage_conv[10-gcn-g1-int32] PASSED           [ 53%]
tests/mxnet/test_nn.py::test_sage_conv[10-gcn-g1-int64] FAILED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-v-shp3-g1] FAILED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-v-shp4-g0] FAILED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-v-shp4-g1] FAILED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-e-shp0-g0] PASSED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-e-shp0-g1] PASSED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-e-shp1-g0] PASSED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-e-shp1-g1] PASSED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-e-shp2-g0] PASSED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-e-shp2-g1] PASSED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-e-shp3-g0] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-e-shp3-g1] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-e-shp4-g0] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-e-shp4-g1] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-u-shp0-g0] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-u-shp0-g1] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-u-shp1-g0] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-u-shp1-g1] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-u-shp2-g0] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-u-shp2-g1] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-u-shp3-g0] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-u-shp3-g1] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-u-shp4-g0] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-u-shp4-g1] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-v-shp0-g0] PASSED           [ 53%]
tests/mxnet/test_nn.py::test_sage_conv[10-gcn-g2-int32] PASSED           [ 54%]
tests/mxnet/test_nn.py::test_sage_conv[10-gcn-g2-int64] PASSED           [ 54%]
tests/mxnet/test_nn.py::test_sage_conv[10-gcn-g3-int32] PASSED           [ 54%]
tests/mxnet/test_nn.py::test_sage_conv[10-gcn-g3-int64] PASSED           [ 54%]
tests/mxnet/test_nn.py::test_sage_conv[10-gcn-g4-int32] PASSED           [ 54%]
tests/mxnet/test_nn.py::test_sage_conv[10-gcn-g4-int64] PASSED           [ 54%]
tests/mxnet/test_nn.py::test_sage_conv[10-gcn-g5-int32] PASSED           [ 54%]
tests/mxnet/test_nn.py::test_sage_conv[10-gcn-g5-int64] PASSED           [ 54%]
tests/mxnet/test_nn.py::test_sage_conv[10-gcn-g6-int32] PASSED           [ 55%]
tests/mxnet/test_nn.py::test_sage_conv[10-gcn-g6-int64] PASSED           [ 55%]
tests/mxnet/test_nn.py::test_sage_conv[10-gcn-g7-int32] PASSED           [ 55%]
tests/mxnet/test_nn.py::test_sage_conv[10-gcn-g7-int64] PASSED           [ 55%]
tests/mxnet/test_nn.py::test_sage_conv_bi[1-mean-g0-int32] PASSED        [ 55%]
tests/mxnet/test_nn.py::test_sage_conv_bi[1-mean-g0-int64] PASSED        [ 55%]
tests/mxnet/test_nn.py::test_sage_conv_bi[1-mean-g1-int32] PASSED        [ 55%]
tests/mxnet/test_nn.py::test_sage_conv_bi[1-mean-g1-int64] PASSED        [ 55%]
tests/mxnet/test_nn.py::test_sage_conv_bi[1-mean-g2-int32] PASSED        [ 56%]
tests/mxnet/test_nn.py::test_sage_conv_bi[1-mean-g2-int64] PASSED        [ 56%]
tests/mxnet/test_nn.py::test_sage_conv_bi[1-pool-g0-int32] PASSED        [ 56%]
tests/mxnet/test_nn.py::test_sage_conv_bi[1-pool-g0-int64] PASSED        [ 56%]
tests/mxnet/test_nn.py::test_sage_conv_bi[1-pool-g1-int32] PASSED        [ 56%]
tests/mxnet/test_nn.py::test_sage_conv_bi[1-pool-g1-int64] PASSED        [ 56%]
tests/mxnet/test_nn.py::test_sage_conv_bi[1-pool-g2-int32] PASSED        [ 56%]
tests/mxnet/test_nn.py::test_sage_conv_bi[1-pool-g2-int64] PASSED        [ 56%]
tests/mxnet/test_nn.py::test_sage_conv_bi[1-gcn-g0-int32] PASSED         [ 57%]
tests/mxnet/test_nn.py::test_sage_conv_bi[1-gcn-g0-int64] PASSED         [ 57%]
tests/mxnet/test_nn.py::test_sage_conv_bi[1-gcn-g1-int32] PASSED         [ 57%]
tests/mxnet/test_nn.py::test_sage_conv_bi[1-gcn-g1-int64] PASSED         [ 57%]
tests/mxnet/test_nn.py::test_sage_conv_bi[1-gcn-g2-int32] PASSED         [ 57%]
tests/mxnet/test_nn.py::test_sage_conv_bi[1-gcn-g2-int64] PASSED         [ 57%]
tests/mxnet/test_nn.py::test_sage_conv_bi[2-mean-g0-int32] PASSED        [ 57%]
tests/mxnet/test_nn.py::test_sage_conv_bi[2-mean-g0-int64] PASSED        [ 57%]
tests/mxnet/test_nn.py::test_sage_conv_bi[2-mean-g1-int32] PASSED        [ 58%]
tests/mxnet/test_nn.py::test_sage_conv_bi[2-mean-g1-int64] PASSED        [ 58%]
tests/mxnet/test_nn.py::test_sage_conv_bi[2-mean-g2-int32] PASSED        [ 58%]
tests/mxnet/test_nn.py::test_sage_conv_bi[2-mean-g2-int64] PASSED        [ 58%]
tests/mxnet/test_nn.py::test_sage_conv_bi[2-pool-g0-int32] PASSED        [ 58%]
tests/mxnet/test_nn.py::test_sage_conv_bi[2-pool-g0-int64] PASSED        [ 58%]
tests/mxnet/test_nn.py::test_sage_conv_bi[2-pool-g1-int32] PASSED        [ 58%]
tests/mxnet/test_nn.py::test_sage_conv_bi[2-pool-g1-int64] PASSED        [ 58%]
tests/mxnet/test_nn.py::test_sage_conv_bi[2-pool-g2-int32] PASSED        [ 59%]
tests/mxnet/test_nn.py::test_sage_conv_bi[2-pool-g2-int64] PASSED        [ 59%]
tests/mxnet/test_nn.py::test_sage_conv_bi[2-gcn-g0-int32] PASSED         [ 59%]
tests/mxnet/test_nn.py::test_sage_conv_bi[2-gcn-g0-int64] PASSED         [ 59%]
tests/mxnet/test_nn.py::test_sage_conv_bi[2-gcn-g1-int32] PASSED         [ 59%]
tests/mxnet/test_nn.py::test_sage_conv_bi[2-gcn-g1-int64] PASSED         [ 59%]
tests/mxnet/test_nn.py::test_sage_conv_bi[2-gcn-g2-int32] PASSED         [ 59%]
tests/mxnet/test_nn.py::test_sage_conv_bi[2-gcn-g2-int64] PASSED         [ 59%]
tests/mxnet/test_nn.py::test_sage_conv_bi2[1-mean-int32] PASSED          [ 60%]
tests/mxnet/test_nn.py::test_sage_conv_bi2[1-mean-int64] FAILED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-v-shp0-g1] FAILED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-v-shp1-g0] FAILED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-v-shp1-g1] FAILED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-v-shp2-g0] FAILED [ 55%]PASSED          [ 60%]
tests/mxnet/test_nn.py::test_sage_conv_bi2[1-pool-int32] PASSED          [ 60%]
tests/mxnet/test_nn.py::test_sage_conv_bi2[1-pool-int64] PASSED          [ 60%]
tests/mxnet/test_nn.py::test_sage_conv_bi2[1-gcn-int32] PASSED           [ 60%]
tests/mxnet/test_nn.py::test_sage_conv_bi2[1-gcn-int64] PASSED           [ 60%]
tests/mxnet/test_nn.py::test_sage_conv_bi2[2-mean-int32] PASSED          [ 60%]
tests/mxnet/test_nn.py::test_sage_conv_bi2[2-mean-int64] PASSED          [ 60%]
tests/mxnet/test_nn.py::test_sage_conv_bi2[2-pool-int32] PASSED          [ 61%]
tests/mxnet/test_nn.py::test_sage_conv_bi2[2-pool-int64] PASSED          [ 61%]
tests/mxnet/test_nn.py::test_sage_conv_bi2[2-gcn-int32] PASSED           [ 61%]
tests/mxnet/test_nn.py::test_sage_conv_bi2[2-gcn-int64] PASSED           [ 61%]
tests/mxnet/test_nn.py::test_gg_conv PASSED                              [ 61%]
tests/mxnet/test_nn.py::test_cheb_conv[1] PASSED                         [ 61%]
tests/mxnet/test_nn.py::test_cheb_conv[20] PASSED                        [ 61%]
tests/mxnet/test_nn.py::test_agnn_conv[g0-int32] PASSED                  [ 61%]
tests/mxnet/test_nn.py::test_agnn_conv[g0-int64] PASSED                  [ 62%]
tests/mxnet/test_nn.py::test_agnn_conv[g1-int32] PASSED                  [ 62%]
tests/mxnet/test_nn.py::test_agnn_conv[g1-int64] PASSED                  [ 62%]
tests/mxnet/test_nn.py::test_agnn_conv[g2-int32] PASSED                  [ 62%]
tests/mxnet/test_nn.py::test_agnn_conv[g2-int64] PASSED                  [ 62%]
tests/mxnet/test_nn.py::test_agnn_conv[g3-int32] PASSED                  [ 62%]
tests/mxnet/test_nn.py::test_agnn_conv[g3-int64] PASSED                  [ 62%]
tests/mxnet/test_nn.py::test_agnn_conv[g4-int32] PASSED                  [ 62%]
tests/mxnet/test_nn.py::test_agnn_conv[g4-int64] PASSED                  [ 63%]
tests/mxnet/test_nn.py::test_agnn_conv[g5-int32] PASSED                  [ 63%]
tests/mxnet/test_nn.py::test_agnn_conv[g5-int64] PASSED                  [ 63%]
tests/mxnet/test_nn.py::test_agnn_conv[g6-int32] PASSED                  [ 63%]
tests/mxnet/test_nn.py::test_agnn_conv[g6-int64] PASSED                  [ 63%]
tests/mxnet/test_nn.py::test_agnn_conv_bi[g0-int32] PASSED               [ 63%]
tests/mxnet/test_nn.py::test_agnn_conv_bi[g0-int64] 
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-v-shp2-g1] FAILED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-v-shp3-g0] FAILED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-v-shp3-g1] FAILED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-v-shp4-g0] FAILED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-v-shp4-g1] FAILED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-e-shp0-g0] PASSED               [ 63%]
tests/mxnet/test_nn.py::test_agnn_conv_bi[g1-int32] PASSED               [ 63%]
tests/mxnet/test_nn.py::test_agnn_conv_bi[g1-int64] PASSED               [ 64%]
tests/mxnet/test_nn.py::test_appnp_conv PASSED                           [ 64%]
tests/mxnet/test_nn.py::test_dense_cheb_conv[1] PASSED                   [ 64%]
tests/mxnet/test_nn.py::test_dense_cheb_conv[2] PASSED                   [ 64%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g0-both-int32] PASSED    [ 64%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g0-both-int64] PASSED    [ 64%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g0-right-int32] PASSED   [ 64%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g0-right-int64] PASSED   [ 64%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g0-none-int32] PASSED    [ 65%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g0-none-int64] PASSED    [ 65%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g1-both-int32] PASSED    [ 65%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g1-both-int64] PASSED    [ 65%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g1-right-int32] PASSED   [ 65%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g1-right-int64] PASSED   [ 65%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g1-none-int32] PASSED    [ 65%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g1-none-int64] PASSED    [ 65%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g2-both-int32] PASSED    [ 66%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g2-both-int64] PASSED    [ 66%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g2-right-int32] PASSED   [ 66%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g2-right-int64] PASSED   [ 66%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g2-none-int32] PASSED    [ 66%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g2-none-int64] PASSED    [ 66%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g3-both-int32] PASSED    [ 66%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g3-both-int64] PASSED    [ 66%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g3-right-int32] PASSED   [ 66%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g3-right-int64] PASSED   [ 67%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g3-none-int32] PASSED    [ 67%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g3-none-int64] PASSED    [ 67%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g4-both-int32] PASSED    [ 67%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g4-both-int64] PASSED    [ 67%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g4-right-int32] FAILED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-e-shp0-g1] FAILED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-e-shp1-g0] FAILED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-e-shp1-g1] FAILED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-e-shp2-g0] PASSED   [ 67%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g4-right-int64] PASSED   [ 67%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g4-none-int32] PASSED    [ 67%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g4-none-int64] PASSED    [ 68%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g5-both-int32] PASSED    [ 68%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g5-both-int64] PASSED    [ 68%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g5-right-int32] PASSED   [ 68%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g5-right-int64] PASSED   [ 68%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g5-none-int32] PASSED    [ 68%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g5-none-int64] PASSED    [ 68%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g6-both-int32] PASSED    [ 68%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g6-both-int64] PASSED    [ 69%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g6-right-int32] PASSED   [ 69%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g6-right-int64] PASSED   [ 69%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g6-none-int32] PASSED    [ 69%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g6-none-int64] PASSED    [ 69%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g0-both-int32] PASSED    [ 69%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g0-both-int64] PASSED    [ 69%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g0-right-int32] PASSED   [ 69%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g0-right-int64] PASSED   [ 70%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g0-none-int32] PASSED    [ 70%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g0-none-int64] PASSED    [ 70%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g1-both-int32] PASSED    [ 70%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g1-both-int64] PASSED    [ 70%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g1-right-int32] PASSED   [ 70%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g1-right-int64] PASSED   [ 70%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g1-none-int32] PASSED    [ 70%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g1-none-int64] PASSED    [ 71%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g2-both-int32] PASSED    [ 71%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g2-both-int64] PASSED    [ 71%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g2-right-int32] PASSED   [ 71%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g2-right-int64] PASSED   [ 71%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g2-none-int32] PASSED    [ 71%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g2-none-int64] PASSED    [ 71%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g3-both-int32] PASSED    [ 71%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g3-both-int64] PASSED    [ 72%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g3-right-int32] PASSED   [ 72%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g3-right-int64] PASSED   [ 72%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g3-none-int32] FAILED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-e-shp2-g1] FAILED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-e-shp3-g0] FAILED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-e-shp3-g1] FAILED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-e-shp4-g0] FAILED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-e-shp4-g1] PASSED    [ 72%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g3-none-int64] PASSED    [ 72%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g4-both-int32] PASSED    [ 72%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g4-both-int64] PASSED    [ 72%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g4-right-int32] PASSED   [ 72%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g4-right-int64] PASSED   [ 73%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g4-none-int32] PASSED    [ 73%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g4-none-int64] PASSED    [ 73%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g5-both-int32] PASSED    [ 73%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g5-both-int64] PASSED    [ 73%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g5-right-int32] PASSED   [ 73%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g5-right-int64] PASSED   [ 73%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g5-none-int32] PASSED    [ 73%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g5-none-int64] PASSED    [ 74%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g6-both-int32] PASSED    [ 74%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g6-both-int64] PASSED    [ 74%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g6-right-int32] PASSED   [ 74%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g6-right-int64] PASSED   [ 74%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g6-none-int32] PASSED    [ 74%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g6-none-int64] PASSED    [ 74%]
tests/mxnet/test_nn.py::test_dense_sage_conv[1-g0-int32] PASSED          [ 74%]
tests/mxnet/test_nn.py::test_dense_sage_conv[1-g0-int64] PASSED          [ 75%]
tests/mxnet/test_nn.py::test_dense_sage_conv[1-g1-int32] PASSED          [ 75%]
tests/mxnet/test_nn.py::test_dense_sage_conv[1-g1-int64] PASSED          [ 75%]
tests/mxnet/test_nn.py::test_dense_sage_conv[1-g2-int32] PASSED          [ 75%]
tests/mxnet/test_nn.py::test_dense_sage_conv[1-g2-int64] PASSED          [ 75%]
tests/mxnet/test_nn.py::test_dense_sage_conv[1-g3-int32] PASSED          [ 75%]
tests/mxnet/test_nn.py::test_dense_sage_conv[1-g3-int64] PASSED          [ 75%]
tests/mxnet/test_nn.py::test_dense_sage_conv[1-g4-int32] PASSED          [ 75%]
tests/mxnet/test_nn.py::test_dense_sage_conv[1-g4-int64] PASSED          [ 76%]
tests/mxnet/test_nn.py::test_dense_sage_conv[1-g5-int32] PASSED          [ 76%]
tests/mxnet/test_nn.py::test_dense_sage_conv[1-g5-int64] PASSED          [ 76%]
tests/mxnet/test_nn.py::test_dense_sage_conv[1-g6-int32] PASSED          [ 76%]
tests/mxnet/test_nn.py::test_dense_sage_conv[1-g6-int64] PASSED          [ 76%]
tests/mxnet/test_nn.py::test_dense_sage_conv[1-g7-int32] PASSED          [ 76%]
tests/mxnet/test_nn.py::test_dense_sage_conv[1-g7-int64] PASSED          [ 76%]
tests/mxnet/test_nn.py::test_dense_sage_conv[1-g8-int32] PASSED          [ 76%]
tests/mxnet/test_nn.py::test_dense_sage_conv[1-g8-int64] PASSED          [ 77%]
tests/mxnet/test_nn.py::test_dense_sage_conv[1-g9-int32] PASSED          [ 77%]
tests/mxnet/test_nn.py::test_dense_sage_conv[1-g9-int64] FAILED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-u-shp0-g0] FAILED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-u-shp0-g1] FAILED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-u-shp1-g0] FAILED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-u-shp1-g1] PASSED          [ 77%]
tests/mxnet/test_nn.py::test_dense_sage_conv[2-g0-int32] PASSED          [ 77%]
tests/mxnet/test_nn.py::test_dense_sage_conv[2-g0-int64] PASSED          [ 77%]
tests/mxnet/test_nn.py::test_dense_sage_conv[2-g1-int32] PASSED          [ 77%]
tests/mxnet/test_nn.py::test_dense_sage_conv[2-g1-int64] PASSED          [ 77%]
tests/mxnet/test_nn.py::test_dense_sage_conv[2-g2-int32] PASSED          [ 77%]
tests/mxnet/test_nn.py::test_dense_sage_conv[2-g2-int64] PASSED          [ 78%]
tests/mxnet/test_nn.py::test_dense_sage_conv[2-g3-int32] PASSED          [ 78%]
tests/mxnet/test_nn.py::test_dense_sage_conv[2-g3-int64] PASSED          [ 78%]
tests/mxnet/test_nn.py::test_dense_sage_conv[2-g4-int32] PASSED          [ 78%]
tests/mxnet/test_nn.py::test_dense_sage_conv[2-g4-int64] PASSED          [ 78%]
tests/mxnet/test_nn.py::test_dense_sage_conv[2-g5-int32] PASSED          [ 78%]
tests/mxnet/test_nn.py::test_dense_sage_conv[2-g5-int64] PASSED          [ 78%]
tests/mxnet/test_nn.py::test_dense_sage_conv[2-g6-int32] PASSED          [ 78%]
tests/mxnet/test_nn.py::test_dense_sage_conv[2-g6-int64] PASSED          [ 79%]
tests/mxnet/test_nn.py::test_dense_sage_conv[2-g7-int32] PASSED          [ 79%]
tests/mxnet/test_nn.py::test_dense_sage_conv[2-g7-int64] PASSED          [ 79%]
tests/mxnet/test_nn.py::test_dense_sage_conv[2-g8-int32] PASSED          [ 79%]
tests/mxnet/test_nn.py::test_dense_sage_conv[2-g8-int64] PASSED          [ 79%]
tests/mxnet/test_nn.py::test_dense_sage_conv[2-g9-int32] PASSED          [ 79%]
tests/mxnet/test_nn.py::test_dense_sage_conv[2-g9-int64] PASSED          [ 79%]
tests/mxnet/test_nn.py::test_edge_conv[1-g0-int32] PASSED                [ 79%]
tests/mxnet/test_nn.py::test_edge_conv[1-g0-int64] PASSED                [ 80%]
tests/mxnet/test_nn.py::test_edge_conv[1-g1-int32] PASSED                [ 80%]
tests/mxnet/test_nn.py::test_edge_conv[1-g1-int64] PASSED                [ 80%]
tests/mxnet/test_nn.py::test_edge_conv[1-g2-int32] PASSED                [ 80%]
tests/mxnet/test_nn.py::test_edge_conv[1-g2-int64] PASSED                [ 80%]
tests/mxnet/test_nn.py::test_edge_conv[1-g3-int32] PASSED                [ 80%]
tests/mxnet/test_nn.py::test_edge_conv[1-g3-int64] PASSED                [ 80%]
tests/mxnet/test_nn.py::test_edge_conv[1-g4-int32] PASSED                [ 80%]
tests/mxnet/test_nn.py::test_edge_conv[1-g4-int64] PASSED                [ 81%]
tests/mxnet/test_nn.py::test_edge_conv[1-g5-int32] PASSED                [ 81%]
tests/mxnet/test_nn.py::test_edge_conv[1-g5-int64] PASSED                [ 81%]
tests/mxnet/test_nn.py::test_edge_conv[1-g6-int32] PASSED                [ 81%]
tests/mxnet/test_nn.py::test_edge_conv[1-g6-int64] PASSED                [ 81%]
tests/mxnet/test_nn.py::test_edge_conv[2-g0-int32] PASSED                [ 81%]
tests/mxnet/test_nn.py::test_edge_conv[2-g0-int64] PASSED                [ 81%]
tests/mxnet/test_nn.py::test_edge_conv[2-g1-int32] PASSED                [ 81%]
tests/mxnet/test_nn.py::test_edge_conv[2-g1-int64] PASSED                [ 82%]
tests/mxnet/test_nn.py::test_edge_conv[2-g2-int32] PASSED                [ 82%]
tests/mxnet/test_nn.py::test_edge_conv[2-g2-int64] PASSED                [ 82%]
tests/mxnet/test_nn.py::test_edge_conv[2-g3-int32] PASSED                [ 82%]
tests/mxnet/test_nn.py::test_edge_conv[2-g3-int64] FAILED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-u-shp2-g0] FAILED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-u-shp2-g1] FAILED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-u-shp3-g0] FAILED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-u-shp3-g1] PASSED                [ 82%]
tests/mxnet/test_nn.py::test_edge_conv[2-g4-int32] PASSED                [ 82%]
tests/mxnet/test_nn.py::test_edge_conv[2-g4-int64] PASSED                [ 82%]
tests/mxnet/test_nn.py::test_edge_conv[2-g5-int32] PASSED                [ 82%]
tests/mxnet/test_nn.py::test_edge_conv[2-g5-int64] PASSED                [ 83%]
tests/mxnet/test_nn.py::test_edge_conv[2-g6-int32] PASSED                [ 83%]
tests/mxnet/test_nn.py::test_edge_conv[2-g6-int64] PASSED                [ 83%]
tests/mxnet/test_nn.py::test_edge_conv_bi[1-g0-int32] PASSED             [ 83%]
tests/mxnet/test_nn.py::test_edge_conv_bi[1-g0-int64] PASSED             [ 83%]
tests/mxnet/test_nn.py::test_edge_conv_bi[1-g1-int32] PASSED             [ 83%]
tests/mxnet/test_nn.py::test_edge_conv_bi[1-g1-int64] PASSED             [ 83%]
tests/mxnet/test_nn.py::test_edge_conv_bi[2-g0-int32] PASSED             [ 83%]
tests/mxnet/test_nn.py::test_edge_conv_bi[2-g0-int64] PASSED             [ 83%]
tests/mxnet/test_nn.py::test_edge_conv_bi[2-g1-int32] PASSED             [ 84%]
tests/mxnet/test_nn.py::test_edge_conv_bi[2-g1-int64] PASSED             [ 84%]
tests/mxnet/test_nn.py::test_gin_conv[mean-g0-int32] PASSED              [ 84%]
tests/mxnet/test_nn.py::test_gin_conv[mean-g0-int64] PASSED              [ 84%]
tests/mxnet/test_nn.py::test_gin_conv[mean-g1-int32] PASSED              [ 84%]
tests/mxnet/test_nn.py::test_gin_conv[mean-g1-int64] PASSED              [ 84%]
tests/mxnet/test_nn.py::test_gin_conv[mean-g2-int32] PASSED              [ 84%]
tests/mxnet/test_nn.py::test_gin_conv[mean-g2-int64] PASSED              [ 84%]
tests/mxnet/test_nn.py::test_gin_conv[mean-g3-int32] PASSED              [ 85%]
tests/mxnet/test_nn.py::test_gin_conv[mean-g3-int64] PASSED              [ 85%]
tests/mxnet/test_nn.py::test_gin_conv[mean-g4-int32] PASSED              [ 85%]
tests/mxnet/test_nn.py::test_gin_conv[mean-g4-int64] PASSED              [ 85%]
tests/mxnet/test_nn.py::test_gin_conv[mean-g5-int32] PASSED              [ 85%]
tests/mxnet/test_nn.py::test_gin_conv[mean-g5-int64] PASSED              [ 85%]
tests/mxnet/test_nn.py::test_gin_conv[mean-g6-int32] PASSED              [ 85%]
tests/mxnet/test_nn.py::test_gin_conv[mean-g6-int64] PASSED              [ 85%]
tests/mxnet/test_nn.py::test_gin_conv[mean-g7-int32] PASSED              [ 86%]
tests/mxnet/test_nn.py::test_gin_conv[mean-g7-int64] PASSED              [ 86%]
tests/mxnet/test_nn.py::test_gin_conv[max-g0-int32] PASSED               [ 86%]
tests/mxnet/test_nn.py::test_gin_conv[max-g0-int64] PASSED               [ 86%]
tests/mxnet/test_nn.py::test_gin_conv[max-g1-int32] PASSED               [ 86%]
tests/mxnet/test_nn.py::test_gin_conv[max-g1-int64] PASSED               [ 86%]
tests/mxnet/test_nn.py::test_gin_conv[max-g2-int32] PASSED               [ 86%]
tests/mxnet/test_nn.py::test_gin_conv[max-g2-int64] PASSED               [ 86%]
tests/mxnet/test_nn.py::test_gin_conv[max-g3-int32] PASSED               [ 87%]
tests/mxnet/test_nn.py::test_gin_conv[max-g3-int64] PASSED               [ 87%]
tests/mxnet/test_nn.py::test_gin_conv[max-g4-int32] PASSED               [ 87%]
tests/mxnet/test_nn.py::test_gin_conv[max-g4-int64] PASSED               [ 87%]
tests/mxnet/test_nn.py::test_gin_conv[max-g5-int32] PASSED               [ 87%]
tests/mxnet/test_nn.py::test_gin_conv[max-g5-int64] PASSED               [ 87%]
tests/mxnet/test_nn.py::test_gin_conv[max-g6-int32] PASSED               [ 87%]
tests/mxnet/test_nn.py::test_gin_conv[max-g6-int64] PASSED               [ 87%]
tests/mxnet/test_nn.py::test_gin_conv[max-g7-int32] PASSED               [ 88%]
tests/mxnet/test_nn.py::test_gin_conv[max-g7-int64] PASSED               [ 88%]
tests/mxnet/test_nn.py::test_gin_conv[sum-g0-int32] PASSED               [ 88%]
tests/mxnet/test_nn.py::test_gin_conv[sum-g0-int64] PASSED               [ 88%]
tests/mxnet/test_nn.py::test_gin_conv[sum-g1-int32] PASSED               [ 88%]
tests/mxnet/test_nn.py::test_gin_conv[sum-g1-int64] PASSED               [ 88%]
tests/mxnet/test_nn.py::test_gin_conv[sum-g2-int32] PASSED               [ 88%]
tests/mxnet/test_nn.py::test_gin_conv[sum-g2-int64] PASSED               [ 88%]
tests/mxnet/test_nn.py::test_gin_conv[sum-g3-int32] PASSED               [ 89%]
tests/mxnet/test_nn.py::test_gin_conv[sum-g3-int64] PASSED               [ 89%]
tests/mxnet/test_nn.py::test_gin_conv[sum-g4-int32] PASSED               [ 89%]
tests/mxnet/test_nn.py::test_gin_conv[sum-g4-int64] PASSED               [ 89%]
tests/mxnet/test_nn.py::test_gin_conv[sum-g5-int32] PASSED               [ 89%]
tests/mxnet/test_nn.py::test_gin_conv[sum-g5-int64] PASSED               [ 89%]
tests/mxnet/test_nn.py::test_gin_conv[sum-g6-int32] PASSED               [ 89%]
tests/mxnet/test_nn.py::test_gin_conv[sum-g6-int64] PASSED               [ 89%]
tests/mxnet/test_nn.py::test_gin_conv[sum-g7-int32] PASSED               [ 90%]
tests/mxnet/test_nn.py::test_gin_conv[sum-g7-int64] PASSED               [ 90%]
tests/mxnet/test_nn.py::test_gin_conv_bi[mean-g0-int32] PASSED           [ 90%]
tests/mxnet/test_nn.py::test_gin_conv_bi[mean-g0-int64] PASSED           [ 90%]
tests/mxnet/test_nn.py::test_gin_conv_bi[mean-g1-int32] PASSED           [ 90%]
tests/mxnet/test_nn.py::test_gin_conv_bi[mean-g1-int64] PASSED           [ 90%]
tests/mxnet/test_nn.py::test_gin_conv_bi[mean-g2-int32] PASSED           [ 90%]FAILED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-u-shp4-g0] FAILED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-u-shp4-g1] FAILED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-v-shp0-g0] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-v-shp0-g1] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-v-shp1-g0] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-v-shp1-g1] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-v-shp2-g0] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-v-shp2-g1] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-v-shp3-g0] PASSED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-v-shp3-g1] PASSED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-v-shp4-g0] PASSED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-v-shp4-g1] PASSED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-e-shp0-g0] FAILED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-e-shp0-g1] FAILED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-e-shp1-g0] Client [1090] waits on 172.17.0.3:60789
Machine (0) group (0) client (0) connect to server successfuly!
Client[0] in group[0] is exiting...
Server (0) shutdown.
Server is exiting...

tests/mxnet/test_nn.py::test_gin_conv_bi[mean-g2-int64] PASSED           [ 90%]
tests/mxnet/test_nn.py::test_gin_conv_bi[max-g0-int32] PASSED            [ 91%]
tests/mxnet/test_nn.py::test_gin_conv_bi[max-g0-int64] PASSED            [ 91%]
tests/mxnet/test_nn.py::test_gin_conv_bi[max-g1-int32] PASSED            [ 91%]
tests/mxnet/test_nn.py::test_gin_conv_bi[max-g1-int64] PASSED            [ 91%]
tests/mxnet/test_nn.py::test_gin_conv_bi[max-g2-int32] PASSED            [ 91%]
tests/mxnet/test_nn.py::test_gin_conv_bi[max-g2-int64] PASSED            [ 91%]
tests/mxnet/test_nn.py::test_gin_conv_bi[sum-g0-int32] PASSED            [ 91%]
tests/mxnet/test_nn.py::test_gin_conv_bi[sum-g0-int64] PASSED            [ 91%]
tests/mxnet/test_nn.py::test_gin_conv_bi[sum-g1-int32] PASSED            [ 92%]
tests/mxnet/test_nn.py::test_gin_conv_bi[sum-g1-int64] PASSED            [ 92%]
tests/mxnet/test_nn.py::test_gin_conv_bi[sum-g2-int32] PASSED            [ 92%]
tests/mxnet/test_nn.py::test_gin_conv_bi[sum-g2-int64] PASSED            [ 92%]
tests/mxnet/test_nn.py::test_gmm_conv[g0-int32] PASSED                   [ 92%]
tests/mxnet/test_nn.py::test_gmm_conv[g0-int64] PASSED                   [ 92%]
tests/mxnet/test_nn.py::test_gmm_conv[g1-int32] PASSED                   [ 92%]
tests/mxnet/test_nn.py::test_gmm_conv[g1-int64] PASSED                   [ 92%]
tests/mxnet/test_nn.py::test_gmm_conv[g2-int32] PASSED                   [ 93%]
tests/mxnet/test_nn.py::test_gmm_conv[g2-int64] PASSED                   [ 93%]
tests/mxnet/test_nn.py::test_gmm_conv[g3-int32] PASSED                   [ 93%]
tests/mxnet/test_nn.py::test_gmm_conv[g3-int64] PASSED                   [ 93%]
tests/mxnet/test_nn.py::test_gmm_conv[g4-int32] PASSED                   [ 93%]
tests/mxnet/test_nn.py::test_gmm_conv[g4-int64] PASSED                   [ 93%]
tests/mxnet/test_nn.py::test_gmm_conv[g5-int32] PASSED                   [ 93%]
tests/mxnet/test_nn.py::test_gmm_conv[g5-int64] PASSED                   [ 93%]
tests/mxnet/test_nn.py::test_gmm_conv[g6-int32] PASSED                   [ 94%]
tests/mxnet/test_nn.py::test_gmm_conv[g6-int64] PASSED                   [ 94%]
tests/mxnet/test_nn.py::test_gmm_conv_bi[g0-int32] PASSED                [ 94%]
tests/mxnet/test_nn.py::test_gmm_conv_bi[g0-int64] PASSED                [ 94%]
tests/mxnet/test_nn.py::test_gmm_conv_bi[g1-int32] PASSED                [ 94%]
tests/mxnet/test_nn.py::test_gmm_conv_bi[g1-int64] PASSED                [ 94%]
tests/mxnet/test_nn.py::test_nn_conv[g0-int32] PASSED                    [ 94%]
tests/mxnet/test_nn.py::test_nn_conv[g0-int64] PASSED                    [ 94%]
tests/mxnet/test_nn.py::test_nn_conv[g1-int32] PASSED                    [ 95%]
tests/mxnet/test_nn.py::test_nn_conv[g1-int64] PASSED                    [ 95%]
tests/mxnet/test_nn.py::test_nn_conv[g2-int32] PASSED                    [ 95%]
tests/mxnet/test_nn.py::test_nn_conv[g2-int64] PASSED                    [ 95%]
tests/mxnet/test_nn.py::test_nn_conv[g3-int32] PASSED                    [ 95%]
tests/mxnet/test_nn.py::test_nn_conv[g3-int64] PASSED                    [ 95%]
tests/mxnet/test_nn.py::test_nn_conv[g4-int32] PASSED                    [ 95%]
tests/mxnet/test_nn.py::test_nn_conv[g4-int64] PASSED                    [ 95%]
tests/mxnet/test_nn.py::test_nn_conv[g5-int32] PASSED                    [ 96%]
tests/mxnet/test_nn.py::test_nn_conv[g5-int64] PASSED                    [ 96%]
tests/mxnet/test_nn.py::test_nn_conv[g6-int32] PASSED                    [ 96%]
tests/mxnet/test_nn.py::test_nn_conv[g6-int64] PASSED                    [ 96%]
tests/mxnet/test_nn.py::test_nn_conv[g7-int32] PASSED                    [ 96%]
tests/mxnet/test_nn.py::test_nn_conv[g7-int64] PASSED                    [ 96%]
tests/mxnet/test_nn.py::test_nn_conv_bi[g0-int32] PASSED                 [ 96%]
tests/mxnet/test_nn.py::test_nn_conv_bi[g0-int64] PASSED                 [ 96%]
tests/mxnet/test_nn.py::test_nn_conv_bi[g1-int32] PASSED                 [ 97%]
tests/mxnet/test_nn.py::test_nn_conv_bi[g1-int64] PASSED                 [ 97%]
tests/mxnet/test_nn.py::test_nn_conv_bi[g2-int32] PASSED                 [ 97%]
tests/mxnet/test_nn.py::test_nn_conv_bi[g2-int64] PASSED                 [ 97%]
tests/mxnet/test_nn.py::test_sg_conv[1] PASSED                           [ 97%]
tests/mxnet/test_nn.py::test_sg_conv[2] PASSED                           [ 97%]
tests/mxnet/test_nn.py::test_set2set FAILED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-e-shp1-g1] FAILED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-e-shp2-g0] FAILED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-e-shp2-g1] FAILED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-e-shp3-g0] PASSED                              [ 97%]
tests/mxnet/test_nn.py::test_glob_att_pool PASSED                        [ 97%]
tests/mxnet/test_nn.py::test_simple_pool PASSED                          [ 98%]
tests/mxnet/test_nn.py::test_rgcn[1] PASSED                              [ 98%]
tests/mxnet/test_nn.py::test_rgcn[2] PASSED                              [ 98%]
tests/mxnet/test_nn.py::test_rgcn[8] PASSED                              [ 98%]
tests/mxnet/test_nn.py::test_sequential PASSED                           [ 98%]
tests/mxnet/test_nn.py::test_hetero_conv[sum-int32] PASSED               [ 98%]
tests/mxnet/test_nn.py::test_hetero_conv[sum-int64] PASSED               [ 98%]
tests/mxnet/test_nn.py::test_hetero_conv[max-int32] PASSED               [ 98%]
tests/mxnet/test_nn.py::test_hetero_conv[max-int64] FAILED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-e-shp3-g1] FAILED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-e-shp4-g0] FAILED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-e-shp4-g1] FAILED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-u-shp0-g0] FAILED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-u-shp0-g1] Server (1) shutdown.
Server is exiting...
Done sampling
PASSED               [ 99%]
tests/mxnet/test_nn.py::test_hetero_conv[min-int32] PASSED               [ 99%]
tests/mxnet/test_nn.py::test_hetero_conv[min-int64] PASSED               [ 99%]
tests/mxnet/test_nn.py::test_hetero_conv[mean-int32] PASSED              [ 99%]
tests/mxnet/test_nn.py::test_hetero_conv[mean-int64] PASSED              [ 99%]
tests/mxnet/test_nn.py::test_hetero_conv[stack-int32] FAILED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-u-shp1-g0] FAILED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-u-shp1-g1] FAILED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-u-shp2-g0] FAILED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-u-shp2-g1] Converting to homogeneous graph takes 0.001s, peak mem: 1.529 GB
Convert a graph into a bidirected graph: 0.000 seconds, peak memory: 1.529 GB
Construct multi-constraint weights: 0.000 seconds, peak memory: 1.529 GB
[05:32:59] /root/jenkins/workspace/dgl_PR-4648/src/graph/transform/metis_partition_hetero.cc:87: Partition a graph with 1500 nodes and 2000 edges into 2 parts and get 31 edge cuts
Metis partitioning: 0.001 seconds, peak memory: 1.529 GB
Assigning nodes to METIS partitions takes 0.002s, peak mem: 1.529 GB
Reshuffle nodes and edges: 0.001 seconds
Split the graph: 0.001 seconds
Construct subgraphs: 0.002 seconds
Splitting the graph into partitions takes 0.004s, peak mem: 1.529 GB
part 0 has 501 nodes of type game and 485 are inside the partition
part 0 has 257 nodes of type user and 243 are inside the partition
part 0 has 476 edges of type buys and 460 are inside the partition
part 1 has 529 nodes of type game and 515 are inside the partition
part 1 has 273 nodes of type user and 257 are inside the partition
part 1 has 555 edges of type buys and 540 are inside the partition
Save partitions: 0.002 seconds, peak memory: 1.529 GB
There are 1000 edges in the graph and 0 edge cuts for 2 partitions.
PASSED             [ 99%]
tests/mxnet/test_nn.py::test_hetero_conv[stack-int64] PASSED             [ 99%]
tests/mxnet/test_nn.py::test_hetero_conv[myagg-int32] PASSED             [ 99%]
tests/mxnet/test_nn.py::test_hetero_conv[myagg-int64] PASSED             [100%]

=============================== warnings summary ===============================
../../../../opt/conda/envs/mxnet-ci/lib/python3.7/site-packages/mxnet/numpy/utils.py:37
  /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages/mxnet/numpy/utils.py:37: DeprecationWarning: `np.bool` is a deprecated alias for the builtin `bool`. To silence this warning, use `bool` by itself. Doing this will not modify any behavior and is safe. If you specifically wanted the numpy scalar type, use `np.bool_` here.
  Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations
    bool = onp.bool

../../../../opt/conda/envs/mxnet-ci/lib/python3.7/site-packages/mxnet/numpy/fallback.py:143
  /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages/mxnet/numpy/fallback.py:143: DeprecationWarning: In accordance with NEP 32, the function mirr was removed from NumPy version 1.20.  A replacement for this function is available in the numpy_financial library: https://pypi.org/project/numpy-financial
    mirr = onp.mirr

../../../../opt/conda/envs/mxnet-ci/lib/python3.7/site-packages/mxnet/numpy/fallback.py:160
  /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages/mxnet/numpy/fallback.py:160: DeprecationWarning: In accordance with NEP 32, the function npv was removed from NumPy version 1.20.  A replacement for this function is available in the numpy_financial library: https://pypi.org/project/numpy-financial
    npv = onp.npv

../../../../opt/conda/envs/mxnet-ci/lib/python3.7/site-packages/mxnet/numpy/fallback.py:164
  /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages/mxnet/numpy/fallback.py:164: DeprecationWarning: In accordance with NEP 32, the function pmt was removed from NumPy version 1.20.  A replacement for this function is available in the numpy_financial library: https://pypi.org/project/numpy-financial
    pmt = onp.pmt

../../../../opt/conda/envs/mxnet-ci/lib/python3.7/site-packages/mxnet/numpy/fallback.py:173
  /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages/mxnet/numpy/fallback.py:173: DeprecationWarning: In accordance with NEP 32, the function ppmt was removed from NumPy version 1.20.  A replacement for this function is available in the numpy_financial library: https://pypi.org/project/numpy-financial
    ppmt = onp.ppmt

../../../../opt/conda/envs/mxnet-ci/lib/python3.7/site-packages/mxnet/numpy/fallback.py:176
  /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages/mxnet/numpy/fallback.py:176: DeprecationWarning: In accordance with NEP 32, the function pv was removed from NumPy version 1.20.  A replacement for this function is available in the numpy_financial library: https://pypi.org/project/numpy-financial
    pv = onp.pv

../../../../opt/conda/envs/mxnet-ci/lib/python3.7/site-packages/mxnet/numpy/fallback.py:177
  /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages/mxnet/numpy/fallback.py:177: DeprecationWarning: In accordance with NEP 32, the function rate was removed from NumPy version 1.20.  A replacement for this function is available in the numpy_financial library: https://pypi.org/project/numpy-financial
    rate = onp.rate

../../../../opt/conda/envs/mxnet-ci/lib/python3.7/site-packages/mxnet/numpy_dispatch_protocol.py:48
  /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages/mxnet/numpy_dispatch_protocol.py:48: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    cur_np_ver = LooseVersion(_np.__version__)

../../../../opt/conda/envs/mxnet-ci/lib/python3.7/site-packages/mxnet/numpy_dispatch_protocol.py:49
  /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages/mxnet/numpy_dispatch_protocol.py:49: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    np_1_17_ver = LooseVersion('1.17')

../../../../opt/conda/envs/mxnet-ci/lib/python3.7/site-packages/mxnet/numpy_dispatch_protocol.py:68
  /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages/mxnet/numpy_dispatch_protocol.py:68: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    cur_np_ver = LooseVersion(_np.__version__)

../../../../opt/conda/envs/mxnet-ci/lib/python3.7/site-packages/mxnet/numpy_dispatch_protocol.py:69
  /opt/conda/envs/mxnet-ci/lib/python3.7/site-packages/mxnet/numpy_dispatch_protocol.py:69: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    np_1_15_ver = LooseVersion('1.15')

python/dgl/backend/mxnet/tensor.py:15
python/dgl/backend/mxnet/tensor.py:15
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/backend/mxnet/tensor.py:15: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    if LooseVersion(mx.__version__) < LooseVersion("1.6.0"):

python/dgl/backend/mxnet/tensor.py:31
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/backend/mxnet/tensor.py:31: DeprecationWarning: `np.bool` is a deprecated alias for the builtin `bool`. To silence this warning, use `bool` by itself. Doing this will not modify any behavior and is safe. If you specifically wanted the numpy scalar type, use `np.bool_` here.
  Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations
    'bool'    : np.bool}  # mxnet does not support bool

python/dgl/backend/mxnet/tensor.py:37: 290 warnings
tests/mxnet/test_nn.py: 388 warnings
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/backend/mxnet/tensor.py:37: DeprecationWarning: `np.bool` is a deprecated alias for the builtin `bool`. To silence this warning, use `bool` by itself. Doing this will not modify any behavior and is safe. If you specifically wanted the numpy scalar type, use `np.bool_` here.
  Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations
    if dtype == np.bool:

python/dgl/backend/mxnet/tensor.py:136: 230 warnings
tests/mxnet/test_nn.py: 464 warnings
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/backend/mxnet/tensor.py:136: DeprecationWarning: `np.bool` is a deprecated alias for the builtin `bool`. To silence this warning, use `bool` by itself. Doing this will not modify any behavior and is safe. If you specifically wanted the numpy scalar type, use `np.bool_` here.
  Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations
    if ty == np.bool:

tests/mxnet/test_nn.py::test_cheb_conv[1]
tests/mxnet/test_nn.py::test_cheb_conv[20]
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/nn/mxnet/conv/chebconv.py:121: DGLWarning: lambda_max is not provided, using default value of 2.  Please use dgl.laplacian_lambda_max to compute the eigenvalues.
    "lambda_max is not provided, using default value of 2.  "

-- Docs: https://docs.pytest.org/en/stable/how-to/capture-warnings.html
-- generated xml file: /root/jenkins/workspace/dgl_PR-4648/pytest_backend.xml --
============================ slowest 100 durations =============================
0.33s call     tests/mxnet/test_nn.py::test_graph_conv[1-int32]
0.32s call     tests/mxnet/test_geometry.py::test_fps
0.06s call     tests/mxnet/test_nn.py::test_gat_conv[5-20-g5-int64]
0.05s call     tests/mxnet/test_nn.py::test_hetero_conv[sum-int32]
0.04s call     tests/mxnet/test_nn.py::test_hetero_conv[min-int64]
0.04s call     tests/mxnet/test_nn.py::test_hetero_conv[mean-int32]
0.04s call     tests/mxnet/test_nn.py::test_hetero_conv[stack-int32]
0.04s call     tests/mxnet/test_nn.py::test_hetero_conv[max-int32]
0.04s call     tests/mxnet/test_nn.py::test_hetero_conv[min-int32]
0.04s call     tests/mxnet/test_nn.py::test_hetero_conv[sum-int64]
0.04s call     tests/mxnet/test_nn.py::test_hetero_conv[mean-int64]
0.04s call     tests/mxnet/test_nn.py::test_hetero_conv[max-int64]
0.04s call     tests/mxnet/test_nn.py::test_hetero_conv[myagg-int32]
0.04s call     tests/mxnet/test_nn.py::test_hetero_conv[myagg-int64]
0.04s call     tests/mxnet/test_nn.py::test_hetero_conv[stack-int64]
0.03s call     tests/mxnet/test_nn.py::test_set2set
0.03s call     tests/mxnet/test_nn.py::test_dense_cheb_conv[1]
0.03s call     tests/mxnet/test_nn.py::test_dense_cheb_conv[2]
0.03s call     tests/mxnet/test_nn.py::test_gg_conv
0.02s call     tests/mxnet/test_nn.py::test_graph_conv[1-int64]
0.02s call     tests/mxnet/test_nn.py::test_graph_conv[2-int64]
0.02s call     tests/mxnet/test_nn.py::test_graph_conv[2-int32]
0.02s call     tests/mxnet/test_nn.py::test_rgcn[8]
0.02s call     tests/mxnet/test_nn.py::test_rgcn[2]
0.01s call     tests/mxnet/test_nn.py::test_simple_pool
0.01s call     tests/mxnet/test_nn.py::test_gat_conv[1-20-g6-int64]
0.01s call     tests/mxnet/test_nn.py::test_gat_conv[5-20-g5-int32]
0.01s call     tests/mxnet/test_nn.py::test_gat_conv[1-1-g5-int32]
0.01s call     tests/mxnet/test_nn.py::test_gat_conv[1-1-g0-int32]
0.01s call     tests/mxnet/test_nn.py::test_gat_conv[5-20-g1-int64]
0.01s call     tests/mxnet/test_nn.py::test_gat_conv[5-20-g6-int32]
0.01s call     tests/mxnet/test_nn.py::test_rgcn[1]
0.01s call     tests/mxnet/test_nn.py::test_gat_conv[5-1-g6-int64]
0.01s call     tests/mxnet/test_nn.py::test_gat_conv[5-1-g1-int64]
0.01s call     tests/mxnet/test_nn.py::test_gat_conv[5-20-g6-int64]
0.01s call     tests/mxnet/test_nn.py::test_gat_conv[1-20-g5-int32]
0.01s call     tests/mxnet/test_nn.py::test_gat_conv[1-1-g6-int64]
0.01s call     tests/mxnet/test_nn.py::test_gat_conv[1-1-g2-int32]
0.01s call     tests/mxnet/test_nn.py::test_gat_conv[1-20-g4-int32]
0.01s call     tests/mxnet/test_nn.py::test_gat_conv[1-20-g6-int32]
0.01s call     tests/mxnet/test_nn.py::test_gat_conv[1-1-g1-int32]
0.01s call     tests/mxnet/test_nn.py::test_gat_conv[5-20-g4-int32]
0.01s call     tests/mxnet/test_nn.py::test_gat_conv[5-1-g5-int32]
0.01s call     tests/mxnet/test_nn.py::test_gat_conv[1-20-g0-int64]
0.01s call     tests/mxnet/test_nn.py::test_gat_conv[1-1-g4-int32]
0.01s call     tests/mxnet/test_nn.py::test_gat_conv[5-20-g0-int64]
0.01s call     tests/mxnet/test_nn.py::test_gat_conv[1-1-g3-int32]
0.01s call     tests/mxnet/test_nn.py::test_gat_conv[5-1-g0-int64]
0.01s call     tests/mxnet/test_nn.py::test_gat_conv[1-20-g3-int32]
0.01s call     tests/mxnet/test_nn.py::test_gat_conv[1-20-g2-int32]
0.01s call     tests/mxnet/test_nn.py::test_gat_conv[1-1-g6-int32]
0.01s call     tests/mxnet/test_nn.py::test_gat_conv[5-20-g3-int32]
0.01s call     tests/mxnet/test_nn.py::test_gat_conv[1-20-g1-int64]
0.01s call     tests/mxnet/test_nn.py::test_gat_conv[5-1-g4-int32]
0.01s call     tests/mxnet/test_nn.py::test_gat_conv[5-1-g6-int32]
0.01s call     tests/mxnet/test_nn.py::test_gat_conv[1-1-g5-int64]
0.01s call     tests/mxnet/test_nn.py::test_gat_conv[5-20-g2-int32]
0.01s call     tests/mxnet/test_nn.py::test_gat_conv[1-20-g5-int64]
0.01s call     tests/mxnet/test_nn.py::test_gat_conv[5-1-g3-int32]
0.01s call     tests/mxnet/test_nn.py::test_gat_conv[5-1-g5-int64]
0.01s call     tests/mxnet/test_nn.py::test_gat_conv[5-1-g2-int32]
0.01s call     tests/mxnet/test_nn.py::test_gat_conv[1-20-g2-int64]
0.01s call     tests/mxnet/test_nn.py::test_gat_conv[5-20-g2-int64]
0.01s call     tests/mxnet/test_nn.py::test_gat_conv[1-20-g4-int64]
0.01s call     tests/mxnet/test_nn.py::test_gat_conv[5-20-g4-int64]
0.01s call     tests/mxnet/test_nn.py::test_gat_conv[1-1-g2-int64]
0.01s call     tests/mxnet/test_nn.py::test_gat_conv[5-1-g0-int32]
0.01s call     tests/mxnet/test_nn.py::test_gat_conv[5-1-g2-int64]
0.01s call     tests/mxnet/test_nn.py::test_gat_conv[1-20-g0-int32]
0.01s call     tests/mxnet/test_nn.py::test_gat_conv[1-1-g4-int64]
0.01s call     tests/mxnet/test_nn.py::test_gat_conv[1-20-g1-int32]
0.01s call     tests/mxnet/test_nn.py::test_gat_conv[5-20-g3-int64]
0.01s call     tests/mxnet/test_nn.py::test_sage_conv_bi2[2-mean-int64]
0.01s call     tests/mxnet/test_nn.py::test_gat_conv[1-1-g0-int64]
0.01s call     tests/mxnet/test_nn.py::test_gat_conv[5-20-g0-int32]
0.01s call     tests/mxnet/test_nn.py::test_gat_conv[1-20-g3-int64]
0.01s call     tests/mxnet/test_nn.py::test_sage_conv_bi2[1-mean-int64]
0.01s call     tests/mxnet/test_nn.py::test_gat_conv[5-1-g1-int32]
0.01s call     tests/mxnet/test_nn.py::test_gat_conv[5-20-g1-int32]
0.01s call     tests/mxnet/test_nn.py::test_gat_conv[1-1-g1-int64]
0.01s call     tests/mxnet/test_nn.py::test_gat_conv[1-1-g3-int64]
0.01s call     tests/mxnet/test_nn.py::test_sage_conv_bi2[2-mean-int32]
0.01s call     tests/mxnet/test_nn.py::test_gat_conv[5-1-g4-int64]
0.01s call     tests/mxnet/test_nn.py::test_gat_conv[5-1-g3-int64]
0.01s call     tests/mxnet/test_nn.py::test_sage_conv_bi2[1-gcn-int64]
0.01s call     tests/mxnet/test_nn.py::test_sequential
0.01s call     tests/mxnet/test_nn.py::test_sage_conv_bi2[1-gcn-int32]
0.01s call     tests/mxnet/test_nn.py::test_sage_conv_bi2[1-pool-int32]
0.01s call     tests/mxnet/test_nn.py::test_sage_conv_bi2[1-mean-int32]
0.01s call     tests/mxnet/test_nn.py::test_sage_conv_bi2[2-gcn-int64]
0.01s call     tests/mxnet/test_nn.py::test_sage_conv_bi2[1-pool-int64]
0.01s call     tests/mxnet/test_nn.py::test_sage_conv_bi2[2-gcn-int32]
0.01s call     tests/mxnet/test_nn.py::test_sage_conv_bi2[2-pool-int64]
0.01s call     tests/mxnet/test_nn.py::test_sage_conv_bi2[2-pool-int32]
0.01s call     tests/mxnet/test_nn.py::test_tagconv[1]
0.01s call     tests/mxnet/test_nn.py::test_tagconv[2]
0.01s call     tests/mxnet/test_nn.py::test_glob_att_pool
0.01s call     tests/mxnet/test_nn.py::test_gat_conv_bi[1-2-g1-int32]
0.01s call     tests/mxnet/test_nn.py::test_gat_conv_bi[1-1-g1-int32]
0.01s call     tests/mxnet/test_nn.py::test_gat_conv_bi[4-1-g1-int32]
===================== 806 passed, 1388 warnings in 11.78s ======================
FAILED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-u-shp3-g0] FAILED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-u-shp3-g1] FAILED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-u-shp4-g0] FAILED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-u-shp4-g1] FAILED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-v-shp0-g0] Error in sys.excepthook:

Original exception was:
Error in sys.excepthook:

Original exception was:
Error in sys.excepthook:

Original exception was:
Error in sys.excepthook:

Original exception was:
Error in sys.excepthook:

Original exception was:
Error in sys.excepthook:

Original exception was:
Error in sys.excepthook:

Original exception was:
Error in sys.excepthook:

Original exception was:
Error in sys.excepthook:

Original exception was:
Error in sys.excepthook:

Original exception was:
Error in sys.excepthook:

Original exception was:
Error in sys.excepthook:

Original exception was:
Error in sys.excepthook:

Original exception was:
Error in sys.excepthook:

Original exception was:
Error in sys.excepthook:

Original exception was:
Error in sys.excepthook:

Original exception was:
Error in sys.excepthook:

Original exception was:
Error in sys.excepthook:

Original exception was:
Error in sys.excepthook:

Original exception was:
Error in sys.excepthook:

Original exception was:
Error in sys.excepthook:

Original exception was:
Error in sys.excepthook:

Original exception was:
Error in sys.excepthook:

Original exception was:
Error in sys.excepthook:

Original exception was:
Error in sys.excepthook:

Original exception was:
Error in sys.excepthook:

Original exception was:
Error in sys.excepthook:

Original exception was:
Error in sys.excepthook:

Original exception was:
Error in sys.excepthook:

Original exception was:
Error in sys.excepthook:

Original exception was:
Error in sys.excepthook:

Original exception was:
Error in sys.excepthook:

Original exception was:
Error in sys.excepthook:

Original exception was:
Error in sys.excepthook:

Original exception was:
Error in sys.excepthook:

Original exception was:
Error in sys.excepthook:

Original exception was:
Error in sys.excepthook:

Original exception was:
Error in sys.excepthook:

Original exception was:
Error in sys.excepthook:

Original exception was:
Error in sys.excepthook:

Original exception was:
Error in sys.excepthook:

Original exception was:
Error in sys.excepthook:

Original exception was:
Error in sys.excepthook:

Original exception was:
Error in sys.excepthook:

Original exception was:
Error in sys.excepthook:

Original exception was:
Error in sys.excepthook:

Original exception was:
FAILED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-v-shp0-g1] FAILED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-v-shp1-g0] FAILED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-v-shp1-g1] FAILED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-v-shp2-g0] FAILED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-v-shp2-g1] FAILED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-v-shp3-g0] FAILED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-v-shp3-g1] FAILED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-v-shp4-g0] FAILED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-v-shp4-g1] [Pipeline] }
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_sampling
start graph service on server 0 for part 0
[05:33:00] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:33:00] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:10976]...
[Pipeline] // timeout
FAILED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-e-shp0-g0] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-e-shp0-g1] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-e-shp1-g0] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-e-shp1-g1] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-e-shp2-g0] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-e-shp2-g1] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-e-shp3-g0] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-e-shp3-g1] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-e-shp4-g0] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-e-shp4-g1] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-u-shp0-g0] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-u-shp0-g1] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-u-shp1-g0] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-u-shp1-g1] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-u-shp2-g0] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-u-shp2-g1] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-u-shp3-g0] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-u-shp3-g1] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-u-shp4-g0] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-u-shp4-g1] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-v-shp0-g0] FAILED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-v-shp0-g1] FAILED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-v-shp1-g0] FAILED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-v-shp1-g1] [Pipeline] }
[Pipeline] // stage
Post stage
[Pipeline] cleanWs
[WS-CLEANUP] Deleting project workspace...
[WS-CLEANUP] Deferred wipeout is disabled by the job configuration...
FAILED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-v-shp2-g0] FAILED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-v-shp2-g1] FAILED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-v-shp3-g0] FAILED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-v-shp3-g1] FAILED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-v-shp4-g0] FAILED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-v-shp4-g1] FAILED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-e-shp0-g0] FAILED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-e-shp0-g1] FAILED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-e-shp1-g0] FAILED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-e-shp1-g1] [WS-CLEANUP] done
[Pipeline] }
$ docker stop --time=1 93431508da8c7bd1ffa88cc07c573b36fd6622338dac6808937ce4f0b46901f0
FAILED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-e-shp2-g0] FAILED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-e-shp2-g1] FAILED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-e-shp3-g0] FAILED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-e-shp3-g1] /root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_sampling
start graph service on server 1 for part 1
[05:33:01] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:33:01] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:10978]...
[05:33:01] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:33:01] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
FAILED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-e-shp4-g0] FAILED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-e-shp4-g1] FAILED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-u-shp0-g0] FAILED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-u-shp0-g1] FAILED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-u-shp1-g0] FAILED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-u-shp1-g1] FAILED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-u-shp2-g0] FAILED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-u-shp2-g1] FAILED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-u-shp3-g0] FAILED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-u-shp3-g1] FAILED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-u-shp4-g0] FAILED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-u-shp4-g1] FAILED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-v-shp0-g0] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-v-shp0-g1] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-v-shp1-g0] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-v-shp1-g1] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-v-shp2-g0] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-v-shp2-g1] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-v-shp3-g0] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-v-shp3-g1] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-v-shp4-g0] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-v-shp4-g1] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-e-shp0-g0] FAILED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-e-shp0-g1] FAILED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-e-shp1-g0] FAILED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-e-shp1-g1] FAILED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-e-shp2-g0] FAILED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-e-shp2-g1] FAILED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-e-shp3-g0] FAILED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-e-shp3-g1] FAILED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-e-shp4-g0] $ docker rm -f 93431508da8c7bd1ffa88cc07c573b36fd6622338dac6808937ce4f0b46901f0
FAILED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-e-shp4-g1] FAILED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-u-shp0-g0] FAILED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-u-shp0-g1] FAILED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-u-shp1-g0] FAILED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-u-shp1-g1] FAILED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-u-shp2-g0] FAILED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-u-shp2-g1] FAILED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-u-shp3-g0] FAILED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-u-shp3-g1] FAILED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-u-shp4-g0] [Pipeline] // withDockerContainer
FAILED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-u-shp4-g1] FAILED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-v-shp0-g0] FAILED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-v-shp0-g1] FAILED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-v-shp1-g0] FAILED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-v-shp1-g1] [Pipeline] }
[Pipeline] // withEnv
[Pipeline] }
[Pipeline] // node
[Pipeline] }
[Pipeline] // stage
[Pipeline] }
FAILED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-v-shp2-g0] FAILED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-v-shp2-g1] FAILED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-v-shp3-g0] FAILED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-v-shp3-g1] FAILED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-v-shp4-g0] FAILED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-v-shp4-g1] FAILED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-e-shp0-g0] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-e-shp0-g1] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-e-shp1-g0] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-e-shp1-g1] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-e-shp2-g0] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-e-shp2-g1] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-e-shp3-g0] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-e-shp3-g1] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-e-shp4-g0] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-e-shp4-g1] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-u-shp0-g0] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-u-shp0-g1] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-u-shp1-g0] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-u-shp1-g1] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-u-shp2-g0] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-u-shp2-g1] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-u-shp3-g0] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-u-shp3-g1] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-u-shp4-g0] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-u-shp4-g1] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-v-shp0-g0] FAILED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-v-shp0-g1] FAILED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-v-shp1-g0] FAILED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-v-shp1-g1] FAILED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-v-shp2-g0] FAILED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-v-shp2-g1] FAILED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-v-shp3-g0] FAILED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-v-shp3-g1] FAILED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-v-shp4-g0] FAILED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-v-shp4-g1] FAILED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-e-shp0-g0] FAILED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-e-shp0-g1] FAILED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-e-shp1-g0] FAILED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-e-shp1-g1] FAILED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-e-shp2-g0] Client [1090] waits on 172.17.0.3:50325
Machine (0) group (0) client (0) connect to server successfuly!
Client[0] in group[0] is exiting...
Server (1) shutdown.
Server is exiting...
Server (0) shutdown.
Server is exiting...
Done sampling
FAILED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-e-shp2-g1] FAILED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-e-shp3-g0] FAILED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-e-shp3-g1] FAILED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-e-shp4-g0] FAILED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-e-shp4-g1] Converting to homogeneous graph takes 0.001s, peak mem: 1.529 GB
Convert a graph into a bidirected graph: 0.000 seconds, peak memory: 1.529 GB
Construct multi-constraint weights: 0.000 seconds, peak memory: 1.529 GB
[05:33:04] /root/jenkins/workspace/dgl_PR-4648/src/graph/transform/metis_partition_hetero.cc:87: Partition a graph with 1500 nodes and 2000 edges into 2 parts and get 27 edge cuts
Metis partitioning: 0.001 seconds, peak memory: 1.529 GB
Assigning nodes to METIS partitions takes 0.002s, peak mem: 1.529 GB
Reshuffle nodes and edges: 0.001 seconds
Split the graph: 0.001 seconds
Construct subgraphs: 0.002 seconds
Splitting the graph into partitions takes 0.003s, peak mem: 1.529 GB
part 0 has 514 nodes of type game and 503 are inside the partition
part 0 has 273 nodes of type user and 257 are inside the partition
part 0 has 523 edges of type buys and 512 are inside the partition
part 1 has 513 nodes of type game and 497 are inside the partition
part 1 has 254 nodes of type user and 243 are inside the partition
part 1 has 504 edges of type buys and 488 are inside the partition
Save partitions: 0.002 seconds, peak memory: 1.529 GB
There are 1000 edges in the graph and 0 edge cuts for 2 partitions.
FAILED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-u-shp0-g0] FAILED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-u-shp0-g1] FAILED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-u-shp1-g0] FAILED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-u-shp1-g1] FAILED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-u-shp2-g0] FAILED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-u-shp2-g1] FAILED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-u-shp3-g0] FAILED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-u-shp3-g1] FAILED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-u-shp4-g0] FAILED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-u-shp4-g1] FAILED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-v-shp0-g0] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-v-shp0-g1] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-v-shp1-g0] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-v-shp1-g1] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-v-shp2-g0] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-v-shp2-g1] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-v-shp3-g0] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-v-shp3-g1] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-v-shp4-g0] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-v-shp4-g1] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-e-shp0-g0] FAILED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-e-shp0-g1] FAILED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-e-shp1-g0] /root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_sampling
start graph service on server 0 for part 0
[05:33:05] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:33:05] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:26559]...
FAILED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-e-shp1-g1] FAILED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-e-shp2-g0] FAILED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-e-shp2-g1] FAILED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-e-shp3-g0] FAILED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-e-shp3-g1] FAILED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-e-shp4-g0] FAILED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-e-shp4-g1] FAILED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-u-shp0-g0] FAILED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-u-shp0-g1] FAILED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-u-shp1-g0] FAILED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-u-shp1-g1] FAILED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-u-shp2-g0] FAILED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-u-shp2-g1] FAILED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-u-shp3-g0] FAILED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-u-shp3-g1] FAILED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-u-shp4-g0] FAILED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-u-shp4-g1] FAILED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-v-shp0-g0] FAILED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-v-shp0-g1] /root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_sampling
start graph service on server 1 for part 1
[05:33:06] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:33:06] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:26561]...
[05:33:06] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:33:06] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
FAILED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-v-shp1-g0] FAILED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-v-shp1-g1] FAILED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-v-shp2-g0] FAILED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-v-shp2-g1] FAILED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-v-shp3-g0] FAILED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-v-shp3-g1] FAILED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-v-shp4-g0] FAILED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-v-shp4-g1] FAILED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-e-shp0-g0] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-e-shp0-g1] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-e-shp1-g0] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-e-shp1-g1] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-e-shp2-g0] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-e-shp2-g1] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-e-shp3-g0] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-e-shp3-g1] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-e-shp4-g0] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-e-shp4-g1] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-u-shp0-g0] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-u-shp0-g1] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-u-shp1-g0] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-u-shp1-g1] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-u-shp2-g0] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-u-shp2-g1] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-u-shp3-g0] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-u-shp3-g1] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-u-shp4-g0] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-u-shp4-g1] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-v-shp0-g0] FAILED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-v-shp0-g1] FAILED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-v-shp1-g0] FAILED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-v-shp1-g1] FAILED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-v-shp2-g0] FAILED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-v-shp2-g1] FAILED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-v-shp3-g0] FAILED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-v-shp3-g1] FAILED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-v-shp4-g0] FAILED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-v-shp4-g1] FAILED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-e-shp0-g0] FAILED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-e-shp0-g1] FAILED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-e-shp1-g0] FAILED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-e-shp1-g1] FAILED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-e-shp2-g0] FAILED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-e-shp2-g1] FAILED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-e-shp3-g0] FAILED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-e-shp3-g1] FAILED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-e-shp4-g0] FAILED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-e-shp4-g1] FAILED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-u-shp0-g0] FAILED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-u-shp0-g1] FAILED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-u-shp1-g0] FAILED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-u-shp1-g1] FAILED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-u-shp2-g0] FAILED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-u-shp2-g1] FAILED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-u-shp3-g0] FAILED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-u-shp3-g1] FAILED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-u-shp4-g0] FAILED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-u-shp4-g1] FAILED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-v-shp0-g0] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-v-shp0-g1] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-v-shp1-g0] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-v-shp1-g1] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-v-shp2-g0] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-v-shp2-g1] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-v-shp3-g0] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-v-shp3-g1] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-v-shp4-g0] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-v-shp4-g1] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-e-shp0-g0] FAILED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-e-shp0-g1] FAILED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-e-shp1-g0] FAILED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-e-shp1-g1] FAILED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-e-shp2-g0] FAILED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-e-shp2-g1] FAILED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-e-shp3-g0] FAILED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-e-shp3-g1] FAILED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-e-shp4-g0] FAILED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-e-shp4-g1] FAILED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-u-shp0-g0] FAILED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-u-shp0-g1] FAILED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-u-shp1-g0] FAILED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-u-shp1-g1] FAILED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-u-shp2-g0] FAILED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-u-shp2-g1] FAILED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-u-shp3-g0] FAILED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-u-shp3-g1] FAILED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-u-shp4-g0] Client [1090] waits on 172.17.0.3:50565
Machine (0) group (0) client (0) connect to server successfuly!
Client[0] in group[0] is exiting...
FAILED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-u-shp4-g1] FAILED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-v-shp0-g0] FAILED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-v-shp0-g1] FAILED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-v-shp1-g0] FAILED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-v-shp1-g1] FAILED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-v-shp2-g0] FAILED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-v-shp2-g1] FAILED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-v-shp3-g0] FAILED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-v-shp3-g1] FAILED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-v-shp4-g0] FAILED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-v-shp4-g1] FAILED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-e-shp0-g0] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-e-shp0-g1] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-e-shp1-g0] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-e-shp1-g1] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-e-shp2-g0] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-e-shp2-g1] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-e-shp3-g0] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-e-shp3-g1] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-e-shp4-g0] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-e-shp4-g1] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-u-shp0-g0] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-u-shp0-g1] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-u-shp1-g0] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-u-shp1-g1] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-u-shp2-g0] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-u-shp2-g1] PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-u-shp3-g0] PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-u-shp3-g1] PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-u-shp4-g0] PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-u-shp4-g1] PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-v-shp0-g0] FAILED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-v-shp0-g1] FAILED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-v-shp1-g0] Done sampling
Server (1) shutdown.Server (0) shutdown.

Server is exiting...
Server is exiting...
FAILED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-v-shp1-g1] FAILED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-v-shp2-g0] FAILED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-v-shp2-g1] FAILED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-v-shp3-g0] PASSED [ 25%]
tests/distributed/test_distributed_sampling.py::test_standalone_sampling   NumNodes: 2708
  NumEdges: 10556
  NumFeats: 1433
  NumClasses: 7
  NumTrainingSamples: 140
  NumValidationSamples: 500
  NumTestSamples: 1000
Done loading data from cached files.
Converting to homogeneous graph takes 0.001s, peak mem: 1.529 GB
Calculate edge assignment: 0.000 seconds
Save partitions: 0.018 seconds, peak memory: 1.529 GB
There are 10556 edges in the graph and 0 edge cuts for 1 partitions.
Client[-1] in group[-1] is exiting...
  NumNodes: 2708
  NumEdges: 10556
  NumFeats: 1433
  NumClasses: 7
  NumTrainingSamples: 140
  NumValidationSamples: 500
  NumTestSamples: 1000
Done loading data from cached files.
Converting to homogeneous graph takes 0.001s, peak mem: 1.529 GB
FAILED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-v-shp3-g1] FAILED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-v-shp4-g0] FAILED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-v-shp4-g1] FAILED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-e-shp0-g0] Save partitions: 0.043 seconds, peak memory: 1.529 GB
There are 10556 edges in the graph and 0 edge cuts for 1 partitions.
Client[-1] in group[-1] is exiting...
PASSED [ 27%]
tests/distributed/test_distributed_sampling.py::test_rpc_in_subgraph   NumNodes: 2708
  NumEdges: 10556
  NumFeats: 1433
  NumClasses: 7
  NumTrainingSamples: 140
  NumValidationSamples: 500
  NumTestSamples: 1000
Done loading data from cached files.
Converting to homogeneous graph takes 0.001s, peak mem: 1.529 GB
Convert a graph into a bidirected graph: 0.001 seconds, peak memory: 1.529 GB
Construct multi-constraint weights: 0.000 seconds, peak memory: 1.529 GB
[05:33:11] /root/jenkins/workspace/dgl_PR-4648/src/graph/transform/metis_partition_hetero.cc:87: Partition a graph with 2708 nodes and 10556 edges into 2 parts and get 260 edge cuts
Metis partitioning: 0.002 seconds, peak memory: 1.529 GB
Assigning nodes to METIS partitions takes 0.003s, peak mem: 1.529 GB
Reshuffle nodes and edges: 0.001 seconds
Split the graph: 0.002 seconds
Construct subgraphs: 0.003 seconds
Splitting the graph into partitions takes 0.006s, peak mem: 1.529 GB
part 0 has 1512 nodes and 1354 are inside the partition
part 0 has 5630 edges and 5370 are inside the partition
part 1 has 1552 nodes and 1354 are inside the partition
part 1 has 5446 edges and 5186 are inside the partition
Save partitions: 0.015 seconds, peak memory: 1.529 GB
There are 10556 edges in the graph and 0 edge cuts for 2 partitions.
FAILED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-e-shp0-g1] FAILED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-e-shp1-g0] FAILED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-e-shp1-g1] FAILED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-e-shp2-g0] FAILED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-e-shp2-g1] FAILED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-e-shp3-g0] FAILED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-e-shp3-g1] FAILED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-e-shp4-g0] FAILED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-e-shp4-g1] FAILED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-u-shp0-g0] FAILED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-u-shp0-g1] FAILED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-u-shp1-g0] FAILED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-u-shp1-g1] FAILED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-u-shp2-g0] FAILED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-u-shp2-g1] FAILED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-u-shp3-g0] FAILED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-u-shp3-g1] FAILED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-u-shp4-g0] FAILED [ 71%]/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_in_subgraph
start graph service on server 0 for part 0
[05:33:12] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:33:12] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:17223]...

tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-u-shp4-g1] FAILED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-v-shp0-g0] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-v-shp0-g1] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-v-shp1-g0] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-v-shp1-g1] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-v-shp2-g0] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-v-shp2-g1] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-v-shp3-g0] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-v-shp3-g1] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-v-shp4-g0] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-v-shp4-g1] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-e-shp0-g0] FAILED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-e-shp0-g1] FAILED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-e-shp1-g0] FAILED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-e-shp1-g1] FAILED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-e-shp2-g0] FAILED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-e-shp2-g1] FAILED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-e-shp3-g0] FAILED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-e-shp3-g1] FAILED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-e-shp4-g0] FAILED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-e-shp4-g1] FAILED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-u-shp0-g0] FAILED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-u-shp0-g1] FAILED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-u-shp1-g0] FAILED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-u-shp1-g1] FAILED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-u-shp2-g0] FAILED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-u-shp2-g1] FAILED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-u-shp3-g0] FAILED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-u-shp3-g1] FAILED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-u-shp4-g0] FAILED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-u-shp4-g1] /root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
[05:33:13] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:33:13] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
load test_in_subgraph
start graph service on server 1 for part 1
[05:33:13] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:33:13] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:17225]...
FAILED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-v-shp0-g0] FAILED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-v-shp0-g1] FAILED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-v-shp1-g0] FAILED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-v-shp1-g1] FAILED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-v-shp2-g0] FAILED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-v-shp2-g1] FAILED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-v-shp3-g0] FAILED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-v-shp3-g1] FAILED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-v-shp4-g0] FAILED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-v-shp4-g1] FAILED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-e-shp0-g0] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-e-shp0-g1] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-e-shp1-g0] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-e-shp1-g1] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-e-shp2-g0] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-e-shp2-g1] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-e-shp3-g0] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-e-shp3-g1] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-e-shp4-g0] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-e-shp4-g1] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-u-shp0-g0] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-u-shp0-g1] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-u-shp1-g0] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-u-shp1-g1] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-u-shp2-g0] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-u-shp2-g1] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-u-shp3-g0] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-u-shp3-g1] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-u-shp4-g0] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-u-shp4-g1] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-v-shp0-g0] FAILED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-v-shp0-g1] FAILED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-v-shp1-g0] FAILED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-v-shp1-g1] FAILED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-v-shp2-g0] FAILED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-v-shp2-g1] FAILED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-v-shp3-g0] FAILED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-v-shp3-g1] FAILED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-v-shp4-g0] FAILED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-v-shp4-g1] FAILED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-e-shp0-g0] FAILED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-e-shp0-g1] FAILED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-e-shp1-g0] FAILED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-e-shp1-g1] FAILED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-e-shp2-g0] FAILED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-e-shp2-g1] FAILED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-e-shp3-g0] FAILED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-e-shp3-g1] FAILED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-e-shp4-g0] FAILED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-e-shp4-g1] FAILED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-u-shp0-g0] FAILED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-u-shp0-g1] FAILED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-u-shp1-g0] FAILED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-u-shp1-g1] FAILED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-u-shp2-g0] FAILED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-u-shp2-g1] FAILED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-u-shp3-g0] FAILED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-u-shp3-g1] FAILED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-u-shp4-g0] FAILED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-u-shp4-g1] FAILED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-v-shp0-g0] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-v-shp0-g1] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-v-shp1-g0] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-v-shp1-g1] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-v-shp2-g0] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-v-shp2-g1] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-v-shp3-g0] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-v-shp3-g1] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-v-shp4-g0] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-v-shp4-g1] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-e-shp0-g0] FAILED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-e-shp0-g1] FAILED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-e-shp1-g0] FAILED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-e-shp1-g1] FAILED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-e-shp2-g0] FAILED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-e-shp2-g1] FAILED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-e-shp3-g0] FAILED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-e-shp3-g1] FAILED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-e-shp4-g0] FAILED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-e-shp4-g1] FAILED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-u-shp0-g0] FAILED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-u-shp0-g1] FAILED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-u-shp1-g0] FAILED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-u-shp1-g1] FAILED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-u-shp2-g0] FAILED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-u-shp2-g1] FAILED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-u-shp3-g0] FAILED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-u-shp3-g1] FAILED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-u-shp4-g0] FAILED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-u-shp4-g1] FAILED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-v-shp0-g0] FAILED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-v-shp0-g1] FAILED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-v-shp1-g0] FAILED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-v-shp1-g1] FAILED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-v-shp2-g0] FAILED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-v-shp2-g1] FAILED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-v-shp3-g0] FAILED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-v-shp3-g1] FAILED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-v-shp4-g0] FAILED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-v-shp4-g1] FAILED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-e-shp0-g0] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-e-shp0-g1] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-e-shp1-g0] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-e-shp1-g1] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-e-shp2-g0] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-e-shp2-g1] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-e-shp3-g0] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-e-shp3-g1] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-e-shp4-g0] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-e-shp4-g1] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-u-shp0-g0] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-u-shp0-g1] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-u-shp1-g0] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-u-shp1-g1] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-u-shp2-g0] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-u-shp2-g1] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-u-shp3-g0] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-u-shp3-g1] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-u-shp4-g0] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-u-shp4-g1] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-v-shp0-g0] FAILED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-v-shp0-g1] FAILED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-v-shp1-g0] FAILED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-v-shp1-g1] FAILED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-v-shp2-g0] FAILED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-v-shp2-g1] FAILED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-v-shp3-g0] FAILED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-v-shp3-g1] FAILED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-v-shp4-g0] FAILED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-v-shp4-g1] FAILED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-e-shp0-g0] FAILED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-e-shp0-g1] FAILED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-e-shp1-g0] FAILED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-e-shp1-g1] FAILED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-e-shp2-g0] FAILED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-e-shp2-g1] FAILED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-e-shp3-g0] FAILED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-e-shp3-g1] FAILED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-e-shp4-g0] FAILED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-e-shp4-g1] FAILED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-u-shp0-g0] FAILED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-u-shp0-g1] FAILED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-u-shp1-g0] FAILED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-u-shp1-g1] FAILED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-u-shp2-g0] FAILED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-u-shp2-g1] FAILED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-u-shp3-g0] FAILED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-u-shp3-g1] FAILED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-u-shp4-g0] FAILED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-u-shp4-g1] FAILED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-v-shp0-g0] PASSED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-v-shp0-g1] PASSED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-v-shp1-g0] PASSED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-v-shp1-g1] PASSED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-v-shp2-g0] PASSED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-v-shp2-g1] PASSED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-v-shp3-g0] PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-v-shp3-g1] PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-v-shp4-g0] PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-v-shp4-g1] PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-e-shp0-g0] FAILED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-e-shp0-g1] FAILED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-e-shp1-g0] FAILED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-e-shp1-g1] FAILED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-e-shp2-g0] FAILED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-e-shp2-g1] FAILED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-e-shp3-g0] FAILED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-e-shp3-g1] FAILED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-e-shp4-g0] FAILED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-e-shp4-g1] FAILED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-u-shp0-g0] FAILED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-u-shp0-g1] FAILED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-u-shp1-g0] FAILED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-u-shp1-g1] Client [1090] waits on 172.17.0.3:39697
Machine (0) group (0) client (0) connect to server successfuly!
Client[0] in group[0] is exiting...
Server (0) shutdown.
Server is exiting...
FAILED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-u-shp2-g0] FAILED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-u-shp2-g1] FAILED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-u-shp3-g0] FAILED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-u-shp3-g1] FAILED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-u-shp4-g0] FAILED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-u-shp4-g1] Server (1) shutdown.
Server is exiting...
FAILED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-v-shp0-g0] FAILED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-v-shp0-g1] FAILED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-v-shp1-g0] FAILED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-v-shp1-g1] PASSED [ 29%]
tests/distributed/test_distributed_sampling.py::test_standalone_etype_sampling   NumNodes: 2708
  NumEdges: 10556
  NumFeats: 1433
  NumClasses: 7
  NumTrainingSamples: 140
  NumValidationSamples: 500
  NumTestSamples: 1000
Done loading data from cached files.
Converting to homogeneous graph takes 0.001s, peak mem: 1.529 GB
Save partitions: 0.002 seconds, peak memory: 1.529 GB
There are 21112 edges in the graph and 0 edge cuts for 1 partitions.
Client[-1] in group[-1] is exiting...
  NumNodes: 2708
  NumEdges: 10556
  NumFeats: 1433
  NumClasses: 7
  NumTrainingSamples: 140
  NumValidationSamples: 500
  NumTestSamples: 1000
Done loading data from cached files.
Converting to homogeneous graph takes 0.001s, peak mem: 1.529 GB
Save partitions: 0.030 seconds, peak memory: 1.529 GB
There are 10556 edges in the graph and 0 edge cuts for 1 partitions.
Client[-1] in group[-1] is exiting...
  NumNodes: 2708
  NumEdges: 10556
  NumFeats: 1433
  NumClasses: 7
  NumTrainingSamples: 140
  NumValidationSamples: 500
  NumTestSamples: 1000
Done loading data from cached files.
Converting to homogeneous graph takes 0.001s, peak mem: 1.529 GB
Calculate edge assignment: 0.000 seconds
FAILED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-v-shp2-g0] FAILED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-v-shp2-g1] FAILED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-v-shp3-g0] FAILED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-v-shp3-g1] Save partitions: 0.026 seconds, peak memory: 1.529 GB
There are 10556 edges in the graph and 0 edge cuts for 1 partitions.
Client[-1] in group[-1] is exiting...
PASSED [ 30%]
tests/distributed/test_mp_dataloader.py::test_standalone   NumNodes: 2708
  NumEdges: 10556
  NumFeats: 1433
  NumClasses: 7
  NumTrainingSamples: 140
  NumValidationSamples: 500
  NumTestSamples: 1000
Done loading data from cached files.
torch.int64
Converting to homogeneous graph takes 0.001s, peak mem: 1.529 GB
Save partitions: 0.015 seconds, peak memory: 1.529 GB
There are 10556 edges in the graph and 0 edge cuts for 1 partitions.
  NumNodes: 2708
  NumEdges: 10556
  NumFeats: 1433
  NumClasses: 7
  NumTrainingSamples: 140
  NumValidationSamples: 500
  NumTestSamples: 1000
Done loading data from cached files.
FAILED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-v-shp4-g0] FAILED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-v-shp4-g1] FAILED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-e-shp0-g0] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-e-shp0-g1] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-e-shp1-g0] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-e-shp1-g1] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-e-shp2-g0] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-e-shp2-g1] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-e-shp3-g0] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-e-shp3-g1] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-e-shp4-g0] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-e-shp4-g1] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-u-shp0-g0] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-u-shp0-g1] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-u-shp1-g0] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-u-shp1-g1] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-u-shp2-g0] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-u-shp2-g1] PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-u-shp3-g0] PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-u-shp3-g1] PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-u-shp4-g0] PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-u-shp4-g1] PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-v-shp0-g0] FAILED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-v-shp0-g1]   NumNodes: 2708
  NumEdges: 10556
  NumFeats: 1433
  NumClasses: 7
  NumTrainingSamples: 140
  NumValidationSamples: 500
  NumTestSamples: 1000
Done loading data from cached files.
Client[-1] in group[-1] is exiting...
PASSED          [ 32%]
tests/distributed/test_mp_dataloader.py::test_dist_dataloader[1-True-True-0-3]   NumNodes: 2708
  NumEdges: 10556
  NumFeats: 1433
  NumClasses: 7
  NumTrainingSamples: 140
  NumValidationSamples: 500
  NumTestSamples: 1000
Done loading data from cached files.
torch.int64
Converting to homogeneous graph takes 0.001s, peak mem: 1.529 GB
Convert a graph into a bidirected graph: 0.001 seconds, peak memory: 1.529 GB
Construct multi-constraint weights: 0.000 seconds, peak memory: 1.529 GB
[05:33:21] /root/jenkins/workspace/dgl_PR-4648/src/graph/transform/metis_partition_hetero.cc:87: Partition a graph with 2708 nodes and 10556 edges into 3 parts and get 325 edge cuts
Metis partitioning: 0.002 seconds, peak memory: 1.529 GB
Assigning nodes to METIS partitions takes 0.003s, peak mem: 1.529 GB
Reshuffle nodes and edges: 0.001 seconds
Split the graph: 0.002 seconds
Construct subgraphs: 0.003 seconds
Splitting the graph into partitions takes 0.006s, peak mem: 1.529 GB
part 0 has 1046 nodes and 903 are inside the partition
part 0 has 3490 edges and 3317 are inside the partition
part 1 has 1064 nodes and 902 are inside the partition
part 1 has 4006 edges and 3739 are inside the partition
part 2 has 1056 nodes and 903 are inside the partition
part 2 has 3710 edges and 3500 are inside the partition
Save partitions: 0.017 seconds, peak memory: 1.529 GB
There are 10556 edges in the graph and 0 edge cuts for 3 partitions.
FAILED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-v-shp1-g0] FAILED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-v-shp1-g1] FAILED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-v-shp2-g0] FAILED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-v-shp2-g1] FAILED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-v-shp3-g0] FAILED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-v-shp3-g1] FAILED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-v-shp4-g0] FAILED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-v-shp4-g1] FAILED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-e-shp0-g0] FAILED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-e-shp0-g1] FAILED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-e-shp1-g0] FAILED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-e-shp1-g1] FAILED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-e-shp2-g0] FAILED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-e-shp2-g1] FAILED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-e-shp3-g0] FAILED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-e-shp3-g1] FAILED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-e-shp4-g0] FAILED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-e-shp4-g1] FAILED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-u-shp0-g0] FAILED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-u-shp0-g1] FAILED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-u-shp1-g0] FAILED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-u-shp1-g1] server: #clients=1
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_sampling
start graph service on server 0 for part 0
[05:33:22] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:33:22] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:23848]...
FAILED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-u-shp2-g0] FAILED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-u-shp2-g1] FAILED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-u-shp3-g0] FAILED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-u-shp3-g1] FAILED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-u-shp4-g0] FAILED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-u-shp4-g1] FAILED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-v-shp0-g0] PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-v-shp0-g1] PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-v-shp1-g0] PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-v-shp1-g1] PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-v-shp2-g0] PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-v-shp2-g1] PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-v-shp3-g0] PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-v-shp3-g1] PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-v-shp4-g0] PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-v-shp4-g1] PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-e-shp0-g0] FAILED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-e-shp0-g1] FAILED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-e-shp1-g0] server: #clients=1
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_sampling
start graph service on server 1 for part 1
[05:33:23] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:33:23] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:23851]...
FAILED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-e-shp1-g1] FAILED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-e-shp2-g0] FAILED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-e-shp2-g1] FAILED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-e-shp3-g0] FAILED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-e-shp3-g1] FAILED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-e-shp4-g0] FAILED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-e-shp4-g1] FAILED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-u-shp0-g0] FAILED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-u-shp0-g1] FAILED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-u-shp1-g0] FAILED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-u-shp1-g1] FAILED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-u-shp2-g0] FAILED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-u-shp2-g1] FAILED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-u-shp3-g0] FAILED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-u-shp3-g1] FAILED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-u-shp4-g0] FAILED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-u-shp4-g1] FAILED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-v-shp0-g0] FAILED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-v-shp0-g1] FAILED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-v-shp1-g0] FAILED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-v-shp1-g1] server: #clients=1
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_sampling
start graph service on server 2 for part 2
[05:33:24] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:33:24] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:23854]...
FAILED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-v-shp2-g0] FAILED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-v-shp2-g1] FAILED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-v-shp3-g0] FAILED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-v-shp3-g1] FAILED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-v-shp4-g0] FAILED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-v-shp4-g1] FAILED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-e-shp0-g0] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-e-shp0-g1] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-e-shp1-g0] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-e-shp1-g1] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-e-shp2-g0] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-e-shp2-g1] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-e-shp3-g0] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-e-shp3-g1] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-e-shp4-g0] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-e-shp4-g1] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-u-shp0-g0] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-u-shp0-g1] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-u-shp1-g0] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-u-shp1-g1] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-u-shp2-g0] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-u-shp2-g1] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-u-shp3-g0] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-u-shp3-g1] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-u-shp4-g0] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-u-shp4-g1] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-v-shp0-g0] FAILED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-v-shp0-g1] FAILED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-v-shp1-g0] FAILED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-v-shp1-g1] FAILED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-v-shp2-g0] FAILED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-v-shp2-g1] FAILED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-v-shp3-g0] FAILED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-v-shp3-g1] FAILED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-v-shp4-g0] FAILED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-v-shp4-g1] FAILED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-e-shp0-g0] FAILED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-e-shp0-g1] FAILED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-e-shp1-g0] [05:33:25] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:33:25] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
FAILED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-e-shp1-g1] FAILED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-e-shp2-g0] FAILED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-e-shp2-g1] FAILED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-e-shp3-g0] FAILED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-e-shp3-g1] FAILED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-e-shp4-g0] FAILED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-e-shp4-g1] FAILED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-u-shp0-g0] FAILED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-u-shp0-g1] FAILED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-u-shp1-g0] FAILED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-u-shp1-g1] FAILED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-u-shp2-g0] FAILED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-u-shp2-g1] FAILED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-u-shp3-g0] FAILED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-u-shp3-g1] FAILED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-u-shp4-g0] FAILED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-u-shp4-g1] FAILED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-v-shp0-g0] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-v-shp0-g1] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-v-shp1-g0] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-v-shp1-g1] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-v-shp2-g0] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-v-shp2-g1] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-v-shp3-g0] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-v-shp3-g1] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-v-shp4-g0] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-v-shp4-g1] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-e-shp0-g0] FAILED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-e-shp0-g1] FAILED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-e-shp1-g0] FAILED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-e-shp1-g1] FAILED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-e-shp2-g0] FAILED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-e-shp2-g1] FAILED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-e-shp3-g0] FAILED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-e-shp3-g1] FAILED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-e-shp4-g0] FAILED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-e-shp4-g1] FAILED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-u-shp0-g0] FAILED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-u-shp0-g1] FAILED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-u-shp1-g0] FAILED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-u-shp1-g1] FAILED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-u-shp2-g0] FAILED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-u-shp2-g1] FAILED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-u-shp3-g0] FAILED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-u-shp3-g1] FAILED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-u-shp4-g0] FAILED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-u-shp4-g1] FAILED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-v-shp0-g0] FAILED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-v-shp0-g1] FAILED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-v-shp1-g0] FAILED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-v-shp1-g1] FAILED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-v-shp2-g0] FAILED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-v-shp2-g1] FAILED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-v-shp3-g0] FAILED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-v-shp3-g1] FAILED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-v-shp4-g0] FAILED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-v-shp4-g1] FAILED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-e-shp0-g0] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-e-shp0-g1] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-e-shp1-g0] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-e-shp1-g1] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-e-shp2-g0] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-e-shp2-g1] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-e-shp3-g0] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-e-shp3-g1] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-e-shp4-g0] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-e-shp4-g1] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-u-shp0-g0] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-u-shp0-g1] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-u-shp1-g0] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-u-shp1-g1] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-u-shp2-g0] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-u-shp2-g1] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-u-shp3-g0] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-u-shp3-g1] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-u-shp4-g0] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-u-shp4-g1] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-v-shp0-g0] FAILED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-v-shp0-g1] FAILED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-v-shp1-g0] FAILED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-v-shp1-g1] FAILED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-v-shp2-g0] FAILED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-v-shp2-g1] FAILED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-v-shp3-g0] FAILED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-v-shp3-g1] FAILED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-v-shp4-g0] Client [1719] waits on 172.17.0.3:55355
Machine (0) group (0) client (0) connect to server successfuly!
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
  NumNodes: 2708
  NumEdges: 10556
  NumFeats: 1433
  NumClasses: 7
  NumTrainingSamples: 140
  NumValidationSamples: 500
  NumTestSamples: 1000
Done loading data from cached files.
  NumNodes: 2708
  NumEdges: 10556
  NumFeats: 1433
  NumClasses: 7
  NumTrainingSamples: 140
  NumValidationSamples: 500
  NumTestSamples: 1000
Done loading data from cached files.
FAILED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-v-shp4-g1] FAILED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-e-shp0-g0] FAILED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-e-shp0-g1] FAILED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-e-shp1-g0] Client[0] in group[0] is exiting...
Server (1) shutdown.
Server is exiting...
Server (0) shutdown.
Server is exiting...
FAILED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-e-shp1-g1] FAILED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-e-shp2-g0] FAILED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-e-shp2-g1] FAILED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-e-shp3-g0] FAILED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-e-shp3-g1] FAILED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-e-shp4-g0] FAILED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-e-shp4-g1] FAILED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-u-shp0-g0] FAILED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-u-shp0-g1] FAILED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-u-shp1-g0] FAILED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-u-shp1-g1] FAILED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-u-shp2-g0] Server (2) shutdown.
Server is exiting...
FAILED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-u-shp2-g1] FAILED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-u-shp3-g0] FAILED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-u-shp3-g1] FAILED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-u-shp4-g0] FAILED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-u-shp4-g1] PASSED [ 34%]
tests/distributed/test_mp_dataloader.py::test_dist_dataloader[1-True-True-4-3]   NumNodes: 2708
  NumEdges: 10556
  NumFeats: 1433
  NumClasses: 7
  NumTrainingSamples: 140
  NumValidationSamples: 500
  NumTestSamples: 1000
Done loading data from cached files.
torch.int64
Converting to homogeneous graph takes 0.001s, peak mem: 1.529 GB
Convert a graph into a bidirected graph: 0.001 seconds, peak memory: 1.529 GB
Construct multi-constraint weights: 0.000 seconds, peak memory: 1.529 GB
[05:33:30] /root/jenkins/workspace/dgl_PR-4648/src/graph/transform/metis_partition_hetero.cc:87: Partition a graph with 2708 nodes and 10556 edges into 3 parts and get 325 edge cuts
Metis partitioning: 0.002 seconds, peak memory: 1.529 GB
Assigning nodes to METIS partitions takes 0.003s, peak mem: 1.529 GB
Reshuffle nodes and edges: 0.001 seconds
Split the graph: 0.002 seconds
Construct subgraphs: 0.003 seconds
Splitting the graph into partitions takes 0.007s, peak mem: 1.529 GB
part 0 has 1046 nodes and 903 are inside the partition
part 0 has 3490 edges and 3317 are inside the partition
part 1 has 1064 nodes and 902 are inside the partition
part 1 has 4006 edges and 3739 are inside the partition
part 2 has 1056 nodes and 903 are inside the partition
part 2 has 3710 edges and 3500 are inside the partition
Save partitions: 0.017 seconds, peak memory: 1.529 GB
There are 10556 edges in the graph and 0 edge cuts for 3 partitions.
FAILED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-v-shp0-g0] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-v-shp0-g1] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-v-shp1-g0] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-v-shp1-g1] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-v-shp2-g0] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-v-shp2-g1] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-v-shp3-g0] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-v-shp3-g1] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-v-shp4-g0] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-v-shp4-g1] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-e-shp0-g0] FAILED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-e-shp0-g1] FAILED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-e-shp1-g0] FAILED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-e-shp1-g1] FAILED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-e-shp2-g0] FAILED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-e-shp2-g1] FAILED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-e-shp3-g0] FAILED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-e-shp3-g1] FAILED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-e-shp4-g0] FAILED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-e-shp4-g1] FAILED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-u-shp0-g0] FAILED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-u-shp0-g1] FAILED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-u-shp1-g0] FAILED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-u-shp1-g1] FAILED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-u-shp2-g0] server: #clients=5
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_sampling
start graph service on server 0 for part 0
[05:33:31] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:33:31] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:21149]...
FAILED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-u-shp2-g1] FAILED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-u-shp3-g0] FAILED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-u-shp3-g1] FAILED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-u-shp4-g0] FAILED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-u-shp4-g1] FAILED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-v-shp0-g0] FAILED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-v-shp0-g1] FAILED [ 91%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-v-shp1-g0] FAILED [ 91%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-v-shp1-g1] FAILED [ 91%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-v-shp2-g0] FAILED [ 91%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-v-shp2-g1] FAILED [ 91%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-v-shp3-g0] FAILED [ 91%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-v-shp3-g1] FAILED [ 91%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-v-shp4-g0] FAILED [ 91%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-v-shp4-g1] FAILED [ 91%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-e-shp0-g0] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-e-shp0-g1] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-e-shp1-g0] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-e-shp1-g1] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-e-shp2-g0] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-e-shp2-g1] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-e-shp3-g0] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-e-shp3-g1] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-e-shp4-g0] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-e-shp4-g1] PASSED [ 91%]
tests/compute/test_sparse.py::test_edge_softmax[idtype0-shp0-src-g0] server: #clients=5
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_sampling
start graph service on server 1 for part 1
[05:33:32] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:33:32] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:21152]...
FAILED [ 91%]
tests/compute/test_sparse.py::test_edge_softmax[idtype0-shp0-dst-g0] FAILED [ 91%]
tests/compute/test_sparse.py::test_edge_softmax[idtype0-shp1-src-g0] FAILED [ 91%]
tests/compute/test_sparse.py::test_edge_softmax[idtype0-shp1-dst-g0] FAILED [ 91%]
tests/compute/test_sparse.py::test_edge_softmax[idtype0-shp2-src-g0] FAILED [ 91%]
tests/compute/test_sparse.py::test_edge_softmax[idtype0-shp2-dst-g0] FAILED [ 91%]
tests/compute/test_sparse.py::test_edge_softmax[idtype1-shp0-src-g0] FAILED [ 92%]
tests/compute/test_sparse.py::test_edge_softmax[idtype1-shp0-dst-g0] FAILED [ 92%]
tests/compute/test_sparse.py::test_edge_softmax[idtype1-shp1-src-g0] FAILED [ 92%]
tests/compute/test_sparse.py::test_edge_softmax[idtype1-shp1-dst-g0] FAILED [ 92%]
tests/compute/test_sparse.py::test_edge_softmax[idtype1-shp2-src-g0] FAILED [ 92%]
tests/compute/test_sparse.py::test_edge_softmax[idtype1-shp2-dst-g0] FAILED [ 92%]
tests/compute/test_sparse.py::test_segment_reduce[sum] FAILED            [ 92%]
tests/compute/test_sparse.py::test_segment_reduce[max] FAILED            [ 92%]
tests/compute/test_sparse.py::test_segment_reduce[min] FAILED            [ 92%]
tests/compute/test_sparse.py::test_segment_reduce[mean] FAILED           [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype0-0.01-1-idtype0] SKIPPED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype0-0.01-1-idtype1] SKIPPED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype0-0.01-8-idtype0] SKIPPED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype0-0.01-8-idtype1] SKIPPED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype0-0.01-16-idtype0] SKIPPED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype0-0.01-16-idtype1] SKIPPED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype0-0.01-64-idtype0] SKIPPED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype0-0.01-64-idtype1] SKIPPED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype0-0.01-256-idtype0] SKIPPED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype0-0.01-256-idtype1] SKIPPED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype1-0.01-1-idtype0] SKIPPED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype1-0.01-1-idtype1] SKIPPED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype1-0.01-8-idtype0] SKIPPED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype1-0.01-8-idtype1] SKIPPED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype1-0.01-16-idtype0] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype1-0.01-16-idtype1] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype1-0.01-64-idtype0] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype1-0.01-64-idtype1] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype1-0.01-256-idtype0] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype1-0.01-256-idtype1] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype2-0.003-1-idtype0] FAILED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype2-0.003-1-idtype1] FAILED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype2-0.003-8-idtype0] FAILED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype2-0.003-8-idtype1] FAILED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype2-0.003-16-idtype0] FAILED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype2-0.003-16-idtype1] FAILED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype2-0.003-64-idtype0] FAILED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype2-0.003-64-idtype1] FAILED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype2-0.003-256-idtype0] FAILED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype2-0.003-256-idtype1] FAILED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype3-0.0001-1-idtype0] FAILED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype3-0.0001-1-idtype1] FAILED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype3-0.0001-8-idtype0] FAILED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype3-0.0001-8-idtype1] FAILED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype3-0.0001-16-idtype0] FAILED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype3-0.0001-16-idtype1] FAILED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype3-0.0001-64-idtype0] FAILED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype3-0.0001-64-idtype1] server: #clients=5
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_sampling
start graph service on server 2 for part 2
[05:33:33] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:33:33] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:21155]...
FAILED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype3-0.0001-256-idtype0] FAILED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype3-0.0001-256-idtype1] FAILED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype0-0.01-1] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype0-0.01-8] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype0-0.01-16] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype0-0.01-64] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype0-0.01-256] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype1-0.02-1] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype1-0.02-8] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype1-0.02-16] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype1-0.02-64] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype1-0.02-256] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype2-0.003-1] FAILED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype2-0.003-8] FAILED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype2-0.003-16] FAILED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype2-0.003-64] FAILED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype2-0.003-256] FAILED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype3-0.0001-1] FAILED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype3-0.0001-8] FAILED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype3-0.0001-16] FAILED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype3-0.0001-64] FAILED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype3-0.0001-256] FAILED [ 94%]
tests/compute/test_sparse.py::test_use_libxsmm_switch FAILED             [ 94%]
tests/compute/test_specialization.py::test_v2v_update_all[idtype0] FAILED [ 94%]
tests/compute/test_specialization.py::test_v2v_update_all[idtype1] FAILED [ 94%]
tests/compute/test_specialization.py::test_v2v_snr[idtype0] [05:33:34] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:33:34] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
FAILED       [ 94%]
tests/compute/test_specialization.py::test_v2v_snr[idtype1] FAILED       [ 95%]
tests/compute/test_specialization.py::test_v2v_pull[idtype0] FAILED      [ 95%]
tests/compute/test_specialization.py::test_v2v_pull[idtype1] FAILED      [ 95%]
tests/compute/test_specialization.py::test_update_all_multi_fallback[idtype0] FAILED [ 95%]
tests/compute/test_specialization.py::test_update_all_multi_fallback[idtype1] [05:33:35] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:33:35] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
[05:33:35] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:33:35] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
[05:33:35] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:33:35] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
[05:33:35] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:33:35] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
FAILED [ 95%]
tests/compute/test_specialization.py::test_pull_multi_fallback[idtype0] FAILED [ 95%]
tests/compute/test_specialization.py::test_pull_multi_fallback[idtype1] FAILED [ 95%]
tests/compute/test_specialization.py::test_spmv_3d_feat[idtype0] FAILED  [ 95%]
tests/compute/test_specialization.py::test_spmv_3d_feat[idtype1] FAILED  [ 95%]
tests/compute/test_subgraph.py::test_edge_subgraph PASSED                [ 95%]
tests/compute/test_subgraph.py::test_subgraph PASSED                     [ 95%]
tests/compute/test_subgraph.py::test_subgraph_mask[idtype0] PASSED       [ 95%]
tests/compute/test_subgraph.py::test_subgraph_mask[idtype1] PASSED       [ 95%]
tests/compute/test_subgraph.py::test_subgraph1[idtype0] PASSED           [ 95%]
tests/compute/test_subgraph.py::test_subgraph1[idtype1] PASSED           [ 95%]
tests/compute/test_subgraph.py::test_in_subgraph[idtype0] PASSED         [ 95%]
tests/compute/test_subgraph.py::test_in_subgraph[idtype1] PASSED         [ 95%]
tests/compute/test_subgraph.py::test_out_subgraph[idtype0] PASSED        [ 95%]
tests/compute/test_subgraph.py::test_out_subgraph[idtype1] PASSED        [ 95%]
tests/compute/test_subgraph.py::test_subgraph_message_passing PASSED     [ 95%]
tests/compute/test_subgraph.py::test_khop_in_subgraph[idtype0] PASSED    [ 95%]
tests/compute/test_subgraph.py::test_khop_in_subgraph[idtype1] PASSED    [ 95%]
tests/compute/test_subgraph.py::test_khop_out_subgraph[idtype0] PASSED   [ 95%]
tests/compute/test_subgraph.py::test_khop_out_subgraph[idtype1] Client [1756] waits on 172.17.0.3:58147Client [1755] waits on 172.17.0.3:52765

Client [1753] waits on 172.17.0.3:32967Client [1754] waits on 172.17.0.3:38817Client [1749] waits on 172.17.0.3:43513


Machine (0) group (0) client (1) connect to server successfuly!Machine (0) group (0) client (0) connect to server successfuly!

Machine (0) group (0) client (3) connect to server successfuly!
Machine (0) group (0) client (2) connect to server successfuly!Machine (0) group (0) client (4) connect to server successfuly!

/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
  NumNodes: 2708
  NumEdges: 10556
  NumFeats: 1433
  NumClasses: 7
  NumTrainingSamples: 140
  NumValidationSamples: 500
  NumTestSamples: 1000
Done loading data from cached files.
  NumNodes: 2708
  NumEdges: 10556
  NumFeats: 1433
  NumClasses: 7
  NumTrainingSamples: 140
  NumValidationSamples: 500
  NumTestSamples: 1000
Done loading data from cached files.
PASSED   [ 95%]
tests/compute/test_subgraph.py::test_subframes[child_device0-parent_idx_device0] SKIPPED [ 96%]
tests/compute/test_subgraph.py::test_subframes[child_device0-parent_idx_device1] SKIPPED [ 96%]
tests/compute/test_subgraph.py::test_subframes[child_device0-parent_idx_device2] SKIPPED [ 96%]
tests/compute/test_subgraph.py::test_subframes[child_device0-parent_idx_device3] SKIPPED [ 96%]
tests/compute/test_subgraph.py::test_subframes[child_device1-parent_idx_device0] SKIPPED [ 96%]
tests/compute/test_subgraph.py::test_subframes[child_device1-parent_idx_device1] SKIPPED [ 96%]
tests/compute/test_subgraph.py::test_subframes[child_device1-parent_idx_device2] SKIPPED [ 96%]
tests/compute/test_subgraph.py::test_subframes[child_device1-parent_idx_device3] SKIPPED [ 96%]
tests/compute/test_subgraph.py::test_uva_subgraph[idtype0-device0] SKIPPED [ 96%]
tests/compute/test_subgraph.py::test_uva_subgraph[idtype0-device1] SKIPPED [ 96%]
tests/compute/test_subgraph.py::test_uva_subgraph[idtype1-device0] SKIPPED [ 96%]
tests/compute/test_subgraph.py::test_uva_subgraph[idtype1-device1] SKIPPED [ 96%]
tests/compute/test_transform.py::test_line_graph1 PASSED                 [ 96%]
tests/compute/test_transform.py::test_line_graph2[idtype0] PASSED        [ 96%]
tests/compute/test_transform.py::test_line_graph2[idtype1] PASSED        [ 96%]
tests/compute/test_transform.py::test_no_backtracking PASSED             [ 96%]
tests/compute/test_transform.py::test_reverse[idtype0] PASSED            [ 96%]
tests/compute/test_transform.py::test_reverse[idtype1] PASSED            [ 96%]
tests/compute/test_transform.py::test_reverse_shared_frames[idtype0] Client[2] in group[0] is exiting...
Client[0] in group[0] is exiting...
PASSED [ 96%]
tests/compute/test_transform.py::test_reverse_shared_frames[idtype1] PASSED [ 96%]
tests/compute/test_transform.py::test_to_bidirected PASSED               [ 96%]
tests/compute/test_transform.py::test_add_reverse_edges PASSED           [ 96%]
tests/compute/test_transform.py::test_simple_graph PASSED                [ 96%]
tests/compute/test_transform.py::test_khop_graph Client[1] in group[0] is exiting...
FAILED                  [ 96%]
tests/compute/test_transform.py::test_khop_adj FAILED                    [ 96%]
tests/compute/test_transform.py::test_laplacian_lambda_max PASSED        [ 97%]
tests/compute/test_transform.py::test_partition_with_halo Client[3] in group[0] is exiting...
PASSED         [ 97%]
tests/compute/test_transform.py::test_metis_partition[idtype0] PASSED    [ 97%]
tests/compute/test_transform.py::test_metis_partition[idtype1] Client[4] in group[0] is exiting...
Server (0) shutdown.
Server is exiting...
Server (1) shutdown.
Server is exiting...
Server (2) shutdown.
Server is exiting...
PASSED [ 36%]
tests/distributed/test_mp_dataloader.py::test_dist_dataloader[1-True-False-0-3]   NumNodes: 2708
  NumEdges: 10556
  NumFeats: 1433
  NumClasses: 7
  NumTrainingSamples: 140
  NumValidationSamples: 500
  NumTestSamples: 1000
Done loading data from cached files.
torch.int64
Converting to homogeneous graph takes 0.001s, peak mem: 1.529 GB
Convert a graph into a bidirected graph: 0.001 seconds, peak memory: 1.529 GB
Construct multi-constraint weights: 0.000 seconds, peak memory: 1.529 GB
PASSED    [ 97%]
tests/compute/test_transform.py::test_reorder_nodes [05:33:42] /root/jenkins/workspace/dgl_PR-4648/src/graph/transform/metis_partition_hetero.cc:87: Partition a graph with 2708 nodes and 10556 edges into 3 parts and get 325 edge cuts
Metis partitioning: 0.003 seconds, peak memory: 1.529 GB
Assigning nodes to METIS partitions takes 0.004s, peak mem: 1.529 GB
Reshuffle nodes and edges: 0.001 seconds
Split the graph: 0.002 seconds
Construct subgraphs: 0.004 seconds
Splitting the graph into partitions takes 0.008s, peak mem: 1.529 GB
part 0 has 1046 nodes and 903 are inside the partition
part 0 has 3490 edges and 3317 are inside the partition
part 1 has 1064 nodes and 902 are inside the partition
part 1 has 4006 edges and 3739 are inside the partition
part 2 has 1056 nodes and 903 are inside the partition
part 2 has 3710 edges and 3500 are inside the partition
Save partitions: 0.017 seconds, peak memory: 1.529 GB
There are 10556 edges in the graph and 0 edge cuts for 3 partitions.
PASSED               [ 97%]
tests/compute/test_transform.py::test_compact[idtype0] PASSED            [ 97%]
tests/compute/test_transform.py::test_compact[idtype1] PASSED            [ 97%]
tests/compute/test_transform.py::test_to_simple[idtype0] FAILED          [ 97%]
tests/compute/test_transform.py::test_to_simple[idtype1] FAILED          [ 97%]
tests/compute/test_transform.py::test_to_block[idtype0] PASSED           [ 97%]
tests/compute/test_transform.py::test_to_block[idtype1] PASSED           [ 97%]
tests/compute/test_transform.py::test_remove_edges[idtype0] FAILED       [ 97%]
tests/compute/test_transform.py::test_remove_edges[idtype1] server: #clients=1
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_sampling
start graph service on server 0 for part 0
[05:33:43] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:33:43] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:19105]...
FAILED       [ 97%]
tests/compute/test_transform.py::test_add_edges[idtype0] PASSED          [ 97%]
tests/compute/test_transform.py::test_add_edges[idtype1] PASSED          [ 97%]
tests/compute/test_transform.py::test_add_nodes[idtype0] PASSED          [ 97%]
tests/compute/test_transform.py::test_add_nodes[idtype1] PASSED          [ 97%]
tests/compute/test_transform.py::test_remove_nodes[idtype0] FAILED       [ 97%]
tests/compute/test_transform.py::test_remove_nodes[idtype1] server: #clients=1
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_sampling
start graph service on server 1 for part 1
[05:33:44] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:33:44] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:19108]...
FAILED       [ 97%]
tests/compute/test_transform.py::test_add_selfloop[idtype0] FAILED       [ 97%]
tests/compute/test_transform.py::test_add_selfloop[idtype1] FAILED       [ 97%]
tests/compute/test_transform.py::test_remove_selfloop[idtype0] server: #clients=1
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_sampling
start graph service on server 2 for part 2
[05:33:45] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:33:45] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:19111]...
FAILED    [ 97%]
tests/compute/test_transform.py::test_remove_selfloop[idtype1] FAILED    [ 97%]
tests/compute/test_transform.py::test_reorder_graph[idtype0] PASSED      [ 97%]
tests/compute/test_transform.py::test_reorder_graph[idtype1] PASSED      [ 98%]
tests/compute/test_transform.py::test_norm_by_dst[idtype0] PASSED        [ 98%]
tests/compute/test_transform.py::test_norm_by_dst[idtype1] PASSED        [ 98%]
tests/compute/test_transform.py::test_module_add_self_loop[idtype0] PASSED [ 98%]
tests/compute/test_transform.py::test_module_add_self_loop[idtype1] PASSED [ 98%]
tests/compute/test_transform.py::test_module_remove_self_loop[idtype0] PASSED [ 98%]
tests/compute/test_transform.py::test_module_remove_self_loop[idtype1] PASSED [ 98%]
tests/compute/test_transform.py::test_module_add_reverse[idtype0] PASSED [ 98%]
tests/compute/test_transform.py::test_module_add_reverse[idtype1] PASSED [ 98%]
tests/compute/test_transform.py::test_module_to_simple[idtype0] PASSED   [ 98%]
tests/compute/test_transform.py::test_module_to_simple[idtype1] PASSED   [ 98%]
tests/compute/test_transform.py::test_module_line_graph[idtype0] PASSED  [ 98%]
tests/compute/test_transform.py::test_module_line_graph[idtype1] PASSED  [ 98%]
tests/compute/test_transform.py::test_module_khop_graph[idtype0] PASSED  [ 98%]
tests/compute/test_transform.py::test_module_khop_graph[idtype1] PASSED  [ 98%]
tests/compute/test_transform.py::test_module_add_metapaths[idtype0] PASSED [ 98%]
tests/compute/test_transform.py::test_module_add_metapaths[idtype1] PASSED [ 98%]
tests/compute/test_transform.py::test_module_compose[idtype0] PASSED     [ 98%]
tests/compute/test_transform.py::test_module_compose[idtype1] PASSED     [ 98%]
tests/compute/test_transform.py::test_module_gcnnorm[idtype0] [05:33:46] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:33:46] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
FAILED     [ 98%]
tests/compute/test_transform.py::test_module_gcnnorm[idtype1] FAILED     [ 98%]
tests/compute/test_transform.py::test_module_ppr[idtype0] PASSED         [ 98%]
tests/compute/test_transform.py::test_module_ppr[idtype1] PASSED         [ 98%]
tests/compute/test_transform.py::test_module_heat_kernel[idtype0] PASSED [ 98%]
tests/compute/test_transform.py::test_module_heat_kernel[idtype1] PASSED [ 98%]
tests/compute/test_transform.py::test_module_gdc[idtype0] PASSED         [ 99%]
tests/compute/test_transform.py::test_module_gdc[idtype1] PASSED         [ 99%]
tests/compute/test_transform.py::test_module_node_shuffle[idtype0] PASSED [ 99%]
tests/compute/test_transform.py::test_module_node_shuffle[idtype1] PASSED [ 99%]
tests/compute/test_transform.py::test_module_drop_node[idtype0] PASSED   [ 99%]
tests/compute/test_transform.py::test_module_drop_node[idtype1] PASSED   [ 99%]
tests/compute/test_transform.py::test_module_drop_edge[idtype0] PASSED   [ 99%]
tests/compute/test_transform.py::test_module_drop_edge[idtype1] PASSED   [ 99%]
tests/compute/test_transform.py::test_module_add_edge[idtype0] PASSED    [ 99%]
tests/compute/test_transform.py::test_module_add_edge[idtype1] PASSED    [ 99%]
tests/compute/test_transform.py::test_module_random_walk_pe[idtype0] PASSED [ 99%]
tests/compute/test_transform.py::test_module_random_walk_pe[idtype1] PASSED [ 99%]
tests/compute/test_transform.py::test_module_laplacian_pe[idtype0] PASSED [ 99%]
tests/compute/test_transform.py::test_module_laplacian_pe[idtype1] PASSED [ 99%]
tests/compute/test_transform.py::test_module_sign[g0] FAILED             [ 99%]
tests/compute/test_transform.py::test_module_row_feat_normalizer[idtype0] PASSED [ 99%]
tests/compute/test_transform.py::test_module_row_feat_normalizer[idtype1] PASSED [ 99%]
tests/compute/test_transform.py::test_module_feat_mask[idtype0] PASSED   [ 99%]
tests/compute/test_transform.py::test_module_feat_mask[idtype1] PASSED   [ 99%]
tests/compute/test_traversal.py::test_bfs[idtype0] PASSED                [ 99%]
tests/compute/test_traversal.py::test_bfs[idtype1] PASSED                [ 99%]
tests/compute/test_traversal.py::test_topological_nodes[idtype0] PASSED  [ 99%]
tests/compute/test_traversal.py::test_topological_nodes[idtype1] PASSED  [ 99%]
tests/compute/test_traversal.py::test_dfs_labeled_edges[idtype0] PASSED  [ 99%]
tests/compute/test_traversal.py::test_dfs_labeled_edges[idtype1] PASSED  [100%]

=================================== FAILURES ===================================
__________________________ test_unary_copy_u[idtype0] __________________________

idtype = torch.int32

    @parametrize_idtype
    def test_unary_copy_u(idtype):
        def _test(mfunc):
    
            g = create_test_heterograph(idtype)
    
            x1 = F.randn((g.num_nodes('user'), feat_size))
            x2 = F.randn((g.num_nodes('developer'), feat_size))
    
            F.attach_grad(x1)
            F.attach_grad(x2)
            g.nodes['user'].data['h'] = x1
            g.nodes['developer'].data['h'] = x2
    
            #################################################################
            #  apply_edges() is called on each relation type separately
            #################################################################
    
            with F.record_grad():
                [g.apply_edges(fn.copy_u('h', 'm'), etype = rel)
                    for rel in g.canonical_etypes]
                r1 = g['plays'].edata['m']
                F.backward(r1, F.ones(r1.shape))
                n_grad1 = F.grad(g.ndata['h']['user'])
            # TODO (Israt): clear not working
            g.edata['m'].clear()
    
            #################################################################
            #  apply_edges() is called on all relation types
            #################################################################
    
            g.apply_edges(fn.copy_u('h', 'm'))
            r2 = g['plays'].edata['m']
            F.backward(r2, F.ones(r2.shape))
            n_grad2 = F.grad(g.nodes['user'].data['h'])
    
            # correctness check
            def _print_error(a, b):
                for i, (x, y) in enumerate(zip(F.asnumpy(a).flatten(), F.asnumpy(b).flatten())):
                    if not np.allclose(x, y):
                        print('@{} {} v.s. {}'.format(i, x, y))
    
            if not F.allclose(r1, r2):
                _print_error(r1, r2)
            assert F.allclose(r1, r2)
            if not F.allclose(n_grad1, n_grad2):
                print('node grad')
                _print_error(n_grad1, n_grad2)
            assert(F.allclose(n_grad1, n_grad2))
    
>       _test(fn.copy_u)

tests/compute/test_apply_edges_hetero.py:92: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
tests/compute/test_apply_edges_hetero.py:62: in _test
    for rel in g.canonical_etypes]
tests/compute/test_apply_edges_hetero.py:62: in <listcomp>
    for rel in g.canonical_etypes]
python/dgl/heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python/dgl/core.py:276: in invoke_gsddmm
    z = op(graph, x)
python/dgl/ops/sddmm.py:164: in copy_u
    return gsddmm(g, 'copy_lhs', x, None)
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72ade9c60>
op = 'copy_lhs'
lhs_data = tensor([[-0.8083,  2.2130],
        [-1.3532, -0.4294]], requires_grad=True)
rhs_data = None, lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
__________________________ test_unary_copy_u[idtype1] __________________________

idtype = torch.int64

    @parametrize_idtype
    def test_unary_copy_u(idtype):
        def _test(mfunc):
    
            g = create_test_heterograph(idtype)
    
            x1 = F.randn((g.num_nodes('user'), feat_size))
            x2 = F.randn((g.num_nodes('developer'), feat_size))
    
            F.attach_grad(x1)
            F.attach_grad(x2)
            g.nodes['user'].data['h'] = x1
            g.nodes['developer'].data['h'] = x2
    
            #################################################################
            #  apply_edges() is called on each relation type separately
            #################################################################
    
            with F.record_grad():
                [g.apply_edges(fn.copy_u('h', 'm'), etype = rel)
                    for rel in g.canonical_etypes]
                r1 = g['plays'].edata['m']
                F.backward(r1, F.ones(r1.shape))
                n_grad1 = F.grad(g.ndata['h']['user'])
            # TODO (Israt): clear not working
            g.edata['m'].clear()
    
            #################################################################
            #  apply_edges() is called on all relation types
            #################################################################
    
            g.apply_edges(fn.copy_u('h', 'm'))
            r2 = g['plays'].edata['m']
            F.backward(r2, F.ones(r2.shape))
            n_grad2 = F.grad(g.nodes['user'].data['h'])
    
            # correctness check
            def _print_error(a, b):
                for i, (x, y) in enumerate(zip(F.asnumpy(a).flatten(), F.asnumpy(b).flatten())):
                    if not np.allclose(x, y):
                        print('@{} {} v.s. {}'.format(i, x, y))
    
            if not F.allclose(r1, r2):
                _print_error(r1, r2)
            assert F.allclose(r1, r2)
            if not F.allclose(n_grad1, n_grad2):
                print('node grad')
                _print_error(n_grad1, n_grad2)
            assert(F.allclose(n_grad1, n_grad2))
    
>       _test(fn.copy_u)

tests/compute/test_apply_edges_hetero.py:92: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
tests/compute/test_apply_edges_hetero.py:62: in _test
    for rel in g.canonical_etypes]
tests/compute/test_apply_edges_hetero.py:62: in <listcomp>
    for rel in g.canonical_etypes]
python/dgl/heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python/dgl/core.py:276: in invoke_gsddmm
    z = op(graph, x)
python/dgl/ops/sddmm.py:164: in copy_u
    return gsddmm(g, 'copy_lhs', x, None)
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72aeba4c8>
op = 'copy_lhs'
lhs_data = tensor([[ 0.7500,  0.1746],
        [-1.8386, -3.9195]], requires_grad=True)
rhs_data = None, lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
___________________________ test_binary_op[idtype0] ____________________________

idtype = torch.int32

    @parametrize_idtype
    def test_binary_op(idtype):
        def _test(lhs, rhs, binary_op):
    
            g = create_test_heterograph(idtype)
    
            n1 = F.randn((g.num_nodes('user'), feat_size))
            n2 = F.randn((g.num_nodes('developer'), feat_size))
            n3 = F.randn((g.num_nodes('game'), feat_size))
    
            x1 = F.randn((g.num_edges('plays'),feat_size))
            x2 = F.randn((g.num_edges('follows'),feat_size))
            x3 = F.randn((g.num_edges('develops'),feat_size))
            x4 = F.randn((g.num_edges('wishes'),feat_size))
    
            builtin_msg_name = "{}_{}_{}".format(lhs, binary_op, rhs)
            builtin_msg = getattr(fn, builtin_msg_name)
    
            #################################################################
            #  apply_edges() is called on each relation type separately
            #################################################################
    
            F.attach_grad(n1)
            F.attach_grad(n2)
            F.attach_grad(n3)
            g.nodes['user'].data['h'] = n1
            g.nodes['developer'].data['h'] = n2
            g.nodes['game'].data['h'] = n3
            F.attach_grad(x1)
            F.attach_grad(x2)
            F.attach_grad(x3)
            F.attach_grad(x4)
            g['plays'].edata['h'] = x1
            g['follows'].edata['h'] = x2
            g['develops'].edata['h'] = x3
            g['wishes'].edata['h'] = x4
    
            with F.record_grad():
                [g.apply_edges(builtin_msg('h', 'h', 'm'), etype = rel)
                    for rel in g.canonical_etypes]
                r1 = g['plays'].edata['m']
                loss = F.sum(r1.view(-1), 0)
                F.backward(loss)
                n_grad1 = F.grad(g.nodes['game'].data['h'])
    
            #################################################################
            #  apply_edges() is called on all relation types
            #################################################################
    
            F.attach_grad(n1)
            F.attach_grad(n2)
            F.attach_grad(n3)
            g.nodes['user'].data['h'] = n1
            g.nodes['developer'].data['h'] = n2
            g.nodes['game'].data['h'] = n3
            F.attach_grad(x1)
            F.attach_grad(x2)
            F.attach_grad(x3)
            F.attach_grad(x4)
            g['plays'].edata['h'] = x1
            g['follows'].edata['h'] = x2
            g['develops'].edata['h'] = x3
            g['wishes'].edata['h'] = x4
    
            with F.record_grad():
                g.apply_edges(builtin_msg('h', 'h', 'm'))
                r2 = g['plays'].edata['m']
                loss = F.sum(r2.view(-1), 0)
                F.backward(loss)
                n_grad2 = F.grad(g.nodes['game'].data['h'])
            # correctness check
            def _print_error(a, b):
                for i, (x, y) in enumerate(zip(F.asnumpy(a).flatten(), F.asnumpy(b).flatten())):
                    if not np.allclose(x, y):
                        print('@{} {} v.s. {}'.format(i, x, y))
    
            if not F.allclose(r1, r2):
                _print_error(r1, r2)
            assert F.allclose(r1, r2)
            if n_grad1 is not None or n_grad2 is not None:
                if not F.allclose(n_grad1, n_grad2):
                    print('node grad')
                    _print_error(n_grad1, n_grad2)
                assert(F.allclose(n_grad1, n_grad2))
    
        target = ["u", "v", "e"]
        for lhs, rhs in product(target, target):
            if lhs == rhs:
                continue
            for binary_op in ["add", "sub", "mul", "div", "dot"]:
                print(lhs, rhs, binary_op)
>               _test(lhs, rhs, binary_op)

tests/compute/test_apply_edges_hetero.py:242: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
tests/compute/test_apply_edges_hetero.py:190: in _test
    for rel in g.canonical_etypes]
tests/compute/test_apply_edges_hetero.py:190: in <listcomp>
    for rel in g.canonical_etypes]
python/dgl/heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python/dgl/core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python/dgl/ops/sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb73313e318>
op = 'add'
lhs_data = tensor([[-1.9796, -0.1490],
        [-1.0488, -0.7427]], requires_grad=True)
rhs_data = tensor([[-0.7523, -0.9474],
        [ 0.0539, -0.9849]], requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
u v add
___________________________ test_binary_op[idtype1] ____________________________

idtype = torch.int64

    @parametrize_idtype
    def test_binary_op(idtype):
        def _test(lhs, rhs, binary_op):
    
            g = create_test_heterograph(idtype)
    
            n1 = F.randn((g.num_nodes('user'), feat_size))
            n2 = F.randn((g.num_nodes('developer'), feat_size))
            n3 = F.randn((g.num_nodes('game'), feat_size))
    
            x1 = F.randn((g.num_edges('plays'),feat_size))
            x2 = F.randn((g.num_edges('follows'),feat_size))
            x3 = F.randn((g.num_edges('develops'),feat_size))
            x4 = F.randn((g.num_edges('wishes'),feat_size))
    
            builtin_msg_name = "{}_{}_{}".format(lhs, binary_op, rhs)
            builtin_msg = getattr(fn, builtin_msg_name)
    
            #################################################################
            #  apply_edges() is called on each relation type separately
            #################################################################
    
            F.attach_grad(n1)
            F.attach_grad(n2)
            F.attach_grad(n3)
            g.nodes['user'].data['h'] = n1
            g.nodes['developer'].data['h'] = n2
            g.nodes['game'].data['h'] = n3
            F.attach_grad(x1)
            F.attach_grad(x2)
            F.attach_grad(x3)
            F.attach_grad(x4)
            g['plays'].edata['h'] = x1
            g['follows'].edata['h'] = x2
            g['develops'].edata['h'] = x3
            g['wishes'].edata['h'] = x4
    
            with F.record_grad():
                [g.apply_edges(builtin_msg('h', 'h', 'm'), etype = rel)
                    for rel in g.canonical_etypes]
                r1 = g['plays'].edata['m']
                loss = F.sum(r1.view(-1), 0)
                F.backward(loss)
                n_grad1 = F.grad(g.nodes['game'].data['h'])
    
            #################################################################
            #  apply_edges() is called on all relation types
            #################################################################
    
            F.attach_grad(n1)
            F.attach_grad(n2)
            F.attach_grad(n3)
            g.nodes['user'].data['h'] = n1
            g.nodes['developer'].data['h'] = n2
            g.nodes['game'].data['h'] = n3
            F.attach_grad(x1)
            F.attach_grad(x2)
            F.attach_grad(x3)
            F.attach_grad(x4)
            g['plays'].edata['h'] = x1
            g['follows'].edata['h'] = x2
            g['develops'].edata['h'] = x3
            g['wishes'].edata['h'] = x4
    
            with F.record_grad():
                g.apply_edges(builtin_msg('h', 'h', 'm'))
                r2 = g['plays'].edata['m']
                loss = F.sum(r2.view(-1), 0)
                F.backward(loss)
                n_grad2 = F.grad(g.nodes['game'].data['h'])
            # correctness check
            def _print_error(a, b):
                for i, (x, y) in enumerate(zip(F.asnumpy(a).flatten(), F.asnumpy(b).flatten())):
                    if not np.allclose(x, y):
                        print('@{} {} v.s. {}'.format(i, x, y))
    
            if not F.allclose(r1, r2):
                _print_error(r1, r2)
            assert F.allclose(r1, r2)
            if n_grad1 is not None or n_grad2 is not None:
                if not F.allclose(n_grad1, n_grad2):
                    print('node grad')
                    _print_error(n_grad1, n_grad2)
                assert(F.allclose(n_grad1, n_grad2))
    
        target = ["u", "v", "e"]
        for lhs, rhs in product(target, target):
            if lhs == rhs:
                continue
            for binary_op in ["add", "sub", "mul", "div", "dot"]:
                print(lhs, rhs, binary_op)
>               _test(lhs, rhs, binary_op)

tests/compute/test_apply_edges_hetero.py:242: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
tests/compute/test_apply_edges_hetero.py:190: in _test
    for rel in g.canonical_etypes]
tests/compute/test_apply_edges_hetero.py:190: in <listcomp>
    for rel in g.canonical_etypes]
python/dgl/heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python/dgl/core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python/dgl/ops/sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72b9a2288>
op = 'add'
lhs_data = tensor([[-0.4311, -1.4427],
        [-1.3631, -0.8587]], requires_grad=True)
rhs_data = tensor([[-0.3154, -0.5503],
        [-0.4826, -0.9824]], requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
u v add
___________________________ test_issue_1088[idtype0] ___________________________

idtype = torch.int32

    @parametrize_idtype
    def test_issue_1088(idtype):
        # This test ensures that message passing on a heterograph with one edge type
        # would not crash (GitHub issue #1088).
        import dgl.function as fn
        g = dgl.heterograph({('U', 'E', 'V'): ([0, 1, 2], [1, 2, 3])}, idtype=idtype, device=F.ctx())
        g.nodes['U'].data['x'] = F.randn((3, 3))
>       g.update_all(fn.copy_u('x', 'm'), fn.sum('m', 'y'))

tests/compute/test_basics.py:646: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python/dgl/core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python/dgl/core.py:332: in invoke_gspmm
    z = op(graph, x)
python/dgl/ops/spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb732ce97e0>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.5470, -1.6287, -1.2363],
        [-1.1114,  1.4881,  0.0528],
        [ 0.9083, -0.6497,  0.6860]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
___________________________ test_issue_1088[idtype1] ___________________________

idtype = torch.int64

    @parametrize_idtype
    def test_issue_1088(idtype):
        # This test ensures that message passing on a heterograph with one edge type
        # would not crash (GitHub issue #1088).
        import dgl.function as fn
        g = dgl.heterograph({('U', 'E', 'V'): ([0, 1, 2], [1, 2, 3])}, idtype=idtype, device=F.ctx())
        g.nodes['U'].data['x'] = F.randn((3, 3))
>       g.update_all(fn.copy_u('x', 'm'), fn.sum('m', 'y'))

tests/compute/test_basics.py:646: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python/dgl/core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python/dgl/core.py:332: in invoke_gspmm
    z = op(graph, x)
python/dgl/ops/spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72b4f13a8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.4199, -0.0981,  2.1831],
        [ 2.6065, -0.8216, -0.6848],
        [-0.1195, -1.4392, -0.7790]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
___________________________ test_issue_2484[idtype0] ___________________________

idtype = torch.int32

    @parametrize_idtype
    def test_issue_2484(idtype):
        import dgl.function as fn
        g = dgl.graph(([0, 1, 2], [1, 2, 3]), idtype=idtype, device=F.ctx())
        x = F.copy_to(F.randn((4,)), F.ctx())
        g.ndata['x'] = x
>       g.pull([2, 1], fn.u_add_v('x', 'x', 'm'), fn.sum('m', 'x'))

tests/compute/test_basics.py:667: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/heterograph.py:4711: in pull
    compute_graph, message_func, reduce_func, apply_node_func)
python/dgl/core.py:362: in message_passing
    msgdata = invoke_gsddmm(g, mfunc)
python/dgl/core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python/dgl/ops/sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72bbcba68>
op = 'add', lhs_data = tensor([ 0.0074, -1.3295])
rhs_data = tensor([-1.3295,  1.4919]), lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
___________________________ test_issue_2484[idtype1] ___________________________

idtype = torch.int64

    @parametrize_idtype
    def test_issue_2484(idtype):
        import dgl.function as fn
        g = dgl.graph(([0, 1, 2], [1, 2, 3]), idtype=idtype, device=F.ctx())
        x = F.copy_to(F.randn((4,)), F.ctx())
        g.ndata['x'] = x
>       g.pull([2, 1], fn.u_add_v('x', 'x', 'm'), fn.sum('m', 'x'))

tests/compute/test_basics.py:667: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/heterograph.py:4711: in pull
    compute_graph, message_func, reduce_func, apply_node_func)
python/dgl/core.py:362: in message_passing
    msgdata = invoke_gsddmm(g, mfunc)
python/dgl/core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python/dgl/ops/sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72bc01d38>
op = 'add', lhs_data = tensor([ 1.0018, -0.3992])
rhs_data = tensor([-0.3992, -0.2018]), lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
_______________________ test_edge_softmax_unidirectional _______________________

    @unittest.skipIf(dgl.backend.backend_name != 'pytorch', reason='Only support PyTorch for now')
    def test_edge_softmax_unidirectional():
        g = dgl.heterograph({
            ('A', 'AB', 'B'): ([1,2,3,1,2,3,1,2,3],[0,0,0,1,1,1,2,2,2]),
            ('B', 'BB', 'B'): ([0,1,2,0,1,2,0,1,2], [0,0,0,1,1,1,2,2,2])})
        g = g.to(F.ctx())
        g.edges['AB'].data['x'] = F.ones(9) * 2
        g.edges['BB'].data['x'] = F.ones(9)
>       result = dgl.ops.edge_softmax(g, {'AB': g.edges['AB'].data['x'], 'BB': g.edges['BB'].data['x']})

tests/compute/test_edge_softmax_hetero.py:48: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/edge_softmax.py:144: in edge_softmax
    eids, norm_by, *logits_tuple)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb705a3c438>
eids = '__ALL__', norm_by = 'dst'
logits = (tensor([2., 2., 2., 2., 2., 2., 2., 2., 2.]), tensor([1., 1., 1., 1., 1., 1., 1., 1., 1.]))
args = (<dgl.heterograph_index.HeteroGraphIndex object at 0x7fb705a3c438>, '__ALL__', 'dst', tensor([2., 2., 2., 2., 2., 2., 2., 2., 2.]), tensor([1., 1., 1., 1., 1., 1., 1., 1., 1.]))

    def edge_softmax_hetero(gidx, eids=ALL, norm_by='dst', *logits):
        args = _cast_if_autocast_enabled(gidx, eids, norm_by, *logits)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:775: TypeError
______________________ test_edge_softmax[idtype0-src-g0] _______________________

g = Graph(num_nodes={'developer': 2, 'game': 2, 'user': 3},
      num_edges={('developer', 'develops', 'game'): 3, ('user'...('developer', 'game', 'develops'), ('user', 'user', 'follows'), ('user', 'game', 'plays'), ('user', 'game', 'wishes')])
norm_by = 'src', idtype = torch.int32

    @unittest.skipIf(dgl.backend.backend_name != 'pytorch', reason='Only support PyTorch for now')
    @pytest.mark.parametrize('g', get_cases(['clique']))
    @pytest.mark.parametrize('norm_by', ['src', 'dst'])
    # @pytest.mark.parametrize('shp', edge_softmax_shapes)
    @parametrize_idtype
    def test_edge_softmax(g, norm_by, idtype):
        print("params", norm_by, idtype)
    
        g = create_test_heterograph(idtype)
    
        x1 = F.randn((g.num_edges('plays'),feat_size))
        x2 = F.randn((g.num_edges('follows'),feat_size))
        x3 = F.randn((g.num_edges('develops'),feat_size))
        x4 = F.randn((g.num_edges('wishes'),feat_size))
    
        F.attach_grad(F.clone(x1))
        F.attach_grad(F.clone(x2))
        F.attach_grad(F.clone(x3))
        F.attach_grad(F.clone(x4))
    
        g['plays'].edata['eid'] = x1
        g['follows'].edata['eid'] = x2
        g['develops'].edata['eid'] = x3
        g['wishes'].edata['eid'] = x4
    
        #################################################################
        #  edge_softmax() on homogeneous graph
        #################################################################
    
        with F.record_grad():
            hm_g = dgl.to_homogeneous(g)
            hm_x = F.cat((x3, x2, x1, x4), 0)
            hm_e = F.attach_grad(F.clone(hm_x))
>           score_hm = edge_softmax(hm_g, hm_e, norm_by=norm_by)

tests/compute/test_edge_softmax_hetero.py:91: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/edge_softmax.py:135: in edge_softmax
    eids=eids, norm_by=norm_by)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb70599eca8>
logits = tensor([[-1.4336,  1.5559],
        [-0.3350,  0.0120],
        [ 1.5648,  0.8156],
        [-1.1123,  1.5411],
      ...057,  0.2814],
        [-0.4346,  0.2468],
        [-0.0544,  0.4573],
        [-0.6307, -1.2937]], requires_grad=True)
eids = '__ALL__', norm_by = 'src'

    def edge_softmax(gidx, logits, eids=ALL, norm_by='dst'):
        args = _cast_if_autocast_enabled(gidx, logits, eids, norm_by)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:770: TypeError
----------------------------- Captured stdout call -----------------------------
params src torch.int32
______________________ test_edge_softmax[idtype0-dst-g0] _______________________

g = Graph(num_nodes={'developer': 2, 'game': 2, 'user': 3},
      num_edges={('developer', 'develops', 'game'): 3, ('user'...('developer', 'game', 'develops'), ('user', 'user', 'follows'), ('user', 'game', 'plays'), ('user', 'game', 'wishes')])
norm_by = 'dst', idtype = torch.int32

    @unittest.skipIf(dgl.backend.backend_name != 'pytorch', reason='Only support PyTorch for now')
    @pytest.mark.parametrize('g', get_cases(['clique']))
    @pytest.mark.parametrize('norm_by', ['src', 'dst'])
    # @pytest.mark.parametrize('shp', edge_softmax_shapes)
    @parametrize_idtype
    def test_edge_softmax(g, norm_by, idtype):
        print("params", norm_by, idtype)
    
        g = create_test_heterograph(idtype)
    
        x1 = F.randn((g.num_edges('plays'),feat_size))
        x2 = F.randn((g.num_edges('follows'),feat_size))
        x3 = F.randn((g.num_edges('develops'),feat_size))
        x4 = F.randn((g.num_edges('wishes'),feat_size))
    
        F.attach_grad(F.clone(x1))
        F.attach_grad(F.clone(x2))
        F.attach_grad(F.clone(x3))
        F.attach_grad(F.clone(x4))
    
        g['plays'].edata['eid'] = x1
        g['follows'].edata['eid'] = x2
        g['develops'].edata['eid'] = x3
        g['wishes'].edata['eid'] = x4
    
        #################################################################
        #  edge_softmax() on homogeneous graph
        #################################################################
    
        with F.record_grad():
            hm_g = dgl.to_homogeneous(g)
            hm_x = F.cat((x3, x2, x1, x4), 0)
            hm_e = F.attach_grad(F.clone(hm_x))
>           score_hm = edge_softmax(hm_g, hm_e, norm_by=norm_by)

tests/compute/test_edge_softmax_hetero.py:91: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/edge_softmax.py:135: in edge_softmax
    eids=eids, norm_by=norm_by)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb705a02b88>
logits = tensor([[ 0.4611, -1.6093],
        [ 0.0317,  0.8821],
        [ 1.5017, -0.5792],
        [ 1.9286,  1.7655],
      ...343,  0.1096],
        [ 0.5236, -0.1736],
        [ 0.6223, -0.6720],
        [-1.5531,  0.9349]], requires_grad=True)
eids = '__ALL__', norm_by = 'dst'

    def edge_softmax(gidx, logits, eids=ALL, norm_by='dst'):
        args = _cast_if_autocast_enabled(gidx, logits, eids, norm_by)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:770: TypeError
----------------------------- Captured stdout call -----------------------------
params dst torch.int32
______________________ test_edge_softmax[idtype1-src-g0] _______________________

g = Graph(num_nodes={'developer': 2, 'game': 2, 'user': 3},
      num_edges={('developer', 'develops', 'game'): 3, ('user'...('developer', 'game', 'develops'), ('user', 'user', 'follows'), ('user', 'game', 'plays'), ('user', 'game', 'wishes')])
norm_by = 'src', idtype = torch.int64

    @unittest.skipIf(dgl.backend.backend_name != 'pytorch', reason='Only support PyTorch for now')
    @pytest.mark.parametrize('g', get_cases(['clique']))
    @pytest.mark.parametrize('norm_by', ['src', 'dst'])
    # @pytest.mark.parametrize('shp', edge_softmax_shapes)
    @parametrize_idtype
    def test_edge_softmax(g, norm_by, idtype):
        print("params", norm_by, idtype)
    
        g = create_test_heterograph(idtype)
    
        x1 = F.randn((g.num_edges('plays'),feat_size))
        x2 = F.randn((g.num_edges('follows'),feat_size))
        x3 = F.randn((g.num_edges('develops'),feat_size))
        x4 = F.randn((g.num_edges('wishes'),feat_size))
    
        F.attach_grad(F.clone(x1))
        F.attach_grad(F.clone(x2))
        F.attach_grad(F.clone(x3))
        F.attach_grad(F.clone(x4))
    
        g['plays'].edata['eid'] = x1
        g['follows'].edata['eid'] = x2
        g['develops'].edata['eid'] = x3
        g['wishes'].edata['eid'] = x4
    
        #################################################################
        #  edge_softmax() on homogeneous graph
        #################################################################
    
        with F.record_grad():
            hm_g = dgl.to_homogeneous(g)
            hm_x = F.cat((x3, x2, x1, x4), 0)
            hm_e = F.attach_grad(F.clone(hm_x))
>           score_hm = edge_softmax(hm_g, hm_e, norm_by=norm_by)

tests/compute/test_edge_softmax_hetero.py:91: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/edge_softmax.py:135: in edge_softmax
    eids=eids, norm_by=norm_by)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb705aa26c0>
logits = tensor([[-3.0921e-01,  8.0471e-01],
        [ 1.6534e+00, -3.8585e-01],
        [-9.6037e-02,  1.2391e+00],
        [ ... 1.3756e-03,  1.1694e+00],
        [ 3.5168e-01, -1.3745e+00],
        [ 1.9409e-01,  1.4674e+00]], requires_grad=True)
eids = '__ALL__', norm_by = 'src'

    def edge_softmax(gidx, logits, eids=ALL, norm_by='dst'):
        args = _cast_if_autocast_enabled(gidx, logits, eids, norm_by)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:770: TypeError
----------------------------- Captured stdout call -----------------------------
params src torch.int64
______________________ test_edge_softmax[idtype1-dst-g0] _______________________

g = Graph(num_nodes={'developer': 2, 'game': 2, 'user': 3},
      num_edges={('developer', 'develops', 'game'): 3, ('user'...('developer', 'game', 'develops'), ('user', 'user', 'follows'), ('user', 'game', 'plays'), ('user', 'game', 'wishes')])
norm_by = 'dst', idtype = torch.int64

    @unittest.skipIf(dgl.backend.backend_name != 'pytorch', reason='Only support PyTorch for now')
    @pytest.mark.parametrize('g', get_cases(['clique']))
    @pytest.mark.parametrize('norm_by', ['src', 'dst'])
    # @pytest.mark.parametrize('shp', edge_softmax_shapes)
    @parametrize_idtype
    def test_edge_softmax(g, norm_by, idtype):
        print("params", norm_by, idtype)
    
        g = create_test_heterograph(idtype)
    
        x1 = F.randn((g.num_edges('plays'),feat_size))
        x2 = F.randn((g.num_edges('follows'),feat_size))
        x3 = F.randn((g.num_edges('develops'),feat_size))
        x4 = F.randn((g.num_edges('wishes'),feat_size))
    
        F.attach_grad(F.clone(x1))
        F.attach_grad(F.clone(x2))
        F.attach_grad(F.clone(x3))
        F.attach_grad(F.clone(x4))
    
        g['plays'].edata['eid'] = x1
        g['follows'].edata['eid'] = x2
        g['develops'].edata['eid'] = x3
        g['wishes'].edata['eid'] = x4
    
        #################################################################
        #  edge_softmax() on homogeneous graph
        #################################################################
    
        with F.record_grad():
            hm_g = dgl.to_homogeneous(g)
            hm_x = F.cat((x3, x2, x1, x4), 0)
            hm_e = F.attach_grad(F.clone(hm_x))
>           score_hm = edge_softmax(hm_g, hm_e, norm_by=norm_by)

tests/compute/test_edge_softmax_hetero.py:91: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/edge_softmax.py:135: in edge_softmax
    eids=eids, norm_by=norm_by)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb705a31510>
logits = tensor([[-0.1608,  0.2121],
        [-0.4321,  0.3399],
        [-0.0639, -0.9893],
        [ 0.6717,  0.0155],
      ...806, -0.8436],
        [ 0.5347,  0.4538],
        [-0.4843,  0.5824],
        [-2.2155, -0.7792]], requires_grad=True)
eids = '__ALL__', norm_by = 'dst'

    def edge_softmax(gidx, logits, eids=ALL, norm_by='dst'):
        args = _cast_if_autocast_enabled(gidx, logits, eids, norm_by)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:770: TypeError
----------------------------- Captured stdout call -----------------------------
params dst torch.int64
____________________________ test_updates[idtype0] _____________________________

idtype = torch.int32

    @parametrize_idtype
    def test_updates(idtype):
        def msg_func(edges):
            return {'m': edges.src['h']}
        def reduce_func(nodes):
            return {'y': F.sum(nodes.mailbox['m'], 1)}
        def apply_func(nodes):
            return {'y': nodes.data['y'] * 2}
        g = create_test_heterograph(idtype)
        x = F.randn((3, 5))
        g.nodes['user'].data['h'] = x
    
        for msg, red, apply in itertools.product(
                [fn.copy_u('h', 'm'), msg_func], [fn.sum('m', 'y'), reduce_func],
                [None, apply_func]):
            multiplier = 1 if apply is None else 2
    
>           g['user', 'plays', 'game'].update_all(msg, red, apply)

tests/compute/test_heterograph.py:1674: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python/dgl/core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python/dgl/core.py:332: in invoke_gspmm
    z = op(graph, x)
python/dgl/ops/spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb7059b3ca8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.1249,  0.5763,  1.4886,  2.4238, -1.5383],
        [-1.4728, -0.9454,  0.7629, -1.1135, -0.2008],
        [-0.1803,  0.4783,  0.1011,  1.6278, -0.1912]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
____________________________ test_updates[idtype1] _____________________________

idtype = torch.int64

    @parametrize_idtype
    def test_updates(idtype):
        def msg_func(edges):
            return {'m': edges.src['h']}
        def reduce_func(nodes):
            return {'y': F.sum(nodes.mailbox['m'], 1)}
        def apply_func(nodes):
            return {'y': nodes.data['y'] * 2}
        g = create_test_heterograph(idtype)
        x = F.randn((3, 5))
        g.nodes['user'].data['h'] = x
    
        for msg, red, apply in itertools.product(
                [fn.copy_u('h', 'm'), msg_func], [fn.sum('m', 'y'), reduce_func],
                [None, apply_func]):
            multiplier = 1 if apply is None else 2
    
>           g['user', 'plays', 'game'].update_all(msg, red, apply)

tests/compute/test_heterograph.py:1674: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python/dgl/core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python/dgl/core.py:332: in invoke_gspmm
    z = op(graph, x)
python/dgl/ops/spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43e56a798>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.4010,  0.5922, -0.0199,  1.0543, -1.0379],
        [-1.5216, -0.4269,  0.1716,  1.6454,  0.7565],
        [ 0.9748,  1.6111,  1.3588,  0.0985, -0.3508]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
____________________________ test_backward[idtype0] ____________________________

idtype = torch.int32

    @parametrize_idtype
    def test_backward(idtype):
        g = create_test_heterograph(idtype)
        x = F.randn((3, 5))
        F.attach_grad(x)
        g.nodes['user'].data['h'] = x
        with F.record_grad():
            g.multi_update_all(
                {'plays' : (fn.copy_u('h', 'm'), fn.sum('m', 'y')),
                 'wishes': (fn.copy_u('h', 'm'), fn.sum('m', 'y'))},
>               'sum')

tests/compute/test_heterograph.py:1709: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/heterograph.py:5023: in multi_update_all
    all_out[dtid].append(core.message_passing(g, mfunc, rfunc, afunc))
python/dgl/core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python/dgl/core.py:332: in invoke_gspmm
    z = op(graph, x)
python/dgl/ops/spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72c239630>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.6385, -1.2521,  0.4248, -2.3375, -0.6380],
        [ 0.0338,  1.4721, -0.7206, -0.1300,  0.5931],
        [ 0.7461,  1.6199, -0.7145, -1.2436, -0.7472]], requires_grad=True)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
____________________________ test_backward[idtype1] ____________________________

idtype = torch.int64

    @parametrize_idtype
    def test_backward(idtype):
        g = create_test_heterograph(idtype)
        x = F.randn((3, 5))
        F.attach_grad(x)
        g.nodes['user'].data['h'] = x
        with F.record_grad():
            g.multi_update_all(
                {'plays' : (fn.copy_u('h', 'm'), fn.sum('m', 'y')),
                 'wishes': (fn.copy_u('h', 'm'), fn.sum('m', 'y'))},
>               'sum')

tests/compute/test_heterograph.py:1709: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/heterograph.py:5023: in multi_update_all
    all_out[dtid].append(core.message_passing(g, mfunc, rfunc, afunc))
python/dgl/core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python/dgl/core.py:332: in invoke_gspmm
    z = op(graph, x)
python/dgl/ops/spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43e528168>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.4332, -1.3325,  0.0030,  0.7070, -1.5994],
        [-2.6495,  2.6074,  1.6508,  0.9886,  2.4109],
        [-0.1265,  0.5570, -1.8708,  0.1544, -0.5959]], requires_grad=True)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
_____________________________ test_copy_src_reduce _____________________________

    def test_copy_src_reduce():
        def _test(red, partial):
            g = dgl.DGLGraph(nx.erdos_renyi_graph(100, 0.1))
            # NOTE(zihao): add self-loop to avoid zero-degree nodes.
            # https://github.com/dmlc/dgl/issues/761
            g.add_edges(g.nodes(), g.nodes())
            g = g.to(F.ctx())
            hu, hv, he = generate_feature(g, 'none', 'none')
            if partial:
                nid = F.tensor(list(range(0, 100, 2)), g.idtype)
    
            g.ndata['u'] = F.attach_grad(F.clone(hu))
            g.ndata['v'] = F.attach_grad(F.clone(hv))
            g.edata['e'] = F.attach_grad(F.clone(he))
    
            with F.record_grad():
                if partial:
                    g.pull(nid, fn.copy_src(src='u', out='m'),
                           builtin[red](msg='m', out='r1'))
                else:
                    g.update_all(fn.copy_src(src='u', out='m'),
                                 builtin[red](msg='m', out='r1'))
                r1 = g.ndata['r1']
                F.backward(F.reduce_sum(r1))
                n_grad1 = F.grad(g.ndata['u'])
    
            # reset grad
            g.ndata['u'] = F.attach_grad(F.clone(hu))
            g.ndata['v'] = F.attach_grad(F.clone(hv))
            g.edata['e'] = F.attach_grad(F.clone(he))
    
            with F.record_grad():
                if partial:
                    g.pull(nid, udf_copy_src, udf_reduce[red])
                else:
                    g.update_all(udf_copy_src, udf_reduce[red])
                r2 = g.ndata['r2']
                F.backward(F.reduce_sum(r2))
                n_grad2 = F.grad(g.ndata['u'])
    
            def _print_error(a, b):
                print("ERROR: Test copy_src_{} partial: {}".
                      format(red, partial))
                for i, (x, y) in enumerate(zip(F.asnumpy(a).flatten(), F.asnumpy(b).flatten())):
                    if not np.allclose(x, y):
                        print('@{} {} v.s. {}'.format(i, x, y))
    
            if not F.allclose(r1, r2):
                _print_error(r1, r2)
            assert F.allclose(r1, r2)
            if not F.allclose(n_grad1, n_grad2):
                print('node grad')
                _print_error(n_grad1, n_grad2)
            assert(F.allclose(n_grad1, n_grad2))
    
>       _test('sum', False)

tests/compute/test_kernel.py:134: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
tests/compute/test_kernel.py:100: in _test
    builtin[red](msg='m', out='r1'))
python/dgl/heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python/dgl/core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python/dgl/core.py:332: in invoke_gspmm
    z = op(graph, x)
python/dgl/ops/spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb732cd5990>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[[[-0.4279,  0.9162,  0.5406,  0.9737],
          [-0.5837, -0.7262,  0.8167, -0.8627],
          [-0.8493,  0...
          [-0.5046, -0.1275,  0.9672,  0.7794],
          [ 0.3485, -0.1620,  0.4156,  0.4446]]]], requires_grad=True)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
____________________________ test_copy_edge_reduce _____________________________

    def test_copy_edge_reduce():
        def _test(red, partial):
            g = dgl.DGLGraph(nx.erdos_renyi_graph(100, 0.1))
            # NOTE(zihao): add self-loop to avoid zero-degree nodes.
            g.add_edges(g.nodes(), g.nodes())
            g = g.to(F.ctx())
            hu, hv, he = generate_feature(g, 'none', 'none')
            if partial:
                nid = F.tensor(list(range(0, 100, 2)), g.idtype)
    
            g.ndata['u'] = F.attach_grad(F.clone(hu))
            g.ndata['v'] = F.attach_grad(F.clone(hv))
            g.edata['e'] = F.attach_grad(F.clone(he))
    
            with F.record_grad():
                if partial:
                    g.pull(nid, fn.copy_edge(edge='e', out='m'),
                           builtin[red](msg='m', out='r1'))
                else:
                    g.update_all(fn.copy_edge(edge='e', out='m'),
                                 builtin[red](msg='m', out='r1'))
                r1 = g.ndata['r1']
                F.backward(F.reduce_sum(r1))
                e_grad1 = F.grad(g.edata['e'])
    
            # reset grad
            g.ndata['u'] = F.attach_grad(F.clone(hu))
            g.ndata['v'] = F.attach_grad(F.clone(hv))
            g.edata['e'] = F.attach_grad(F.clone(he))
    
            with F.record_grad():
                if partial:
                    g.pull(nid, udf_copy_edge, udf_reduce[red])
                else:
                    g.update_all(udf_copy_edge, udf_reduce[red])
                r2 = g.ndata['r2']
                F.backward(F.reduce_sum(r2))
                e_grad2 = F.grad(g.edata['e'])
    
            def _print_error(a, b):
                print("ERROR: Test copy_edge_{} partial: {}".
                      format(red, partial))
                return
                for i, (x, y) in enumerate(zip(F.asnumpy(a).flatten(), F.asnumpy(b).flatten())):
                    if not np.allclose(x, y):
                        print('@{} {} v.s. {}'.format(i, x, y))
    
            if not F.allclose(r1, r2):
                _print_error(r1, r2)
            assert F.allclose(r1, r2)
            if not F.allclose(e_grad1, e_grad2):
                print('edge gradient')
                _print_error(e_grad1, e_grad2)
            assert(F.allclose(e_grad1, e_grad2))
    
>       _test('sum', False)

tests/compute/test_kernel.py:197: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
tests/compute/test_kernel.py:162: in _test
    builtin[red](msg='m', out='r1'))
python/dgl/heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python/dgl/core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python/dgl/core.py:332: in invoke_gspmm
    z = op(graph, x)
python/dgl/ops/spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72c0741b0>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([[[[ 4.6492e-05,  7.2036e-02,  3.4002e-01,  6.9137e-01],
          [ 2.1845e-01,  5.2046e-01, -6.6986e-01,  3.1...1.2087e-01, -2.6184e-01],
          [-5.5105e-01,  6.1630e-01, -8.6822e-01,  5.0840e-01]]]],
       requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
___________________________ test_all_binary_builtins ___________________________

    def test_all_binary_builtins():
        def _test(g, lhs, rhs, binary_op, reducer, partial, nid, broadcast='none'):
            # initialize node/edge features with uniform(-1, 1)
            hu, hv, he = generate_feature(g, broadcast, binary_op)
            if binary_op == 'div':
                # op = div
                # lhs range: [-1, 1]
                # rhs range: [1, 2]
                # result range: [-1, 1]
                if rhs == 'u':
                    hu = (hu + 3) / 2
                elif rhs == 'v':
                    hv = (hv + 3) / 2
                elif rhs == 'e':
                    he = (he + 3) / 2
    
            if binary_op == 'add' or binary_op == 'sub':
                # op = add, sub
                # lhs range: [-1/2, 1/2]
                # rhs range: [-1/2, 1/2]
                # result range: [-1, 1]
                hu = hu / 2
                hv = hv / 2
                he = he / 2
    
            g.ndata['u'] = F.attach_grad(F.clone(hu))
            g.ndata['v'] = F.attach_grad(F.clone(hv))
            g.edata['e'] = F.attach_grad(F.clone(he))
    
            builtin_msg_name = "{}_{}_{}".format(lhs, binary_op, rhs)
            builtin_msg = getattr(fn, builtin_msg_name)
            builtin_red = getattr(fn, reducer)
    
            def target_feature_switch(g, target):
                if target == "u":
                    return g.ndata["u"]
                elif target == "v":
                    return g.ndata["v"]
                else:
                    return g.edata["e"]
    
            with F.record_grad():
                if partial:
                    g.pull(nid, builtin_msg(lhs, rhs, 'm'), builtin_red('m', 'r1'))
                else:
                    g.update_all(builtin_msg(lhs, rhs, 'm'), builtin_red('m', 'r1'))
                r1 = g.ndata.pop('r1')
                F.backward(F.reduce_sum(r1))
                lhs_grad_1 = F.grad(target_feature_switch(g, lhs))
                rhs_grad_1 = F.grad(target_feature_switch(g, rhs))
    
            # reset grad
            g.ndata['u'] = F.attach_grad(F.clone(hu))
            g.ndata['v'] = F.attach_grad(F.clone(hv))
            g.edata['e'] = F.attach_grad(F.clone(he))
    
            def target_switch(edges, target):
                if target == "u":
                    return edges.src
                elif target == "v":
                    return edges.dst
                elif target == "e":
                    return edges.data
                else:
                    assert(0), "Unknown target {}".format(target)
    
            def mfunc(edges):
                op = getattr(F, binary_op)
                lhs_data = target_switch(edges, lhs)[lhs]
                rhs_data = target_switch(edges, rhs)[rhs]
                # NOTE(zihao): we need to do batched broadcast
                # e.g. (68, 3, 1) op (68, 5, 3, 4)
                while F.ndim(lhs_data) < F.ndim(rhs_data):
                    lhs_data = F.unsqueeze(lhs_data, 1)
                while F.ndim(rhs_data) < F.ndim(lhs_data):
                    rhs_data = F.unsqueeze(rhs_data, 1)
                return {"m": op(lhs_data, rhs_data)}
    
            def rfunc(nodes):
                op = getattr(F, reducer)
                return {"r2": op(nodes.mailbox['m'], 1)}
    
            with F.record_grad():
                if partial:
                    g.pull(nid, mfunc, rfunc)
                else:
                    g.update_all(mfunc, rfunc)
                r2 = g.ndata.pop('r2')
                F.backward(F.reduce_sum(r2), F.tensor([1.]))
                lhs_grad_2 = F.grad(target_feature_switch(g, lhs))
                rhs_grad_2 = F.grad(target_feature_switch(g, rhs))
    
            rtol = 1e-4
            atol = 1e-4
    
            def _print_error(a, b):
                print("ERROR: Test {}_{}_{}_{} broadcast: {} partial: {}".
                      format(lhs, binary_op, rhs, reducer, broadcast, partial))
                return
                if lhs == 'u':
                    lhs_data = hu
                elif lhs == 'v':
                    lhs_data = hv
                elif lhs == 'e':
                    lhs_data = he
    
                if rhs == 'u':
                    rhs_data = hu
                elif rhs == 'v':
                    rhs_data = hv
                elif rhs == 'e':
                    rhs_data = he
                print("lhs", F.asnumpy(lhs_data).tolist())
                print("rhs", F.asnumpy(rhs_data).tolist())
                for i, (x, y) in enumerate(zip(F.asnumpy(a).flatten(), F.asnumpy(b).flatten())):
                    if not np.allclose(x, y, rtol, atol):
                        print('@{} {} v.s. {}'.format(i, x, y))
    
            if not F.allclose(r1, r2, rtol, atol):
                _print_error(r1, r2)
            assert F.allclose(r1, r2, rtol, atol)
    
            if not F.allclose(lhs_grad_1, lhs_grad_2, rtol, atol):
                print("left grad")
                _print_error(lhs_grad_1, lhs_grad_2)
            assert(F.allclose(lhs_grad_1, lhs_grad_2, rtol, atol))
    
            if not F.allclose(rhs_grad_1, rhs_grad_2, rtol, atol):
                print("right grad")
                _print_error(rhs_grad_1, rhs_grad_2)
            assert(F.allclose(rhs_grad_1, rhs_grad_2, rtol, atol))
    
        g = dgl.DGLGraph()
        g.add_nodes(20)
        # NOTE(zihao): add self-loop to avoid zero-degree nodes.
        g.add_edges(g.nodes(), g.nodes())
        for i in range(2, 18):
            g.add_edge(0, i)
            g.add_edge(1, i)
            g.add_edge(i, 18)
            g.add_edge(i, 19)
        g.add_edge(18, 0)
        g.add_edge(18, 1)
        g.add_edge(19, 0)
        g.add_edge(19, 1)
        g = g.to(F.ctx())
        nid = F.tensor([0, 1, 4, 5, 7, 12, 14, 15, 18, 19], g.idtype)
        target = ["u", "v", "e"]
    
        for lhs, rhs in product(target, target):
            if lhs == rhs:
                continue
            for binary_op in ["add", "sub", "mul", "div"]:
                for reducer in ["sum", "max", "min", "mean"]:
                    for broadcast in ["none", lhs, rhs]:
                        for partial in [False, True]:
                            print(lhs, rhs, binary_op, reducer, broadcast, partial)
                            _test(g, lhs, rhs, binary_op, reducer, partial, nid,
>                                 broadcast=broadcast)

tests/compute/test_kernel.py:363: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
tests/compute/test_kernel.py:250: in _test
    g.update_all(builtin_msg(lhs, rhs, 'm'), builtin_red('m', 'r1'))
python/dgl/heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python/dgl/core.py:362: in message_passing
    msgdata = invoke_gsddmm(g, mfunc)
python/dgl/core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python/dgl/ops/sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb7058d5240>
op = 'add'
lhs_data = tensor([[[[-0.2139,  0.4581,  0.2703,  0.4869],
          [-0.2918, -0.3631,  0.4084, -0.4314],
          [-0.4247,  0...
          [ 0.2560,  0.2322,  0.2193, -0.4617],
          [ 0.3370, -0.1000,  0.4172,  0.3279]]]], requires_grad=True)
rhs_data = tensor([[[[-2.4563e-01, -4.7115e-02, -3.0174e-01, -3.5703e-01],
          [ 4.8955e-01, -4.1437e-01,  1.3228e-01, -1.4...5.2199e-02, -3.8938e-01],
          [ 1.5332e-01,  5.9728e-02, -6.3271e-02, -3.7373e-01]]]],
       requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
u v add sum none False
______________________ test_mean_zero_degree[g0-idtype0] _______________________

g = Graph(num_nodes=5, num_edges=6,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int32

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo-zero-degree']))
    def test_mean_zero_degree(g, idtype):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.ones((g.number_of_nodes(), 3))
>       g.update_all(fn.copy_u('h', 'm'), fn.mean('m', 'x'))

tests/compute/test_kernel.py:370: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python/dgl/core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python/dgl/core.py:332: in invoke_gspmm
    z = op(graph, x)
python/dgl/ops/spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43eb61b40>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[1., 1., 1.],
        [1., 1., 1.],
        [1., 1., 1.],
        [1., 1., 1.],
        [1., 1., 1.]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
______________________ test_mean_zero_degree[g0-idtype1] _______________________

g = Graph(num_nodes=5, num_edges=6,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32)}
      edata_schemes={})
idtype = torch.int64

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo-zero-degree']))
    def test_mean_zero_degree(g, idtype):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.ones((g.number_of_nodes(), 3))
>       g.update_all(fn.copy_u('h', 'm'), fn.mean('m', 'x'))

tests/compute/test_kernel.py:370: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python/dgl/core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python/dgl/core.py:332: in invoke_gspmm
    z = op(graph, x)
python/dgl/ops/spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43e8dd9d8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[1., 1., 1.],
        [1., 1., 1.],
        [1., 1., 1.],
        [1., 1., 1.],
        [1., 1., 1.]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
__________________________ test_unary_copy_u[idtype0] __________________________

idtype = torch.int32

    @parametrize_idtype
    def test_unary_copy_u(idtype):
        def _test(mfunc, rfunc):
            g = create_test_heterograph_2(idtype)
            g0 = create_test_heterograph(idtype)
            g1 = create_test_heterograph_large(idtype)
            cross_reducer = rfunc.__name__
            x1 = F.randn((g.num_nodes('user'), feat_size))
            x2 = F.randn((g.num_nodes('developer'), feat_size))
            F.attach_grad(x1)
            F.attach_grad(x2)
            g.nodes['user'].data['h'] = x1
            g.nodes['developer'].data['h'] = x2
    
            #################################################################
            #  multi_update_all(): call msg_passing separately for each etype
            #################################################################
    
            with F.record_grad():
                g.multi_update_all(
                    {etype : (mfunc('h', 'm'), rfunc('m', 'y'))
                        for etype in g.canonical_etypes},
                    cross_reducer)
                r1 = g.nodes['game'].data['y'].clone()
                r2 = g.nodes['user'].data['y'].clone()
                r3 = g.nodes['player'].data['y'].clone()
                loss = r1.sum() + r2.sum() + r3.sum()
                F.backward(loss)
                n_grad1 = F.grad(g.nodes['user'].data['h']).clone()
                n_grad2 = F.grad(g.nodes['developer'].data['h']).clone()
    
            g.nodes['user'].data.clear()
            g.nodes['developer'].data.clear()
            g.nodes['game'].data.clear()
            g.nodes['player'].data.clear()
    
            #################################################################
            #  update_all(): call msg_passing for all etypes
            #################################################################
    
            F.attach_grad(x1)
            F.attach_grad(x2)
            g.nodes['user'].data['h'] = x1
            g.nodes['developer'].data['h'] = x2
    
            with F.record_grad():
                g.update_all(mfunc('h', 'm'), rfunc('m', 'y'))
                r4 = g.nodes['game'].data['y']
                r5 = g.nodes['user'].data['y']
                r6 = g.nodes['player'].data['y']
                loss = r4.sum() + r5.sum() + r6.sum()
                F.backward(loss)
                n_grad3 = F.grad(g.nodes['user'].data['h'])
                n_grad4 = F.grad(g.nodes['developer'].data['h'])
    
            assert F.allclose(r1, r4)
            assert F.allclose(r2, r5)
            assert F.allclose(r3, r6)
            assert(F.allclose(n_grad1, n_grad3))
            assert(F.allclose(n_grad2, n_grad4))
>       _test(fn.copy_u, fn.sum)

tests/compute/test_new_update_all_hetero.py:132: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
tests/compute/test_new_update_all_hetero.py:94: in _test
    cross_reducer)
python/dgl/heterograph.py:5023: in multi_update_all
    all_out[dtid].append(core.message_passing(g, mfunc, rfunc, afunc))
python/dgl/core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python/dgl/core.py:332: in invoke_gspmm
    z = op(graph, x)
python/dgl/ops/spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb7058fe0d8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.2540e+00, -9.2633e-01],
        [-1.7815e+00, -1.5279e+00],
        [-7.5881e-01,  9.6034e-01],
        [ ...-1.6601e+00, -1.6416e+00],
        [-1.4363e+00,  4.7783e-01],
        [ 1.0850e+00,  1.3749e-01]], requires_grad=True)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
__________________________ test_unary_copy_u[idtype1] __________________________

idtype = torch.int64

    @parametrize_idtype
    def test_unary_copy_u(idtype):
        def _test(mfunc, rfunc):
            g = create_test_heterograph_2(idtype)
            g0 = create_test_heterograph(idtype)
            g1 = create_test_heterograph_large(idtype)
            cross_reducer = rfunc.__name__
            x1 = F.randn((g.num_nodes('user'), feat_size))
            x2 = F.randn((g.num_nodes('developer'), feat_size))
            F.attach_grad(x1)
            F.attach_grad(x2)
            g.nodes['user'].data['h'] = x1
            g.nodes['developer'].data['h'] = x2
    
            #################################################################
            #  multi_update_all(): call msg_passing separately for each etype
            #################################################################
    
            with F.record_grad():
                g.multi_update_all(
                    {etype : (mfunc('h', 'm'), rfunc('m', 'y'))
                        for etype in g.canonical_etypes},
                    cross_reducer)
                r1 = g.nodes['game'].data['y'].clone()
                r2 = g.nodes['user'].data['y'].clone()
                r3 = g.nodes['player'].data['y'].clone()
                loss = r1.sum() + r2.sum() + r3.sum()
                F.backward(loss)
                n_grad1 = F.grad(g.nodes['user'].data['h']).clone()
                n_grad2 = F.grad(g.nodes['developer'].data['h']).clone()
    
            g.nodes['user'].data.clear()
            g.nodes['developer'].data.clear()
            g.nodes['game'].data.clear()
            g.nodes['player'].data.clear()
    
            #################################################################
            #  update_all(): call msg_passing for all etypes
            #################################################################
    
            F.attach_grad(x1)
            F.attach_grad(x2)
            g.nodes['user'].data['h'] = x1
            g.nodes['developer'].data['h'] = x2
    
            with F.record_grad():
                g.update_all(mfunc('h', 'm'), rfunc('m', 'y'))
                r4 = g.nodes['game'].data['y']
                r5 = g.nodes['user'].data['y']
                r6 = g.nodes['player'].data['y']
                loss = r4.sum() + r5.sum() + r6.sum()
                F.backward(loss)
                n_grad3 = F.grad(g.nodes['user'].data['h'])
                n_grad4 = F.grad(g.nodes['developer'].data['h'])
    
            assert F.allclose(r1, r4)
            assert F.allclose(r2, r5)
            assert F.allclose(r3, r6)
            assert(F.allclose(n_grad1, n_grad3))
            assert(F.allclose(n_grad2, n_grad4))
>       _test(fn.copy_u, fn.sum)

tests/compute/test_new_update_all_hetero.py:132: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
tests/compute/test_new_update_all_hetero.py:94: in _test
    cross_reducer)
python/dgl/heterograph.py:5023: in multi_update_all
    all_out[dtid].append(core.message_passing(g, mfunc, rfunc, afunc))
python/dgl/core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python/dgl/core.py:332: in invoke_gspmm
    z = op(graph, x)
python/dgl/ops/spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb705a11870>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 8.7427e-01,  9.6819e-01],
        [ 1.1609e+00, -9.4198e-01],
        [-2.0202e-01,  1.3531e+00],
        [-... 8.8002e-01,  5.7814e-01],
        [-1.3564e+00, -3.6910e-01],
        [ 8.1194e-01, -1.6130e+00]], requires_grad=True)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
__________________________ test_unary_copy_e[idtype0] __________________________

idtype = torch.int32

    @parametrize_idtype
    def test_unary_copy_e(idtype):
        def _test(mfunc, rfunc):
    
            g = create_test_heterograph_large(idtype)
            g0 = create_test_heterograph_2(idtype)
            g1 = create_test_heterograph(idtype)
            cross_reducer = rfunc.__name__
            x1 = F.randn((g.num_edges('plays'),feat_size))
            x2 = F.randn((g.num_edges('follows'),feat_size))
            x3 = F.randn((g.num_edges('develops'),feat_size))
            x4 = F.randn((g.num_edges('wishes'),feat_size))
            F.attach_grad(x1)
            F.attach_grad(x2)
            F.attach_grad(x3)
            F.attach_grad(x4)
            g['plays'].edata['eid'] = x1
            g['follows'].edata['eid'] = x2
            g['develops'].edata['eid'] = x3
            g['wishes'].edata['eid'] = x4
    
            #################################################################
            #  multi_update_all(): call msg_passing separately for each etype
            #################################################################
    
            with F.record_grad():
                g.multi_update_all(
                    {'plays' : (mfunc('eid', 'm'), rfunc('m', 'y')),
                    'follows': (mfunc('eid', 'm'), rfunc('m', 'y')),
                    'develops': (mfunc('eid', 'm'), rfunc('m', 'y')),
                    'wishes': (mfunc('eid', 'm'), rfunc('m', 'y'))},
                    cross_reducer)
                r1 = g.nodes['game'].data['y'].clone()
                r2 = g.nodes['user'].data['y'].clone()
                loss = r1.sum() + r2.sum()
                F.backward(loss)
                e_grad1 = F.grad(g['develops'].edata['eid']).clone()
                e_grad2 = F.grad(g['plays'].edata['eid']).clone()
                e_grad3 = F.grad(g['wishes'].edata['eid']).clone()
                e_grad4 = F.grad(g['follows'].edata['eid']).clone()
            {etype : (g[etype].edata.clear())
                for _, etype, _ in g.canonical_etypes},
    
            #################################################################
            #  update_all(): call msg_passing for all etypes
            #################################################################
    
            # TODO(Israt): output type can be None in multi_update and empty
            F.attach_grad(x1)
            F.attach_grad(x2)
            F.attach_grad(x3)
            F.attach_grad(x4)
    
            g['plays'].edata['eid'] = x1
            g['follows'].edata['eid'] = x2
            g['develops'].edata['eid'] = x3
            g['wishes'].edata['eid'] = x4
    
            with F.record_grad():
                g.update_all(mfunc('eid', 'm'), rfunc('m', 'y'))
                r3 = g.nodes['game'].data['y']
                r4 = g.nodes['user'].data['y']
                loss = r3.sum() + r4.sum()
                F.backward(loss)
                e_grad5 = F.grad(g['develops'].edata['eid'])
                e_grad6 = F.grad(g['plays'].edata['eid'])
                e_grad7 = F.grad(g['wishes'].edata['eid'])
                e_grad8 = F.grad(g['follows'].edata['eid'])
            # # correctness check
            def _print_error(a, b):
                for i, (x, y) in enumerate(zip(F.asnumpy(a).flatten(), F.asnumpy(b).flatten())):
                    if not np.allclose(x, y):
                       print('@{} {} v.s. {}'.format(i, x, y))
    
            assert F.allclose(r1, r3)
            assert F.allclose(r2, r4)
            assert(F.allclose(e_grad1, e_grad5))
            assert(F.allclose(e_grad2, e_grad6))
            assert(F.allclose(e_grad3, e_grad7))
            assert(F.allclose(e_grad4, e_grad8))
>       _test(fn.copy_e, fn.sum)

tests/compute/test_new_update_all_hetero.py:217: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
tests/compute/test_new_update_all_hetero.py:168: in _test
    cross_reducer)
python/dgl/heterograph.py:5023: in multi_update_all
    all_out[dtid].append(core.message_passing(g, mfunc, rfunc, afunc))
python/dgl/core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python/dgl/core.py:332: in invoke_gspmm
    z = op(graph, x)
python/dgl/ops/spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb457c93870>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([[ 0.6293,  1.5118],
        [ 0.4452,  0.8534],
        [-1.0854, -0.2637],
        ...,
        [ 1.5691, -1.2203],
        [-0.5798,  0.1200],
        [-0.7006, -1.1834]], requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
__________________________ test_unary_copy_e[idtype1] __________________________

idtype = torch.int64

    @parametrize_idtype
    def test_unary_copy_e(idtype):
        def _test(mfunc, rfunc):
    
            g = create_test_heterograph_large(idtype)
            g0 = create_test_heterograph_2(idtype)
            g1 = create_test_heterograph(idtype)
            cross_reducer = rfunc.__name__
            x1 = F.randn((g.num_edges('plays'),feat_size))
            x2 = F.randn((g.num_edges('follows'),feat_size))
            x3 = F.randn((g.num_edges('develops'),feat_size))
            x4 = F.randn((g.num_edges('wishes'),feat_size))
            F.attach_grad(x1)
            F.attach_grad(x2)
            F.attach_grad(x3)
            F.attach_grad(x4)
            g['plays'].edata['eid'] = x1
            g['follows'].edata['eid'] = x2
            g['develops'].edata['eid'] = x3
            g['wishes'].edata['eid'] = x4
    
            #################################################################
            #  multi_update_all(): call msg_passing separately for each etype
            #################################################################
    
            with F.record_grad():
                g.multi_update_all(
                    {'plays' : (mfunc('eid', 'm'), rfunc('m', 'y')),
                    'follows': (mfunc('eid', 'm'), rfunc('m', 'y')),
                    'develops': (mfunc('eid', 'm'), rfunc('m', 'y')),
                    'wishes': (mfunc('eid', 'm'), rfunc('m', 'y'))},
                    cross_reducer)
                r1 = g.nodes['game'].data['y'].clone()
                r2 = g.nodes['user'].data['y'].clone()
                loss = r1.sum() + r2.sum()
                F.backward(loss)
                e_grad1 = F.grad(g['develops'].edata['eid']).clone()
                e_grad2 = F.grad(g['plays'].edata['eid']).clone()
                e_grad3 = F.grad(g['wishes'].edata['eid']).clone()
                e_grad4 = F.grad(g['follows'].edata['eid']).clone()
            {etype : (g[etype].edata.clear())
                for _, etype, _ in g.canonical_etypes},
    
            #################################################################
            #  update_all(): call msg_passing for all etypes
            #################################################################
    
            # TODO(Israt): output type can be None in multi_update and empty
            F.attach_grad(x1)
            F.attach_grad(x2)
            F.attach_grad(x3)
            F.attach_grad(x4)
    
            g['plays'].edata['eid'] = x1
            g['follows'].edata['eid'] = x2
            g['develops'].edata['eid'] = x3
            g['wishes'].edata['eid'] = x4
    
            with F.record_grad():
                g.update_all(mfunc('eid', 'm'), rfunc('m', 'y'))
                r3 = g.nodes['game'].data['y']
                r4 = g.nodes['user'].data['y']
                loss = r3.sum() + r4.sum()
                F.backward(loss)
                e_grad5 = F.grad(g['develops'].edata['eid'])
                e_grad6 = F.grad(g['plays'].edata['eid'])
                e_grad7 = F.grad(g['wishes'].edata['eid'])
                e_grad8 = F.grad(g['follows'].edata['eid'])
            # # correctness check
            def _print_error(a, b):
                for i, (x, y) in enumerate(zip(F.asnumpy(a).flatten(), F.asnumpy(b).flatten())):
                    if not np.allclose(x, y):
                       print('@{} {} v.s. {}'.format(i, x, y))
    
            assert F.allclose(r1, r3)
            assert F.allclose(r2, r4)
            assert(F.allclose(e_grad1, e_grad5))
            assert(F.allclose(e_grad2, e_grad6))
            assert(F.allclose(e_grad3, e_grad7))
            assert(F.allclose(e_grad4, e_grad8))
>       _test(fn.copy_e, fn.sum)

tests/compute/test_new_update_all_hetero.py:217: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
tests/compute/test_new_update_all_hetero.py:168: in _test
    cross_reducer)
python/dgl/heterograph.py:5023: in multi_update_all
    all_out[dtid].append(core.message_passing(g, mfunc, rfunc, afunc))
python/dgl/core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python/dgl/core.py:332: in invoke_gspmm
    z = op(graph, x)
python/dgl/ops/spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43eb504c8>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([[-0.0052,  2.0500],
        [-0.7121, -0.0811],
        [-0.1134, -0.1402],
        ...,
        [-0.6976, -0.4215],
        [ 0.9653, -0.8273],
        [-0.4365, -1.6862]], requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
___________________________ test_binary_op[idtype0] ____________________________

idtype = torch.int32

    @parametrize_idtype
    def test_binary_op(idtype):
        def _test(lhs, rhs, binary_op, reducer):
    
            g = create_test_heterograph(idtype)
    
            x1 = F.randn((g.num_nodes('user'), feat_size))
            x2 = F.randn((g.num_nodes('developer'), feat_size))
            x3 = F.randn((g.num_nodes('game'), feat_size))
    
            F.attach_grad(x1)
            F.attach_grad(x2)
            F.attach_grad(x3)
            g.nodes['user'].data['h'] = x1
            g.nodes['developer'].data['h'] = x2
            g.nodes['game'].data['h'] = x3
    
            x1 = F.randn((4,feat_size))
            x2 = F.randn((4,feat_size))
            x3 = F.randn((3,feat_size))
            x4 = F.randn((3,feat_size))
            F.attach_grad(x1)
            F.attach_grad(x2)
            F.attach_grad(x3)
            F.attach_grad(x4)
            g['plays'].edata['h'] = x1
            g['follows'].edata['h'] = x2
            g['develops'].edata['h'] = x3
            g['wishes'].edata['h'] = x4
    
            builtin_msg_name = "{}_{}_{}".format(lhs, binary_op, rhs)
            builtin_msg = getattr(fn, builtin_msg_name)
            builtin_red = getattr(fn, reducer)
    
            #################################################################
            #  multi_update_all(): call msg_passing separately for each etype
            #################################################################
    
            with F.record_grad():
                g.multi_update_all(
                    {etype : (builtin_msg('h', 'h', 'm'), builtin_red('m', 'y'))
                        for etype in g.canonical_etypes},
                    'sum')
                r1 = g.nodes['game'].data['y']
                F.backward(r1, F.ones(r1.shape))
                n_grad1 = F.grad(r1)
    
            #################################################################
            #  update_all(): call msg_passing for all etypes
            #################################################################
    
            g.update_all(builtin_msg('h', 'h', 'm'), builtin_red('m', 'y'))
            r2 = g.nodes['game'].data['y']
            F.backward(r2, F.ones(r2.shape))
            n_grad2 = F.grad(r2)
            # correctness check
            def _print_error(a, b):
                for i, (x, y) in enumerate(zip(F.asnumpy(a).flatten(), F.asnumpy(b).flatten())):
                    if not np.allclose(x, y):
                        print('@{} {} v.s. {}'.format(i, x, y))
    
            if not F.allclose(r1, r2):
                _print_error(r1, r2)
            assert F.allclose(r1, r2)
            # TODO (Israt): r1 and r2 have different frad func associated with
            # if not F.allclose(n_grad1, n_grad2):
            #     print('node grad')
            #     _print_error(n_grad1, n_grad2)
            # assert(F.allclose(n_grad1, n_grad2))
    
        target = ["u", "v", "e"]
        for lhs, rhs in product(target, target):
            if lhs == rhs:
                continue
            for binary_op in ["add", "sub", "mul", "div"]:
                # TODO(Israt) :Add support for reduce func "max", "min", "mean"
                for reducer in ["sum"]:
                    print(lhs, rhs, binary_op, reducer)
>                   _test(lhs, rhs, binary_op, reducer)

tests/compute/test_new_update_all_hetero.py:300: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
tests/compute/test_new_update_all_hetero.py:264: in _test
    'sum')
python/dgl/heterograph.py:5023: in multi_update_all
    all_out[dtid].append(core.message_passing(g, mfunc, rfunc, afunc))
python/dgl/core.py:362: in message_passing
    msgdata = invoke_gsddmm(g, mfunc)
python/dgl/core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python/dgl/ops/sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43eb5d630>
op = 'add'
lhs_data = tensor([[ 1.7124,  0.1878],
        [ 1.1194, -1.8616]], requires_grad=True)
rhs_data = tensor([[ 0.3744,  1.5753],
        [-0.3141, -0.7006]], requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
u v add sum
___________________________ test_binary_op[idtype1] ____________________________

idtype = torch.int64

    @parametrize_idtype
    def test_binary_op(idtype):
        def _test(lhs, rhs, binary_op, reducer):
    
            g = create_test_heterograph(idtype)
    
            x1 = F.randn((g.num_nodes('user'), feat_size))
            x2 = F.randn((g.num_nodes('developer'), feat_size))
            x3 = F.randn((g.num_nodes('game'), feat_size))
    
            F.attach_grad(x1)
            F.attach_grad(x2)
            F.attach_grad(x3)
            g.nodes['user'].data['h'] = x1
            g.nodes['developer'].data['h'] = x2
            g.nodes['game'].data['h'] = x3
    
            x1 = F.randn((4,feat_size))
            x2 = F.randn((4,feat_size))
            x3 = F.randn((3,feat_size))
            x4 = F.randn((3,feat_size))
            F.attach_grad(x1)
            F.attach_grad(x2)
            F.attach_grad(x3)
            F.attach_grad(x4)
            g['plays'].edata['h'] = x1
            g['follows'].edata['h'] = x2
            g['develops'].edata['h'] = x3
            g['wishes'].edata['h'] = x4
    
            builtin_msg_name = "{}_{}_{}".format(lhs, binary_op, rhs)
            builtin_msg = getattr(fn, builtin_msg_name)
            builtin_red = getattr(fn, reducer)
    
            #################################################################
            #  multi_update_all(): call msg_passing separately for each etype
            #################################################################
    
            with F.record_grad():
                g.multi_update_all(
                    {etype : (builtin_msg('h', 'h', 'm'), builtin_red('m', 'y'))
                        for etype in g.canonical_etypes},
                    'sum')
                r1 = g.nodes['game'].data['y']
                F.backward(r1, F.ones(r1.shape))
                n_grad1 = F.grad(r1)
    
            #################################################################
            #  update_all(): call msg_passing for all etypes
            #################################################################
    
            g.update_all(builtin_msg('h', 'h', 'm'), builtin_red('m', 'y'))
            r2 = g.nodes['game'].data['y']
            F.backward(r2, F.ones(r2.shape))
            n_grad2 = F.grad(r2)
            # correctness check
            def _print_error(a, b):
                for i, (x, y) in enumerate(zip(F.asnumpy(a).flatten(), F.asnumpy(b).flatten())):
                    if not np.allclose(x, y):
                        print('@{} {} v.s. {}'.format(i, x, y))
    
            if not F.allclose(r1, r2):
                _print_error(r1, r2)
            assert F.allclose(r1, r2)
            # TODO (Israt): r1 and r2 have different frad func associated with
            # if not F.allclose(n_grad1, n_grad2):
            #     print('node grad')
            #     _print_error(n_grad1, n_grad2)
            # assert(F.allclose(n_grad1, n_grad2))
    
        target = ["u", "v", "e"]
        for lhs, rhs in product(target, target):
            if lhs == rhs:
                continue
            for binary_op in ["add", "sub", "mul", "div"]:
                # TODO(Israt) :Add support for reduce func "max", "min", "mean"
                for reducer in ["sum"]:
                    print(lhs, rhs, binary_op, reducer)
>                   _test(lhs, rhs, binary_op, reducer)

tests/compute/test_new_update_all_hetero.py:300: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
tests/compute/test_new_update_all_hetero.py:264: in _test
    'sum')
python/dgl/heterograph.py:5023: in multi_update_all
    all_out[dtid].append(core.message_passing(g, mfunc, rfunc, afunc))
python/dgl/core.py:362: in message_passing
    msgdata = invoke_gsddmm(g, mfunc)
python/dgl/core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python/dgl/ops/sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb705998510>
op = 'add'
lhs_data = tensor([[-0.4307, -0.1323],
        [-0.5036,  0.5018]], requires_grad=True)
rhs_data = tensor([[ 0.1916, -0.3977],
        [-0.0392, -0.7487]], requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
u v add sum
___________________________ test_sum_case1[idtype0] ____________________________

idtype = torch.int32

    @parametrize_idtype
    def test_sum_case1(idtype):
        # NOTE: If you want to update this test case, remember to update the docstring
        #  example too!!!
        g1 = dgl.graph(([0, 1], [1, 0]), idtype=idtype, device=F.ctx())
        g1.ndata['h'] = F.tensor([1., 2.])
        g2 = dgl.graph(([0, 1], [1, 2]), idtype=idtype, device=F.ctx())
        g2.ndata['h'] = F.tensor([1., 2., 3.])
        bg = dgl.batch([g1, g2])
        bg.ndata['w'] = F.tensor([.1, .2, .1, .5, .2])
>       assert F.allclose(F.tensor([3.]), dgl.sum_nodes(g1, 'h'))

tests/compute/test_readout.py:20: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:180: in sum_nodes
    return readout_nodes(graph, feat, weight, ntype=ntype, op='sum')
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'sum', x = tensor([1., 2.]), offsets = tensor([0, 2])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
___________________________ test_sum_case1[idtype1] ____________________________

idtype = torch.int64

    @parametrize_idtype
    def test_sum_case1(idtype):
        # NOTE: If you want to update this test case, remember to update the docstring
        #  example too!!!
        g1 = dgl.graph(([0, 1], [1, 0]), idtype=idtype, device=F.ctx())
        g1.ndata['h'] = F.tensor([1., 2.])
        g2 = dgl.graph(([0, 1], [1, 2]), idtype=idtype, device=F.ctx())
        g2.ndata['h'] = F.tensor([1., 2., 3.])
        bg = dgl.batch([g1, g2])
        bg.ndata['w'] = F.tensor([.1, .2, .1, .5, .2])
>       assert F.allclose(F.tensor([3.]), dgl.sum_nodes(g1, 'h'))

tests/compute/test_readout.py:20: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:180: in sum_nodes
    return readout_nodes(graph, feat, weight, ntype=ntype, op='sum')
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'sum', x = tensor([1., 2.]), offsets = tensor([0, 2])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
_____________________ test_reduce_readout[sum-g0-idtype0] ______________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32)}
      edata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int32, reducer = 'sum'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', op=reducer)

tests/compute/test_readout.py:33: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'sum'
x = tensor([[-2.8364, -1.0353, -0.4919],
        [-1.1800,  0.6876,  1.3861],
        [ 0.3296,  0.4225,  1.9144],
       ....7921],
        [-0.8496,  0.1152,  1.1855],
        [-1.7852,  0.0727, -0.8690],
        [-1.4106, -0.8359,  0.8846]])
offsets = tensor([0, 4, 7, 9])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
_____________________ test_reduce_readout[sum-g0-idtype1] ______________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32)}
      edata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int64, reducer = 'sum'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', op=reducer)

tests/compute/test_readout.py:33: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'sum'
x = tensor([[-1.0782,  0.6476, -0.8538],
        [-0.5916, -0.9811,  0.9358],
        [ 0.7833, -2.0918, -2.3020],
       ....5363],
        [-0.3733, -1.6018,  0.4269],
        [-0.1997, -1.3495, -1.6612],
        [-0.7860,  0.0297,  0.8318]])
offsets = tensor([0, 4, 7, 9])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
_____________________ test_reduce_readout[sum-g1-idtype0] ______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32)}
      edata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int32, reducer = 'sum'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', op=reducer)

tests/compute/test_readout.py:33: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'sum'
x = tensor([[-1.0739,  1.6836, -0.8769],
        [ 0.9023,  0.4539, -0.1565],
        [-0.3857, -2.4024, -0.7915],
       ....0675],
        [ 2.0740, -0.9526,  1.2514],
        [-1.9887, -0.1769,  0.3457],
        [ 0.5959,  1.6257, -0.4776]])
offsets = tensor([ 0, 10])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
_____________________ test_reduce_readout[sum-g1-idtype1] ______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32)}
      edata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int64, reducer = 'sum'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', op=reducer)

tests/compute/test_readout.py:33: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'sum'
x = tensor([[ 0.7592, -1.9412, -0.2083],
        [ 1.6470,  0.3246,  0.7294],
        [-0.4415,  1.6107,  0.7801],
       ....0334],
        [ 1.1004, -0.0910,  1.1176],
        [-0.2845, -0.2347, -0.4729],
        [-0.5175,  0.0828,  0.1533]])
offsets = tensor([ 0, 10])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
_____________________ test_reduce_readout[sum-g2-idtype0] ______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32)}
      edata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int32, reducer = 'sum'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', op=reducer)

tests/compute/test_readout.py:33: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'sum'
x = tensor([[-1.4882, -0.2959,  0.1846],
        [-0.4123,  0.2662,  1.7657],
        [ 0.6045, -1.1730,  1.1152],
       ....0098],
        [-0.2651,  0.1838,  0.1709],
        [ 1.1290, -1.6773,  1.0562],
        [ 0.5711, -0.2179, -0.0673]])
offsets = tensor([ 0, 10])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
_____________________ test_reduce_readout[sum-g2-idtype1] ______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32)}
      edata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int64, reducer = 'sum'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', op=reducer)

tests/compute/test_readout.py:33: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'sum'
x = tensor([[-1.0299,  0.2888, -1.1753],
        [-0.2036,  1.4174,  0.6362],
        [ 1.0356, -0.2404,  0.5936],
       ....9475],
        [ 1.3080,  0.1385,  0.0250],
        [-0.3133,  0.9439,  2.1647],
        [-0.3130,  0.2468, -1.6515]])
offsets = tensor([ 0, 10])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
_____________________ test_reduce_readout[sum-g3-idtype0] ______________________

g = Graph(num_nodes=5, num_edges=6,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32)}
      edata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int32, reducer = 'sum'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', op=reducer)

tests/compute/test_readout.py:33: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'sum'
x = tensor([[ 0.5612, -0.8833, -0.0137],
        [ 0.3736,  1.3757,  0.5183],
        [ 0.1549,  0.3168,  0.8680],
        [-0.4952,  1.3115, -3.0724],
        [-0.1829, -0.1217,  0.0108]])
offsets = tensor([0, 5])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
_____________________ test_reduce_readout[sum-g3-idtype1] ______________________

g = Graph(num_nodes=5, num_edges=6,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32)}
      edata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int64, reducer = 'sum'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', op=reducer)

tests/compute/test_readout.py:33: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'sum'
x = tensor([[-0.0659,  1.7096,  1.7523],
        [-0.6923,  0.2334,  1.0874],
        [ 0.2114, -1.9964,  1.1410],
        [ 1.1908,  0.4091, -0.5994],
        [ 1.4762, -0.1851, -1.1883]])
offsets = tensor([0, 5])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
_____________________ test_reduce_readout[sum-g4-idtype0] ______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32)}
      edata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int32, reducer = 'sum'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', op=reducer)

tests/compute/test_readout.py:33: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'sum'
x = tensor([[-0.9318,  1.3094, -0.7576],
        [-1.1828, -1.2134, -1.5229],
        [-0.4690, -1.6581,  0.7433],
       ....0472],
        [-1.2740,  0.4992,  0.0797],
        [-1.5507, -0.7189, -0.5618],
        [-0.7115, -0.3167,  1.2862]])
offsets = tensor([ 0, 10])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
_____________________ test_reduce_readout[sum-g4-idtype1] ______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32)}
      edata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int64, reducer = 'sum'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', op=reducer)

tests/compute/test_readout.py:33: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'sum'
x = tensor([[-1.2567,  0.6579,  1.4155],
        [ 0.0826,  0.0606, -0.2362],
        [-0.9078, -1.1937, -0.4678],
       ....1282],
        [-1.6351,  0.3001,  1.5766],
        [-2.1818, -0.6137,  0.6100],
        [ 0.5433,  1.5816, -0.6341]])
offsets = tensor([ 0, 10])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
_____________________ test_reduce_readout[sum-g5-idtype0] ______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32), 'h': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int32, reducer = 'sum'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', op=reducer)

tests/compute/test_readout.py:33: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'sum'
x = tensor([[-0.9684, -0.0986, -2.1648],
        [ 1.4855,  0.5254,  0.1432],
        [ 0.2317,  1.1700, -0.7859],
       ....5633],
        [ 0.1775,  0.5587, -0.2780],
        [-1.2737,  0.3181, -0.4133],
        [ 0.6354,  0.4366,  0.1938]])
offsets = tensor([ 0, 10])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
_____________________ test_reduce_readout[sum-g5-idtype1] ______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32), 'h': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int64, reducer = 'sum'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', op=reducer)

tests/compute/test_readout.py:33: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'sum'
x = tensor([[ 0.2859, -2.1922, -0.7229],
        [-1.8786,  1.4057, -0.7536],
        [-0.4350, -0.3864, -0.0662],
       ....0198],
        [ 0.4340, -0.5945,  0.4388],
        [-0.5108, -2.4906, -0.4912],
        [-0.2969, -1.1603, -0.8026]])
offsets = tensor([ 0, 10])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
_____________________ test_reduce_readout[sum-g6-idtype0] ______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), 'h': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int32, reducer = 'sum'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', op=reducer)

tests/compute/test_readout.py:33: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'sum'
x = tensor([[ 1.4495,  1.3584,  0.2090],
        [-0.5432,  0.9094,  0.4453],
        [ 1.5122, -0.4356,  0.0107],
       ....3282],
        [-1.0932,  1.0396,  0.7743],
        [ 1.1792, -0.6445, -0.1553],
        [-0.6456,  0.0351,  0.3879]])
offsets = tensor([ 0, 10])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
_____________________ test_reduce_readout[sum-g6-idtype1] ______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), 'h': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int64, reducer = 'sum'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', op=reducer)

tests/compute/test_readout.py:33: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'sum'
x = tensor([[-0.9179, -0.2448,  0.7839],
        [-0.2048, -1.8162, -0.1200],
        [-1.8167,  0.9537, -0.3597],
       ....4443],
        [ 1.2055, -0.3588,  1.2357],
        [ 1.3663, -0.4251,  0.0981],
        [-0.1938, -1.0618, -0.2631]])
offsets = tensor([ 0, 10])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
_____________________ test_reduce_readout[max-g0-idtype0] ______________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32)}
      edata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int32, reducer = 'max'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', op=reducer)

tests/compute/test_readout.py:33: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'max'
x = tensor([[ 1.7384,  0.7851, -1.3229],
        [-2.5093, -0.6567, -0.5050],
        [ 0.0304, -0.7496,  0.6551],
       ....6980],
        [ 0.8913, -0.7262, -0.7476],
        [ 1.5244, -2.2297,  0.0631],
        [ 1.1767,  0.2626,  0.1108]])
offsets = tensor([0, 4, 7, 9])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
_____________________ test_reduce_readout[max-g0-idtype1] ______________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32)}
      edata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int64, reducer = 'max'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', op=reducer)

tests/compute/test_readout.py:33: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'max'
x = tensor([[ 0.1971,  1.3030,  1.5339],
        [ 0.7505,  0.9339,  1.2065],
        [-0.5800,  0.1585, -0.1759],
       ....4128],
        [ 0.3614, -1.5498,  0.3387],
        [-1.4349, -0.4409,  1.1713],
        [-0.1676, -1.7781,  1.3261]])
offsets = tensor([0, 4, 7, 9])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
_____________________ test_reduce_readout[max-g1-idtype0] ______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32)}
      edata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int32, reducer = 'max'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', op=reducer)

tests/compute/test_readout.py:33: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'max'
x = tensor([[ 1.4524,  1.0101, -1.4148],
        [-0.0199,  0.8429,  0.1273],
        [-1.0890,  1.2716, -1.2364],
       ....4446],
        [-1.2056, -0.1659, -0.1270],
        [ 2.1496, -0.6172, -0.1679],
        [-1.2433, -1.9676, -0.6706]])
offsets = tensor([ 0, 10])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
_____________________ test_reduce_readout[max-g1-idtype1] ______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32)}
      edata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int64, reducer = 'max'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', op=reducer)

tests/compute/test_readout.py:33: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'max'
x = tensor([[-0.1172, -1.5030, -0.5167],
        [-0.1370, -0.5816,  1.4138],
        [ 1.6629,  0.8165,  0.4193],
       ....0616],
        [-0.7221,  0.2972, -2.5609],
        [ 1.4580, -0.8247,  0.6954],
        [ 0.0260,  0.3529,  1.6332]])
offsets = tensor([ 0, 10])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
_____________________ test_reduce_readout[max-g2-idtype0] ______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32)}
      edata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int32, reducer = 'max'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', op=reducer)

tests/compute/test_readout.py:33: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'max'
x = tensor([[-0.1360,  0.0952,  1.0855],
        [-0.9518, -0.8557, -0.7348],
        [ 0.9089,  0.6412,  0.8596],
       ....2154],
        [ 0.7787, -1.1165, -0.6668],
        [-0.5932, -1.4121,  0.5853],
        [-0.9795,  2.1147,  0.2790]])
offsets = tensor([ 0, 10])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
_____________________ test_reduce_readout[max-g2-idtype1] ______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32)}
      edata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int64, reducer = 'max'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', op=reducer)

tests/compute/test_readout.py:33: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'max'
x = tensor([[ 0.5900, -0.2583, -0.4099],
        [ 1.1172, -2.4816,  1.7347],
        [-1.7869, -1.1873, -0.5998],
       ....1463],
        [ 0.2329, -1.2721,  0.6777],
        [ 0.3531, -0.7692,  0.2427],
        [-0.1919, -1.6321, -1.4145]])
offsets = tensor([ 0, 10])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
_____________________ test_reduce_readout[max-g3-idtype0] ______________________

g = Graph(num_nodes=5, num_edges=6,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32)}
      edata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int32, reducer = 'max'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', op=reducer)

tests/compute/test_readout.py:33: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'max'
x = tensor([[ 0.7597, -0.1085,  0.9324],
        [-1.1951, -0.7605,  0.3853],
        [-0.7983, -1.5080, -2.0970],
        [-0.0378, -0.4593,  1.2322],
        [-0.5752, -0.5472, -0.1621]])
offsets = tensor([0, 5])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
_____________________ test_reduce_readout[max-g3-idtype1] ______________________

g = Graph(num_nodes=5, num_edges=6,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32)}
      edata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int64, reducer = 'max'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', op=reducer)

tests/compute/test_readout.py:33: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'max'
x = tensor([[ 0.3216, -0.0237,  1.2853],
        [ 0.4177,  0.8905, -0.2841],
        [ 0.8723,  0.1649,  0.0287],
        [-0.0258, -1.0771,  0.4485],
        [ 1.2325, -1.3257,  0.9312]])
offsets = tensor([0, 5])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
_____________________ test_reduce_readout[max-g4-idtype0] ______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32)}
      edata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int32, reducer = 'max'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', op=reducer)

tests/compute/test_readout.py:33: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'max'
x = tensor([[-1.1656, -1.5166, -1.1086],
        [ 1.2280,  0.2463, -1.3887],
        [ 1.4164,  0.7573,  0.8159],
       ....3492],
        [ 0.3382,  1.1115, -0.4812],
        [-0.0650, -0.4075,  1.6826],
        [ 0.2590,  1.5538, -0.3788]])
offsets = tensor([ 0, 10])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
_____________________ test_reduce_readout[max-g4-idtype1] ______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32)}
      edata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int64, reducer = 'max'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', op=reducer)

tests/compute/test_readout.py:33: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'max'
x = tensor([[-1.0307, -2.4815, -1.1040],
        [-0.7126,  0.2800,  0.9599],
        [ 1.2104, -0.8818, -0.2031],
       ....8811],
        [ 0.5869,  0.0654, -1.0220],
        [-0.3045,  0.9668,  1.8083],
        [ 0.5610,  0.0356, -0.5141]])
offsets = tensor([ 0, 10])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
_____________________ test_reduce_readout[max-g5-idtype0] ______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32), 'h': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int32, reducer = 'max'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', op=reducer)

tests/compute/test_readout.py:33: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'max'
x = tensor([[ 0.5707,  0.3315,  0.2343],
        [ 0.5113, -0.6062, -0.2481],
        [ 1.8136, -1.6279,  0.8133],
       ....9845],
        [ 1.0555, -0.6158, -0.6891],
        [ 0.8164, -1.1297,  0.7604],
        [-0.9752,  1.0832,  1.3712]])
offsets = tensor([ 0, 10])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
_____________________ test_reduce_readout[max-g5-idtype1] ______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32), 'h': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int64, reducer = 'max'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', op=reducer)

tests/compute/test_readout.py:33: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'max'
x = tensor([[-0.7425, -0.2770, -0.1798],
        [ 0.3136, -1.6532, -0.2974],
        [ 1.3246, -1.0963,  0.7130],
       ....1270],
        [ 0.5998,  1.5538,  1.1494],
        [-1.0788,  1.7510,  0.7316],
        [ 0.3548,  2.8994, -1.6148]])
offsets = tensor([ 0, 10])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
_____________________ test_reduce_readout[max-g6-idtype0] ______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), 'h': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int32, reducer = 'max'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', op=reducer)

tests/compute/test_readout.py:33: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'max'
x = tensor([[ 0.0824,  0.4856,  1.3446],
        [-1.4574,  0.2896, -1.1125],
        [ 1.5695, -0.3545,  0.5547],
       ....5782],
        [ 0.0834,  0.4634, -0.3549],
        [-0.7650,  0.1975,  0.1324],
        [ 1.7680, -1.9302,  0.0125]])
offsets = tensor([ 0, 10])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
_____________________ test_reduce_readout[max-g6-idtype1] ______________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), 'h': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int64, reducer = 'max'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', op=reducer)

tests/compute/test_readout.py:33: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'max'
x = tensor([[-1.4912, -1.1968, -0.8199],
        [-1.2686,  0.4707,  0.8057],
        [-1.2792, -0.0106, -0.6686],
       ....6055],
        [ 1.4023,  1.2641, -1.6673],
        [ 0.1817, -0.2396, -1.5206],
        [ 0.3942,  0.3143, -1.0493]])
offsets = tensor([ 0, 10])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
_____________________ test_reduce_readout[mean-g0-idtype0] _____________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32)}
      edata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int32, reducer = 'mean'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', op=reducer)

tests/compute/test_readout.py:33: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:46: in segment_reduce
    rst = F.segment_reduce('sum', value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'sum'
x = tensor([[-0.8727, -0.0405, -0.3006],
        [-1.3837, -0.9498, -0.4831],
        [ 0.5086, -0.6589,  0.3464],
       ....1500],
        [ 0.8424,  0.5004, -0.5893],
        [ 0.5795, -0.6771,  0.7701],
        [ 1.4044, -1.3909,  0.4566]])
offsets = tensor([0, 4, 7, 9])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
_____________________ test_reduce_readout[mean-g0-idtype1] _____________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32)}
      edata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int64, reducer = 'mean'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', op=reducer)

tests/compute/test_readout.py:33: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:46: in segment_reduce
    rst = F.segment_reduce('sum', value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'sum'
x = tensor([[-0.1080,  1.4928,  0.7905],
        [ 0.5214,  1.0339,  0.1882],
        [ 0.2188, -0.0341,  0.3122],
       ....5653],
        [ 0.9583, -1.7611,  0.6450],
        [ 0.7494,  1.1224,  0.1799],
        [ 0.7639,  0.3151,  0.3090]])
offsets = tensor([0, 4, 7, 9])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
_____________________ test_reduce_readout[mean-g1-idtype0] _____________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32)}
      edata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int32, reducer = 'mean'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', op=reducer)

tests/compute/test_readout.py:33: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:46: in segment_reduce
    rst = F.segment_reduce('sum', value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'sum'
x = tensor([[-1.6878,  0.5801,  0.6904],
        [-0.4028,  0.5830, -0.4169],
        [ 0.2618, -0.3212,  0.0323],
       ....0605],
        [ 0.9768, -0.1534, -0.2137],
        [-1.0656, -1.1702,  0.1521],
        [-0.3097, -1.1240, -2.4908]])
offsets = tensor([ 0, 10])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
_____________________ test_reduce_readout[mean-g1-idtype1] _____________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32)}
      edata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int64, reducer = 'mean'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', op=reducer)

tests/compute/test_readout.py:33: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:46: in segment_reduce
    rst = F.segment_reduce('sum', value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'sum'
x = tensor([[-0.0266,  0.3205,  0.6258],
        [-0.4271, -1.4287,  1.6424],
        [ 0.5641, -1.0218, -0.5197],
       ....0323],
        [-1.7294,  0.1164, -1.0827],
        [-0.4149,  1.2868,  0.0692],
        [-0.8454,  1.0687,  1.2230]])
offsets = tensor([ 0, 10])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
_____________________ test_reduce_readout[mean-g2-idtype0] _____________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32)}
      edata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int32, reducer = 'mean'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', op=reducer)

tests/compute/test_readout.py:33: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:46: in segment_reduce
    rst = F.segment_reduce('sum', value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'sum'
x = tensor([[-0.3977, -0.7196,  0.4419],
        [ 0.4705,  1.0076,  0.7802],
        [ 0.0820,  0.1748, -1.2564],
       ....5503],
        [ 1.4653, -0.4887, -1.5526],
        [-0.1370, -0.8826,  0.7302],
        [ 1.1000, -2.1532, -0.0598]])
offsets = tensor([ 0, 10])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
_____________________ test_reduce_readout[mean-g2-idtype1] _____________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32)}
      edata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int64, reducer = 'mean'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', op=reducer)

tests/compute/test_readout.py:33: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:46: in segment_reduce
    rst = F.segment_reduce('sum', value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'sum'
x = tensor([[-0.1221, -0.0688,  0.5091],
        [-0.3279,  1.0434,  0.3342],
        [ 1.7591, -0.9773,  0.2410],
       ....0199],
        [ 1.0838,  1.5097, -1.4376],
        [ 1.4138,  1.0552,  0.1734],
        [ 0.0873, -2.3705,  0.1102]])
offsets = tensor([ 0, 10])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
_____________________ test_reduce_readout[mean-g3-idtype0] _____________________

g = Graph(num_nodes=5, num_edges=6,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32)}
      edata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int32, reducer = 'mean'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', op=reducer)

tests/compute/test_readout.py:33: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:46: in segment_reduce
    rst = F.segment_reduce('sum', value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'sum'
x = tensor([[ 0.8321,  1.3178,  0.5952],
        [ 0.5472, -0.8827, -0.0050],
        [-0.9443, -3.5292, -0.5378],
        [-1.3447,  0.2092, -1.0280],
        [-1.4643,  1.1571,  0.0487]])
offsets = tensor([0, 5])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
_____________________ test_reduce_readout[mean-g3-idtype1] _____________________

g = Graph(num_nodes=5, num_edges=6,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32)}
      edata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int64, reducer = 'mean'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', op=reducer)

tests/compute/test_readout.py:33: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:46: in segment_reduce
    rst = F.segment_reduce('sum', value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'sum'
x = tensor([[-1.6457, -0.5760, -0.4526],
        [ 0.3032,  0.0465, -1.0869],
        [ 1.3555, -0.1523, -1.5895],
        [ 2.7079, -0.9616,  0.1727],
        [ 0.8902, -1.3105,  0.5742]])
offsets = tensor([0, 5])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
_____________________ test_reduce_readout[mean-g4-idtype0] _____________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32)}
      edata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int32, reducer = 'mean'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', op=reducer)

tests/compute/test_readout.py:33: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:46: in segment_reduce
    rst = F.segment_reduce('sum', value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'sum'
x = tensor([[-1.6001, -1.5004,  0.7876],
        [ 0.1058,  0.9451, -1.5262],
        [ 0.0306,  1.6976,  0.3043],
       ....2891],
        [ 1.0479,  0.1818,  0.8173],
        [-0.2006, -0.2820, -0.7820],
        [ 0.2447, -0.2099, -0.1746]])
offsets = tensor([ 0, 10])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
_____________________ test_reduce_readout[mean-g4-idtype1] _____________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32)}
      edata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int64, reducer = 'mean'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', op=reducer)

tests/compute/test_readout.py:33: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:46: in segment_reduce
    rst = F.segment_reduce('sum', value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'sum'
x = tensor([[ 0.8169,  0.1152, -1.2259],
        [ 0.5931, -0.8041, -0.5575],
        [-0.6959,  1.7790, -0.9705],
       ....1865],
        [-0.6379, -1.3150,  0.4085],
        [-0.9715,  0.4371, -0.5713],
        [ 0.5112, -0.2477,  1.4977]])
offsets = tensor([ 0, 10])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
_____________________ test_reduce_readout[mean-g5-idtype0] _____________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32), 'h': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int32, reducer = 'mean'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', op=reducer)

tests/compute/test_readout.py:33: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:46: in segment_reduce
    rst = F.segment_reduce('sum', value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'sum'
x = tensor([[ 1.0167, -0.5613, -0.2234],
        [ 1.6827,  0.1945,  0.6440],
        [-0.2404, -2.0453,  0.7849],
       ....1146],
        [-0.3541, -0.9578,  0.2126],
        [-0.3597,  2.5869,  1.0755],
        [-1.1646,  1.9828, -0.1719]])
offsets = tensor([ 0, 10])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
_____________________ test_reduce_readout[mean-g5-idtype1] _____________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32), 'h': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int64, reducer = 'mean'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', op=reducer)

tests/compute/test_readout.py:33: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:46: in segment_reduce
    rst = F.segment_reduce('sum', value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'sum'
x = tensor([[ 0.0517,  0.5746, -1.0183],
        [ 0.2621,  1.1710, -0.7309],
        [ 0.2695, -0.8435, -0.3373],
       ....7484],
        [ 1.0682,  0.1011,  0.7573],
        [ 0.3541, -0.0729, -0.0553],
        [-0.3816, -1.1214,  1.6027]])
offsets = tensor([ 0, 10])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
_____________________ test_reduce_readout[mean-g6-idtype0] _____________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), 'h': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int32, reducer = 'mean'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', op=reducer)

tests/compute/test_readout.py:33: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:46: in segment_reduce
    rst = F.segment_reduce('sum', value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'sum'
x = tensor([[ 1.8807,  0.0703, -0.2096],
        [ 0.3712, -2.3564, -1.2583],
        [ 1.5721, -0.7247, -0.9661],
       ....1051],
        [-0.6164, -0.6354, -0.6497],
        [-0.1298,  0.0505, -0.0793],
        [ 0.4194, -0.2703,  0.5401]])
offsets = tensor([ 0, 10])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
_____________________ test_reduce_readout[mean-g6-idtype1] _____________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), 'h': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int64, reducer = 'mean'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', op=reducer)

tests/compute/test_readout.py:33: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:46: in segment_reduce
    rst = F.segment_reduce('sum', value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'sum'
x = tensor([[-0.3477,  0.7987,  1.4922],
        [ 0.3608,  0.7453, -0.3156],
        [-0.3867, -0.3508,  0.2761],
       ....6832],
        [-1.5721,  0.8892,  0.6916],
        [ 0.9180, -1.0732,  0.1784],
        [ 0.8905,  1.9989, -0.5284]])
offsets = tensor([ 0, 10])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
_________________ test_weighted_reduce_readout[sum-g0-idtype0] _________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32), 'w': Scheme(shape=...t32)}
      edata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'w': Scheme(shape=(1,), dtype=torch.float32)})
idtype = torch.int32, reducer = 'sum'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_weighted_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.ndata['w'] = F.randn((g.number_of_nodes(), 1))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
        g.edata['w'] = F.randn((g.number_of_edges(), 1))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', 'w', op=reducer)

tests/compute/test_readout.py:81: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'sum'
x = tensor([[ 0.3261,  1.5683,  0.3717],
        [-0.0552, -0.0162,  0.0074],
        [-1.0557,  0.1204, -0.1036],
       ....0230],
        [-0.0046,  0.3150, -0.2855],
        [ 1.0539,  0.7194, -0.0299],
        [-0.3368, -1.1975,  0.5502]])
offsets = tensor([0, 4, 7, 9])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
_________________ test_weighted_reduce_readout[sum-g0-idtype1] _________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32), 'w': Scheme(shape=...t32)}
      edata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'w': Scheme(shape=(1,), dtype=torch.float32)})
idtype = torch.int64, reducer = 'sum'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_weighted_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.ndata['w'] = F.randn((g.number_of_nodes(), 1))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
        g.edata['w'] = F.randn((g.number_of_edges(), 1))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', 'w', op=reducer)

tests/compute/test_readout.py:81: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'sum'
x = tensor([[-2.3089e-01,  1.7783e+00,  4.7157e-01],
        [ 4.2602e-01, -5.2966e-01, -5.1168e-01],
        [ 3.2777e-01...4e-03,  3.5090e-02],
        [-7.8913e-01, -7.0228e-01, -1.2504e-01],
        [ 2.4140e+00, -3.1584e+00, -3.2709e-01]])
offsets = tensor([0, 4, 7, 9])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
_________________ test_weighted_reduce_readout[sum-g1-idtype0] _________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32), 'w': Scheme(shape...t32)}
      edata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'w': Scheme(shape=(1,), dtype=torch.float32)})
idtype = torch.int32, reducer = 'sum'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_weighted_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.ndata['w'] = F.randn((g.number_of_nodes(), 1))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
        g.edata['w'] = F.randn((g.number_of_edges(), 1))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', 'w', op=reducer)

tests/compute/test_readout.py:81: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'sum'
x = tensor([[-2.7431e-01, -1.0017e+00, -1.4377e+00],
        [-1.5117e-02, -1.5229e-01, -9.6422e-02],
        [ 1.1204e-01...1e-01, -1.4309e-01],
        [ 3.1655e-01,  4.8792e-02,  5.0287e-02],
        [ 1.8599e+00, -4.2643e-01,  1.5963e+00]])
offsets = tensor([ 0, 10])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
_________________ test_weighted_reduce_readout[sum-g1-idtype1] _________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32), 'w': Scheme(shape...t32)}
      edata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'w': Scheme(shape=(1,), dtype=torch.float32)})
idtype = torch.int64, reducer = 'sum'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_weighted_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.ndata['w'] = F.randn((g.number_of_nodes(), 1))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
        g.edata['w'] = F.randn((g.number_of_edges(), 1))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', 'w', op=reducer)

tests/compute/test_readout.py:81: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'sum'
x = tensor([[ 0.0609, -0.7648,  0.3178],
        [ 0.1062,  0.0182,  0.1126],
        [ 0.4784, -0.2483,  0.3041],
       ....8023],
        [ 0.0274, -0.1061, -0.0986],
        [ 4.1400, -0.6954,  1.9493],
        [-0.2705,  0.0893,  0.1558]])
offsets = tensor([ 0, 10])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
_________________ test_weighted_reduce_readout[sum-g2-idtype0] _________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32), 'w': Scheme(shape...t32)}
      edata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'w': Scheme(shape=(1,), dtype=torch.float32)})
idtype = torch.int32, reducer = 'sum'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_weighted_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.ndata['w'] = F.randn((g.number_of_nodes(), 1))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
        g.edata['w'] = F.randn((g.number_of_edges(), 1))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', 'w', op=reducer)

tests/compute/test_readout.py:81: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'sum'
x = tensor([[-2.4056e-01, -2.5571e-03,  1.0755e-01],
        [ 8.3282e-01,  1.8654e+00,  6.0927e-01],
        [ 1.1808e-01...2e+00, -5.3770e-01],
        [-3.6995e-03,  1.5944e-01, -1.0481e+00],
        [ 2.8367e-02, -1.8220e-02, -2.9318e-02]])
offsets = tensor([ 0, 10])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
_________________ test_weighted_reduce_readout[sum-g2-idtype1] _________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32), 'w': Scheme(shape...t32)}
      edata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'w': Scheme(shape=(1,), dtype=torch.float32)})
idtype = torch.int64, reducer = 'sum'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_weighted_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.ndata['w'] = F.randn((g.number_of_nodes(), 1))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
        g.edata['w'] = F.randn((g.number_of_edges(), 1))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', 'w', op=reducer)

tests/compute/test_readout.py:81: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'sum'
x = tensor([[-4.2779e-01,  2.7114e-01,  3.3167e-01],
        [-2.4456e-03, -1.6760e-03, -5.3544e-03],
        [ 6.8029e-01...9e-01,  9.4353e-02],
        [-2.4849e+00,  3.9758e-01, -6.7391e-01],
        [-8.5060e-01,  9.6941e-02,  1.3614e+00]])
offsets = tensor([ 0, 10])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
_________________ test_weighted_reduce_readout[sum-g3-idtype0] _________________

g = Graph(num_nodes=5, num_edges=6,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32), 'w': Scheme(shape=(...t32)}
      edata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'w': Scheme(shape=(1,), dtype=torch.float32)})
idtype = torch.int32, reducer = 'sum'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_weighted_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.ndata['w'] = F.randn((g.number_of_nodes(), 1))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
        g.edata['w'] = F.randn((g.number_of_edges(), 1))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', 'w', op=reducer)

tests/compute/test_readout.py:81: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'sum'
x = tensor([[-0.3566, -0.5339, -0.9003],
        [-1.6285,  0.8738,  0.7481],
        [-1.8880, -0.4768, -0.4826],
        [ 0.2898, -0.2470, -0.3093],
        [-1.0060,  0.6507, -1.0606]])
offsets = tensor([0, 5])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
_________________ test_weighted_reduce_readout[sum-g3-idtype1] _________________

g = Graph(num_nodes=5, num_edges=6,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32), 'w': Scheme(shape=(...t32)}
      edata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'w': Scheme(shape=(1,), dtype=torch.float32)})
idtype = torch.int64, reducer = 'sum'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_weighted_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.ndata['w'] = F.randn((g.number_of_nodes(), 1))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
        g.edata['w'] = F.randn((g.number_of_edges(), 1))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', 'w', op=reducer)

tests/compute/test_readout.py:81: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'sum'
x = tensor([[ 0.2539,  0.2932,  0.0576],
        [ 0.0495,  0.1193, -0.1960],
        [-0.4644, -0.1863, -0.1056],
        [-1.3326, -0.5687,  0.5925],
        [ 0.1628,  0.2018, -0.2711]])
offsets = tensor([0, 5])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
_________________ test_weighted_reduce_readout[sum-g4-idtype0] _________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32), 'w': Scheme(shape...t32)}
      edata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'w': Scheme(shape=(1,), dtype=torch.float32)})
idtype = torch.int32, reducer = 'sum'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_weighted_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.ndata['w'] = F.randn((g.number_of_nodes(), 1))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
        g.edata['w'] = F.randn((g.number_of_edges(), 1))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', 'w', op=reducer)

tests/compute/test_readout.py:81: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'sum'
x = tensor([[ 0.1003, -0.0900, -0.1079],
        [ 0.7191, -0.5974,  0.0046],
        [ 0.1975,  0.1559,  0.0793],
       ....9948],
        [-0.5396, -1.6410,  0.0145],
        [ 0.6199, -0.8014,  0.2945],
        [ 0.0652, -0.1054, -0.0825]])
offsets = tensor([ 0, 10])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
_________________ test_weighted_reduce_readout[sum-g4-idtype1] _________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32), 'w': Scheme(shape...t32)}
      edata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'w': Scheme(shape=(1,), dtype=torch.float32)})
idtype = torch.int64, reducer = 'sum'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_weighted_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.ndata['w'] = F.randn((g.number_of_nodes(), 1))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
        g.edata['w'] = F.randn((g.number_of_edges(), 1))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', 'w', op=reducer)

tests/compute/test_readout.py:81: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'sum'
x = tensor([[-0.4924,  0.7035,  0.9131],
        [-0.3724, -0.0819,  0.2000],
        [ 0.0489,  0.0233, -0.0277],
       ....2523],
        [-0.4642,  0.2772, -0.7558],
        [ 0.5147, -0.7762, -2.0423],
        [ 1.2267, -1.1129,  0.2893]])
offsets = tensor([ 0, 10])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
_________________ test_weighted_reduce_readout[sum-g5-idtype0] _________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32), 'w': Scheme(shape...t32)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float32), 'h': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int32, reducer = 'sum'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_weighted_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.ndata['w'] = F.randn((g.number_of_nodes(), 1))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
        g.edata['w'] = F.randn((g.number_of_edges(), 1))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', 'w', op=reducer)

tests/compute/test_readout.py:81: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'sum'
x = tensor([[-0.1222, -0.5543, -0.0077],
        [-0.7472,  0.5584,  0.4276],
        [ 0.8668,  0.8769, -1.4978],
       ....0272],
        [ 1.1129, -0.0422,  0.9648],
        [ 0.9933, -0.6892, -0.6968],
        [ 0.0500,  0.1541,  0.0567]])
offsets = tensor([ 0, 10])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
_________________ test_weighted_reduce_readout[sum-g5-idtype1] _________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32), 'w': Scheme(shape...t32)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float32), 'h': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int64, reducer = 'sum'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_weighted_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.ndata['w'] = F.randn((g.number_of_nodes(), 1))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
        g.edata['w'] = F.randn((g.number_of_edges(), 1))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', 'w', op=reducer)

tests/compute/test_readout.py:81: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'sum'
x = tensor([[-0.5806, -0.6718,  0.3574],
        [-1.0148, -0.6845,  2.7806],
        [-0.2204, -0.4630,  0.1793],
       ....0114],
        [ 1.9290,  2.8398,  3.1674],
        [ 0.2915,  1.0609,  1.3786],
        [ 1.6323, -2.5928, -0.6880]])
offsets = tensor([ 0, 10])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
_________________ test_weighted_reduce_readout[sum-g6-idtype0] _________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32), 'w': Scheme(shape...=(), dtype=torch.float32), 'h': Scheme(shape=(2,), dtype=torch.float32), 'w': Scheme(shape=(1,), dtype=torch.float32)})
idtype = torch.int32, reducer = 'sum'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_weighted_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.ndata['w'] = F.randn((g.number_of_nodes(), 1))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
        g.edata['w'] = F.randn((g.number_of_edges(), 1))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', 'w', op=reducer)

tests/compute/test_readout.py:81: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'sum'
x = tensor([[ 0.9680, -1.0404,  0.8732],
        [-0.0295,  1.0369, -0.2261],
        [-0.7792,  0.5111, -0.0585],
       ....5771],
        [ 1.0211, -0.5391, -0.6120],
        [-0.1570,  2.5063,  1.9306],
        [ 0.5131,  0.9295, -0.8811]])
offsets = tensor([ 0, 10])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
_________________ test_weighted_reduce_readout[sum-g6-idtype1] _________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32), 'w': Scheme(shape...=(), dtype=torch.float32), 'h': Scheme(shape=(2,), dtype=torch.float32), 'w': Scheme(shape=(1,), dtype=torch.float32)})
idtype = torch.int64, reducer = 'sum'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_weighted_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.ndata['w'] = F.randn((g.number_of_nodes(), 1))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
        g.edata['w'] = F.randn((g.number_of_edges(), 1))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', 'w', op=reducer)

tests/compute/test_readout.py:81: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'sum'
x = tensor([[ 0.5698, -0.1317, -0.5012],
        [-0.2583, -0.4771, -1.7902],
        [-0.0925,  0.1902,  0.7287],
       ....3116],
        [ 0.9801, -0.0865,  0.9227],
        [ 0.4284,  5.6297,  0.1502],
        [ 0.2716,  0.7330,  0.3709]])
offsets = tensor([ 0, 10])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
_________________ test_weighted_reduce_readout[max-g0-idtype0] _________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32), 'w': Scheme(shape=...t32)}
      edata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'w': Scheme(shape=(1,), dtype=torch.float32)})
idtype = torch.int32, reducer = 'max'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_weighted_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.ndata['w'] = F.randn((g.number_of_nodes(), 1))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
        g.edata['w'] = F.randn((g.number_of_edges(), 1))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', 'w', op=reducer)

tests/compute/test_readout.py:81: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'max'
x = tensor([[-0.0879, -0.0826,  0.0020],
        [-0.4365,  0.3302, -0.1149],
        [-0.2789, -0.4708,  0.7520],
       ....5525],
        [ 1.5727, -1.0589, -0.8505],
        [-0.3607, -0.0251, -0.2194],
        [ 0.7491,  0.9505, -1.2048]])
offsets = tensor([0, 4, 7, 9])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
_________________ test_weighted_reduce_readout[max-g0-idtype1] _________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32), 'w': Scheme(shape=...t32)}
      edata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'w': Scheme(shape=(1,), dtype=torch.float32)})
idtype = torch.int64, reducer = 'max'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_weighted_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.ndata['w'] = F.randn((g.number_of_nodes(), 1))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
        g.edata['w'] = F.randn((g.number_of_edges(), 1))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', 'w', op=reducer)

tests/compute/test_readout.py:81: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'max'
x = tensor([[ 0.4921, -0.2597, -0.1347],
        [-3.4025, -0.0409,  1.7869],
        [-3.8097,  1.1494,  2.0548],
       ....7685],
        [-0.7364,  0.2780,  0.0973],
        [-0.3737,  0.0868, -0.0499],
        [ 0.1862, -2.4873, -2.3914]])
offsets = tensor([0, 4, 7, 9])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
_________________ test_weighted_reduce_readout[max-g1-idtype0] _________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32), 'w': Scheme(shape...t32)}
      edata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'w': Scheme(shape=(1,), dtype=torch.float32)})
idtype = torch.int32, reducer = 'max'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_weighted_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.ndata['w'] = F.randn((g.number_of_nodes(), 1))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
        g.edata['w'] = F.randn((g.number_of_edges(), 1))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', 'w', op=reducer)

tests/compute/test_readout.py:81: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'max'
x = tensor([[-1.9412, -0.6770,  0.7690],
        [ 0.3755, -1.4493, -0.5529],
        [ 1.0548,  0.2091,  0.0991],
       ....3125],
        [-0.3521,  0.1380, -0.0551],
        [-0.0052,  0.1446, -0.0127],
        [ 0.2298,  0.1160, -0.2141]])
offsets = tensor([ 0, 10])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
_________________ test_weighted_reduce_readout[max-g1-idtype1] _________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32), 'w': Scheme(shape...t32)}
      edata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'w': Scheme(shape=(1,), dtype=torch.float32)})
idtype = torch.int64, reducer = 'max'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_weighted_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.ndata['w'] = F.randn((g.number_of_nodes(), 1))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
        g.edata['w'] = F.randn((g.number_of_edges(), 1))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', 'w', op=reducer)

tests/compute/test_readout.py:81: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'max'
x = tensor([[-4.0073e-01, -7.7875e-01, -3.5032e-01],
        [-3.3865e-01, -3.4744e-01, -2.0073e+00],
        [-3.3429e-01...7e-02,  1.9831e-01],
        [-6.3930e-02,  5.9300e-01,  1.9979e+00],
        [ 6.4975e-01, -6.8334e-02, -8.4643e-01]])
offsets = tensor([ 0, 10])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
_________________ test_weighted_reduce_readout[max-g2-idtype0] _________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32), 'w': Scheme(shape...t32)}
      edata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'w': Scheme(shape=(1,), dtype=torch.float32)})
idtype = torch.int32, reducer = 'max'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_weighted_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.ndata['w'] = F.randn((g.number_of_nodes(), 1))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
        g.edata['w'] = F.randn((g.number_of_edges(), 1))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', 'w', op=reducer)

tests/compute/test_readout.py:81: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'max'
x = tensor([[ 1.7226,  0.4079,  0.8019],
        [ 0.1403,  0.0506,  0.4294],
        [-0.2672, -0.1885, -0.0904],
       ....7886],
        [-0.3799,  0.6025,  2.4005],
        [ 0.1458,  0.2526,  0.3558],
        [ 1.5102,  1.8874,  0.0850]])
offsets = tensor([ 0, 10])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
_________________ test_weighted_reduce_readout[max-g2-idtype1] _________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32), 'w': Scheme(shape...t32)}
      edata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'w': Scheme(shape=(1,), dtype=torch.float32)})
idtype = torch.int64, reducer = 'max'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_weighted_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.ndata['w'] = F.randn((g.number_of_nodes(), 1))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
        g.edata['w'] = F.randn((g.number_of_edges(), 1))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', 'w', op=reducer)

tests/compute/test_readout.py:81: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'max'
x = tensor([[ 0.6061, -0.7247, -0.8487],
        [-0.0567,  0.0166,  0.1725],
        [ 0.6027,  1.9727, -0.1648],
       ....7824],
        [ 0.0152, -0.0384, -0.0352],
        [-1.9871,  0.3338,  0.9092],
        [ 0.0436, -0.3636,  0.3027]])
offsets = tensor([ 0, 10])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
_________________ test_weighted_reduce_readout[max-g3-idtype0] _________________

g = Graph(num_nodes=5, num_edges=6,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32), 'w': Scheme(shape=(...t32)}
      edata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'w': Scheme(shape=(1,), dtype=torch.float32)})
idtype = torch.int32, reducer = 'max'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_weighted_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.ndata['w'] = F.randn((g.number_of_nodes(), 1))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
        g.edata['w'] = F.randn((g.number_of_edges(), 1))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', 'w', op=reducer)

tests/compute/test_readout.py:81: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'max'
x = tensor([[-1.7234,  0.5737, -0.2500],
        [ 0.0092, -0.0608,  0.0025],
        [ 0.5151,  0.3537,  0.1296],
        [-0.0327,  0.0353,  0.0146],
        [-0.2840, -0.0938, -0.1171]])
offsets = tensor([0, 5])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
_________________ test_weighted_reduce_readout[max-g3-idtype1] _________________

g = Graph(num_nodes=5, num_edges=6,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32), 'w': Scheme(shape=(...t32)}
      edata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'w': Scheme(shape=(1,), dtype=torch.float32)})
idtype = torch.int64, reducer = 'max'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_weighted_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.ndata['w'] = F.randn((g.number_of_nodes(), 1))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
        g.edata['w'] = F.randn((g.number_of_edges(), 1))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', 'w', op=reducer)

tests/compute/test_readout.py:81: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'max'
x = tensor([[-0.0271,  0.1810,  0.1791],
        [ 0.3765, -0.1090, -0.1773],
        [-0.6442,  0.8015,  0.8410],
        [ 0.1916,  0.2012, -0.0722],
        [ 1.4329,  2.6177, -0.5049]])
offsets = tensor([0, 5])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
_________________ test_weighted_reduce_readout[max-g4-idtype0] _________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32), 'w': Scheme(shape...t32)}
      edata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'w': Scheme(shape=(1,), dtype=torch.float32)})
idtype = torch.int32, reducer = 'max'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_weighted_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.ndata['w'] = F.randn((g.number_of_nodes(), 1))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
        g.edata['w'] = F.randn((g.number_of_edges(), 1))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', 'w', op=reducer)

tests/compute/test_readout.py:81: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'max'
x = tensor([[-0.0607, -0.0991, -0.1694],
        [-0.2003, -0.4722,  0.4881],
        [ 0.2429,  0.1810,  0.1022],
       ....0652],
        [-0.0423, -0.4084, -0.1811],
        [-2.5105,  1.8462,  0.2897],
        [-0.4721, -1.4015,  0.3187]])
offsets = tensor([ 0, 10])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
_________________ test_weighted_reduce_readout[max-g4-idtype1] _________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32), 'w': Scheme(shape...t32)}
      edata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'w': Scheme(shape=(1,), dtype=torch.float32)})
idtype = torch.int64, reducer = 'max'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_weighted_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.ndata['w'] = F.randn((g.number_of_nodes(), 1))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
        g.edata['w'] = F.randn((g.number_of_edges(), 1))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', 'w', op=reducer)

tests/compute/test_readout.py:81: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'max'
x = tensor([[-2.8877,  0.6455, -3.0974],
        [ 0.3863, -2.0574,  2.6114],
        [-0.3358,  0.8232, -0.1851],
       ....3492],
        [ 0.4503,  0.0439, -0.0213],
        [-0.0146,  0.2251,  0.0260],
        [ 0.0678,  0.0282, -0.0149]])
offsets = tensor([ 0, 10])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
_________________ test_weighted_reduce_readout[max-g5-idtype0] _________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32), 'w': Scheme(shape...t32)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float32), 'h': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int32, reducer = 'max'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_weighted_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.ndata['w'] = F.randn((g.number_of_nodes(), 1))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
        g.edata['w'] = F.randn((g.number_of_edges(), 1))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', 'w', op=reducer)

tests/compute/test_readout.py:81: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'max'
x = tensor([[-1.0212,  0.0493, -0.2019],
        [ 0.3759, -0.0021,  0.2231],
        [ 0.0205,  0.5992,  0.1468],
       ....6854],
        [-0.0567, -0.1725,  0.0459],
        [ 0.0893, -0.3036,  0.7135],
        [ 0.0231, -0.6701,  0.1761]])
offsets = tensor([ 0, 10])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
_________________ test_weighted_reduce_readout[max-g5-idtype1] _________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32), 'w': Scheme(shape...t32)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float32), 'h': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int64, reducer = 'max'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_weighted_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.ndata['w'] = F.randn((g.number_of_nodes(), 1))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
        g.edata['w'] = F.randn((g.number_of_edges(), 1))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', 'w', op=reducer)

tests/compute/test_readout.py:81: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'max'
x = tensor([[ 0.0308, -0.0115,  0.0434],
        [-0.4963,  0.5435, -1.3471],
        [ 0.0979, -0.2925, -0.2294],
       ....4143],
        [-0.0064,  0.0050, -0.0085],
        [-0.4626,  3.2726,  0.3691],
        [ 0.4667, -0.7063, -0.5483]])
offsets = tensor([ 0, 10])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
_________________ test_weighted_reduce_readout[max-g6-idtype0] _________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32), 'w': Scheme(shape...=(), dtype=torch.float32), 'h': Scheme(shape=(2,), dtype=torch.float32), 'w': Scheme(shape=(1,), dtype=torch.float32)})
idtype = torch.int32, reducer = 'max'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_weighted_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.ndata['w'] = F.randn((g.number_of_nodes(), 1))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
        g.edata['w'] = F.randn((g.number_of_edges(), 1))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', 'w', op=reducer)

tests/compute/test_readout.py:81: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'max'
x = tensor([[-0.4035,  2.9957,  0.5774],
        [ 0.5056, -0.3924, -0.5739],
        [-1.4574,  1.2098, -0.4511],
       ....0713],
        [ 1.2320, -0.3895, -0.0845],
        [ 2.9534, -3.7960, -1.5338],
        [-0.0540,  0.4729,  0.6305]])
offsets = tensor([ 0, 10])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
_________________ test_weighted_reduce_readout[max-g6-idtype1] _________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32), 'w': Scheme(shape...=(), dtype=torch.float32), 'h': Scheme(shape=(2,), dtype=torch.float32), 'w': Scheme(shape=(1,), dtype=torch.float32)})
idtype = torch.int64, reducer = 'max'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_weighted_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.ndata['w'] = F.randn((g.number_of_nodes(), 1))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
        g.edata['w'] = F.randn((g.number_of_edges(), 1))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', 'w', op=reducer)

tests/compute/test_readout.py:81: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'max'
x = tensor([[-0.3097, -0.5468,  0.4783],
        [ 0.6620, -0.0558, -0.2572],
        [ 0.1138, -0.0472, -0.0407],
       ....4437],
        [-0.1779,  0.1042,  0.2034],
        [ 0.0322, -0.0650,  0.2189],
        [-0.0704, -1.0004,  0.3870]])
offsets = tensor([ 0, 10])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
________________ test_weighted_reduce_readout[mean-g0-idtype0] _________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32), 'w': Scheme(shape=...t32)}
      edata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'w': Scheme(shape=(1,), dtype=torch.float32)})
idtype = torch.int32, reducer = 'mean'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_weighted_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.ndata['w'] = F.randn((g.number_of_nodes(), 1))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
        g.edata['w'] = F.randn((g.number_of_edges(), 1))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', 'w', op=reducer)

tests/compute/test_readout.py:81: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:46: in segment_reduce
    rst = F.segment_reduce('sum', value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'sum'
x = tensor([[-4.0132e-01, -8.1206e-03,  1.9132e-01],
        [ 6.0274e-01,  1.2888e+00,  9.5444e-02],
        [-2.2819e+00...4e+00, -4.6069e+00],
        [-3.0383e-02,  1.5486e-01, -3.0273e-01],
        [-1.2977e+00,  2.8006e-01, -2.5003e+00]])
offsets = tensor([0, 4, 7, 9])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
________________ test_weighted_reduce_readout[mean-g0-idtype1] _________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32), 'w': Scheme(shape=...t32)}
      edata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'w': Scheme(shape=(1,), dtype=torch.float32)})
idtype = torch.int64, reducer = 'mean'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_weighted_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.ndata['w'] = F.randn((g.number_of_nodes(), 1))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
        g.edata['w'] = F.randn((g.number_of_edges(), 1))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', 'w', op=reducer)

tests/compute/test_readout.py:81: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:46: in segment_reduce
    rst = F.segment_reduce('sum', value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'sum'
x = tensor([[ 0.7265, -1.3938, -0.3944],
        [-0.8316,  0.6839,  1.1794],
        [-1.7994,  0.6211,  3.5598],
       ....2780],
        [ 0.8004, -0.0206,  0.0812],
        [ 0.2971,  0.0442,  0.5900],
        [ 0.7341, -1.0076,  0.3092]])
offsets = tensor([0, 4, 7, 9])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
________________ test_weighted_reduce_readout[mean-g1-idtype0] _________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32), 'w': Scheme(shape...t32)}
      edata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'w': Scheme(shape=(1,), dtype=torch.float32)})
idtype = torch.int32, reducer = 'mean'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_weighted_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.ndata['w'] = F.randn((g.number_of_nodes(), 1))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
        g.edata['w'] = F.randn((g.number_of_edges(), 1))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', 'w', op=reducer)

tests/compute/test_readout.py:81: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:46: in segment_reduce
    rst = F.segment_reduce('sum', value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'sum'
x = tensor([[ 4.3781e-01, -1.9306e-01,  1.0732e+00],
        [ 1.5201e-01, -4.9974e-03,  9.5987e-03],
        [ 8.4022e-02...9e-01, -1.0259e+00],
        [ 1.3449e-02, -3.6590e-02,  1.0202e-02],
        [ 1.3410e-03, -1.9022e-01, -1.8057e-02]])
offsets = tensor([ 0, 10])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
________________ test_weighted_reduce_readout[mean-g1-idtype1] _________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32), 'w': Scheme(shape...t32)}
      edata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'w': Scheme(shape=(1,), dtype=torch.float32)})
idtype = torch.int64, reducer = 'mean'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_weighted_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.ndata['w'] = F.randn((g.number_of_nodes(), 1))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
        g.edata['w'] = F.randn((g.number_of_edges(), 1))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', 'w', op=reducer)

tests/compute/test_readout.py:81: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:46: in segment_reduce
    rst = F.segment_reduce('sum', value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'sum'
x = tensor([[ 2.1439,  2.0308, -1.0265],
        [-0.0661,  0.0459,  0.1228],
        [-0.6473, -0.5812, -0.1082],
       ....7476],
        [ 0.0767, -0.1040, -0.0318],
        [-0.6893, -0.2154, -0.4976],
        [ 0.0307, -0.6631,  0.5475]])
offsets = tensor([ 0, 10])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
________________ test_weighted_reduce_readout[mean-g2-idtype0] _________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32), 'w': Scheme(shape...t32)}
      edata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'w': Scheme(shape=(1,), dtype=torch.float32)})
idtype = torch.int32, reducer = 'mean'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_weighted_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.ndata['w'] = F.randn((g.number_of_nodes(), 1))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
        g.edata['w'] = F.randn((g.number_of_edges(), 1))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', 'w', op=reducer)

tests/compute/test_readout.py:81: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:46: in segment_reduce
    rst = F.segment_reduce('sum', value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'sum'
x = tensor([[ 1.8865,  5.6161, -1.2546],
        [ 0.0079,  0.0154,  0.0075],
        [-0.0181,  0.2520,  0.2798],
       ....0702],
        [ 0.5145, -0.1976, -0.2253],
        [-0.1450, -0.2248,  0.1099],
        [ 0.0996,  0.0460,  0.0462]])
offsets = tensor([ 0, 10])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
________________ test_weighted_reduce_readout[mean-g2-idtype1] _________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32), 'w': Scheme(shape...t32)}
      edata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'w': Scheme(shape=(1,), dtype=torch.float32)})
idtype = torch.int64, reducer = 'mean'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_weighted_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.ndata['w'] = F.randn((g.number_of_nodes(), 1))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
        g.edata['w'] = F.randn((g.number_of_edges(), 1))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', 'w', op=reducer)

tests/compute/test_readout.py:81: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:46: in segment_reduce
    rst = F.segment_reduce('sum', value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'sum'
x = tensor([[ 0.1665, -0.0128, -0.0827],
        [ 0.0415, -0.1121, -0.1093],
        [ 0.1783, -0.0734,  0.0322],
       ....4245],
        [-0.0314,  1.1551, -0.8841],
        [-0.1896, -0.0453, -0.0073],
        [ 0.5840,  0.1745,  0.2435]])
offsets = tensor([ 0, 10])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
________________ test_weighted_reduce_readout[mean-g3-idtype0] _________________

g = Graph(num_nodes=5, num_edges=6,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32), 'w': Scheme(shape=(...t32)}
      edata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'w': Scheme(shape=(1,), dtype=torch.float32)})
idtype = torch.int32, reducer = 'mean'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_weighted_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.ndata['w'] = F.randn((g.number_of_nodes(), 1))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
        g.edata['w'] = F.randn((g.number_of_edges(), 1))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', 'w', op=reducer)

tests/compute/test_readout.py:81: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:46: in segment_reduce
    rst = F.segment_reduce('sum', value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'sum'
x = tensor([[-0.1922,  1.4144,  0.8833],
        [-0.1818,  0.7012, -0.6259],
        [-0.0160,  0.0318, -0.1081],
        [ 1.0078, -0.1794,  1.2910],
        [-1.2566, -0.7348,  0.7364]])
offsets = tensor([0, 5])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
________________ test_weighted_reduce_readout[mean-g3-idtype1] _________________

g = Graph(num_nodes=5, num_edges=6,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32), 'w': Scheme(shape=(...t32)}
      edata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'w': Scheme(shape=(1,), dtype=torch.float32)})
idtype = torch.int64, reducer = 'mean'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_weighted_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.ndata['w'] = F.randn((g.number_of_nodes(), 1))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
        g.edata['w'] = F.randn((g.number_of_edges(), 1))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', 'w', op=reducer)

tests/compute/test_readout.py:81: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:46: in segment_reduce
    rst = F.segment_reduce('sum', value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'sum'
x = tensor([[-0.1942, -0.2705,  0.4601],
        [-0.1262,  0.3511, -0.2058],
        [-0.3586,  0.2104,  1.5683],
        [ 1.0828,  0.3755,  0.0134],
        [ 0.9884,  2.1038, -0.0933]])
offsets = tensor([0, 5])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
________________ test_weighted_reduce_readout[mean-g4-idtype0] _________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32), 'w': Scheme(shape...t32)}
      edata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'w': Scheme(shape=(1,), dtype=torch.float32)})
idtype = torch.int32, reducer = 'mean'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_weighted_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.ndata['w'] = F.randn((g.number_of_nodes(), 1))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
        g.edata['w'] = F.randn((g.number_of_edges(), 1))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', 'w', op=reducer)

tests/compute/test_readout.py:81: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:46: in segment_reduce
    rst = F.segment_reduce('sum', value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'sum'
x = tensor([[ 2.2951e+00, -1.1828e+00,  2.5587e-01],
        [-4.9547e-02, -6.4390e-02, -9.1058e-02],
        [-7.8794e-01...0e-01, -4.4991e-01],
        [-1.9321e-03,  2.8101e-03, -8.0224e-04],
        [-9.9687e-01,  5.3161e-02, -6.3797e-02]])
offsets = tensor([ 0, 10])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
________________ test_weighted_reduce_readout[mean-g4-idtype1] _________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32), 'w': Scheme(shape...t32)}
      edata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'w': Scheme(shape=(1,), dtype=torch.float32)})
idtype = torch.int64, reducer = 'mean'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_weighted_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.ndata['w'] = F.randn((g.number_of_nodes(), 1))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
        g.edata['w'] = F.randn((g.number_of_edges(), 1))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', 'w', op=reducer)

tests/compute/test_readout.py:81: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:46: in segment_reduce
    rst = F.segment_reduce('sum', value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'sum'
x = tensor([[-0.4826, -2.2434, -0.0068],
        [ 0.4720, -1.4129, -0.9489],
        [-0.3964,  0.9945,  0.4331],
       ....3243],
        [ 0.2590,  0.2810,  1.9278],
        [ 0.0173,  0.0087,  0.0111],
        [-3.0435,  0.9749, -0.9331]])
offsets = tensor([ 0, 10])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
________________ test_weighted_reduce_readout[mean-g5-idtype0] _________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32), 'w': Scheme(shape...t32)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float32), 'h': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int32, reducer = 'mean'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_weighted_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.ndata['w'] = F.randn((g.number_of_nodes(), 1))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
        g.edata['w'] = F.randn((g.number_of_edges(), 1))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', 'w', op=reducer)

tests/compute/test_readout.py:81: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:46: in segment_reduce
    rst = F.segment_reduce('sum', value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'sum'
x = tensor([[-1.7193e-01,  5.8989e-01, -1.5774e+00],
        [ 1.0644e-02,  1.8932e-01, -1.1007e-01],
        [ 1.1406e-01...0e+00,  4.0478e+00],
        [ 4.2893e-02,  1.4893e-01, -5.5555e-01],
        [-4.6487e-01,  1.8672e-01,  1.2928e+00]])
offsets = tensor([ 0, 10])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
________________ test_weighted_reduce_readout[mean-g5-idtype1] _________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32), 'w': Scheme(shape...t32)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float32), 'h': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int64, reducer = 'mean'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_weighted_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.ndata['w'] = F.randn((g.number_of_nodes(), 1))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
        g.edata['w'] = F.randn((g.number_of_edges(), 1))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', 'w', op=reducer)

tests/compute/test_readout.py:81: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:46: in segment_reduce
    rst = F.segment_reduce('sum', value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'sum'
x = tensor([[ 1.2014, -0.5730, -0.5063],
        [ 0.4955, -0.2203, -0.7970],
        [ 0.2087,  0.3711, -0.3147],
       ....8257],
        [-4.2873,  1.0210,  1.0732],
        [ 0.4456, -0.3542,  0.1283],
        [ 0.4178, -0.9784, -0.8325]])
offsets = tensor([ 0, 10])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
________________ test_weighted_reduce_readout[mean-g6-idtype0] _________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32), 'w': Scheme(shape...=(), dtype=torch.float32), 'h': Scheme(shape=(2,), dtype=torch.float32), 'w': Scheme(shape=(1,), dtype=torch.float32)})
idtype = torch.int32, reducer = 'mean'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_weighted_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.ndata['w'] = F.randn((g.number_of_nodes(), 1))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
        g.edata['w'] = F.randn((g.number_of_edges(), 1))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', 'w', op=reducer)

tests/compute/test_readout.py:81: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:46: in segment_reduce
    rst = F.segment_reduce('sum', value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'sum'
x = tensor([[ 0.7105, -0.8271,  0.8896],
        [ 0.0520, -0.0729,  0.1912],
        [-0.0373,  0.1154,  0.0228],
       ....9000],
        [ 1.0504,  1.7663, -0.4968],
        [ 2.6396, -1.0414,  0.0641],
        [ 0.0183, -0.0402,  0.1095]])
offsets = tensor([ 0, 10])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
________________ test_weighted_reduce_readout[mean-g6-idtype1] _________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32), 'w': Scheme(shape...=(), dtype=torch.float32), 'h': Scheme(shape=(2,), dtype=torch.float32), 'w': Scheme(shape=(1,), dtype=torch.float32)})
idtype = torch.int64, reducer = 'mean'

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    @pytest.mark.parametrize('reducer', ['sum', 'max', 'mean'])
    def test_weighted_reduce_readout(g, idtype, reducer):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.ndata['w'] = F.randn((g.number_of_nodes(), 1))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
        g.edata['w'] = F.randn((g.number_of_edges(), 1))
    
        # Test.1: node readout
>       x = dgl.readout_nodes(g, 'h', 'w', op=reducer)

tests/compute/test_readout.py:81: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:88: in readout_nodes
    return segment.segment_reduce(graph.batch_num_nodes(ntype), x, reducer=op)
python/dgl/ops/segment.py:46: in segment_reduce
    rst = F.segment_reduce('sum', value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'sum'
x = tensor([[-0.1080,  0.3017,  0.0071],
        [-0.3957,  0.9836, -0.4031],
        [ 1.2907,  0.3195,  0.8144],
       ....1357],
        [ 0.8530,  0.5273, -2.2813],
        [ 0.9468,  0.4147,  5.6728],
        [-0.3218,  0.1146, -1.3979]])
offsets = tensor([ 0, 10])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
___________________________ test_softmax[g0-idtype0] ___________________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32)}
      edata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int32

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    def test_softmax(g, idtype):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
    
        # Test.1: node readout
>       x = dgl.softmax_nodes(g, 'h')

tests/compute/test_readout.py:192: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:284: in softmax_nodes
    return segment.segment_softmax(graph.batch_num_nodes(ntype), x)
python/dgl/ops/segment.py:98: in segment_softmax
    value_max = segment_reduce(seglen, value, reducer='max')
python/dgl/ops/segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'max'
x = tensor([[ 0.1967, -0.6725,  2.4015],
        [-0.5067, -1.1432,  0.3985],
        [ 0.6828, -1.0130,  1.7629],
       ....2395],
        [ 0.7006,  0.0037,  0.1807],
        [ 0.1524,  0.8077,  1.3621],
        [-1.6351,  0.5322,  0.0109]])
offsets = tensor([0, 4, 7, 9])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
___________________________ test_softmax[g0-idtype1] ___________________________

g = Graph(num_nodes=9, num_edges=15,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32)}
      edata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int64

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    def test_softmax(g, idtype):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
    
        # Test.1: node readout
>       x = dgl.softmax_nodes(g, 'h')

tests/compute/test_readout.py:192: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:284: in softmax_nodes
    return segment.segment_softmax(graph.batch_num_nodes(ntype), x)
python/dgl/ops/segment.py:98: in segment_softmax
    value_max = segment_reduce(seglen, value, reducer='max')
python/dgl/ops/segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'max'
x = tensor([[ 1.9846,  0.3620, -0.8714],
        [-0.0357,  0.2804, -0.6070],
        [-0.1033, -1.3570, -1.0853],
       ....2408],
        [ 0.6750,  0.1024, -0.9977],
        [-0.6691, -0.2334,  0.2348],
        [ 0.1748, -0.4843,  0.3442]])
offsets = tensor([0, 4, 7, 9])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
___________________________ test_softmax[g1-idtype0] ___________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32)}
      edata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int32

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    def test_softmax(g, idtype):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
    
        # Test.1: node readout
>       x = dgl.softmax_nodes(g, 'h')

tests/compute/test_readout.py:192: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:284: in softmax_nodes
    return segment.segment_softmax(graph.batch_num_nodes(ntype), x)
python/dgl/ops/segment.py:98: in segment_softmax
    value_max = segment_reduce(seglen, value, reducer='max')
python/dgl/ops/segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'max'
x = tensor([[-0.0597,  1.3889, -0.6126],
        [-0.5570,  1.0783,  0.9740],
        [ 1.2673,  0.1520,  1.5479],
       ....8510],
        [ 1.8830,  0.3579, -1.4092],
        [ 0.9364,  0.2506, -2.0436],
        [-1.1689, -1.1922, -0.5850]])
offsets = tensor([ 0, 10])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
___________________________ test_softmax[g1-idtype1] ___________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32)}
      edata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int64

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    def test_softmax(g, idtype):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
    
        # Test.1: node readout
>       x = dgl.softmax_nodes(g, 'h')

tests/compute/test_readout.py:192: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:284: in softmax_nodes
    return segment.segment_softmax(graph.batch_num_nodes(ntype), x)
python/dgl/ops/segment.py:98: in segment_softmax
    value_max = segment_reduce(seglen, value, reducer='max')
python/dgl/ops/segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'max'
x = tensor([[ 0.4272,  1.1201,  0.2017],
        [-0.3338, -0.3309, -0.2551],
        [-0.6120,  1.1263, -0.5782],
       ....8505],
        [ 0.8496,  0.8871, -0.0713],
        [-0.4711,  1.1334, -0.8519],
        [ 0.8204,  1.3008, -0.2745]])
offsets = tensor([ 0, 10])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
___________________________ test_softmax[g2-idtype0] ___________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32)}
      edata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int32

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    def test_softmax(g, idtype):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
    
        # Test.1: node readout
>       x = dgl.softmax_nodes(g, 'h')

tests/compute/test_readout.py:192: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:284: in softmax_nodes
    return segment.segment_softmax(graph.batch_num_nodes(ntype), x)
python/dgl/ops/segment.py:98: in segment_softmax
    value_max = segment_reduce(seglen, value, reducer='max')
python/dgl/ops/segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'max'
x = tensor([[-0.1719,  0.3248, -0.6994],
        [ 1.2418, -0.6625,  0.0904],
        [ 1.6104,  1.0404, -0.6698],
       ....0333],
        [-1.2233,  0.6460,  0.8322],
        [-0.7795, -0.4347, -0.0181],
        [-1.4169,  0.6527,  2.5954]])
offsets = tensor([ 0, 10])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
___________________________ test_softmax[g2-idtype1] ___________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32)}
      edata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int64

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    def test_softmax(g, idtype):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
    
        # Test.1: node readout
>       x = dgl.softmax_nodes(g, 'h')

tests/compute/test_readout.py:192: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:284: in softmax_nodes
    return segment.segment_softmax(graph.batch_num_nodes(ntype), x)
python/dgl/ops/segment.py:98: in segment_softmax
    value_max = segment_reduce(seglen, value, reducer='max')
python/dgl/ops/segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'max'
x = tensor([[-0.4628,  1.1958, -0.2905],
        [-0.6382, -1.2386,  0.8152],
        [-1.2265, -0.8516,  0.9339],
       ....0838],
        [-1.3084, -1.4412,  0.2435],
        [ 0.3760,  0.4614, -0.3473],
        [ 2.3081,  1.5383,  2.0966]])
offsets = tensor([ 0, 10])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
___________________________ test_softmax[g3-idtype0] ___________________________

g = Graph(num_nodes=5, num_edges=6,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32)}
      edata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int32

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    def test_softmax(g, idtype):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
    
        # Test.1: node readout
>       x = dgl.softmax_nodes(g, 'h')

tests/compute/test_readout.py:192: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:284: in softmax_nodes
    return segment.segment_softmax(graph.batch_num_nodes(ntype), x)
python/dgl/ops/segment.py:98: in segment_softmax
    value_max = segment_reduce(seglen, value, reducer='max')
python/dgl/ops/segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'max'
x = tensor([[-5.1826e-01,  1.7971e+00, -1.6173e-01],
        [ 8.5485e-01, -1.0742e+00, -8.9070e-01],
        [-3.2164e-01...7e-01,  1.0889e-03],
        [-1.6908e-01, -3.1009e-03,  7.4899e-01],
        [ 2.6297e-01,  8.9996e-02, -1.4489e+00]])
offsets = tensor([0, 5])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
___________________________ test_softmax[g3-idtype1] ___________________________

g = Graph(num_nodes=5, num_edges=6,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32)}
      edata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int64

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    def test_softmax(g, idtype):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
    
        # Test.1: node readout
>       x = dgl.softmax_nodes(g, 'h')

tests/compute/test_readout.py:192: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:284: in softmax_nodes
    return segment.segment_softmax(graph.batch_num_nodes(ntype), x)
python/dgl/ops/segment.py:98: in segment_softmax
    value_max = segment_reduce(seglen, value, reducer='max')
python/dgl/ops/segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'max'
x = tensor([[-0.3507,  0.7733, -0.3646],
        [-0.1068,  1.1619, -0.3371],
        [-0.8295,  1.1541,  1.0698],
        [-0.1500,  1.3992, -0.9389],
        [ 0.0714,  0.4344, -1.0874]])
offsets = tensor([0, 5])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
___________________________ test_softmax[g4-idtype0] ___________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32)}
      edata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int32

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    def test_softmax(g, idtype):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
    
        # Test.1: node readout
>       x = dgl.softmax_nodes(g, 'h')

tests/compute/test_readout.py:192: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:284: in softmax_nodes
    return segment.segment_softmax(graph.batch_num_nodes(ntype), x)
python/dgl/ops/segment.py:98: in segment_softmax
    value_max = segment_reduce(seglen, value, reducer='max')
python/dgl/ops/segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'max'
x = tensor([[ 0.0303, -0.9491,  0.5930],
        [-0.5404, -0.2061, -0.9835],
        [ 0.6187,  0.9843, -1.3149],
       ....5153],
        [-0.2069,  0.5945,  2.2292],
        [ 0.1288,  0.0337,  0.1120],
        [-0.0752, -1.3872, -1.1065]])
offsets = tensor([ 0, 10])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
___________________________ test_softmax[g4-idtype1] ___________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32)}
      edata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int64

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    def test_softmax(g, idtype):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
    
        # Test.1: node readout
>       x = dgl.softmax_nodes(g, 'h')

tests/compute/test_readout.py:192: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:284: in softmax_nodes
    return segment.segment_softmax(graph.batch_num_nodes(ntype), x)
python/dgl/ops/segment.py:98: in segment_softmax
    value_max = segment_reduce(seglen, value, reducer='max')
python/dgl/ops/segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'max'
x = tensor([[-0.7190, -1.3882, -0.1445],
        [ 0.2737,  1.1646, -0.5754],
        [ 0.0918,  1.5806, -0.5103],
       ....0502],
        [ 1.7123, -2.2311, -0.3382],
        [ 0.4354,  1.7656, -0.0854],
        [ 3.0380, -0.2942,  1.7542]])
offsets = tensor([ 0, 10])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
___________________________ test_softmax[g5-idtype0] ___________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32), 'h': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int32

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    def test_softmax(g, idtype):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
    
        # Test.1: node readout
>       x = dgl.softmax_nodes(g, 'h')

tests/compute/test_readout.py:192: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:284: in softmax_nodes
    return segment.segment_softmax(graph.batch_num_nodes(ntype), x)
python/dgl/ops/segment.py:98: in segment_softmax
    value_max = segment_reduce(seglen, value, reducer='max')
python/dgl/ops/segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'max'
x = tensor([[ 0.8460, -1.9264,  0.2734],
        [ 0.8382,  1.3117, -0.3001],
        [-0.0914, -0.3878,  0.5347],
       ....2002],
        [ 0.9320, -0.7572, -0.9204],
        [-0.2679,  1.9172,  1.6622],
        [ 1.7413, -0.2491,  0.5903]])
offsets = tensor([ 0, 10])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
___________________________ test_softmax[g5-idtype1] ___________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float32), 'h': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int64

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    def test_softmax(g, idtype):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
    
        # Test.1: node readout
>       x = dgl.softmax_nodes(g, 'h')

tests/compute/test_readout.py:192: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:284: in softmax_nodes
    return segment.segment_softmax(graph.batch_num_nodes(ntype), x)
python/dgl/ops/segment.py:98: in segment_softmax
    value_max = segment_reduce(seglen, value, reducer='max')
python/dgl/ops/segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'max'
x = tensor([[-0.1390, -0.3065,  0.4020],
        [-0.4218, -1.2013, -1.1217],
        [-0.2980, -1.4770,  0.8092],
       ....2537],
        [-0.6519,  1.6374, -1.7824],
        [-1.6721,  2.1290, -0.7148],
        [-0.2255, -0.4443,  1.0330]])
offsets = tensor([ 0, 10])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
___________________________ test_softmax[g6-idtype0] ___________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), 'h': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int32

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    def test_softmax(g, idtype):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
    
        # Test.1: node readout
>       x = dgl.softmax_nodes(g, 'h')

tests/compute/test_readout.py:192: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:284: in softmax_nodes
    return segment.segment_softmax(graph.batch_num_nodes(ntype), x)
python/dgl/ops/segment.py:98: in segment_softmax
    value_max = segment_reduce(seglen, value, reducer='max')
python/dgl/ops/segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'max'
x = tensor([[-1.6747, -0.7123,  0.6029],
        [ 2.0078,  0.8582,  0.4981],
        [ 0.9494, -1.2424,  0.5749],
       ....4684],
        [-0.4354, -0.9709,  0.6248],
        [ 0.6197,  0.8547,  0.4541],
        [ 0.0598,  0.4052, -0.9054]])
offsets = tensor([ 0, 10])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
___________________________ test_softmax[g6-idtype1] ___________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(3,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32), 'h': Scheme(shape=(2,), dtype=torch.float32)})
idtype = torch.int64

    @parametrize_idtype
    @pytest.mark.parametrize('g', get_cases(['homo'], exclude=['dglgraph']))
    def test_softmax(g, idtype):
        g = g.astype(idtype).to(F.ctx())
        g.ndata['h'] = F.randn((g.number_of_nodes(), 3))
        g.edata['h'] = F.randn((g.number_of_edges(), 2))
    
        # Test.1: node readout
>       x = dgl.softmax_nodes(g, 'h')

tests/compute/test_readout.py:192: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/readout.py:284: in softmax_nodes
    return segment.segment_softmax(graph.batch_num_nodes(ntype), x)
python/dgl/ops/segment.py:98: in segment_softmax
    value_max = segment_reduce(seglen, value, reducer='max')
python/dgl/ops/segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'max'
x = tensor([[ 0.0224,  0.0601,  0.8085],
        [ 1.0668,  0.7681,  0.9605],
        [ 1.2172, -2.2719,  1.0764],
       ....1963],
        [-1.2099, -0.9392, -0.9143],
        [ 0.8604,  0.4746, -2.2591],
        [ 1.9504,  1.0406, -0.5393]])
offsets = tensor([ 0, 10])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
______________________ test_spmm[idtype0-sum-add-shp0-g0] ______________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), msg = 'add', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43e8fd3f0>
op = 'add', reduce_op = 'sum'
lhs_data = tensor([[[[[[1.7988],
            [1.9738],
            [1.8492]]],


          [[[1.7822],
            [1.9642],
    ...1]]],


          [[[1.1523],
            [1.8091],
            [1.4629]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.8393]],

           [[1.1441]],

           [[1.7414]]]],



         [[[[1.7275]],

           [[1.667...



         [[[[1.9997]],

           [[1.3344]],

           [[1.4938]]]]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={}
      edata_schemes={})
torch.int32
u shape: torch.Size([30, 1, 2, 1, 3, 1]), e shape: torch.Size([100, 4, 1, 3, 1, 1])
SpMM(message func: add, reduce func: sum)
______________________ test_spmm[idtype0-sum-add-shp0-g1] ______________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), msg = 'add', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72ab4c3a8>
op = 'add', reduce_op = 'sum'
lhs_data = tensor([[[[[[1.6991],
            [1.3707],
            [1.4296]]],


          [[[1.5545],
            [1.1239],
    ...7]]],


          [[[1.8556],
            [1.3231],
            [1.3123]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.3227]],

           [[1.1139]],

           [[1.6915]]]],



         [[[[1.2954]],

           [[1.140...



         [[[[1.5742]],

           [[1.7554]],

           [[1.6285]]]]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 1, 2, 1, 3, 1]), e shape: torch.Size([300, 4, 1, 3, 1, 1])
SpMM(message func: add, reduce func: sum)
______________________ test_spmm[idtype0-sum-add-shp1-g0] ______________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1, 3), dtype=torch.float64)})
shp = ((3, 3), (1, 3)), msg = 'add', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb7058a0ee8>
op = 'add', reduce_op = 'sum'
lhs_data = tensor([[[1.6375, 1.7040, 1.9483],
         [1.6280, 1.7046, 1.6080],
         [1.4972, 1.8846, 1.8243]],

        [[1...7053],
         [1.7492, 1.3863, 1.6557],
         [1.0881, 1.8065, 1.9345]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[1.6039, 1.7525, 1.4452]],

        [[1.0998, 1.4284, 1.1972]],

        [[1.9479, 1.7439, 1.4954]],

       ...]],

        [[1.5731, 1.8088, 1.6805]],

        [[1.1558, 1.4835, 1.8496]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 3, 3]), e shape: torch.Size([100, 1, 3])
SpMM(message func: add, reduce func: sum)
______________________ test_spmm[idtype0-sum-add-shp1-g1] ______________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3, 3), (1, 3)), msg = 'add', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb7059d2ca8>
op = 'add', reduce_op = 'sum'
lhs_data = tensor([[[1.1872, 1.7318, 1.1416],
         [1.2360, 1.5106, 1.5875],
         [1.0379, 1.6934, 1.5637]],

        [[1...8325],
         [1.9628, 1.7323, 1.7679],
         [1.0295, 1.8598, 1.2713]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[1.3805, 1.9984, 1.3804]],

        [[1.2945, 1.3385, 1.1105]],

        [[1.1655, 1.7655, 1.7866]],

       ...]],

        [[1.7121, 1.4845, 1.9628]],

        [[1.2331, 1.8344, 1.5940]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 3, 3]), e shape: torch.Size([300, 1, 3])
SpMM(message func: add, reduce func: sum)
______________________ test_spmm[idtype0-sum-add-shp2-g0] ______________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float64)})
shp = ((1,), (3,)), msg = 'add', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43e528798>
op = 'add', reduce_op = 'sum'
lhs_data = tensor([[1.9404],
        [1.5371],
        [1.8836],
        [1.2111],
        [1.4610],
        [1.9024],
        [1...981],
        [1.6544],
        [1.4571],
        [1.8558],
        [1.3309]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.2955, 1.4762, 1.6936],
        [1.0636, 1.0177, 1.2249],
        [1.7148, 1.1689, 1.1909],
        [1.5305,... 1.3840],
        [1.0711, 1.1178, 1.5841],
        [1.2570, 1.6639, 1.4979]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1, 3), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 1]), e shape: torch.Size([100, 3])
SpMM(message func: add, reduce func: sum)
______________________ test_spmm[idtype0-sum-add-shp2-g1] ______________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (3,)), msg = 'add', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb7049c5ca8>
op = 'add', reduce_op = 'sum'
lhs_data = tensor([[1.0196],
        [1.5406],
        [1.3206],
        [1.6903],
        [1.2002],
        [1.1520],
        [1...255],
        [1.9055],
        [1.0519],
        [1.1033],
        [1.8468]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.2513, 1.3992, 1.1699],
        [1.1009, 1.7682, 1.4494],
        [1.7635, 1.5298, 1.4288],
        [1.3746,... 1.7827],
        [1.0477, 1.9134, 1.2878],
        [1.4969, 1.6770, 1.7768]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 1]), e shape: torch.Size([300, 3])
SpMM(message func: add, reduce func: sum)
______________________ test_spmm[idtype0-sum-add-shp3-g0] ______________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (1,)), msg = 'add', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72b278480>
op = 'add', reduce_op = 'sum'
lhs_data = tensor([[1.5347, 1.8398, 1.8706],
        [1.1133, 1.4291, 1.9496],
        [1.7728, 1.0964, 1.1372],
        [1.1396,... 1.9298],
        [1.7124, 1.3413, 1.8794],
        [1.7815, 1.6723, 1.7826]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.6344],
        [1.2153],
        [1.5949],
        [1.7928],
        [1.7063],
        [1.5364],
        [1...660],
        [1.6560],
        [1.2985],
        [1.5984],
        [1.2720]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 3]), e shape: torch.Size([100, 1])
SpMM(message func: add, reduce func: sum)
______________________ test_spmm[idtype0-sum-add-shp3-g1] ______________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (1,)), msg = 'add', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72b278168>
op = 'add', reduce_op = 'sum'
lhs_data = tensor([[1.6106, 1.5596, 1.6960],
        [1.3212, 1.1348, 1.2166],
        [1.7009, 1.1504, 1.2053],
        [1.6894,... 1.6111],
        [1.2993, 1.8395, 1.9852],
        [1.4233, 1.7940, 1.7565]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.9598],
        [1.1862],
        [1.1602],
        [1.7684],
        [1.8662],
        [1.8374],
        [1...194],
        [1.0936],
        [1.9684],
        [1.4152],
        [1.7579]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 3]), e shape: torch.Size([300, 1])
SpMM(message func: add, reduce func: sum)
______________________ test_spmm[idtype0-sum-add-shp4-g0] ______________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), msg = 'add', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb705a22510>
op = 'add', reduce_op = 'sum'
lhs_data = tensor([[1.6792],
        [1.7253],
        [1.9835],
        [1.5296],
        [1.9047],
        [1.0949],
        [1...156],
        [1.5139],
        [1.2748],
        [1.0495],
        [1.2839]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.8760],
        [1.2263],
        [1.1637],
        [1.3442],
        [1.8018],
        [1.1687],
        [1...451],
        [1.9836],
        [1.2814],
        [1.8288],
        [1.1753]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 1]), e shape: torch.Size([100, 1])
SpMM(message func: add, reduce func: sum)
______________________ test_spmm[idtype0-sum-add-shp4-g1] ______________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), msg = 'add', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72b1b8b88>
op = 'add', reduce_op = 'sum'
lhs_data = tensor([[1.6298],
        [1.3763],
        [1.5548],
        [1.0676],
        [1.0084],
        [1.4410],
        [1...982],
        [1.9721],
        [1.0902],
        [1.9185],
        [1.1756]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.6589],
        [1.6065],
        [1.0723],
        [1.4644],
        [1.5275],
        [1.4710],
        [1...053],
        [1.2265],
        [1.3697],
        [1.8278],
        [1.1642]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 1]), e shape: torch.Size([300, 1])
SpMM(message func: add, reduce func: sum)
______________________ test_spmm[idtype0-sum-add-shp5-g0] ______________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
shp = ((), ()), msg = 'add', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72b1b8b40>
op = 'add', reduce_op = 'sum'
lhs_data = tensor([1.3055, 1.4002, 1.0308, 1.8993, 1.3034, 1.1198, 1.2363, 1.1388, 1.4134,
        1.8861, 1.8042, 1.0242, 1.4917...2283, 1.5901, 1.5505, 1.5083, 1.6590, 1.1661,
        1.1199, 1.7411, 1.4893], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([1.0858, 1.3805, 1.4047, 1.2790, 1.7066, 1.8015, 1.4226, 1.9157, 1.0914,
        1.5754, 1.7787, 1.1304, 1.1554...7728, 1.1475, 1.7822, 1.0086, 1.5114, 1.0413, 1.7404, 1.2910,
        1.5359], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30]), e shape: torch.Size([100])
SpMM(message func: add, reduce func: sum)
______________________ test_spmm[idtype0-sum-add-shp5-g1] ______________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((), ()), msg = 'add', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43ebb8900>
op = 'add', reduce_op = 'sum'
lhs_data = tensor([1.7120, 1.6719, 1.2374, 1.4007, 1.3862, 1.4946, 1.4730, 1.4063, 1.0453,
        1.6854, 1.8517, 1.0627, 1.8465...7072, 1.8136, 1.2734, 1.9800, 1.5682, 1.4515,
        1.5760, 1.2391, 1.9564], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([1.7235, 1.0480, 1.3563, 1.5806, 1.6711, 1.2613, 1.7576, 1.4966, 1.1601,
        1.6759, 1.5343, 1.8077, 1.0834...3140, 1.5604, 1.4091, 1.7468, 1.8287, 1.2366,
        1.4802, 1.9232, 1.4034], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30]), e shape: torch.Size([300])
SpMM(message func: add, reduce func: sum)
______________________ test_spmm[idtype0-sum-sub-shp0-g0] ______________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), msg = 'sub', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43eb7c708>
op = 'add', reduce_op = 'sum'
lhs_data = tensor([[[[[[1.4805],
            [1.4058],
            [1.6997]]],


          [[[1.8374],
            [1.6162],
    ...8]]],


          [[[1.2626],
            [1.4467],
            [1.5986]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[-1.7703]],

           [[-1.4230]],

           [[-1.4124]]]],



         [[[[-1.1790]],

           [[-...       [[[[-1.5550]],

           [[-1.7774]],

           [[-1.3281]]]]]], dtype=torch.float64, grad_fn=<NegBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 1, 2, 1, 3, 1]), e shape: torch.Size([100, 4, 1, 3, 1, 1])
SpMM(message func: sub, reduce func: sum)
______________________ test_spmm[idtype0-sum-sub-shp0-g1] ______________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), msg = 'sub', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43eb3e240>
op = 'add', reduce_op = 'sum'
lhs_data = tensor([[[[[[1.2516],
            [1.3863],
            [1.2933]]],


          [[[1.1787],
            [1.9850],
    ...7]]],


          [[[1.0899],
            [1.6965],
            [1.1708]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[-1.1369]],

           [[-1.7458]],

           [[-1.3634]]]],



         [[[[-1.3718]],

           [[-...       [[[[-1.1812]],

           [[-1.4360]],

           [[-1.7928]]]]]], dtype=torch.float64, grad_fn=<NegBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 1, 2, 1, 3, 1]), e shape: torch.Size([300, 4, 1, 3, 1, 1])
SpMM(message func: sub, reduce func: sum)
______________________ test_spmm[idtype0-sum-sub-shp1-g0] ______________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1, 3), dtype=torch.float64)})
shp = ((3, 3), (1, 3)), msg = 'sub', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb7049c1cf0>
op = 'add', reduce_op = 'sum'
lhs_data = tensor([[[1.1791, 1.3372, 1.8695],
         [1.5076, 1.2234, 1.2826],
         [1.4083, 1.5960, 1.7498]],

        [[1...2752],
         [1.2401, 1.8662, 1.9926],
         [1.9347, 1.0054, 1.3359]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[-1.9264, -1.8565, -1.3119]],

        [[-1.1963, -1.8883, -1.7505]],

        [[-1.8197, -1.4651, -1.8237]],...1.4999, -1.9605, -1.4011]],

        [[-1.6672, -1.4501, -1.4816]]], dtype=torch.float64,
       grad_fn=<NegBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 3, 3]), e shape: torch.Size([100, 1, 3])
SpMM(message func: sub, reduce func: sum)
______________________ test_spmm[idtype0-sum-sub-shp1-g1] ______________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3, 3), (1, 3)), msg = 'sub', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb705accab0>
op = 'add', reduce_op = 'sum'
lhs_data = tensor([[[1.5206, 1.9339, 1.0833],
         [1.2695, 1.4418, 1.4243],
         [1.1122, 1.4374, 1.2612]],

        [[1...5994],
         [1.6962, 1.0114, 1.4815],
         [1.8979, 1.9793, 1.3215]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[-1.1712, -1.4582, -1.4906]],

        [[-1.1365, -1.9007, -1.7136]],

        [[-1.6691, -1.4544, -1.2016]],...1.2868, -1.5125, -1.1145]],

        [[-1.3599, -1.6483, -1.7072]]], dtype=torch.float64,
       grad_fn=<NegBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 3, 3]), e shape: torch.Size([300, 1, 3])
SpMM(message func: sub, reduce func: sum)
______________________ test_spmm[idtype0-sum-sub-shp2-g0] ______________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float64)})
shp = ((1,), (3,)), msg = 'sub', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43e528708>
op = 'add', reduce_op = 'sum'
lhs_data = tensor([[1.1123],
        [1.9208],
        [1.5630],
        [1.2850],
        [1.6032],
        [1.3009],
        [1...404],
        [1.4406],
        [1.9476],
        [1.4543],
        [1.3749]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.8453, -1.1886, -1.5568],
        [-1.3840, -1.6391, -1.1400],
        [-1.4007, -1.5679, -1.0582],
       ...  [-1.8283, -1.9624, -1.2588],
        [-1.1225, -1.4817, -1.4477]], dtype=torch.float64,
       grad_fn=<NegBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1, 3), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 1]), e shape: torch.Size([100, 3])
SpMM(message func: sub, reduce func: sum)
______________________ test_spmm[idtype0-sum-sub-shp2-g1] ______________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (3,)), msg = 'sub', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43e549678>
op = 'add', reduce_op = 'sum'
lhs_data = tensor([[1.0784],
        [1.4055],
        [1.9026],
        [1.6320],
        [1.3372],
        [1.3157],
        [1...719],
        [1.9869],
        [1.1495],
        [1.5442],
        [1.5053]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.3930, -1.4457, -1.7824],
        [-1.0067, -1.1247, -1.8807],
        [-1.0412, -1.4318, -1.1551],
       ...  [-1.3699, -1.1305, -1.9950],
        [-1.9934, -1.5551, -1.2210]], dtype=torch.float64,
       grad_fn=<NegBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 1]), e shape: torch.Size([300, 3])
SpMM(message func: sub, reduce func: sum)
______________________ test_spmm[idtype0-sum-sub-shp3-g0] ______________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (1,)), msg = 'sub', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb7059bd5e8>
op = 'add', reduce_op = 'sum'
lhs_data = tensor([[1.1425, 1.0545, 1.6091],
        [1.4111, 1.2635, 1.5326],
        [1.2408, 1.6082, 1.0603],
        [1.4984,... 1.5712],
        [1.0820, 1.2713, 1.5015],
        [1.9589, 1.0775, 1.9009]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.2473],
        [-1.9782],
        [-1.4914],
        [-1.8235],
        [-1.3040],
        [-1.4999],
    ...       [-1.8185],
        [-1.0769],
        [-1.0625],
        [-1.8108]], dtype=torch.float64, grad_fn=<NegBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 3]), e shape: torch.Size([100, 1])
SpMM(message func: sub, reduce func: sum)
______________________ test_spmm[idtype0-sum-sub-shp3-g1] ______________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (1,)), msg = 'sub', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43e8bd7e0>
op = 'add', reduce_op = 'sum'
lhs_data = tensor([[1.6933, 1.9630, 1.2219],
        [1.8793, 1.5275, 1.2808],
        [1.1967, 1.5776, 1.8460],
        [1.3277,... 1.8764],
        [1.5420, 1.8778, 1.6654],
        [1.2725, 1.7009, 1.3823]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.2569],
        [-1.8180],
        [-1.3741],
        [-1.3436],
        [-1.2602],
        [-1.2169],
    ...       [-1.9400],
        [-1.9570],
        [-1.4328],
        [-1.9951]], dtype=torch.float64, grad_fn=<NegBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 3]), e shape: torch.Size([300, 1])
SpMM(message func: sub, reduce func: sum)
______________________ test_spmm[idtype0-sum-sub-shp4-g0] ______________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), msg = 'sub', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb705a6be10>
op = 'add', reduce_op = 'sum'
lhs_data = tensor([[1.2760],
        [1.0470],
        [1.1414],
        [1.7306],
        [1.7177],
        [1.4677],
        [1...510],
        [1.2984],
        [1.1867],
        [1.1724],
        [1.9103]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.9894],
        [-1.1105],
        [-1.3630],
        [-1.7839],
        [-1.6529],
        [-1.8446],
    ...       [-1.5321],
        [-1.2838],
        [-1.5264],
        [-1.7479]], dtype=torch.float64, grad_fn=<NegBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 1]), e shape: torch.Size([100, 1])
SpMM(message func: sub, reduce func: sum)
______________________ test_spmm[idtype0-sum-sub-shp4-g1] ______________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), msg = 'sub', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72c05d288>
op = 'add', reduce_op = 'sum'
lhs_data = tensor([[1.6244],
        [1.5533],
        [1.8909],
        [1.6866],
        [1.3093],
        [1.5936],
        [1...122],
        [1.8332],
        [1.9697],
        [1.3892],
        [1.3096]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.9767],
        [-1.8773],
        [-1.5309],
        [-1.1591],
        [-1.4181],
        [-1.7704],
    ...       [-1.1628],
        [-1.9562],
        [-1.4900],
        [-1.2409]], dtype=torch.float64, grad_fn=<NegBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 1]), e shape: torch.Size([300, 1])
SpMM(message func: sub, reduce func: sum)
______________________ test_spmm[idtype0-sum-sub-shp5-g0] ______________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
shp = ((), ()), msg = 'sub', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb7058f3240>
op = 'add', reduce_op = 'sum'
lhs_data = tensor([1.3037, 1.0742, 1.1153, 1.5434, 1.8342, 1.8203, 1.8247, 1.9058, 1.9678,
        1.0746, 1.7946, 1.6702, 1.1102...0085, 1.8335, 1.4440, 1.4577, 1.0415, 1.4778,
        1.9637, 1.5857, 1.1518], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([-1.0224, -1.3550, -1.6578, -1.5635, -1.8304, -1.5426, -1.3127, -1.7301,
        -1.5008, -1.0747, -1.3295, -1....9896, -1.8214, -1.3959,
        -1.0630, -1.9159, -1.9049, -1.1866], dtype=torch.float64,
       grad_fn=<NegBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30]), e shape: torch.Size([100])
SpMM(message func: sub, reduce func: sum)
______________________ test_spmm[idtype0-sum-sub-shp5-g1] ______________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((), ()), msg = 'sub', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb44ebc4318>
op = 'add', reduce_op = 'sum'
lhs_data = tensor([1.7508, 1.9074, 1.1235, 1.1536, 1.4819, 1.6190, 1.8066, 1.1496, 1.8230,
        1.8883, 1.1991, 1.3882, 1.5839...8291, 1.3387, 1.2565, 1.0445, 1.6726, 1.8034,
        1.9525, 1.0970, 1.3303], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([-1.2986, -1.2146, -1.6727, -1.6179, -1.1804, -1.6827, -1.2872, -1.7017,
        -1.4716, -1.5743, -1.9886, -1....8991, -1.4532, -1.6248,
        -1.0588, -1.4869, -1.8691, -1.5706], dtype=torch.float64,
       grad_fn=<NegBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30]), e shape: torch.Size([300])
SpMM(message func: sub, reduce func: sum)
______________________ test_spmm[idtype0-sum-mul-shp0-g0] ______________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), msg = 'mul', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72c191cf0>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[[[[[1.0140],
            [1.8979],
            [1.2494]]],


          [[[1.0599],
            [1.1319],
    ...0]]],


          [[[1.6851],
            [1.6047],
            [1.5512]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.2067]],

           [[1.7983]],

           [[1.0364]]]],



         [[[[1.7821]],

           [[1.704...



         [[[[1.8652]],

           [[1.2216]],

           [[1.7332]]]]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 1, 2, 1, 3, 1]), e shape: torch.Size([100, 4, 1, 3, 1, 1])
SpMM(message func: mul, reduce func: sum)
______________________ test_spmm[idtype0-sum-mul-shp0-g1] ______________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), msg = 'mul', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43e8c8288>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[[[[[1.5903],
            [1.9912],
            [1.7561]]],


          [[[1.5368],
            [1.1804],
    ...0]]],


          [[[1.6191],
            [1.6275],
            [1.8715]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.3511]],

           [[1.9281]],

           [[1.9746]]]],



         [[[[1.6182]],

           [[1.147...



         [[[[1.8363]],

           [[1.1885]],

           [[1.0999]]]]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 1, 2, 1, 3, 1]), e shape: torch.Size([300, 4, 1, 3, 1, 1])
SpMM(message func: mul, reduce func: sum)
______________________ test_spmm[idtype0-sum-mul-shp1-g0] ______________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1, 3), dtype=torch.float64)})
shp = ((3, 3), (1, 3)), msg = 'mul', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb705a5df30>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[[1.4824, 1.9225, 1.8137],
         [1.8712, 1.9302, 1.9007],
         [1.5347, 1.7434, 1.3775]],

        [[1...6349],
         [1.9670, 1.9923, 1.0215],
         [1.2903, 1.8675, 1.0671]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[1.1824, 1.1320, 1.1169]],

        [[1.0353, 1.3255, 1.8268]],

        [[1.9570, 1.7670, 1.3943]],

       ...]],

        [[1.3645, 1.6154, 1.1872]],

        [[1.8106, 1.2568, 1.1374]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 3, 3]), e shape: torch.Size([100, 1, 3])
SpMM(message func: mul, reduce func: sum)
______________________ test_spmm[idtype0-sum-mul-shp1-g1] ______________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3, 3), (1, 3)), msg = 'mul', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb7059c6630>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[[1.7139, 1.3087, 1.1682],
         [1.1509, 1.4921, 1.4540],
         [1.4166, 1.6132, 1.3436]],

        [[1...8489],
         [1.5328, 1.6525, 1.5892],
         [1.7500, 1.4858, 1.8740]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[1.1014, 1.4299, 1.1164]],

        [[1.8178, 1.3236, 1.9239]],

        [[1.9086, 1.6516, 1.9862]],

       ...]],

        [[1.7102, 1.5560, 1.2902]],

        [[1.2849, 1.1691, 1.5025]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 3, 3]), e shape: torch.Size([300, 1, 3])
SpMM(message func: mul, reduce func: sum)
______________________ test_spmm[idtype0-sum-mul-shp2-g0] ______________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float64)})
shp = ((1,), (3,)), msg = 'mul', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72aa9abd0>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[1.5549],
        [1.5018],
        [1.2902],
        [1.4946],
        [1.4263],
        [1.7023],
        [1...145],
        [1.7701],
        [1.9742],
        [1.2869],
        [1.3221]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.5514, 1.6477, 1.8085],
        [1.6523, 1.4142, 1.7243],
        [1.0711, 1.0453, 1.7096],
        [1.9693,... 1.8290],
        [1.1880, 1.0347, 1.1098],
        [1.3033, 1.2928, 1.8547]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1, 3), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 1]), e shape: torch.Size([100, 3])
SpMM(message func: mul, reduce func: sum)
______________________ test_spmm[idtype0-sum-mul-shp2-g1] ______________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (3,)), msg = 'mul', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72aa9a558>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[1.2846],
        [1.0639],
        [1.2068],
        [1.3675],
        [1.4064],
        [1.5291],
        [1...626],
        [1.0486],
        [1.6086],
        [1.0466],
        [1.9739]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.9435, 1.1584, 1.4398],
        [1.6969, 1.2328, 1.6207],
        [1.5595, 1.1947, 1.2301],
        [1.7986,... 1.8437],
        [1.6271, 1.6477, 1.4210],
        [1.1381, 1.2383, 1.1529]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 1]), e shape: torch.Size([300, 3])
SpMM(message func: mul, reduce func: sum)
______________________ test_spmm[idtype0-sum-mul-shp3-g0] ______________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (1,)), msg = 'mul', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb70599bbd0>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[1.2918, 1.5036, 1.2547],
        [1.0635, 1.5564, 1.4829],
        [1.5631, 1.7641, 1.5696],
        [1.6122,... 1.7212],
        [1.7771, 1.5790, 1.1386],
        [1.6392, 1.7817, 1.5953]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.7887],
        [1.7018],
        [1.1979],
        [1.5527],
        [1.9940],
        [1.9516],
        [1...506],
        [1.9117],
        [1.3456],
        [1.5112],
        [1.5642]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 3]), e shape: torch.Size([100, 1])
SpMM(message func: mul, reduce func: sum)
______________________ test_spmm[idtype0-sum-mul-shp3-g1] ______________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (1,)), msg = 'mul', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb7059eb6c0>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[1.8916, 1.6317, 1.8441],
        [1.1313, 1.1656, 1.0077],
        [1.7796, 1.4778, 1.9393],
        [1.3855,... 1.2277],
        [1.5208, 1.9864, 1.6008],
        [1.5470, 1.5941, 1.8383]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.7527],
        [1.8805],
        [1.2918],
        [1.1471],
        [1.2333],
        [1.6424],
        [1...431],
        [1.7083],
        [1.9557],
        [1.6104],
        [1.2041]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 3]), e shape: torch.Size([300, 1])
SpMM(message func: mul, reduce func: sum)
______________________ test_spmm[idtype0-sum-mul-shp4-g0] ______________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), msg = 'mul', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb705a60480>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[1.7251],
        [1.3680],
        [1.1051],
        [1.7769],
        [1.0065],
        [1.7938],
        [1...096],
        [1.7214],
        [1.6984],
        [1.4404],
        [1.8507]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.5406],
        [1.4890],
        [1.0032],
        [1.0295],
        [1.2767],
        [1.0226],
        [1...663],
        [1.1327],
        [1.2312],
        [1.9686],
        [1.7805]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 1]), e shape: torch.Size([100, 1])
SpMM(message func: mul, reduce func: sum)
______________________ test_spmm[idtype0-sum-mul-shp4-g1] ______________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), msg = 'mul', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb705a371f8>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[1.3252],
        [1.2135],
        [1.3169],
        [1.2841],
        [1.1833],
        [1.2280],
        [1...694],
        [1.6213],
        [1.8859],
        [1.8936],
        [1.7161]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.7259],
        [1.9144],
        [1.8004],
        [1.8633],
        [1.2997],
        [1.7188],
        [1...321],
        [1.3693],
        [1.3387],
        [1.9556],
        [1.8871]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 1]), e shape: torch.Size([300, 1])
SpMM(message func: mul, reduce func: sum)
______________________ test_spmm[idtype0-sum-mul-shp5-g0] ______________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
shp = ((), ()), msg = 'mul', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72ac2b480>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([1.7029, 1.3322, 1.5399, 1.7353, 1.6370, 1.8041, 1.8184, 1.1030, 1.1398,
        1.4152, 1.5305, 1.7750, 1.7584...4194, 1.0309, 1.9746, 1.5267, 1.5729, 1.0277,
        1.0949, 1.3212, 1.4989], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([1.5648, 1.8834, 1.0281, 1.1945, 1.0030, 1.5596, 1.6731, 1.7885, 1.0137,
        1.3970, 1.4375, 1.7686, 1.3601...0877, 1.8754, 1.6112, 1.5903, 1.9558, 1.2606, 1.9699, 1.1672,
        1.3533], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30]), e shape: torch.Size([100])
SpMM(message func: mul, reduce func: sum)
______________________ test_spmm[idtype0-sum-mul-shp5-g1] ______________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((), ()), msg = 'mul', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72ac2b288>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([1.2106, 1.1638, 1.7502, 1.7752, 1.1824, 1.9559, 1.9792, 1.1555, 1.9308,
        1.1553, 1.9382, 1.2445, 1.1724...2464, 1.5592, 1.2332, 1.3012, 1.4791, 1.9062,
        1.6181, 1.5803, 1.7808], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([1.1990, 1.3084, 1.1776, 1.5920, 1.6564, 1.6520, 1.9476, 1.9358, 1.0350,
        1.7133, 1.6433, 1.5288, 1.3158...6278, 1.2669, 1.7353, 1.4297, 1.2671, 1.3093,
        1.2830, 1.5184, 1.5004], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30]), e shape: torch.Size([300])
SpMM(message func: mul, reduce func: sum)
______________________ test_spmm[idtype0-sum-div-shp0-g0] ______________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), msg = 'div', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72b191e58>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[[[[[1.9814],
            [1.9324],
            [1.2653]]],


          [[[1.5933],
            [1.8191],
    ...3]]],


          [[[1.7742],
            [1.9108],
            [1.2632]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[0.9068]],

           [[0.7897]],

           [[0.9791]]]],



         [[[[0.7116]],

           [[0.823...         [[[[0.6842]],

           [[0.7121]],

           [[0.5817]]]]]], dtype=torch.float64, grad_fn=<MulBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 1, 2, 1, 3, 1]), e shape: torch.Size([100, 4, 1, 3, 1, 1])
SpMM(message func: div, reduce func: sum)
______________________ test_spmm[idtype0-sum-div-shp0-g1] ______________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), msg = 'div', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72c01aa20>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[[[[[1.7379],
            [1.8346],
            [1.3522]]],


          [[[1.8086],
            [1.1045],
    ...2]]],


          [[[1.7047],
            [1.7217],
            [1.0543]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[0.6817]],

           [[0.8234]],

           [[0.6176]]]],



         [[[[0.5502]],

           [[0.555...         [[[[0.5851]],

           [[0.9794]],

           [[0.9820]]]]]], dtype=torch.float64, grad_fn=<MulBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 1, 2, 1, 3, 1]), e shape: torch.Size([300, 4, 1, 3, 1, 1])
SpMM(message func: div, reduce func: sum)
______________________ test_spmm[idtype0-sum-div-shp1-g0] ______________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1, 3), dtype=torch.float64)})
shp = ((3, 3), (1, 3)), msg = 'div', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb705a313a8>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[[1.2131, 1.4117, 1.2080],
         [1.1355, 1.9336, 1.0511],
         [1.8324, 1.5133, 1.9061]],

        [[1...8289],
         [1.1204, 1.0569, 1.3141],
         [1.9185, 1.3790, 1.5953]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[0.5015, 0.6270, 0.9297]],

        [[0.8215, 0.7039, 0.9560]],

        [[0.6560, 0.5652, 0.5484]],

       ...  [[0.5059, 0.5406, 0.5858]],

        [[0.5400, 0.5726, 0.5682]]], dtype=torch.float64,
       grad_fn=<MulBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 3, 3]), e shape: torch.Size([100, 1, 3])
SpMM(message func: div, reduce func: sum)
______________________ test_spmm[idtype0-sum-div-shp1-g1] ______________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3, 3), (1, 3)), msg = 'div', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb7059f44c8>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[[1.7545, 1.6713, 1.1842],
         [1.1259, 1.3147, 1.8533],
         [1.5601, 1.3238, 1.6004]],

        [[1...5455],
         [1.0855, 1.8315, 1.3317],
         [1.4602, 1.8077, 1.4559]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[0.9910, 0.7775, 0.5682]],

        [[0.5167, 0.6960, 0.6588]],

        [[0.5315, 0.6402, 0.7677]],

       ...  [[0.6230, 0.9383, 0.5791]],

        [[0.7394, 0.7732, 0.9708]]], dtype=torch.float64,
       grad_fn=<MulBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 3, 3]), e shape: torch.Size([300, 1, 3])
SpMM(message func: div, reduce func: sum)
______________________ test_spmm[idtype0-sum-div-shp2-g0] ______________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float64)})
shp = ((1,), (3,)), msg = 'div', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb705a60480>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[1.1670],
        [1.3602],
        [1.2536],
        [1.5956],
        [1.7176],
        [1.0454],
        [1...736],
        [1.0338],
        [1.2382],
        [1.2489],
        [1.8060]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.9913, 0.7978, 0.6708],
        [0.8024, 0.6068, 0.6247],
        [0.9971, 0.6277, 0.5638],
        [0.5901,...106],
        [0.5312, 0.7094, 0.8725],
        [0.5162, 0.6937, 0.6908]], dtype=torch.float64, grad_fn=<MulBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1, 3), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 1]), e shape: torch.Size([100, 3])
SpMM(message func: div, reduce func: sum)
______________________ test_spmm[idtype0-sum-div-shp2-g1] ______________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (3,)), msg = 'div', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb705a033f0>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[1.5734],
        [1.6739],
        [1.8208],
        [1.8869],
        [1.7064],
        [1.6033],
        [1...384],
        [1.6099],
        [1.0142],
        [1.6083],
        [1.0194]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.5283, 0.8915, 0.7014],
        [0.6830, 0.9152, 0.5703],
        [0.5962, 0.5433, 0.8491],
        [0.7636,...396],
        [0.7938, 0.6626, 0.7125],
        [0.6083, 0.5288, 0.8110]], dtype=torch.float64, grad_fn=<MulBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 1]), e shape: torch.Size([300, 3])
SpMM(message func: div, reduce func: sum)
______________________ test_spmm[idtype0-sum-div-shp3-g0] ______________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (1,)), msg = 'div', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43e5687e0>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[1.9801, 1.3402, 1.4063],
        [1.4479, 1.2688, 1.9987],
        [1.7176, 1.4663, 1.8850],
        [1.9183,... 1.6661],
        [1.1677, 1.1264, 1.2122],
        [1.4312, 1.9284, 1.4458]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.9307],
        [0.5597],
        [0.5058],
        [0.9243],
        [0.5155],
        [0.7755],
        [0...,
        [0.6310],
        [0.7515],
        [0.8391],
        [0.6801]], dtype=torch.float64, grad_fn=<MulBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 3]), e shape: torch.Size([100, 1])
SpMM(message func: div, reduce func: sum)
______________________ test_spmm[idtype0-sum-div-shp3-g1] ______________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (1,)), msg = 'div', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72ab3fd80>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[1.9750, 1.6273, 1.2460],
        [1.8043, 1.3368, 1.6432],
        [1.2404, 1.7898, 1.6111],
        [1.8996,... 1.5291],
        [1.1550, 1.4309, 1.7765],
        [1.3788, 1.7686, 1.6795]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.5229],
        [0.6807],
        [0.5414],
        [0.9952],
        [0.7800],
        [0.6691],
        [0...,
        [0.5331],
        [0.5626],
        [0.5067],
        [0.5445]], dtype=torch.float64, grad_fn=<MulBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 3]), e shape: torch.Size([300, 1])
SpMM(message func: div, reduce func: sum)
______________________ test_spmm[idtype0-sum-div-shp4-g0] ______________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), msg = 'div', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72a9137e0>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[1.1620],
        [1.7116],
        [1.8465],
        [1.6238],
        [1.0184],
        [1.7162],
        [1...866],
        [1.5766],
        [1.9688],
        [1.0304],
        [1.7980]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.8123],
        [0.8452],
        [0.7237],
        [0.8438],
        [0.6898],
        [0.5194],
        [0...,
        [0.8157],
        [0.6327],
        [0.5311],
        [0.9213]], dtype=torch.float64, grad_fn=<MulBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 1]), e shape: torch.Size([100, 1])
SpMM(message func: div, reduce func: sum)
______________________ test_spmm[idtype0-sum-div-shp4-g1] ______________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), msg = 'div', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb705acc900>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[1.0546],
        [1.5534],
        [1.7112],
        [1.8227],
        [1.9266],
        [1.9127],
        [1...348],
        [1.4884],
        [1.6006],
        [1.2941],
        [1.8381]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.7310],
        [0.8876],
        [0.5888],
        [0.5168],
        [0.7770],
        [0.5890],
        [0...,
        [0.6311],
        [0.6944],
        [0.5178],
        [0.7979]], dtype=torch.float64, grad_fn=<MulBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 1]), e shape: torch.Size([300, 1])
SpMM(message func: div, reduce func: sum)
______________________ test_spmm[idtype0-sum-div-shp5-g0] ______________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
shp = ((), ()), msg = 'div', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43ebbd5e8>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([1.6844, 1.3725, 1.7066, 1.5253, 1.1052, 1.1859, 1.3106, 1.7606, 1.6430,
        1.8171, 1.9313, 1.6043, 1.4135...5946, 1.4130, 1.0098, 1.6895, 1.0750, 1.8854,
        1.9464, 1.1018, 1.0511], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([0.8853, 0.6472, 0.5255, 0.9254, 0.7176, 0.6729, 0.8022, 0.6200, 0.7154,
        0.7479, 0.7791, 0.6791, 0.5294..., 0.8327, 0.5068, 0.6025, 0.5513, 0.8490, 0.7701, 0.9602,
        0.8285], dtype=torch.float64, grad_fn=<MulBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30]), e shape: torch.Size([100])
SpMM(message func: div, reduce func: sum)
______________________ test_spmm[idtype0-sum-div-shp5-g1] ______________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((), ()), msg = 'div', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72c05f3f0>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([1.1721, 1.5724, 1.5739, 1.1240, 1.5747, 1.6038, 1.7709, 1.5490, 1.8300,
        1.8641, 1.2472, 1.3143, 1.7850...3320, 1.3772, 1.9583, 1.5476, 1.3027, 1.9849,
        1.1265, 1.3499, 1.9441], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([0.5075, 0.7333, 0.7726, 0.7481, 0.7334, 0.8213, 0.9013, 0.6387, 0.6288,
        0.9491, 0.5956, 0.6901, 0.6190..., 0.5009, 0.7801, 0.6849, 0.5232, 0.6308,
        0.6621, 0.5240, 0.6413], dtype=torch.float64, grad_fn=<MulBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30]), e shape: torch.Size([300])
SpMM(message func: div, reduce func: sum)
___________________ test_spmm[idtype0-sum-copy_lhs-shp0-g0] ____________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), msg = 'copy_lhs', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72b726558>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[[[[[1.8452],
            [1.7821],
            [1.5476]]],


          [[[1.5195],
            [1.6259],
    ...2]]],


          [[[1.7249],
            [1.0986],
            [1.4275]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.2940]],

           [[1.0926]],

           [[1.2477]]]],



         [[[[1.8982]],

           [[1.977...



         [[[[1.2939]],

           [[1.1651]],

           [[1.2239]]]]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 1, 2, 1, 3, 1]), e shape: torch.Size([100, 4, 1, 3, 1, 1])
SpMM(message func: copy_lhs, reduce func: sum)
___________________ test_spmm[idtype0-sum-copy_lhs-shp0-g1] ____________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), msg = 'copy_lhs', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72c0c5af8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[[[[[1.6214],
            [1.5827],
            [1.7938]]],


          [[[1.1637],
            [1.7092],
    ...6]]],


          [[[1.5474],
            [1.4579],
            [1.8201]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.9394]],

           [[1.0724]],

           [[1.5861]]]],



         [[[[1.6206]],

           [[1.377...



         [[[[1.0879]],

           [[1.3401]],

           [[1.8074]]]]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 1, 2, 1, 3, 1]), e shape: torch.Size([300, 4, 1, 3, 1, 1])
SpMM(message func: copy_lhs, reduce func: sum)
___________________ test_spmm[idtype0-sum-copy_lhs-shp1-g0] ____________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1, 3), dtype=torch.float64)})
shp = ((3, 3), (1, 3)), msg = 'copy_lhs', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb705a62558>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[[1.5997, 1.8513, 1.7983],
         [1.9605, 1.3498, 1.3835],
         [1.3884, 1.0152, 1.3431]],

        [[1...8978],
         [1.9586, 1.5630, 1.6509],
         [1.1073, 1.3057, 1.5721]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[1.3926, 1.2279, 1.5436]],

        [[1.9142, 1.7783, 1.5315]],

        [[1.8053, 1.3856, 1.1803]],

       ...]],

        [[1.2818, 1.9247, 1.3314]],

        [[1.7629, 1.4039, 1.4817]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 3, 3]), e shape: torch.Size([100, 1, 3])
SpMM(message func: copy_lhs, reduce func: sum)
___________________ test_spmm[idtype0-sum-copy_lhs-shp1-g1] ____________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3, 3), (1, 3)), msg = 'copy_lhs', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb7059b3948>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[[1.4779, 1.3780, 1.8392],
         [1.0997, 1.1402, 1.9693],
         [1.7200, 1.0185, 1.6612]],

        [[1...4226],
         [1.5895, 1.4566, 1.8223],
         [1.1345, 1.6480, 1.5332]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[1.8881, 1.3324, 1.2988]],

        [[1.3590, 1.5581, 1.2008]],

        [[1.6410, 1.7804, 1.3370]],

       ...]],

        [[1.6402, 1.2671, 1.9508]],

        [[1.6335, 1.2026, 1.1520]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 3, 3]), e shape: torch.Size([300, 1, 3])
SpMM(message func: copy_lhs, reduce func: sum)
___________________ test_spmm[idtype0-sum-copy_lhs-shp2-g0] ____________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float64)})
shp = ((1,), (3,)), msg = 'copy_lhs', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72a93b288>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[1.4593],
        [1.8479],
        [1.6698],
        [1.4151],
        [1.3091],
        [1.8983],
        [1...055],
        [1.7282],
        [1.4490],
        [1.6849],
        [1.7787]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.2791, 1.5762, 1.6729],
        [1.3757, 1.9369, 1.3147],
        [1.2084, 1.2129, 1.9397],
        [1.0443,... 1.3443],
        [1.9673, 1.0693, 1.0730],
        [1.1513, 1.1696, 1.2334]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1, 3), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 1]), e shape: torch.Size([100, 3])
SpMM(message func: copy_lhs, reduce func: sum)
___________________ test_spmm[idtype0-sum-copy_lhs-shp2-g1] ____________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (3,)), msg = 'copy_lhs', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72c22e5e8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[1.8449],
        [1.7321],
        [1.9656],
        [1.6878],
        [1.7632],
        [1.9886],
        [1...581],
        [1.0305],
        [1.2214],
        [1.3983],
        [1.9188]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.7275, 1.5659, 1.2324],
        [1.4721, 1.1070, 1.8263],
        [1.2426, 1.2408, 1.6990],
        [1.4391,... 1.8544],
        [1.3335, 1.3326, 1.6916],
        [1.0626, 1.7277, 1.5237]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 1]), e shape: torch.Size([300, 3])
SpMM(message func: copy_lhs, reduce func: sum)
___________________ test_spmm[idtype0-sum-copy_lhs-shp3-g0] ____________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (1,)), msg = 'copy_lhs', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72932d900>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[1.4010, 1.1068, 1.1218],
        [1.2847, 1.0731, 1.4992],
        [1.3375, 1.2383, 1.8182],
        [1.3825,... 1.0318],
        [1.5544, 1.2142, 1.7238],
        [1.4998, 1.0758, 1.0448]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.5598],
        [1.0299],
        [1.8050],
        [1.8472],
        [1.0688],
        [1.1781],
        [1...552],
        [1.4060],
        [1.3798],
        [1.2749],
        [1.8655]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 3]), e shape: torch.Size([100, 1])
SpMM(message func: copy_lhs, reduce func: sum)
___________________ test_spmm[idtype0-sum-copy_lhs-shp3-g1] ____________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (1,)), msg = 'copy_lhs', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72c1be480>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[1.3793, 1.7244, 1.7848],
        [1.7753, 1.9518, 1.5453],
        [1.2922, 1.5085, 1.4221],
        [1.8086,... 1.2929],
        [1.4976, 1.3039, 1.9827],
        [1.5732, 1.5599, 1.7200]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.6071],
        [1.9353],
        [1.0459],
        [1.9373],
        [1.0379],
        [1.0515],
        [1...198],
        [1.9897],
        [1.4742],
        [1.6116],
        [1.0560]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 3]), e shape: torch.Size([300, 1])
SpMM(message func: copy_lhs, reduce func: sum)
___________________ test_spmm[idtype0-sum-copy_lhs-shp4-g0] ____________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), msg = 'copy_lhs', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43eb21f78>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[1.3715],
        [1.1650],
        [1.3288],
        [1.6154],
        [1.6781],
        [1.4083],
        [1...551],
        [1.3351],
        [1.3725],
        [1.8450],
        [1.0347]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.1439],
        [1.3960],
        [1.5681],
        [1.0697],
        [1.5837],
        [1.5640],
        [1...335],
        [1.2699],
        [1.1305],
        [1.6417],
        [1.3109]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 1]), e shape: torch.Size([100, 1])
SpMM(message func: copy_lhs, reduce func: sum)
___________________ test_spmm[idtype0-sum-copy_lhs-shp4-g1] ____________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), msg = 'copy_lhs', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43e56a240>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[1.0093],
        [1.7665],
        [1.9353],
        [1.7445],
        [1.9016],
        [1.3253],
        [1...363],
        [1.3317],
        [1.8038],
        [1.3437],
        [1.0866]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.9193],
        [1.2632],
        [1.0993],
        [1.5874],
        [1.7641],
        [1.2434],
        [1...262],
        [1.2883],
        [1.1883],
        [1.2261],
        [1.7014]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 1]), e shape: torch.Size([300, 1])
SpMM(message func: copy_lhs, reduce func: sum)
___________________ test_spmm[idtype0-sum-copy_lhs-shp5-g0] ____________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
shp = ((), ()), msg = 'copy_lhs', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb7055f21b0>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([1.4804, 1.6692, 1.4475, 1.2569, 1.9885, 1.0378, 1.1180, 1.7455, 1.7528,
        1.2320, 1.3078, 1.5787, 1.5410...1153, 1.2060, 1.6675, 1.0087, 1.3364, 1.1109,
        1.1755, 1.4941, 1.1814], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([1.4830, 1.6412, 1.4840, 1.3205, 1.7716, 1.3604, 1.5086, 1.7159, 1.1792,
        1.9845, 1.2988, 1.6342, 1.7570...2785, 1.4247, 1.9105, 1.5178, 1.4165, 1.4882, 1.1144, 1.6186,
        1.9452], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30]), e shape: torch.Size([100])
SpMM(message func: copy_lhs, reduce func: sum)
___________________ test_spmm[idtype0-sum-copy_lhs-shp5-g1] ____________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((), ()), msg = 'copy_lhs', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72bf67438>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([1.2539, 1.2497, 1.9109, 1.7717, 1.5408, 1.9682, 1.0594, 1.2744, 1.2356,
        1.4819, 1.7918, 1.7128, 1.0037...8494, 1.5952, 1.0836, 1.7642, 1.0113, 1.4698,
        1.7854, 1.3585, 1.6025], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([1.0103, 1.1843, 1.1112, 1.4352, 1.1494, 1.2022, 1.3400, 1.4542, 1.7155,
        1.7699, 1.5144, 1.5511, 1.7937...2854, 1.2491, 1.4714, 1.2622, 1.8843, 1.5420,
        1.9357, 1.8957, 1.5234], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30]), e shape: torch.Size([300])
SpMM(message func: copy_lhs, reduce func: sum)
___________________ test_spmm[idtype0-sum-copy_rhs-shp0-g0] ____________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), msg = 'copy_rhs', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72c1fd9d8>
op = 'copy_rhs', reduce_op = 'sum'
lhs_data = tensor([[[[[[1.9372],
            [1.5062],
            [1.7634]]],


          [[[1.5915],
            [1.7052],
    ...5]]],


          [[[1.9581],
            [1.8569],
            [1.6066]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.1487]],

           [[1.2994]],

           [[1.4454]]]],



         [[[[1.0139]],

           [[1.875...



         [[[[1.6152]],

           [[1.5851]],

           [[1.4825]]]]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 1, 2, 1, 3, 1]), e shape: torch.Size([100, 4, 1, 3, 1, 1])
SpMM(message func: copy_rhs, reduce func: sum)
___________________ test_spmm[idtype0-sum-copy_rhs-shp0-g1] ____________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), msg = 'copy_rhs', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72c1fd708>
op = 'copy_rhs', reduce_op = 'sum'
lhs_data = tensor([[[[[[1.2726],
            [1.0043],
            [1.9217]]],


          [[[1.8386],
            [1.3569],
    ...3]]],


          [[[1.8109],
            [1.4932],
            [1.8310]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.3101]],

           [[1.3799]],

           [[1.7502]]]],



         [[[[1.8358]],

           [[1.788...



         [[[[1.2082]],

           [[1.7897]],

           [[1.4945]]]]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 1, 2, 1, 3, 1]), e shape: torch.Size([300, 4, 1, 3, 1, 1])
SpMM(message func: copy_rhs, reduce func: sum)
___________________ test_spmm[idtype0-sum-copy_rhs-shp1-g0] ____________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1, 3), dtype=torch.float64)})
shp = ((3, 3), (1, 3)), msg = 'copy_rhs', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43e56d480>
op = 'copy_rhs', reduce_op = 'sum'
lhs_data = tensor([[[1.4471, 1.7304, 1.2646],
         [1.1210, 1.0169, 1.7056],
         [1.2252, 1.9382, 1.8745]],

        [[1...6772],
         [1.3509, 1.0493, 1.8754],
         [1.5648, 1.3071, 1.7694]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[1.4915, 1.3172, 1.4518]],

        [[1.5083, 1.9429, 1.4559]],

        [[1.5533, 1.8534, 1.3843]],

       ...]],

        [[1.9044, 1.4913, 1.5877]],

        [[1.5760, 1.1878, 1.8721]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 3, 3]), e shape: torch.Size([100, 1, 3])
SpMM(message func: copy_rhs, reduce func: sum)
___________________ test_spmm[idtype0-sum-copy_rhs-shp1-g1] ____________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3, 3), (1, 3)), msg = 'copy_rhs', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43eba1a68>
op = 'copy_rhs', reduce_op = 'sum'
lhs_data = tensor([[[1.9471, 1.9225, 1.2470],
         [1.3350, 1.4292, 1.9243],
         [1.6086, 1.4803, 1.6614]],

        [[1...2617],
         [1.3292, 1.8883, 1.1294],
         [1.8375, 1.2971, 1.2105]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[1.1019, 1.9543, 1.6099]],

        [[1.7072, 1.9094, 1.9836]],

        [[1.0363, 1.1956, 1.9940]],

       ...]],

        [[1.9793, 1.5195, 1.9470]],

        [[1.6636, 1.8484, 1.4366]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 3, 3]), e shape: torch.Size([300, 1, 3])
SpMM(message func: copy_rhs, reduce func: sum)
___________________ test_spmm[idtype0-sum-copy_rhs-shp2-g0] ____________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float64)})
shp = ((1,), (3,)), msg = 'copy_rhs', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72a922ee8>
op = 'copy_rhs', reduce_op = 'sum'
lhs_data = tensor([[1.7549],
        [1.1102],
        [1.8974],
        [1.2575],
        [1.1825],
        [1.4893],
        [1...990],
        [1.3192],
        [1.2096],
        [1.6141],
        [1.2589]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.9012, 1.6160, 1.7841],
        [1.5285, 1.4537, 1.2490],
        [1.0023, 1.4502, 1.3083],
        [1.9884,... 1.7266],
        [1.6124, 1.0339, 1.5981],
        [1.4972, 1.7703, 1.7003]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1, 3), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 1]), e shape: torch.Size([100, 3])
SpMM(message func: copy_rhs, reduce func: sum)
___________________ test_spmm[idtype0-sum-copy_rhs-shp2-g1] ____________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (3,)), msg = 'copy_rhs', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43e38dc60>
op = 'copy_rhs', reduce_op = 'sum'
lhs_data = tensor([[1.2147],
        [1.7237],
        [1.5705],
        [1.6183],
        [1.0065],
        [1.3441],
        [1...160],
        [1.6707],
        [1.4900],
        [1.1663],
        [1.7360]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.0355, 1.9598, 1.5015],
        [1.8614, 1.1366, 1.9378],
        [1.4584, 1.1857, 1.6562],
        [1.4596,... 1.4233],
        [1.6531, 1.8560, 1.3650],
        [1.2197, 1.9577, 1.8141]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 1]), e shape: torch.Size([300, 3])
SpMM(message func: copy_rhs, reduce func: sum)
___________________ test_spmm[idtype0-sum-copy_rhs-shp3-g0] ____________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (1,)), msg = 'copy_rhs', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb705893678>
op = 'copy_rhs', reduce_op = 'sum'
lhs_data = tensor([[1.5813, 1.4839, 1.9807],
        [1.3423, 1.1027, 1.5193],
        [1.0617, 1.2221, 1.5883],
        [1.1558,... 1.8095],
        [1.5119, 1.7975, 1.5558],
        [1.9037, 1.8777, 1.1878]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.8334],
        [1.3905],
        [1.8482],
        [1.7682],
        [1.3395],
        [1.7422],
        [1...843],
        [1.7022],
        [1.3949],
        [1.3744],
        [1.3580]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 3]), e shape: torch.Size([100, 1])
SpMM(message func: copy_rhs, reduce func: sum)
___________________ test_spmm[idtype0-sum-copy_rhs-shp3-g1] ____________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (1,)), msg = 'copy_rhs', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72c1c4240>
op = 'copy_rhs', reduce_op = 'sum'
lhs_data = tensor([[1.4036, 1.2223, 1.2545],
        [1.7175, 1.0229, 1.8528],
        [1.3221, 1.6562, 1.4356],
        [1.1433,... 1.2122],
        [1.6242, 1.2114, 1.0879],
        [1.5247, 1.6059, 1.5591]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.9316],
        [1.5436],
        [1.3437],
        [1.2476],
        [1.4127],
        [1.6249],
        [1...268],
        [1.9566],
        [1.5833],
        [1.2654],
        [1.6868]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 3]), e shape: torch.Size([300, 1])
SpMM(message func: copy_rhs, reduce func: sum)
___________________ test_spmm[idtype0-sum-copy_rhs-shp4-g0] ____________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), msg = 'copy_rhs', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb7058f3828>
op = 'copy_rhs', reduce_op = 'sum'
lhs_data = tensor([[1.0060],
        [1.4769],
        [1.6433],
        [1.0928],
        [1.1657],
        [1.8613],
        [1...121],
        [1.8718],
        [1.5393],
        [1.5300],
        [1.1627]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.6606],
        [1.1083],
        [1.8680],
        [1.7075],
        [1.3026],
        [1.6015],
        [1...460],
        [1.7485],
        [1.9546],
        [1.4918],
        [1.2120]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 1]), e shape: torch.Size([100, 1])
SpMM(message func: copy_rhs, reduce func: sum)
___________________ test_spmm[idtype0-sum-copy_rhs-shp4-g1] ____________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), msg = 'copy_rhs', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb44ebc2900>
op = 'copy_rhs', reduce_op = 'sum'
lhs_data = tensor([[1.5196],
        [1.7044],
        [1.9794],
        [1.2643],
        [1.9136],
        [1.6683],
        [1...572],
        [1.4076],
        [1.1400],
        [1.1320],
        [1.5795]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.8273],
        [1.0045],
        [1.0098],
        [1.2901],
        [1.6475],
        [1.7772],
        [1...389],
        [1.8228],
        [1.0419],
        [1.0085],
        [1.5687]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 1]), e shape: torch.Size([300, 1])
SpMM(message func: copy_rhs, reduce func: sum)
___________________ test_spmm[idtype0-sum-copy_rhs-shp5-g0] ____________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
shp = ((), ()), msg = 'copy_rhs', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb705a7dea0>
op = 'copy_rhs', reduce_op = 'sum'
lhs_data = tensor([1.6120, 1.3162, 1.6949, 1.8935, 1.2788, 1.9166, 1.0778, 1.6760, 1.7876,
        1.4875, 1.1897, 1.5618, 1.4345...7248, 1.6266, 1.2074, 1.4787, 1.7711, 1.2018,
        1.1761, 1.5323, 1.8047], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([1.3214, 1.4891, 1.9626, 1.5169, 1.0849, 1.9361, 1.5042, 1.7019, 1.1987,
        1.8670, 1.0259, 1.3939, 1.6852...5950, 1.7366, 1.6739, 1.0696, 1.4035, 1.6044, 1.6364, 1.7999,
        1.2649], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30]), e shape: torch.Size([100])
SpMM(message func: copy_rhs, reduce func: sum)
___________________ test_spmm[idtype0-sum-copy_rhs-shp5-g1] ____________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((), ()), msg = 'copy_rhs', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72aa21900>
op = 'copy_rhs', reduce_op = 'sum'
lhs_data = tensor([1.3657, 1.0981, 1.0405, 1.1034, 1.2694, 1.4782, 1.0436, 1.9800, 1.5801,
        1.9809, 1.5041, 1.9831, 1.9126...1679, 1.7116, 1.4544, 1.8331, 1.5969, 1.6357,
        1.4404, 1.6976, 1.9940], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([1.7419, 1.0377, 1.4278, 1.6932, 1.5612, 1.9461, 1.9855, 1.1802, 1.5541,
        1.3788, 1.0787, 1.6651, 1.7787...6768, 1.3516, 1.8943, 1.7430, 1.8030, 1.5826,
        1.3795, 1.7043, 1.3873], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30]), e shape: torch.Size([300])
SpMM(message func: copy_rhs, reduce func: sum)
______________________ test_spmm[idtype0-min-add-shp0-g0] ______________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), msg = 'add', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb705a021b0>
op = 'add', reduce_op = 'min'
lhs_data = tensor([[[[[[1.8331],
            [1.9795],
            [1.0447]]],


          [[[1.3753],
            [1.3522],
    ...4]]],


          [[[1.0548],
            [1.4134],
            [1.9425]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.8490]],

           [[1.2035]],

           [[1.4295]]]],



         [[[[1.0357]],

           [[1.180...



         [[[[1.7697]],

           [[1.0179]],

           [[1.9555]]]]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 1, 2, 1, 3, 1]), e shape: torch.Size([100, 4, 1, 3, 1, 1])
SpMM(message func: add, reduce func: min)
______________________ test_spmm[idtype0-min-add-shp0-g1] ______________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), msg = 'add', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43eba1a68>
op = 'add', reduce_op = 'min'
lhs_data = tensor([[[[[[1.7598],
            [1.7258],
            [1.8360]]],


          [[[1.7943],
            [1.6991],
    ...8]]],


          [[[1.6300],
            [1.8575],
            [1.1807]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.3379]],

           [[1.5832]],

           [[1.1017]]]],



         [[[[1.5772]],

           [[1.861...



         [[[[1.5361]],

           [[1.2985]],

           [[1.3251]]]]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 1, 2, 1, 3, 1]), e shape: torch.Size([300, 4, 1, 3, 1, 1])
SpMM(message func: add, reduce func: min)
______________________ test_spmm[idtype0-min-add-shp1-g0] ______________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1, 3), dtype=torch.float64)})
shp = ((3, 3), (1, 3)), msg = 'add', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43e8fdab0>
op = 'add', reduce_op = 'min'
lhs_data = tensor([[[1.5435, 1.5837, 1.7738],
         [1.2792, 1.6432, 1.0881],
         [1.4289, 1.8463, 1.2535]],

        [[1...6749],
         [1.1597, 1.8680, 1.6745],
         [1.9518, 1.2500, 1.9831]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[1.5431, 1.7479, 1.9989]],

        [[1.7546, 1.4159, 1.1350]],

        [[1.8649, 1.0416, 1.7173]],

       ...]],

        [[1.4174, 1.9016, 1.6696]],

        [[1.9209, 1.9884, 1.7468]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 3, 3]), e shape: torch.Size([100, 1, 3])
SpMM(message func: add, reduce func: min)
______________________ test_spmm[idtype0-min-add-shp1-g1] ______________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3, 3), (1, 3)), msg = 'add', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb705a62bd0>
op = 'add', reduce_op = 'min'
lhs_data = tensor([[[1.4407, 1.8599, 1.1724],
         [1.3676, 1.7848, 1.8312],
         [1.5670, 1.7303, 1.1400]],

        [[1...0320],
         [1.4921, 1.8360, 1.8549],
         [1.2586, 1.8199, 1.7974]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[1.6271, 1.4883, 1.6088]],

        [[1.4620, 1.3887, 1.6374]],

        [[1.1160, 1.8566, 1.6175]],

       ...]],

        [[1.1234, 1.0635, 1.4670]],

        [[1.9441, 1.8978, 1.2791]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 3, 3]), e shape: torch.Size([300, 1, 3])
SpMM(message func: add, reduce func: min)
______________________ test_spmm[idtype0-min-add-shp2-g0] ______________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float64)})
shp = ((1,), (3,)), msg = 'add', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72c078048>
op = 'add', reduce_op = 'min'
lhs_data = tensor([[1.9701],
        [1.6144],
        [1.6555],
        [1.3278],
        [1.8142],
        [1.6918],
        [1...681],
        [1.3357],
        [1.7781],
        [1.1575],
        [1.7445]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.8908, 1.0547, 1.1668],
        [1.5015, 1.8768, 1.0935],
        [1.9803, 1.1676, 1.4444],
        [1.8894,... 1.7807],
        [1.2352, 1.1157, 1.8739],
        [1.8211, 1.9524, 1.3333]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1, 3), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 1]), e shape: torch.Size([100, 3])
SpMM(message func: add, reduce func: min)
______________________ test_spmm[idtype0-min-add-shp2-g1] ______________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (3,)), msg = 'add', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72c078708>
op = 'add', reduce_op = 'min'
lhs_data = tensor([[1.2207],
        [1.5799],
        [1.9575],
        [1.2226],
        [1.3904],
        [1.3452],
        [1...205],
        [1.3084],
        [1.9328],
        [1.9173],
        [1.5066]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.5864, 1.6932, 1.1702],
        [1.7157, 1.5332, 1.6856],
        [1.4493, 1.1798, 1.5199],
        [1.4583,... 1.3053],
        [1.8092, 1.5761, 1.0460],
        [1.1014, 1.3297, 1.0413]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 1]), e shape: torch.Size([300, 3])
SpMM(message func: add, reduce func: min)
______________________ test_spmm[idtype0-min-add-shp3-g0] ______________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (1,)), msg = 'add', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72ab03048>
op = 'add', reduce_op = 'min'
lhs_data = tensor([[1.4218, 1.6684, 1.9967],
        [1.4755, 1.4515, 1.0504],
        [1.8077, 1.8866, 1.3000],
        [1.1277,... 1.3125],
        [1.0462, 1.6108, 1.2255],
        [1.7605, 1.4075, 1.5006]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.3694],
        [1.1933],
        [1.9002],
        [1.1960],
        [1.1058],
        [1.9496],
        [1...269],
        [1.6318],
        [1.8448],
        [1.4308],
        [1.7821]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 3]), e shape: torch.Size([100, 1])
SpMM(message func: add, reduce func: min)
______________________ test_spmm[idtype0-min-add-shp3-g1] ______________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (1,)), msg = 'add', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb705a123a8>
op = 'add', reduce_op = 'min'
lhs_data = tensor([[1.3533, 1.4449, 1.0412],
        [1.1458, 1.5488, 1.5492],
        [1.2461, 1.8352, 1.0107],
        [1.0541,... 1.7122],
        [1.7265, 1.6297, 1.1646],
        [1.3310, 1.7517, 1.7742]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.8748],
        [1.1020],
        [1.5378],
        [1.0238],
        [1.9516],
        [1.4934],
        [1...725],
        [1.4555],
        [1.3058],
        [1.7346],
        [1.8286]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 3]), e shape: torch.Size([300, 1])
SpMM(message func: add, reduce func: min)
______________________ test_spmm[idtype0-min-add-shp4-g0] ______________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), msg = 'add', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43eb71828>
op = 'add', reduce_op = 'min'
lhs_data = tensor([[1.6326],
        [1.4979],
        [1.5020],
        [1.8831],
        [1.0286],
        [1.0454],
        [1...594],
        [1.3666],
        [1.2028],
        [1.6072],
        [1.1485]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.2220],
        [1.9500],
        [1.4927],
        [1.0658],
        [1.1785],
        [1.6288],
        [1...541],
        [1.2058],
        [1.7027],
        [1.6747],
        [1.4677]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 1]), e shape: torch.Size([100, 1])
SpMM(message func: add, reduce func: min)
______________________ test_spmm[idtype0-min-add-shp4-g1] ______________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), msg = 'add', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72af10828>
op = 'add', reduce_op = 'min'
lhs_data = tensor([[1.8015],
        [1.3381],
        [1.1840],
        [1.9815],
        [1.7479],
        [1.6291],
        [1...168],
        [1.4647],
        [1.3511],
        [1.6023],
        [1.8381]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.0278],
        [1.0715],
        [1.0467],
        [1.0356],
        [1.4242],
        [1.0124],
        [1...805],
        [1.6877],
        [1.4760],
        [1.8783],
        [1.5676]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 1]), e shape: torch.Size([300, 1])
SpMM(message func: add, reduce func: min)
______________________ test_spmm[idtype0-min-add-shp5-g0] ______________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
shp = ((), ()), msg = 'add', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72c126c18>
op = 'add', reduce_op = 'min'
lhs_data = tensor([1.1008, 1.0523, 1.6248, 1.7496, 1.9695, 1.6037, 1.7050, 1.1929, 1.9138,
        1.7469, 1.3544, 1.3188, 1.2597...2484, 1.6488, 1.1484, 1.7387, 1.9176, 1.2203,
        1.6183, 1.8329, 1.4124], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([1.3082, 1.4870, 1.2327, 1.3116, 1.6167, 1.5235, 1.1195, 1.7236, 1.0522,
        1.4300, 1.7373, 1.1238, 1.7254...1975, 1.7756, 1.4343, 1.9038, 1.1785, 1.3676, 1.7546, 1.7039,
        1.8295], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30]), e shape: torch.Size([100])
SpMM(message func: add, reduce func: min)
______________________ test_spmm[idtype0-min-add-shp5-g1] ______________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((), ()), msg = 'add', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb7058de438>
op = 'add', reduce_op = 'min'
lhs_data = tensor([1.2745, 1.4446, 1.3047, 1.5860, 1.5518, 1.2751, 1.8491, 1.7185, 1.3725,
        1.6673, 1.9251, 1.7839, 1.5065...0298, 1.4580, 1.1515, 1.3048, 1.7336, 1.6770,
        1.5970, 1.4968, 1.3797], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([1.9333, 1.9221, 1.7948, 1.6241, 1.1608, 1.8610, 1.1260, 1.9925, 1.7447,
        1.3619, 1.5636, 1.7703, 1.1443...1332, 1.8395, 1.6590, 1.9270, 1.1636, 1.2094,
        1.8527, 1.4667, 1.6586], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30]), e shape: torch.Size([300])
SpMM(message func: add, reduce func: min)
______________________ test_spmm[idtype0-min-sub-shp0-g0] ______________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), msg = 'sub', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb70597f7e0>
op = 'add', reduce_op = 'min'
lhs_data = tensor([[[[[[1.3514],
            [1.5375],
            [1.6178]]],


          [[[1.7128],
            [1.7941],
    ...1]]],


          [[[1.0381],
            [1.2439],
            [1.9254]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[-1.6686]],

           [[-1.8564]],

           [[-1.0032]]]],



         [[[[-1.6100]],

           [[-...       [[[[-1.2270]],

           [[-1.0472]],

           [[-1.3987]]]]]], dtype=torch.float64, grad_fn=<NegBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 1, 2, 1, 3, 1]), e shape: torch.Size([100, 4, 1, 3, 1, 1])
SpMM(message func: sub, reduce func: min)
______________________ test_spmm[idtype0-min-sub-shp0-g1] ______________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), msg = 'sub', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb7059bd558>
op = 'add', reduce_op = 'min'
lhs_data = tensor([[[[[[1.4312],
            [1.1724],
            [1.6604]]],


          [[[1.8548],
            [1.9040],
    ...4]]],


          [[[1.8733],
            [1.5254],
            [1.0065]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[-1.7280]],

           [[-1.8349]],

           [[-1.3683]]]],



         [[[[-1.8225]],

           [[-...       [[[[-1.0842]],

           [[-1.8898]],

           [[-1.1048]]]]]], dtype=torch.float64, grad_fn=<NegBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 1, 2, 1, 3, 1]), e shape: torch.Size([300, 4, 1, 3, 1, 1])
SpMM(message func: sub, reduce func: min)
______________________ test_spmm[idtype0-min-sub-shp1-g0] ______________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1, 3), dtype=torch.float64)})
shp = ((3, 3), (1, 3)), msg = 'sub', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43eb574c8>
op = 'add', reduce_op = 'min'
lhs_data = tensor([[[1.7320, 1.3761, 1.2906],
         [1.4485, 1.4773, 1.1799],
         [1.8023, 1.7670, 1.5560]],

        [[1...3757],
         [1.8242, 1.8138, 1.6500],
         [1.3730, 1.6684, 1.1015]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[-1.6311, -1.7257, -1.7990]],

        [[-1.5951, -1.0861, -1.8905]],

        [[-1.9447, -1.6224, -1.1484]],...1.2855, -1.6390, -1.4123]],

        [[-1.2682, -1.1543, -1.6796]]], dtype=torch.float64,
       grad_fn=<NegBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 3, 3]), e shape: torch.Size([100, 1, 3])
SpMM(message func: sub, reduce func: min)
______________________ test_spmm[idtype0-min-sub-shp1-g1] ______________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3, 3), (1, 3)), msg = 'sub', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb7058febd0>
op = 'add', reduce_op = 'min'
lhs_data = tensor([[[1.1611, 1.0155, 1.7183],
         [1.1319, 1.1977, 1.1885],
         [1.8994, 1.2824, 1.1078]],

        [[1...2488],
         [1.8440, 1.9789, 1.8949],
         [1.1208, 1.0673, 1.6736]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[-1.6182, -1.3635, -1.9018]],

        [[-1.8643, -1.5363, -1.1462]],

        [[-1.9545, -1.5106, -1.8125]],...1.9415, -1.5686, -1.8972]],

        [[-1.0342, -1.8136, -1.7133]]], dtype=torch.float64,
       grad_fn=<NegBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 3, 3]), e shape: torch.Size([300, 1, 3])
SpMM(message func: sub, reduce func: min)
______________________ test_spmm[idtype0-min-sub-shp2-g0] ______________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float64)})
shp = ((1,), (3,)), msg = 'sub', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72bfbe828>
op = 'add', reduce_op = 'min'
lhs_data = tensor([[1.6845],
        [1.5344],
        [1.1162],
        [1.5949],
        [1.3401],
        [1.6216],
        [1...447],
        [1.3013],
        [1.2708],
        [1.2812],
        [1.5030]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.0639, -1.1143, -1.1664],
        [-1.3622, -1.6741, -1.3363],
        [-1.5725, -1.1180, -1.0609],
       ...  [-1.8670, -1.7154, -1.7603],
        [-1.3835, -1.6139, -1.1620]], dtype=torch.float64,
       grad_fn=<NegBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1, 3), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 1]), e shape: torch.Size([100, 3])
SpMM(message func: sub, reduce func: min)
______________________ test_spmm[idtype0-min-sub-shp2-g1] ______________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (3,)), msg = 'sub', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb70598ab88>
op = 'add', reduce_op = 'min'
lhs_data = tensor([[1.1626],
        [1.7283],
        [1.7441],
        [1.4268],
        [1.2400],
        [1.4884],
        [1...211],
        [1.6391],
        [1.0343],
        [1.3863],
        [1.6503]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.8947, -1.5271, -1.9227],
        [-1.5697, -1.2404, -1.0653],
        [-1.9688, -1.5758, -1.5447],
       ...  [-1.7592, -1.0681, -1.5523],
        [-1.9421, -1.4276, -1.4928]], dtype=torch.float64,
       grad_fn=<NegBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 1]), e shape: torch.Size([300, 3])
SpMM(message func: sub, reduce func: min)
______________________ test_spmm[idtype0-min-sub-shp3-g0] ______________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (1,)), msg = 'sub', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72932d5a0>
op = 'add', reduce_op = 'min'
lhs_data = tensor([[1.2231, 1.3956, 1.0095],
        [1.3488, 1.2321, 1.3779],
        [1.6827, 1.8295, 1.1623],
        [1.6806,... 1.0154],
        [1.1838, 1.1142, 1.6111],
        [1.8896, 1.7841, 1.7062]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.5936],
        [-1.9246],
        [-1.4426],
        [-1.0135],
        [-1.4035],
        [-1.9186],
    ...       [-1.1363],
        [-1.5750],
        [-1.6423],
        [-1.2860]], dtype=torch.float64, grad_fn=<NegBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 3]), e shape: torch.Size([100, 1])
SpMM(message func: sub, reduce func: min)
______________________ test_spmm[idtype0-min-sub-shp3-g1] ______________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (1,)), msg = 'sub', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72bf67168>
op = 'add', reduce_op = 'min'
lhs_data = tensor([[1.3648, 1.3387, 1.7525],
        [1.7478, 1.9100, 1.9172],
        [1.4694, 1.4009, 1.1797],
        [1.0128,... 1.7727],
        [1.6752, 1.7547, 1.5282],
        [1.1029, 1.6945, 1.5492]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.7977],
        [-1.2859],
        [-1.3162],
        [-1.9863],
        [-1.6940],
        [-1.4588],
    ...       [-1.9548],
        [-1.7385],
        [-1.1114],
        [-1.1942]], dtype=torch.float64, grad_fn=<NegBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 3]), e shape: torch.Size([300, 1])
SpMM(message func: sub, reduce func: min)
______________________ test_spmm[idtype0-min-sub-shp4-g0] ______________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), msg = 'sub', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb705915cf0>
op = 'add', reduce_op = 'min'
lhs_data = tensor([[1.3806],
        [1.1330],
        [1.1165],
        [1.8482],
        [1.2378],
        [1.7999],
        [1...858],
        [1.8786],
        [1.5580],
        [1.9433],
        [1.8936]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.1589],
        [-1.5319],
        [-1.7057],
        [-1.7941],
        [-1.9985],
        [-1.7153],
    ...       [-1.7015],
        [-1.4511],
        [-1.9180],
        [-1.7006]], dtype=torch.float64, grad_fn=<NegBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 1]), e shape: torch.Size([100, 1])
SpMM(message func: sub, reduce func: min)
______________________ test_spmm[idtype0-min-sub-shp4-g1] ______________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), msg = 'sub', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb705915750>
op = 'add', reduce_op = 'min'
lhs_data = tensor([[1.6428],
        [1.5059],
        [1.4235],
        [1.4180],
        [1.5165],
        [1.9007],
        [1...467],
        [1.3290],
        [1.8098],
        [1.0262],
        [1.0055]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.3738],
        [-1.2945],
        [-1.2212],
        [-1.6249],
        [-1.6990],
        [-1.9817],
    ...       [-1.4983],
        [-1.0987],
        [-1.6287],
        [-1.6828]], dtype=torch.float64, grad_fn=<NegBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 1]), e shape: torch.Size([300, 1])
SpMM(message func: sub, reduce func: min)
______________________ test_spmm[idtype0-min-sub-shp5-g0] ______________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
shp = ((), ()), msg = 'sub', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72ab29090>
op = 'add', reduce_op = 'min'
lhs_data = tensor([1.3283, 1.1499, 1.9372, 1.6175, 1.1832, 1.0075, 1.3367, 1.6092, 1.7783,
        1.5867, 1.2794, 1.2947, 1.9811...6413, 1.6789, 1.3062, 1.9078, 1.2195, 1.5339,
        1.1671, 1.0140, 1.7579], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([-1.6945, -1.9605, -1.9347, -1.3747, -1.0393, -1.4547, -1.1764, -1.1987,
        -1.7835, -1.3870, -1.1389, -1....4250, -1.9392, -1.2195,
        -1.3517, -1.5885, -1.2311, -1.7991], dtype=torch.float64,
       grad_fn=<NegBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30]), e shape: torch.Size([100])
SpMM(message func: sub, reduce func: min)
______________________ test_spmm[idtype0-min-sub-shp5-g1] ______________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((), ()), msg = 'sub', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72ab29360>
op = 'add', reduce_op = 'min'
lhs_data = tensor([1.7186, 1.7097, 1.0614, 1.5061, 1.0384, 1.4516, 1.2893, 1.1022, 1.7407,
        1.9788, 1.1215, 1.0483, 1.9106...0422, 1.4995, 1.4196, 1.1074, 1.6233, 1.5614,
        1.1640, 1.0416, 1.0911], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([-1.2345, -1.0668, -1.1419, -1.4954, -1.8500, -1.5042, -1.2209, -1.2984,
        -1.1438, -1.1802, -1.8769, -1....5866, -1.3025, -1.1669,
        -1.1558, -1.4362, -1.8451, -1.8220], dtype=torch.float64,
       grad_fn=<NegBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30]), e shape: torch.Size([300])
SpMM(message func: sub, reduce func: min)
______________________ test_spmm[idtype0-min-mul-shp0-g0] ______________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), msg = 'mul', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43ea0d900>
op = 'mul', reduce_op = 'min'
lhs_data = tensor([[[[[[1.1094],
            [1.1577],
            [1.0158]]],


          [[[1.0955],
            [1.3705],
    ...1]]],


          [[[1.2389],
            [1.3454],
            [1.2483]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.8801]],

           [[1.0881]],

           [[1.2644]]]],



         [[[[1.0247]],

           [[1.514...



         [[[[1.7822]],

           [[1.5983]],

           [[1.2552]]]]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 1, 2, 1, 3, 1]), e shape: torch.Size([100, 4, 1, 3, 1, 1])
SpMM(message func: mul, reduce func: min)
______________________ test_spmm[idtype0-min-mul-shp0-g1] ______________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), msg = 'mul', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43eb91168>
op = 'mul', reduce_op = 'min'
lhs_data = tensor([[[[[[1.7606],
            [1.8144],
            [1.4749]]],


          [[[1.2901],
            [1.9895],
    ...1]]],


          [[[1.6079],
            [1.9016],
            [1.5327]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.7770]],

           [[1.0506]],

           [[1.9817]]]],



         [[[[1.1020]],

           [[1.406...



         [[[[1.5667]],

           [[1.1099]],

           [[1.6353]]]]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 1, 2, 1, 3, 1]), e shape: torch.Size([300, 4, 1, 3, 1, 1])
SpMM(message func: mul, reduce func: min)
______________________ test_spmm[idtype0-min-mul-shp1-g0] ______________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1, 3), dtype=torch.float64)})
shp = ((3, 3), (1, 3)), msg = 'mul', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72b2993f0>
op = 'mul', reduce_op = 'min'
lhs_data = tensor([[[1.6401, 1.9532, 1.2558],
         [1.4173, 1.1467, 1.0517],
         [1.5136, 1.4918, 1.3467]],

        [[1...8810],
         [1.8554, 1.6190, 1.7464],
         [1.8361, 1.1985, 1.0727]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[1.4391, 1.2748, 1.8309]],

        [[1.0521, 1.3079, 1.1467]],

        [[1.8286, 1.4462, 1.0532]],

       ...]],

        [[1.1692, 1.0901, 1.1074]],

        [[1.6884, 1.4915, 1.8326]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 3, 3]), e shape: torch.Size([100, 1, 3])
SpMM(message func: mul, reduce func: min)
______________________ test_spmm[idtype0-min-mul-shp1-g1] ______________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3, 3), (1, 3)), msg = 'mul', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72b299750>
op = 'mul', reduce_op = 'min'
lhs_data = tensor([[[1.8555, 1.3776, 1.8515],
         [1.1993, 1.1307, 1.7017],
         [1.1510, 1.2103, 1.2269]],

        [[1...2465],
         [1.0904, 1.0035, 1.5136],
         [1.0908, 1.6034, 1.1005]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[1.2124, 1.4891, 1.8288]],

        [[1.5399, 1.4555, 1.8356]],

        [[1.8392, 1.0377, 1.1803]],

       ...]],

        [[1.5023, 1.4982, 1.3307]],

        [[1.8430, 1.9342, 1.5776]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 3, 3]), e shape: torch.Size([300, 1, 3])
SpMM(message func: mul, reduce func: min)
______________________ test_spmm[idtype0-min-mul-shp2-g0] ______________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float64)})
shp = ((1,), (3,)), msg = 'mul', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb705938798>
op = 'mul', reduce_op = 'min'
lhs_data = tensor([[1.4534],
        [1.2493],
        [1.3906],
        [1.9280],
        [1.3044],
        [1.6196],
        [1...406],
        [1.2074],
        [1.9935],
        [1.8287],
        [1.6576]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.2954, 1.1932, 1.1982],
        [1.3975, 1.7715, 1.4566],
        [1.9144, 1.5380, 1.4101],
        [1.8543,... 1.9252],
        [1.6589, 1.9409, 1.6025],
        [1.7075, 1.7750, 1.4058]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1, 3), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 1]), e shape: torch.Size([100, 3])
SpMM(message func: mul, reduce func: min)
______________________ test_spmm[idtype0-min-mul-shp2-g1] ______________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (3,)), msg = 'mul', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43e71c750>
op = 'mul', reduce_op = 'min'
lhs_data = tensor([[1.8577],
        [1.4468],
        [1.7156],
        [1.2442],
        [1.8403],
        [1.7911],
        [1...523],
        [1.2846],
        [1.9968],
        [1.5554],
        [1.0376]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.8526, 1.1920, 1.6067],
        [1.0452, 1.3839, 1.5136],
        [1.3186, 1.5975, 1.2141],
        [1.4315,... 1.5244],
        [1.4728, 1.0942, 1.1273],
        [1.7602, 1.5139, 1.9166]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 1]), e shape: torch.Size([300, 3])
SpMM(message func: mul, reduce func: min)
______________________ test_spmm[idtype0-min-mul-shp3-g0] ______________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (1,)), msg = 'mul', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72c01b6c0>
op = 'mul', reduce_op = 'min'
lhs_data = tensor([[1.2695, 1.4507, 1.9591],
        [1.7409, 1.9274, 1.2853],
        [1.0555, 1.3629, 1.8810],
        [1.3217,... 1.4797],
        [1.1902, 1.2625, 1.4448],
        [1.1351, 1.5790, 1.5343]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.4696],
        [1.5453],
        [1.6263],
        [1.5489],
        [1.2393],
        [1.2301],
        [1...252],
        [1.0971],
        [1.5067],
        [1.2049],
        [1.1038]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 3]), e shape: torch.Size([100, 1])
SpMM(message func: mul, reduce func: min)
______________________ test_spmm[idtype0-min-mul-shp3-g1] ______________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (1,)), msg = 'mul', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72c01b8b8>
op = 'mul', reduce_op = 'min'
lhs_data = tensor([[1.8630, 1.3216, 1.7894],
        [1.2093, 1.5291, 1.3904],
        [1.4678, 1.2692, 1.9398],
        [1.8948,... 1.5221],
        [1.5688, 1.8623, 1.2004],
        [1.3724, 1.4753, 1.5450]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.8584],
        [1.1360],
        [1.0961],
        [1.8880],
        [1.3657],
        [1.1984],
        [1...377],
        [1.8909],
        [1.3968],
        [1.9293],
        [1.3526]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 3]), e shape: torch.Size([300, 1])
SpMM(message func: mul, reduce func: min)
______________________ test_spmm[idtype0-min-mul-shp4-g0] ______________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), msg = 'mul', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43ebae708>
op = 'mul', reduce_op = 'min'
lhs_data = tensor([[1.0659],
        [1.2545],
        [1.7910],
        [1.5625],
        [1.4345],
        [1.6125],
        [1...695],
        [1.0737],
        [1.4050],
        [1.8741],
        [1.4402]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.0583],
        [1.9795],
        [1.2976],
        [1.7550],
        [1.7640],
        [1.4871],
        [1...115],
        [1.6221],
        [1.1724],
        [1.3289],
        [1.5920]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 1]), e shape: torch.Size([100, 1])
SpMM(message func: mul, reduce func: min)
______________________ test_spmm[idtype0-min-mul-shp4-g1] ______________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), msg = 'mul', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb705a60678>
op = 'mul', reduce_op = 'min'
lhs_data = tensor([[1.6384],
        [1.3057],
        [1.8569],
        [1.2275],
        [1.9012],
        [1.0821],
        [1...514],
        [1.9201],
        [1.7979],
        [1.9199],
        [1.5082]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.3756],
        [1.4100],
        [1.5017],
        [1.2107],
        [1.1631],
        [1.7269],
        [1...765],
        [1.3175],
        [1.3241],
        [1.3228],
        [1.0044]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 1]), e shape: torch.Size([300, 1])
SpMM(message func: mul, reduce func: min)
______________________ test_spmm[idtype0-min-mul-shp5-g0] ______________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
shp = ((), ()), msg = 'mul', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72c006a20>
op = 'mul', reduce_op = 'min'
lhs_data = tensor([1.9991, 1.7531, 1.2939, 1.8333, 1.4182, 1.0288, 1.4599, 1.1317, 1.1531,
        1.1563, 1.3830, 1.0487, 1.4476...5815, 1.0442, 1.0620, 1.5130, 1.3501, 1.9202,
        1.8254, 1.5714, 1.2011], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([1.9746, 1.9457, 1.4082, 1.2389, 1.7288, 1.3519, 1.7175, 1.9306, 1.0590,
        1.5110, 1.4003, 1.0192, 1.3998...7307, 1.5935, 1.8833, 1.0861, 1.0037, 1.9302, 1.1350, 1.1015,
        1.8643], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30]), e shape: torch.Size([100])
SpMM(message func: mul, reduce func: min)
______________________ test_spmm[idtype0-min-mul-shp5-g1] ______________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((), ()), msg = 'mul', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72c0065e8>
op = 'mul', reduce_op = 'min'
lhs_data = tensor([1.5561, 1.0071, 1.0717, 1.0897, 1.5560, 1.1646, 1.1865, 1.3131, 1.8647,
        1.0211, 1.2740, 1.9726, 1.0695...5843, 1.5390, 1.5151, 1.6871, 1.3024, 1.9913,
        1.3063, 1.4451, 1.8041], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([1.1942, 1.4060, 1.9566, 1.5567, 1.1590, 1.4907, 1.1329, 1.0052, 1.7790,
        1.1886, 1.6181, 1.2258, 1.2114...8703, 1.6968, 1.0459, 1.8607, 1.1233, 1.9670,
        1.1578, 1.4099, 1.7395], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30]), e shape: torch.Size([300])
SpMM(message func: mul, reduce func: min)
______________________ test_spmm[idtype0-min-div-shp0-g0] ______________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), msg = 'div', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43e5725e8>
op = 'mul', reduce_op = 'min'
lhs_data = tensor([[[[[[1.1138],
            [1.0022],
            [1.9241]]],


          [[[1.4423],
            [1.1856],
    ...2]]],


          [[[1.9894],
            [1.8014],
            [1.6993]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[0.6494]],

           [[0.6875]],

           [[0.9919]]]],



         [[[[0.5182]],

           [[0.986...         [[[[0.5131]],

           [[0.5730]],

           [[0.8784]]]]]], dtype=torch.float64, grad_fn=<MulBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 1, 2, 1, 3, 1]), e shape: torch.Size([100, 4, 1, 3, 1, 1])
SpMM(message func: div, reduce func: min)
______________________ test_spmm[idtype0-min-div-shp0-g1] ______________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), msg = 'div', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb45517a360>
op = 'mul', reduce_op = 'min'
lhs_data = tensor([[[[[[1.0881],
            [1.5295],
            [1.6439]]],


          [[[1.1347],
            [1.1431],
    ...7]]],


          [[[1.4702],
            [1.0810],
            [1.5049]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[0.8315]],

           [[0.8681]],

           [[0.5367]]]],



         [[[[0.5080]],

           [[0.572...         [[[[0.5615]],

           [[0.5595]],

           [[0.5683]]]]]], dtype=torch.float64, grad_fn=<MulBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 1, 2, 1, 3, 1]), e shape: torch.Size([300, 4, 1, 3, 1, 1])
SpMM(message func: div, reduce func: min)
______________________ test_spmm[idtype0-min-div-shp1-g0] ______________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1, 3), dtype=torch.float64)})
shp = ((3, 3), (1, 3)), msg = 'div', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb705a885a0>
op = 'mul', reduce_op = 'min'
lhs_data = tensor([[[1.6326, 1.2039, 1.1745],
         [1.0549, 1.9031, 1.5341],
         [1.4877, 1.7274, 1.7217]],

        [[1...2215],
         [1.5981, 1.6551, 1.1043],
         [1.2382, 1.3143, 1.4383]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[0.6554, 0.6850, 0.7219]],

        [[0.6504, 0.6280, 0.5868]],

        [[0.5611, 0.5057, 0.8433]],

       ...  [[0.6169, 0.9513, 0.8281]],

        [[0.9744, 0.6721, 0.9025]]], dtype=torch.float64,
       grad_fn=<MulBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 3, 3]), e shape: torch.Size([100, 1, 3])
SpMM(message func: div, reduce func: min)
______________________ test_spmm[idtype0-min-div-shp1-g1] ______________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3, 3), (1, 3)), msg = 'div', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43e56aaf8>
op = 'mul', reduce_op = 'min'
lhs_data = tensor([[[1.7129, 1.3390, 1.9207],
         [1.2493, 1.3820, 1.1657],
         [1.4051, 1.6324, 1.2700]],

        [[1...6488],
         [1.4589, 1.0403, 1.1572],
         [1.8744, 1.3302, 1.7853]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[0.5747, 0.7841, 0.6623]],

        [[0.6335, 0.5194, 0.9935]],

        [[0.6843, 0.7845, 0.7297]],

       ...  [[0.7136, 0.7043, 0.8721]],

        [[0.5394, 0.8735, 0.6149]]], dtype=torch.float64,
       grad_fn=<MulBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 3, 3]), e shape: torch.Size([300, 1, 3])
SpMM(message func: div, reduce func: min)
______________________ test_spmm[idtype0-min-div-shp2-g0] ______________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float64)})
shp = ((1,), (3,)), msg = 'div', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72ab11090>
op = 'mul', reduce_op = 'min'
lhs_data = tensor([[1.4762],
        [1.7209],
        [1.9148],
        [1.4821],
        [1.1799],
        [1.3826],
        [1...463],
        [1.9981],
        [1.8153],
        [1.6818],
        [1.0280]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.7431, 0.7950, 0.5887],
        [0.7882, 0.6318, 0.5305],
        [0.7151, 0.5491, 0.8115],
        [0.5437,...222],
        [0.5374, 0.5333, 0.7308],
        [0.8079, 0.6935, 0.5146]], dtype=torch.float64, grad_fn=<MulBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1, 3), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 1]), e shape: torch.Size([100, 3])
SpMM(message func: div, reduce func: min)
______________________ test_spmm[idtype0-min-div-shp2-g1] ______________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (3,)), msg = 'div', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72ab11af8>
op = 'mul', reduce_op = 'min'
lhs_data = tensor([[1.8446],
        [1.2595],
        [1.5330],
        [1.0019],
        [1.8563],
        [1.3785],
        [1...470],
        [1.8698],
        [1.1476],
        [1.1656],
        [1.7829]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.9706, 0.5676, 0.6208],
        [0.5998, 0.6292, 0.8583],
        [0.9958, 0.9696, 0.8232],
        [0.5306,...830],
        [0.7192, 0.5214, 0.9936],
        [0.5726, 0.8819, 0.6156]], dtype=torch.float64, grad_fn=<MulBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 1]), e shape: torch.Size([300, 3])
SpMM(message func: div, reduce func: min)
______________________ test_spmm[idtype0-min-div-shp3-g0] ______________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (1,)), msg = 'div', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72ab29318>
op = 'mul', reduce_op = 'min'
lhs_data = tensor([[1.3345, 1.0574, 1.3575],
        [1.2572, 1.1618, 1.2650],
        [1.3651, 1.4718, 1.6712],
        [1.7170,... 1.8375],
        [1.8265, 1.2033, 1.4783],
        [1.5954, 1.0222, 1.4144]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.7375],
        [0.5867],
        [0.5311],
        [0.8601],
        [0.9993],
        [0.8844],
        [0...,
        [0.5005],
        [0.8325],
        [0.6216],
        [0.5055]], dtype=torch.float64, grad_fn=<MulBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 3]), e shape: torch.Size([100, 1])
SpMM(message func: div, reduce func: min)
______________________ test_spmm[idtype0-min-div-shp3-g1] ______________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (1,)), msg = 'div', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb705aa0480>
op = 'mul', reduce_op = 'min'
lhs_data = tensor([[1.0291, 1.6080, 1.5929],
        [1.6941, 1.4055, 1.6903],
        [1.7402, 1.5274, 1.0813],
        [1.8183,... 1.6998],
        [1.3428, 1.6187, 1.2212],
        [1.7766, 1.2219, 1.8398]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.5773],
        [0.6553],
        [0.8424],
        [0.5639],
        [0.5369],
        [0.8558],
        [0...,
        [0.6927],
        [0.7270],
        [0.6374],
        [0.9975]], dtype=torch.float64, grad_fn=<MulBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 3]), e shape: torch.Size([300, 1])
SpMM(message func: div, reduce func: min)
______________________ test_spmm[idtype0-min-div-shp4-g0] ______________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), msg = 'div', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72ab092d0>
op = 'mul', reduce_op = 'min'
lhs_data = tensor([[1.2958],
        [1.8718],
        [1.8766],
        [1.6618],
        [1.4314],
        [1.7738],
        [1...740],
        [1.7972],
        [1.8501],
        [1.0837],
        [1.8284]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.9643],
        [0.6293],
        [0.5654],
        [0.7310],
        [0.6283],
        [0.5749],
        [0...,
        [0.5406],
        [0.5780],
        [0.5099],
        [0.7606]], dtype=torch.float64, grad_fn=<MulBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 1]), e shape: torch.Size([100, 1])
SpMM(message func: div, reduce func: min)
______________________ test_spmm[idtype0-min-div-shp4-g1] ______________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), msg = 'div', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72ab09558>
op = 'mul', reduce_op = 'min'
lhs_data = tensor([[1.0724],
        [1.1543],
        [1.0248],
        [1.1978],
        [1.1577],
        [1.6484],
        [1...950],
        [1.8805],
        [1.0211],
        [1.1439],
        [1.9464]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.5079],
        [0.8889],
        [0.6705],
        [0.8538],
        [0.8992],
        [0.5584],
        [0...,
        [0.8948],
        [0.7090],
        [0.7816],
        [0.6349]], dtype=torch.float64, grad_fn=<MulBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 1]), e shape: torch.Size([300, 1])
SpMM(message func: div, reduce func: min)
______________________ test_spmm[idtype0-min-div-shp5-g0] ______________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
shp = ((), ()), msg = 'div', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43ea3f510>
op = 'mul', reduce_op = 'min'
lhs_data = tensor([1.3543, 1.4228, 1.4527, 1.2140, 1.7171, 1.1715, 1.4447, 1.7856, 1.2859,
        1.6294, 1.7122, 1.9256, 1.8243...5188, 1.7902, 1.5827, 1.5346, 1.6453, 1.3571,
        1.5679, 1.4737, 1.2501], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([0.6166, 0.8431, 0.6601, 0.5638, 0.6587, 0.9966, 0.6093, 0.8652, 0.5851,
        0.5543, 0.5872, 0.6393, 0.8996..., 0.6784, 0.6023, 0.5794, 0.5955, 0.5119, 0.5048, 0.6913,
        0.5360], dtype=torch.float64, grad_fn=<MulBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30]), e shape: torch.Size([100])
SpMM(message func: div, reduce func: min)
______________________ test_spmm[idtype0-min-div-shp5-g1] ______________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((), ()), msg = 'div', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43eb9e480>
op = 'mul', reduce_op = 'min'
lhs_data = tensor([1.4133, 1.7780, 1.7752, 1.4689, 1.7268, 1.2512, 1.9331, 1.3180, 1.2548,
        1.6282, 1.6988, 1.1262, 1.1126...6810, 1.3637, 1.6592, 1.9873, 1.1246, 1.6865,
        1.3687, 1.0176, 1.6033], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([0.8898, 0.5066, 0.6875, 0.9768, 0.5125, 0.5610, 0.7921, 0.5266, 0.8717,
        0.6381, 0.6893, 0.6207, 0.9531..., 0.9252, 0.8380, 0.7402, 0.5831, 0.6122,
        0.6134, 0.6012, 0.7883], dtype=torch.float64, grad_fn=<MulBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30]), e shape: torch.Size([300])
SpMM(message func: div, reduce func: min)
___________________ test_spmm[idtype0-min-copy_lhs-shp0-g0] ____________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), msg = 'copy_lhs', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb732ce4ee8>
op = 'copy_lhs', reduce_op = 'min'
lhs_data = tensor([[[[[[1.4864],
            [1.0332],
            [1.3697]]],


          [[[1.7931],
            [1.4794],
    ...8]]],


          [[[1.2944],
            [1.9707],
            [1.5755]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.2865]],

           [[1.1476]],

           [[1.2312]]]],



         [[[[1.5051]],

           [[1.064...



         [[[[1.7603]],

           [[1.2995]],

           [[1.7055]]]]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 1, 2, 1, 3, 1]), e shape: torch.Size([100, 4, 1, 3, 1, 1])
SpMM(message func: copy_lhs, reduce func: min)
___________________ test_spmm[idtype0-min-copy_lhs-shp0-g1] ____________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), msg = 'copy_lhs', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb7049c5120>
op = 'copy_lhs', reduce_op = 'min'
lhs_data = tensor([[[[[[1.7879],
            [1.2010],
            [1.7030]]],


          [[[1.2291],
            [1.5786],
    ...8]]],


          [[[1.9657],
            [1.0880],
            [1.2586]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.0220]],

           [[1.6451]],

           [[1.5775]]]],



         [[[[1.8598]],

           [[1.953...



         [[[[1.4793]],

           [[1.6151]],

           [[1.9346]]]]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 1, 2, 1, 3, 1]), e shape: torch.Size([300, 4, 1, 3, 1, 1])
SpMM(message func: copy_lhs, reduce func: min)
___________________ test_spmm[idtype0-min-copy_lhs-shp1-g0] ____________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1, 3), dtype=torch.float64)})
shp = ((3, 3), (1, 3)), msg = 'copy_lhs', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72ab1f1b0>
op = 'copy_lhs', reduce_op = 'min'
lhs_data = tensor([[[1.6567, 1.1607, 1.0953],
         [1.1791, 1.6763, 1.6591],
         [1.0671, 1.7188, 1.7130]],

        [[1...3990],
         [1.8905, 1.9012, 1.6384],
         [1.3949, 1.7025, 1.0475]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[1.6336, 1.0768, 1.7676]],

        [[1.3118, 1.0929, 1.4314]],

        [[1.1557, 1.2726, 1.1903]],

       ...]],

        [[1.0049, 1.2379, 1.7082]],

        [[1.8118, 1.9856, 1.5737]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 3, 3]), e shape: torch.Size([100, 1, 3])
SpMM(message func: copy_lhs, reduce func: min)
___________________ test_spmm[idtype0-min-copy_lhs-shp1-g1] ____________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3, 3), (1, 3)), msg = 'copy_lhs', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72ab1f3f0>
op = 'copy_lhs', reduce_op = 'min'
lhs_data = tensor([[[1.2807, 1.7475, 1.6994],
         [1.0755, 1.3219, 1.7354],
         [1.0304, 1.2293, 1.1335]],

        [[1...7250],
         [1.5421, 1.6717, 1.9188],
         [1.0872, 1.8892, 1.0019]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[1.4151, 1.1433, 1.3580]],

        [[1.2472, 1.1119, 1.3119]],

        [[1.6937, 1.1085, 1.3096]],

       ...]],

        [[1.5026, 1.7068, 1.7178]],

        [[1.3235, 1.4363, 1.7216]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 3, 3]), e shape: torch.Size([300, 1, 3])
SpMM(message func: copy_lhs, reduce func: min)
___________________ test_spmm[idtype0-min-copy_lhs-shp2-g0] ____________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float64)})
shp = ((1,), (3,)), msg = 'copy_lhs', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72bf9f990>
op = 'copy_lhs', reduce_op = 'min'
lhs_data = tensor([[1.2940],
        [1.1722],
        [1.9179],
        [1.2632],
        [1.8029],
        [1.2306],
        [1...191],
        [1.0817],
        [1.8098],
        [1.2779],
        [1.8459]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.0197, 1.0459, 1.1780],
        [1.1862, 1.3184, 1.5801],
        [1.0240, 1.4399, 1.0381],
        [1.9369,... 1.4293],
        [1.6705, 1.2382, 1.3451],
        [1.1051, 1.5993, 1.7518]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1, 3), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 1]), e shape: torch.Size([100, 3])
SpMM(message func: copy_lhs, reduce func: min)
___________________ test_spmm[idtype0-min-copy_lhs-shp2-g1] ____________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (3,)), msg = 'copy_lhs', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb7059d5630>
op = 'copy_lhs', reduce_op = 'min'
lhs_data = tensor([[1.0320],
        [1.8765],
        [1.1268],
        [1.8408],
        [1.8076],
        [1.1097],
        [1...840],
        [1.5753],
        [1.5955],
        [1.5770],
        [1.0560]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.9808, 1.5866, 1.0256],
        [1.7597, 1.4265, 1.7167],
        [1.0475, 1.7607, 1.6256],
        [1.2318,... 1.9269],
        [1.7434, 1.7852, 1.5972],
        [1.7010, 1.2172, 1.5527]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 1]), e shape: torch.Size([300, 3])
SpMM(message func: copy_lhs, reduce func: min)
___________________ test_spmm[idtype0-min-copy_lhs-shp3-g0] ____________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (1,)), msg = 'copy_lhs', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43e38ad38>
op = 'copy_lhs', reduce_op = 'min'
lhs_data = tensor([[1.9029, 1.4097, 1.3046],
        [1.3639, 1.8489, 1.1953],
        [1.9702, 1.1912, 1.6449],
        [1.9946,... 1.9386],
        [1.1181, 1.2967, 1.4620],
        [1.5490, 1.7773, 1.5782]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.5456],
        [1.9721],
        [1.6301],
        [1.8505],
        [1.4244],
        [1.6296],
        [1...218],
        [1.0535],
        [1.4711],
        [1.8235],
        [1.6254]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 3]), e shape: torch.Size([100, 1])
SpMM(message func: copy_lhs, reduce func: min)
___________________ test_spmm[idtype0-min-copy_lhs-shp3-g1] ____________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (1,)), msg = 'copy_lhs', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43e3c8e58>
op = 'copy_lhs', reduce_op = 'min'
lhs_data = tensor([[1.4454, 1.1053, 1.0226],
        [1.7260, 1.5270, 1.1793],
        [1.6601, 1.3207, 1.6422],
        [1.3360,... 1.3785],
        [1.9916, 1.1507, 1.2681],
        [1.5969, 1.5697, 1.1212]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.6433],
        [1.9308],
        [1.9981],
        [1.8541],
        [1.5525],
        [1.2599],
        [1...521],
        [1.1475],
        [1.8754],
        [1.1404],
        [1.1632]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 3]), e shape: torch.Size([300, 1])
SpMM(message func: copy_lhs, reduce func: min)
___________________ test_spmm[idtype0-min-copy_lhs-shp4-g0] ____________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), msg = 'copy_lhs', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb706547318>
op = 'copy_lhs', reduce_op = 'min'
lhs_data = tensor([[1.5549],
        [1.1664],
        [1.5616],
        [1.8898],
        [1.8246],
        [1.4588],
        [1...790],
        [1.6818],
        [1.0905],
        [1.5345],
        [1.4956]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.6412],
        [1.5511],
        [1.6272],
        [1.3848],
        [1.4163],
        [1.1454],
        [1...068],
        [1.8417],
        [1.8293],
        [1.3225],
        [1.5047]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 1]), e shape: torch.Size([100, 1])
SpMM(message func: copy_lhs, reduce func: min)
___________________ test_spmm[idtype0-min-copy_lhs-shp4-g1] ____________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), msg = 'copy_lhs', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72b250048>
op = 'copy_lhs', reduce_op = 'min'
lhs_data = tensor([[1.7606],
        [1.3806],
        [1.6949],
        [1.7318],
        [1.6746],
        [1.6239],
        [1...024],
        [1.5262],
        [1.0185],
        [1.5883],
        [1.4894]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.9256],
        [1.7758],
        [1.3901],
        [1.6491],
        [1.0080],
        [1.8986],
        [1...353],
        [1.8069],
        [1.0451],
        [1.5015],
        [1.2048]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 1]), e shape: torch.Size([300, 1])
SpMM(message func: copy_lhs, reduce func: min)
___________________ test_spmm[idtype0-min-copy_lhs-shp5-g0] ____________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
shp = ((), ()), msg = 'copy_lhs', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43eb49708>
op = 'copy_lhs', reduce_op = 'min'
lhs_data = tensor([1.3518, 1.3069, 1.4534, 1.0618, 1.9028, 1.3276, 1.4347, 1.2671, 1.1136,
        1.4444, 1.0371, 1.6555, 1.1926...7236, 1.3595, 1.2539, 1.4153, 1.0054, 1.6603,
        1.3831, 1.7773, 1.5697], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([1.6965, 1.5458, 1.7217, 1.4152, 1.2376, 1.6111, 1.8805, 1.4344, 1.7412,
        1.9396, 1.5357, 1.4802, 1.5816...8641, 1.4186, 1.9937, 1.6841, 1.0408, 1.6298, 1.9915, 1.1443,
        1.0874], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30]), e shape: torch.Size([100])
SpMM(message func: copy_lhs, reduce func: min)
___________________ test_spmm[idtype0-min-copy_lhs-shp5-g1] ____________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((), ()), msg = 'copy_lhs', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72a934828>
op = 'copy_lhs', reduce_op = 'min'
lhs_data = tensor([1.6642, 1.1389, 1.9626, 1.2112, 1.2117, 1.2243, 1.0199, 1.4497, 1.0322,
        1.7969, 1.5776, 1.0493, 1.0259...7929, 1.6374, 1.3503, 1.4400, 1.9170, 1.0800,
        1.4239, 1.3555, 1.6342], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([1.4269, 1.4065, 1.1330, 1.1856, 1.3226, 1.8293, 1.6538, 1.2012, 1.9114,
        1.5239, 1.4352, 1.5743, 1.8570...7755, 1.0767, 1.9778, 1.8014, 1.1638, 1.1122,
        1.4180, 1.3864, 1.4314], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30]), e shape: torch.Size([300])
SpMM(message func: copy_lhs, reduce func: min)
___________________ test_spmm[idtype0-min-copy_rhs-shp0-g0] ____________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), msg = 'copy_rhs', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb705ab3ea0>
op = 'copy_rhs', reduce_op = 'min'
lhs_data = tensor([[[[[[1.2861],
            [1.4740],
            [1.5132]]],


          [[[1.7331],
            [1.9137],
    ...1]]],


          [[[1.2524],
            [1.9724],
            [1.2768]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.6795]],

           [[1.1561]],

           [[1.8566]]]],



         [[[[1.2878]],

           [[1.983...



         [[[[1.1341]],

           [[1.7462]],

           [[1.8752]]]]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 1, 2, 1, 3, 1]), e shape: torch.Size([100, 4, 1, 3, 1, 1])
SpMM(message func: copy_rhs, reduce func: min)
___________________ test_spmm[idtype0-min-copy_rhs-shp0-g1] ____________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), msg = 'copy_rhs', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb7059d2828>
op = 'copy_rhs', reduce_op = 'min'
lhs_data = tensor([[[[[[1.3154],
            [1.8356],
            [1.6957]]],


          [[[1.4993],
            [1.9694],
    ...1]]],


          [[[1.7450],
            [1.2841],
            [1.1533]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.2506]],

           [[1.2305]],

           [[1.2623]]]],



         [[[[1.0162]],

           [[1.585...



         [[[[1.7416]],

           [[1.5601]],

           [[1.4736]]]]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 1, 2, 1, 3, 1]), e shape: torch.Size([300, 4, 1, 3, 1, 1])
SpMM(message func: copy_rhs, reduce func: min)
___________________ test_spmm[idtype0-min-copy_rhs-shp1-g0] ____________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1, 3), dtype=torch.float64)})
shp = ((3, 3), (1, 3)), msg = 'copy_rhs', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72af18990>
op = 'copy_rhs', reduce_op = 'min'
lhs_data = tensor([[[1.0000, 1.5257, 1.8336],
         [1.1057, 1.8962, 1.5127],
         [1.8652, 1.5493, 1.8373]],

        [[1...4147],
         [1.8580, 1.2324, 1.6349],
         [1.3625, 1.1889, 1.9191]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[1.6235, 1.8867, 1.0460]],

        [[1.1021, 1.9747, 1.2705]],

        [[1.6694, 1.1850, 1.4201]],

       ...]],

        [[1.1787, 1.0944, 1.6087]],

        [[1.0481, 1.1148, 1.4603]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 3, 3]), e shape: torch.Size([100, 1, 3])
SpMM(message func: copy_rhs, reduce func: min)
___________________ test_spmm[idtype0-min-copy_rhs-shp1-g1] ____________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3, 3), (1, 3)), msg = 'copy_rhs', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43eba1168>
op = 'copy_rhs', reduce_op = 'min'
lhs_data = tensor([[[1.3223, 1.6392, 1.4283],
         [1.9686, 1.4472, 1.5462],
         [1.5962, 1.9364, 1.8772]],

        [[1...3752],
         [1.2524, 1.1085, 1.2061],
         [1.7635, 1.6548, 1.7962]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[1.5426, 1.3834, 1.2039]],

        [[1.7254, 1.7935, 1.1352]],

        [[1.0422, 1.3360, 1.1733]],

       ...]],

        [[1.9447, 1.6623, 1.0403]],

        [[1.5260, 1.0270, 1.8514]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 3, 3]), e shape: torch.Size([300, 1, 3])
SpMM(message func: copy_rhs, reduce func: min)
___________________ test_spmm[idtype0-min-copy_rhs-shp2-g0] ____________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float64)})
shp = ((1,), (3,)), msg = 'copy_rhs', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72b24ff30>
op = 'copy_rhs', reduce_op = 'min'
lhs_data = tensor([[1.9825],
        [1.7323],
        [1.0089],
        [1.5063],
        [1.8715],
        [1.8382],
        [1...621],
        [1.5353],
        [1.5865],
        [1.3765],
        [1.9438]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.6470, 1.8324, 1.5301],
        [1.2997, 1.1876, 1.8170],
        [1.8179, 1.6092, 1.4850],
        [1.8303,... 1.9133],
        [1.2363, 1.9386, 1.1686],
        [1.8832, 1.8759, 1.9830]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1, 3), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 1]), e shape: torch.Size([100, 3])
SpMM(message func: copy_rhs, reduce func: min)
___________________ test_spmm[idtype0-min-copy_rhs-shp2-g1] ____________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (3,)), msg = 'copy_rhs', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72b24f708>
op = 'copy_rhs', reduce_op = 'min'
lhs_data = tensor([[1.9582],
        [1.5477],
        [1.1914],
        [1.1608],
        [1.4415],
        [1.5580],
        [1...557],
        [1.2004],
        [1.8216],
        [1.5399],
        [1.6995]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.2331, 1.5841, 1.9947],
        [1.5420, 1.1626, 1.1918],
        [1.1729, 1.4134, 1.5961],
        [1.4395,... 1.4353],
        [1.3045, 1.9080, 1.3258],
        [1.4593, 1.6232, 1.1285]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 1]), e shape: torch.Size([300, 3])
SpMM(message func: copy_rhs, reduce func: min)
___________________ test_spmm[idtype0-min-copy_rhs-shp3-g0] ____________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (1,)), msg = 'copy_rhs', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb70596ca20>
op = 'copy_rhs', reduce_op = 'min'
lhs_data = tensor([[1.5618, 1.6646, 1.0994],
        [1.0111, 1.4802, 1.3180],
        [1.9564, 1.9270, 1.9588],
        [1.3931,... 1.1766],
        [1.6301, 1.6650, 1.5692],
        [1.5305, 1.3456, 1.1910]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.5940],
        [1.5223],
        [1.9485],
        [1.0218],
        [1.3918],
        [1.1374],
        [1...399],
        [1.1036],
        [1.5259],
        [1.1591],
        [1.8347]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 3]), e shape: torch.Size([100, 1])
SpMM(message func: copy_rhs, reduce func: min)
___________________ test_spmm[idtype0-min-copy_rhs-shp3-g1] ____________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (1,)), msg = 'copy_rhs', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72ac2b828>
op = 'copy_rhs', reduce_op = 'min'
lhs_data = tensor([[1.4721, 1.8037, 1.4493],
        [1.9854, 1.1126, 1.4460],
        [1.4376, 1.3245, 1.6171],
        [1.2408,... 1.1595],
        [1.7619, 1.2215, 1.6391],
        [1.1307, 1.3027, 1.8680]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.1352],
        [1.2954],
        [1.5011],
        [1.4194],
        [1.6593],
        [1.3915],
        [1...953],
        [1.3539],
        [1.5897],
        [1.2885],
        [1.6908]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 3]), e shape: torch.Size([300, 1])
SpMM(message func: copy_rhs, reduce func: min)
___________________ test_spmm[idtype0-min-copy_rhs-shp4-g0] ____________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), msg = 'copy_rhs', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb7059924c8>
op = 'copy_rhs', reduce_op = 'min'
lhs_data = tensor([[1.5118],
        [1.4998],
        [1.1480],
        [1.6254],
        [1.4252],
        [1.5371],
        [1...190],
        [1.0962],
        [1.8882],
        [1.1042],
        [1.6544]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.8193],
        [1.1478],
        [1.4216],
        [1.7995],
        [1.5188],
        [1.3667],
        [1...359],
        [1.2684],
        [1.9288],
        [1.1181],
        [1.2467]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 1]), e shape: torch.Size([100, 1])
SpMM(message func: copy_rhs, reduce func: min)
___________________ test_spmm[idtype0-min-copy_rhs-shp4-g1] ____________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), msg = 'copy_rhs', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72932d360>
op = 'copy_rhs', reduce_op = 'min'
lhs_data = tensor([[1.0551],
        [1.6503],
        [1.1148],
        [1.9920],
        [1.0727],
        [1.2034],
        [1...584],
        [1.9880],
        [1.2535],
        [1.4955],
        [1.6233]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.8831],
        [1.4872],
        [1.9566],
        [1.5082],
        [1.0831],
        [1.8553],
        [1...989],
        [1.5484],
        [1.5196],
        [1.3693],
        [1.6919]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 1]), e shape: torch.Size([300, 1])
SpMM(message func: copy_rhs, reduce func: min)
___________________ test_spmm[idtype0-min-copy_rhs-shp5-g0] ____________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
shp = ((), ()), msg = 'copy_rhs', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb705a83af8>
op = 'copy_rhs', reduce_op = 'min'
lhs_data = tensor([1.2304, 1.4292, 1.9073, 1.3457, 1.9933, 1.8010, 1.8677, 1.8019, 1.7165,
        1.3694, 1.9239, 1.6534, 1.0068...8661, 1.6950, 1.5235, 1.3807, 1.3332, 1.4287,
        1.0687, 1.4150, 1.6943], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([1.6859, 1.7809, 1.7460, 1.9935, 1.8777, 1.8977, 1.6988, 1.1317, 1.7765,
        1.8282, 1.1768, 1.1380, 1.4596...0710, 1.7306, 1.1559, 1.4019, 1.1524, 1.2075, 1.5762, 1.0416,
        1.8902], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30]), e shape: torch.Size([100])
SpMM(message func: copy_rhs, reduce func: min)
___________________ test_spmm[idtype0-min-copy_rhs-shp5-g1] ____________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((), ()), msg = 'copy_rhs', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72b27d1b0>
op = 'copy_rhs', reduce_op = 'min'
lhs_data = tensor([1.9295, 1.2920, 1.6236, 1.1593, 1.7224, 1.1127, 1.7076, 1.7660, 1.5338,
        1.9190, 1.8216, 1.0588, 1.7317...8444, 1.5344, 1.9091, 1.4396, 1.6889, 1.0222,
        1.0162, 1.3150, 1.1554], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([1.6416, 1.5811, 1.6590, 1.6119, 1.9101, 1.0338, 1.1441, 1.3019, 1.3451,
        1.0929, 1.8555, 1.4152, 1.5931...1206, 1.1285, 1.5967, 1.4010, 1.3293, 1.3883,
        1.3298, 1.8272, 1.5288], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30]), e shape: torch.Size([300])
SpMM(message func: copy_rhs, reduce func: min)
______________________ test_spmm[idtype0-max-add-shp0-g0] ______________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), msg = 'add', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb729301d80>
op = 'add', reduce_op = 'max'
lhs_data = tensor([[[[[[1.6043],
            [1.6876],
            [1.6699]]],


          [[[1.8457],
            [1.6440],
    ...4]]],


          [[[1.3541],
            [1.6723],
            [1.7950]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.4516]],

           [[1.0661]],

           [[1.1743]]]],



         [[[[1.2535]],

           [[1.197...



         [[[[1.1510]],

           [[1.0158]],

           [[1.3630]]]]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 1, 2, 1, 3, 1]), e shape: torch.Size([100, 4, 1, 3, 1, 1])
SpMM(message func: add, reduce func: max)
______________________ test_spmm[idtype0-max-add-shp0-g1] ______________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), msg = 'add', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43e568708>
op = 'add', reduce_op = 'max'
lhs_data = tensor([[[[[[1.3413],
            [1.8904],
            [1.4366]]],


          [[[1.7681],
            [1.8938],
    ...1]]],


          [[[1.1234],
            [1.6714],
            [1.8472]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.3815]],

           [[1.7025]],

           [[1.7428]]]],



         [[[[1.4077]],

           [[1.033...



         [[[[1.9601]],

           [[1.1381]],

           [[1.3480]]]]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 1, 2, 1, 3, 1]), e shape: torch.Size([300, 4, 1, 3, 1, 1])
SpMM(message func: add, reduce func: max)
______________________ test_spmm[idtype0-max-add-shp1-g0] ______________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1, 3), dtype=torch.float64)})
shp = ((3, 3), (1, 3)), msg = 'add', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72a931318>
op = 'add', reduce_op = 'max'
lhs_data = tensor([[[1.0281, 1.2349, 1.8988],
         [1.6012, 1.0631, 1.0675],
         [1.4586, 1.1925, 1.5404]],

        [[1...9137],
         [1.2043, 1.6751, 1.9481],
         [1.7065, 1.2178, 1.5582]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[1.4874, 1.9086, 1.9045]],

        [[1.0983, 1.6715, 1.4013]],

        [[1.4935, 1.3226, 1.9880]],

       ...]],

        [[1.1885, 1.3207, 1.9555]],

        [[1.6941, 1.7552, 1.7328]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 3, 3]), e shape: torch.Size([100, 1, 3])
SpMM(message func: add, reduce func: max)
______________________ test_spmm[idtype0-max-add-shp1-g1] ______________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3, 3), (1, 3)), msg = 'add', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72af181f8>
op = 'add', reduce_op = 'max'
lhs_data = tensor([[[1.2563, 1.4003, 1.6367],
         [1.3839, 1.6973, 1.3616],
         [1.8952, 1.2297, 1.5386]],

        [[1...0983],
         [1.2578, 1.2060, 1.4956],
         [1.9623, 1.3857, 1.1778]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[1.5351, 1.2870, 1.1302]],

        [[1.7773, 1.9411, 1.4194]],

        [[1.0009, 1.5971, 1.5210]],

       ...]],

        [[1.4130, 1.1265, 1.3854]],

        [[1.7044, 1.3819, 1.6447]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 3, 3]), e shape: torch.Size([300, 1, 3])
SpMM(message func: add, reduce func: max)
______________________ test_spmm[idtype0-max-add-shp2-g0] ______________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float64)})
shp = ((1,), (3,)), msg = 'add', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43e5498b8>
op = 'add', reduce_op = 'max'
lhs_data = tensor([[1.2739],
        [1.9907],
        [1.7108],
        [1.4494],
        [1.7854],
        [1.3175],
        [1...781],
        [1.5551],
        [1.9174],
        [1.6490],
        [1.8310]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.5755, 1.8684, 1.4957],
        [1.7181, 1.8645, 1.4433],
        [1.0523, 1.2875, 1.9412],
        [1.4421,... 1.4753],
        [1.6464, 1.4678, 1.9822],
        [1.4903, 1.5819, 1.1340]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1, 3), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 1]), e shape: torch.Size([100, 3])
SpMM(message func: add, reduce func: max)
______________________ test_spmm[idtype0-max-add-shp2-g1] ______________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (3,)), msg = 'add', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72b2810d8>
op = 'add', reduce_op = 'max'
lhs_data = tensor([[1.4273],
        [1.2721],
        [1.6440],
        [1.1935],
        [1.3824],
        [1.2633],
        [1...130],
        [1.9070],
        [1.4969],
        [1.3015],
        [1.6641]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.7214, 1.7241, 1.4306],
        [1.9133, 1.6394, 1.7716],
        [1.4056, 1.3562, 1.1700],
        [1.6935,... 1.6851],
        [1.0095, 1.3532, 1.5881],
        [1.6983, 1.7902, 1.7770]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 1]), e shape: torch.Size([300, 3])
SpMM(message func: add, reduce func: max)
______________________ test_spmm[idtype0-max-add-shp3-g0] ______________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (1,)), msg = 'add', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb7058a9948>
op = 'add', reduce_op = 'max'
lhs_data = tensor([[1.3216, 1.9615, 1.3173],
        [1.8955, 1.8481, 1.8115],
        [1.6751, 1.5220, 1.7750],
        [1.7419,... 1.6787],
        [1.6633, 1.1499, 1.7859],
        [1.7906, 1.8313, 1.9635]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.8106],
        [1.3868],
        [1.1465],
        [1.0231],
        [1.3065],
        [1.1119],
        [1...812],
        [1.3021],
        [1.2004],
        [1.5025],
        [1.8443]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 3]), e shape: torch.Size([100, 1])
SpMM(message func: add, reduce func: max)
______________________ test_spmm[idtype0-max-add-shp3-g1] ______________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (1,)), msg = 'add', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb457bee288>
op = 'add', reduce_op = 'max'
lhs_data = tensor([[1.5217, 1.3594, 1.7827],
        [1.7802, 1.5787, 1.0712],
        [1.6978, 1.7020, 1.3062],
        [1.1989,... 1.2604],
        [1.0105, 1.8433, 1.9876],
        [1.6258, 1.8054, 1.5159]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.2166],
        [1.5300],
        [1.1877],
        [1.6076],
        [1.4640],
        [1.3451],
        [1...825],
        [1.2531],
        [1.5587],
        [1.4323],
        [1.9294]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 3]), e shape: torch.Size([300, 1])
SpMM(message func: add, reduce func: max)
______________________ test_spmm[idtype0-max-add-shp4-g0] ______________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), msg = 'add', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb705a4ea68>
op = 'add', reduce_op = 'max'
lhs_data = tensor([[1.7328],
        [1.0308],
        [1.0721],
        [1.2534],
        [1.2157],
        [1.9000],
        [1...953],
        [1.9884],
        [1.7190],
        [1.9095],
        [1.3094]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.9688],
        [1.4122],
        [1.7085],
        [1.0934],
        [1.6539],
        [1.5382],
        [1...892],
        [1.1639],
        [1.6475],
        [1.1266],
        [1.9194]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 1]), e shape: torch.Size([100, 1])
SpMM(message func: add, reduce func: max)
______________________ test_spmm[idtype0-max-add-shp4-g1] ______________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), msg = 'add', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43ebab990>
op = 'add', reduce_op = 'max'
lhs_data = tensor([[1.8163],
        [1.9563],
        [1.9519],
        [1.0179],
        [1.0182],
        [1.3515],
        [1...282],
        [1.3424],
        [1.7160],
        [1.2056],
        [1.4275]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.3545],
        [1.2097],
        [1.7338],
        [1.9635],
        [1.5591],
        [1.4940],
        [1...371],
        [1.2051],
        [1.1214],
        [1.3412],
        [1.4935]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 1]), e shape: torch.Size([300, 1])
SpMM(message func: add, reduce func: max)
______________________ test_spmm[idtype0-max-add-shp5-g0] ______________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
shp = ((), ()), msg = 'add', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43e792ca8>
op = 'add', reduce_op = 'max'
lhs_data = tensor([1.5412, 1.7817, 1.8495, 1.0589, 1.8504, 1.0300, 1.2370, 1.1027, 1.1399,
        1.7765, 1.8337, 1.7660, 1.1554...7215, 1.0784, 1.5627, 1.8300, 1.9660, 1.7088,
        1.4339, 1.8865, 1.3138], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([1.4495, 1.2577, 1.0542, 1.0880, 1.2047, 1.8762, 1.6219, 1.1892, 1.9954,
        1.1483, 1.5179, 1.7609, 1.3176...9331, 1.7277, 1.0170, 1.1251, 1.4778, 1.1254, 1.7598, 1.4081,
        1.8603], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30]), e shape: torch.Size([100])
SpMM(message func: add, reduce func: max)
______________________ test_spmm[idtype0-max-add-shp5-g1] ______________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((), ()), msg = 'add', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43eb43798>
op = 'add', reduce_op = 'max'
lhs_data = tensor([1.4063, 1.3830, 1.4981, 1.7725, 1.4007, 1.5434, 1.5645, 1.5404, 1.3260,
        1.4118, 1.3630, 1.1936, 1.1949...6332, 1.2509, 1.3742, 1.0020, 1.6533, 1.8552,
        1.6636, 1.3759, 1.7270], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([1.1042, 1.0551, 1.3690, 1.4470, 1.5422, 1.7476, 1.4803, 1.1073, 1.2068,
        1.3859, 1.5582, 1.8621, 1.7468...6782, 1.6427, 1.9378, 1.6333, 1.8549, 1.4233,
        1.0341, 1.0981, 1.3032], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30]), e shape: torch.Size([300])
SpMM(message func: add, reduce func: max)
______________________ test_spmm[idtype0-max-sub-shp0-g0] ______________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), msg = 'sub', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72b256558>
op = 'add', reduce_op = 'max'
lhs_data = tensor([[[[[[1.2712],
            [1.4794],
            [1.6621]]],


          [[[1.8743],
            [1.7420],
    ...8]]],


          [[[1.1658],
            [1.7488],
            [1.2261]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[-1.3494]],

           [[-1.2876]],

           [[-1.0105]]]],



         [[[[-1.0053]],

           [[-...       [[[[-1.5635]],

           [[-1.8458]],

           [[-1.4316]]]]]], dtype=torch.float64, grad_fn=<NegBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 1, 2, 1, 3, 1]), e shape: torch.Size([100, 4, 1, 3, 1, 1])
SpMM(message func: sub, reduce func: max)
______________________ test_spmm[idtype0-max-sub-shp0-g1] ______________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), msg = 'sub', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72b256480>
op = 'add', reduce_op = 'max'
lhs_data = tensor([[[[[[1.9185],
            [1.7220],
            [1.9158]]],


          [[[1.0820],
            [1.6220],
    ...2]]],


          [[[1.2471],
            [1.3585],
            [1.9873]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[-1.9148]],

           [[-1.3888]],

           [[-1.3484]]]],



         [[[[-1.5951]],

           [[-...       [[[[-1.4918]],

           [[-1.2524]],

           [[-1.3086]]]]]], dtype=torch.float64, grad_fn=<NegBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 1, 2, 1, 3, 1]), e shape: torch.Size([300, 4, 1, 3, 1, 1])
SpMM(message func: sub, reduce func: max)
______________________ test_spmm[idtype0-max-sub-shp1-g0] ______________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1, 3), dtype=torch.float64)})
shp = ((3, 3), (1, 3)), msg = 'sub', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb705a43ea0>
op = 'add', reduce_op = 'max'
lhs_data = tensor([[[1.5875, 1.6568, 1.4121],
         [1.0362, 1.4537, 1.9956],
         [1.4592, 1.6266, 1.9625]],

        [[1...0990],
         [1.4022, 1.6478, 1.0537],
         [1.5993, 1.3719, 1.0393]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[-1.0443, -1.7883, -1.8467]],

        [[-1.2268, -1.0067, -1.2243]],

        [[-1.4635, -1.4763, -1.5797]],...1.5758, -1.2245, -1.3437]],

        [[-1.0822, -1.2541, -1.7349]]], dtype=torch.float64,
       grad_fn=<NegBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 3, 3]), e shape: torch.Size([100, 1, 3])
SpMM(message func: sub, reduce func: max)
______________________ test_spmm[idtype0-max-sub-shp1-g1] ______________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3, 3), (1, 3)), msg = 'sub', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72bf932d0>
op = 'add', reduce_op = 'max'
lhs_data = tensor([[[1.7851, 1.0406, 1.5278],
         [1.1075, 1.1988, 1.9011],
         [1.9069, 1.7544, 1.1572]],

        [[1...0556],
         [1.8551, 1.3292, 1.9384],
         [1.5410, 1.5350, 1.2063]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[-1.3493, -1.3500, -1.4668]],

        [[-1.9450, -1.4350, -1.0254]],

        [[-1.9567, -1.3435, -1.2441]],...1.9397, -1.6907, -1.4869]],

        [[-1.7284, -1.4853, -1.5539]]], dtype=torch.float64,
       grad_fn=<NegBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 3, 3]), e shape: torch.Size([300, 1, 3])
SpMM(message func: sub, reduce func: max)
______________________ test_spmm[idtype0-max-sub-shp2-g0] ______________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float64)})
shp = ((1,), (3,)), msg = 'sub', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43eb29af8>
op = 'add', reduce_op = 'max'
lhs_data = tensor([[1.0144],
        [1.6239],
        [1.3819],
        [1.3991],
        [1.4652],
        [1.9368],
        [1...317],
        [1.9398],
        [1.5033],
        [1.8227],
        [1.8471]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.2182, -1.2502, -1.3634],
        [-1.0293, -1.2039, -1.9643],
        [-1.0753, -1.4201, -1.4781],
       ...  [-1.7728, -1.0712, -1.4941],
        [-1.3075, -1.2019, -1.8277]], dtype=torch.float64,
       grad_fn=<NegBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1, 3), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 1]), e shape: torch.Size([100, 3])
SpMM(message func: sub, reduce func: max)
______________________ test_spmm[idtype0-max-sub-shp2-g1] ______________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (3,)), msg = 'sub', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43eb6e3a8>
op = 'add', reduce_op = 'max'
lhs_data = tensor([[1.3372],
        [1.2482],
        [1.7977],
        [1.6185],
        [1.0351],
        [1.8196],
        [1...420],
        [1.9152],
        [1.6929],
        [1.9283],
        [1.3846]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.3921, -1.4249, -1.4012],
        [-1.5084, -1.1827, -1.8158],
        [-1.4739, -1.1013, -1.1790],
       ...  [-1.8358, -1.5271, -1.1155],
        [-1.8576, -1.3404, -1.7350]], dtype=torch.float64,
       grad_fn=<NegBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 1]), e shape: torch.Size([300, 3])
SpMM(message func: sub, reduce func: max)
______________________ test_spmm[idtype0-max-sub-shp3-g0] ______________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (1,)), msg = 'sub', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72b26bb88>
op = 'add', reduce_op = 'max'
lhs_data = tensor([[1.7150, 1.9349, 1.1802],
        [1.4885, 1.0165, 1.8286],
        [1.5390, 1.9419, 1.6723],
        [1.5401,... 1.2907],
        [1.3154, 1.7509, 1.3359],
        [1.4569, 1.0246, 1.3703]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.4252],
        [-1.1983],
        [-1.8137],
        [-1.3008],
        [-1.8251],
        [-1.9239],
    ...       [-1.7677],
        [-1.9888],
        [-1.0225],
        [-1.3400]], dtype=torch.float64, grad_fn=<NegBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 3]), e shape: torch.Size([100, 1])
SpMM(message func: sub, reduce func: max)
______________________ test_spmm[idtype0-max-sub-shp3-g1] ______________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (1,)), msg = 'sub', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72b26b1b0>
op = 'add', reduce_op = 'max'
lhs_data = tensor([[1.3990, 1.1509, 1.2573],
        [1.3798, 1.2323, 1.1532],
        [1.4858, 1.5461, 1.8151],
        [1.4374,... 1.4095],
        [1.8179, 1.2997, 1.1341],
        [1.9947, 1.4188, 1.2126]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.3901],
        [-1.2037],
        [-1.1142],
        [-1.0586],
        [-1.6723],
        [-1.4457],
    ...       [-1.6002],
        [-1.4959],
        [-1.4474],
        [-1.0033]], dtype=torch.float64, grad_fn=<NegBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 3]), e shape: torch.Size([300, 1])
SpMM(message func: sub, reduce func: max)
______________________ test_spmm[idtype0-max-sub-shp4-g0] ______________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), msg = 'sub', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43e6135e8>
op = 'add', reduce_op = 'max'
lhs_data = tensor([[1.9320],
        [1.0766],
        [1.2509],
        [1.9864],
        [1.7195],
        [1.2525],
        [1...033],
        [1.9184],
        [1.9789],
        [1.4373],
        [1.6461]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.9965],
        [-1.5284],
        [-1.2159],
        [-1.5186],
        [-1.6239],
        [-1.0104],
    ...       [-1.1538],
        [-1.3720],
        [-1.8042],
        [-1.9574]], dtype=torch.float64, grad_fn=<NegBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 1]), e shape: torch.Size([100, 1])
SpMM(message func: sub, reduce func: max)
______________________ test_spmm[idtype0-max-sub-shp4-g1] ______________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), msg = 'sub', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb705aa02d0>
op = 'add', reduce_op = 'max'
lhs_data = tensor([[1.7155],
        [1.4480],
        [1.7019],
        [1.6417],
        [1.6891],
        [1.6751],
        [1...633],
        [1.8876],
        [1.8737],
        [1.3142],
        [1.8154]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.1052],
        [-1.9650],
        [-1.4293],
        [-1.6492],
        [-1.0204],
        [-1.1236],
    ...       [-1.4364],
        [-1.1503],
        [-1.7635],
        [-1.7693]], dtype=torch.float64, grad_fn=<NegBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 1]), e shape: torch.Size([300, 1])
SpMM(message func: sub, reduce func: max)
______________________ test_spmm[idtype0-max-sub-shp5-g0] ______________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
shp = ((), ()), msg = 'sub', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72b24e090>
op = 'add', reduce_op = 'max'
lhs_data = tensor([1.5175, 1.1506, 1.0445, 1.1496, 1.5334, 1.6856, 1.4515, 1.2659, 1.7874,
        1.1435, 1.3699, 1.4695, 1.0419...8641, 1.6705, 1.0787, 1.3841, 1.2191, 1.9608,
        1.6233, 1.8435, 1.6743], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([-1.7569, -1.4238, -1.7704, -1.3059, -1.9850, -1.1502, -1.2466, -1.7237,
        -1.1774, -1.3277, -1.4171, -1....8511, -1.6954, -1.4308,
        -1.5712, -1.3378, -1.7359, -1.7962], dtype=torch.float64,
       grad_fn=<NegBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30]), e shape: torch.Size([100])
SpMM(message func: sub, reduce func: max)
______________________ test_spmm[idtype0-max-sub-shp5-g1] ______________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((), ()), msg = 'sub', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb7058f3b40>
op = 'add', reduce_op = 'max'
lhs_data = tensor([1.5633, 1.7182, 1.2122, 1.4843, 1.6634, 1.1849, 1.4289, 1.8970, 1.3097,
        1.6854, 1.3761, 1.7545, 1.4253...8295, 1.2332, 1.0882, 1.2984, 1.3948, 1.7682,
        1.7613, 1.1876, 1.4397], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([-1.4771, -1.3277, -1.0077, -1.4428, -1.5778, -1.9583, -1.9517, -1.5341,
        -1.3671, -1.6054, -1.2842, -1....7387, -1.1167, -1.5986,
        -1.6349, -1.4961, -1.0087, -1.4304], dtype=torch.float64,
       grad_fn=<NegBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30]), e shape: torch.Size([300])
SpMM(message func: sub, reduce func: max)
______________________ test_spmm[idtype0-max-mul-shp0-g0] ______________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), msg = 'mul', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72b282d80>
op = 'mul', reduce_op = 'max'
lhs_data = tensor([[[[[[1.3384],
            [1.3777],
            [1.8374]]],


          [[[1.7099],
            [1.5344],
    ...3]]],


          [[[1.1507],
            [1.7018],
            [1.6464]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.5020]],

           [[1.2979]],

           [[1.3698]]]],



         [[[[1.1362]],

           [[1.281...



         [[[[1.9521]],

           [[1.3032]],

           [[1.6398]]]]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 1, 2, 1, 3, 1]), e shape: torch.Size([100, 4, 1, 3, 1, 1])
SpMM(message func: mul, reduce func: max)
______________________ test_spmm[idtype0-max-mul-shp0-g1] ______________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), msg = 'mul', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72b282870>
op = 'mul', reduce_op = 'max'
lhs_data = tensor([[[[[[1.3364],
            [1.1958],
            [1.3392]]],


          [[[1.3041],
            [1.5241],
    ...7]]],


          [[[1.8621],
            [1.8137],
            [1.7288]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.9704]],

           [[1.1453]],

           [[1.2892]]]],



         [[[[1.2404]],

           [[1.060...



         [[[[1.4362]],

           [[1.4576]],

           [[1.2926]]]]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 1, 2, 1, 3, 1]), e shape: torch.Size([300, 4, 1, 3, 1, 1])
SpMM(message func: mul, reduce func: max)
______________________ test_spmm[idtype0-max-mul-shp1-g0] ______________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1, 3), dtype=torch.float64)})
shp = ((3, 3), (1, 3)), msg = 'mul', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72b282708>
op = 'mul', reduce_op = 'max'
lhs_data = tensor([[[1.4843, 1.0222, 1.4591],
         [1.4326, 1.6433, 1.4465],
         [1.8117, 1.6602, 1.5397]],

        [[1...1394],
         [1.3471, 1.8673, 1.5068],
         [1.8070, 1.6199, 1.1492]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[1.7921, 1.4512, 1.7257]],

        [[1.1861, 1.0018, 1.7530]],

        [[1.7835, 1.2899, 1.6118]],

       ...]],

        [[1.3556, 1.8269, 1.0807]],

        [[1.5538, 1.4740, 1.8244]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 3, 3]), e shape: torch.Size([100, 1, 3])
SpMM(message func: mul, reduce func: max)
______________________ test_spmm[idtype0-max-mul-shp1-g1] ______________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3, 3), (1, 3)), msg = 'mul', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72b2829d8>
op = 'mul', reduce_op = 'max'
lhs_data = tensor([[[1.0052, 1.2981, 1.9887],
         [1.9113, 1.8018, 1.4297],
         [1.4276, 1.2351, 1.1197]],

        [[1...8075],
         [1.4351, 1.7341, 1.2152],
         [1.4994, 1.9150, 1.5186]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[1.9542, 1.6503, 1.7380]],

        [[1.4694, 1.5248, 1.3434]],

        [[1.3053, 1.4629, 1.7323]],

       ...]],

        [[1.7833, 1.3262, 1.7846]],

        [[1.0873, 1.8522, 1.3156]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 3, 3]), e shape: torch.Size([300, 1, 3])
SpMM(message func: mul, reduce func: max)
______________________ test_spmm[idtype0-max-mul-shp2-g0] ______________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float64)})
shp = ((1,), (3,)), msg = 'mul', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72ae7d4c8>
op = 'mul', reduce_op = 'max'
lhs_data = tensor([[1.5116],
        [1.0514],
        [1.0876],
        [1.0077],
        [1.5933],
        [1.1446],
        [1...406],
        [1.1226],
        [1.2962],
        [1.9754],
        [1.0542]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.7558, 1.0754, 1.2241],
        [1.2976, 1.6328, 1.7057],
        [1.0352, 1.7697, 1.5591],
        [1.5782,... 1.1229],
        [1.9557, 1.5998, 1.2629],
        [1.0345, 1.1396, 1.8267]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1, 3), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 1]), e shape: torch.Size([100, 3])
SpMM(message func: mul, reduce func: max)
______________________ test_spmm[idtype0-max-mul-shp2-g1] ______________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (3,)), msg = 'mul', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43e8e5318>
op = 'mul', reduce_op = 'max'
lhs_data = tensor([[1.5867],
        [1.5849],
        [1.2086],
        [1.7655],
        [1.9723],
        [1.6365],
        [1...481],
        [1.3336],
        [1.3153],
        [1.6916],
        [1.0614]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.8403, 1.8126, 1.8392],
        [1.8783, 1.6265, 1.6571],
        [1.7354, 1.1338, 1.5931],
        [1.6463,... 1.8540],
        [1.1153, 1.8574, 1.1779],
        [1.2257, 1.5403, 1.9887]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 1]), e shape: torch.Size([300, 3])
SpMM(message func: mul, reduce func: max)
______________________ test_spmm[idtype0-max-mul-shp3-g0] ______________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (1,)), msg = 'mul', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72b2430d8>
op = 'mul', reduce_op = 'max'
lhs_data = tensor([[1.1615, 1.2613, 1.6297],
        [1.0523, 1.6618, 1.5543],
        [1.7560, 1.3979, 1.7666],
        [1.9303,... 1.0475],
        [1.7448, 1.4306, 1.3983],
        [1.6332, 1.8569, 1.9249]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.2183],
        [1.6383],
        [1.9280],
        [1.8380],
        [1.5889],
        [1.9209],
        [1...020],
        [1.5064],
        [1.3640],
        [1.2880],
        [1.0348]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 3]), e shape: torch.Size([100, 1])
SpMM(message func: mul, reduce func: max)
______________________ test_spmm[idtype0-max-mul-shp3-g1] ______________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (1,)), msg = 'mul', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72b2433f0>
op = 'mul', reduce_op = 'max'
lhs_data = tensor([[1.0310, 1.1184, 1.6185],
        [1.1680, 1.0299, 1.9023],
        [1.8661, 1.2108, 1.3288],
        [1.6900,... 1.4794],
        [1.6772, 1.6250, 1.2951],
        [1.0197, 1.7047, 1.6659]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.4846],
        [1.3366],
        [1.6799],
        [1.2955],
        [1.1397],
        [1.0478],
        [1...442],
        [1.6722],
        [1.3674],
        [1.4672],
        [1.8525]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 3]), e shape: torch.Size([300, 1])
SpMM(message func: mul, reduce func: max)
______________________ test_spmm[idtype0-max-mul-shp4-g0] ______________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), msg = 'mul', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb61439faf8>
op = 'mul', reduce_op = 'max'
lhs_data = tensor([[1.6600],
        [1.1677],
        [1.4272],
        [1.7254],
        [1.0293],
        [1.1737],
        [1...033],
        [1.0519],
        [1.8423],
        [1.6517],
        [1.3495]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.2099],
        [1.5100],
        [1.2586],
        [1.0733],
        [1.9432],
        [1.0222],
        [1...260],
        [1.1665],
        [1.6293],
        [1.1930],
        [1.3240]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 1]), e shape: torch.Size([100, 1])
SpMM(message func: mul, reduce func: max)
______________________ test_spmm[idtype0-max-mul-shp4-g1] ______________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), msg = 'mul', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72b1b8b88>
op = 'mul', reduce_op = 'max'
lhs_data = tensor([[1.6858],
        [1.6293],
        [1.6561],
        [1.9367],
        [1.4207],
        [1.1865],
        [1...594],
        [1.9042],
        [1.5079],
        [1.4289],
        [1.4490]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.6194],
        [1.8559],
        [1.5311],
        [1.8276],
        [1.6283],
        [1.2096],
        [1...249],
        [1.6587],
        [1.4228],
        [1.8076],
        [1.4938]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 1]), e shape: torch.Size([300, 1])
SpMM(message func: mul, reduce func: max)
______________________ test_spmm[idtype0-max-mul-shp5-g0] ______________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
shp = ((), ()), msg = 'mul', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb7059d5ea0>
op = 'mul', reduce_op = 'max'
lhs_data = tensor([1.7390, 1.8184, 1.0022, 1.7023, 1.9772, 1.5539, 1.5687, 1.1383, 1.6269,
        1.8598, 1.5472, 1.4398, 1.2368...9801, 1.4493, 1.8109, 1.7794, 1.1311, 1.3044,
        1.1038, 1.2230, 1.7086], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([1.0173, 1.7250, 1.9276, 1.1970, 1.0850, 1.1061, 1.5007, 1.1779, 1.2405,
        1.1464, 1.9379, 1.7185, 1.2322...9375, 1.9161, 1.7018, 1.2650, 1.3835, 1.2762, 1.7100, 1.3665,
        1.8027], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30]), e shape: torch.Size([100])
SpMM(message func: mul, reduce func: max)
______________________ test_spmm[idtype0-max-mul-shp5-g1] ______________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((), ()), msg = 'mul', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72ae7d438>
op = 'mul', reduce_op = 'max'
lhs_data = tensor([1.7283, 1.1122, 1.7515, 1.5091, 1.0567, 1.8180, 1.6410, 1.6929, 1.4745,
        1.9529, 1.6848, 1.4201, 1.2018...9104, 1.4461, 1.0315, 1.2788, 1.9321, 1.3439,
        1.1456, 1.2659, 1.5925], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([1.8752, 1.7782, 1.3396, 1.9926, 1.2828, 1.6160, 1.7757, 1.2260, 1.9685,
        1.6916, 1.7629, 1.0831, 1.4620...1576, 1.3390, 1.0072, 1.2470, 1.0805, 1.3343,
        1.5332, 1.9645, 1.8619], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30]), e shape: torch.Size([300])
SpMM(message func: mul, reduce func: max)
______________________ test_spmm[idtype0-max-div-shp0-g0] ______________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), msg = 'div', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72b24a1f8>
op = 'mul', reduce_op = 'max'
lhs_data = tensor([[[[[[1.2471],
            [1.8453],
            [1.2893]]],


          [[[1.9798],
            [1.5650],
    ...9]]],


          [[[1.7688],
            [1.5081],
            [1.6749]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[0.7047]],

           [[0.6783]],

           [[0.8943]]]],



         [[[[0.7110]],

           [[0.590...         [[[[0.5352]],

           [[0.5795]],

           [[0.8714]]]]]], dtype=torch.float64, grad_fn=<MulBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 1, 2, 1, 3, 1]), e shape: torch.Size([100, 4, 1, 3, 1, 1])
SpMM(message func: div, reduce func: max)
______________________ test_spmm[idtype0-max-div-shp0-g1] ______________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), msg = 'div', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72b24a7e0>
op = 'mul', reduce_op = 'max'
lhs_data = tensor([[[[[[1.7685],
            [1.4749],
            [1.7441]]],


          [[[1.9962],
            [1.4189],
    ...6]]],


          [[[1.6869],
            [1.6635],
            [1.7210]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[0.7932]],

           [[0.6988]],

           [[0.9588]]]],



         [[[[0.5611]],

           [[0.630...         [[[[0.5166]],

           [[0.7604]],

           [[0.8788]]]]]], dtype=torch.float64, grad_fn=<MulBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 1, 2, 1, 3, 1]), e shape: torch.Size([300, 4, 1, 3, 1, 1])
SpMM(message func: div, reduce func: max)
______________________ test_spmm[idtype0-max-div-shp1-g0] ______________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1, 3), dtype=torch.float64)})
shp = ((3, 3), (1, 3)), msg = 'div', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72a92f240>
op = 'mul', reduce_op = 'max'
lhs_data = tensor([[[1.6071, 1.6815, 1.1389],
         [1.8109, 1.1794, 1.0046],
         [1.3537, 1.1834, 1.2734]],

        [[1...8599],
         [1.6369, 1.8951, 1.5758],
         [1.8754, 1.9009, 1.4676]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[0.5794, 0.6545, 0.6904]],

        [[0.5717, 0.6504, 0.5553]],

        [[0.7135, 0.5107, 0.5211]],

       ...  [[0.6638, 0.6435, 0.8232]],

        [[0.5585, 0.5649, 0.7271]]], dtype=torch.float64,
       grad_fn=<MulBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 3, 3]), e shape: torch.Size([100, 1, 3])
SpMM(message func: div, reduce func: max)
______________________ test_spmm[idtype0-max-div-shp1-g1] ______________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3, 3), (1, 3)), msg = 'div', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72a9a0ea0>
op = 'mul', reduce_op = 'max'
lhs_data = tensor([[[1.1846, 1.0133, 1.0273],
         [1.4928, 1.0909, 1.4066],
         [1.8139, 1.4269, 1.9989]],

        [[1...9758],
         [1.0754, 1.6533, 1.7957],
         [1.3055, 1.7279, 1.2886]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[0.5623, 0.5084, 0.7150]],

        [[0.6253, 0.5528, 0.5756]],

        [[0.5618, 0.7331, 0.7193]],

       ...  [[0.9243, 0.9479, 0.9956]],

        [[0.7292, 0.6753, 0.5750]]], dtype=torch.float64,
       grad_fn=<MulBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 3, 3]), e shape: torch.Size([300, 1, 3])
SpMM(message func: div, reduce func: max)
______________________ test_spmm[idtype0-max-div-shp2-g0] ______________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float64)})
shp = ((1,), (3,)), msg = 'div', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72c1e0d38>
op = 'mul', reduce_op = 'max'
lhs_data = tensor([[1.6045],
        [1.5783],
        [1.3732],
        [1.9559],
        [1.2235],
        [1.1237],
        [1...435],
        [1.4190],
        [1.1845],
        [1.4162],
        [1.7927]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.6328, 0.5482, 0.6232],
        [0.5452, 0.5364, 0.6508],
        [0.6497, 0.6103, 0.6070],
        [0.7770,...829],
        [0.5753, 0.5214, 0.5092],
        [0.7536, 0.5032, 0.7291]], dtype=torch.float64, grad_fn=<MulBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1, 3), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 1]), e shape: torch.Size([100, 3])
SpMM(message func: div, reduce func: max)
______________________ test_spmm[idtype0-max-div-shp2-g1] ______________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (3,)), msg = 'div', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb7295a8438>
op = 'mul', reduce_op = 'max'
lhs_data = tensor([[1.1873],
        [1.4048],
        [1.4414],
        [1.8782],
        [1.0316],
        [1.2334],
        [1...510],
        [1.4583],
        [1.0353],
        [1.9617],
        [1.2606]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.6839, 0.6237, 0.9912],
        [0.6308, 0.7816, 0.5106],
        [0.5754, 0.5731, 0.5451],
        [0.6091,...056],
        [0.9096, 0.5807, 0.5265],
        [0.7957, 0.6191, 0.6110]], dtype=torch.float64, grad_fn=<MulBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 1]), e shape: torch.Size([300, 3])
SpMM(message func: div, reduce func: max)
______________________ test_spmm[idtype0-max-div-shp3-g0] ______________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (1,)), msg = 'div', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb45517aa68>
op = 'mul', reduce_op = 'max'
lhs_data = tensor([[1.8004, 1.6731, 1.8183],
        [1.1278, 1.8581, 1.6110],
        [1.3449, 1.0106, 1.6240],
        [1.1257,... 1.4590],
        [1.8462, 1.2020, 1.5627],
        [1.8040, 1.9130, 1.0948]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.5750],
        [0.7785],
        [0.6971],
        [0.5087],
        [0.6220],
        [0.7962],
        [0...,
        [0.5701],
        [0.5427],
        [0.7714],
        [0.7827]], dtype=torch.float64, grad_fn=<MulBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 3]), e shape: torch.Size([100, 1])
SpMM(message func: div, reduce func: max)
______________________ test_spmm[idtype0-max-div-shp3-g1] ______________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (1,)), msg = 'div', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72b25e288>
op = 'mul', reduce_op = 'max'
lhs_data = tensor([[1.7848, 1.6837, 1.3920],
        [1.8869, 1.2801, 1.7656],
        [1.5420, 1.0244, 1.6464],
        [1.8012,... 1.9104],
        [1.4313, 1.3464, 1.7096],
        [1.9819, 1.0388, 1.9589]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.9811],
        [0.5560],
        [0.5818],
        [0.6023],
        [0.5144],
        [0.5293],
        [0...,
        [0.5630],
        [0.5864],
        [0.8666],
        [0.8229]], dtype=torch.float64, grad_fn=<MulBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 3]), e shape: torch.Size([300, 1])
SpMM(message func: div, reduce func: max)
______________________ test_spmm[idtype0-max-div-shp4-g0] ______________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), msg = 'div', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb7059aedc8>
op = 'mul', reduce_op = 'max'
lhs_data = tensor([[1.1585],
        [1.3826],
        [1.6215],
        [1.7210],
        [1.2405],
        [1.6052],
        [1...861],
        [1.1894],
        [1.2331],
        [1.0870],
        [1.5253]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.9786],
        [0.6985],
        [0.5122],
        [0.6725],
        [0.7168],
        [0.9164],
        [0...,
        [0.5238],
        [0.7112],
        [0.6212],
        [0.8422]], dtype=torch.float64, grad_fn=<MulBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 1]), e shape: torch.Size([100, 1])
SpMM(message func: div, reduce func: max)
______________________ test_spmm[idtype0-max-div-shp4-g1] ______________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), msg = 'div', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb705aa2900>
op = 'mul', reduce_op = 'max'
lhs_data = tensor([[1.6571],
        [1.8944],
        [1.6842],
        [1.1502],
        [1.0124],
        [1.7699],
        [1...949],
        [1.9638],
        [1.5713],
        [1.3112],
        [1.9936]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.5024],
        [0.6754],
        [0.5492],
        [0.7338],
        [0.9664],
        [0.8668],
        [0...,
        [0.7718],
        [0.9472],
        [0.7042],
        [0.7292]], dtype=torch.float64, grad_fn=<MulBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 1]), e shape: torch.Size([300, 1])
SpMM(message func: div, reduce func: max)
______________________ test_spmm[idtype0-max-div-shp5-g0] ______________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
shp = ((), ()), msg = 'div', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72958f828>
op = 'mul', reduce_op = 'max'
lhs_data = tensor([1.5332, 1.5201, 1.4454, 1.1001, 1.2318, 1.5097, 1.0765, 1.2196, 1.6092,
        1.7288, 1.9291, 1.5800, 1.7816...7797, 1.2764, 1.3129, 1.6692, 1.1337, 1.2378,
        1.3687, 1.4140, 1.4611], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([0.9128, 0.9112, 0.8612, 0.5586, 0.8351, 0.5495, 0.9017, 0.5980, 0.5166,
        0.7106, 0.9464, 0.7328, 0.6230..., 0.8882, 0.5538, 0.6030, 0.5326, 0.7440, 0.6979, 0.6456,
        0.5965], dtype=torch.float64, grad_fn=<MulBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30]), e shape: torch.Size([100])
SpMM(message func: div, reduce func: max)
______________________ test_spmm[idtype0-max-div-shp5-g1] ______________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((), ()), msg = 'div', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb455172f30>
op = 'mul', reduce_op = 'max'
lhs_data = tensor([1.1185, 1.4804, 1.0801, 1.2872, 1.8984, 1.2138, 1.7913, 1.9121, 1.1607,
        1.3808, 1.8768, 1.0997, 1.9325...5409, 1.7809, 1.2608, 1.3507, 1.8783, 1.0274,
        1.5886, 1.3949, 1.3027], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([0.7827, 0.5207, 0.6622, 0.6201, 0.5258, 0.8746, 0.5414, 0.5161, 0.5927,
        0.6058, 0.7551, 0.5707, 0.6253..., 0.5441, 0.8333, 0.7913, 0.9820, 0.5898,
        0.7832, 0.6423, 0.7200], dtype=torch.float64, grad_fn=<MulBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30]), e shape: torch.Size([300])
SpMM(message func: div, reduce func: max)
___________________ test_spmm[idtype0-max-copy_lhs-shp0-g0] ____________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), msg = 'copy_lhs', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb7059ebbd0>
op = 'copy_lhs', reduce_op = 'max'
lhs_data = tensor([[[[[[1.8557],
            [1.3782],
            [1.6370]]],


          [[[1.2109],
            [1.6080],
    ...0]]],


          [[[1.0707],
            [1.0542],
            [1.1916]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.2547]],

           [[1.6391]],

           [[1.3923]]]],



         [[[[1.7800]],

           [[1.592...



         [[[[1.1226]],

           [[1.7332]],

           [[1.2772]]]]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 1, 2, 1, 3, 1]), e shape: torch.Size([100, 4, 1, 3, 1, 1])
SpMM(message func: copy_lhs, reduce func: max)
___________________ test_spmm[idtype0-max-copy_lhs-shp0-g1] ____________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), msg = 'copy_lhs', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72b18b1b0>
op = 'copy_lhs', reduce_op = 'max'
lhs_data = tensor([[[[[[1.9290],
            [1.7372],
            [1.2465]]],


          [[[1.0420],
            [1.1243],
    ...4]]],


          [[[1.3352],
            [1.2140],
            [1.5078]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.2052]],

           [[1.2812]],

           [[1.1959]]]],



         [[[[1.3888]],

           [[1.679...



         [[[[1.1693]],

           [[1.6509]],

           [[1.3873]]]]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 1, 2, 1, 3, 1]), e shape: torch.Size([300, 4, 1, 3, 1, 1])
SpMM(message func: copy_lhs, reduce func: max)
___________________ test_spmm[idtype0-max-copy_lhs-shp1-g0] ____________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1, 3), dtype=torch.float64)})
shp = ((3, 3), (1, 3)), msg = 'copy_lhs', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb61dcfc5a0>
op = 'copy_lhs', reduce_op = 'max'
lhs_data = tensor([[[1.9586, 1.4001, 1.0363],
         [1.3538, 1.0634, 1.2170],
         [1.6339, 1.3949, 1.8994]],

        [[1...7975],
         [1.6327, 1.3844, 1.2411],
         [1.0113, 1.1761, 1.7143]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[1.7706, 1.4119, 1.0584]],

        [[1.2617, 1.6390, 1.4696]],

        [[1.1904, 1.8668, 1.8117]],

       ...]],

        [[1.1825, 1.4651, 1.3179]],

        [[1.5594, 1.0765, 1.3467]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 3, 3]), e shape: torch.Size([100, 1, 3])
SpMM(message func: copy_lhs, reduce func: max)
___________________ test_spmm[idtype0-max-copy_lhs-shp1-g1] ____________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3, 3), (1, 3)), msg = 'copy_lhs', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43e8e5f78>
op = 'copy_lhs', reduce_op = 'max'
lhs_data = tensor([[[1.3552, 1.5302, 1.3097],
         [1.1976, 1.1386, 1.2172],
         [1.2556, 1.3798, 1.8829]],

        [[1...3458],
         [1.3066, 1.3085, 1.8999],
         [1.8419, 1.4895, 1.4541]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[1.7019, 1.8894, 1.0857]],

        [[1.1004, 1.4132, 1.6321]],

        [[1.8595, 1.2493, 1.9058]],

       ...]],

        [[1.5726, 1.9769, 1.8953]],

        [[1.5859, 1.6568, 1.2652]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 3, 3]), e shape: torch.Size([300, 1, 3])
SpMM(message func: copy_lhs, reduce func: max)
___________________ test_spmm[idtype0-max-copy_lhs-shp2-g0] ____________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float64)})
shp = ((1,), (3,)), msg = 'copy_lhs', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb7059d7e58>
op = 'copy_lhs', reduce_op = 'max'
lhs_data = tensor([[1.0172],
        [1.2563],
        [1.2927],
        [1.6495],
        [1.2118],
        [1.4789],
        [1...797],
        [1.0060],
        [1.9540],
        [1.5394],
        [1.0384]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.3144, 1.4085, 1.6472],
        [1.1967, 1.7414, 1.1968],
        [1.4629, 1.0177, 1.7287],
        [1.9500,... 1.8202],
        [1.5742, 1.2605, 1.5702],
        [1.8549, 1.7707, 1.1461]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1, 3), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 1]), e shape: torch.Size([100, 3])
SpMM(message func: copy_lhs, reduce func: max)
___________________ test_spmm[idtype0-max-copy_lhs-shp2-g1] ____________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (3,)), msg = 'copy_lhs', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72b1911f8>
op = 'copy_lhs', reduce_op = 'max'
lhs_data = tensor([[1.7629],
        [1.0822],
        [1.4096],
        [1.9348],
        [1.7594],
        [1.4220],
        [1...941],
        [1.7248],
        [1.3497],
        [1.0425],
        [1.6349]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.2233, 1.8076, 1.8601],
        [1.9226, 1.9567, 1.3510],
        [1.5600, 1.8173, 1.3451],
        [1.9922,... 1.2845],
        [1.5284, 1.6864, 1.9418],
        [1.4130, 1.2405, 1.9699]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 1]), e shape: torch.Size([300, 3])
SpMM(message func: copy_lhs, reduce func: max)
___________________ test_spmm[idtype0-max-copy_lhs-shp3-g0] ____________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (1,)), msg = 'copy_lhs', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43e719fc0>
op = 'copy_lhs', reduce_op = 'max'
lhs_data = tensor([[1.6936, 1.4210, 1.5262],
        [1.5869, 1.0020, 1.4415],
        [1.6106, 1.3497, 1.7641],
        [1.7250,... 1.8052],
        [1.7457, 1.5230, 1.6647],
        [1.3196, 1.1171, 1.7515]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.9735],
        [1.1846],
        [1.7524],
        [1.2910],
        [1.5107],
        [1.2062],
        [1...813],
        [1.7986],
        [1.0666],
        [1.3502],
        [1.8126]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 3]), e shape: torch.Size([100, 1])
SpMM(message func: copy_lhs, reduce func: max)
___________________ test_spmm[idtype0-max-copy_lhs-shp3-g1] ____________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (1,)), msg = 'copy_lhs', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb45403e948>
op = 'copy_lhs', reduce_op = 'max'
lhs_data = tensor([[1.2447, 1.3793, 1.9509],
        [1.1258, 1.3346, 1.6841],
        [1.7169, 1.2138, 1.4740],
        [1.2883,... 1.0676],
        [1.1325, 1.6658, 1.8523],
        [1.5435, 1.2352, 1.6464]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.9388],
        [1.1394],
        [1.8445],
        [1.3334],
        [1.7890],
        [1.3686],
        [1...961],
        [1.2018],
        [1.6246],
        [1.9792],
        [1.4081]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 3]), e shape: torch.Size([300, 1])
SpMM(message func: copy_lhs, reduce func: max)
___________________ test_spmm[idtype0-max-copy_lhs-shp4-g0] ____________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), msg = 'copy_lhs', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72b1901b0>
op = 'copy_lhs', reduce_op = 'max'
lhs_data = tensor([[1.9547],
        [1.1594],
        [1.0984],
        [1.2676],
        [1.2246],
        [1.3607],
        [1...997],
        [1.7647],
        [1.1706],
        [1.7373],
        [1.6103]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.7115],
        [1.3302],
        [1.9348],
        [1.2353],
        [1.6989],
        [1.5143],
        [1...618],
        [1.3858],
        [1.4908],
        [1.9168],
        [1.3557]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 1]), e shape: torch.Size([100, 1])
SpMM(message func: copy_lhs, reduce func: max)
___________________ test_spmm[idtype0-max-copy_lhs-shp4-g1] ____________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), msg = 'copy_lhs', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72b1903f0>
op = 'copy_lhs', reduce_op = 'max'
lhs_data = tensor([[1.8550],
        [1.9131],
        [1.2888],
        [1.7848],
        [1.6058],
        [1.1588],
        [1...925],
        [1.2656],
        [1.0036],
        [1.2322],
        [1.6880]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.7658],
        [1.3081],
        [1.1305],
        [1.3972],
        [1.6224],
        [1.2001],
        [1...569],
        [1.6520],
        [1.3972],
        [1.0894],
        [1.2386]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 1]), e shape: torch.Size([300, 1])
SpMM(message func: copy_lhs, reduce func: max)
___________________ test_spmm[idtype0-max-copy_lhs-shp5-g0] ____________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
shp = ((), ()), msg = 'copy_lhs', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72c0787e0>
op = 'copy_lhs', reduce_op = 'max'
lhs_data = tensor([1.6549, 1.0410, 1.1282, 1.1575, 1.5230, 1.4727, 1.6556, 1.5115, 1.9761,
        1.2806, 1.4091, 1.4066, 1.8613...3601, 1.6245, 1.9686, 1.0859, 1.9055, 1.8670,
        1.1565, 1.2928, 1.2731], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([1.0396, 1.7046, 1.4045, 1.7233, 1.3256, 1.1508, 1.9587, 1.5670, 1.7239,
        1.4728, 1.2002, 1.7366, 1.4047...5571, 1.3806, 1.9014, 1.0308, 1.9779, 1.3547, 1.6706, 1.8434,
        1.0029], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30]), e shape: torch.Size([100])
SpMM(message func: copy_lhs, reduce func: max)
___________________ test_spmm[idtype0-max-copy_lhs-shp5-g1] ____________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((), ()), msg = 'copy_lhs', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43e792120>
op = 'copy_lhs', reduce_op = 'max'
lhs_data = tensor([1.5548, 1.8828, 1.3486, 1.4642, 1.7283, 1.3010, 1.3972, 1.4866, 1.5815,
        1.8775, 1.2983, 1.0779, 1.5725...7269, 1.2946, 1.0091, 1.6211, 1.5468, 1.2387,
        1.1386, 1.0587, 1.2180], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([1.2013, 1.0485, 1.2032, 1.4787, 1.4942, 1.6647, 1.7072, 1.1744, 1.9542,
        1.4039, 1.1993, 1.1829, 1.2555...2109, 1.5066, 1.1039, 1.5587, 1.4618, 1.5827,
        1.2000, 1.0134, 1.4513], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30]), e shape: torch.Size([300])
SpMM(message func: copy_lhs, reduce func: max)
___________________ test_spmm[idtype0-max-copy_rhs-shp0-g0] ____________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), msg = 'copy_rhs', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb45402afc0>
op = 'copy_rhs', reduce_op = 'max'
lhs_data = tensor([[[[[[1.7748],
            [1.4194],
            [1.6031]]],


          [[[1.6095],
            [1.8141],
    ...2]]],


          [[[1.1950],
            [1.5572],
            [1.8135]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.1939]],

           [[1.3210]],

           [[1.4601]]]],



         [[[[1.9744]],

           [[1.287...



         [[[[1.8486]],

           [[1.9241]],

           [[1.5460]]]]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 1, 2, 1, 3, 1]), e shape: torch.Size([100, 4, 1, 3, 1, 1])
SpMM(message func: copy_rhs, reduce func: max)
___________________ test_spmm[idtype0-max-copy_rhs-shp0-g1] ____________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), msg = 'copy_rhs', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb7058fe798>
op = 'copy_rhs', reduce_op = 'max'
lhs_data = tensor([[[[[[1.2859],
            [1.0838],
            [1.5940]]],


          [[[1.4581],
            [1.2119],
    ...5]]],


          [[[1.7423],
            [1.6716],
            [1.5827]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.4198]],

           [[1.3369]],

           [[1.5251]]]],



         [[[[1.0402]],

           [[1.854...



         [[[[1.7444]],

           [[1.9230]],

           [[1.2698]]]]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 1, 2, 1, 3, 1]), e shape: torch.Size([300, 4, 1, 3, 1, 1])
SpMM(message func: copy_rhs, reduce func: max)
___________________ test_spmm[idtype0-max-copy_rhs-shp1-g0] ____________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1, 3), dtype=torch.float64)})
shp = ((3, 3), (1, 3)), msg = 'copy_rhs', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43ea21b88>
op = 'copy_rhs', reduce_op = 'max'
lhs_data = tensor([[[1.3714, 1.4014, 1.9761],
         [1.9993, 1.5537, 1.6914],
         [1.0573, 1.3395, 1.7001]],

        [[1...8216],
         [1.0425, 1.6170, 1.4352],
         [1.4921, 1.6984, 1.6449]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[1.8329, 1.8482, 1.0133]],

        [[1.4595, 1.6175, 1.4524]],

        [[1.0651, 1.9387, 1.2262]],

       ...]],

        [[1.9083, 1.2817, 1.7250]],

        [[1.0961, 1.2242, 1.3015]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 3, 3]), e shape: torch.Size([100, 1, 3])
SpMM(message func: copy_rhs, reduce func: max)
___________________ test_spmm[idtype0-max-copy_rhs-shp1-g1] ____________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3, 3), (1, 3)), msg = 'copy_rhs', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72b1b01b0>
op = 'copy_rhs', reduce_op = 'max'
lhs_data = tensor([[[1.2709, 1.5212, 1.7502],
         [1.0007, 1.1331, 1.8916],
         [1.4736, 1.8831, 1.4634]],

        [[1...5446],
         [1.7907, 1.5652, 1.3110],
         [1.2057, 1.7891, 1.5762]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[1.1248, 1.1289, 1.1253]],

        [[1.8959, 1.6283, 1.4443]],

        [[1.5553, 1.0877, 1.5808]],

       ...]],

        [[1.9308, 1.7482, 1.6999]],

        [[1.2333, 1.2498, 1.4930]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 3, 3]), e shape: torch.Size([300, 1, 3])
SpMM(message func: copy_rhs, reduce func: max)
___________________ test_spmm[idtype0-max-copy_rhs-shp2-g0] ____________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float64)})
shp = ((1,), (3,)), msg = 'copy_rhs', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb732ce49d8>
op = 'copy_rhs', reduce_op = 'max'
lhs_data = tensor([[1.2867],
        [1.6235],
        [1.5133],
        [1.6254],
        [1.9541],
        [1.3302],
        [1...589],
        [1.7100],
        [1.0595],
        [1.5173],
        [1.8566]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.4095, 1.6785, 1.4640],
        [1.5784, 1.4875, 1.7035],
        [1.7531, 1.5251, 1.6965],
        [1.1804,... 1.2584],
        [1.8459, 1.0736, 1.7982],
        [1.7326, 1.5845, 1.0106]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1, 3), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 1]), e shape: torch.Size([100, 3])
SpMM(message func: copy_rhs, reduce func: max)
___________________ test_spmm[idtype0-max-copy_rhs-shp2-g1] ____________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (3,)), msg = 'copy_rhs', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43e513a68>
op = 'copy_rhs', reduce_op = 'max'
lhs_data = tensor([[1.3267],
        [1.3109],
        [1.4414],
        [1.4565],
        [1.3597],
        [1.5789],
        [1...649],
        [1.9799],
        [1.6551],
        [1.9962],
        [1.7384]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.3601, 1.3408, 1.7471],
        [1.3631, 1.9631, 1.7610],
        [1.0283, 1.4686, 1.7756],
        [1.2092,... 1.8092],
        [1.9578, 1.4974, 1.9805],
        [1.7869, 1.3296, 1.0686]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 1]), e shape: torch.Size([300, 3])
SpMM(message func: copy_rhs, reduce func: max)
___________________ test_spmm[idtype0-max-copy_rhs-shp3-g0] ____________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (1,)), msg = 'copy_rhs', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72b143168>
op = 'copy_rhs', reduce_op = 'max'
lhs_data = tensor([[1.1632, 1.6662, 1.3166],
        [1.3862, 1.6820, 1.6806],
        [1.8191, 1.3018, 1.2840],
        [1.6399,... 1.0092],
        [1.8077, 1.1123, 1.8079],
        [1.2413, 1.0060, 1.4218]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.8514],
        [1.4549],
        [1.4747],
        [1.9970],
        [1.1620],
        [1.2223],
        [1...396],
        [1.3946],
        [1.3854],
        [1.9016],
        [1.7262]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 3]), e shape: torch.Size([100, 1])
SpMM(message func: copy_rhs, reduce func: max)
___________________ test_spmm[idtype0-max-copy_rhs-shp3-g1] ____________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (1,)), msg = 'copy_rhs', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43eb223a8>
op = 'copy_rhs', reduce_op = 'max'
lhs_data = tensor([[1.1036, 1.2782, 1.7402],
        [1.9434, 1.1302, 1.7559],
        [1.6385, 1.4987, 1.4572],
        [1.2217,... 1.2525],
        [1.2952, 1.3906, 1.6127],
        [1.2043, 1.7781, 1.8771]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.2115],
        [1.4252],
        [1.2264],
        [1.6249],
        [1.6005],
        [1.3187],
        [1...908],
        [1.9020],
        [1.4636],
        [1.6750],
        [1.0893]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 3]), e shape: torch.Size([300, 1])
SpMM(message func: copy_rhs, reduce func: max)
___________________ test_spmm[idtype0-max-copy_rhs-shp4-g0] ____________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), msg = 'copy_rhs', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb729332750>
op = 'copy_rhs', reduce_op = 'max'
lhs_data = tensor([[1.0752],
        [1.8211],
        [1.2676],
        [1.0775],
        [1.5246],
        [1.8727],
        [1...745],
        [1.4357],
        [1.8042],
        [1.7643],
        [1.5967]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.4779],
        [1.2015],
        [1.2862],
        [1.7438],
        [1.6801],
        [1.1645],
        [1...570],
        [1.1161],
        [1.5073],
        [1.3563],
        [1.6189]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30, 1]), e shape: torch.Size([100, 1])
SpMM(message func: copy_rhs, reduce func: max)
___________________ test_spmm[idtype0-max-copy_rhs-shp4-g1] ____________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), msg = 'copy_rhs', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72aa21870>
op = 'copy_rhs', reduce_op = 'max'
lhs_data = tensor([[1.9732],
        [1.3294],
        [1.0238],
        [1.4987],
        [1.7005],
        [1.0264],
        [1...411],
        [1.2375],
        [1.9828],
        [1.7104],
        [1.3851]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.0920],
        [1.7312],
        [1.5843],
        [1.1044],
        [1.8233],
        [1.0173],
        [1...516],
        [1.0609],
        [1.2289],
        [1.9156],
        [1.8980]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30, 1]), e shape: torch.Size([300, 1])
SpMM(message func: copy_rhs, reduce func: max)
___________________ test_spmm[idtype0-max-copy_rhs-shp5-g0] ____________________

idtype = torch.int32
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
shp = ((), ()), msg = 'copy_rhs', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72b1a31b0>
op = 'copy_rhs', reduce_op = 'max'
lhs_data = tensor([1.8754, 1.8211, 1.0742, 1.1501, 1.2496, 1.0345, 1.6998, 1.9336, 1.9475,
        1.7190, 1.9787, 1.5645, 1.8272...4583, 1.4130, 1.2605, 1.3066, 1.9594, 1.2147,
        1.0318, 1.0781, 1.3308], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([1.2299, 1.6775, 1.1920, 1.9270, 1.1895, 1.6898, 1.5972, 1.9073, 1.8616,
        1.1886, 1.8586, 1.6501, 1.0375...2806, 1.8451, 1.0185, 1.3740, 1.7539, 1.3936, 1.2735, 1.3822,
        1.0950], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
u shape: torch.Size([30]), e shape: torch.Size([100])
SpMM(message func: copy_rhs, reduce func: max)
___________________ test_spmm[idtype0-max-copy_rhs-shp5-g1] ____________________

idtype = torch.int32
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((), ()), msg = 'copy_rhs', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72b1a33f0>
op = 'copy_rhs', reduce_op = 'max'
lhs_data = tensor([1.7403, 1.9285, 1.3275, 1.1240, 1.5780, 1.5464, 1.6707, 1.7432, 1.5645,
        1.8457, 1.6308, 1.7242, 1.7206...5987, 1.8551, 1.4477, 1.5800, 1.4665, 1.6824,
        1.4540, 1.8206, 1.6907], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([1.5246, 1.7338, 1.2168, 1.3272, 1.7921, 1.9183, 1.7908, 1.0481, 1.0852,
        1.9479, 1.4978, 1.4498, 1.9464...7350, 1.0411, 1.2815, 1.5690, 1.0771, 1.1450,
        1.7969, 1.0667, 1.4873], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
u shape: torch.Size([30]), e shape: torch.Size([300])
SpMM(message func: copy_rhs, reduce func: max)
______________________ test_spmm[idtype1-sum-add-shp0-g0] ______________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), msg = 'add', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add', reduce_op = 'sum'
lhs_data = tensor([[[[[[1.9590],
            [1.3813],
            [1.4641]]],


          [[[1.5620],
            [1.5445],
    ...2]]],


          [[[1.8293],
            [1.7029],
            [1.9603]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.9900]],

           [[1.8275]],

           [[1.6995]]]],



         [[[[1.1445]],

           [[1.685...



         [[[[1.2008]],

           [[1.5227]],

           [[1.2191]]]]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 1, 2, 1, 3, 1]), e shape: torch.Size([100, 4, 1, 3, 1, 1])
SpMM(message func: add, reduce func: sum)
______________________ test_spmm[idtype1-sum-add-shp0-g1] ______________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), msg = 'add', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add', reduce_op = 'sum'
lhs_data = tensor([[[[[[1.7196],
            [1.0863],
            [1.5461]]],


          [[[1.4638],
            [1.3980],
    ...2]]],


          [[[1.5326],
            [1.4283],
            [1.5090]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.3184]],

           [[1.5259]],

           [[1.6907]]]],



         [[[[1.0793]],

           [[1.207...



         [[[[1.4529]],

           [[1.6614]],

           [[1.7628]]]]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 1, 2, 1, 3, 1]), e shape: torch.Size([300, 4, 1, 3, 1, 1])
SpMM(message func: add, reduce func: sum)
______________________ test_spmm[idtype1-sum-add-shp1-g0] ______________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1, 3), dtype=torch.float64)})
shp = ((3, 3), (1, 3)), msg = 'add', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add', reduce_op = 'sum'
lhs_data = tensor([[[1.9465, 1.5019, 1.9380],
         [1.1939, 1.9451, 1.9655],
         [1.0798, 1.2653, 1.9047]],

        [[1...3687],
         [1.4358, 1.6922, 1.7388],
         [1.9975, 1.9660, 1.4456]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[1.9018, 1.7721, 1.1130]],

        [[1.0444, 1.5432, 1.1949]],

        [[1.4461, 1.3231, 1.4779]],

       ...]],

        [[1.1177, 1.4325, 1.5220]],

        [[1.7967, 1.8107, 1.3115]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 3, 3]), e shape: torch.Size([100, 1, 3])
SpMM(message func: add, reduce func: sum)
______________________ test_spmm[idtype1-sum-add-shp1-g1] ______________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3, 3), (1, 3)), msg = 'add', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add', reduce_op = 'sum'
lhs_data = tensor([[[1.0080, 1.8381, 1.3458],
         [1.6152, 1.0019, 1.6085],
         [1.6218, 1.1245, 1.9025]],

        [[1...9798],
         [1.2207, 1.1349, 1.5382],
         [1.5135, 1.1109, 1.7371]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[1.7292, 1.2594, 1.2661]],

        [[1.6163, 1.2889, 1.8563]],

        [[1.2680, 1.1005, 1.0056]],

       ...]],

        [[1.2425, 1.3642, 1.5386]],

        [[1.7940, 1.6761, 1.9921]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 3, 3]), e shape: torch.Size([300, 1, 3])
SpMM(message func: add, reduce func: sum)
______________________ test_spmm[idtype1-sum-add-shp2-g0] ______________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float64)})
shp = ((1,), (3,)), msg = 'add', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add', reduce_op = 'sum'
lhs_data = tensor([[1.5339],
        [1.9754],
        [1.5769],
        [1.8768],
        [1.9460],
        [1.4810],
        [1...629],
        [1.4685],
        [1.4404],
        [1.0102],
        [1.2066]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.9636, 1.1689, 1.9173],
        [1.1647, 1.9737, 1.4318],
        [1.1844, 1.1581, 1.0655],
        [1.7096,... 1.5382],
        [1.4963, 1.5669, 1.9194],
        [1.5355, 1.9141, 1.9981]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1, 3), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 1]), e shape: torch.Size([100, 3])
SpMM(message func: add, reduce func: sum)
______________________ test_spmm[idtype1-sum-add-shp2-g1] ______________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (3,)), msg = 'add', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add', reduce_op = 'sum'
lhs_data = tensor([[1.8236],
        [1.6962],
        [1.1676],
        [1.5271],
        [1.5589],
        [1.6043],
        [1...051],
        [1.3447],
        [1.1225],
        [1.4677],
        [1.0953]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.0663, 1.9224, 1.2218],
        [1.4590, 1.3918, 1.6436],
        [1.1498, 1.0902, 1.0309],
        [1.9932,... 1.3018],
        [1.7125, 1.6099, 1.1223],
        [1.0478, 1.6962, 1.9119]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 1]), e shape: torch.Size([300, 3])
SpMM(message func: add, reduce func: sum)
______________________ test_spmm[idtype1-sum-add-shp3-g0] ______________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (1,)), msg = 'add', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add', reduce_op = 'sum'
lhs_data = tensor([[1.4106, 1.0011, 1.6612],
        [1.6426, 1.4639, 1.3215],
        [1.9818, 1.4399, 1.8933],
        [1.5671,... 1.8889],
        [1.2385, 1.4962, 1.0406],
        [1.6918, 1.5600, 1.7445]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.9069],
        [1.4577],
        [1.6101],
        [1.5220],
        [1.7614],
        [1.9455],
        [1...830],
        [1.0519],
        [1.9957],
        [1.9703],
        [1.4833]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 3]), e shape: torch.Size([100, 1])
SpMM(message func: add, reduce func: sum)
______________________ test_spmm[idtype1-sum-add-shp3-g1] ______________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (1,)), msg = 'add', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add', reduce_op = 'sum'
lhs_data = tensor([[1.2425, 1.0916, 1.1371],
        [1.5163, 1.9590, 1.6362],
        [1.4230, 1.8243, 1.3559],
        [1.1948,... 1.3989],
        [1.0130, 1.5549, 1.8336],
        [1.7381, 1.5061, 1.8212]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.9846],
        [1.4744],
        [1.0386],
        [1.2194],
        [1.3553],
        [1.0072],
        [1...655],
        [1.5706],
        [1.1338],
        [1.7391],
        [1.5881]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 3]), e shape: torch.Size([300, 1])
SpMM(message func: add, reduce func: sum)
______________________ test_spmm[idtype1-sum-add-shp4-g0] ______________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), msg = 'add', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add', reduce_op = 'sum'
lhs_data = tensor([[1.6979],
        [1.4976],
        [1.1769],
        [1.2819],
        [1.9991],
        [1.2906],
        [1...742],
        [1.8459],
        [1.6854],
        [1.3552],
        [1.8876]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.6297],
        [1.8047],
        [1.4807],
        [1.1301],
        [1.3427],
        [1.2611],
        [1...466],
        [1.0430],
        [1.2389],
        [1.4661],
        [1.6237]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 1]), e shape: torch.Size([100, 1])
SpMM(message func: add, reduce func: sum)
______________________ test_spmm[idtype1-sum-add-shp4-g1] ______________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), msg = 'add', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add', reduce_op = 'sum'
lhs_data = tensor([[1.4676],
        [1.8453],
        [1.2619],
        [1.7172],
        [1.5700],
        [1.6069],
        [1...272],
        [1.5323],
        [1.9195],
        [1.8061],
        [1.6204]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.3990],
        [1.1973],
        [1.3856],
        [1.3935],
        [1.1838],
        [1.5023],
        [1...617],
        [1.4686],
        [1.9449],
        [1.9466],
        [1.3712]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 1]), e shape: torch.Size([300, 1])
SpMM(message func: add, reduce func: sum)
______________________ test_spmm[idtype1-sum-add-shp5-g0] ______________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
shp = ((), ()), msg = 'add', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add', reduce_op = 'sum'
lhs_data = tensor([1.0566, 1.6046, 1.0657, 1.7811, 1.6443, 1.6422, 1.6756, 1.2653, 1.4132,
        1.0449, 1.5618, 1.9736, 1.2573...8284, 1.5187, 1.1639, 1.1414, 1.0887, 1.9097,
        1.3085, 1.9753, 1.0912], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([1.7370, 1.9829, 1.5376, 1.5724, 1.6898, 1.3567, 1.7435, 1.3544, 1.2646,
        1.6375, 1.5233, 1.3608, 1.1654...0587, 1.0287, 1.3322, 1.3305, 1.6188, 1.3851, 1.4217, 1.2990,
        1.2648], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30]), e shape: torch.Size([100])
SpMM(message func: add, reduce func: sum)
______________________ test_spmm[idtype1-sum-add-shp5-g1] ______________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((), ()), msg = 'add', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add', reduce_op = 'sum'
lhs_data = tensor([1.9104, 1.4930, 1.5941, 1.7849, 1.8509, 1.4751, 1.4084, 1.6963, 1.3984,
        1.0809, 1.9582, 1.2274, 1.9283...8990, 1.3016, 1.8106, 1.0469, 1.9544, 1.4349,
        1.2761, 1.8001, 1.4912], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([1.6607, 1.9699, 1.7756, 1.5934, 1.4169, 1.0750, 1.6201, 1.6987, 1.8453,
        1.5798, 1.9245, 1.0506, 1.1664...6890, 1.2962, 1.4046, 1.5960, 1.4912, 1.3843,
        1.1693, 1.5746, 1.9482], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30]), e shape: torch.Size([300])
SpMM(message func: add, reduce func: sum)
______________________ test_spmm[idtype1-sum-sub-shp0-g0] ______________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), msg = 'sub', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add', reduce_op = 'sum'
lhs_data = tensor([[[[[[1.4017],
            [1.3021],
            [1.4060]]],


          [[[1.9332],
            [1.1315],
    ...5]]],


          [[[1.0123],
            [1.7103],
            [1.7058]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[-1.0203]],

           [[-1.4523]],

           [[-1.2732]]]],



         [[[[-1.2772]],

           [[-...       [[[[-1.9409]],

           [[-1.1286]],

           [[-1.1472]]]]]], dtype=torch.float64, grad_fn=<NegBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 1, 2, 1, 3, 1]), e shape: torch.Size([100, 4, 1, 3, 1, 1])
SpMM(message func: sub, reduce func: sum)
______________________ test_spmm[idtype1-sum-sub-shp0-g1] ______________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), msg = 'sub', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add', reduce_op = 'sum'
lhs_data = tensor([[[[[[1.8348],
            [1.0356],
            [1.0929]]],


          [[[1.5014],
            [1.9266],
    ...6]]],


          [[[1.1606],
            [1.6438],
            [1.7354]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[-1.6229]],

           [[-1.4845]],

           [[-1.7832]]]],



         [[[[-1.0560]],

           [[-...       [[[[-1.4471]],

           [[-1.5106]],

           [[-1.4801]]]]]], dtype=torch.float64, grad_fn=<NegBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 1, 2, 1, 3, 1]), e shape: torch.Size([300, 4, 1, 3, 1, 1])
SpMM(message func: sub, reduce func: sum)
______________________ test_spmm[idtype1-sum-sub-shp1-g0] ______________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1, 3), dtype=torch.float64)})
shp = ((3, 3), (1, 3)), msg = 'sub', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add', reduce_op = 'sum'
lhs_data = tensor([[[1.3813, 1.1439, 1.1474],
         [1.5304, 1.4805, 1.3921],
         [1.9278, 1.8214, 1.4486]],

        [[1...1382],
         [1.4072, 1.5694, 1.0207],
         [1.0535, 1.8369, 1.8469]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[-1.9666, -1.2291, -1.3630]],

        [[-1.8404, -1.1437, -1.3897]],

        [[-1.4856, -1.4653, -1.2057]],...1.3892, -1.5502, -1.2166]],

        [[-1.8367, -1.5608, -1.1268]]], dtype=torch.float64,
       grad_fn=<NegBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 3, 3]), e shape: torch.Size([100, 1, 3])
SpMM(message func: sub, reduce func: sum)
______________________ test_spmm[idtype1-sum-sub-shp1-g1] ______________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3, 3), (1, 3)), msg = 'sub', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add', reduce_op = 'sum'
lhs_data = tensor([[[1.0819, 1.6953, 1.6903],
         [1.6790, 1.6554, 1.3711],
         [1.4616, 1.1941, 1.5664]],

        [[1...8060],
         [1.7620, 1.9121, 1.8570],
         [1.0681, 1.1227, 1.4061]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[-1.7827, -1.0772, -1.5389]],

        [[-1.7541, -1.6162, -1.5942]],

        [[-1.4964, -1.4012, -1.6526]],...1.5314, -1.0840, -1.8791]],

        [[-1.2470, -1.0677, -1.8311]]], dtype=torch.float64,
       grad_fn=<NegBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 3, 3]), e shape: torch.Size([300, 1, 3])
SpMM(message func: sub, reduce func: sum)
______________________ test_spmm[idtype1-sum-sub-shp2-g0] ______________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float64)})
shp = ((1,), (3,)), msg = 'sub', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add', reduce_op = 'sum'
lhs_data = tensor([[1.1221],
        [1.6230],
        [1.3191],
        [1.1245],
        [1.3120],
        [1.1018],
        [1...852],
        [1.4298],
        [1.9545],
        [1.7264],
        [1.9931]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.8124, -1.7985, -1.4755],
        [-1.7191, -1.4346, -1.0410],
        [-1.0353, -1.1614, -1.9693],
       ...  [-1.5248, -1.2576, -1.4448],
        [-1.3623, -1.4248, -1.6111]], dtype=torch.float64,
       grad_fn=<NegBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1, 3), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 1]), e shape: torch.Size([100, 3])
SpMM(message func: sub, reduce func: sum)
______________________ test_spmm[idtype1-sum-sub-shp2-g1] ______________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (3,)), msg = 'sub', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add', reduce_op = 'sum'
lhs_data = tensor([[1.3708],
        [1.4207],
        [1.0720],
        [1.5377],
        [1.6571],
        [1.3789],
        [1...310],
        [1.8326],
        [1.3604],
        [1.8040],
        [1.3998]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.9836, -1.4379, -1.8448],
        [-1.5777, -1.8902, -1.8768],
        [-1.7533, -1.8706, -1.6267],
       ...  [-1.7475, -1.4675, -1.9558],
        [-1.1816, -1.7097, -1.5514]], dtype=torch.float64,
       grad_fn=<NegBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 1]), e shape: torch.Size([300, 3])
SpMM(message func: sub, reduce func: sum)
______________________ test_spmm[idtype1-sum-sub-shp3-g0] ______________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (1,)), msg = 'sub', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add', reduce_op = 'sum'
lhs_data = tensor([[1.7504, 1.7120, 1.1756],
        [1.9901, 1.0901, 1.4582],
        [1.2202, 1.7173, 1.8631],
        [1.5366,... 1.9680],
        [1.1770, 1.9975, 1.5464],
        [1.2292, 1.5722, 1.2913]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.3902],
        [-1.6250],
        [-1.5946],
        [-1.1517],
        [-1.6409],
        [-1.6333],
    ...       [-1.8791],
        [-1.5591],
        [-1.0976],
        [-1.7082]], dtype=torch.float64, grad_fn=<NegBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 3]), e shape: torch.Size([100, 1])
SpMM(message func: sub, reduce func: sum)
______________________ test_spmm[idtype1-sum-sub-shp3-g1] ______________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (1,)), msg = 'sub', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add', reduce_op = 'sum'
lhs_data = tensor([[1.9437, 1.3265, 1.9900],
        [1.1226, 1.3537, 1.6065],
        [1.3489, 1.9140, 1.8898],
        [1.6430,... 1.6326],
        [1.5264, 1.5909, 1.2813],
        [1.3065, 1.5928, 1.0477]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.1767],
        [-1.5158],
        [-1.6504],
        [-1.7435],
        [-1.2121],
        [-1.1788],
    ...       [-1.0987],
        [-1.1166],
        [-1.7169],
        [-1.1293]], dtype=torch.float64, grad_fn=<NegBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 3]), e shape: torch.Size([300, 1])
SpMM(message func: sub, reduce func: sum)
______________________ test_spmm[idtype1-sum-sub-shp4-g0] ______________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), msg = 'sub', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add', reduce_op = 'sum'
lhs_data = tensor([[1.1972],
        [1.0537],
        [1.5912],
        [1.5107],
        [1.4551],
        [1.7715],
        [1...426],
        [1.0088],
        [1.6175],
        [1.0794],
        [1.3400]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.6006],
        [-1.5744],
        [-1.6139],
        [-1.1359],
        [-1.4277],
        [-1.7456],
    ...       [-1.6696],
        [-1.9944],
        [-1.3820],
        [-1.4215]], dtype=torch.float64, grad_fn=<NegBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 1]), e shape: torch.Size([100, 1])
SpMM(message func: sub, reduce func: sum)
______________________ test_spmm[idtype1-sum-sub-shp4-g1] ______________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), msg = 'sub', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add', reduce_op = 'sum'
lhs_data = tensor([[1.1355],
        [1.0976],
        [1.6875],
        [1.5762],
        [1.2718],
        [1.6151],
        [1...787],
        [1.0494],
        [1.0230],
        [1.4540],
        [1.2801]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.4029],
        [-1.0429],
        [-1.8518],
        [-1.9069],
        [-1.4328],
        [-1.1749],
    ...       [-1.6455],
        [-1.2488],
        [-1.8961],
        [-1.4914]], dtype=torch.float64, grad_fn=<NegBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 1]), e shape: torch.Size([300, 1])
SpMM(message func: sub, reduce func: sum)
______________________ test_spmm[idtype1-sum-sub-shp5-g0] ______________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
shp = ((), ()), msg = 'sub', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add', reduce_op = 'sum'
lhs_data = tensor([1.2712, 1.0310, 1.3699, 1.0337, 1.6787, 1.3711, 1.3328, 1.9993, 1.6272,
        1.7678, 1.7575, 1.7615, 1.1168...3504, 1.4146, 1.6309, 1.4586, 1.7031, 1.5621,
        1.8300, 1.0398, 1.2227], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([-1.2878, -1.8963, -1.7098, -1.0219, -1.2627, -1.3822, -1.7609, -1.2647,
        -1.8582, -1.7254, -1.3636, -1....9334, -1.8816, -1.4167,
        -1.4158, -1.0699, -1.5390, -1.6575], dtype=torch.float64,
       grad_fn=<NegBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30]), e shape: torch.Size([100])
SpMM(message func: sub, reduce func: sum)
______________________ test_spmm[idtype1-sum-sub-shp5-g1] ______________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((), ()), msg = 'sub', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add', reduce_op = 'sum'
lhs_data = tensor([1.7103, 1.7761, 1.4541, 1.1740, 1.3146, 1.8828, 1.2236, 1.5058, 1.5811,
        1.0827, 1.8687, 1.7723, 1.4835...4525, 1.4561, 1.4894, 1.1691, 1.2874, 1.6025,
        1.3527, 1.9124, 1.8299], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([-1.6108, -1.2814, -1.6885, -1.0054, -1.6392, -1.0609, -1.5444, -1.9433,
        -1.1341, -1.0589, -1.1218, -1....5558, -1.3582, -1.5204,
        -1.2757, -1.6942, -1.9377, -1.7919], dtype=torch.float64,
       grad_fn=<NegBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30]), e shape: torch.Size([300])
SpMM(message func: sub, reduce func: sum)
______________________ test_spmm[idtype1-sum-mul-shp0-g0] ______________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), msg = 'mul', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[[[[[1.4385],
            [1.7007],
            [1.6448]]],


          [[[1.3995],
            [1.0889],
    ...5]]],


          [[[1.8936],
            [1.9006],
            [1.1164]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.4722]],

           [[1.0473]],

           [[1.0957]]]],



         [[[[1.1487]],

           [[1.963...



         [[[[1.7808]],

           [[1.0508]],

           [[1.5724]]]]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 1, 2, 1, 3, 1]), e shape: torch.Size([100, 4, 1, 3, 1, 1])
SpMM(message func: mul, reduce func: sum)
______________________ test_spmm[idtype1-sum-mul-shp0-g1] ______________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), msg = 'mul', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[[[[[1.5449],
            [1.4723],
            [1.6888]]],


          [[[1.2473],
            [1.4352],
    ...8]]],


          [[[1.6659],
            [1.4314],
            [1.1698]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.2735]],

           [[1.6168]],

           [[1.0888]]]],



         [[[[1.1797]],

           [[1.438...



         [[[[1.9650]],

           [[1.7153]],

           [[1.8157]]]]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 1, 2, 1, 3, 1]), e shape: torch.Size([300, 4, 1, 3, 1, 1])
SpMM(message func: mul, reduce func: sum)
______________________ test_spmm[idtype1-sum-mul-shp1-g0] ______________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1, 3), dtype=torch.float64)})
shp = ((3, 3), (1, 3)), msg = 'mul', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[[1.9862, 1.2386, 1.5903],
         [1.8031, 1.3953, 1.6375],
         [1.3190, 1.2847, 1.7156]],

        [[1...9726],
         [1.1223, 1.7131, 1.4010],
         [1.0354, 1.8621, 1.1066]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[1.9512, 1.6853, 1.9368]],

        [[1.0624, 1.2073, 1.1803]],

        [[1.4618, 1.8211, 1.2640]],

       ...]],

        [[1.9426, 1.3838, 1.9163]],

        [[1.2791, 1.1296, 1.7781]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 3, 3]), e shape: torch.Size([100, 1, 3])
SpMM(message func: mul, reduce func: sum)
______________________ test_spmm[idtype1-sum-mul-shp1-g1] ______________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3, 3), (1, 3)), msg = 'mul', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[[1.5819, 1.0015, 1.1361],
         [1.3415, 1.7774, 1.4701],
         [1.5070, 1.5408, 1.6602]],

        [[1...8217],
         [1.0256, 1.2239, 1.9355],
         [1.2985, 1.6669, 1.3882]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[1.8736, 1.6301, 1.3182]],

        [[1.1271, 1.4420, 1.1706]],

        [[1.7582, 1.9688, 1.2796]],

       ...]],

        [[1.9424, 1.8408, 1.5395]],

        [[1.8296, 1.7860, 1.1986]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 3, 3]), e shape: torch.Size([300, 1, 3])
SpMM(message func: mul, reduce func: sum)
______________________ test_spmm[idtype1-sum-mul-shp2-g0] ______________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float64)})
shp = ((1,), (3,)), msg = 'mul', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[1.6026],
        [1.1780],
        [1.7234],
        [1.3812],
        [1.1804],
        [1.0828],
        [1...697],
        [1.7235],
        [1.0519],
        [1.6541],
        [1.9086]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.3962, 1.8464, 1.2005],
        [1.2285, 1.3697, 1.0651],
        [1.9317, 1.3835, 1.1697],
        [1.4913,... 1.2771],
        [1.3437, 1.8098, 1.0819],
        [1.1743, 1.4121, 1.1606]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1, 3), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 1]), e shape: torch.Size([100, 3])
SpMM(message func: mul, reduce func: sum)
______________________ test_spmm[idtype1-sum-mul-shp2-g1] ______________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (3,)), msg = 'mul', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[1.2339],
        [1.5265],
        [1.6499],
        [1.5901],
        [1.6148],
        [1.2467],
        [1...962],
        [1.4892],
        [1.5193],
        [1.7435],
        [1.7881]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.5030, 1.2868, 1.2985],
        [1.8815, 1.3756, 1.2761],
        [1.3366, 1.1614, 1.6079],
        [1.5954,... 1.8972],
        [1.5664, 1.4177, 1.8358],
        [1.3895, 1.8056, 1.4485]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 1]), e shape: torch.Size([300, 3])
SpMM(message func: mul, reduce func: sum)
______________________ test_spmm[idtype1-sum-mul-shp3-g0] ______________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (1,)), msg = 'mul', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[1.1183, 1.0680, 1.4964],
        [1.7711, 1.1008, 1.5546],
        [1.6423, 1.9512, 1.5329],
        [1.6648,... 1.3739],
        [1.2981, 1.8005, 1.1427],
        [1.8512, 1.1015, 1.8171]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.1579],
        [1.1475],
        [1.2842],
        [1.2111],
        [1.0553],
        [1.4200],
        [1...761],
        [1.7249],
        [1.3724],
        [1.4161],
        [1.8300]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 3]), e shape: torch.Size([100, 1])
SpMM(message func: mul, reduce func: sum)
______________________ test_spmm[idtype1-sum-mul-shp3-g1] ______________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (1,)), msg = 'mul', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[1.5482, 1.4010, 1.9620],
        [1.8252, 1.9546, 1.0203],
        [1.0198, 1.4791, 1.0050],
        [1.6436,... 1.4173],
        [1.6547, 1.1332, 1.4478],
        [1.3360, 1.9226, 1.5354]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.5535],
        [1.6841],
        [1.7008],
        [1.3310],
        [1.1638],
        [1.0454],
        [1...280],
        [1.8878],
        [1.6588],
        [1.4414],
        [1.7579]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 3]), e shape: torch.Size([300, 1])
SpMM(message func: mul, reduce func: sum)
______________________ test_spmm[idtype1-sum-mul-shp4-g0] ______________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), msg = 'mul', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[1.6683],
        [1.9697],
        [1.3946],
        [1.6888],
        [1.3422],
        [1.5867],
        [1...629],
        [1.7310],
        [1.5094],
        [1.9804],
        [1.8539]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.9398],
        [1.4507],
        [1.3131],
        [1.8713],
        [1.3989],
        [1.7352],
        [1...331],
        [1.8321],
        [1.6142],
        [1.9013],
        [1.3965]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 1]), e shape: torch.Size([100, 1])
SpMM(message func: mul, reduce func: sum)
______________________ test_spmm[idtype1-sum-mul-shp4-g1] ______________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), msg = 'mul', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[1.8493],
        [1.2547],
        [1.4678],
        [1.5707],
        [1.1762],
        [1.7696],
        [1...160],
        [1.1654],
        [1.2978],
        [1.4658],
        [1.3484]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.9784],
        [1.6214],
        [1.8322],
        [1.6529],
        [1.6067],
        [1.1147],
        [1...666],
        [1.0670],
        [1.9236],
        [1.9535],
        [1.3905]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 1]), e shape: torch.Size([300, 1])
SpMM(message func: mul, reduce func: sum)
______________________ test_spmm[idtype1-sum-mul-shp5-g0] ______________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
shp = ((), ()), msg = 'mul', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([1.0900, 1.5596, 1.0163, 1.2366, 1.1149, 1.0651, 1.3200, 1.0131, 1.1100,
        1.8606, 1.8904, 1.4584, 1.1638...4573, 1.0903, 1.2263, 1.9710, 1.5769, 1.9625,
        1.7841, 1.4869, 1.3011], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([1.8594, 1.4158, 1.0343, 1.1527, 1.4824, 1.9724, 1.5400, 1.2552, 1.0039,
        1.7558, 1.0647, 1.8020, 1.5410...9588, 1.9700, 1.7919, 1.6260, 1.6866, 1.7335, 1.4692, 1.1104,
        1.6920], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30]), e shape: torch.Size([100])
SpMM(message func: mul, reduce func: sum)
______________________ test_spmm[idtype1-sum-mul-shp5-g1] ______________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((), ()), msg = 'mul', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([1.8864, 1.7254, 1.2853, 1.5635, 1.8530, 1.2006, 1.7270, 1.3487, 1.4432,
        1.2469, 1.9619, 1.2554, 1.6753...2551, 1.2187, 1.3963, 1.9705, 1.2969, 1.7317,
        1.8432, 1.5580, 1.2253], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([1.9463, 1.0393, 1.5744, 1.8254, 1.7578, 1.4840, 1.1708, 1.6098, 1.8541,
        1.0537, 1.5115, 1.9501, 1.2951...3765, 1.4009, 1.4053, 1.9185, 1.4224, 1.3911,
        1.7552, 1.2900, 1.4991], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30]), e shape: torch.Size([300])
SpMM(message func: mul, reduce func: sum)
______________________ test_spmm[idtype1-sum-div-shp0-g0] ______________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), msg = 'div', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[[[[[1.3799],
            [1.6150],
            [1.0065]]],


          [[[1.1020],
            [1.2968],
    ...3]]],


          [[[1.5334],
            [1.7261],
            [1.4070]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[0.5200]],

           [[0.6409]],

           [[0.8493]]]],



         [[[[0.6101]],

           [[0.509...         [[[[0.7261]],

           [[0.5533]],

           [[0.6747]]]]]], dtype=torch.float64, grad_fn=<MulBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 1, 2, 1, 3, 1]), e shape: torch.Size([100, 4, 1, 3, 1, 1])
SpMM(message func: div, reduce func: sum)
______________________ test_spmm[idtype1-sum-div-shp0-g1] ______________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), msg = 'div', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[[[[[1.4798],
            [1.7354],
            [1.9145]]],


          [[[1.5692],
            [1.6292],
    ...7]]],


          [[[1.3515],
            [1.7379],
            [1.8931]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[0.5726]],

           [[0.8688]],

           [[0.7166]]]],



         [[[[0.5324]],

           [[0.504...         [[[[0.8489]],

           [[0.5195]],

           [[0.9099]]]]]], dtype=torch.float64, grad_fn=<MulBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 1, 2, 1, 3, 1]), e shape: torch.Size([300, 4, 1, 3, 1, 1])
SpMM(message func: div, reduce func: sum)
______________________ test_spmm[idtype1-sum-div-shp1-g0] ______________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1, 3), dtype=torch.float64)})
shp = ((3, 3), (1, 3)), msg = 'div', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[[1.8159, 1.5376, 1.7199],
         [1.0274, 1.3888, 1.7630],
         [1.5737, 1.0614, 1.1303]],

        [[1...3613],
         [1.3806, 1.6053, 1.0124],
         [1.3637, 1.5941, 1.8091]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[0.8715, 0.7792, 0.5211]],

        [[0.7082, 0.5267, 0.9934]],

        [[0.9772, 0.5194, 0.9411]],

       ...  [[0.7167, 0.8200, 0.7908]],

        [[0.6562, 0.6958, 0.6318]]], dtype=torch.float64,
       grad_fn=<MulBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 3, 3]), e shape: torch.Size([100, 1, 3])
SpMM(message func: div, reduce func: sum)
______________________ test_spmm[idtype1-sum-div-shp1-g1] ______________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3, 3), (1, 3)), msg = 'div', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[[1.1268, 1.0420, 1.3624],
         [1.1015, 1.2097, 1.4692],
         [1.6067, 1.0514, 1.6071]],

        [[1...6788],
         [1.5679, 1.8355, 1.8997],
         [1.5042, 1.5453, 1.7477]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[0.5457, 0.9158, 0.6392]],

        [[0.5956, 0.6642, 0.6786]],

        [[0.5556, 0.6246, 0.6471]],

       ...  [[0.7507, 0.6567, 0.6867]],

        [[0.5462, 0.7397, 0.7714]]], dtype=torch.float64,
       grad_fn=<MulBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 3, 3]), e shape: torch.Size([300, 1, 3])
SpMM(message func: div, reduce func: sum)
______________________ test_spmm[idtype1-sum-div-shp2-g0] ______________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float64)})
shp = ((1,), (3,)), msg = 'div', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[1.8442],
        [1.0511],
        [1.8672],
        [1.1656],
        [1.3264],
        [1.1233],
        [1...758],
        [1.8897],
        [1.8559],
        [1.8904],
        [1.4276]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.5664, 0.7647, 0.6801],
        [0.9580, 0.5788, 0.7333],
        [0.9582, 0.6912, 0.7593],
        [0.5909,...257],
        [0.8243, 0.5953, 0.6771],
        [0.5814, 0.5174, 0.6744]], dtype=torch.float64, grad_fn=<MulBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1, 3), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 1]), e shape: torch.Size([100, 3])
SpMM(message func: div, reduce func: sum)
______________________ test_spmm[idtype1-sum-div-shp2-g1] ______________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (3,)), msg = 'div', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[1.8763],
        [1.9628],
        [1.3925],
        [1.9086],
        [1.2764],
        [1.4012],
        [1...492],
        [1.2418],
        [1.9014],
        [1.8320],
        [1.2668]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.7082, 0.9146, 0.5707],
        [0.6167, 0.7269, 0.6892],
        [0.7979, 0.9655, 0.7932],
        [0.5022,...238],
        [0.6508, 0.7432, 0.7589],
        [0.6378, 0.8237, 0.9224]], dtype=torch.float64, grad_fn=<MulBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 1]), e shape: torch.Size([300, 3])
SpMM(message func: div, reduce func: sum)
______________________ test_spmm[idtype1-sum-div-shp3-g0] ______________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (1,)), msg = 'div', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[1.2943, 1.7322, 1.3901],
        [1.5139, 1.3192, 1.8698],
        [1.3034, 1.1595, 1.3942],
        [1.7655,... 1.3730],
        [1.8400, 1.6642, 1.0901],
        [1.4289, 1.4916, 1.4630]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.6401],
        [0.8621],
        [0.7813],
        [0.5103],
        [0.5189],
        [0.5073],
        [0...,
        [0.5073],
        [0.7184],
        [0.7099],
        [0.6830]], dtype=torch.float64, grad_fn=<MulBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 3]), e shape: torch.Size([100, 1])
SpMM(message func: div, reduce func: sum)
______________________ test_spmm[idtype1-sum-div-shp3-g1] ______________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (1,)), msg = 'div', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[1.1817, 1.7159, 1.7889],
        [1.3671, 1.5372, 1.2907],
        [1.7830, 1.9551, 1.8556],
        [1.0553,... 1.0780],
        [1.0417, 1.0749, 1.4547],
        [1.7992, 1.1053, 1.1828]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.6548],
        [0.7000],
        [0.6429],
        [0.7911],
        [0.6993],
        [0.8968],
        [0...,
        [0.7133],
        [0.5982],
        [0.6422],
        [0.5552]], dtype=torch.float64, grad_fn=<MulBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 3]), e shape: torch.Size([300, 1])
SpMM(message func: div, reduce func: sum)
______________________ test_spmm[idtype1-sum-div-shp4-g0] ______________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), msg = 'div', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[1.9722],
        [1.1213],
        [1.1088],
        [1.7359],
        [1.8794],
        [1.7525],
        [1...988],
        [1.5390],
        [1.2658],
        [1.3736],
        [1.9788]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.7489],
        [0.6259],
        [0.6415],
        [0.6672],
        [0.9453],
        [0.6035],
        [0...,
        [0.8960],
        [0.5421],
        [0.5528],
        [0.7000]], dtype=torch.float64, grad_fn=<MulBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 1]), e shape: torch.Size([100, 1])
SpMM(message func: div, reduce func: sum)
______________________ test_spmm[idtype1-sum-div-shp4-g1] ______________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), msg = 'div', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[1.1799],
        [1.1100],
        [1.4682],
        [1.6186],
        [1.7700],
        [1.1562],
        [1...588],
        [1.3699],
        [1.3363],
        [1.8482],
        [1.8526]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.6047],
        [0.8038],
        [0.7028],
        [0.5133],
        [0.5064],
        [0.9456],
        [0...,
        [0.5768],
        [0.5190],
        [0.9973],
        [0.9444]], dtype=torch.float64, grad_fn=<MulBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 1]), e shape: torch.Size([300, 1])
SpMM(message func: div, reduce func: sum)
______________________ test_spmm[idtype1-sum-div-shp5-g0] ______________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
shp = ((), ()), msg = 'div', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([1.9472, 1.5994, 1.6640, 1.5543, 1.8757, 1.2315, 1.1515, 1.3723, 1.2283,
        1.3012, 1.1891, 1.3534, 1.8446...4655, 1.3585, 1.3223, 1.7618, 1.0561, 1.2168,
        1.7424, 1.1325, 1.3242], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([0.9605, 0.8879, 0.5772, 0.6767, 0.5919, 0.8494, 0.5639, 0.5617, 0.9977,
        0.9070, 0.7286, 0.5567, 0.8567..., 0.6590, 0.6702, 0.7103, 0.5191, 0.5807, 0.6962, 0.7916,
        0.5462], dtype=torch.float64, grad_fn=<MulBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30]), e shape: torch.Size([100])
SpMM(message func: div, reduce func: sum)
______________________ test_spmm[idtype1-sum-div-shp5-g1] ______________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((), ()), msg = 'div', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([1.1421, 1.3980, 1.3958, 1.1674, 1.4836, 1.8906, 1.2427, 1.8017, 1.1345,
        1.8977, 1.6261, 1.9619, 1.0013...9388, 1.9209, 1.7353, 1.9747, 1.4079, 1.6914,
        1.6582, 1.0033, 1.5640], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([0.5378, 0.5013, 0.7749, 0.9032, 0.9497, 0.8555, 0.5447, 0.5307, 0.5015,
        0.5926, 0.8675, 0.6255, 0.8815..., 0.5427, 0.7440, 0.8837, 0.6241, 0.6629,
        0.5095, 0.9686, 0.9230], dtype=torch.float64, grad_fn=<MulBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30]), e shape: torch.Size([300])
SpMM(message func: div, reduce func: sum)
___________________ test_spmm[idtype1-sum-copy_lhs-shp0-g0] ____________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), msg = 'copy_lhs', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[[[[[1.9656],
            [1.1263],
            [1.8371]]],


          [[[1.8688],
            [1.0640],
    ...5]]],


          [[[1.8659],
            [1.8259],
            [1.5949]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.4279]],

           [[1.7153]],

           [[1.5640]]]],



         [[[[1.3120]],

           [[1.512...



         [[[[1.3728]],

           [[1.7397]],

           [[1.4268]]]]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 1, 2, 1, 3, 1]), e shape: torch.Size([100, 4, 1, 3, 1, 1])
SpMM(message func: copy_lhs, reduce func: sum)
___________________ test_spmm[idtype1-sum-copy_lhs-shp0-g1] ____________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), msg = 'copy_lhs', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[[[[[1.2846],
            [1.5743],
            [1.9885]]],


          [[[1.7539],
            [1.7718],
    ...8]]],


          [[[1.5873],
            [1.8830],
            [1.9861]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.1291]],

           [[1.9296]],

           [[1.6768]]]],



         [[[[1.0048]],

           [[1.306...



         [[[[1.9846]],

           [[1.4790]],

           [[1.0806]]]]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 1, 2, 1, 3, 1]), e shape: torch.Size([300, 4, 1, 3, 1, 1])
SpMM(message func: copy_lhs, reduce func: sum)
___________________ test_spmm[idtype1-sum-copy_lhs-shp1-g0] ____________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1, 3), dtype=torch.float64)})
shp = ((3, 3), (1, 3)), msg = 'copy_lhs', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[[1.6395, 1.6252, 1.8576],
         [1.1737, 1.5628, 1.1700],
         [1.8471, 1.2822, 1.6228]],

        [[1...4093],
         [1.1318, 1.4383, 1.3144],
         [1.8556, 1.2756, 1.2838]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[1.4989, 1.5987, 1.1602]],

        [[1.5305, 1.8715, 1.6367]],

        [[1.7419, 1.1818, 1.8809]],

       ...]],

        [[1.3107, 1.9559, 1.3080]],

        [[1.8068, 1.6875, 1.7043]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 3, 3]), e shape: torch.Size([100, 1, 3])
SpMM(message func: copy_lhs, reduce func: sum)
___________________ test_spmm[idtype1-sum-copy_lhs-shp1-g1] ____________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3, 3), (1, 3)), msg = 'copy_lhs', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[[1.5425, 1.0112, 1.3100],
         [1.5294, 1.7102, 1.8570],
         [1.5733, 1.4406, 1.8329]],

        [[1...7502],
         [1.0756, 1.0151, 1.0475],
         [1.1834, 1.5998, 1.8751]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[1.9234, 1.1191, 1.7297]],

        [[1.9583, 1.7130, 1.4492]],

        [[1.4947, 1.1560, 1.5657]],

       ...]],

        [[1.7480, 1.7266, 1.9032]],

        [[1.6557, 1.2166, 1.3794]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 3, 3]), e shape: torch.Size([300, 1, 3])
SpMM(message func: copy_lhs, reduce func: sum)
___________________ test_spmm[idtype1-sum-copy_lhs-shp2-g0] ____________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float64)})
shp = ((1,), (3,)), msg = 'copy_lhs', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[1.2849],
        [1.9471],
        [1.2280],
        [1.0398],
        [1.1139],
        [1.2108],
        [1...637],
        [1.1524],
        [1.0120],
        [1.3354],
        [1.2541]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.2137, 1.6744, 1.4004],
        [1.8149, 1.0704, 1.9954],
        [1.0102, 1.0363, 1.4953],
        [1.0253,... 1.1827],
        [1.4201, 1.3642, 1.9882],
        [1.2645, 1.3888, 1.6087]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1, 3), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 1]), e shape: torch.Size([100, 3])
SpMM(message func: copy_lhs, reduce func: sum)
___________________ test_spmm[idtype1-sum-copy_lhs-shp2-g1] ____________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (3,)), msg = 'copy_lhs', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[1.2235],
        [1.4667],
        [1.1582],
        [1.0059],
        [1.5625],
        [1.8636],
        [1...172],
        [1.8308],
        [1.4035],
        [1.6593],
        [1.7022]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.5205, 1.1763, 1.6230],
        [1.2805, 1.9518, 1.5077],
        [1.3653, 1.6416, 1.3277],
        [1.1121,... 1.8070],
        [1.8039, 1.2296, 1.9037],
        [1.0199, 1.8401, 1.9346]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 1]), e shape: torch.Size([300, 3])
SpMM(message func: copy_lhs, reduce func: sum)
___________________ test_spmm[idtype1-sum-copy_lhs-shp3-g0] ____________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (1,)), msg = 'copy_lhs', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[1.2030, 1.1628, 1.4151],
        [1.0991, 1.3831, 1.4098],
        [1.8092, 1.5743, 1.1144],
        [1.2254,... 1.4681],
        [1.9398, 1.9435, 1.8548],
        [1.2849, 1.0713, 1.2894]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.8526],
        [1.0118],
        [1.7216],
        [1.4453],
        [1.8715],
        [1.6908],
        [1...989],
        [1.9659],
        [1.5678],
        [1.9465],
        [1.1170]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 3]), e shape: torch.Size([100, 1])
SpMM(message func: copy_lhs, reduce func: sum)
___________________ test_spmm[idtype1-sum-copy_lhs-shp3-g1] ____________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (1,)), msg = 'copy_lhs', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[1.2375, 1.6829, 1.7505],
        [1.6399, 1.0153, 1.7637],
        [1.8919, 1.0466, 1.7417],
        [1.7962,... 1.6201],
        [1.3385, 1.8322, 1.3304],
        [1.6596, 1.8411, 1.1979]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.0040],
        [1.2383],
        [1.1180],
        [1.8387],
        [1.4332],
        [1.9495],
        [1...847],
        [1.5549],
        [1.6257],
        [1.5654],
        [1.8516]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 3]), e shape: torch.Size([300, 1])
SpMM(message func: copy_lhs, reduce func: sum)
___________________ test_spmm[idtype1-sum-copy_lhs-shp4-g0] ____________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), msg = 'copy_lhs', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[1.8999],
        [1.4458],
        [1.7357],
        [1.3922],
        [1.9630],
        [1.8944],
        [1...887],
        [1.4317],
        [1.7635],
        [1.1341],
        [1.7529]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.9710],
        [1.2617],
        [1.2387],
        [1.4737],
        [1.2326],
        [1.1427],
        [1...347],
        [1.6674],
        [1.5093],
        [1.6918],
        [1.8887]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 1]), e shape: torch.Size([100, 1])
SpMM(message func: copy_lhs, reduce func: sum)
___________________ test_spmm[idtype1-sum-copy_lhs-shp4-g1] ____________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), msg = 'copy_lhs', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[1.5662],
        [1.6983],
        [1.8147],
        [1.3186],
        [1.7071],
        [1.7492],
        [1...148],
        [1.8578],
        [1.7017],
        [1.0520],
        [1.0366]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.8457],
        [1.1255],
        [1.6835],
        [1.1320],
        [1.7651],
        [1.1515],
        [1...919],
        [1.1819],
        [1.6263],
        [1.3923],
        [1.2164]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 1]), e shape: torch.Size([300, 1])
SpMM(message func: copy_lhs, reduce func: sum)
___________________ test_spmm[idtype1-sum-copy_lhs-shp5-g0] ____________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
shp = ((), ()), msg = 'copy_lhs', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([1.4397, 1.5967, 1.8072, 1.0524, 1.7289, 1.5063, 1.1009, 1.4451, 1.9126,
        1.1190, 1.6235, 1.2619, 1.1558...4384, 1.9773, 1.2082, 1.2912, 1.4578, 1.1719,
        1.4640, 1.7758, 1.6153], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([1.7867, 1.6823, 1.0589, 1.7804, 1.2350, 1.1154, 1.1219, 1.7200, 1.6248,
        1.9741, 1.4602, 1.3868, 1.9751...4220, 1.9897, 1.6066, 1.8044, 1.7518, 1.5158, 1.9954, 1.8651,
        1.4057], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30]), e shape: torch.Size([100])
SpMM(message func: copy_lhs, reduce func: sum)
___________________ test_spmm[idtype1-sum-copy_lhs-shp5-g1] ____________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((), ()), msg = 'copy_lhs', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([1.1732, 1.6152, 1.9398, 1.4631, 1.9603, 1.8152, 1.9534, 1.1104, 1.7235,
        1.2067, 1.9356, 1.3969, 1.7886...3025, 1.0327, 1.2162, 1.6816, 1.9424, 1.2937,
        1.4508, 1.3610, 1.8060], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([1.5405, 1.2184, 1.8584, 1.0676, 1.8614, 1.8849, 1.9240, 1.5430, 1.0997,
        1.9175, 1.0570, 1.2426, 1.3183...1589, 1.3351, 1.3697, 1.4507, 1.6558, 1.0853,
        1.7856, 1.9017, 1.9853], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30]), e shape: torch.Size([300])
SpMM(message func: copy_lhs, reduce func: sum)
___________________ test_spmm[idtype1-sum-copy_rhs-shp0-g0] ____________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), msg = 'copy_rhs', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_rhs', reduce_op = 'sum'
lhs_data = tensor([[[[[[1.7501],
            [1.5715],
            [1.9112]]],


          [[[1.7885],
            [1.5950],
    ...6]]],


          [[[1.5544],
            [1.8755],
            [1.0282]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.7314]],

           [[1.0892]],

           [[1.3785]]]],



         [[[[1.7192]],

           [[1.479...



         [[[[1.0676]],

           [[1.1762]],

           [[1.1476]]]]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 1, 2, 1, 3, 1]), e shape: torch.Size([100, 4, 1, 3, 1, 1])
SpMM(message func: copy_rhs, reduce func: sum)
___________________ test_spmm[idtype1-sum-copy_rhs-shp0-g1] ____________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), msg = 'copy_rhs', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_rhs', reduce_op = 'sum'
lhs_data = tensor([[[[[[1.6862],
            [1.6579],
            [1.9593]]],


          [[[1.8568],
            [1.8319],
    ...7]]],


          [[[1.3825],
            [1.1748],
            [1.7879]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.2593]],

           [[1.1093]],

           [[1.8061]]]],



         [[[[1.7069]],

           [[1.029...



         [[[[1.3702]],

           [[1.3992]],

           [[1.2459]]]]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 1, 2, 1, 3, 1]), e shape: torch.Size([300, 4, 1, 3, 1, 1])
SpMM(message func: copy_rhs, reduce func: sum)
___________________ test_spmm[idtype1-sum-copy_rhs-shp1-g0] ____________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1, 3), dtype=torch.float64)})
shp = ((3, 3), (1, 3)), msg = 'copy_rhs', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_rhs', reduce_op = 'sum'
lhs_data = tensor([[[1.1500, 1.3343, 1.7756],
         [1.8189, 1.8814, 1.4228],
         [1.5445, 1.2928, 1.2122]],

        [[1...4407],
         [1.9386, 1.6416, 1.6561],
         [1.7920, 1.1707, 1.0474]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[1.0360, 1.0357, 1.3306]],

        [[1.1459, 1.7509, 1.3289]],

        [[1.2994, 1.4550, 1.1867]],

       ...]],

        [[1.0866, 1.9236, 1.3418]],

        [[1.0353, 1.1426, 1.9460]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 3, 3]), e shape: torch.Size([100, 1, 3])
SpMM(message func: copy_rhs, reduce func: sum)
___________________ test_spmm[idtype1-sum-copy_rhs-shp1-g1] ____________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3, 3), (1, 3)), msg = 'copy_rhs', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_rhs', reduce_op = 'sum'
lhs_data = tensor([[[1.2631, 1.9168, 1.5887],
         [1.1123, 1.7291, 1.8944],
         [1.1037, 1.7952, 1.1625]],

        [[1...3452],
         [1.0557, 1.7501, 1.5561],
         [1.5298, 1.3565, 1.3975]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[1.0540, 1.4496, 1.1693]],

        [[1.5534, 1.8674, 1.1064]],

        [[1.0240, 1.8367, 1.9743]],

       ...]],

        [[1.3109, 1.4165, 1.1420]],

        [[1.4634, 1.8744, 1.3742]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 3, 3]), e shape: torch.Size([300, 1, 3])
SpMM(message func: copy_rhs, reduce func: sum)
___________________ test_spmm[idtype1-sum-copy_rhs-shp2-g0] ____________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float64)})
shp = ((1,), (3,)), msg = 'copy_rhs', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_rhs', reduce_op = 'sum'
lhs_data = tensor([[1.1089],
        [1.4120],
        [1.1556],
        [1.3201],
        [1.8525],
        [1.1020],
        [1...242],
        [1.5511],
        [1.9383],
        [1.7984],
        [1.0562]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.9280, 1.5896, 1.0235],
        [1.8002, 1.5186, 1.9182],
        [1.8257, 1.7104, 1.4510],
        [1.5118,... 1.8043],
        [1.2972, 1.0286, 1.3006],
        [1.5840, 1.6892, 1.0126]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1, 3), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 1]), e shape: torch.Size([100, 3])
SpMM(message func: copy_rhs, reduce func: sum)
___________________ test_spmm[idtype1-sum-copy_rhs-shp2-g1] ____________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (3,)), msg = 'copy_rhs', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_rhs', reduce_op = 'sum'
lhs_data = tensor([[1.4957],
        [1.7674],
        [1.2941],
        [1.7876],
        [1.3636],
        [1.0150],
        [1...099],
        [1.4470],
        [1.8882],
        [1.8910],
        [1.0219]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.6880, 1.1646, 1.9590],
        [1.4536, 1.5842, 1.8182],
        [1.4575, 1.9571, 1.0425],
        [1.5290,... 1.1937],
        [1.6935, 1.4754, 1.2056],
        [1.7008, 1.5021, 1.2136]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 1]), e shape: torch.Size([300, 3])
SpMM(message func: copy_rhs, reduce func: sum)
___________________ test_spmm[idtype1-sum-copy_rhs-shp3-g0] ____________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (1,)), msg = 'copy_rhs', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_rhs', reduce_op = 'sum'
lhs_data = tensor([[1.6090, 1.0799, 1.4881],
        [1.1757, 1.2642, 1.7783],
        [1.3282, 1.6448, 2.0000],
        [1.8021,... 1.9838],
        [1.4876, 1.1194, 1.3461],
        [1.9431, 1.5817, 1.3629]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.1644],
        [1.4646],
        [1.4633],
        [1.8686],
        [1.3885],
        [1.3249],
        [1...990],
        [1.5472],
        [1.9084],
        [1.6722],
        [1.6511]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 3]), e shape: torch.Size([100, 1])
SpMM(message func: copy_rhs, reduce func: sum)
___________________ test_spmm[idtype1-sum-copy_rhs-shp3-g1] ____________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (1,)), msg = 'copy_rhs', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_rhs', reduce_op = 'sum'
lhs_data = tensor([[1.9546, 1.6114, 1.1086],
        [1.8216, 1.1505, 1.9370],
        [1.7012, 1.0445, 1.7861],
        [1.4284,... 1.5568],
        [1.0342, 1.2805, 1.2094],
        [1.4141, 1.6782, 1.1892]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.2427],
        [1.8736],
        [1.5317],
        [1.7974],
        [1.3775],
        [1.6441],
        [1...222],
        [1.1379],
        [1.2171],
        [1.9604],
        [1.7835]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 3]), e shape: torch.Size([300, 1])
SpMM(message func: copy_rhs, reduce func: sum)
___________________ test_spmm[idtype1-sum-copy_rhs-shp4-g0] ____________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), msg = 'copy_rhs', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_rhs', reduce_op = 'sum'
lhs_data = tensor([[1.2293],
        [1.1290],
        [1.2241],
        [1.8260],
        [1.0473],
        [1.2407],
        [1...062],
        [1.4298],
        [1.6843],
        [1.6183],
        [1.3251]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.0229],
        [1.3506],
        [1.2539],
        [1.0700],
        [1.1630],
        [1.9627],
        [1...091],
        [1.1501],
        [1.0034],
        [1.8029],
        [1.0454]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 1]), e shape: torch.Size([100, 1])
SpMM(message func: copy_rhs, reduce func: sum)
___________________ test_spmm[idtype1-sum-copy_rhs-shp4-g1] ____________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), msg = 'copy_rhs', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_rhs', reduce_op = 'sum'
lhs_data = tensor([[1.6017],
        [1.2612],
        [1.4638],
        [1.4273],
        [1.1989],
        [1.5673],
        [1...083],
        [1.0274],
        [1.4967],
        [1.3953],
        [1.2881]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.2453],
        [1.4729],
        [1.0863],
        [1.6248],
        [1.7602],
        [1.0603],
        [1...914],
        [1.1655],
        [1.7968],
        [1.8463],
        [1.0369]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 1]), e shape: torch.Size([300, 1])
SpMM(message func: copy_rhs, reduce func: sum)
___________________ test_spmm[idtype1-sum-copy_rhs-shp5-g0] ____________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
shp = ((), ()), msg = 'copy_rhs', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_rhs', reduce_op = 'sum'
lhs_data = tensor([1.8845, 1.7952, 1.2221, 1.7357, 1.1579, 1.8971, 1.4628, 1.7156, 1.9189,
        1.7702, 1.6391, 1.5604, 1.0444...3942, 1.7537, 1.9665, 1.7342, 1.9722, 1.8343,
        1.2691, 1.8725, 1.4181], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([1.5586, 1.4287, 1.2979, 1.7531, 1.2304, 1.3777, 1.8594, 1.9815, 1.4219,
        1.2679, 1.4820, 1.4887, 1.9938...7738, 1.3959, 1.6758, 1.6481, 1.6847, 1.2095, 1.4866, 1.6150,
        1.0682], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30]), e shape: torch.Size([100])
SpMM(message func: copy_rhs, reduce func: sum)
___________________ test_spmm[idtype1-sum-copy_rhs-shp5-g1] ____________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((), ()), msg = 'copy_rhs', reducer = 'sum'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_rhs', reduce_op = 'sum'
lhs_data = tensor([1.3158, 1.1246, 1.5951, 1.4251, 1.0783, 1.5706, 1.3217, 1.3423, 1.3151,
        1.4528, 1.4928, 1.6552, 1.4010...8836, 1.1844, 1.1143, 1.3958, 1.4157, 1.6525,
        1.8381, 1.7564, 1.0059], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([1.8282, 1.0095, 1.6740, 1.9940, 1.2973, 1.4253, 1.3036, 1.8156, 1.0124,
        1.5477, 1.4778, 1.6547, 1.8698...1566, 1.1537, 1.0246, 1.2479, 1.6438, 1.7058,
        1.6035, 1.9154, 1.8042], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30]), e shape: torch.Size([300])
SpMM(message func: copy_rhs, reduce func: sum)
______________________ test_spmm[idtype1-min-add-shp0-g0] ______________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), msg = 'add', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add', reduce_op = 'min'
lhs_data = tensor([[[[[[1.7863],
            [1.9733],
            [1.2908]]],


          [[[1.8563],
            [1.4619],
    ...4]]],


          [[[1.5360],
            [1.2621],
            [1.9066]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.3431]],

           [[1.7339]],

           [[1.1363]]]],



         [[[[1.1050]],

           [[1.668...



         [[[[1.7111]],

           [[1.8518]],

           [[1.7723]]]]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 1, 2, 1, 3, 1]), e shape: torch.Size([100, 4, 1, 3, 1, 1])
SpMM(message func: add, reduce func: min)
______________________ test_spmm[idtype1-min-add-shp0-g1] ______________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), msg = 'add', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add', reduce_op = 'min'
lhs_data = tensor([[[[[[1.4596],
            [1.6091],
            [1.0703]]],


          [[[1.2706],
            [1.5354],
    ...4]]],


          [[[1.8847],
            [1.7672],
            [1.1616]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.9535]],

           [[1.4220]],

           [[1.4534]]]],



         [[[[1.8113]],

           [[1.921...



         [[[[1.6969]],

           [[1.3244]],

           [[1.8695]]]]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 1, 2, 1, 3, 1]), e shape: torch.Size([300, 4, 1, 3, 1, 1])
SpMM(message func: add, reduce func: min)
______________________ test_spmm[idtype1-min-add-shp1-g0] ______________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1, 3), dtype=torch.float64)})
shp = ((3, 3), (1, 3)), msg = 'add', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add', reduce_op = 'min'
lhs_data = tensor([[[1.1305, 1.3726, 1.7812],
         [1.0212, 1.0640, 1.8234],
         [1.3354, 1.4905, 1.4655]],

        [[1...7009],
         [1.8148, 1.6987, 1.2123],
         [1.2292, 1.9980, 1.9290]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[1.8286, 1.7842, 1.1345]],

        [[1.4894, 1.6708, 1.3640]],

        [[1.8371, 1.1287, 1.2263]],

       ...]],

        [[1.0952, 1.8119, 1.0144]],

        [[1.8524, 1.0381, 1.4766]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 3, 3]), e shape: torch.Size([100, 1, 3])
SpMM(message func: add, reduce func: min)
______________________ test_spmm[idtype1-min-add-shp1-g1] ______________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3, 3), (1, 3)), msg = 'add', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add', reduce_op = 'min'
lhs_data = tensor([[[1.3584, 1.6938, 1.1930],
         [1.5138, 1.7168, 1.3783],
         [1.7482, 1.2080, 1.8110]],

        [[1...3464],
         [1.2629, 1.9246, 1.9509],
         [1.7930, 1.5335, 1.4229]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[1.9488, 1.1751, 1.9396]],

        [[1.9080, 1.8904, 1.9726]],

        [[1.9493, 1.7334, 1.0078]],

       ...]],

        [[1.8341, 1.2008, 1.5922]],

        [[1.8461, 1.3156, 1.0577]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 3, 3]), e shape: torch.Size([300, 1, 3])
SpMM(message func: add, reduce func: min)
______________________ test_spmm[idtype1-min-add-shp2-g0] ______________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float64)})
shp = ((1,), (3,)), msg = 'add', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add', reduce_op = 'min'
lhs_data = tensor([[1.2949],
        [1.1805],
        [1.3332],
        [1.6845],
        [1.5164],
        [1.2397],
        [1...785],
        [1.9539],
        [1.7731],
        [1.3985],
        [1.0774]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.7068, 1.3119, 1.6463],
        [1.3915, 1.2731, 1.3493],
        [1.9395, 1.5912, 1.3531],
        [1.5068,... 1.8494],
        [1.3636, 1.5990, 1.3398],
        [1.5070, 1.6512, 1.1571]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1, 3), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 1]), e shape: torch.Size([100, 3])
SpMM(message func: add, reduce func: min)
______________________ test_spmm[idtype1-min-add-shp2-g1] ______________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (3,)), msg = 'add', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add', reduce_op = 'min'
lhs_data = tensor([[1.2428],
        [1.0831],
        [1.0862],
        [1.0935],
        [1.9921],
        [1.4539],
        [1...339],
        [1.4943],
        [1.2691],
        [1.6382],
        [1.3724]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.9813, 1.6664, 1.9832],
        [1.2013, 1.2083, 1.1618],
        [1.4904, 1.2059, 1.3680],
        [1.4474,... 1.8487],
        [1.1080, 1.8423, 1.4414],
        [1.6413, 1.9538, 1.5092]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 1]), e shape: torch.Size([300, 3])
SpMM(message func: add, reduce func: min)
______________________ test_spmm[idtype1-min-add-shp3-g0] ______________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (1,)), msg = 'add', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add', reduce_op = 'min'
lhs_data = tensor([[1.2799, 1.8299, 1.4178],
        [1.2259, 1.4449, 1.5913],
        [1.8020, 1.8521, 1.9351],
        [1.3948,... 1.2391],
        [1.6026, 1.1393, 1.0473],
        [1.7296, 1.6818, 1.4620]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.5353],
        [1.7216],
        [1.5549],
        [1.6380],
        [1.4469],
        [1.3115],
        [1...266],
        [1.0584],
        [1.3803],
        [1.7577],
        [1.4740]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 3]), e shape: torch.Size([100, 1])
SpMM(message func: add, reduce func: min)
______________________ test_spmm[idtype1-min-add-shp3-g1] ______________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (1,)), msg = 'add', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add', reduce_op = 'min'
lhs_data = tensor([[1.4582, 1.9538, 1.9882],
        [1.1622, 1.4435, 1.5964],
        [1.3660, 1.6947, 1.4631],
        [1.1534,... 1.2735],
        [1.5745, 1.0938, 1.3485],
        [1.3410, 1.0639, 1.5164]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.6643],
        [1.7597],
        [1.8893],
        [1.7508],
        [1.0303],
        [1.2661],
        [1...244],
        [1.7019],
        [1.5782],
        [1.2326],
        [1.5855]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 3]), e shape: torch.Size([300, 1])
SpMM(message func: add, reduce func: min)
______________________ test_spmm[idtype1-min-add-shp4-g0] ______________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), msg = 'add', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add', reduce_op = 'min'
lhs_data = tensor([[1.5603],
        [1.2375],
        [1.7112],
        [1.0302],
        [1.0684],
        [1.4512],
        [1...580],
        [1.4283],
        [1.2597],
        [1.8296],
        [1.8174]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.9682],
        [1.8133],
        [1.4432],
        [1.1159],
        [1.1753],
        [1.9987],
        [1...681],
        [1.9563],
        [1.9551],
        [1.8359],
        [1.1028]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 1]), e shape: torch.Size([100, 1])
SpMM(message func: add, reduce func: min)
______________________ test_spmm[idtype1-min-add-shp4-g1] ______________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), msg = 'add', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add', reduce_op = 'min'
lhs_data = tensor([[1.8077],
        [1.4855],
        [1.4417],
        [1.4461],
        [1.3657],
        [1.1106],
        [1...838],
        [1.9480],
        [1.6002],
        [1.3375],
        [1.8298]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.8167],
        [1.6060],
        [1.0002],
        [1.5336],
        [1.6964],
        [1.4879],
        [1...632],
        [1.6280],
        [1.6335],
        [1.7881],
        [1.6955]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 1]), e shape: torch.Size([300, 1])
SpMM(message func: add, reduce func: min)
______________________ test_spmm[idtype1-min-add-shp5-g0] ______________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
shp = ((), ()), msg = 'add', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add', reduce_op = 'min'
lhs_data = tensor([1.9882, 1.3705, 1.9824, 1.2338, 1.4456, 1.6783, 1.7172, 1.3214, 1.1105,
        1.1770, 1.9605, 1.6384, 1.8654...2338, 1.2762, 1.8317, 1.8304, 1.8710, 1.2585,
        1.4213, 1.8213, 1.9976], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([1.3744, 1.2163, 1.8413, 1.6030, 1.8236, 1.0125, 1.6684, 1.8377, 1.1318,
        1.5802, 1.1978, 1.9531, 1.2199...3969, 1.5985, 1.9951, 1.9629, 1.5330, 1.2488, 1.2144, 1.5143,
        1.8740], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30]), e shape: torch.Size([100])
SpMM(message func: add, reduce func: min)
______________________ test_spmm[idtype1-min-add-shp5-g1] ______________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((), ()), msg = 'add', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add', reduce_op = 'min'
lhs_data = tensor([1.4976, 1.8731, 1.4528, 1.7531, 1.9699, 1.5962, 1.3387, 1.7555, 1.4868,
        1.0324, 1.3024, 1.1484, 1.8398...3719, 1.1070, 1.2897, 1.7213, 1.1287, 1.8756,
        1.2568, 1.4970, 1.6760], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([1.8489, 1.6336, 1.0979, 1.1728, 1.7234, 1.3679, 1.8790, 1.5064, 1.4108,
        1.2337, 1.6397, 1.2184, 1.8209...2060, 1.0992, 1.0982, 1.8601, 1.8337, 1.1286,
        1.1324, 1.0777, 1.5493], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30]), e shape: torch.Size([300])
SpMM(message func: add, reduce func: min)
______________________ test_spmm[idtype1-min-sub-shp0-g0] ______________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), msg = 'sub', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add', reduce_op = 'min'
lhs_data = tensor([[[[[[1.0954],
            [1.1833],
            [1.5109]]],


          [[[1.1341],
            [1.2921],
    ...2]]],


          [[[1.6139],
            [1.1328],
            [1.7618]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[-1.6133]],

           [[-1.8201]],

           [[-1.2430]]]],



         [[[[-1.0244]],

           [[-...       [[[[-1.7729]],

           [[-1.7789]],

           [[-1.4292]]]]]], dtype=torch.float64, grad_fn=<NegBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 1, 2, 1, 3, 1]), e shape: torch.Size([100, 4, 1, 3, 1, 1])
SpMM(message func: sub, reduce func: min)
______________________ test_spmm[idtype1-min-sub-shp0-g1] ______________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), msg = 'sub', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add', reduce_op = 'min'
lhs_data = tensor([[[[[[1.3503],
            [1.5213],
            [1.7971]]],


          [[[1.5618],
            [1.1607],
    ...9]]],


          [[[1.6969],
            [1.0352],
            [1.3095]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[-1.1252]],

           [[-1.8176]],

           [[-1.8519]]]],



         [[[[-1.2481]],

           [[-...       [[[[-1.7766]],

           [[-1.6847]],

           [[-1.4375]]]]]], dtype=torch.float64, grad_fn=<NegBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 1, 2, 1, 3, 1]), e shape: torch.Size([300, 4, 1, 3, 1, 1])
SpMM(message func: sub, reduce func: min)
______________________ test_spmm[idtype1-min-sub-shp1-g0] ______________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1, 3), dtype=torch.float64)})
shp = ((3, 3), (1, 3)), msg = 'sub', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add', reduce_op = 'min'
lhs_data = tensor([[[1.2520, 1.2700, 1.6748],
         [1.5990, 1.6844, 1.6038],
         [1.6880, 1.8502, 1.9459]],

        [[1...4010],
         [1.8764, 1.5912, 1.7242],
         [1.7740, 1.8834, 1.0525]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[-1.2534, -1.3245, -1.6296]],

        [[-1.0119, -1.6818, -1.4600]],

        [[-1.4929, -1.2280, -1.8037]],...1.0317, -1.3838, -1.4002]],

        [[-1.5361, -1.1120, -1.1419]]], dtype=torch.float64,
       grad_fn=<NegBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 3, 3]), e shape: torch.Size([100, 1, 3])
SpMM(message func: sub, reduce func: min)
______________________ test_spmm[idtype1-min-sub-shp1-g1] ______________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3, 3), (1, 3)), msg = 'sub', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add', reduce_op = 'min'
lhs_data = tensor([[[1.1037, 1.4217, 1.7989],
         [1.8899, 1.3542, 1.0560],
         [1.3775, 1.0741, 1.5777]],

        [[1...2072],
         [1.2889, 1.9613, 1.1899],
         [1.7151, 1.3571, 1.0588]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[-1.9220, -1.1152, -1.9870]],

        [[-1.4924, -1.3643, -1.2723]],

        [[-1.5560, -1.4278, -1.8244]],...1.9944, -1.3685, -1.0937]],

        [[-1.6366, -1.0992, -1.7066]]], dtype=torch.float64,
       grad_fn=<NegBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 3, 3]), e shape: torch.Size([300, 1, 3])
SpMM(message func: sub, reduce func: min)
______________________ test_spmm[idtype1-min-sub-shp2-g0] ______________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float64)})
shp = ((1,), (3,)), msg = 'sub', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add', reduce_op = 'min'
lhs_data = tensor([[1.8432],
        [1.7917],
        [1.0293],
        [1.6452],
        [1.1834],
        [1.4613],
        [1...152],
        [1.0864],
        [1.1842],
        [1.4948],
        [1.2876]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.7504, -1.7719, -1.8410],
        [-1.5272, -1.4582, -1.0290],
        [-1.3695, -1.8879, -1.2959],
       ...  [-1.4170, -1.5573, -1.7634],
        [-1.4727, -1.0040, -1.3776]], dtype=torch.float64,
       grad_fn=<NegBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1, 3), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 1]), e shape: torch.Size([100, 3])
SpMM(message func: sub, reduce func: min)
______________________ test_spmm[idtype1-min-sub-shp2-g1] ______________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (3,)), msg = 'sub', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add', reduce_op = 'min'
lhs_data = tensor([[1.6009],
        [1.7284],
        [1.5734],
        [1.8791],
        [1.7574],
        [1.0237],
        [1...833],
        [1.0924],
        [1.1716],
        [1.5934],
        [1.0635]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.9924, -1.8616, -1.9345],
        [-1.7263, -1.2404, -1.2162],
        [-1.8577, -1.2826, -1.3843],
       ...  [-1.0400, -1.8275, -1.6731],
        [-1.3129, -1.1464, -1.0631]], dtype=torch.float64,
       grad_fn=<NegBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 1]), e shape: torch.Size([300, 3])
SpMM(message func: sub, reduce func: min)
______________________ test_spmm[idtype1-min-sub-shp3-g0] ______________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (1,)), msg = 'sub', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add', reduce_op = 'min'
lhs_data = tensor([[1.9066, 1.3944, 1.3750],
        [1.4626, 1.4326, 1.2302],
        [1.6360, 1.3427, 1.3476],
        [1.2541,... 1.0103],
        [1.7177, 1.8876, 1.8550],
        [1.2694, 1.9259, 1.6674]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.9370],
        [-1.7052],
        [-1.6200],
        [-1.5079],
        [-1.4864],
        [-1.8032],
    ...       [-1.5731],
        [-1.3560],
        [-1.9193],
        [-1.2528]], dtype=torch.float64, grad_fn=<NegBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 3]), e shape: torch.Size([100, 1])
SpMM(message func: sub, reduce func: min)
______________________ test_spmm[idtype1-min-sub-shp3-g1] ______________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (1,)), msg = 'sub', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add', reduce_op = 'min'
lhs_data = tensor([[1.0327, 1.4269, 1.8122],
        [1.8974, 1.7307, 1.9873],
        [1.7264, 1.1628, 1.3787],
        [1.2209,... 1.4145],
        [1.8271, 1.8953, 1.1488],
        [1.3180, 1.2232, 1.5065]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.4551],
        [-1.7746],
        [-1.2859],
        [-1.0365],
        [-1.5731],
        [-1.2606],
    ...       [-1.2583],
        [-1.0579],
        [-1.5285],
        [-1.7018]], dtype=torch.float64, grad_fn=<NegBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 3]), e shape: torch.Size([300, 1])
SpMM(message func: sub, reduce func: min)
______________________ test_spmm[idtype1-min-sub-shp4-g0] ______________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), msg = 'sub', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add', reduce_op = 'min'
lhs_data = tensor([[1.8778],
        [1.7210],
        [1.3519],
        [1.3545],
        [1.8002],
        [1.2613],
        [1...530],
        [1.3335],
        [1.0403],
        [1.6029],
        [1.5227]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.9884],
        [-1.7386],
        [-1.8216],
        [-1.0185],
        [-1.6740],
        [-1.4735],
    ...       [-1.1330],
        [-1.1535],
        [-1.0121],
        [-1.7835]], dtype=torch.float64, grad_fn=<NegBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 1]), e shape: torch.Size([100, 1])
SpMM(message func: sub, reduce func: min)
______________________ test_spmm[idtype1-min-sub-shp4-g1] ______________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), msg = 'sub', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add', reduce_op = 'min'
lhs_data = tensor([[1.9986],
        [1.8125],
        [1.3348],
        [1.7963],
        [1.1611],
        [1.7911],
        [1...939],
        [1.4351],
        [1.1083],
        [1.2284],
        [1.3573]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.4144],
        [-1.0371],
        [-1.1180],
        [-1.3692],
        [-1.5780],
        [-1.9155],
    ...       [-1.1991],
        [-1.8910],
        [-1.0555],
        [-1.0572]], dtype=torch.float64, grad_fn=<NegBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 1]), e shape: torch.Size([300, 1])
SpMM(message func: sub, reduce func: min)
______________________ test_spmm[idtype1-min-sub-shp5-g0] ______________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
shp = ((), ()), msg = 'sub', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add', reduce_op = 'min'
lhs_data = tensor([1.1461, 1.1765, 1.2557, 1.6465, 1.4332, 1.7700, 1.4815, 1.3497, 1.5859,
        1.5385, 1.1634, 1.9201, 1.4048...7537, 1.2347, 1.3305, 1.6345, 1.3430, 1.4267,
        1.8037, 1.6086, 1.8527], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([-1.9981, -1.9564, -1.6551, -1.6568, -1.8752, -1.4625, -1.9855, -1.5691,
        -1.8962, -1.7212, -1.6127, -1....9247, -1.2235, -1.1584,
        -1.6070, -1.8529, -1.7784, -1.1181], dtype=torch.float64,
       grad_fn=<NegBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30]), e shape: torch.Size([100])
SpMM(message func: sub, reduce func: min)
______________________ test_spmm[idtype1-min-sub-shp5-g1] ______________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((), ()), msg = 'sub', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add', reduce_op = 'min'
lhs_data = tensor([1.7750, 1.9395, 1.1849, 1.6844, 1.3380, 1.6133, 1.9441, 1.6864, 1.0906,
        1.9165, 1.1398, 1.0766, 1.9080...7926, 1.0838, 1.0041, 1.3682, 1.3248, 1.4277,
        1.0329, 1.6358, 1.1179], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([-1.1747, -1.8830, -1.1794, -1.8805, -1.6203, -1.2063, -1.4479, -1.3211,
        -1.3823, -1.0206, -1.2670, -1....7043, -1.5902, -1.3293,
        -1.5698, -1.2277, -1.9291, -1.9564], dtype=torch.float64,
       grad_fn=<NegBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30]), e shape: torch.Size([300])
SpMM(message func: sub, reduce func: min)
______________________ test_spmm[idtype1-min-mul-shp0-g0] ______________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), msg = 'mul', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul', reduce_op = 'min'
lhs_data = tensor([[[[[[1.5028],
            [1.1395],
            [1.5477]]],


          [[[1.5039],
            [1.7530],
    ...0]]],


          [[[1.5772],
            [1.0942],
            [1.8468]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.9255]],

           [[1.2015]],

           [[1.2636]]]],



         [[[[1.8007]],

           [[1.659...



         [[[[1.2962]],

           [[1.1973]],

           [[1.7227]]]]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 1, 2, 1, 3, 1]), e shape: torch.Size([100, 4, 1, 3, 1, 1])
SpMM(message func: mul, reduce func: min)
______________________ test_spmm[idtype1-min-mul-shp0-g1] ______________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), msg = 'mul', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul', reduce_op = 'min'
lhs_data = tensor([[[[[[1.7696],
            [1.5711],
            [1.8251]]],


          [[[1.7042],
            [1.4652],
    ...5]]],


          [[[1.7939],
            [1.2731],
            [1.4290]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.1830]],

           [[1.9736]],

           [[1.0368]]]],



         [[[[1.8312]],

           [[1.312...



         [[[[1.2439]],

           [[1.7455]],

           [[1.2490]]]]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 1, 2, 1, 3, 1]), e shape: torch.Size([300, 4, 1, 3, 1, 1])
SpMM(message func: mul, reduce func: min)
______________________ test_spmm[idtype1-min-mul-shp1-g0] ______________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1, 3), dtype=torch.float64)})
shp = ((3, 3), (1, 3)), msg = 'mul', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul', reduce_op = 'min'
lhs_data = tensor([[[1.1308, 1.6524, 1.4822],
         [1.9385, 1.1816, 1.8033],
         [1.2791, 1.3926, 1.7491]],

        [[1...3653],
         [1.4078, 1.8104, 1.5847],
         [1.9114, 1.4403, 1.1345]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[1.5971, 1.5217, 1.3460]],

        [[1.0293, 1.8709, 1.9116]],

        [[1.4942, 1.9856, 1.3570]],

       ...]],

        [[1.5107, 1.0673, 1.4484]],

        [[1.9973, 1.3712, 1.6643]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 3, 3]), e shape: torch.Size([100, 1, 3])
SpMM(message func: mul, reduce func: min)
______________________ test_spmm[idtype1-min-mul-shp1-g1] ______________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3, 3), (1, 3)), msg = 'mul', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul', reduce_op = 'min'
lhs_data = tensor([[[1.6581, 1.9723, 1.9719],
         [1.2104, 1.1670, 1.3672],
         [1.6064, 1.9390, 1.8737]],

        [[1...5840],
         [1.7983, 1.8366, 1.0651],
         [1.9240, 1.3286, 1.9550]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[1.5891, 1.0498, 1.3713]],

        [[1.6269, 1.8040, 1.3283]],

        [[1.4566, 1.3452, 1.3013]],

       ...]],

        [[1.4412, 1.6571, 1.5188]],

        [[1.4223, 1.2200, 1.5114]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 3, 3]), e shape: torch.Size([300, 1, 3])
SpMM(message func: mul, reduce func: min)
______________________ test_spmm[idtype1-min-mul-shp2-g0] ______________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float64)})
shp = ((1,), (3,)), msg = 'mul', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul', reduce_op = 'min'
lhs_data = tensor([[1.2910],
        [1.9446],
        [1.0622],
        [1.2689],
        [1.5834],
        [1.0667],
        [1...937],
        [1.9719],
        [1.0118],
        [1.7459],
        [1.9834]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.3767, 1.5692, 1.8267],
        [1.5609, 1.2438, 1.3664],
        [1.6017, 1.1091, 1.9430],
        [1.1032,... 1.0301],
        [1.0407, 1.7588, 1.3801],
        [1.1191, 1.8753, 1.7560]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1, 3), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 1]), e shape: torch.Size([100, 3])
SpMM(message func: mul, reduce func: min)
______________________ test_spmm[idtype1-min-mul-shp2-g1] ______________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (3,)), msg = 'mul', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul', reduce_op = 'min'
lhs_data = tensor([[1.2758],
        [1.7173],
        [1.5135],
        [1.0006],
        [1.5529],
        [1.0452],
        [1...703],
        [1.8801],
        [1.9652],
        [1.3973],
        [1.5315]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.9544, 1.8540, 1.0481],
        [1.9131, 1.8163, 1.2943],
        [1.3622, 1.5772, 1.5474],
        [1.1681,... 1.6377],
        [1.3226, 1.0227, 1.2294],
        [1.7265, 1.7555, 1.5638]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 1]), e shape: torch.Size([300, 3])
SpMM(message func: mul, reduce func: min)
______________________ test_spmm[idtype1-min-mul-shp3-g0] ______________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (1,)), msg = 'mul', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul', reduce_op = 'min'
lhs_data = tensor([[1.3848, 1.6472, 1.4687],
        [1.9921, 1.9838, 1.7610],
        [1.3150, 1.1022, 1.5023],
        [1.8276,... 1.0122],
        [1.6329, 1.0970, 1.7823],
        [1.2344, 1.7628, 1.7074]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.7315],
        [1.8633],
        [1.4294],
        [1.1130],
        [1.1697],
        [1.8808],
        [1...653],
        [1.7885],
        [1.6456],
        [1.1009],
        [1.2342]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 3]), e shape: torch.Size([100, 1])
SpMM(message func: mul, reduce func: min)
______________________ test_spmm[idtype1-min-mul-shp3-g1] ______________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (1,)), msg = 'mul', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul', reduce_op = 'min'
lhs_data = tensor([[1.1959, 1.1775, 1.1123],
        [1.4725, 1.0311, 1.1489],
        [1.4850, 1.1660, 1.1101],
        [1.6711,... 1.8470],
        [1.5391, 1.9659, 1.4608],
        [1.2370, 1.2580, 1.2714]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.0309],
        [1.8841],
        [1.1403],
        [1.0350],
        [1.8449],
        [1.0440],
        [1...235],
        [1.5192],
        [1.5290],
        [1.4279],
        [1.7869]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 3]), e shape: torch.Size([300, 1])
SpMM(message func: mul, reduce func: min)
______________________ test_spmm[idtype1-min-mul-shp4-g0] ______________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), msg = 'mul', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul', reduce_op = 'min'
lhs_data = tensor([[1.3519],
        [1.3939],
        [1.6735],
        [1.2460],
        [1.9999],
        [1.8093],
        [1...551],
        [1.5808],
        [1.0612],
        [1.2360],
        [1.6507]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.1287],
        [1.6548],
        [1.6422],
        [1.9908],
        [1.9935],
        [1.8444],
        [1...851],
        [1.1661],
        [1.9893],
        [1.8229],
        [1.7888]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 1]), e shape: torch.Size([100, 1])
SpMM(message func: mul, reduce func: min)
______________________ test_spmm[idtype1-min-mul-shp4-g1] ______________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), msg = 'mul', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul', reduce_op = 'min'
lhs_data = tensor([[1.1065],
        [1.1075],
        [1.1215],
        [1.1859],
        [1.2554],
        [1.3146],
        [1...994],
        [1.4488],
        [1.6995],
        [1.6901],
        [1.4876]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.6798],
        [1.8731],
        [1.0956],
        [1.6834],
        [1.1068],
        [1.1642],
        [1...998],
        [1.9732],
        [1.8178],
        [1.2184],
        [1.3875]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 1]), e shape: torch.Size([300, 1])
SpMM(message func: mul, reduce func: min)
______________________ test_spmm[idtype1-min-mul-shp5-g0] ______________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
shp = ((), ()), msg = 'mul', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul', reduce_op = 'min'
lhs_data = tensor([1.6063, 1.5314, 1.0479, 1.9152, 1.8522, 1.3913, 1.7159, 1.5087, 1.8078,
        1.8172, 1.8817, 1.4343, 1.6147...1098, 1.1022, 1.1622, 1.2310, 1.3203, 1.8717,
        1.3150, 1.7530, 1.4576], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([1.8835, 1.1938, 1.0169, 1.6819, 1.4608, 1.9024, 1.6934, 1.9831, 1.9746,
        1.1744, 1.8635, 1.4116, 1.4612...2194, 1.7818, 1.6285, 1.5893, 1.7531, 1.0942, 1.8572, 1.7827,
        1.7678], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30]), e shape: torch.Size([100])
SpMM(message func: mul, reduce func: min)
______________________ test_spmm[idtype1-min-mul-shp5-g1] ______________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((), ()), msg = 'mul', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul', reduce_op = 'min'
lhs_data = tensor([1.5190, 1.1598, 1.2749, 1.9526, 1.3789, 1.6662, 1.2876, 1.9316, 1.2347,
        1.6927, 1.2405, 1.4224, 1.9774...8699, 1.6683, 1.1561, 1.4240, 1.0873, 1.9789,
        1.4701, 1.7918, 1.3935], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([1.2095, 1.8841, 1.5353, 1.9424, 1.9198, 1.9943, 1.7078, 1.0867, 1.5494,
        1.5239, 1.6607, 1.1116, 1.6259...6471, 1.4871, 1.9793, 1.2255, 1.6120, 1.4764,
        1.7329, 1.9199, 1.4729], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30]), e shape: torch.Size([300])
SpMM(message func: mul, reduce func: min)
______________________ test_spmm[idtype1-min-div-shp0-g0] ______________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), msg = 'div', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul', reduce_op = 'min'
lhs_data = tensor([[[[[[1.5782],
            [1.4698],
            [1.4654]]],


          [[[1.9503],
            [1.2268],
    ...0]]],


          [[[1.2978],
            [1.9345],
            [1.8658]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[0.5076]],

           [[0.6682]],

           [[0.5465]]]],



         [[[[0.5043]],

           [[0.673...         [[[[0.5025]],

           [[0.8602]],

           [[0.6039]]]]]], dtype=torch.float64, grad_fn=<MulBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 1, 2, 1, 3, 1]), e shape: torch.Size([100, 4, 1, 3, 1, 1])
SpMM(message func: div, reduce func: min)
______________________ test_spmm[idtype1-min-div-shp0-g1] ______________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), msg = 'div', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul', reduce_op = 'min'
lhs_data = tensor([[[[[[1.1794],
            [1.1806],
            [1.0298]]],


          [[[1.2813],
            [1.4987],
    ...2]]],


          [[[1.6214],
            [1.6495],
            [1.8810]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[0.9996]],

           [[0.6779]],

           [[0.5509]]]],



         [[[[0.6312]],

           [[0.523...         [[[[0.5449]],

           [[0.6950]],

           [[0.8257]]]]]], dtype=torch.float64, grad_fn=<MulBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 1, 2, 1, 3, 1]), e shape: torch.Size([300, 4, 1, 3, 1, 1])
SpMM(message func: div, reduce func: min)
______________________ test_spmm[idtype1-min-div-shp1-g0] ______________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1, 3), dtype=torch.float64)})
shp = ((3, 3), (1, 3)), msg = 'div', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul', reduce_op = 'min'
lhs_data = tensor([[[1.4417, 1.8383, 1.2883],
         [1.4916, 1.8323, 1.3929],
         [1.0946, 1.4812, 1.1570]],

        [[1...8331],
         [1.9333, 1.7136, 1.5581],
         [1.0500, 1.4380, 1.3367]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[0.5832, 0.5245, 0.9123]],

        [[0.5833, 0.6744, 0.6416]],

        [[0.7159, 0.6909, 0.5235]],

       ...  [[0.9237, 0.8273, 0.9850]],

        [[0.9937, 0.7834, 0.5746]]], dtype=torch.float64,
       grad_fn=<MulBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 3, 3]), e shape: torch.Size([100, 1, 3])
SpMM(message func: div, reduce func: min)
______________________ test_spmm[idtype1-min-div-shp1-g1] ______________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3, 3), (1, 3)), msg = 'div', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul', reduce_op = 'min'
lhs_data = tensor([[[1.7867, 1.5842, 1.7257],
         [1.4484, 1.8710, 1.6905],
         [1.7836, 1.7616, 1.6451]],

        [[1...9377],
         [1.4128, 1.3771, 1.7721],
         [1.8890, 1.4450, 1.8662]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[0.8494, 0.7951, 0.9046]],

        [[0.6853, 0.5082, 0.9857]],

        [[0.6279, 0.9035, 0.7032]],

       ...  [[0.7108, 0.8735, 0.5000]],

        [[0.6131, 0.6485, 0.5545]]], dtype=torch.float64,
       grad_fn=<MulBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 3, 3]), e shape: torch.Size([300, 1, 3])
SpMM(message func: div, reduce func: min)
______________________ test_spmm[idtype1-min-div-shp2-g0] ______________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float64)})
shp = ((1,), (3,)), msg = 'div', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul', reduce_op = 'min'
lhs_data = tensor([[1.9490],
        [1.0723],
        [1.3938],
        [1.2146],
        [1.9175],
        [1.6947],
        [1...148],
        [1.8243],
        [1.6763],
        [1.4103],
        [1.3015]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.5887, 0.6614, 0.7435],
        [0.6677, 0.5864, 0.7458],
        [0.7118, 0.5980, 0.8889],
        [0.8137,...441],
        [0.6077, 0.6446, 0.7506],
        [0.5549, 0.5672, 0.6855]], dtype=torch.float64, grad_fn=<MulBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1, 3), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 1]), e shape: torch.Size([100, 3])
SpMM(message func: div, reduce func: min)
______________________ test_spmm[idtype1-min-div-shp2-g1] ______________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (3,)), msg = 'div', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul', reduce_op = 'min'
lhs_data = tensor([[1.9449],
        [1.9702],
        [1.2595],
        [1.6013],
        [1.1717],
        [1.3144],
        [1...692],
        [1.9622],
        [1.3183],
        [1.1681],
        [1.1738]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.7630, 0.6112, 0.7536],
        [0.5253, 0.7900, 0.7866],
        [0.7601, 0.6238, 0.5602],
        [0.6543,...440],
        [0.7512, 0.5970, 0.8155],
        [0.7143, 0.9618, 0.8482]], dtype=torch.float64, grad_fn=<MulBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 1]), e shape: torch.Size([300, 3])
SpMM(message func: div, reduce func: min)
______________________ test_spmm[idtype1-min-div-shp3-g0] ______________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (1,)), msg = 'div', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul', reduce_op = 'min'
lhs_data = tensor([[1.1642, 1.8647, 1.9574],
        [1.0845, 1.2068, 1.9841],
        [1.2847, 1.6358, 1.9695],
        [1.5949,... 1.1331],
        [1.7113, 1.0480, 1.5493],
        [1.7062, 1.8005, 1.3184]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.8391],
        [0.5738],
        [0.5847],
        [0.5069],
        [0.8383],
        [0.5038],
        [0...,
        [0.8779],
        [0.9443],
        [0.9161],
        [0.7172]], dtype=torch.float64, grad_fn=<MulBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 3]), e shape: torch.Size([100, 1])
SpMM(message func: div, reduce func: min)
______________________ test_spmm[idtype1-min-div-shp3-g1] ______________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (1,)), msg = 'div', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul', reduce_op = 'min'
lhs_data = tensor([[1.3692, 1.2088, 1.0534],
        [1.7448, 1.2694, 1.4594],
        [1.3351, 1.8891, 1.1808],
        [1.9347,... 1.9258],
        [1.1573, 1.2319, 1.0810],
        [1.3283, 1.6375, 1.7773]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.7337],
        [0.5192],
        [0.6752],
        [0.8047],
        [0.5161],
        [0.5581],
        [0...,
        [0.8635],
        [0.9713],
        [0.5685],
        [0.6726]], dtype=torch.float64, grad_fn=<MulBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 3]), e shape: torch.Size([300, 1])
SpMM(message func: div, reduce func: min)
______________________ test_spmm[idtype1-min-div-shp4-g0] ______________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), msg = 'div', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul', reduce_op = 'min'
lhs_data = tensor([[1.0189],
        [1.3910],
        [1.0815],
        [1.8586],
        [1.8685],
        [1.7798],
        [1...670],
        [1.7110],
        [1.0603],
        [1.8084],
        [1.0530]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.8755],
        [0.7561],
        [0.7670],
        [0.7628],
        [0.5351],
        [0.7238],
        [0...,
        [0.5004],
        [0.6089],
        [0.5264],
        [0.5264]], dtype=torch.float64, grad_fn=<MulBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 1]), e shape: torch.Size([100, 1])
SpMM(message func: div, reduce func: min)
______________________ test_spmm[idtype1-min-div-shp4-g1] ______________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), msg = 'div', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul', reduce_op = 'min'
lhs_data = tensor([[1.0562],
        [1.6118],
        [1.8316],
        [1.7364],
        [1.3097],
        [1.4181],
        [1...268],
        [1.9401],
        [1.0951],
        [1.5323],
        [1.1096]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.5730],
        [0.8823],
        [0.5160],
        [0.5215],
        [0.9520],
        [0.6194],
        [0...,
        [0.5722],
        [0.8919],
        [0.6950],
        [0.5551]], dtype=torch.float64, grad_fn=<MulBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 1]), e shape: torch.Size([300, 1])
SpMM(message func: div, reduce func: min)
______________________ test_spmm[idtype1-min-div-shp5-g0] ______________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
shp = ((), ()), msg = 'div', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul', reduce_op = 'min'
lhs_data = tensor([1.4553, 1.6450, 1.0299, 1.8166, 1.3746, 1.4226, 1.1231, 1.2326, 1.6206,
        1.0592, 1.9474, 1.0576, 1.6390...8156, 1.4985, 1.9339, 1.8782, 1.3323, 1.5676,
        1.2036, 1.8896, 1.0610], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([0.9678, 0.9457, 0.9971, 0.5159, 0.8650, 0.5159, 0.8171, 0.7252, 0.5386,
        0.8242, 0.7588, 0.9079, 0.5681..., 0.6819, 0.5668, 0.8617, 0.9034, 0.5388, 0.5477, 0.6471,
        0.6079], dtype=torch.float64, grad_fn=<MulBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30]), e shape: torch.Size([100])
SpMM(message func: div, reduce func: min)
______________________ test_spmm[idtype1-min-div-shp5-g1] ______________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((), ()), msg = 'div', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul', reduce_op = 'min'
lhs_data = tensor([1.8464, 1.5417, 1.7753, 1.3473, 1.0474, 1.4243, 1.6334, 1.3641, 1.9769,
        1.2155, 1.5389, 1.0582, 1.5559...3749, 1.9029, 1.0000, 1.4540, 1.4167, 1.1120,
        1.9240, 1.2361, 1.8058], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([0.8875, 0.7632, 0.7632, 0.6213, 0.7335, 0.5304, 0.5905, 0.8545, 0.5260,
        0.5320, 0.5892, 0.5295, 0.5878..., 0.7336, 0.6220, 0.8466, 0.5988, 0.9230,
        0.8126, 0.7073, 0.6340], dtype=torch.float64, grad_fn=<MulBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30]), e shape: torch.Size([300])
SpMM(message func: div, reduce func: min)
___________________ test_spmm[idtype1-min-copy_lhs-shp0-g0] ____________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), msg = 'copy_lhs', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_lhs', reduce_op = 'min'
lhs_data = tensor([[[[[[1.4698],
            [1.6758],
            [1.1701]]],


          [[[1.5883],
            [1.8868],
    ...3]]],


          [[[1.5831],
            [1.5050],
            [1.8720]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.3223]],

           [[1.4970]],

           [[1.3549]]]],



         [[[[1.5449]],

           [[1.151...



         [[[[1.0044]],

           [[1.2770]],

           [[1.2772]]]]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 1, 2, 1, 3, 1]), e shape: torch.Size([100, 4, 1, 3, 1, 1])
SpMM(message func: copy_lhs, reduce func: min)
___________________ test_spmm[idtype1-min-copy_lhs-shp0-g1] ____________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), msg = 'copy_lhs', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_lhs', reduce_op = 'min'
lhs_data = tensor([[[[[[1.0544],
            [1.4520],
            [1.2569]]],


          [[[1.7994],
            [1.4689],
    ...5]]],


          [[[1.3492],
            [1.6295],
            [1.0159]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.6464]],

           [[1.8576]],

           [[1.7414]]]],



         [[[[1.3481]],

           [[1.060...



         [[[[1.0482]],

           [[1.3678]],

           [[1.7029]]]]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 1, 2, 1, 3, 1]), e shape: torch.Size([300, 4, 1, 3, 1, 1])
SpMM(message func: copy_lhs, reduce func: min)
___________________ test_spmm[idtype1-min-copy_lhs-shp1-g0] ____________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1, 3), dtype=torch.float64)})
shp = ((3, 3), (1, 3)), msg = 'copy_lhs', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_lhs', reduce_op = 'min'
lhs_data = tensor([[[1.1401, 1.4461, 1.5517],
         [1.9527, 1.6932, 1.8780],
         [1.3393, 1.6551, 1.9997]],

        [[1...7618],
         [1.8550, 1.1264, 1.7317],
         [1.3374, 1.0422, 1.3387]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[1.3410, 1.8648, 1.2066]],

        [[1.8449, 1.2494, 1.1910]],

        [[1.7414, 1.3033, 1.7529]],

       ...]],

        [[1.2379, 1.5680, 1.5766]],

        [[1.0677, 1.9889, 1.4553]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 3, 3]), e shape: torch.Size([100, 1, 3])
SpMM(message func: copy_lhs, reduce func: min)
___________________ test_spmm[idtype1-min-copy_lhs-shp1-g1] ____________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3, 3), (1, 3)), msg = 'copy_lhs', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_lhs', reduce_op = 'min'
lhs_data = tensor([[[1.2264, 1.9157, 1.3594],
         [1.8637, 1.6513, 1.1963],
         [1.0207, 1.0042, 1.8475]],

        [[1...0310],
         [1.8902, 1.5384, 1.7677],
         [1.8900, 1.9841, 1.3078]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[1.3429, 1.0498, 1.6213]],

        [[1.7488, 1.8049, 1.6619]],

        [[1.3878, 1.5887, 1.0006]],

       ...]],

        [[1.6801, 1.5001, 1.5560]],

        [[1.2544, 1.4122, 1.4485]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 3, 3]), e shape: torch.Size([300, 1, 3])
SpMM(message func: copy_lhs, reduce func: min)
___________________ test_spmm[idtype1-min-copy_lhs-shp2-g0] ____________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float64)})
shp = ((1,), (3,)), msg = 'copy_lhs', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_lhs', reduce_op = 'min'
lhs_data = tensor([[1.1635],
        [1.7334],
        [1.1679],
        [1.9089],
        [1.6525],
        [1.4220],
        [1...709],
        [1.2079],
        [1.0054],
        [1.5926],
        [1.2578]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.1319, 1.1667, 1.8400],
        [1.0585, 1.7362, 1.5529],
        [1.9751, 1.2388, 1.1922],
        [1.1223,... 1.9834],
        [1.7421, 1.2118, 1.3424],
        [1.2630, 1.9982, 1.5790]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1, 3), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 1]), e shape: torch.Size([100, 3])
SpMM(message func: copy_lhs, reduce func: min)
___________________ test_spmm[idtype1-min-copy_lhs-shp2-g1] ____________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (3,)), msg = 'copy_lhs', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_lhs', reduce_op = 'min'
lhs_data = tensor([[1.0364],
        [1.1667],
        [1.3700],
        [1.8795],
        [1.9200],
        [1.4214],
        [1...306],
        [1.8933],
        [1.9417],
        [1.9415],
        [1.2614]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.2513, 1.5535, 1.1005],
        [1.0859, 1.2331, 1.6529],
        [1.4963, 1.6002, 1.8162],
        [1.1797,... 1.0111],
        [1.2107, 1.9713, 1.8421],
        [1.1012, 1.4994, 1.6605]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 1]), e shape: torch.Size([300, 3])
SpMM(message func: copy_lhs, reduce func: min)
___________________ test_spmm[idtype1-min-copy_lhs-shp3-g0] ____________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (1,)), msg = 'copy_lhs', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_lhs', reduce_op = 'min'
lhs_data = tensor([[1.7323, 1.5339, 1.0591],
        [1.6515, 1.9890, 1.4325],
        [1.3845, 1.1650, 1.5343],
        [1.5069,... 1.0488],
        [1.4560, 1.9273, 1.0440],
        [1.4523, 1.8733, 1.8973]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.2134],
        [1.2408],
        [1.7802],
        [1.3356],
        [1.6940],
        [1.6777],
        [1...775],
        [1.5110],
        [1.5774],
        [1.8248],
        [1.8826]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 3]), e shape: torch.Size([100, 1])
SpMM(message func: copy_lhs, reduce func: min)
___________________ test_spmm[idtype1-min-copy_lhs-shp3-g1] ____________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (1,)), msg = 'copy_lhs', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_lhs', reduce_op = 'min'
lhs_data = tensor([[1.0074, 1.5769, 1.1522],
        [1.9842, 1.7374, 1.3026],
        [1.6795, 1.2604, 1.8884],
        [1.1000,... 1.8732],
        [1.3409, 1.0075, 1.0909],
        [1.6210, 1.1251, 1.8986]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.1603],
        [1.7922],
        [1.6762],
        [1.8914],
        [1.5166],
        [1.1719],
        [1...832],
        [1.4251],
        [1.3700],
        [1.1555],
        [1.7818]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 3]), e shape: torch.Size([300, 1])
SpMM(message func: copy_lhs, reduce func: min)
___________________ test_spmm[idtype1-min-copy_lhs-shp4-g0] ____________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), msg = 'copy_lhs', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_lhs', reduce_op = 'min'
lhs_data = tensor([[1.4024],
        [1.8838],
        [1.8069],
        [1.0961],
        [1.9117],
        [1.6675],
        [1...136],
        [1.8810],
        [1.1095],
        [1.0717],
        [1.9764]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.3771],
        [1.6478],
        [1.4127],
        [1.9121],
        [1.8586],
        [1.9375],
        [1...305],
        [1.9958],
        [1.7706],
        [1.0423],
        [1.1884]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 1]), e shape: torch.Size([100, 1])
SpMM(message func: copy_lhs, reduce func: min)
___________________ test_spmm[idtype1-min-copy_lhs-shp4-g1] ____________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), msg = 'copy_lhs', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_lhs', reduce_op = 'min'
lhs_data = tensor([[1.0551],
        [1.1372],
        [1.3340],
        [1.8910],
        [1.7807],
        [1.7669],
        [1...464],
        [1.9775],
        [1.0631],
        [1.2254],
        [1.0852]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.2187],
        [1.2493],
        [1.0095],
        [1.2499],
        [1.2666],
        [1.6742],
        [1...942],
        [1.9027],
        [1.3516],
        [1.2667],
        [1.8667]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 1]), e shape: torch.Size([300, 1])
SpMM(message func: copy_lhs, reduce func: min)
___________________ test_spmm[idtype1-min-copy_lhs-shp5-g0] ____________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
shp = ((), ()), msg = 'copy_lhs', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_lhs', reduce_op = 'min'
lhs_data = tensor([1.3819, 1.3537, 1.8367, 1.6676, 1.6437, 1.4322, 1.4954, 1.5305, 1.7147,
        1.3316, 1.2755, 1.6280, 1.9083...5189, 1.2686, 1.7234, 1.4954, 1.0061, 1.6008,
        1.8153, 1.4176, 1.2287], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([1.3007, 1.1017, 1.1284, 1.6021, 1.7636, 1.3947, 1.7483, 1.6663, 1.7609,
        1.9515, 1.5901, 1.0143, 1.7956...2054, 1.9309, 1.7720, 1.2798, 1.0137, 1.4349, 1.5046, 1.7386,
        1.8266], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30]), e shape: torch.Size([100])
SpMM(message func: copy_lhs, reduce func: min)
___________________ test_spmm[idtype1-min-copy_lhs-shp5-g1] ____________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((), ()), msg = 'copy_lhs', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_lhs', reduce_op = 'min'
lhs_data = tensor([1.3845, 1.8064, 1.9886, 1.9422, 1.9741, 1.6275, 1.6192, 1.5651, 1.5417,
        1.9614, 1.2178, 1.0666, 1.4455...9363, 1.2494, 1.5982, 1.9690, 1.8919, 1.8301,
        1.7414, 1.9348, 1.0716], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([1.1570, 1.5223, 1.9285, 1.2977, 1.8203, 1.6550, 1.7038, 1.5375, 1.7615,
        1.1670, 1.8517, 1.5483, 1.6079...8330, 1.9910, 1.2502, 1.0829, 1.4734, 1.2155,
        1.8590, 1.5332, 1.5931], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30]), e shape: torch.Size([300])
SpMM(message func: copy_lhs, reduce func: min)
___________________ test_spmm[idtype1-min-copy_rhs-shp0-g0] ____________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), msg = 'copy_rhs', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_rhs', reduce_op = 'min'
lhs_data = tensor([[[[[[1.2629],
            [1.3935],
            [1.2400]]],


          [[[1.1028],
            [1.0576],
    ...1]]],


          [[[1.7065],
            [1.7581],
            [1.6051]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.4082]],

           [[1.9934]],

           [[1.8732]]]],



         [[[[1.3711]],

           [[1.538...



         [[[[1.3282]],

           [[1.0583]],

           [[1.6408]]]]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 1, 2, 1, 3, 1]), e shape: torch.Size([100, 4, 1, 3, 1, 1])
SpMM(message func: copy_rhs, reduce func: min)
___________________ test_spmm[idtype1-min-copy_rhs-shp0-g1] ____________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), msg = 'copy_rhs', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_rhs', reduce_op = 'min'
lhs_data = tensor([[[[[[1.7283],
            [1.5982],
            [1.6122]]],


          [[[1.8736],
            [1.2188],
    ...7]]],


          [[[1.0719],
            [1.9560],
            [1.7160]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.5582]],

           [[1.5179]],

           [[1.9930]]]],



         [[[[1.3210]],

           [[1.657...



         [[[[1.6504]],

           [[1.0596]],

           [[1.5476]]]]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 1, 2, 1, 3, 1]), e shape: torch.Size([300, 4, 1, 3, 1, 1])
SpMM(message func: copy_rhs, reduce func: min)
___________________ test_spmm[idtype1-min-copy_rhs-shp1-g0] ____________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1, 3), dtype=torch.float64)})
shp = ((3, 3), (1, 3)), msg = 'copy_rhs', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_rhs', reduce_op = 'min'
lhs_data = tensor([[[1.0852, 1.9055, 1.1564],
         [1.7019, 1.2669, 1.2236],
         [1.1772, 1.8350, 1.8998]],

        [[1...1747],
         [1.6866, 1.0194, 1.5073],
         [1.4242, 1.7979, 1.4916]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[1.1986, 1.1907, 1.2488]],

        [[1.2947, 1.9048, 1.4898]],

        [[1.4536, 1.8872, 1.0128]],

       ...]],

        [[1.7277, 1.5750, 1.4647]],

        [[1.2179, 1.8270, 1.5975]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 3, 3]), e shape: torch.Size([100, 1, 3])
SpMM(message func: copy_rhs, reduce func: min)
___________________ test_spmm[idtype1-min-copy_rhs-shp1-g1] ____________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3, 3), (1, 3)), msg = 'copy_rhs', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_rhs', reduce_op = 'min'
lhs_data = tensor([[[1.6813, 1.9052, 1.3397],
         [1.7971, 1.8565, 1.7826],
         [1.8265, 1.1161, 1.5932]],

        [[1...1293],
         [1.4141, 1.2444, 1.2820],
         [1.5430, 1.1822, 1.6299]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[1.2960, 1.1640, 1.5996]],

        [[1.7601, 1.7145, 1.7196]],

        [[1.4589, 1.3251, 1.7385]],

       ...]],

        [[1.2254, 1.2238, 1.9775]],

        [[1.5252, 1.6853, 1.0737]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 3, 3]), e shape: torch.Size([300, 1, 3])
SpMM(message func: copy_rhs, reduce func: min)
___________________ test_spmm[idtype1-min-copy_rhs-shp2-g0] ____________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float64)})
shp = ((1,), (3,)), msg = 'copy_rhs', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_rhs', reduce_op = 'min'
lhs_data = tensor([[1.7929],
        [1.8574],
        [1.3899],
        [1.2359],
        [1.3534],
        [1.2626],
        [1...573],
        [1.7743],
        [1.0658],
        [1.4861],
        [1.1029]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.8420, 1.9710, 1.1057],
        [1.4568, 1.8469, 1.7010],
        [1.4698, 1.8206, 1.4787],
        [1.7484,... 1.2614],
        [1.8696, 1.7088, 1.6043],
        [1.9656, 1.6855, 1.0892]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1, 3), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 1]), e shape: torch.Size([100, 3])
SpMM(message func: copy_rhs, reduce func: min)
___________________ test_spmm[idtype1-min-copy_rhs-shp2-g1] ____________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (3,)), msg = 'copy_rhs', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_rhs', reduce_op = 'min'
lhs_data = tensor([[1.5116],
        [1.4084],
        [1.3064],
        [1.9686],
        [1.9940],
        [1.1870],
        [1...809],
        [1.3358],
        [1.7935],
        [1.5248],
        [1.0246]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.8831, 1.7231, 1.4364],
        [1.0809, 1.0974, 1.3198],
        [1.2490, 1.8348, 1.9483],
        [1.3861,... 1.9182],
        [1.7269, 1.4706, 1.1893],
        [1.1243, 1.8843, 1.0773]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 1]), e shape: torch.Size([300, 3])
SpMM(message func: copy_rhs, reduce func: min)
___________________ test_spmm[idtype1-min-copy_rhs-shp3-g0] ____________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (1,)), msg = 'copy_rhs', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_rhs', reduce_op = 'min'
lhs_data = tensor([[1.5589, 1.3107, 1.7207],
        [1.9089, 1.3185, 1.0669],
        [1.3572, 1.3282, 1.3475],
        [1.4461,... 1.6560],
        [1.8142, 1.2257, 1.5088],
        [1.1086, 1.1866, 1.6242]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.4640],
        [1.9530],
        [1.1028],
        [1.3960],
        [1.5133],
        [1.6845],
        [1...558],
        [1.6299],
        [1.3971],
        [1.5049],
        [1.1325]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 3]), e shape: torch.Size([100, 1])
SpMM(message func: copy_rhs, reduce func: min)
___________________ test_spmm[idtype1-min-copy_rhs-shp3-g1] ____________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (1,)), msg = 'copy_rhs', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_rhs', reduce_op = 'min'
lhs_data = tensor([[1.9777, 1.2364, 1.9991],
        [1.3785, 1.2802, 1.0445],
        [1.0074, 1.9090, 1.5136],
        [1.9186,... 1.3771],
        [1.3433, 1.0264, 1.4463],
        [1.8245, 1.7812, 1.2656]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.8017],
        [1.4641],
        [1.8586],
        [1.8566],
        [1.7588],
        [1.8954],
        [1...083],
        [1.5111],
        [1.2840],
        [1.8831],
        [1.1990]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 3]), e shape: torch.Size([300, 1])
SpMM(message func: copy_rhs, reduce func: min)
___________________ test_spmm[idtype1-min-copy_rhs-shp4-g0] ____________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), msg = 'copy_rhs', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_rhs', reduce_op = 'min'
lhs_data = tensor([[1.8361],
        [1.0660],
        [1.1171],
        [1.1019],
        [1.6053],
        [1.4307],
        [1...110],
        [1.0925],
        [1.9717],
        [1.5921],
        [1.4461]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.2482],
        [1.9155],
        [1.7232],
        [1.9971],
        [1.6388],
        [1.0006],
        [1...781],
        [1.8823],
        [1.1390],
        [1.4112],
        [1.0863]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 1]), e shape: torch.Size([100, 1])
SpMM(message func: copy_rhs, reduce func: min)
___________________ test_spmm[idtype1-min-copy_rhs-shp4-g1] ____________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), msg = 'copy_rhs', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_rhs', reduce_op = 'min'
lhs_data = tensor([[1.5196],
        [1.8575],
        [1.9656],
        [1.0882],
        [1.0862],
        [1.6789],
        [1...999],
        [1.4194],
        [1.1624],
        [1.7637],
        [1.5416]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.4618],
        [1.9174],
        [1.4837],
        [1.8789],
        [1.6275],
        [1.3855],
        [1...244],
        [1.9547],
        [1.0205],
        [1.5536],
        [1.5043]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 1]), e shape: torch.Size([300, 1])
SpMM(message func: copy_rhs, reduce func: min)
___________________ test_spmm[idtype1-min-copy_rhs-shp5-g0] ____________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
shp = ((), ()), msg = 'copy_rhs', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_rhs', reduce_op = 'min'
lhs_data = tensor([1.0167, 1.7561, 1.4168, 1.8017, 1.6744, 1.5869, 1.3544, 1.8271, 1.2650,
        1.6990, 1.4032, 1.5081, 1.0160...3691, 1.1598, 1.4310, 1.4310, 1.3839, 1.2875,
        1.9730, 1.0088, 1.9026], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([1.7456, 1.5091, 1.6620, 1.8315, 1.1908, 1.3253, 1.8656, 1.1113, 1.5011,
        1.2387, 1.6240, 1.7704, 1.4378...8345, 1.7751, 1.2881, 1.5729, 1.0745, 1.8837, 1.8234, 1.6562,
        1.6031], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30]), e shape: torch.Size([100])
SpMM(message func: copy_rhs, reduce func: min)
___________________ test_spmm[idtype1-min-copy_rhs-shp5-g1] ____________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((), ()), msg = 'copy_rhs', reducer = 'min'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_rhs', reduce_op = 'min'
lhs_data = tensor([1.4106, 1.2477, 1.8027, 1.3892, 1.4971, 1.5010, 1.4925, 1.0311, 1.9911,
        1.4394, 1.1871, 1.0901, 1.0051...4826, 1.5299, 1.9416, 1.6407, 1.5007, 1.4142,
        1.5663, 1.7212, 1.8126], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([1.5497, 1.0365, 1.8301, 1.2734, 1.6391, 1.6306, 1.4686, 1.1738, 1.9476,
        1.0660, 1.9547, 1.2288, 1.6953...9699, 1.9940, 1.8504, 1.4435, 1.3932, 1.6242,
        1.1291, 1.0382, 1.4125], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30]), e shape: torch.Size([300])
SpMM(message func: copy_rhs, reduce func: min)
______________________ test_spmm[idtype1-max-add-shp0-g0] ______________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), msg = 'add', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add', reduce_op = 'max'
lhs_data = tensor([[[[[[1.6484],
            [1.0304],
            [1.2755]]],


          [[[1.0347],
            [1.9206],
    ...5]]],


          [[[1.4237],
            [1.8964],
            [1.3571]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.1205]],

           [[1.9561]],

           [[1.0350]]]],



         [[[[1.9828]],

           [[1.017...



         [[[[1.7945]],

           [[1.0779]],

           [[1.8006]]]]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 1, 2, 1, 3, 1]), e shape: torch.Size([100, 4, 1, 3, 1, 1])
SpMM(message func: add, reduce func: max)
______________________ test_spmm[idtype1-max-add-shp0-g1] ______________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), msg = 'add', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add', reduce_op = 'max'
lhs_data = tensor([[[[[[1.8842],
            [1.2344],
            [1.1518]]],


          [[[1.1828],
            [1.6961],
    ...4]]],


          [[[1.1978],
            [1.5806],
            [1.1465]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.8844]],

           [[1.6772]],

           [[1.5577]]]],



         [[[[1.6448]],

           [[1.136...



         [[[[1.2873]],

           [[1.9667]],

           [[1.2276]]]]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 1, 2, 1, 3, 1]), e shape: torch.Size([300, 4, 1, 3, 1, 1])
SpMM(message func: add, reduce func: max)
______________________ test_spmm[idtype1-max-add-shp1-g0] ______________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1, 3), dtype=torch.float64)})
shp = ((3, 3), (1, 3)), msg = 'add', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add', reduce_op = 'max'
lhs_data = tensor([[[1.2515, 1.4783, 1.2254],
         [1.4750, 1.2457, 1.5504],
         [1.7088, 1.7924, 1.8932]],

        [[1...0706],
         [1.7180, 1.1052, 1.7974],
         [1.3869, 1.5922, 1.0174]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[1.1973, 1.4290, 1.7468]],

        [[1.8599, 1.3034, 1.4423]],

        [[1.8831, 1.3353, 1.8491]],

       ...]],

        [[1.5022, 1.3694, 1.2080]],

        [[1.4894, 1.4533, 1.2087]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 3, 3]), e shape: torch.Size([100, 1, 3])
SpMM(message func: add, reduce func: max)
______________________ test_spmm[idtype1-max-add-shp1-g1] ______________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3, 3), (1, 3)), msg = 'add', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add', reduce_op = 'max'
lhs_data = tensor([[[1.1009, 1.8221, 1.1422],
         [1.0256, 1.6379, 1.4837],
         [1.0683, 1.4023, 1.2615]],

        [[1...5465],
         [1.2935, 1.7305, 1.4314],
         [1.3528, 1.3273, 1.8523]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[1.3810, 1.6259, 1.0275]],

        [[1.4572, 1.6007, 1.4000]],

        [[1.4644, 1.5316, 1.7036]],

       ...]],

        [[1.3295, 1.0956, 1.2527]],

        [[1.6802, 1.8789, 1.1223]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 3, 3]), e shape: torch.Size([300, 1, 3])
SpMM(message func: add, reduce func: max)
______________________ test_spmm[idtype1-max-add-shp2-g0] ______________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float64)})
shp = ((1,), (3,)), msg = 'add', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add', reduce_op = 'max'
lhs_data = tensor([[1.2429],
        [1.9164],
        [1.6170],
        [1.2755],
        [1.5579],
        [1.7431],
        [1...312],
        [1.2524],
        [1.8075],
        [1.8955],
        [1.9470]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.4278, 1.0479, 1.1723],
        [1.0271, 1.2008, 1.6296],
        [1.2769, 1.9274, 1.1456],
        [1.0639,... 1.3438],
        [1.1465, 1.6115, 1.4028],
        [1.6362, 1.4307, 1.5875]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1, 3), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 1]), e shape: torch.Size([100, 3])
SpMM(message func: add, reduce func: max)
______________________ test_spmm[idtype1-max-add-shp2-g1] ______________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (3,)), msg = 'add', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add', reduce_op = 'max'
lhs_data = tensor([[1.7222],
        [1.6960],
        [1.5294],
        [1.5871],
        [1.3925],
        [1.6335],
        [1...772],
        [1.2467],
        [1.8966],
        [1.4203],
        [1.2294]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.5352, 1.1034, 1.3126],
        [1.9690, 1.5661, 1.9415],
        [1.0854, 1.9986, 1.8343],
        [1.2603,... 1.6305],
        [1.9329, 1.1132, 1.1413],
        [1.2644, 1.0556, 1.2300]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 1]), e shape: torch.Size([300, 3])
SpMM(message func: add, reduce func: max)
______________________ test_spmm[idtype1-max-add-shp3-g0] ______________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (1,)), msg = 'add', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add', reduce_op = 'max'
lhs_data = tensor([[1.6029, 1.2697, 1.8503],
        [1.0109, 1.0956, 1.9009],
        [1.2697, 1.4627, 1.3621],
        [1.0659,... 1.6414],
        [1.0331, 1.2981, 1.5971],
        [1.4372, 1.7759, 1.4741]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.9205],
        [1.0302],
        [1.5533],
        [1.1146],
        [1.1601],
        [1.8411],
        [1...608],
        [1.8055],
        [1.4697],
        [1.4972],
        [1.3064]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 3]), e shape: torch.Size([100, 1])
SpMM(message func: add, reduce func: max)
______________________ test_spmm[idtype1-max-add-shp3-g1] ______________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (1,)), msg = 'add', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add', reduce_op = 'max'
lhs_data = tensor([[1.6549, 1.9449, 1.1695],
        [1.9661, 1.3288, 1.6818],
        [1.1717, 1.6149, 1.0752],
        [1.0964,... 1.6961],
        [1.8251, 1.6347, 1.0674],
        [1.4174, 1.2251, 1.2003]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.3558],
        [1.7475],
        [1.0836],
        [1.9427],
        [1.4695],
        [1.0492],
        [1...238],
        [1.0008],
        [1.3191],
        [1.2500],
        [1.1136]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 3]), e shape: torch.Size([300, 1])
SpMM(message func: add, reduce func: max)
______________________ test_spmm[idtype1-max-add-shp4-g0] ______________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), msg = 'add', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add', reduce_op = 'max'
lhs_data = tensor([[1.7887],
        [1.0927],
        [1.9458],
        [1.4607],
        [1.3541],
        [1.1942],
        [1...452],
        [1.3808],
        [1.0745],
        [1.1748],
        [1.6427]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.0169],
        [1.1444],
        [1.2907],
        [1.7950],
        [1.4361],
        [1.4511],
        [1...873],
        [1.3556],
        [1.8842],
        [1.6126],
        [1.1453]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 1]), e shape: torch.Size([100, 1])
SpMM(message func: add, reduce func: max)
______________________ test_spmm[idtype1-max-add-shp4-g1] ______________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), msg = 'add', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add', reduce_op = 'max'
lhs_data = tensor([[1.5903],
        [1.7993],
        [1.5298],
        [1.9509],
        [1.5383],
        [1.4867],
        [1...804],
        [1.1430],
        [1.8269],
        [1.6247],
        [1.2844]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.2649],
        [1.6362],
        [1.1449],
        [1.0393],
        [1.5450],
        [1.5004],
        [1...927],
        [1.1991],
        [1.6898],
        [1.0456],
        [1.6805]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 1]), e shape: torch.Size([300, 1])
SpMM(message func: add, reduce func: max)
______________________ test_spmm[idtype1-max-add-shp5-g0] ______________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
shp = ((), ()), msg = 'add', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add', reduce_op = 'max'
lhs_data = tensor([1.3143, 1.1159, 1.9778, 1.8013, 1.4823, 1.5125, 1.4562, 1.2324, 1.3510,
        1.9554, 1.0228, 1.1616, 1.7007...6620, 1.6756, 1.8937, 1.3661, 1.8465, 1.2889,
        1.5916, 1.0558, 1.9082], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([1.8623, 1.2938, 1.1503, 1.2338, 1.1660, 1.1811, 1.4269, 1.1161, 1.8911,
        1.6928, 1.2857, 1.7126, 1.9865...4164, 1.0757, 1.2195, 1.5740, 1.1135, 1.1774, 1.4546, 1.2151,
        1.7460], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30]), e shape: torch.Size([100])
SpMM(message func: add, reduce func: max)
______________________ test_spmm[idtype1-max-add-shp5-g1] ______________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((), ()), msg = 'add', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add', reduce_op = 'max'
lhs_data = tensor([1.4069, 1.7375, 1.4316, 1.5201, 1.2057, 1.4948, 1.2030, 1.8539, 1.9017,
        1.6513, 1.5824, 1.0560, 1.4313...3759, 1.1348, 1.6728, 1.9832, 1.3891, 1.6625,
        1.8278, 1.4361, 1.6079], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([1.4609, 1.0908, 1.8458, 1.3594, 1.7643, 1.0395, 1.6861, 1.9208, 1.4579,
        1.3449, 1.2688, 1.9385, 1.8294...7935, 1.8717, 1.2304, 1.3193, 1.7149, 1.5796,
        1.9270, 1.4824, 1.2656], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30]), e shape: torch.Size([300])
SpMM(message func: add, reduce func: max)
______________________ test_spmm[idtype1-max-sub-shp0-g0] ______________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), msg = 'sub', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add', reduce_op = 'max'
lhs_data = tensor([[[[[[1.6710],
            [1.1993],
            [1.0159]]],


          [[[1.4273],
            [1.0016],
    ...2]]],


          [[[1.0295],
            [1.0606],
            [1.4223]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[-1.2083]],

           [[-1.6583]],

           [[-1.8273]]]],



         [[[[-1.7555]],

           [[-...       [[[[-1.8051]],

           [[-1.6480]],

           [[-1.3343]]]]]], dtype=torch.float64, grad_fn=<NegBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 1, 2, 1, 3, 1]), e shape: torch.Size([100, 4, 1, 3, 1, 1])
SpMM(message func: sub, reduce func: max)
______________________ test_spmm[idtype1-max-sub-shp0-g1] ______________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), msg = 'sub', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add', reduce_op = 'max'
lhs_data = tensor([[[[[[1.7029],
            [1.8835],
            [1.6903]]],


          [[[1.6594],
            [1.9518],
    ...6]]],


          [[[1.1303],
            [1.2447],
            [1.1095]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[-1.8827]],

           [[-1.1118]],

           [[-1.8283]]]],



         [[[[-1.5953]],

           [[-...       [[[[-1.8634]],

           [[-1.8155]],

           [[-1.9792]]]]]], dtype=torch.float64, grad_fn=<NegBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 1, 2, 1, 3, 1]), e shape: torch.Size([300, 4, 1, 3, 1, 1])
SpMM(message func: sub, reduce func: max)
______________________ test_spmm[idtype1-max-sub-shp1-g0] ______________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1, 3), dtype=torch.float64)})
shp = ((3, 3), (1, 3)), msg = 'sub', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add', reduce_op = 'max'
lhs_data = tensor([[[1.6894, 1.4038, 1.8861],
         [1.2866, 1.0783, 1.1571],
         [1.5827, 1.0922, 1.8288]],

        [[1...1657],
         [1.8921, 1.2126, 1.0891],
         [1.0354, 1.5819, 1.9253]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[-1.3932, -1.2954, -1.8085]],

        [[-1.7726, -1.9924, -1.3538]],

        [[-1.9603, -1.3349, -1.0110]],...1.8100, -1.6053, -1.5374]],

        [[-1.4049, -1.4442, -1.0912]]], dtype=torch.float64,
       grad_fn=<NegBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 3, 3]), e shape: torch.Size([100, 1, 3])
SpMM(message func: sub, reduce func: max)
______________________ test_spmm[idtype1-max-sub-shp1-g1] ______________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3, 3), (1, 3)), msg = 'sub', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add', reduce_op = 'max'
lhs_data = tensor([[[1.7460, 1.0200, 1.4781],
         [1.2512, 1.0960, 1.3278],
         [1.5788, 1.2799, 1.7301]],

        [[1...4963],
         [1.8149, 1.5038, 1.3788],
         [1.6144, 1.8743, 1.4866]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[-1.4096, -1.7699, -1.0804]],

        [[-1.2115, -1.7428, -1.7657]],

        [[-1.0424, -1.9679, -1.8557]],...1.7218, -1.2556, -1.7638]],

        [[-1.2145, -1.6339, -1.4866]]], dtype=torch.float64,
       grad_fn=<NegBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 3, 3]), e shape: torch.Size([300, 1, 3])
SpMM(message func: sub, reduce func: max)
______________________ test_spmm[idtype1-max-sub-shp2-g0] ______________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float64)})
shp = ((1,), (3,)), msg = 'sub', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add', reduce_op = 'max'
lhs_data = tensor([[1.5233],
        [1.0383],
        [1.6685],
        [1.3123],
        [1.6971],
        [1.7692],
        [1...275],
        [1.5620],
        [1.4277],
        [1.9463],
        [1.2396]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.8849, -1.2206, -1.9895],
        [-1.3651, -1.4547, -1.4787],
        [-1.4676, -1.1346, -1.9223],
       ...  [-1.4135, -1.0864, -1.0676],
        [-1.6173, -1.3113, -1.0025]], dtype=torch.float64,
       grad_fn=<NegBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1, 3), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 1]), e shape: torch.Size([100, 3])
SpMM(message func: sub, reduce func: max)
______________________ test_spmm[idtype1-max-sub-shp2-g1] ______________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (3,)), msg = 'sub', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add', reduce_op = 'max'
lhs_data = tensor([[1.3269],
        [1.8170],
        [1.2932],
        [1.3458],
        [1.4548],
        [1.9844],
        [1...777],
        [1.2890],
        [1.3291],
        [1.1746],
        [1.8865]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.8781, -1.5377, -1.9643],
        [-1.0779, -1.1751, -1.8633],
        [-1.5891, -1.4920, -1.3847],
       ...  [-1.6692, -1.9191, -1.8837],
        [-1.6118, -1.2407, -1.4425]], dtype=torch.float64,
       grad_fn=<NegBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 1]), e shape: torch.Size([300, 3])
SpMM(message func: sub, reduce func: max)
______________________ test_spmm[idtype1-max-sub-shp3-g0] ______________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (1,)), msg = 'sub', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add', reduce_op = 'max'
lhs_data = tensor([[1.8000, 1.3116, 1.6744],
        [1.1415, 1.6735, 1.1104],
        [1.8738, 1.9800, 1.9700],
        [1.5741,... 1.5641],
        [1.3120, 1.9904, 1.3634],
        [1.0349, 1.8742, 1.0389]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.7586],
        [-1.5014],
        [-1.3815],
        [-1.0268],
        [-1.6477],
        [-1.5634],
    ...       [-1.4503],
        [-1.4542],
        [-1.1879],
        [-1.1530]], dtype=torch.float64, grad_fn=<NegBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 3]), e shape: torch.Size([100, 1])
SpMM(message func: sub, reduce func: max)
______________________ test_spmm[idtype1-max-sub-shp3-g1] ______________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (1,)), msg = 'sub', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add', reduce_op = 'max'
lhs_data = tensor([[1.1802, 1.2759, 1.1957],
        [1.3816, 1.6498, 1.6544],
        [1.8046, 1.2842, 1.9946],
        [1.1655,... 1.6813],
        [1.0078, 1.3762, 1.4047],
        [1.0819, 1.2983, 1.5378]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.7239],
        [-1.6005],
        [-1.5605],
        [-1.9915],
        [-1.4893],
        [-1.2960],
    ...       [-1.8262],
        [-1.9855],
        [-1.8040],
        [-1.8762]], dtype=torch.float64, grad_fn=<NegBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 3]), e shape: torch.Size([300, 1])
SpMM(message func: sub, reduce func: max)
______________________ test_spmm[idtype1-max-sub-shp4-g0] ______________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), msg = 'sub', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add', reduce_op = 'max'
lhs_data = tensor([[1.7100],
        [1.5252],
        [1.4182],
        [1.1332],
        [1.5762],
        [1.3340],
        [1...863],
        [1.3995],
        [1.0864],
        [1.7921],
        [1.6610]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.7362],
        [-1.7287],
        [-1.5887],
        [-1.0973],
        [-1.4916],
        [-1.9457],
    ...       [-1.0751],
        [-1.6149],
        [-1.6467],
        [-1.2077]], dtype=torch.float64, grad_fn=<NegBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 1]), e shape: torch.Size([100, 1])
SpMM(message func: sub, reduce func: max)
______________________ test_spmm[idtype1-max-sub-shp4-g1] ______________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), msg = 'sub', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add', reduce_op = 'max'
lhs_data = tensor([[1.2048],
        [1.7449],
        [1.3245],
        [1.6570],
        [1.3350],
        [1.4174],
        [1...675],
        [1.4222],
        [1.9315],
        [1.5076],
        [1.7682]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.5364],
        [-1.2326],
        [-1.4377],
        [-1.4171],
        [-1.2268],
        [-1.7164],
    ...       [-1.7603],
        [-1.3743],
        [-1.5605],
        [-1.0198]], dtype=torch.float64, grad_fn=<NegBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 1]), e shape: torch.Size([300, 1])
SpMM(message func: sub, reduce func: max)
______________________ test_spmm[idtype1-max-sub-shp5-g0] ______________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
shp = ((), ()), msg = 'sub', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add', reduce_op = 'max'
lhs_data = tensor([1.5983, 1.1329, 1.7750, 1.2139, 1.3339, 1.0927, 1.0970, 1.5895, 1.4347,
        1.7503, 1.9357, 1.1900, 1.5401...6060, 1.9909, 1.5719, 1.0493, 1.7641, 1.9698,
        1.2038, 1.4722, 1.0548], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([-1.4369, -1.0159, -1.8098, -1.5402, -1.1044, -1.9850, -1.6287, -1.0428,
        -1.4526, -1.5362, -1.2344, -1....8536, -1.6362, -1.7124,
        -1.4280, -1.1415, -1.9517, -1.2309], dtype=torch.float64,
       grad_fn=<NegBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30]), e shape: torch.Size([100])
SpMM(message func: sub, reduce func: max)
______________________ test_spmm[idtype1-max-sub-shp5-g1] ______________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((), ()), msg = 'sub', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add', reduce_op = 'max'
lhs_data = tensor([1.1225, 1.1868, 1.7393, 1.7282, 1.9819, 1.9571, 1.2376, 1.7159, 1.0739,
        1.2552, 1.8539, 1.3830, 1.2693...3839, 1.7508, 1.7325, 1.6026, 1.3937, 1.2544,
        1.0734, 1.1874, 1.7093], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([-1.9403, -1.7250, -1.6978, -1.1687, -1.0307, -1.0651, -1.5353, -1.4374,
        -1.9746, -1.4406, -1.4079, -1....6111, -1.1127, -1.0668,
        -1.5089, -1.2140, -1.2908, -1.1077], dtype=torch.float64,
       grad_fn=<NegBackward>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30]), e shape: torch.Size([300])
SpMM(message func: sub, reduce func: max)
______________________ test_spmm[idtype1-max-mul-shp0-g0] ______________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), msg = 'mul', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul', reduce_op = 'max'
lhs_data = tensor([[[[[[1.9098],
            [1.5737],
            [1.8178]]],


          [[[1.0649],
            [1.0954],
    ...1]]],


          [[[1.0763],
            [1.6155],
            [1.0357]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.5813]],

           [[1.3061]],

           [[1.5431]]]],



         [[[[1.5215]],

           [[1.956...



         [[[[1.6801]],

           [[1.0502]],

           [[1.4200]]]]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 1, 2, 1, 3, 1]), e shape: torch.Size([100, 4, 1, 3, 1, 1])
SpMM(message func: mul, reduce func: max)
______________________ test_spmm[idtype1-max-mul-shp0-g1] ______________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), msg = 'mul', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul', reduce_op = 'max'
lhs_data = tensor([[[[[[1.3739],
            [1.5778],
            [1.4334]]],


          [[[1.5437],
            [1.0221],
    ...6]]],


          [[[1.9586],
            [1.1774],
            [1.0344]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.2042]],

           [[1.3376]],

           [[1.8695]]]],



         [[[[1.2767]],

           [[1.157...



         [[[[1.4955]],

           [[1.9839]],

           [[1.8578]]]]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 1, 2, 1, 3, 1]), e shape: torch.Size([300, 4, 1, 3, 1, 1])
SpMM(message func: mul, reduce func: max)
______________________ test_spmm[idtype1-max-mul-shp1-g0] ______________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1, 3), dtype=torch.float64)})
shp = ((3, 3), (1, 3)), msg = 'mul', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul', reduce_op = 'max'
lhs_data = tensor([[[1.4787, 1.5902, 1.4231],
         [1.0788, 1.0627, 1.9366],
         [1.8821, 1.9349, 1.2776]],

        [[1...4976],
         [1.4234, 1.8311, 1.1166],
         [1.5282, 1.2209, 1.8966]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[1.5182, 1.8592, 1.7181]],

        [[1.6052, 1.7119, 1.0001]],

        [[1.8543, 1.8931, 1.4878]],

       ...]],

        [[1.2731, 1.3555, 1.6728]],

        [[1.0597, 1.5325, 1.5273]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 3, 3]), e shape: torch.Size([100, 1, 3])
SpMM(message func: mul, reduce func: max)
______________________ test_spmm[idtype1-max-mul-shp1-g1] ______________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3, 3), (1, 3)), msg = 'mul', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul', reduce_op = 'max'
lhs_data = tensor([[[1.9579, 1.3735, 1.9051],
         [1.5647, 1.7732, 1.0863],
         [1.3590, 1.9908, 1.2813]],

        [[1...3151],
         [1.5983, 1.1743, 1.3600],
         [1.9148, 1.9345, 1.9410]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[1.3212, 1.8615, 1.8953]],

        [[1.6423, 1.6517, 1.1731]],

        [[1.1222, 1.0155, 1.2929]],

       ...]],

        [[1.5164, 1.0748, 1.9708]],

        [[1.9055, 1.4012, 1.1642]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 3, 3]), e shape: torch.Size([300, 1, 3])
SpMM(message func: mul, reduce func: max)
______________________ test_spmm[idtype1-max-mul-shp2-g0] ______________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float64)})
shp = ((1,), (3,)), msg = 'mul', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul', reduce_op = 'max'
lhs_data = tensor([[1.0201],
        [1.1572],
        [1.6627],
        [1.5802],
        [1.0173],
        [1.8224],
        [1...653],
        [1.2132],
        [1.8409],
        [1.8934],
        [1.7112]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.9936, 1.3945, 1.0713],
        [1.0171, 1.0462, 1.6008],
        [1.5121, 1.6242, 1.1807],
        [1.6687,... 1.5105],
        [1.5716, 1.1216, 1.0793],
        [1.6869, 1.0195, 1.3753]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1, 3), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 1]), e shape: torch.Size([100, 3])
SpMM(message func: mul, reduce func: max)
______________________ test_spmm[idtype1-max-mul-shp2-g1] ______________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (3,)), msg = 'mul', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul', reduce_op = 'max'
lhs_data = tensor([[1.2333],
        [1.4924],
        [1.3779],
        [1.1696],
        [1.6526],
        [1.7366],
        [1...587],
        [1.6424],
        [1.8781],
        [1.6527],
        [1.3478]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.3368, 1.6306, 1.7414],
        [1.6420, 1.5462, 1.9042],
        [1.9471, 1.6855, 1.9585],
        [1.1235,... 1.8532],
        [1.9351, 1.7820, 1.9041],
        [1.5047, 1.4704, 1.5321]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 1]), e shape: torch.Size([300, 3])
SpMM(message func: mul, reduce func: max)
______________________ test_spmm[idtype1-max-mul-shp3-g0] ______________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (1,)), msg = 'mul', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul', reduce_op = 'max'
lhs_data = tensor([[1.6656, 1.7254, 1.8844],
        [1.2399, 1.1594, 1.5333],
        [1.6897, 1.4970, 1.0427],
        [1.6956,... 1.0559],
        [1.3066, 1.2406, 1.3969],
        [1.1751, 1.4406, 1.6484]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.8189],
        [1.5455],
        [1.3785],
        [1.3565],
        [1.1803],
        [1.2639],
        [1...327],
        [1.7538],
        [1.9322],
        [1.7509],
        [1.3283]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 3]), e shape: torch.Size([100, 1])
SpMM(message func: mul, reduce func: max)
______________________ test_spmm[idtype1-max-mul-shp3-g1] ______________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (1,)), msg = 'mul', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul', reduce_op = 'max'
lhs_data = tensor([[1.8312, 1.9835, 1.9149],
        [1.3803, 1.8956, 1.4052],
        [1.8990, 1.8760, 1.9748],
        [1.0122,... 1.3484],
        [1.9224, 1.2670, 1.3520],
        [1.7785, 1.6861, 1.2813]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.0812],
        [1.2130],
        [1.3799],
        [1.3315],
        [1.4084],
        [1.9466],
        [1...796],
        [1.1258],
        [1.5135],
        [1.5839],
        [1.4384]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 3]), e shape: torch.Size([300, 1])
SpMM(message func: mul, reduce func: max)
______________________ test_spmm[idtype1-max-mul-shp4-g0] ______________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), msg = 'mul', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul', reduce_op = 'max'
lhs_data = tensor([[1.4553],
        [1.8559],
        [1.8673],
        [1.3657],
        [1.3738],
        [1.9656],
        [1...767],
        [1.7624],
        [1.6802],
        [1.2979],
        [1.9216]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.2836],
        [1.3746],
        [1.9275],
        [1.6442],
        [1.1895],
        [1.6975],
        [1...476],
        [1.3855],
        [1.9597],
        [1.6156],
        [1.2376]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 1]), e shape: torch.Size([100, 1])
SpMM(message func: mul, reduce func: max)
______________________ test_spmm[idtype1-max-mul-shp4-g1] ______________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), msg = 'mul', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul', reduce_op = 'max'
lhs_data = tensor([[1.1072],
        [1.8898],
        [1.4726],
        [1.4847],
        [1.7249],
        [1.2015],
        [1...258],
        [1.9639],
        [1.2709],
        [1.1926],
        [1.2269]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.3875],
        [1.3281],
        [1.0289],
        [1.8613],
        [1.7606],
        [1.0527],
        [1...497],
        [1.4708],
        [1.8926],
        [1.2815],
        [1.4875]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 1]), e shape: torch.Size([300, 1])
SpMM(message func: mul, reduce func: max)
______________________ test_spmm[idtype1-max-mul-shp5-g0] ______________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
shp = ((), ()), msg = 'mul', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul', reduce_op = 'max'
lhs_data = tensor([1.8228, 1.4628, 1.0627, 1.9842, 1.8110, 1.9584, 1.6966, 1.3928, 1.3049,
        1.0431, 1.8943, 1.8421, 1.3337...6775, 1.4457, 1.3016, 1.2051, 1.7284, 1.7926,
        1.8388, 1.7243, 1.3664], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([1.6584, 1.1367, 1.3239, 1.4219, 1.8897, 1.5108, 1.2809, 1.1678, 1.7880,
        1.4216, 1.9889, 1.1868, 1.5045...6594, 1.0755, 1.8693, 1.7436, 1.7697, 1.0679, 1.7268, 1.6327,
        1.2088], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30]), e shape: torch.Size([100])
SpMM(message func: mul, reduce func: max)
______________________ test_spmm[idtype1-max-mul-shp5-g1] ______________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((), ()), msg = 'mul', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul', reduce_op = 'max'
lhs_data = tensor([1.1289, 1.9071, 1.7042, 1.7680, 1.1742, 1.3646, 1.6350, 1.5063, 1.0565,
        1.1061, 1.5828, 1.2087, 1.6920...0777, 1.0799, 1.7273, 1.2265, 1.1330, 1.7749,
        1.2509, 1.6410, 1.3465], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([1.1887, 1.1863, 1.9326, 1.1716, 1.3959, 1.0906, 1.1791, 1.6346, 1.9712,
        1.7942, 1.8512, 1.0829, 1.4627...5418, 1.4749, 1.8373, 1.3611, 1.6667, 1.6000,
        1.1178, 1.5259, 1.9340], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30]), e shape: torch.Size([300])
SpMM(message func: mul, reduce func: max)
______________________ test_spmm[idtype1-max-div-shp0-g0] ______________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), msg = 'div', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul', reduce_op = 'max'
lhs_data = tensor([[[[[[1.5065],
            [1.5879],
            [1.8572]]],


          [[[1.7271],
            [1.0511],
    ...1]]],


          [[[1.7830],
            [1.0518],
            [1.6875]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[0.8295]],

           [[0.5608]],

           [[0.6024]]]],



         [[[[0.8998]],

           [[0.762...         [[[[0.5992]],

           [[0.9854]],

           [[0.5038]]]]]], dtype=torch.float64, grad_fn=<MulBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 1, 2, 1, 3, 1]), e shape: torch.Size([100, 4, 1, 3, 1, 1])
SpMM(message func: div, reduce func: max)
______________________ test_spmm[idtype1-max-div-shp0-g1] ______________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), msg = 'div', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul', reduce_op = 'max'
lhs_data = tensor([[[[[[1.0469],
            [1.2695],
            [1.1033]]],


          [[[1.6891],
            [1.6108],
    ...9]]],


          [[[1.7035],
            [1.9591],
            [1.3243]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[0.7181]],

           [[0.5536]],

           [[0.5720]]]],



         [[[[0.6747]],

           [[0.807...         [[[[0.5243]],

           [[0.5440]],

           [[0.5116]]]]]], dtype=torch.float64, grad_fn=<MulBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 1, 2, 1, 3, 1]), e shape: torch.Size([300, 4, 1, 3, 1, 1])
SpMM(message func: div, reduce func: max)
______________________ test_spmm[idtype1-max-div-shp1-g0] ______________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1, 3), dtype=torch.float64)})
shp = ((3, 3), (1, 3)), msg = 'div', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul', reduce_op = 'max'
lhs_data = tensor([[[1.8480, 1.5246, 1.8162],
         [1.9912, 1.7947, 1.1639],
         [1.6294, 1.9936, 1.9350]],

        [[1...2541],
         [1.5638, 1.7415, 1.1112],
         [1.0334, 1.5712, 1.4737]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[0.8791, 0.5833, 0.9851]],

        [[0.6638, 0.5310, 0.6871]],

        [[0.6567, 0.7700, 0.6086]],

       ...  [[0.7471, 0.6909, 0.6189]],

        [[0.9870, 0.5232, 0.7054]]], dtype=torch.float64,
       grad_fn=<MulBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 3, 3]), e shape: torch.Size([100, 1, 3])
SpMM(message func: div, reduce func: max)
______________________ test_spmm[idtype1-max-div-shp1-g1] ______________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3, 3), (1, 3)), msg = 'div', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul', reduce_op = 'max'
lhs_data = tensor([[[1.8785, 1.8400, 1.3872],
         [1.9358, 1.8541, 1.4249],
         [1.1472, 1.9783, 1.2147]],

        [[1...8789],
         [1.6483, 1.6969, 1.0370],
         [1.3156, 1.9904, 1.0736]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[0.6247, 0.6437, 0.8087]],

        [[0.5649, 0.5420, 0.5590]],

        [[0.6325, 0.9158, 0.5496]],

       ...  [[0.9229, 0.5861, 0.5790]],

        [[0.6952, 0.6505, 0.6407]]], dtype=torch.float64,
       grad_fn=<MulBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 3, 3]), e shape: torch.Size([300, 1, 3])
SpMM(message func: div, reduce func: max)
______________________ test_spmm[idtype1-max-div-shp2-g0] ______________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float64)})
shp = ((1,), (3,)), msg = 'div', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul', reduce_op = 'max'
lhs_data = tensor([[1.6058],
        [1.2850],
        [1.7664],
        [1.0440],
        [1.1379],
        [1.5469],
        [1...724],
        [1.9767],
        [1.8668],
        [1.0180],
        [1.1543]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.9361, 0.8939, 0.8985],
        [0.9070, 0.5157, 0.8261],
        [0.7529, 0.5484, 0.5751],
        [0.5264,...439],
        [0.7259, 0.5097, 0.5092],
        [0.8728, 0.8070, 0.6599]], dtype=torch.float64, grad_fn=<MulBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1, 3), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 1]), e shape: torch.Size([100, 3])
SpMM(message func: div, reduce func: max)
______________________ test_spmm[idtype1-max-div-shp2-g1] ______________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (3,)), msg = 'div', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul', reduce_op = 'max'
lhs_data = tensor([[1.3993],
        [1.3127],
        [1.9887],
        [1.2151],
        [1.9212],
        [1.1342],
        [1...639],
        [1.1285],
        [1.3661],
        [1.1923],
        [1.9079]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.9755, 0.7047, 0.8047],
        [0.9307, 0.6794, 0.6001],
        [0.8105, 0.6918, 0.6278],
        [0.9147,...465],
        [0.6554, 0.5553, 0.5074],
        [0.6860, 0.6799, 0.9490]], dtype=torch.float64, grad_fn=<MulBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 1]), e shape: torch.Size([300, 3])
SpMM(message func: div, reduce func: max)
______________________ test_spmm[idtype1-max-div-shp3-g0] ______________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (1,)), msg = 'div', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul', reduce_op = 'max'
lhs_data = tensor([[1.7220, 1.0853, 1.9818],
        [1.6752, 1.4155, 1.6388],
        [1.7989, 1.4143, 1.4347],
        [1.0349,... 1.0431],
        [1.1862, 1.5197, 1.5672],
        [1.4070, 1.1962, 1.6013]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.5746],
        [0.8018],
        [0.7010],
        [0.7842],
        [0.8923],
        [0.6273],
        [0...,
        [0.9487],
        [0.5450],
        [0.6363],
        [0.9175]], dtype=torch.float64, grad_fn=<MulBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 3]), e shape: torch.Size([100, 1])
SpMM(message func: div, reduce func: max)
______________________ test_spmm[idtype1-max-div-shp3-g1] ______________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (1,)), msg = 'div', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul', reduce_op = 'max'
lhs_data = tensor([[1.0140, 1.2150, 1.5346],
        [1.4177, 1.7949, 1.4927],
        [1.2190, 1.1698, 1.5771],
        [1.7999,... 1.8808],
        [1.3981, 1.7044, 1.5008],
        [1.6933, 1.1393, 1.4932]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.7741],
        [0.5554],
        [0.7943],
        [0.6672],
        [0.9120],
        [0.7979],
        [0...,
        [0.7899],
        [0.5359],
        [0.5827],
        [0.7615]], dtype=torch.float64, grad_fn=<MulBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 3]), e shape: torch.Size([300, 1])
SpMM(message func: div, reduce func: max)
______________________ test_spmm[idtype1-max-div-shp4-g0] ______________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), msg = 'div', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul', reduce_op = 'max'
lhs_data = tensor([[1.1506],
        [1.0079],
        [1.5863],
        [1.4433],
        [1.5960],
        [1.8890],
        [1...922],
        [1.7279],
        [1.7274],
        [1.0397],
        [1.3814]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.7358],
        [0.7896],
        [0.5068],
        [0.6050],
        [0.7768],
        [0.8811],
        [0...,
        [0.5378],
        [0.5518],
        [0.6372],
        [0.5237]], dtype=torch.float64, grad_fn=<MulBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 1]), e shape: torch.Size([100, 1])
SpMM(message func: div, reduce func: max)
______________________ test_spmm[idtype1-max-div-shp4-g1] ______________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), msg = 'div', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul', reduce_op = 'max'
lhs_data = tensor([[1.4116],
        [1.2072],
        [1.5707],
        [1.7063],
        [1.0849],
        [1.3534],
        [1...862],
        [1.2946],
        [1.0582],
        [1.3900],
        [1.0174]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.7741],
        [0.8095],
        [0.7370],
        [0.6995],
        [0.5593],
        [0.5312],
        [0...,
        [0.5977],
        [0.7064],
        [0.5108],
        [0.8706]], dtype=torch.float64, grad_fn=<MulBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 1]), e shape: torch.Size([300, 1])
SpMM(message func: div, reduce func: max)
______________________ test_spmm[idtype1-max-div-shp5-g0] ______________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
shp = ((), ()), msg = 'div', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul', reduce_op = 'max'
lhs_data = tensor([1.2840, 1.5758, 1.8906, 1.5873, 1.6298, 1.1616, 1.8868, 1.1453, 1.8178,
        1.1098, 1.4267, 1.3660, 1.8151...0248, 1.6071, 1.7271, 1.2412, 1.6134, 1.4371,
        1.2423, 1.6250, 1.7911], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([0.5889, 0.9095, 0.7789, 0.6681, 0.8375, 0.8557, 0.7819, 0.9678, 0.5665,
        0.9792, 0.7127, 0.6073, 0.5057..., 0.7798, 0.6902, 0.5120, 0.9838, 0.7712, 0.5186, 0.7021,
        0.7612], dtype=torch.float64, grad_fn=<MulBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30]), e shape: torch.Size([100])
SpMM(message func: div, reduce func: max)
______________________ test_spmm[idtype1-max-div-shp5-g1] ______________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((), ()), msg = 'div', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul', reduce_op = 'max'
lhs_data = tensor([1.7436, 1.1943, 1.3728, 1.6367, 1.9111, 1.6654, 1.0459, 1.8163, 1.5111,
        1.8469, 1.8584, 1.3269, 1.9050...3599, 1.7262, 1.8937, 1.1168, 1.8266, 1.8010,
        1.2281, 1.4239, 1.0357], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([0.9574, 0.5579, 0.9291, 0.7337, 0.9720, 0.5486, 0.5915, 0.6935, 0.6016,
        0.5948, 0.8707, 0.9778, 0.8813..., 0.8513, 0.6341, 0.8844, 0.8091, 0.6132,
        0.6292, 0.5389, 0.6101], dtype=torch.float64, grad_fn=<MulBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30]), e shape: torch.Size([300])
SpMM(message func: div, reduce func: max)
___________________ test_spmm[idtype1-max-copy_lhs-shp0-g0] ____________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), msg = 'copy_lhs', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_lhs', reduce_op = 'max'
lhs_data = tensor([[[[[[1.5762],
            [1.1993],
            [1.6049]]],


          [[[1.5063],
            [1.9017],
    ...2]]],


          [[[1.8009],
            [1.5730],
            [1.2475]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.6567]],

           [[1.2313]],

           [[1.9511]]]],



         [[[[1.4108]],

           [[1.993...



         [[[[1.6033]],

           [[1.4304]],

           [[1.6380]]]]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 1, 2, 1, 3, 1]), e shape: torch.Size([100, 4, 1, 3, 1, 1])
SpMM(message func: copy_lhs, reduce func: max)
___________________ test_spmm[idtype1-max-copy_lhs-shp0-g1] ____________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), msg = 'copy_lhs', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_lhs', reduce_op = 'max'
lhs_data = tensor([[[[[[1.4200],
            [1.2955],
            [1.4388]]],


          [[[1.6827],
            [1.9414],
    ...6]]],


          [[[1.2641],
            [1.1193],
            [1.8508]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.1072]],

           [[1.8057]],

           [[1.4448]]]],



         [[[[1.9204]],

           [[1.718...



         [[[[1.7673]],

           [[1.0985]],

           [[1.7253]]]]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 1, 2, 1, 3, 1]), e shape: torch.Size([300, 4, 1, 3, 1, 1])
SpMM(message func: copy_lhs, reduce func: max)
___________________ test_spmm[idtype1-max-copy_lhs-shp1-g0] ____________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1, 3), dtype=torch.float64)})
shp = ((3, 3), (1, 3)), msg = 'copy_lhs', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_lhs', reduce_op = 'max'
lhs_data = tensor([[[1.1098, 1.5718, 1.3898],
         [1.4553, 1.4995, 1.0077],
         [1.5388, 1.2813, 1.4717]],

        [[1...3816],
         [1.4732, 1.3292, 1.0252],
         [1.4471, 1.3823, 1.0789]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[1.7284, 1.9422, 1.3451]],

        [[1.4416, 1.7289, 1.6900]],

        [[1.6003, 1.5733, 1.1725]],

       ...]],

        [[1.4106, 1.3942, 1.7273]],

        [[1.4111, 1.8861, 1.2580]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 3, 3]), e shape: torch.Size([100, 1, 3])
SpMM(message func: copy_lhs, reduce func: max)
___________________ test_spmm[idtype1-max-copy_lhs-shp1-g1] ____________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3, 3), (1, 3)), msg = 'copy_lhs', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_lhs', reduce_op = 'max'
lhs_data = tensor([[[1.4826, 1.7108, 1.1790],
         [1.9217, 1.7443, 1.1297],
         [1.7688, 1.1305, 1.8011]],

        [[1...6918],
         [1.9311, 1.1984, 1.7968],
         [1.9070, 1.5666, 1.0936]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[1.1752, 1.5311, 1.5241]],

        [[1.2303, 1.3519, 1.8960]],

        [[1.9373, 1.6842, 1.9673]],

       ...]],

        [[1.5745, 1.1065, 1.9418]],

        [[1.9953, 1.0088, 1.0581]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 3, 3]), e shape: torch.Size([300, 1, 3])
SpMM(message func: copy_lhs, reduce func: max)
___________________ test_spmm[idtype1-max-copy_lhs-shp2-g0] ____________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float64)})
shp = ((1,), (3,)), msg = 'copy_lhs', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_lhs', reduce_op = 'max'
lhs_data = tensor([[1.4019],
        [1.3641],
        [1.7696],
        [1.0823],
        [1.9762],
        [1.4675],
        [1...420],
        [1.8417],
        [1.3024],
        [1.4801],
        [1.1841]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.0447, 1.9978, 1.1113],
        [1.9258, 1.8309, 1.9743],
        [1.5400, 1.0470, 1.0627],
        [1.5777,... 1.0082],
        [1.2106, 1.0517, 1.9206],
        [1.4622, 1.5070, 1.3354]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1, 3), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 1]), e shape: torch.Size([100, 3])
SpMM(message func: copy_lhs, reduce func: max)
___________________ test_spmm[idtype1-max-copy_lhs-shp2-g1] ____________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (3,)), msg = 'copy_lhs', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_lhs', reduce_op = 'max'
lhs_data = tensor([[1.5005],
        [1.2116],
        [1.1906],
        [1.3601],
        [1.9966],
        [1.0834],
        [1...413],
        [1.7849],
        [1.8476],
        [1.3743],
        [1.4578]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.7152, 1.4286, 1.3644],
        [1.1512, 1.9979, 1.7252],
        [1.3022, 1.9409, 1.3333],
        [1.2775,... 1.3597],
        [1.1966, 1.9924, 1.7431],
        [1.6741, 1.5632, 1.7706]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 1]), e shape: torch.Size([300, 3])
SpMM(message func: copy_lhs, reduce func: max)
___________________ test_spmm[idtype1-max-copy_lhs-shp3-g0] ____________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (1,)), msg = 'copy_lhs', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_lhs', reduce_op = 'max'
lhs_data = tensor([[1.8277, 1.4389, 1.5491],
        [1.2772, 1.8915, 1.4389],
        [1.7355, 1.7258, 1.5632],
        [1.5374,... 1.5375],
        [1.1187, 1.1519, 1.3438],
        [1.6218, 1.6354, 1.1951]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.4946],
        [1.7981],
        [1.5549],
        [1.9183],
        [1.9975],
        [1.8257],
        [1...390],
        [1.1491],
        [1.5506],
        [1.1595],
        [1.6848]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 3]), e shape: torch.Size([100, 1])
SpMM(message func: copy_lhs, reduce func: max)
___________________ test_spmm[idtype1-max-copy_lhs-shp3-g1] ____________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (1,)), msg = 'copy_lhs', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_lhs', reduce_op = 'max'
lhs_data = tensor([[1.8943, 1.4179, 1.8289],
        [1.7115, 1.6309, 1.7848],
        [1.4635, 1.6862, 1.4046],
        [1.8447,... 1.9818],
        [1.8169, 1.1324, 1.8279],
        [1.8363, 1.9885, 1.4305]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.9651],
        [1.8914],
        [1.9935],
        [1.3784],
        [1.0007],
        [1.8512],
        [1...028],
        [1.5418],
        [1.0176],
        [1.4288],
        [1.9665]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 3]), e shape: torch.Size([300, 1])
SpMM(message func: copy_lhs, reduce func: max)
___________________ test_spmm[idtype1-max-copy_lhs-shp4-g0] ____________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), msg = 'copy_lhs', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_lhs', reduce_op = 'max'
lhs_data = tensor([[1.9108],
        [1.9152],
        [1.2647],
        [1.6213],
        [1.8683],
        [1.3377],
        [1...391],
        [1.7813],
        [1.6688],
        [1.0444],
        [1.9414]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.4416],
        [1.7016],
        [1.1309],
        [1.5423],
        [1.2940],
        [1.8413],
        [1...559],
        [1.8790],
        [1.2780],
        [1.7550],
        [1.8977]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 1]), e shape: torch.Size([100, 1])
SpMM(message func: copy_lhs, reduce func: max)
___________________ test_spmm[idtype1-max-copy_lhs-shp4-g1] ____________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), msg = 'copy_lhs', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_lhs', reduce_op = 'max'
lhs_data = tensor([[1.9174],
        [1.4229],
        [1.8334],
        [1.1161],
        [1.2392],
        [1.6320],
        [1...061],
        [1.0785],
        [1.8489],
        [1.6682],
        [1.8143]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.5653],
        [1.4571],
        [1.3716],
        [1.0650],
        [1.8752],
        [1.7198],
        [1...138],
        [1.6927],
        [1.0969],
        [1.5324],
        [1.9261]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 1]), e shape: torch.Size([300, 1])
SpMM(message func: copy_lhs, reduce func: max)
___________________ test_spmm[idtype1-max-copy_lhs-shp5-g0] ____________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
shp = ((), ()), msg = 'copy_lhs', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_lhs', reduce_op = 'max'
lhs_data = tensor([1.0966, 1.7699, 1.4928, 1.4720, 1.9015, 1.9101, 1.0270, 1.1513, 1.6360,
        1.1785, 1.6189, 1.8550, 1.8124...6803, 1.7963, 1.7009, 1.9737, 1.8786, 1.2489,
        1.8793, 1.4399, 1.6331], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([1.8297, 1.2791, 1.6426, 1.1416, 1.5940, 1.9716, 1.6360, 1.1922, 1.8324,
        1.4719, 1.5379, 1.5539, 1.8881...2612, 1.7572, 1.9295, 1.5970, 1.0389, 1.1823, 1.6100, 1.9948,
        1.1991], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30]), e shape: torch.Size([100])
SpMM(message func: copy_lhs, reduce func: max)
___________________ test_spmm[idtype1-max-copy_lhs-shp5-g1] ____________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((), ()), msg = 'copy_lhs', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_lhs', reduce_op = 'max'
lhs_data = tensor([1.2424, 1.1982, 1.7685, 1.7637, 1.9998, 1.1921, 1.2401, 1.4309, 1.9515,
        1.4201, 1.8512, 1.3927, 1.8375...1372, 1.5675, 1.3492, 1.8912, 1.0722, 1.5696,
        1.5731, 1.5788, 1.3661], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([1.5413, 1.4234, 1.6399, 1.0058, 1.5044, 1.5776, 1.9651, 1.2191, 1.7505,
        1.1045, 1.3384, 1.4103, 1.5368...6049, 1.9995, 1.6561, 1.3742, 1.8489, 1.3933,
        1.4316, 1.1692, 1.7627], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30]), e shape: torch.Size([300])
SpMM(message func: copy_lhs, reduce func: max)
___________________ test_spmm[idtype1-max-copy_rhs-shp0-g0] ____________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), msg = 'copy_rhs', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_rhs', reduce_op = 'max'
lhs_data = tensor([[[[[[1.5077],
            [1.9991],
            [1.9250]]],


          [[[1.0276],
            [1.4633],
    ...6]]],


          [[[1.3975],
            [1.6377],
            [1.4991]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.7558]],

           [[1.3417]],

           [[1.6443]]]],



         [[[[1.8493]],

           [[1.475...



         [[[[1.2692]],

           [[1.9579]],

           [[1.2139]]]]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 1, 2, 1, 3, 1]), e shape: torch.Size([100, 4, 1, 3, 1, 1])
SpMM(message func: copy_rhs, reduce func: max)
___________________ test_spmm[idtype1-max-copy_rhs-shp0-g1] ____________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), msg = 'copy_rhs', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_rhs', reduce_op = 'max'
lhs_data = tensor([[[[[[1.3478],
            [1.2068],
            [1.1538]]],


          [[[1.1824],
            [1.7510],
    ...6]]],


          [[[1.5803],
            [1.2701],
            [1.8796]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.1601]],

           [[1.5845]],

           [[1.7727]]]],



         [[[[1.4687]],

           [[1.316...



         [[[[1.6594]],

           [[1.2555]],

           [[1.2373]]]]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 1, 2, 1, 3, 1]), e shape: torch.Size([300, 4, 1, 3, 1, 1])
SpMM(message func: copy_rhs, reduce func: max)
___________________ test_spmm[idtype1-max-copy_rhs-shp1-g0] ____________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1, 3), dtype=torch.float64)})
shp = ((3, 3), (1, 3)), msg = 'copy_rhs', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_rhs', reduce_op = 'max'
lhs_data = tensor([[[1.4646, 1.1296, 1.0295],
         [1.6238, 1.3505, 1.0072],
         [1.1973, 1.8653, 1.2577]],

        [[1...6131],
         [1.2690, 1.3072, 1.2922],
         [1.9269, 1.8914, 1.0231]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[1.0262, 1.9915, 1.8939]],

        [[1.4457, 1.8454, 1.9997]],

        [[1.9102, 1.0329, 1.3452]],

       ...]],

        [[1.6842, 1.3629, 1.9304]],

        [[1.0061, 1.1989, 1.2092]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 3, 3]), e shape: torch.Size([100, 1, 3])
SpMM(message func: copy_rhs, reduce func: max)
___________________ test_spmm[idtype1-max-copy_rhs-shp1-g1] ____________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3, 3), (1, 3)), msg = 'copy_rhs', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_rhs', reduce_op = 'max'
lhs_data = tensor([[[1.6051, 1.8884, 1.4286],
         [1.0376, 1.6663, 1.0623],
         [1.9677, 1.6928, 1.7559]],

        [[1...1862],
         [1.1822, 1.2915, 1.0238],
         [1.3759, 1.2778, 1.6381]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[1.9959, 1.5547, 1.6437]],

        [[1.9862, 1.7733, 1.5614]],

        [[1.1274, 1.4943, 1.0812]],

       ...]],

        [[1.7800, 1.9794, 1.2560]],

        [[1.2800, 1.4094, 1.0507]]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 3, 3]), e shape: torch.Size([300, 1, 3])
SpMM(message func: copy_rhs, reduce func: max)
___________________ test_spmm[idtype1-max-copy_rhs-shp2-g0] ____________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float64)})
shp = ((1,), (3,)), msg = 'copy_rhs', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_rhs', reduce_op = 'max'
lhs_data = tensor([[1.9353],
        [1.9317],
        [1.5458],
        [1.9329],
        [1.2794],
        [1.5357],
        [1...738],
        [1.8230],
        [1.5120],
        [1.6478],
        [1.8997]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.0527, 1.9523, 1.1929],
        [1.9636, 1.1363, 1.8293],
        [1.7396, 1.6875, 1.0390],
        [1.2036,... 1.4851],
        [1.7741, 1.3465, 1.5398],
        [1.8116, 1.7231, 1.4823]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1, 3), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 1]), e shape: torch.Size([100, 3])
SpMM(message func: copy_rhs, reduce func: max)
___________________ test_spmm[idtype1-max-copy_rhs-shp2-g1] ____________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (3,)), msg = 'copy_rhs', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_rhs', reduce_op = 'max'
lhs_data = tensor([[1.1865],
        [1.9323],
        [1.8352],
        [1.8995],
        [1.2474],
        [1.8089],
        [1...320],
        [1.1013],
        [1.2392],
        [1.9436],
        [1.4115]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.3970, 1.6539, 1.7090],
        [1.9466, 1.7660, 1.7247],
        [1.6889, 1.5839, 1.5958],
        [1.9168,... 1.4057],
        [1.7744, 1.6280, 1.4668],
        [1.0317, 1.5707, 1.8093]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 1]), e shape: torch.Size([300, 3])
SpMM(message func: copy_rhs, reduce func: max)
___________________ test_spmm[idtype1-max-copy_rhs-shp3-g0] ____________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (1,)), msg = 'copy_rhs', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_rhs', reduce_op = 'max'
lhs_data = tensor([[1.0169, 1.9545, 1.0320],
        [1.8227, 1.3584, 1.6235],
        [1.4482, 1.0361, 1.2083],
        [1.4852,... 1.3414],
        [1.6327, 1.5423, 1.4297],
        [1.3469, 1.1528, 1.9186]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.2045],
        [1.8481],
        [1.0052],
        [1.1498],
        [1.5188],
        [1.9305],
        [1...405],
        [1.4945],
        [1.7254],
        [1.8118],
        [1.3171]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(3,), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 3]), e shape: torch.Size([100, 1])
SpMM(message func: copy_rhs, reduce func: max)
___________________ test_spmm[idtype1-max-copy_rhs-shp3-g1] ____________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (1,)), msg = 'copy_rhs', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_rhs', reduce_op = 'max'
lhs_data = tensor([[1.2070, 1.7472, 1.8957],
        [1.0413, 1.0136, 1.6912],
        [1.3976, 1.3221, 1.4551],
        [1.1243,... 1.1634],
        [1.9549, 1.1574, 1.5699],
        [1.6044, 1.0293, 1.4115]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.1381],
        [1.6963],
        [1.9592],
        [1.7828],
        [1.7304],
        [1.1813],
        [1...186],
        [1.6917],
        [1.6023],
        [1.5720],
        [1.1087]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 3]), e shape: torch.Size([300, 1])
SpMM(message func: copy_rhs, reduce func: max)
___________________ test_spmm[idtype1-max-copy_rhs-shp4-g0] ____________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), msg = 'copy_rhs', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_rhs', reduce_op = 'max'
lhs_data = tensor([[1.0520],
        [1.6411],
        [1.7548],
        [1.6818],
        [1.3309],
        [1.1343],
        [1...871],
        [1.0398],
        [1.0995],
        [1.3611],
        [1.4542]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.0912],
        [1.0418],
        [1.4437],
        [1.9030],
        [1.4809],
        [1.8104],
        [1...604],
        [1.0460],
        [1.5723],
        [1.5485],
        [1.4447]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30, 1]), e shape: torch.Size([100, 1])
SpMM(message func: copy_rhs, reduce func: max)
___________________ test_spmm[idtype1-max-copy_rhs-shp4-g1] ____________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), msg = 'copy_rhs', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_rhs', reduce_op = 'max'
lhs_data = tensor([[1.1395],
        [1.7073],
        [1.6761],
        [1.2410],
        [1.1904],
        [1.8738],
        [1...394],
        [1.9651],
        [1.4256],
        [1.4418],
        [1.5279]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.5907],
        [1.1851],
        [1.3082],
        [1.3810],
        [1.4143],
        [1.3668],
        [1...016],
        [1.6338],
        [1.4838],
        [1.2226],
        [1.4372]], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30, 1]), e shape: torch.Size([300, 1])
SpMM(message func: copy_rhs, reduce func: max)
___________________ test_spmm[idtype1-max-copy_rhs-shp5-g0] ____________________

idtype = torch.int64
g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
shp = ((), ()), msg = 'copy_rhs', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_rhs', reduce_op = 'max'
lhs_data = tensor([1.8493, 1.3279, 1.2895, 1.1608, 1.6293, 1.9074, 1.2303, 1.4470, 1.0331,
        1.2174, 1.6124, 1.4632, 1.8037...5681, 1.0571, 1.3035, 1.0409, 1.0726, 1.1097,
        1.1778, 1.9023, 1.4969], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([1.5407, 1.1753, 1.2624, 1.9827, 1.2945, 1.8554, 1.6569, 1.8298, 1.6474,
        1.8429, 1.7575, 1.0565, 1.2916...1905, 1.4068, 1.9685, 1.0425, 1.0003, 1.9981, 1.4703, 1.1477,
        1.5241], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
u shape: torch.Size([30]), e shape: torch.Size([100])
SpMM(message func: copy_rhs, reduce func: max)
___________________ test_spmm[idtype1-max-copy_rhs-shp5-g1] ____________________

idtype = torch.int64
g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((), ()), msg = 'copy_rhs', reducer = 'max'

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', spmm_shapes)
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'copy_lhs', 'copy_rhs'])
    @pytest.mark.parametrize('reducer', ['sum', 'min', 'max'])
    @parametrize_idtype
    def test_spmm(idtype, g, shp, msg, reducer):
        g = g.astype(idtype).to(F.ctx())
        print(g)
        print(g.idtype)
    
        hu = F.tensor(np.random.rand(*((g.number_of_src_nodes(),) + shp[0])) + 1)
        he = F.tensor(np.random.rand(*((g.number_of_edges(),) + shp[1])) + 1)
        print('u shape: {}, e shape: {}'.format(F.shape(hu), F.shape(he)))
    
        g.srcdata['x'] = F.attach_grad(F.clone(hu))
        g.edata['w'] = F.attach_grad(F.clone(he))
        print('SpMM(message func: {}, reduce func: {})'.format(msg, reducer))
    
        u = F.attach_grad(F.clone(hu))
        e = F.attach_grad(F.clone(he))
        with F.record_grad():
>           v = gspmm(g, msg, reducer, u, e)

tests/compute/test_sparse.py:120: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_rhs', reduce_op = 'max'
lhs_data = tensor([1.1272, 1.1530, 1.3140, 1.9274, 1.7517, 1.6864, 1.3085, 1.9002, 1.7216,
        1.3683, 1.5411, 1.4379, 1.1584...2297, 1.5651, 1.5517, 1.8666, 1.2353, 1.5947,
        1.8187, 1.9068, 1.2176], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([1.6642, 1.6310, 1.0311, 1.4416, 1.7053, 1.6824, 1.0054, 1.9024, 1.2391,
        1.6129, 1.9408, 1.4665, 1.5774...6047, 1.1117, 1.0133, 1.9453, 1.5470, 1.2875,
        1.2749, 1.1516, 1.9865], dtype=torch.float64, requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
u shape: torch.Size([30]), e shape: torch.Size([300])
SpMM(message func: copy_rhs, reduce func: max)
_____________________ test_sddmm[idtype0-add-u-v-shp0-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'v', rhs_target = 'u'
msg = 'add', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72ae4f090>
op = 'add'
lhs_data = tensor([[[[[[1.8359],
            [1.4310],
            [1.2108]]],


          [[[1.4269],
            [1.5427],
    ...0]]],


          [[[1.4066],
            [1.5816],
            [1.1979]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.6500]],

           [[1.1763]],

           [[1.1556]]]],



         [[[[1.6716]],

           [[1.750...



         [[[[1.0850]],

           [[1.9148]],

           [[1.2829]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: v_add_u)
_____________________ test_sddmm[idtype0-add-u-v-shp0-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'v', rhs_target = 'u'
msg = 'add', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72ae4b708>
op = 'add'
lhs_data = tensor([[[[[[1.6809],
            [1.9177],
            [1.6446]]],


          [[[1.6902],
            [1.6200],
    ...4]]],


          [[[1.6185],
            [1.8230],
            [1.2936]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.3226]],

           [[1.2616]],

           [[1.2172]]]],



         [[[[1.2470]],

           [[1.554...



         [[[[1.0080]],

           [[1.0156]],

           [[1.0358]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([40, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: v_add_u)
_____________________ test_sddmm[idtype0-add-u-v-shp1-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'v', rhs_target = 'u'
msg = 'add', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb705c06798>
op = 'add'
lhs_data = tensor([[[[[1.3031, 1.2238, 1.5944,  ..., 1.0581, 1.0186, 1.4038]],

          [[1.9181, 1.8232, 1.9124,  ..., 1.5521,...
          [[1.7145, 1.3003, 1.8675,  ..., 1.2381, 1.1495, 1.0157]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.4180, 1.1464, 1.9356,  ..., 1.9569, 1.0690, 1.1504],
           [1.8139, 1.8424, 1.8897,  ..., 1.0390, 1...
           [1.3406, 1.0087, 1.8938,  ..., 1.8151, 1.5438, 1.2001]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: v_add_u)
_____________________ test_sddmm[idtype0-add-u-v-shp1-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'v', rhs_target = 'u'
msg = 'add', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43e6cb750>
op = 'add'
lhs_data = tensor([[[[[1.4291, 1.2911, 1.3286,  ..., 1.2814, 1.5526, 1.7078]],

          [[1.0098, 1.0926, 1.1066,  ..., 1.9850,...
          [[1.1461, 1.0378, 1.2832,  ..., 1.3910, 1.6486, 1.6279]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.4000, 1.2748, 1.6569,  ..., 1.9805, 1.6038, 1.1115],
           [1.0172, 1.1459, 1.5732,  ..., 1.3681, 1...
           [1.7943, 1.7346, 1.2096,  ..., 1.3983, 1.1555, 1.6346]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([40, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: v_add_u)
_____________________ test_sddmm[idtype0-add-u-v-shp2-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'v', rhs_target = 'u', msg = 'add'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72be55d38>
op = 'add'
lhs_data = tensor([[[[1.2798, 1.4185, 1.5339],
          [1.2361, 1.0130, 1.2795],
          [1.4693, 1.6072, 1.6701]]],


      ...7],
          [1.6377, 1.8780, 1.2169],
          [1.8144, 1.5647, 1.4318]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.4879, 1.4659, 1.9348]],

         [[1.6964, 1.3684, 1.1631]],

         [[1.8221, 1.9239, 1.9851]],

    ...

         [[1.4035, 1.8341, 1.2474]],

         [[1.1138, 1.5864, 1.1728]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: v_add_u)
_____________________ test_sddmm[idtype0-add-u-v-shp2-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'v', rhs_target = 'u', msg = 'add'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72ae652d0>
op = 'add'
lhs_data = tensor([[[[1.9534, 1.4132, 1.4038],
          [1.9710, 1.7728, 1.9806],
          [1.2928, 1.9335, 1.7768]]],


      ...9],
          [1.7887, 1.5742, 1.6692],
          [1.9082, 1.4226, 1.9209]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.8778, 1.6821, 1.4559]],

         [[1.9817, 1.8216, 1.2082]],

         [[1.9773, 1.2188, 1.3031]],

    ...

         [[1.0013, 1.2226, 1.2590]],

         [[1.6931, 1.5742, 1.5459]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([40, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: v_add_u)
_____________________ test_sddmm[idtype0-add-u-v-shp3-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'v', rhs_target = 'u', msg = 'add'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72ae55870>
op = 'add'
lhs_data = tensor([[1.5448, 1.3314, 1.2040],
        [1.4703, 1.0980, 1.7199],
        [1.1437, 1.5473, 1.1323],
        [1.5056,... 1.9422],
        [1.4349, 1.9105, 1.2195],
        [1.9406, 1.7291, 1.7322]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.9010, 1.6635, 1.8458],
        [1.5432, 1.4341, 1.5912],
        [1.1183, 1.7080, 1.8978],
        [1.4062,... 1.6491],
        [1.5781, 1.5862, 1.8169],
        [1.2750, 1.8078, 1.8564]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: v_add_u)
_____________________ test_sddmm[idtype0-add-u-v-shp3-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'v', rhs_target = 'u', msg = 'add'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72ae556c0>
op = 'add'
lhs_data = tensor([[1.1929, 1.2458, 1.7290],
        [1.7940, 1.5281, 1.6899],
        [1.2063, 1.7154, 1.2333],
        [1.0516,... 1.8003],
        [1.1743, 1.3603, 1.6044],
        [1.1837, 1.2694, 1.0415]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.9464, 1.3926, 1.9113],
        [1.6657, 1.6324, 1.8518],
        [1.1292, 1.4452, 1.8811],
        [1.6080,... 1.8177],
        [1.9664, 1.1811, 1.8015],
        [1.7820, 1.5891, 1.9479]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([40, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: v_add_u)
_____________________ test_sddmm[idtype0-add-u-v-shp4-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'v', rhs_target = 'u', msg = 'add'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72ae591f8>
op = 'add'
lhs_data = tensor([[1.7174],
        [1.7400],
        [1.3941],
        [1.3929],
        [1.6896],
        [1.1994],
        [1...636],
        [1.4418],
        [1.6839],
        [1.8561],
        [1.9133]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.8435],
        [1.7878],
        [1.7554],
        [1.2093],
        [1.0422],
        [1.9048],
        [1...564],
        [1.1978],
        [1.3405],
        [1.8498],
        [1.7777]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: v_add_u)
_____________________ test_sddmm[idtype0-add-u-v-shp4-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'v', rhs_target = 'u', msg = 'add'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72959d750>
op = 'add'
lhs_data = tensor([[1.8147],
        [1.7229],
        [1.3779],
        [1.9968],
        [1.1666],
        [1.6118],
        [1...063],
        [1.7209],
        [1.1640],
        [1.9226],
        [1.0912]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.6949],
        [1.0458],
        [1.6857],
        [1.3732],
        [1.9773],
        [1.3937],
        [1...114],
        [1.4225],
        [1.6308],
        [1.2919],
        [1.8140]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([40, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: v_add_u)
_____________________ test_sddmm[idtype0-add-u-e-shp0-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...   edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'e', rhs_target = 'u'
msg = 'add', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb705a84f30>
op = 'add'
lhs_data = tensor([[[[[[1.8416],
            [1.7451],
            [1.1828]]],


          [[[1.9120],
            [1.4824],
    ...9]]],


          [[[1.5341],
            [1.6782],
            [1.3322]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.8472]],

           [[1.2927]],

           [[1.3825]]]],



         [[[[1.9857]],

           [[1.204...



         [[[[1.7390]],

           [[1.7611]],

           [[1.9820]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([100, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: e_add_u)
_____________________ test_sddmm[idtype0-add-u-e-shp0-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'e', rhs_target = 'u'
msg = 'add', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43e96ecf0>
op = 'add'
lhs_data = tensor([[[[[[1.7177],
            [1.7597],
            [1.0669]]],


          [[[1.6947],
            [1.9365],
    ...4]]],


          [[[1.2413],
            [1.5285],
            [1.5667]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.3162]],

           [[1.9923]],

           [[1.0590]]]],



         [[[[1.3361]],

           [[1.168...



         [[[[1.4483]],

           [[1.1261]],

           [[1.7266]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([300, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: e_add_u)
_____________________ test_sddmm[idtype0-add-u-e-shp1-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'e', rhs_target = 'u'
msg = 'add', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72ae44af8>
op = 'add'
lhs_data = tensor([[[[[1.8726, 1.3992, 1.5259,  ..., 1.4883, 1.5476, 1.5043]],

          [[1.6301, 1.5875, 1.0300,  ..., 1.3956,...
          [[1.1336, 1.9592, 1.9632,  ..., 1.2289, 1.1241, 1.0071]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.6147, 1.2744, 1.4549,  ..., 1.7327, 1.9400, 1.2206],
           [1.0304, 1.0472, 1.2055,  ..., 1.9126, 1...
           [1.4652, 1.6232, 1.8283,  ..., 1.0992, 1.7763, 1.9747]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([100, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: e_add_u)
_____________________ test_sddmm[idtype0-add-u-e-shp1-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'e', rhs_target = 'u'
msg = 'add', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72ae44c60>
op = 'add'
lhs_data = tensor([[[[[1.0488, 1.8265, 1.8601,  ..., 1.2549, 1.0371, 1.2858]],

          [[1.7043, 1.0306, 1.9673,  ..., 1.5553,...
          [[1.0894, 1.1446, 1.4774,  ..., 1.2791, 1.1638, 1.1551]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.2022, 1.1100, 1.8475,  ..., 1.1508, 1.5684, 1.6108],
           [1.5413, 1.9142, 1.5643,  ..., 1.3332, 1...
           [1.4284, 1.2672, 1.3725,  ..., 1.0818, 1.9778, 1.6328]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([300, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: e_add_u)
_____________________ test_sddmm[idtype0-add-u-e-shp2-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1, 3, 3), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'e', rhs_target = 'u', msg = 'add'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb7295b0870>
op = 'add'
lhs_data = tensor([[[[1.6975, 1.2203, 1.2841],
          [1.2838, 1.3535, 1.1028],
          [1.7459, 1.5555, 1.7464]]],


      ...1],
          [1.4128, 1.8123, 1.9640],
          [1.8887, 1.3374, 1.2137]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.8155, 1.5865, 1.6584]],

         [[1.9617, 1.4822, 1.8847]],

         [[1.1407, 1.8855, 1.8534]],

    ...

         [[1.3135, 1.1021, 1.0944]],

         [[1.2955, 1.2001, 1.4978]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([100, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: e_add_u)
_____________________ test_sddmm[idtype0-add-u-e-shp2-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'e', rhs_target = 'u', msg = 'add'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72c1d9fc0>
op = 'add'
lhs_data = tensor([[[[1.3758, 1.8711, 1.2947],
          [1.8792, 1.8383, 1.1650],
          [1.9994, 1.0185, 1.2790]]],


      ...7],
          [1.3280, 1.7114, 1.8727],
          [1.2514, 1.3733, 1.7533]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.9342, 1.2110, 1.4657]],

         [[1.5014, 1.7505, 1.1415]],

         [[1.9846, 1.7400, 1.4357]],

    ...

         [[1.0199, 1.5805, 1.4450]],

         [[1.9090, 1.0612, 1.9326]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([300, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: e_add_u)
_____________________ test_sddmm[idtype0-add-u-e-shp3-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...oat64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(3,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'e', rhs_target = 'u', msg = 'add'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb705ae5ea0>
op = 'add'
lhs_data = tensor([[1.4191, 1.5461, 1.5294],
        [1.3016, 1.2080, 1.6682],
        [1.7263, 1.0811, 1.9747],
        [1.1981,... 1.1303],
        [1.5304, 1.6140, 1.1568],
        [1.0741, 1.2795, 1.3087]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.5088, 1.4496, 1.5377],
        [1.8666, 1.7781, 1.2280],
        [1.6166, 1.7209, 1.5250],
        [1.6172,... 1.5719],
        [1.5015, 1.7006, 1.3041],
        [1.1946, 1.6119, 1.8396]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1, 3, 3), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([100, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: e_add_u)
_____________________ test_sddmm[idtype0-add-u-e-shp3-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'e', rhs_target = 'u', msg = 'add'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb44767afc0>
op = 'add'
lhs_data = tensor([[1.5037, 1.3300, 1.7545],
        [1.6087, 1.4279, 1.3494],
        [1.0710, 1.1794, 1.0113],
        [1.4892,... 1.4339],
        [1.9532, 1.8988, 1.1946],
        [1.4970, 1.4307, 1.7400]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.9499, 1.6962, 1.3213],
        [1.9697, 1.7532, 1.0522],
        [1.6274, 1.5883, 1.1131],
        [1.2591,... 1.4200],
        [1.9621, 1.6811, 1.9538],
        [1.6598, 1.9181, 1.8434]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([300, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: e_add_u)
_____________________ test_sddmm[idtype0-add-u-e-shp4-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...oat64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'e', rhs_target = 'u', msg = 'add'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd82d0>
op = 'add'
lhs_data = tensor([[1.7855],
        [1.2114],
        [1.0233],
        [1.9180],
        [1.1531],
        [1.5277],
        [1...444],
        [1.2176],
        [1.2375],
        [1.3924],
        [1.4410]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.8389],
        [1.5678],
        [1.1868],
        [1.1853],
        [1.2061],
        [1.6218],
        [1...924],
        [1.6991],
        [1.1442],
        [1.8868],
        [1.6223]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(3,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([100, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: e_add_u)
_____________________ test_sddmm[idtype0-add-u-e-shp4-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'e', rhs_target = 'u', msg = 'add'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd8510>
op = 'add'
lhs_data = tensor([[1.0078],
        [1.7341],
        [1.0185],
        [1.0480],
        [1.6414],
        [1.2923],
        [1...908],
        [1.2159],
        [1.7952],
        [1.5262],
        [1.6154]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.5128],
        [1.9214],
        [1.2815],
        [1.1572],
        [1.9905],
        [1.9704],
        [1...692],
        [1.2700],
        [1.5861],
        [1.0919],
        [1.9468]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([300, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: e_add_u)
_____________________ test_sddmm[idtype0-add-v-u-shp0-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': ...oat64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'u', rhs_target = 'v'
msg = 'add', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72ae4fcf0>
op = 'add'
lhs_data = tensor([[[[[[1.9567],
            [1.8737],
            [1.6358]]],


          [[[1.3749],
            [1.7472],
    ...1]]],


          [[[1.9226],
            [1.1402],
            [1.7328]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.8509]],

           [[1.4331]],

           [[1.8933]]]],



         [[[[1.3092]],

           [[1.998...



         [[[[1.6774]],

           [[1.0543]],

           [[1.0867]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: u_add_v)
_____________________ test_sddmm[idtype0-add-v-u-shp0-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'u', rhs_target = 'v'
msg = 'add', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72ab60120>
op = 'add'
lhs_data = tensor([[[[[[1.9250],
            [1.6382],
            [1.0213]]],


          [[[1.3354],
            [1.1147],
    ...7]]],


          [[[1.9414],
            [1.7082],
            [1.1554]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.7757]],

           [[1.1341]],

           [[1.0502]]]],



         [[[[1.6224]],

           [[1.281...



         [[[[1.5697]],

           [[1.6469]],

           [[1.7265]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([40, 4, 1, 3, 1, 1])
SDDMM(message func: u_add_v)
_____________________ test_sddmm[idtype0-add-v-u-shp1-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Sch...oat64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'u', rhs_target = 'v'
msg = 'add', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72ab603f0>
op = 'add'
lhs_data = tensor([[[[[1.1243, 1.1527, 1.3101,  ..., 1.6840, 1.5638, 1.5521]],

          [[1.1848, 1.5404, 1.3460,  ..., 1.7100,...
          [[1.5499, 1.6683, 1.4806,  ..., 1.3767, 1.3060, 1.4148]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.7090, 1.0981, 1.2779,  ..., 1.8151, 1.5244, 1.1344],
           [1.7141, 1.6868, 1.8750,  ..., 1.0729, 1...
           [1.7753, 1.3926, 1.4929,  ..., 1.5204, 1.1154, 1.4568]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: u_add_v)
_____________________ test_sddmm[idtype0-add-v-u-shp1-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'u', rhs_target = 'v'
msg = 'add', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72ab60678>
op = 'add'
lhs_data = tensor([[[[[1.1538, 1.1012, 1.4490,  ..., 1.7771, 1.6525, 1.9886]],

          [[1.0263, 1.2944, 1.4714,  ..., 1.6066,...
          [[1.2419, 1.5463, 1.8135,  ..., 1.0718, 1.6890, 1.9449]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.8005, 1.7702, 1.6003,  ..., 1.3688, 1.6425, 1.8139],
           [1.6533, 1.3866, 1.8309,  ..., 1.4326, 1...
           [1.5551, 1.7148, 1.9013,  ..., 1.0680, 1.9719, 1.4990]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([40, 1, 3, 7, 7])
SDDMM(message func: u_add_v)
_____________________ test_sddmm[idtype0-add-v-u-shp2-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme...oat64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'u', rhs_target = 'v', msg = 'add'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb7058a9510>
op = 'add'
lhs_data = tensor([[[[1.7380, 1.4873, 1.3962],
          [1.2878, 1.6336, 1.4275],
          [1.6374, 1.5678, 1.4996]]],


      ...5],
          [1.7210, 1.7499, 1.4188],
          [1.0540, 1.8867, 1.9607]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.3642, 1.4802, 1.3580]],

         [[1.7799, 1.4222, 1.1960]],

         [[1.1043, 1.1521, 1.2609]],

    ...

         [[1.4549, 1.2572, 1.0855]],

         [[1.3074, 1.7075, 1.6027]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: u_add_v)
_____________________ test_sddmm[idtype0-add-v-u-shp2-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'u', rhs_target = 'v', msg = 'add'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72be673f0>
op = 'add'
lhs_data = tensor([[[[1.2837, 1.9466, 1.1785],
          [1.2740, 1.4659, 1.8227],
          [1.6424, 1.6604, 1.2553]]],


      ...5],
          [1.5501, 1.7726, 1.4151],
          [1.3469, 1.6255, 1.8233]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.4544, 1.2328, 1.0590]],

         [[1.6981, 1.8692, 1.4903]],

         [[1.2374, 1.0117, 1.2878]],

    ...

         [[1.8997, 1.5434, 1.2474]],

         [[1.0170, 1.0172, 1.1259]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([40, 4, 1, 3])
SDDMM(message func: u_add_v)
_____________________ test_sddmm[idtype0-add-v-u-shp3-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shap...oat64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'u', rhs_target = 'v', msg = 'add'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb705eb67e0>
op = 'add'
lhs_data = tensor([[1.9630, 1.5424, 1.1294],
        [1.5116, 1.2992, 1.6958],
        [1.6111, 1.6575, 1.8277],
        [1.1278,... 1.9328],
        [1.9040, 1.0812, 1.4000],
        [1.6977, 1.7979, 1.6230]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.8542, 1.1590, 1.4684],
        [1.2662, 1.1176, 1.1915],
        [1.1098, 1.4203, 1.9821],
        [1.0650,... 1.7917],
        [1.8930, 1.2744, 1.0372],
        [1.5143, 1.0951, 1.8238]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: u_add_v)
_____________________ test_sddmm[idtype0-add-v-u-shp3-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'u', rhs_target = 'v', msg = 'add'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72ab770d8>
op = 'add'
lhs_data = tensor([[1.3009, 1.6470, 1.3698],
        [1.0625, 1.9079, 1.4566],
        [1.8825, 1.1185, 1.9489],
        [1.5760,... 1.0705],
        [1.7345, 1.7494, 1.5528],
        [1.9797, 1.7175, 1.6708]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.2957, 1.0841, 1.3347],
        [1.7947, 1.3145, 1.5822],
        [1.2484, 1.8566, 1.1756],
        [1.6312,... 1.5118],
        [1.6252, 1.8418, 1.6607],
        [1.5560, 1.8869, 1.9699]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([40, 3])
SDDMM(message func: u_add_v)
_____________________ test_sddmm[idtype0-add-v-u-shp4-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...oat64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'u', rhs_target = 'v', msg = 'add'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72c1eb828>
op = 'add'
lhs_data = tensor([[1.0514],
        [1.7405],
        [1.5707],
        [1.1335],
        [1.5358],
        [1.9162],
        [1...002],
        [1.2999],
        [1.7387],
        [1.3456],
        [1.7520]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.9114],
        [1.6256],
        [1.5578],
        [1.4830],
        [1.0198],
        [1.2491],
        [1...301],
        [1.7675],
        [1.6689],
        [1.5014],
        [1.0462]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: u_add_v)
_____________________ test_sddmm[idtype0-add-v-u-shp4-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'u', rhs_target = 'v', msg = 'add'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43e502240>
op = 'add'
lhs_data = tensor([[1.2814],
        [1.5008],
        [1.8876],
        [1.1230],
        [1.9805],
        [1.8600],
        [1...433],
        [1.2718],
        [1.5478],
        [1.4223],
        [1.2122]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.9984],
        [1.1118],
        [1.7580],
        [1.0512],
        [1.8434],
        [1.0435],
        [1...787],
        [1.0996],
        [1.9548],
        [1.6559],
        [1.1939]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([40, 1])
SDDMM(message func: u_add_v)
_____________________ test_sddmm[idtype0-add-v-e-shp0-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...   edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'e', rhs_target = 'v'
msg = 'add', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb45403efc0>
op = 'add'
lhs_data = tensor([[[[[[1.8348],
            [1.2017],
            [1.2853]]],


          [[[1.1527],
            [1.4213],
    ...2]]],


          [[[1.2267],
            [1.2164],
            [1.0098]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.0514]],

           [[1.7855]],

           [[1.6194]]]],



         [[[[1.9627]],

           [[1.848...



         [[[[1.4108]],

           [[1.3004]],

           [[1.4338]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([100, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: e_add_v)
_____________________ test_sddmm[idtype0-add-v-e-shp0-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'e', rhs_target = 'v'
msg = 'add', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72ab79480>
op = 'add'
lhs_data = tensor([[[[[[1.9786],
            [1.6414],
            [1.8295]]],


          [[[1.8069],
            [1.8186],
    ...3]]],


          [[[1.6786],
            [1.6462],
            [1.3912]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.3598]],

           [[1.2807]],

           [[1.6749]]]],



         [[[[1.6971]],

           [[1.559...



         [[[[1.1644]],

           [[1.3895]],

           [[1.7157]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([300, 1, 2, 1, 3, 1]), rhs shape: torch.Size([40, 4, 1, 3, 1, 1])
SDDMM(message func: e_add_v)
_____________________ test_sddmm[idtype0-add-v-e-shp1-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'e', rhs_target = 'v'
msg = 'add', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb45403e630>
op = 'add'
lhs_data = tensor([[[[[1.9010, 1.6847, 1.7533,  ..., 1.5834, 1.0659, 1.0702]],

          [[1.2109, 1.8056, 1.6480,  ..., 1.7236,...
          [[1.0198, 1.4905, 1.4074,  ..., 1.9250, 1.7386, 1.5887]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.6652, 1.7560, 1.4818,  ..., 1.1206, 1.9778, 1.2700],
           [1.4747, 1.7307, 1.3616,  ..., 1.0228, 1...
           [1.3968, 1.7973, 1.1483,  ..., 1.0940, 1.4536, 1.7569]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([100, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: e_add_v)
_____________________ test_sddmm[idtype0-add-v-e-shp1-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'e', rhs_target = 'v'
msg = 'add', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb705bec510>
op = 'add'
lhs_data = tensor([[[[[1.6189, 1.5536, 1.1857,  ..., 1.5209, 1.7754, 1.6996]],

          [[1.8784, 1.0455, 1.2865,  ..., 1.5965,...
          [[1.8087, 1.5185, 1.4866,  ..., 1.0271, 1.4395, 1.5471]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.2417, 1.2277, 1.0286,  ..., 1.0465, 1.5529, 1.5912],
           [1.6433, 1.9731, 1.7465,  ..., 1.3891, 1...
           [1.9356, 1.9918, 1.1826,  ..., 1.4598, 1.4077, 1.8870]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([300, 5, 3, 1, 7]), rhs shape: torch.Size([40, 1, 3, 7, 7])
SDDMM(message func: e_add_v)
_____________________ test_sddmm[idtype0-add-v-e-shp2-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1, 3, 3), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'e', rhs_target = 'v', msg = 'add'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43e964dc8>
op = 'add'
lhs_data = tensor([[[[1.2689, 1.3641, 1.6528],
          [1.3216, 1.8237, 1.0103],
          [1.7499, 1.0792, 1.6611]]],


      ...9],
          [1.9066, 1.6713, 1.0165],
          [1.1664, 1.5703, 1.3968]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.1734, 1.2236, 1.8234]],

         [[1.4926, 1.3726, 1.2404]],

         [[1.9823, 1.1630, 1.1617]],

    ...

         [[1.4622, 1.3394, 1.7800]],

         [[1.3954, 1.9458, 1.8851]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([100, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: e_add_v)
_____________________ test_sddmm[idtype0-add-v-e-shp2-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'e', rhs_target = 'v', msg = 'add'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72ae4b900>
op = 'add'
lhs_data = tensor([[[[1.4721, 1.6217, 1.2488],
          [1.2645, 1.1408, 1.7224],
          [1.6077, 1.9375, 1.6520]]],


      ...1],
          [1.2980, 1.0875, 1.0107],
          [1.0501, 1.5420, 1.0437]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.8635, 1.1245, 1.7292]],

         [[1.9708, 1.7757, 1.7939]],

         [[1.7459, 1.2101, 1.7835]],

    ...

         [[1.7843, 1.0234, 1.7616]],

         [[1.8896, 1.2307, 1.7727]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([300, 1, 3, 3]), rhs shape: torch.Size([40, 4, 1, 3])
SDDMM(message func: e_add_v)
_____________________ test_sddmm[idtype0-add-v-e-shp3-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...oat64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(3,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'e', rhs_target = 'v', msg = 'add'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72ab70168>
op = 'add'
lhs_data = tensor([[1.4212, 1.2400, 1.6342],
        [1.4740, 1.2618, 1.4753],
        [1.3159, 1.1574, 1.0400],
        [1.8511,... 1.0276],
        [1.1664, 1.2244, 1.7467],
        [1.8909, 1.7304, 1.1327]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.9218, 1.2511, 1.3874],
        [1.4529, 1.2632, 1.9143],
        [1.2529, 1.9165, 1.7794],
        [1.5148,... 1.1943],
        [1.3177, 1.1569, 1.9293],
        [1.7101, 1.6853, 1.9070]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1, 3, 3), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([100, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: e_add_v)
_____________________ test_sddmm[idtype0-add-v-e-shp3-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'e', rhs_target = 'v', msg = 'add'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72ab70438>
op = 'add'
lhs_data = tensor([[1.4955, 1.5018, 1.8180],
        [1.6861, 1.6213, 1.6839],
        [1.8650, 1.4055, 1.4727],
        [1.9067,... 1.1526],
        [1.7702, 1.2675, 1.6253],
        [1.4623, 1.4108, 1.5381]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.4586, 1.0319, 1.0246],
        [1.4173, 1.6825, 1.3357],
        [1.5985, 1.1848, 1.2461],
        [1.2314,... 1.8365],
        [1.3689, 1.1988, 1.7032],
        [1.6507, 1.7138, 1.3947]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([300, 3]), rhs shape: torch.Size([40, 3])
SDDMM(message func: e_add_v)
_____________________ test_sddmm[idtype0-add-v-e-shp4-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...oat64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'e', rhs_target = 'v', msg = 'add'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72b398708>
op = 'add'
lhs_data = tensor([[1.8919],
        [1.0691],
        [1.5827],
        [1.3674],
        [1.7022],
        [1.5541],
        [1...699],
        [1.8622],
        [1.9631],
        [1.7535],
        [1.2969]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.6937],
        [1.0336],
        [1.9698],
        [1.3445],
        [1.7812],
        [1.5435],
        [1...898],
        [1.4933],
        [1.5754],
        [1.4490],
        [1.2062]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(3,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([100, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: e_add_v)
_____________________ test_sddmm[idtype0-add-v-e-shp4-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'e', rhs_target = 'v', msg = 'add'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72ae4c948>
op = 'add'
lhs_data = tensor([[1.1057],
        [1.1243],
        [1.1071],
        [1.5136],
        [1.4289],
        [1.2135],
        [1...757],
        [1.5487],
        [1.1781],
        [1.3181],
        [1.9594]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.3934],
        [1.2186],
        [1.3258],
        [1.1676],
        [1.4463],
        [1.2933],
        [1...447],
        [1.5527],
        [1.2380],
        [1.0162],
        [1.5596]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([300, 1]), rhs shape: torch.Size([40, 1])
SDDMM(message func: e_add_v)
_____________________ test_sddmm[idtype0-add-e-u-shp0-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': ...torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'u', rhs_target = 'e'
msg = 'add', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43e94ad38>
op = 'add'
lhs_data = tensor([[[[[[1.3182],
            [1.7467],
            [1.8196]]],


          [[[1.8710],
            [1.3245],
    ...3]]],


          [[[1.3035],
            [1.4923],
            [1.3409]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.0560]],

           [[1.8389]],

           [[1.8469]]]],



         [[[[1.4838]],

           [[1.251...



         [[[[1.9405]],

           [[1.3624]],

           [[1.7716]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([100, 4, 1, 3, 1, 1])
SDDMM(message func: u_add_e)
_____________________ test_sddmm[idtype0-add-e-u-shp0-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'u', rhs_target = 'e'
msg = 'add', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd8798>
op = 'add'
lhs_data = tensor([[[[[[1.2427],
            [1.8508],
            [1.5183]]],


          [[[1.9052],
            [1.7929],
    ...9]]],


          [[[1.7909],
            [1.0677],
            [1.1393]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.7622]],

           [[1.0383]],

           [[1.7905]]]],



         [[[[1.4409]],

           [[1.725...



         [[[[1.4596]],

           [[1.6818]],

           [[1.4708]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([300, 4, 1, 3, 1, 1])
SDDMM(message func: u_add_e)
_____________________ test_sddmm[idtype0-add-e-u-shp1-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Sch...pe=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'u', rhs_target = 'e'
msg = 'add', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72b7b46c0>
op = 'add'
lhs_data = tensor([[[[[1.3765, 1.5122, 1.0759,  ..., 1.1604, 1.0545, 1.2619]],

          [[1.1839, 1.7075, 1.0729,  ..., 1.0561,...
          [[1.2405, 1.6166, 1.6727,  ..., 1.9083, 1.4142, 1.5215]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.5225, 1.2520, 1.9882,  ..., 1.3434, 1.8837, 1.2536],
           [1.6408, 1.9410, 1.3518,  ..., 1.7068, 1...
           [1.2195, 1.8964, 1.3815,  ..., 1.3302, 1.6177, 1.2800]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([100, 1, 3, 7, 7])
SDDMM(message func: u_add_e)
_____________________ test_sddmm[idtype0-add-e-u-shp1-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'u', rhs_target = 'e'
msg = 'add', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72b7b41f8>
op = 'add'
lhs_data = tensor([[[[[1.4264, 1.1700, 1.3388,  ..., 1.3167, 1.3740, 1.1690]],

          [[1.1355, 1.6637, 1.6083,  ..., 1.3083,...
          [[1.2077, 1.8140, 1.1236,  ..., 1.1342, 1.6144, 1.0424]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.0250, 1.4934, 1.4195,  ..., 1.1888, 1.0988, 1.5471],
           [1.5330, 1.1044, 1.1095,  ..., 1.9804, 1...
           [1.2404, 1.9778, 1.6176,  ..., 1.5283, 1.9249, 1.4107]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([300, 1, 3, 7, 7])
SDDMM(message func: u_add_e)
_____________________ test_sddmm[idtype0-add-e-u-shp2-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme...dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'u', rhs_target = 'e', msg = 'add'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72959df30>
op = 'add'
lhs_data = tensor([[[[1.6728, 1.5199, 1.6890],
          [1.0570, 1.7991, 1.2200],
          [1.6365, 1.5917, 1.8294]]],


      ...3],
          [1.4660, 1.4284, 1.0933],
          [1.0637, 1.8455, 1.4445]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.5093, 1.2897, 1.8073]],

         [[1.5111, 1.0799, 1.6403]],

         [[1.7333, 1.6471, 1.9006]],

    ...

         [[1.2750, 1.9119, 1.4064]],

         [[1.4162, 1.8137, 1.7570]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([100, 4, 1, 3])
SDDMM(message func: u_add_e)
_____________________ test_sddmm[idtype0-add-e-u-shp2-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'u', rhs_target = 'e', msg = 'add'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72be4f870>
op = 'add'
lhs_data = tensor([[[[1.3608, 1.4903, 1.8132],
          [1.6964, 1.7953, 1.8057],
          [1.7030, 1.3190, 1.6544]]],


      ...4],
          [1.3493, 1.3269, 1.6550],
          [1.7677, 1.2500, 1.3304]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.9435, 1.7228, 1.5689]],

         [[1.6011, 1.7866, 1.0538]],

         [[1.8682, 1.7605, 1.3843]],

    ...

         [[1.1161, 1.3753, 1.6806]],

         [[1.0675, 1.5902, 1.6380]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([300, 4, 1, 3])
SDDMM(message func: u_add_e)
_____________________ test_sddmm[idtype0-add-e-u-shp3-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'u', rhs_target = 'e', msg = 'add'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb705a84828>
op = 'add'
lhs_data = tensor([[1.1106, 1.7558, 1.7219],
        [1.3467, 1.7263, 1.6285],
        [1.0207, 1.6002, 1.9305],
        [1.4801,... 1.7211],
        [1.6178, 1.0188, 1.8904],
        [1.6707, 1.8992, 1.6858]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.9395, 1.6470, 1.8619],
        [1.8267, 1.1218, 1.1443],
        [1.9013, 1.5941, 1.9913],
        [1.0502,... 1.0614],
        [1.7917, 1.6913, 1.4375],
        [1.4071, 1.0323, 1.8221]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([100, 3])
SDDMM(message func: u_add_e)
_____________________ test_sddmm[idtype0-add-e-u-shp3-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'u', rhs_target = 'e', msg = 'add'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72be481b0>
op = 'add'
lhs_data = tensor([[1.8790, 1.4365, 1.7450],
        [1.4910, 1.5349, 1.9491],
        [1.8633, 1.7817, 1.7897],
        [1.5309,... 1.6617],
        [1.8079, 1.8648, 1.6105],
        [1.3278, 1.0782, 1.5566]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.6801, 1.3952, 1.5587],
        [1.2780, 1.3047, 1.9477],
        [1.9791, 1.8801, 1.0297],
        [1.1928,... 1.1253],
        [1.8956, 1.0096, 1.9076],
        [1.5535, 1.3646, 1.7100]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([300, 3])
SDDMM(message func: u_add_e)
_____________________ test_sddmm[idtype0-add-e-u-shp4-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'u', rhs_target = 'e', msg = 'add'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72b79c090>
op = 'add'
lhs_data = tensor([[1.5515],
        [1.6934],
        [1.4825],
        [1.8167],
        [1.8272],
        [1.7604],
        [1...744],
        [1.1864],
        [1.8745],
        [1.1153],
        [1.6488]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.3276],
        [1.7462],
        [1.8108],
        [1.5461],
        [1.5792],
        [1.6030],
        [1...647],
        [1.7861],
        [1.3674],
        [1.6739],
        [1.4727]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([100, 1])
SDDMM(message func: u_add_e)
_____________________ test_sddmm[idtype0-add-e-u-shp4-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'u', rhs_target = 'e', msg = 'add'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72b79c318>
op = 'add'
lhs_data = tensor([[1.5466],
        [1.2443],
        [1.2504],
        [1.0087],
        [1.3137],
        [1.2640],
        [1...392],
        [1.2068],
        [1.8752],
        [1.8615],
        [1.8382]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.0225],
        [1.6585],
        [1.3485],
        [1.7222],
        [1.1808],
        [1.8202],
        [1...582],
        [1.3201],
        [1.1676],
        [1.5675],
        [1.1608]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([300, 1])
SDDMM(message func: u_add_e)
_____________________ test_sddmm[idtype0-add-e-v-shp0-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': ...torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'v', rhs_target = 'e'
msg = 'add', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43e6ce6c0>
op = 'add'
lhs_data = tensor([[[[[[1.3885],
            [1.5606],
            [1.4436]]],


          [[[1.4819],
            [1.7341],
    ...6]]],


          [[[1.2047],
            [1.9665],
            [1.7889]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.2827]],

           [[1.0668]],

           [[1.2797]]]],



         [[[[1.7287]],

           [[1.306...



         [[[[1.7929]],

           [[1.5924]],

           [[1.6002]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([100, 4, 1, 3, 1, 1])
SDDMM(message func: v_add_e)
_____________________ test_sddmm[idtype0-add-e-v-shp0-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'v', rhs_target = 'e'
msg = 'add', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43e94a048>
op = 'add'
lhs_data = tensor([[[[[[1.6949],
            [1.0547],
            [1.2245]]],


          [[[1.8198],
            [1.5672],
    ...2]]],


          [[[1.4554],
            [1.0871],
            [1.0510]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.4745]],

           [[1.3380]],

           [[1.9316]]]],



         [[[[1.3467]],

           [[1.848...



         [[[[1.8327]],

           [[1.3381]],

           [[1.7677]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([40, 1, 2, 1, 3, 1]), rhs shape: torch.Size([300, 4, 1, 3, 1, 1])
SDDMM(message func: v_add_e)
_____________________ test_sddmm[idtype0-add-e-v-shp1-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Sch...pe=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'v', rhs_target = 'e'
msg = 'add', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb70581ba20>
op = 'add'
lhs_data = tensor([[[[[1.9607, 1.7992, 1.8218,  ..., 1.0765, 1.1108, 1.4619]],

          [[1.7810, 1.2431, 1.4474,  ..., 1.9725,...
          [[1.9884, 1.7395, 1.3100,  ..., 1.1142, 1.9321, 1.4091]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.7343, 1.8412, 1.3918,  ..., 1.4117, 1.4401, 1.1439],
           [1.8906, 1.3137, 1.3277,  ..., 1.4540, 1...
           [1.2840, 1.7196, 1.5690,  ..., 1.7900, 1.8975, 1.2990]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([100, 1, 3, 7, 7])
SDDMM(message func: v_add_e)
_____________________ test_sddmm[idtype0-add-e-v-shp1-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'v', rhs_target = 'e'
msg = 'add', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb705bf59d8>
op = 'add'
lhs_data = tensor([[[[[1.5887, 1.4807, 1.8561,  ..., 1.4600, 1.3384, 1.6648]],

          [[1.4489, 1.2764, 1.9063,  ..., 1.3104,...
          [[1.5641, 1.3932, 1.5482,  ..., 1.8161, 1.2781, 1.4642]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.7264, 1.6589, 1.8781,  ..., 1.0546, 1.2430, 1.9720],
           [1.3804, 1.8149, 1.7901,  ..., 1.9633, 1...
           [1.1517, 1.7804, 1.7061,  ..., 1.8603, 1.4115, 1.8675]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([40, 5, 3, 1, 7]), rhs shape: torch.Size([300, 1, 3, 7, 7])
SDDMM(message func: v_add_e)
_____________________ test_sddmm[idtype0-add-e-v-shp2-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme...dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'v', rhs_target = 'e', msg = 'add'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb705add900>
op = 'add'
lhs_data = tensor([[[[1.4317, 1.7778, 1.9126],
          [1.0413, 1.9944, 1.1500],
          [1.1820, 1.4158, 1.6368]]],


      ...4],
          [1.3945, 1.1196, 1.6662],
          [1.4486, 1.4908, 1.5227]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.2106, 1.2897, 1.7018]],

         [[1.5544, 1.1736, 1.0461]],

         [[1.2291, 1.5262, 1.3196]],

    ...

         [[1.5052, 1.8467, 1.6291]],

         [[1.3836, 1.1778, 1.9288]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([100, 4, 1, 3])
SDDMM(message func: v_add_e)
_____________________ test_sddmm[idtype0-add-e-v-shp2-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'v', rhs_target = 'e', msg = 'add'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72b786288>
op = 'add'
lhs_data = tensor([[[[1.0326, 1.3314, 1.6695],
          [1.2188, 1.4740, 1.9369],
          [1.0520, 1.7883, 1.6307]]],


      ...1],
          [1.7007, 1.4183, 1.8906],
          [1.1979, 1.7953, 1.2726]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.4650, 1.4645, 1.5722]],

         [[1.0848, 1.9536, 1.9555]],

         [[1.2539, 1.7311, 1.5875]],

    ...

         [[1.5339, 1.2661, 1.2501]],

         [[1.1439, 1.4322, 1.3522]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([40, 1, 3, 3]), rhs shape: torch.Size([300, 4, 1, 3])
SDDMM(message func: v_add_e)
_____________________ test_sddmm[idtype0-add-e-v-shp3-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'v', rhs_target = 'e', msg = 'add'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb705addcf0>
op = 'add'
lhs_data = tensor([[1.1756, 1.5400, 1.3763],
        [1.7192, 1.0583, 1.0300],
        [1.9592, 1.7213, 1.9990],
        [1.6791,... 1.5240],
        [1.2370, 1.7238, 1.3083],
        [1.7558, 1.8783, 1.4392]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.1220, 1.7448, 1.5011],
        [1.5140, 1.7289, 1.1700],
        [1.8325, 1.8836, 1.8819],
        [1.5871,... 1.4215],
        [1.8532, 1.6484, 1.9266],
        [1.4379, 1.6479, 1.5266]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([100, 3])
SDDMM(message func: v_add_e)
_____________________ test_sddmm[idtype0-add-e-v-shp3-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'v', rhs_target = 'e', msg = 'add'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43e6d65e8>
op = 'add'
lhs_data = tensor([[1.9384, 1.7409, 1.5343],
        [1.4391, 1.7512, 1.1807],
        [1.8738, 1.3219, 1.0369],
        [1.8073,... 1.6120],
        [1.2867, 1.4543, 1.5620],
        [1.0279, 1.4152, 1.9085]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.2564, 1.9507, 1.0500],
        [1.7466, 1.3935, 1.8556],
        [1.7761, 1.0994, 1.6592],
        [1.3661,... 1.6491],
        [1.3991, 1.1034, 1.7104],
        [1.3774, 1.5824, 1.7532]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([40, 3]), rhs shape: torch.Size([300, 3])
SDDMM(message func: v_add_e)
_____________________ test_sddmm[idtype0-add-e-v-shp4-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'v', rhs_target = 'e', msg = 'add'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb61f174c18>
op = 'add'
lhs_data = tensor([[1.0953],
        [1.4424],
        [1.0460],
        [1.3874],
        [1.2766],
        [1.0361],
        [1...885],
        [1.5673],
        [1.1326],
        [1.6869],
        [1.5008]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.7124],
        [1.4094],
        [1.3679],
        [1.2933],
        [1.0417],
        [1.4763],
        [1...711],
        [1.5248],
        [1.2009],
        [1.0568],
        [1.8730]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([100, 1])
SDDMM(message func: v_add_e)
_____________________ test_sddmm[idtype0-add-e-v-shp4-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'v', rhs_target = 'e', msg = 'add'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72ab60d80>
op = 'add'
lhs_data = tensor([[1.4844],
        [1.3888],
        [1.2782],
        [1.2396],
        [1.3067],
        [1.1250],
        [1...391],
        [1.6139],
        [1.6406],
        [1.8656],
        [1.2149]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.0968],
        [1.5469],
        [1.6676],
        [1.0499],
        [1.8701],
        [1.0500],
        [1...326],
        [1.1035],
        [1.0365],
        [1.5324],
        [1.5119]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([40, 1]), rhs shape: torch.Size([300, 1])
SDDMM(message func: v_add_e)
_____________________ test_sddmm[idtype0-sub-u-v-shp0-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': ...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'v', rhs_target = 'u'
msg = 'sub', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43e6c1678>
op = 'add'
lhs_data = tensor([[[[[[1.7140],
            [1.4623],
            [1.0240]]],


          [[[1.8692],
            [1.3161],
    ...3]]],


          [[[1.6545],
            [1.8443],
            [1.8956]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[-1.9901]],

           [[-1.3673]],

           [[-1.5700]]]],



         [[[[-1.4474]],

           [[-...       [[[[-1.8920]],

           [[-1.4425]],

           [[-1.8456]]]]]], dtype=torch.float64, grad_fn=<NegBackward>)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: v_sub_u)
_____________________ test_sddmm[idtype0-sub-u-v-shp0-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'v', rhs_target = 'u'
msg = 'sub', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72aceb558>
op = 'add'
lhs_data = tensor([[[[[[1.9886],
            [1.5255],
            [1.1917]]],


          [[[1.2358],
            [1.2023],
    ...1]]],


          [[[1.4054],
            [1.8964],
            [1.2379]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[-1.7722]],

           [[-1.0630]],

           [[-1.4461]]]],



         [[[[-1.0546]],

           [[-...       [[[[-1.4637]],

           [[-1.0492]],

           [[-1.0446]]]]]], dtype=torch.float64, grad_fn=<NegBackward>)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([40, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: v_sub_u)
_____________________ test_sddmm[idtype0-sub-u-v-shp1-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Sch...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'v', rhs_target = 'u'
msg = 'sub', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb457b56240>
op = 'add'
lhs_data = tensor([[[[[1.2816, 1.0215, 1.1330,  ..., 1.3306, 1.4379, 1.4088]],

          [[1.7173, 1.1584, 1.9245,  ..., 1.5517,...
          [[1.8020, 1.9259, 1.4166,  ..., 1.4450, 1.3227, 1.1492]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[-1.6460, -1.6524, -1.7613,  ..., -1.6273, -1.4486, -1.2625],
           [-1.5540, -1.5062, -1.0727,  ..., ...   [-1.2919, -1.5689, -1.3366,  ..., -1.5550, -1.3652, -1.6965]]]]],
       dtype=torch.float64, grad_fn=<NegBackward>)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: v_sub_u)
_____________________ test_sddmm[idtype0-sub-u-v-shp1-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'v', rhs_target = 'u'
msg = 'sub', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb447663480>
op = 'add'
lhs_data = tensor([[[[[1.3105, 1.0224, 1.4465,  ..., 1.8099, 1.2377, 1.6382]],

          [[1.1774, 1.9240, 1.5733,  ..., 1.0018,...
          [[1.7662, 1.4832, 1.0225,  ..., 1.8007, 1.1609, 1.4435]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[-1.8232, -1.9410, -1.3119,  ..., -1.6215, -1.2623, -1.6926],
           [-1.2476, -1.7712, -1.1300,  ..., ...   [-1.3166, -1.4520, -1.3450,  ..., -1.4764, -1.2231, -1.9802]]]]],
       dtype=torch.float64, grad_fn=<NegBackward>)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([40, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: v_sub_u)
_____________________ test_sddmm[idtype0-sub-u-v-shp2-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'v', rhs_target = 'u', msg = 'sub'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72b25e4c8>
op = 'add'
lhs_data = tensor([[[[1.3809, 1.4098, 1.5973],
          [1.7185, 1.1275, 1.1631],
          [1.7810, 1.2994, 1.2133]]],


      ...4],
          [1.6574, 1.8567, 1.4692],
          [1.2205, 1.1798, 1.4425]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[-1.1347, -1.2555, -1.2995]],

         [[-1.8279, -1.4552, -1.4824]],

         [[-1.7445, -1.2367, -1.5375...9478, -1.7650, -1.5422]],

         [[-1.6757, -1.5310, -1.7782]]]], dtype=torch.float64,
       grad_fn=<NegBackward>)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: v_sub_u)
_____________________ test_sddmm[idtype0-sub-u-v-shp2-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'v', rhs_target = 'u', msg = 'sub'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72b786630>
op = 'add'
lhs_data = tensor([[[[1.5953, 1.9727, 1.5552],
          [1.0004, 1.4514, 1.3318],
          [1.4430, 1.9176, 1.5854]]],


      ...8],
          [1.1954, 1.4314, 1.2788],
          [1.9671, 1.4402, 1.5781]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[-1.8191, -1.9559, -1.3273]],

         [[-1.7004, -1.0025, -1.8311]],

         [[-1.2558, -1.1849, -1.6456...1341, -1.1333, -1.4152]],

         [[-1.1560, -1.2788, -1.2003]]]], dtype=torch.float64,
       grad_fn=<NegBackward>)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([40, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: v_sub_u)
_____________________ test_sddmm[idtype0-sub-u-v-shp3-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'v', rhs_target = 'u', msg = 'sub'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acf2288>
op = 'add'
lhs_data = tensor([[1.9199, 1.9664, 1.2405],
        [1.8858, 1.9105, 1.8692],
        [1.0698, 1.7128, 1.4158],
        [1.3388,... 1.8260],
        [1.5388, 1.6964, 1.9341],
        [1.8112, 1.1176, 1.4489]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.9275, -1.7431, -1.5677],
        [-1.2185, -1.5232, -1.3316],
        [-1.0886, -1.8577, -1.8855],
       ...  [-1.1456, -1.7395, -1.9932],
        [-1.1605, -1.6553, -1.2659]], dtype=torch.float64,
       grad_fn=<NegBackward>)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: v_sub_u)
_____________________ test_sddmm[idtype0-sub-u-v-shp3-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'v', rhs_target = 'u', msg = 'sub'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acf2510>
op = 'add'
lhs_data = tensor([[1.4628, 1.0107, 1.7200],
        [1.2198, 1.2974, 1.3906],
        [1.4891, 1.1114, 1.7112],
        [1.3836,... 1.4484],
        [1.5310, 1.8565, 1.8642],
        [1.7278, 1.9260, 1.7245]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.3354, -1.4297, -1.4259],
        [-1.6993, -1.3022, -1.5048],
        [-1.5636, -1.0359, -1.6039],
       ...  [-1.0027, -1.1009, -1.8303],
        [-1.8741, -1.7820, -1.6216]], dtype=torch.float64,
       grad_fn=<NegBackward>)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([40, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: v_sub_u)
_____________________ test_sddmm[idtype0-sub-u-v-shp4-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'v', rhs_target = 'u', msg = 'sub'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb61f16a1b0>
op = 'add'
lhs_data = tensor([[1.8066],
        [1.7569],
        [1.8783],
        [1.3247],
        [1.9516],
        [1.3286],
        [1...014],
        [1.8699],
        [1.9959],
        [1.2787],
        [1.6449]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.8544],
        [-1.7816],
        [-1.6934],
        [-1.0280],
        [-1.1927],
        [-1.0965],
    ...       [-1.7039],
        [-1.3924],
        [-1.3948],
        [-1.0451]], dtype=torch.float64, grad_fn=<NegBackward>)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: v_sub_u)
_____________________ test_sddmm[idtype0-sub-u-v-shp4-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'v', rhs_target = 'u', msg = 'sub'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43e6d6438>
op = 'add'
lhs_data = tensor([[1.1734],
        [1.4784],
        [1.9976],
        [1.7887],
        [1.3108],
        [1.8983],
        [1...025],
        [1.2204],
        [1.5255],
        [1.6550],
        [1.1919]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.8278],
        [-1.9221],
        [-1.6205],
        [-1.9797],
        [-1.7465],
        [-1.9034],
    ...       [-1.6404],
        [-1.6407],
        [-1.7963],
        [-1.2850]], dtype=torch.float64, grad_fn=<NegBackward>)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([40, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: v_sub_u)
_____________________ test_sddmm[idtype0-sub-u-e-shp0-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...torch.float64), 'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'e', rhs_target = 'u'
msg = 'sub', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb705bec990>
op = 'add'
lhs_data = tensor([[[[[[1.1326],
            [1.5078],
            [1.0284]]],


          [[[1.5024],
            [1.9602],
    ...0]]],


          [[[1.3609],
            [1.4917],
            [1.9039]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[-1.8540]],

           [[-1.0611]],

           [[-1.8639]]]],



         [[[[-1.5523]],

           [[-...       [[[[-1.3935]],

           [[-1.9957]],

           [[-1.7640]]]]]], dtype=torch.float64, grad_fn=<NegBackward>)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([100, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: e_sub_u)
_____________________ test_sddmm[idtype0-sub-u-e-shp0-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'e', rhs_target = 'u'
msg = 'sub', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72ab56948>
op = 'add'
lhs_data = tensor([[[[[[1.2045],
            [1.2974],
            [1.8061]]],


          [[[1.8753],
            [1.2742],
    ...0]]],


          [[[1.7259],
            [1.5829],
            [1.6447]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[-1.7509]],

           [[-1.7344]],

           [[-1.2542]]]],



         [[[[-1.2360]],

           [[-...       [[[[-1.7859]],

           [[-1.6052]],

           [[-1.9099]]]]]], dtype=torch.float64, grad_fn=<NegBackward>)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([300, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: e_sub_u)
_____________________ test_sddmm[idtype0-sub-u-e-shp1-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...pe=torch.float64), 'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'e', rhs_target = 'u'
msg = 'sub', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acf31b0>
op = 'add'
lhs_data = tensor([[[[[1.3858, 1.2767, 1.7062,  ..., 1.6433, 1.3054, 1.6435]],

          [[1.3181, 1.0246, 1.9823,  ..., 1.3077,...
          [[1.3210, 1.4520, 1.0499,  ..., 1.0967, 1.5443, 1.2168]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[-1.2816, -1.7829, -1.2513,  ..., -1.3956, -1.9991, -1.1937],
           [-1.4111, -1.9037, -1.1999,  ..., ...   [-1.1115, -1.7049, -1.3863,  ..., -1.2395, -1.2385, -1.1670]]]]],
       dtype=torch.float64, grad_fn=<NegBackward>)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([100, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: e_sub_u)
_____________________ test_sddmm[idtype0-sub-u-e-shp1-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'e', rhs_target = 'u'
msg = 'sub', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acf3438>
op = 'add'
lhs_data = tensor([[[[[1.5803, 1.6559, 1.2028,  ..., 1.5431, 1.8840, 1.1356]],

          [[1.8486, 1.9700, 1.6580,  ..., 1.9245,...
          [[1.3931, 1.6218, 1.5771,  ..., 1.6866, 1.2030, 1.5753]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[-1.7111, -1.1975, -1.7693,  ..., -1.4811, -1.8071, -1.4275],
           [-1.7779, -1.5464, -1.8810,  ..., ...   [-1.4096, -1.1263, -1.0972,  ..., -1.9600, -1.0074, -1.8000]]]]],
       dtype=torch.float64, grad_fn=<NegBackward>)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([300, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: e_sub_u)
_____________________ test_sddmm[idtype0-sub-u-e-shp2-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...dtype=torch.float64), 'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'e', rhs_target = 'u', msg = 'sub'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72c1f1ea0>
op = 'add'
lhs_data = tensor([[[[1.5070, 1.8985, 1.3580],
          [1.0577, 1.8342, 1.6111],
          [1.7037, 1.7832, 1.5078]]],


      ...5],
          [1.5075, 1.5201, 1.7774],
          [1.9907, 1.5142, 1.2260]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[-1.9077, -1.9390, -1.2679]],

         [[-1.2849, -1.4536, -1.1528]],

         [[-1.0330, -1.4107, -1.4119...0954, -1.8068, -1.8094]],

         [[-1.6255, -1.5560, -1.3984]]]], dtype=torch.float64,
       grad_fn=<NegBackward>)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([100, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: e_sub_u)
_____________________ test_sddmm[idtype0-sub-u-e-shp2-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'e', rhs_target = 'u', msg = 'sub'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72be605e8>
op = 'add'
lhs_data = tensor([[[[1.8100, 1.9988, 1.2428],
          [1.6300, 1.3758, 1.3934],
          [1.9459, 1.0296, 1.3165]]],


      ...0],
          [1.9432, 1.1649, 1.4549],
          [1.1405, 1.8687, 1.4798]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[-1.7357, -1.0781, -1.4468]],

         [[-1.0036, -1.1657, -1.0049]],

         [[-1.0611, -1.2706, -1.5746...6739, -1.4804, -1.1277]],

         [[-1.4679, -1.2661, -1.5980]]]], dtype=torch.float64,
       grad_fn=<NegBackward>)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([300, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: e_sub_u)
_____________________ test_sddmm[idtype0-sub-u-e-shp3-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'e', rhs_target = 'u', msg = 'sub'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb455169a68>
op = 'add'
lhs_data = tensor([[1.7209, 1.2186, 1.4605],
        [1.7579, 1.5613, 1.0189],
        [1.8783, 1.8844, 1.0266],
        [1.4332,... 1.5734],
        [1.5511, 1.8638, 1.5453],
        [1.5880, 1.1228, 1.7772]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.3697, -1.9308, -1.5695],
        [-1.1532, -1.7045, -1.0612],
        [-1.5865, -1.4985, -1.8168],
       ...  [-1.4845, -1.5549, -1.5169],
        [-1.7183, -1.4201, -1.7051]], dtype=torch.float64,
       grad_fn=<NegBackward>)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([100, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: e_sub_u)
_____________________ test_sddmm[idtype0-sub-u-e-shp3-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'e', rhs_target = 'u', msg = 'sub'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72951eaf8>
op = 'add'
lhs_data = tensor([[1.3053, 1.0704, 1.6603],
        [1.0259, 1.8147, 1.1821],
        [1.4831, 1.5738, 1.5009],
        [1.4151,... 1.0321],
        [1.0499, 1.8114, 1.6448],
        [1.5473, 1.8926, 1.2115]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.8676, -1.8558, -1.2979],
        [-1.0912, -1.6381, -1.7749],
        [-1.2295, -1.3842, -1.2967],
       ...  [-1.5675, -1.1446, -1.6194],
        [-1.8675, -1.6849, -1.1766]], dtype=torch.float64,
       grad_fn=<NegBackward>)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([300, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: e_sub_u)
_____________________ test_sddmm[idtype0-sub-u-e-shp4-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'e', rhs_target = 'u', msg = 'sub'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acf8120>
op = 'add'
lhs_data = tensor([[1.0061],
        [1.1296],
        [1.3401],
        [1.7204],
        [1.7252],
        [1.1249],
        [1...056],
        [1.0773],
        [1.1206],
        [1.4145],
        [1.1420]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.5325],
        [-1.5684],
        [-1.5157],
        [-1.5332],
        [-1.6570],
        [-1.1807],
    ...       [-1.2791],
        [-1.4990],
        [-1.0937],
        [-1.8953]], dtype=torch.float64, grad_fn=<NegBackward>)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([100, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: e_sub_u)
_____________________ test_sddmm[idtype0-sub-u-e-shp4-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'e', rhs_target = 'u', msg = 'sub'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acf83a8>
op = 'add'
lhs_data = tensor([[1.8047],
        [1.9109],
        [1.2445],
        [1.1650],
        [1.3552],
        [1.9893],
        [1...630],
        [1.7468],
        [1.2085],
        [1.0908],
        [1.7873]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.4198],
        [-1.5899],
        [-1.5218],
        [-1.4942],
        [-1.5583],
        [-1.4868],
    ...       [-1.2091],
        [-1.2143],
        [-1.1258],
        [-1.5939]], dtype=torch.float64, grad_fn=<NegBackward>)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([300, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: e_sub_u)
_____________________ test_sddmm[idtype0-sub-v-u-shp0-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': ...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'u', rhs_target = 'v'
msg = 'sub', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb705815fc0>
op = 'add'
lhs_data = tensor([[[[[[1.8081],
            [1.4536],
            [1.4497]]],


          [[[1.8999],
            [1.1807],
    ...2]]],


          [[[1.3628],
            [1.2081],
            [1.9578]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[-1.1370]],

           [[-1.1055]],

           [[-1.1089]]]],



         [[[[-1.9001]],

           [[-...       [[[[-1.4882]],

           [[-1.3484]],

           [[-1.5080]]]]]], dtype=torch.float64, grad_fn=<NegBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: u_sub_v)
_____________________ test_sddmm[idtype0-sub-v-u-shp0-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'u', rhs_target = 'v'
msg = 'sub', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb70623e090>
op = 'add'
lhs_data = tensor([[[[[[1.6168],
            [1.2553],
            [1.8081]]],


          [[[1.5411],
            [1.9248],
    ...7]]],


          [[[1.7961],
            [1.8892],
            [1.5590]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[-1.0837]],

           [[-1.6445]],

           [[-1.1309]]]],



         [[[[-1.2349]],

           [[-...       [[[[-1.4796]],

           [[-1.2780]],

           [[-1.4309]]]]]], dtype=torch.float64, grad_fn=<NegBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([40, 4, 1, 3, 1, 1])
SDDMM(message func: u_sub_v)
_____________________ test_sddmm[idtype0-sub-v-u-shp1-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Sch...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'u', rhs_target = 'v'
msg = 'sub', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb455169a68>
op = 'add'
lhs_data = tensor([[[[[1.0089, 1.0754, 1.3368,  ..., 1.0694, 1.4001, 1.0843]],

          [[1.8735, 1.0639, 1.8514,  ..., 1.3515,...
          [[1.5492, 1.1995, 1.5861,  ..., 1.2342, 1.1959, 1.7999]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[-1.0322, -1.9287, -1.2554,  ..., -1.5493, -1.7238, -1.8060],
           [-1.6334, -1.6815, -1.2033,  ..., ...   [-1.9476, -1.2942, -1.4769,  ..., -1.1402, -1.0428, -1.5715]]]]],
       dtype=torch.float64, grad_fn=<NegBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: u_sub_v)
_____________________ test_sddmm[idtype0-sub-v-u-shp1-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'u', rhs_target = 'v'
msg = 'sub', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb705b25e10>
op = 'add'
lhs_data = tensor([[[[[1.9427, 1.3570, 1.3517,  ..., 1.8819, 1.8076, 1.8767]],

          [[1.4544, 1.3232, 1.4537,  ..., 1.8796,...
          [[1.5264, 1.6469, 1.2291,  ..., 1.9659, 1.8957, 1.4669]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[-1.6779, -1.5449, -1.0382,  ..., -1.2904, -1.1138, -1.3468],
           [-1.3496, -1.4456, -1.5459,  ..., ...   [-1.4773, -1.1194, -1.0116,  ..., -1.5467, -1.6157, -1.9734]]]]],
       dtype=torch.float64, grad_fn=<NegBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([40, 1, 3, 7, 7])
SDDMM(message func: u_sub_v)
_____________________ test_sddmm[idtype0-sub-v-u-shp2-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'u', rhs_target = 'v', msg = 'sub'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acfc048>
op = 'add'
lhs_data = tensor([[[[1.7458, 1.7657, 1.4151],
          [1.2600, 1.4569, 1.5367],
          [1.3295, 1.9821, 1.5931]]],


      ...1],
          [1.0123, 1.0301, 1.5919],
          [1.3665, 1.2468, 1.8996]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[-1.2060, -1.3859, -1.0921]],

         [[-1.1290, -1.2300, -1.2926]],

         [[-1.6298, -1.6866, -1.7358...5441, -1.1778, -1.5040]],

         [[-1.3027, -1.5394, -1.3624]]]], dtype=torch.float64,
       grad_fn=<NegBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: u_sub_v)
_____________________ test_sddmm[idtype0-sub-v-u-shp2-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'u', rhs_target = 'v', msg = 'sub'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acfc318>
op = 'add'
lhs_data = tensor([[[[1.5321, 1.0305, 1.5327],
          [1.6850, 1.6536, 1.1687],
          [1.9196, 1.7736, 1.1196]]],


      ...5],
          [1.1595, 1.2387, 1.4325],
          [1.5122, 1.4423, 1.0098]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[-1.4588, -1.9205, -1.0348]],

         [[-1.7850, -1.9364, -1.7897]],

         [[-1.7497, -1.2994, -1.7457...6998, -1.3310, -1.3543]],

         [[-1.7408, -1.8697, -1.7401]]]], dtype=torch.float64,
       grad_fn=<NegBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([40, 4, 1, 3])
SDDMM(message func: u_sub_v)
_____________________ test_sddmm[idtype0-sub-v-u-shp3-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'u', rhs_target = 'v', msg = 'sub'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72ab70f78>
op = 'add'
lhs_data = tensor([[1.6208, 1.9368, 1.9398],
        [1.9322, 1.3534, 1.5727],
        [1.2422, 1.1381, 1.2221],
        [1.6868,... 1.6796],
        [1.5874, 1.9893, 1.5712],
        [1.7673, 1.1076, 1.7648]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.9574, -1.1564, -1.2574],
        [-1.0205, -1.3573, -1.4705],
        [-1.2945, -1.1003, -1.7481],
       ...  [-1.2133, -1.2058, -1.0585],
        [-1.1631, -1.9594, -1.1228]], dtype=torch.float64,
       grad_fn=<NegBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: u_sub_v)
_____________________ test_sddmm[idtype0-sub-v-u-shp3-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'u', rhs_target = 'v', msg = 'sub'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72b398870>
op = 'add'
lhs_data = tensor([[1.7514, 1.1417, 1.6500],
        [1.2653, 1.0678, 1.8890],
        [1.7940, 1.8100, 1.4603],
        [1.2650,... 1.7796],
        [1.3602, 1.9827, 1.0712],
        [1.4779, 1.9577, 1.1282]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.2395, -1.5719, -1.0772],
        [-1.5042, -1.5249, -1.8229],
        [-1.0387, -1.6722, -1.5680],
       ...  [-1.4353, -1.8991, -1.3293],
        [-1.3211, -1.1740, -1.4684]], dtype=torch.float64,
       grad_fn=<NegBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([40, 3])
SDDMM(message func: u_sub_v)
_____________________ test_sddmm[idtype0-sub-v-u-shp4-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'u', rhs_target = 'v', msg = 'sub'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb705b2b900>
op = 'add'
lhs_data = tensor([[1.2030],
        [1.0913],
        [1.4274],
        [1.8483],
        [1.2501],
        [1.5615],
        [1...957],
        [1.5223],
        [1.9396],
        [1.6775],
        [1.1842]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.1089],
        [-1.8899],
        [-1.4904],
        [-1.8176],
        [-1.9490],
        [-1.2959],
    ...       [-1.0110],
        [-1.8145],
        [-1.1015],
        [-1.4067]], dtype=torch.float64, grad_fn=<NegBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: u_sub_v)
_____________________ test_sddmm[idtype0-sub-v-u-shp4-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'u', rhs_target = 'v', msg = 'sub'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb705c04558>
op = 'add'
lhs_data = tensor([[1.7648],
        [1.8107],
        [1.0529],
        [1.4188],
        [1.8539],
        [1.3416],
        [1...053],
        [1.2170],
        [1.1587],
        [1.4561],
        [1.9027]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.1669],
        [-1.6292],
        [-1.7404],
        [-1.0103],
        [-1.9296],
        [-1.1864],
    ...       [-1.3927],
        [-1.2513],
        [-1.4913],
        [-1.9809]], dtype=torch.float64, grad_fn=<NegBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([40, 1])
SDDMM(message func: u_sub_v)
_____________________ test_sddmm[idtype0-sub-v-e-shp0-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...torch.float64), 'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'e', rhs_target = 'v'
msg = 'sub', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb61dd1bea0>
op = 'add'
lhs_data = tensor([[[[[[1.4307],
            [1.2255],
            [1.9464]]],


          [[[1.8208],
            [1.8948],
    ...3]]],


          [[[1.1838],
            [1.6442],
            [1.5326]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[-1.4348]],

           [[-1.4004]],

           [[-1.8251]]]],



         [[[[-1.6433]],

           [[-...       [[[[-1.8061]],

           [[-1.0103]],

           [[-1.1006]]]]]], dtype=torch.float64, grad_fn=<NegBackward>)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([100, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: e_sub_v)
_____________________ test_sddmm[idtype0-sub-v-e-shp0-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'e', rhs_target = 'v'
msg = 'sub', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd3c18>
op = 'add'
lhs_data = tensor([[[[[[1.1560],
            [1.8049],
            [1.8268]]],


          [[[1.5685],
            [1.7378],
    ...7]]],


          [[[1.6850],
            [1.9892],
            [1.9028]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[-1.6812]],

           [[-1.1216]],

           [[-1.1940]]]],



         [[[[-1.1013]],

           [[-...       [[[[-1.9405]],

           [[-1.8228]],

           [[-1.3331]]]]]], dtype=torch.float64, grad_fn=<NegBackward>)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([300, 1, 2, 1, 3, 1]), rhs shape: torch.Size([40, 4, 1, 3, 1, 1])
SDDMM(message func: e_sub_v)
_____________________ test_sddmm[idtype0-sub-v-e-shp1-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...pe=torch.float64), 'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'e', rhs_target = 'v'
msg = 'sub', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72be48288>
op = 'add'
lhs_data = tensor([[[[[1.2375, 1.9205, 1.4429,  ..., 1.5872, 1.1565, 1.1972]],

          [[1.9490, 1.9156, 1.6149,  ..., 1.4956,...
          [[1.2211, 1.1575, 1.7708,  ..., 1.6560, 1.6581, 1.0635]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[-1.8236, -1.4624, -1.8960,  ..., -1.4054, -1.7823, -1.3413],
           [-1.0317, -1.8734, -1.9155,  ..., ...   [-1.2895, -1.4679, -1.4172,  ..., -1.6373, -1.3458, -1.5519]]]]],
       dtype=torch.float64, grad_fn=<NegBackward>)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([100, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: e_sub_v)
_____________________ test_sddmm[idtype0-sub-v-e-shp1-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'e', rhs_target = 'v'
msg = 'sub', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72ad01168>
op = 'add'
lhs_data = tensor([[[[[1.0277, 1.5589, 1.4055,  ..., 1.1990, 1.3343, 1.7606]],

          [[1.4875, 1.3194, 1.2416,  ..., 1.4541,...
          [[1.9767, 1.1350, 1.3671,  ..., 1.9145, 1.5117, 1.2325]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[-1.2792, -1.9323, -1.0463,  ..., -1.4622, -1.6796, -1.8928],
           [-1.3387, -1.7153, -1.9928,  ..., ...   [-1.7626, -1.1240, -1.8576,  ..., -1.3892, -1.0745, -1.0700]]]]],
       dtype=torch.float64, grad_fn=<NegBackward>)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([300, 5, 3, 1, 7]), rhs shape: torch.Size([40, 1, 3, 7, 7])
SDDMM(message func: e_sub_v)
_____________________ test_sddmm[idtype0-sub-v-e-shp2-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...dtype=torch.float64), 'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'e', rhs_target = 'v', msg = 'sub'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb70584c090>
op = 'add'
lhs_data = tensor([[[[1.8448, 1.0490, 1.9188],
          [1.7362, 1.1335, 1.6391],
          [1.5041, 1.7695, 1.8473]]],


      ...6],
          [1.5217, 1.7575, 1.9225],
          [1.4399, 1.9825, 1.6804]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[-1.1776, -1.8140, -1.9528]],

         [[-1.4693, -1.1260, -1.8132]],

         [[-1.9682, -1.1734, -1.9685...5300, -1.8994, -1.4045]],

         [[-1.9913, -1.1112, -1.5052]]]], dtype=torch.float64,
       grad_fn=<NegBackward>)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([100, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: e_sub_v)
_____________________ test_sddmm[idtype0-sub-v-e-shp2-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'e', rhs_target = 'v', msg = 'sub'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb61f15fa68>
op = 'add'
lhs_data = tensor([[[[1.2061, 1.2850, 1.5168],
          [1.9213, 1.0192, 1.8230],
          [1.8577, 1.6882, 1.1419]]],


      ...1],
          [1.9567, 1.9147, 1.0811],
          [1.9820, 1.4205, 1.7724]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[-1.2487, -1.4802, -1.0189]],

         [[-1.9741, -1.3896, -1.6785]],

         [[-1.4235, -1.0481, -1.5631...8703, -1.5060, -1.7861]],

         [[-1.7890, -1.5064, -1.0124]]]], dtype=torch.float64,
       grad_fn=<NegBackward>)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([300, 1, 3, 3]), rhs shape: torch.Size([40, 4, 1, 3])
SDDMM(message func: e_sub_v)
_____________________ test_sddmm[idtype0-sub-v-e-shp3-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'e', rhs_target = 'v', msg = 'sub'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb4551662d0>
op = 'add'
lhs_data = tensor([[1.5157, 1.2284, 1.5483],
        [1.5135, 1.7591, 1.5663],
        [1.3108, 1.2684, 1.0985],
        [1.0280,... 1.5126],
        [1.7437, 1.5501, 1.5671],
        [1.6449, 1.9118, 1.7475]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.8201, -1.5074, -1.0954],
        [-1.8977, -1.9382, -1.1252],
        [-1.8844, -1.6681, -1.4401],
       ...  [-1.8838, -1.2508, -1.3143],
        [-1.9232, -1.1838, -1.2590]], dtype=torch.float64,
       grad_fn=<NegBackward>)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([100, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: e_sub_v)
_____________________ test_sddmm[idtype0-sub-v-e-shp3-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'e', rhs_target = 'v', msg = 'sub'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72b3c2288>
op = 'add'
lhs_data = tensor([[1.7815, 1.3237, 1.1306],
        [1.9882, 1.7099, 1.0903],
        [1.3611, 1.7088, 1.9029],
        [1.1769,... 1.0613],
        [1.9354, 1.1731, 1.1064],
        [1.8829, 1.6280, 1.1951]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.6253, -1.9086, -1.8840],
        [-1.3809, -1.6925, -1.7389],
        [-1.1782, -1.1114, -1.2690],
       ...  [-1.1974, -1.4174, -1.7551],
        [-1.7545, -1.0210, -1.7438]], dtype=torch.float64,
       grad_fn=<NegBackward>)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([300, 3]), rhs shape: torch.Size([40, 3])
SDDMM(message func: e_sub_v)
_____________________ test_sddmm[idtype0-sub-v-e-shp4-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'e', rhs_target = 'v', msg = 'sub'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb457b5f318>
op = 'add'
lhs_data = tensor([[1.5129],
        [1.9186],
        [1.2115],
        [1.4729],
        [1.8979],
        [1.3045],
        [1...054],
        [1.0511],
        [1.6317],
        [1.8599],
        [1.2764]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.1489],
        [-1.6308],
        [-1.5941],
        [-1.8378],
        [-1.1118],
        [-1.5635],
    ...       [-1.4903],
        [-1.3743],
        [-1.8853],
        [-1.7077]], dtype=torch.float64, grad_fn=<NegBackward>)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([100, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: e_sub_v)
_____________________ test_sddmm[idtype0-sub-v-e-shp4-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'e', rhs_target = 'v', msg = 'sub'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72aa9c090>
op = 'add'
lhs_data = tensor([[1.5222],
        [1.7535],
        [1.0655],
        [1.1533],
        [1.7426],
        [1.5606],
        [1...285],
        [1.1362],
        [1.5242],
        [1.7668],
        [1.7805]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.4523],
        [-1.3801],
        [-1.6198],
        [-1.0440],
        [-1.3525],
        [-1.9313],
    ...       [-1.4147],
        [-1.5066],
        [-1.1817],
        [-1.2045]], dtype=torch.float64, grad_fn=<NegBackward>)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([300, 1]), rhs shape: torch.Size([40, 1])
SDDMM(message func: e_sub_v)
_____________________ test_sddmm[idtype0-sub-e-u-shp0-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': ...torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'u', rhs_target = 'e'
msg = 'sub', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72aa9c2d0>
op = 'add'
lhs_data = tensor([[[[[[1.7449],
            [1.7635],
            [1.8669]]],


          [[[1.1938],
            [1.3712],
    ...4]]],


          [[[1.6383],
            [1.6873],
            [1.3037]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[-1.2037]],

           [[-1.2268]],

           [[-1.5247]]]],



         [[[[-1.3501]],

           [[-...       [[[[-1.2458]],

           [[-1.9777]],

           [[-1.2258]]]]]], dtype=torch.float64, grad_fn=<NegBackward>)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([100, 4, 1, 3, 1, 1])
SDDMM(message func: u_sub_e)
_____________________ test_sddmm[idtype0-sub-e-u-shp0-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'u', rhs_target = 'e'
msg = 'sub', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72aa9c5e8>
op = 'add'
lhs_data = tensor([[[[[[1.2269],
            [1.7936],
            [1.1759]]],


          [[[1.1719],
            [1.4462],
    ...7]]],


          [[[1.3307],
            [1.6151],
            [1.4919]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[-1.6927]],

           [[-1.1266]],

           [[-1.2108]]]],



         [[[[-1.1311]],

           [[-...       [[[[-1.6490]],

           [[-1.3085]],

           [[-1.8337]]]]]], dtype=torch.float64, grad_fn=<NegBackward>)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([300, 4, 1, 3, 1, 1])
SDDMM(message func: u_sub_e)
_____________________ test_sddmm[idtype0-sub-e-u-shp1-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Sch...pe=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'u', rhs_target = 'e'
msg = 'sub', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb61f153ab0>
op = 'add'
lhs_data = tensor([[[[[1.6187, 1.1569, 1.3560,  ..., 1.2407, 1.2833, 1.0873]],

          [[1.6851, 1.4458, 1.0688,  ..., 1.9424,...
          [[1.2099, 1.6670, 1.6875,  ..., 1.8023, 1.8026, 1.5795]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[-1.1043, -1.4621, -1.4407,  ..., -1.2924, -1.5041, -1.3680],
           [-1.2097, -1.3834, -1.0276,  ..., ...   [-1.0314, -1.6095, -1.4041,  ..., -1.6749, -1.4250, -1.5672]]]]],
       dtype=torch.float64, grad_fn=<NegBackward>)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([100, 1, 3, 7, 7])
SDDMM(message func: u_sub_e)
_____________________ test_sddmm[idtype0-sub-e-u-shp1-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'u', rhs_target = 'e'
msg = 'sub', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43e6e6900>
op = 'add'
lhs_data = tensor([[[[[1.4718, 1.9919, 1.8539,  ..., 1.6008, 1.7537, 1.8268]],

          [[1.8985, 1.8986, 1.9053,  ..., 1.3601,...
          [[1.0590, 1.0174, 1.4985,  ..., 1.5533, 1.2952, 1.1889]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[-1.9965, -1.1440, -1.2350,  ..., -1.6257, -1.6736, -1.1314],
           [-1.3482, -1.9512, -1.7668,  ..., ...   [-1.7035, -1.7202, -1.5859,  ..., -1.8977, -1.8309, -1.3668]]]]],
       dtype=torch.float64, grad_fn=<NegBackward>)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([300, 1, 3, 7, 7])
SDDMM(message func: u_sub_e)
_____________________ test_sddmm[idtype0-sub-e-u-shp2-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme...dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'u', rhs_target = 'e', msg = 'sub'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb729666990>
op = 'add'
lhs_data = tensor([[[[1.9236, 1.7623, 1.7527],
          [1.5204, 1.9112, 1.8592],
          [1.2829, 1.1724, 1.6406]]],


      ...2],
          [1.6559, 1.2268, 1.0099],
          [1.0276, 1.6253, 1.7161]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[-1.3853, -1.4566, -1.2765]],

         [[-1.3921, -1.1360, -1.3648]],

         [[-1.2924, -1.8072, -1.7361...1739, -1.7188, -1.2760]],

         [[-1.1073, -1.9200, -1.7284]]]], dtype=torch.float64,
       grad_fn=<NegBackward>)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([100, 4, 1, 3])
SDDMM(message func: u_sub_e)
_____________________ test_sddmm[idtype0-sub-e-u-shp2-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'u', rhs_target = 'e', msg = 'sub'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb457b61c18>
op = 'add'
lhs_data = tensor([[[[1.0152, 1.4875, 1.6766],
          [1.6526, 1.4471, 1.9118],
          [1.9330, 1.3505, 1.1367]]],


      ...5],
          [1.6521, 1.2959, 1.8317],
          [1.0876, 1.2707, 1.8425]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[-1.9121, -1.4096, -1.2757]],

         [[-1.8097, -1.1082, -1.4018]],

         [[-1.3705, -1.8105, -1.4635...4402, -1.5613, -1.5677]],

         [[-1.9975, -1.9768, -1.5101]]]], dtype=torch.float64,
       grad_fn=<NegBackward>)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([300, 4, 1, 3])
SDDMM(message func: u_sub_e)
_____________________ test_sddmm[idtype0-sub-e-u-shp3-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'u', rhs_target = 'e', msg = 'sub'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72aaae2d0>
op = 'add'
lhs_data = tensor([[1.1255, 1.4699, 1.0952],
        [1.8186, 1.7236, 1.0363],
        [1.7792, 1.3878, 1.3477],
        [1.9196,... 1.0492],
        [1.1547, 1.7338, 1.1795],
        [1.5857, 1.8285, 1.2182]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.7229, -1.5014, -1.6560],
        [-1.3038, -1.6378, -1.7558],
        [-1.9026, -1.1985, -1.0516],
       ...  [-1.8703, -1.1101, -1.3628],
        [-1.3953, -1.4985, -1.6028]], dtype=torch.float64,
       grad_fn=<NegBackward>)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([100, 3])
SDDMM(message func: u_sub_e)
_____________________ test_sddmm[idtype0-sub-e-u-shp3-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'u', rhs_target = 'e', msg = 'sub'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72aaae558>
op = 'add'
lhs_data = tensor([[1.8126, 1.8137, 1.2258],
        [1.8314, 1.9695, 1.9665],
        [1.2486, 1.9177, 1.8290],
        [1.6046,... 1.5357],
        [1.2190, 1.1568, 1.9296],
        [1.5993, 1.6460, 1.4020]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.7793, -1.1808, -1.5945],
        [-1.8587, -1.2755, -1.6402],
        [-1.1299, -1.3786, -1.8830],
       ...  [-1.4689, -1.0061, -1.5741],
        [-1.6151, -1.7291, -1.9376]], dtype=torch.float64,
       grad_fn=<NegBackward>)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([300, 3])
SDDMM(message func: u_sub_e)
_____________________ test_sddmm[idtype0-sub-e-u-shp4-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'u', rhs_target = 'e', msg = 'sub'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43e86dc18>
op = 'add'
lhs_data = tensor([[1.5571],
        [1.8154],
        [1.8804],
        [1.7105],
        [1.3621],
        [1.2741],
        [1...935],
        [1.6973],
        [1.3097],
        [1.5630],
        [1.0679]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.2101],
        [-1.9144],
        [-1.2962],
        [-1.7500],
        [-1.4202],
        [-1.6153],
    ...       [-1.8881],
        [-1.5175],
        [-1.8139],
        [-1.8347]], dtype=torch.float64, grad_fn=<NegBackward>)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([100, 1])
SDDMM(message func: u_sub_e)
_____________________ test_sddmm[idtype0-sub-e-u-shp4-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'u', rhs_target = 'e', msg = 'sub'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb7294c0558>
op = 'add'
lhs_data = tensor([[1.8143],
        [1.5945],
        [1.4227],
        [1.0942],
        [1.1285],
        [1.2407],
        [1...207],
        [1.7838],
        [1.5970],
        [1.4553],
        [1.2508]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.4265],
        [-1.9583],
        [-1.7181],
        [-1.2552],
        [-1.1216],
        [-1.9776],
    ...       [-1.5056],
        [-1.3607],
        [-1.7031],
        [-1.2049]], dtype=torch.float64, grad_fn=<NegBackward>)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([300, 1])
SDDMM(message func: u_sub_e)
_____________________ test_sddmm[idtype0-sub-e-v-shp0-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': ...torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'v', rhs_target = 'e'
msg = 'sub', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72949c168>
op = 'add'
lhs_data = tensor([[[[[[1.5079],
            [1.4000],
            [1.9335]]],


          [[[1.9383],
            [1.5437],
    ...8]]],


          [[[1.9664],
            [1.8202],
            [1.4638]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[-1.2075]],

           [[-1.5678]],

           [[-1.9009]]]],



         [[[[-1.1652]],

           [[-...       [[[[-1.2293]],

           [[-1.7755]],

           [[-1.8813]]]]]], dtype=torch.float64, grad_fn=<NegBackward>)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([100, 4, 1, 3, 1, 1])
SDDMM(message func: v_sub_e)
_____________________ test_sddmm[idtype0-sub-e-v-shp0-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'v', rhs_target = 'e'
msg = 'sub', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb705aeefc0>
op = 'add'
lhs_data = tensor([[[[[[1.2962],
            [1.0353],
            [1.9149]]],


          [[[1.9003],
            [1.1320],
    ...3]]],


          [[[1.4035],
            [1.2797],
            [1.0303]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[-1.9336]],

           [[-1.7002]],

           [[-1.9657]]]],



         [[[[-1.8078]],

           [[-...       [[[[-1.0974]],

           [[-1.1726]],

           [[-1.0726]]]]]], dtype=torch.float64, grad_fn=<NegBackward>)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([40, 1, 2, 1, 3, 1]), rhs shape: torch.Size([300, 4, 1, 3, 1, 1])
SDDMM(message func: v_sub_e)
_____________________ test_sddmm[idtype0-sub-e-v-shp1-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Sch...pe=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'v', rhs_target = 'e'
msg = 'sub', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72aa9b240>
op = 'add'
lhs_data = tensor([[[[[1.1719, 1.8724, 1.8231,  ..., 1.9349, 1.0470, 1.4771]],

          [[1.2238, 1.9581, 1.0159,  ..., 1.3621,...
          [[1.1410, 1.5489, 1.1582,  ..., 1.0887, 1.9237, 1.2107]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[-1.7591, -1.8692, -1.9027,  ..., -1.2651, -1.5781, -1.7155],
           [-1.4262, -1.2508, -1.7612,  ..., ...   [-1.6187, -1.6674, -1.7329,  ..., -1.6512, -1.0597, -1.0828]]]]],
       dtype=torch.float64, grad_fn=<NegBackward>)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([100, 1, 3, 7, 7])
SDDMM(message func: v_sub_e)
_____________________ test_sddmm[idtype0-sub-e-v-shp1-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'v', rhs_target = 'e'
msg = 'sub', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72aa9b4c8>
op = 'add'
lhs_data = tensor([[[[[1.2609, 1.2447, 1.2461,  ..., 1.1007, 1.1796, 1.3989]],

          [[1.3717, 1.8930, 1.8988,  ..., 1.9137,...
          [[1.8410, 1.0951, 1.0935,  ..., 1.4584, 1.7870, 1.5326]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[-1.0897, -1.6390, -1.7608,  ..., -1.9627, -1.3771, -1.5201],
           [-1.4182, -1.1056, -1.3893,  ..., ...   [-1.1565, -1.1560, -1.8686,  ..., -1.5659, -1.1738, -1.2796]]]]],
       dtype=torch.float64, grad_fn=<NegBackward>)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([40, 5, 3, 1, 7]), rhs shape: torch.Size([300, 1, 3, 7, 7])
SDDMM(message func: v_sub_e)
_____________________ test_sddmm[idtype0-sub-e-v-shp2-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme...dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'v', rhs_target = 'e', msg = 'sub'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb705ae5120>
op = 'add'
lhs_data = tensor([[[[1.3774, 1.9364, 1.5854],
          [1.6931, 1.5305, 1.2546],
          [1.6619, 1.2000, 1.9785]]],


      ...5],
          [1.0922, 1.2489, 1.0642],
          [1.1694, 1.9373, 1.0951]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[-1.9911, -1.6350, -1.9326]],

         [[-1.9933, -1.7398, -1.9372]],

         [[-1.9157, -1.3498, -1.9232...0380, -1.6191, -1.2145]],

         [[-1.1348, -1.0058, -1.1735]]]], dtype=torch.float64,
       grad_fn=<NegBackward>)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([100, 4, 1, 3])
SDDMM(message func: v_sub_e)
_____________________ test_sddmm[idtype0-sub-e-v-shp2-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'v', rhs_target = 'e', msg = 'sub'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72950b168>
op = 'add'
lhs_data = tensor([[[[1.6164, 1.0290, 1.6125],
          [1.0701, 1.7821, 1.1912],
          [1.5355, 1.0683, 1.8794]]],


      ...6],
          [1.3866, 1.8491, 1.6895],
          [1.5823, 1.1928, 1.5527]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[-1.1849, -1.4106, -1.0212]],

         [[-1.0388, -1.5373, -1.9555]],

         [[-1.6817, -1.9028, -1.3804...2894, -1.7692, -1.5882]],

         [[-1.5208, -1.0351, -1.6937]]]], dtype=torch.float64,
       grad_fn=<NegBackward>)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([40, 1, 3, 3]), rhs shape: torch.Size([300, 4, 1, 3])
SDDMM(message func: v_sub_e)
_____________________ test_sddmm[idtype0-sub-e-v-shp3-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'v', rhs_target = 'e', msg = 'sub'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb705bec438>
op = 'add'
lhs_data = tensor([[1.0990, 1.6970, 1.1433],
        [1.9439, 1.7318, 1.0281],
        [1.3959, 1.7948, 1.2402],
        [1.2035,... 1.9103],
        [1.5973, 1.7107, 1.9550],
        [1.9265, 1.2730, 1.7552]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.5213, -1.2444, -1.3199],
        [-1.7123, -1.9139, -1.6038],
        [-1.7356, -1.2189, -1.3561],
       ...  [-1.4453, -1.2536, -1.7217],
        [-1.2138, -1.4201, -1.5267]], dtype=torch.float64,
       grad_fn=<NegBackward>)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([100, 3])
SDDMM(message func: v_sub_e)
_____________________ test_sddmm[idtype0-sub-e-v-shp3-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'v', rhs_target = 'e', msg = 'sub'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43e85b240>
op = 'add'
lhs_data = tensor([[1.9780, 1.0161, 1.3339],
        [1.3811, 1.4450, 1.2837],
        [1.7070, 1.6992, 1.2057],
        [1.2372,... 1.7804],
        [1.4705, 1.8394, 1.8979],
        [1.8011, 1.7762, 1.7827]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.2205, -1.1402, -1.7505],
        [-1.8027, -1.0023, -1.7252],
        [-1.6599, -1.5741, -1.3290],
       ...  [-1.7672, -1.7581, -1.1136],
        [-1.5279, -1.2056, -1.8245]], dtype=torch.float64,
       grad_fn=<NegBackward>)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([40, 3]), rhs shape: torch.Size([300, 3])
SDDMM(message func: v_sub_e)
_____________________ test_sddmm[idtype0-sub-e-v-shp4-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'v', rhs_target = 'e', msg = 'sub'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72aa9e1b0>
op = 'add'
lhs_data = tensor([[1.4584],
        [1.8718],
        [1.9374],
        [1.8633],
        [1.1555],
        [1.1532],
        [1...776],
        [1.9149],
        [1.0426],
        [1.1254],
        [1.6656]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.6242],
        [-1.1407],
        [-1.2399],
        [-1.8214],
        [-1.2202],
        [-1.7181],
    ...       [-1.6561],
        [-1.4180],
        [-1.6066],
        [-1.2577]], dtype=torch.float64, grad_fn=<NegBackward>)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([100, 1])
SDDMM(message func: v_sub_e)
_____________________ test_sddmm[idtype0-sub-e-v-shp4-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'v', rhs_target = 'e', msg = 'sub'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72aa9e438>
op = 'add'
lhs_data = tensor([[1.0549],
        [1.7039],
        [1.5695],
        [1.5167],
        [1.1730],
        [1.7569],
        [1...028],
        [1.1599],
        [1.6418],
        [1.3746],
        [1.6912]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.7099],
        [-1.9367],
        [-1.2382],
        [-1.7006],
        [-1.6878],
        [-1.4353],
    ...       [-1.7910],
        [-1.6263],
        [-1.4964],
        [-1.2293]], dtype=torch.float64, grad_fn=<NegBackward>)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([40, 1]), rhs shape: torch.Size([300, 1])
SDDMM(message func: v_sub_e)
_____________________ test_sddmm[idtype0-mul-u-v-shp0-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': ...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'v', rhs_target = 'u'
msg = 'mul', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72951fc18>
op = 'mul'
lhs_data = tensor([[[[[[1.2134],
            [1.6927],
            [1.1663]]],


          [[[1.0484],
            [1.5480],
    ...5]]],


          [[[1.7529],
            [1.9093],
            [1.9264]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.0028]],

           [[1.3396]],

           [[1.3365]]]],



         [[[[1.7038]],

           [[1.929...



         [[[[1.0211]],

           [[1.6583]],

           [[1.2251]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: v_mul_u)
_____________________ test_sddmm[idtype0-mul-u-v-shp0-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'v', rhs_target = 'u'
msg = 'mul', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72aa85cf0>
op = 'mul'
lhs_data = tensor([[[[[[1.8124],
            [1.4372],
            [1.5706]]],


          [[[1.0559],
            [1.0412],
    ...7]]],


          [[[1.8111],
            [1.1538],
            [1.3779]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.5152]],

           [[1.5933]],

           [[1.2429]]]],



         [[[[1.1496]],

           [[1.499...



         [[[[1.1042]],

           [[1.1431]],

           [[1.7290]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([40, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: v_mul_u)
_____________________ test_sddmm[idtype0-mul-u-v-shp1-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Sch...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'v', rhs_target = 'u'
msg = 'mul', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72be4c4c8>
op = 'mul'
lhs_data = tensor([[[[[1.2091, 1.0849, 1.1892,  ..., 1.1624, 1.1788, 1.0308]],

          [[1.7473, 1.4397, 1.4412,  ..., 1.1906,...
          [[1.6193, 1.9102, 1.4762,  ..., 1.4386, 1.4529, 1.3085]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.4135, 1.3691, 1.6967,  ..., 1.5760, 1.4393, 1.5354],
           [1.5953, 1.5285, 1.0346,  ..., 1.4106, 1...
           [1.3440, 1.9955, 1.6573,  ..., 1.0119, 1.1849, 1.6692]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: v_mul_u)
_____________________ test_sddmm[idtype0-mul-u-v-shp1-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'v', rhs_target = 'u'
msg = 'mul', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72aaa01b0>
op = 'mul'
lhs_data = tensor([[[[[1.1761, 1.9842, 1.3760,  ..., 1.6216, 1.1509, 1.1530]],

          [[1.0518, 1.8946, 1.2883,  ..., 1.8590,...
          [[1.8772, 1.5533, 1.3045,  ..., 1.6520, 1.4407, 1.5422]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.6305, 1.1622, 1.6379,  ..., 1.2834, 1.0041, 1.5206],
           [1.0747, 1.7404, 1.5065,  ..., 1.0867, 1...
           [1.2680, 1.0189, 1.6629,  ..., 1.9595, 1.4079, 1.5980]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([40, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: v_mul_u)
_____________________ test_sddmm[idtype0-mul-u-v-shp2-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'v', rhs_target = 'u', msg = 'mul'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72aaa04c8>
op = 'mul'
lhs_data = tensor([[[[1.6189, 1.0619, 1.7755],
          [1.4444, 1.4630, 1.5520],
          [1.5291, 1.8417, 1.0868]]],


      ...3],
          [1.7410, 1.3870, 1.4729],
          [1.7755, 1.7466, 1.6346]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.6165, 1.1898, 1.1937]],

         [[1.9212, 1.8140, 1.4219]],

         [[1.7562, 1.2732, 1.0687]],

    ...

         [[1.0762, 1.0375, 1.1871]],

         [[1.0939, 1.0398, 1.4344]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: v_mul_u)
_____________________ test_sddmm[idtype0-mul-u-v-shp2-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'v', rhs_target = 'u', msg = 'mul'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72aaa0708>
op = 'mul'
lhs_data = tensor([[[[1.1531, 1.1349, 1.4260],
          [1.7951, 1.6538, 1.8236],
          [1.7919, 1.0299, 1.3691]]],


      ...9],
          [1.2319, 1.8481, 1.5585],
          [1.6313, 1.5973, 1.8582]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.5877, 1.9732, 1.6806]],

         [[1.9310, 1.0194, 1.5979]],

         [[1.2982, 1.8118, 1.9418]],

    ...

         [[1.3513, 1.4033, 1.6056]],

         [[1.0812, 1.2319, 1.1395]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([40, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: v_mul_u)
_____________________ test_sddmm[idtype0-mul-u-v-shp3-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'v', rhs_target = 'u', msg = 'mul'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72b398b88>
op = 'mul'
lhs_data = tensor([[1.8215, 1.1495, 1.8998],
        [1.8968, 1.4857, 1.2333],
        [1.5748, 1.5866, 1.5953],
        [1.4353,... 1.6227],
        [1.1555, 1.8048, 1.1937],
        [1.0075, 1.7446, 1.4644]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.9929, 1.0011, 1.3960],
        [1.6141, 1.7884, 1.0476],
        [1.2252, 1.3016, 1.8981],
        [1.7989,... 1.9804],
        [1.6108, 1.2625, 1.3904],
        [1.6495, 1.5807, 1.9888]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: v_mul_u)
_____________________ test_sddmm[idtype0-mul-u-v-shp3-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'v', rhs_target = 'u', msg = 'mul'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb729530360>
op = 'mul'
lhs_data = tensor([[1.3418, 1.0421, 1.8527],
        [1.9822, 1.8347, 1.6376],
        [1.8090, 1.2274, 1.0674],
        [1.3438,... 1.1976],
        [1.3345, 1.7242, 1.6068],
        [1.0552, 1.9081, 1.2790]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.0685, 1.5585, 1.9651],
        [1.0611, 1.2028, 1.2411],
        [1.3928, 1.4456, 1.2212],
        [1.9867,... 1.1360],
        [1.9507, 1.2104, 1.5330],
        [1.2131, 1.1469, 1.0591]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([40, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: v_mul_u)
_____________________ test_sddmm[idtype0-mul-u-v-shp4-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'v', rhs_target = 'u', msg = 'mul'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb45514eb40>
op = 'mul'
lhs_data = tensor([[1.9901],
        [1.0099],
        [1.2603],
        [1.4604],
        [1.9943],
        [1.1959],
        [1...313],
        [1.0507],
        [1.9165],
        [1.3667],
        [1.2114]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.7121],
        [1.7124],
        [1.5202],
        [1.3195],
        [1.8775],
        [1.3350],
        [1...968],
        [1.6054],
        [1.2468],
        [1.4611],
        [1.5100]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: v_mul_u)
_____________________ test_sddmm[idtype0-mul-u-v-shp4-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'v', rhs_target = 'u', msg = 'mul'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72aabc168>
op = 'mul'
lhs_data = tensor([[1.8475],
        [1.7706],
        [1.0445],
        [1.7648],
        [1.7366],
        [1.7456],
        [1...912],
        [1.6841],
        [1.3253],
        [1.3462],
        [1.4826]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.9941],
        [1.1150],
        [1.8879],
        [1.9967],
        [1.7072],
        [1.2826],
        [1...858],
        [1.2684],
        [1.6162],
        [1.7221],
        [1.2993]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([40, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: v_mul_u)
_____________________ test_sddmm[idtype0-mul-u-e-shp0-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...torch.float64), 'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'e', rhs_target = 'u'
msg = 'mul', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb4551662d0>
op = 'mul'
lhs_data = tensor([[[[[[1.5797],
            [1.9666],
            [1.2427]]],


          [[[1.5013],
            [1.3110],
    ...0]]],


          [[[1.3627],
            [1.1619],
            [1.5314]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.8395]],

           [[1.4502]],

           [[1.1652]]]],



         [[[[1.8136]],

           [[1.450...



         [[[[1.2402]],

           [[1.8023]],

           [[1.1287]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([100, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: e_mul_u)
_____________________ test_sddmm[idtype0-mul-u-e-shp0-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'e', rhs_target = 'u'
msg = 'mul', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb705ae4120>
op = 'mul'
lhs_data = tensor([[[[[[1.5343],
            [1.5562],
            [1.2622]]],


          [[[1.9095],
            [1.3036],
    ...0]]],


          [[[1.5853],
            [1.1440],
            [1.5747]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.4318]],

           [[1.6737]],

           [[1.6752]]]],



         [[[[1.3846]],

           [[1.864...



         [[[[1.8436]],

           [[1.5494]],

           [[1.7254]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([300, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: e_mul_u)
_____________________ test_sddmm[idtype0-mul-u-e-shp1-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...pe=torch.float64), 'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'e', rhs_target = 'u'
msg = 'mul', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb61cf9a630>
op = 'mul'
lhs_data = tensor([[[[[1.1007, 1.7470, 1.3938,  ..., 1.6578, 1.2180, 1.1623]],

          [[1.8032, 1.5490, 1.1946,  ..., 1.5221,...
          [[1.0997, 1.2849, 1.1970,  ..., 1.9571, 1.6170, 1.6252]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.4485, 1.3094, 1.7124,  ..., 1.9970, 1.4404, 1.2315],
           [1.5086, 1.7793, 1.3079,  ..., 1.1855, 1...
           [1.6540, 1.6115, 1.0852,  ..., 1.5910, 1.6503, 1.5416]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([100, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: e_mul_u)
_____________________ test_sddmm[idtype0-mul-u-e-shp1-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'e', rhs_target = 'u'
msg = 'mul', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acebca8>
op = 'mul'
lhs_data = tensor([[[[[1.5777, 1.3026, 1.6051,  ..., 1.2941, 1.0048, 1.3647]],

          [[1.6322, 1.3447, 1.2878,  ..., 1.1573,...
          [[1.4425, 1.0080, 1.2549,  ..., 1.7957, 1.8067, 1.8318]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.8213, 1.1308, 1.4317,  ..., 1.7527, 1.2441, 1.7350],
           [1.6399, 1.9137, 1.5965,  ..., 1.3235, 1...
           [1.5097, 1.5476, 1.8155,  ..., 1.8486, 1.9823, 1.3170]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([300, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: e_mul_u)
_____________________ test_sddmm[idtype0-mul-u-e-shp2-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...dtype=torch.float64), 'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'e', rhs_target = 'u', msg = 'mul'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72c1d92d0>
op = 'mul'
lhs_data = tensor([[[[1.7700, 1.7520, 1.4531],
          [1.2431, 1.3474, 1.0914],
          [1.7379, 1.2985, 1.2161]]],


      ...5],
          [1.9506, 1.9100, 1.8332],
          [1.9846, 1.6485, 1.2340]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.1568, 1.4477, 1.1027]],

         [[1.8659, 1.3733, 1.7064]],

         [[1.8554, 1.6635, 1.8083]],

    ...

         [[1.6604, 1.8937, 1.0280]],

         [[1.1437, 1.6765, 1.9576]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([100, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: e_mul_u)
_____________________ test_sddmm[idtype0-mul-u-e-shp2-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'e', rhs_target = 'u', msg = 'mul'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72aab20d8>
op = 'mul'
lhs_data = tensor([[[[1.8741, 1.6802, 1.1296],
          [1.6912, 1.2217, 1.0754],
          [1.7648, 1.5982, 1.9413]]],


      ...0],
          [1.5531, 1.2532, 1.8947],
          [1.1127, 1.1183, 1.4011]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.3271, 1.2632, 1.7216]],

         [[1.9594, 1.2752, 1.5259]],

         [[1.7936, 1.0132, 1.8602]],

    ...

         [[1.5070, 1.3562, 1.3276]],

         [[1.7682, 1.5432, 1.6112]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([300, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: e_mul_u)
_____________________ test_sddmm[idtype0-mul-u-e-shp3-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'e', rhs_target = 'u', msg = 'mul'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acf27e0>
op = 'mul'
lhs_data = tensor([[1.4779, 1.7472, 1.0544],
        [1.2223, 1.6876, 1.2872],
        [1.5915, 1.7222, 1.1078],
        [1.7720,... 1.8788],
        [1.4587, 1.5107, 1.2179],
        [1.3613, 1.2706, 1.2179]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.5259, 1.8050, 1.3955],
        [1.7819, 1.9196, 1.1062],
        [1.8441, 1.9457, 1.1391],
        [1.4194,... 1.2173],
        [1.7636, 1.0176, 1.0733],
        [1.1939, 1.0752, 1.5212]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([100, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: e_mul_u)
_____________________ test_sddmm[idtype0-mul-u-e-shp3-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'e', rhs_target = 'u', msg = 'mul'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb61cfa4288>
op = 'mul'
lhs_data = tensor([[1.8270, 1.9664, 1.5982],
        [1.2623, 1.9285, 1.8091],
        [1.2986, 1.1988, 1.1452],
        [1.7210,... 1.3768],
        [1.1949, 1.3745, 1.7839],
        [1.6226, 1.1053, 1.4167]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.1025, 1.2757, 1.4300],
        [1.9608, 1.6722, 1.2336],
        [1.2058, 1.2818, 1.4334],
        [1.3391,... 1.7659],
        [1.7472, 1.2533, 1.4170],
        [1.4822, 1.4429, 1.5920]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([300, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: e_mul_u)
_____________________ test_sddmm[idtype0-mul-u-e-shp4-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'e', rhs_target = 'u', msg = 'mul'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb457b5f048>
op = 'mul'
lhs_data = tensor([[1.5916],
        [1.4189],
        [1.2001],
        [1.1418],
        [1.0474],
        [1.6239],
        [1...816],
        [1.7355],
        [1.7824],
        [1.2911],
        [1.5173]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.2780],
        [1.3702],
        [1.5535],
        [1.3454],
        [1.9052],
        [1.7023],
        [1...332],
        [1.1624],
        [1.3874],
        [1.0450],
        [1.8962]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([100, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: e_mul_u)
_____________________ test_sddmm[idtype0-mul-u-e-shp4-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'e', rhs_target = 'u', msg = 'mul'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43e6d23f0>
op = 'mul'
lhs_data = tensor([[1.4979],
        [1.5909],
        [1.0408],
        [1.2086],
        [1.4024],
        [1.4681],
        [1...629],
        [1.8595],
        [1.8932],
        [1.3436],
        [1.6003]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.6362],
        [1.4365],
        [1.8130],
        [1.5843],
        [1.2571],
        [1.3285],
        [1...493],
        [1.9271],
        [1.2590],
        [1.8946],
        [1.9490]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([300, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: e_mul_u)
_____________________ test_sddmm[idtype0-mul-v-u-shp0-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': ...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'u', rhs_target = 'v'
msg = 'mul', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43e6e6870>
op = 'mul'
lhs_data = tensor([[[[[[1.1392],
            [1.6155],
            [1.2149]]],


          [[[1.1130],
            [1.1667],
    ...3]]],


          [[[1.1554],
            [1.8648],
            [1.8932]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.0735]],

           [[1.2980]],

           [[1.8560]]]],



         [[[[1.8961]],

           [[1.736...



         [[[[1.1301]],

           [[1.6002]],

           [[1.1161]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: u_mul_v)
_____________________ test_sddmm[idtype0-mul-v-u-shp0-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'u', rhs_target = 'v'
msg = 'mul', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb61f15fa68>
op = 'mul'
lhs_data = tensor([[[[[[1.6066],
            [1.6647],
            [1.4467]]],


          [[[1.4352],
            [1.8905],
    ...8]]],


          [[[1.3805],
            [1.6044],
            [1.6892]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.8849]],

           [[1.7262]],

           [[1.6378]]]],



         [[[[1.2565]],

           [[1.955...



         [[[[1.2007]],

           [[1.2880]],

           [[1.9900]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([40, 4, 1, 3, 1, 1])
SDDMM(message func: u_mul_v)
_____________________ test_sddmm[idtype0-mul-v-u-shp1-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Sch...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'u', rhs_target = 'v'
msg = 'mul', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72aab42d0>
op = 'mul'
lhs_data = tensor([[[[[1.7254, 1.6355, 1.0465,  ..., 1.1372, 1.7414, 1.4065]],

          [[1.7178, 1.3973, 1.4853,  ..., 1.0294,...
          [[1.7971, 1.4275, 1.0468,  ..., 1.6589, 1.8429, 1.5153]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.2343, 1.8983, 1.0370,  ..., 1.1321, 1.5935, 1.8846],
           [1.7339, 1.0833, 1.7517,  ..., 1.8339, 1...
           [1.0796, 1.3766, 1.0487,  ..., 1.2202, 1.6661, 1.8691]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: u_mul_v)
_____________________ test_sddmm[idtype0-mul-v-u-shp1-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'u', rhs_target = 'v'
msg = 'mul', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72aab4558>
op = 'mul'
lhs_data = tensor([[[[[1.5261, 1.3369, 1.1080,  ..., 1.7421, 1.6405, 1.1153]],

          [[1.4193, 1.8942, 1.9111,  ..., 1.9738,...
          [[1.1715, 1.3667, 1.3108,  ..., 1.3649, 1.5390, 1.7172]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.1519, 1.9878, 1.9865,  ..., 1.4047, 1.6576, 1.1886],
           [1.6331, 1.6331, 1.0236,  ..., 1.1522, 1...
           [1.6836, 1.9985, 1.5890,  ..., 1.8129, 1.9061, 1.0408]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([40, 1, 3, 7, 7])
SDDMM(message func: u_mul_v)
_____________________ test_sddmm[idtype0-mul-v-u-shp2-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'u', rhs_target = 'v', msg = 'mul'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb457b76bd0>
op = 'mul'
lhs_data = tensor([[[[1.5903, 1.2385, 1.5796],
          [1.2301, 1.0793, 1.2229],
          [1.2572, 1.0428, 1.9580]]],


      ...0],
          [1.2781, 1.6750, 1.9855],
          [1.6559, 1.5300, 1.9880]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.4013, 1.6375, 1.8113]],

         [[1.9821, 1.3685, 1.8000]],

         [[1.9631, 1.2024, 1.0593]],

    ...

         [[1.8099, 1.8005, 1.2011]],

         [[1.6635, 1.0638, 1.2064]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: u_mul_v)
_____________________ test_sddmm[idtype0-mul-v-u-shp2-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'u', rhs_target = 'v', msg = 'mul'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb61f15aea0>
op = 'mul'
lhs_data = tensor([[[[1.6890, 1.2641, 1.2064],
          [1.2670, 1.8662, 1.7952],
          [1.7735, 1.8103, 1.5439]]],


      ...2],
          [1.2924, 1.1844, 1.6015],
          [1.6630, 1.3261, 1.6926]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.4580, 1.8963, 1.5111]],

         [[1.7040, 1.4660, 1.9452]],

         [[1.6927, 1.0419, 1.5889]],

    ...

         [[1.9356, 1.5489, 1.9919]],

         [[1.8441, 1.6354, 1.4700]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([40, 4, 1, 3])
SDDMM(message func: u_mul_v)
_____________________ test_sddmm[idtype0-mul-v-u-shp3-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'u', rhs_target = 'v', msg = 'mul'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb61cfb5120>
op = 'mul'
lhs_data = tensor([[1.7364, 1.2589, 1.9642],
        [1.8319, 1.9675, 1.2334],
        [1.6157, 1.4322, 1.7812],
        [1.7950,... 1.1609],
        [1.2212, 1.2352, 1.8276],
        [1.7630, 1.7445, 1.7538]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.7712, 1.4441, 1.8348],
        [1.3232, 1.7858, 1.2215],
        [1.9160, 1.9674, 1.9596],
        [1.5363,... 1.9783],
        [1.3758, 1.9904, 1.8596],
        [1.3909, 1.5570, 1.3388]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: u_mul_v)
_____________________ test_sddmm[idtype0-mul-v-u-shp3-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'u', rhs_target = 'v', msg = 'mul'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72949cf30>
op = 'mul'
lhs_data = tensor([[1.6979, 1.4592, 1.5575],
        [1.6044, 1.4044, 1.0300],
        [1.5838, 1.9546, 1.1346],
        [1.0080,... 1.6506],
        [1.5091, 1.7109, 1.5651],
        [1.7867, 1.3418, 1.1336]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.3482, 1.2118, 1.2790],
        [1.2772, 1.1153, 1.0483],
        [1.8384, 1.9193, 1.8559],
        [1.2240,... 1.5710],
        [1.6515, 1.3912, 1.1337],
        [1.0881, 1.2699, 1.2054]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([40, 3])
SDDMM(message func: u_mul_v)
_____________________ test_sddmm[idtype0-mul-v-u-shp4-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'u', rhs_target = 'v', msg = 'mul'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72aaba240>
op = 'mul'
lhs_data = tensor([[1.9980],
        [1.4273],
        [1.3783],
        [1.3429],
        [1.9074],
        [1.2742],
        [1...482],
        [1.1656],
        [1.3613],
        [1.3674],
        [1.3873]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.9406],
        [1.3865],
        [1.6577],
        [1.2894],
        [1.9294],
        [1.2442],
        [1...214],
        [1.8093],
        [1.7464],
        [1.4702],
        [1.3710]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: u_mul_v)
_____________________ test_sddmm[idtype0-mul-v-u-shp4-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'u', rhs_target = 'v', msg = 'mul'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72aaba4c8>
op = 'mul'
lhs_data = tensor([[1.4056],
        [1.6843],
        [1.9831],
        [1.0975],
        [1.9192],
        [1.6647],
        [1...045],
        [1.0570],
        [1.6374],
        [1.6481],
        [1.0612]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.8677],
        [1.0517],
        [1.4482],
        [1.2381],
        [1.2830],
        [1.0219],
        [1...837],
        [1.6794],
        [1.2546],
        [1.6241],
        [1.1607]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([40, 1])
SDDMM(message func: u_mul_v)
_____________________ test_sddmm[idtype0-mul-v-e-shp0-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...torch.float64), 'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'e', rhs_target = 'v'
msg = 'mul', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72bd4ed80>
op = 'mul'
lhs_data = tensor([[[[[[1.3748],
            [1.3817],
            [1.4792]]],


          [[[1.0104],
            [1.6704],
    ...2]]],


          [[[1.9434],
            [1.1472],
            [1.8164]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.8079]],

           [[1.9936]],

           [[1.9247]]]],



         [[[[1.7477]],

           [[1.009...



         [[[[1.4565]],

           [[1.9245]],

           [[1.2575]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([100, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: e_mul_v)
_____________________ test_sddmm[idtype0-mul-v-e-shp0-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'e', rhs_target = 'v'
msg = 'mul', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72c0593a8>
op = 'mul'
lhs_data = tensor([[[[[[1.6458],
            [1.0392],
            [1.0429]]],


          [[[1.1165],
            [1.5710],
    ...0]]],


          [[[1.3844],
            [1.8812],
            [1.1960]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.8562]],

           [[1.2964]],

           [[1.0601]]]],



         [[[[1.9892]],

           [[1.258...



         [[[[1.9132]],

           [[1.0616]],

           [[1.8392]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([300, 1, 2, 1, 3, 1]), rhs shape: torch.Size([40, 4, 1, 3, 1, 1])
SDDMM(message func: e_mul_v)
_____________________ test_sddmm[idtype0-mul-v-e-shp1-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...pe=torch.float64), 'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'e', rhs_target = 'v'
msg = 'mul', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb7294a3870>
op = 'mul'
lhs_data = tensor([[[[[1.3702, 1.9488, 1.3191,  ..., 1.1618, 1.6973, 1.0026]],

          [[1.0481, 1.8692, 1.4743,  ..., 1.4111,...
          [[1.2945, 1.5375, 1.8787,  ..., 1.9438, 1.4265, 1.7877]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.3718, 1.1680, 1.4834,  ..., 1.5842, 1.7346, 1.1188],
           [1.8824, 1.9586, 1.5324,  ..., 1.0645, 1...
           [1.0798, 1.8629, 1.2088,  ..., 1.6136, 1.8836, 1.0767]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([100, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: e_mul_v)
_____________________ test_sddmm[idtype0-mul-v-e-shp1-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'e', rhs_target = 'v'
msg = 'mul', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb70584d7e0>
op = 'mul'
lhs_data = tensor([[[[[1.3218, 1.4833, 1.5011,  ..., 1.6949, 1.7228, 1.8635]],

          [[1.5348, 1.1614, 1.3238,  ..., 1.5270,...
          [[1.7260, 1.4691, 1.1485,  ..., 1.4666, 1.9524, 1.5599]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.1004, 1.0616, 1.8746,  ..., 1.9884, 1.4734, 1.1603],
           [1.9122, 1.1881, 1.5319,  ..., 1.1466, 1...
           [1.3074, 1.7341, 1.1171,  ..., 1.0870, 1.8793, 1.5882]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([300, 5, 3, 1, 7]), rhs shape: torch.Size([40, 1, 3, 7, 7])
SDDMM(message func: e_mul_v)
_____________________ test_sddmm[idtype0-mul-v-e-shp2-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...dtype=torch.float64), 'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'e', rhs_target = 'v', msg = 'mul'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43e85b5a0>
op = 'mul'
lhs_data = tensor([[[[1.1938, 1.6845, 1.1217],
          [1.0258, 1.0165, 1.1324],
          [1.0129, 1.8475, 1.2167]]],


      ...0],
          [1.3707, 1.1277, 1.1724],
          [1.0434, 1.1374, 1.5299]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.7053, 1.5624, 1.2460]],

         [[1.2740, 1.5178, 1.4410]],

         [[1.7574, 1.7797, 1.7229]],

    ...

         [[1.9836, 1.0942, 1.2160]],

         [[1.3473, 1.5360, 1.8952]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([100, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: e_mul_v)
_____________________ test_sddmm[idtype0-mul-v-e-shp2-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'e', rhs_target = 'v', msg = 'mul'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb457b70ea0>
op = 'mul'
lhs_data = tensor([[[[1.3425, 1.8598, 1.3484],
          [1.9765, 1.3926, 1.4796],
          [1.8511, 1.1295, 1.1848]]],


      ...5],
          [1.7091, 1.5219, 1.1467],
          [1.1310, 1.7235, 1.5321]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.7308, 1.8088, 1.8945]],

         [[1.8694, 1.1523, 1.4650]],

         [[1.1219, 1.5921, 1.5905]],

    ...

         [[1.5459, 1.5700, 1.4220]],

         [[1.7689, 1.5129, 1.1816]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([300, 1, 3, 3]), rhs shape: torch.Size([40, 4, 1, 3])
SDDMM(message func: e_mul_v)
_____________________ test_sddmm[idtype0-mul-v-e-shp3-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'e', rhs_target = 'v', msg = 'mul'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72c052048>
op = 'mul'
lhs_data = tensor([[1.5467, 1.5286, 1.7264],
        [1.1726, 1.0548, 1.3566],
        [1.2517, 1.5119, 1.5066],
        [1.7340,... 1.7654],
        [1.1969, 1.1769, 1.5616],
        [1.5520, 1.5859, 1.8023]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.6189, 1.9072, 1.9266],
        [1.1187, 1.5503, 1.0251],
        [1.7442, 1.4871, 1.6692],
        [1.6484,... 1.8881],
        [1.2789, 1.2769, 1.0588],
        [1.3638, 1.7591, 1.2655]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([100, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: e_mul_v)
_____________________ test_sddmm[idtype0-mul-v-e-shp3-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'e', rhs_target = 'v', msg = 'mul'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72c052288>
op = 'mul'
lhs_data = tensor([[1.5829, 1.5634, 1.5521],
        [1.5847, 1.6746, 1.6888],
        [1.6371, 1.5277, 1.0083],
        [1.4660,... 1.6998],
        [1.6988, 1.8791, 1.8534],
        [1.6827, 1.1968, 1.0039]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.3961, 1.4057, 1.3251],
        [1.7312, 1.5299, 1.3832],
        [1.1578, 1.7748, 1.4104],
        [1.5558,... 1.7955],
        [1.3854, 1.1648, 1.6940],
        [1.5721, 1.1498, 1.0138]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([300, 3]), rhs shape: torch.Size([40, 3])
SDDMM(message func: e_mul_v)
_____________________ test_sddmm[idtype0-mul-v-e-shp4-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'e', rhs_target = 'v', msg = 'mul'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb7294a3d80>
op = 'mul'
lhs_data = tensor([[1.9298],
        [1.3953],
        [1.7581],
        [1.9799],
        [1.9749],
        [1.6717],
        [1...590],
        [1.1925],
        [1.4553],
        [1.8178],
        [1.6407]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.7276],
        [1.0655],
        [1.4373],
        [1.6895],
        [1.3256],
        [1.1902],
        [1...248],
        [1.1730],
        [1.0325],
        [1.4388],
        [1.9845]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([100, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: e_mul_v)
_____________________ test_sddmm[idtype0-mul-v-e-shp4-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'e', rhs_target = 'v', msg = 'mul'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43e85b3a8>
op = 'mul'
lhs_data = tensor([[1.5520],
        [1.7171],
        [1.7093],
        [1.1534],
        [1.4518],
        [1.0282],
        [1...948],
        [1.4260],
        [1.9233],
        [1.5457],
        [1.2305]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.4120],
        [1.8860],
        [1.9619],
        [1.0533],
        [1.2793],
        [1.5232],
        [1...721],
        [1.5878],
        [1.9918],
        [1.0132],
        [1.9564]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([300, 1]), rhs shape: torch.Size([40, 1])
SDDMM(message func: e_mul_v)
_____________________ test_sddmm[idtype0-mul-e-u-shp0-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': ...torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'u', rhs_target = 'e'
msg = 'mul', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb7295345a0>
op = 'mul'
lhs_data = tensor([[[[[[1.4664],
            [1.6914],
            [1.5437]]],


          [[[1.7265],
            [1.1967],
    ...5]]],


          [[[1.8934],
            [1.3102],
            [1.8538]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.4904]],

           [[1.8725]],

           [[1.5990]]]],



         [[[[1.1443]],

           [[1.811...



         [[[[1.7009]],

           [[1.1514]],

           [[1.9840]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([100, 4, 1, 3, 1, 1])
SDDMM(message func: u_mul_e)
_____________________ test_sddmm[idtype0-mul-e-u-shp0-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'u', rhs_target = 'e'
msg = 'mul', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb70583dee8>
op = 'mul'
lhs_data = tensor([[[[[[1.1144],
            [1.0343],
            [1.2415]]],


          [[[1.5507],
            [1.2021],
    ...7]]],


          [[[1.6033],
            [1.6170],
            [1.4110]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.8241]],

           [[1.6830]],

           [[1.9180]]]],



         [[[[1.5085]],

           [[1.597...



         [[[[1.1959]],

           [[1.1387]],

           [[1.4174]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([300, 4, 1, 3, 1, 1])
SDDMM(message func: u_mul_e)
_____________________ test_sddmm[idtype0-mul-e-u-shp1-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Sch...pe=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'u', rhs_target = 'e'
msg = 'mul', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb706233fc0>
op = 'mul'
lhs_data = tensor([[[[[1.2755, 1.4904, 1.4674,  ..., 1.0711, 1.2805, 1.3667]],

          [[1.2688, 1.3233, 1.9057,  ..., 1.3667,...
          [[1.9994, 1.4350, 1.6658,  ..., 1.9135, 1.8815, 1.7668]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.5162, 1.0792, 1.5115,  ..., 1.6078, 1.6591, 1.7418],
           [1.0835, 1.8042, 1.5951,  ..., 1.6102, 1...
           [1.0492, 1.9652, 1.5004,  ..., 1.2636, 1.9799, 1.8994]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([100, 1, 3, 7, 7])
SDDMM(message func: u_mul_e)
_____________________ test_sddmm[idtype0-mul-e-u-shp1-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'u', rhs_target = 'e'
msg = 'mul', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72c05e288>
op = 'mul'
lhs_data = tensor([[[[[1.9607, 1.6323, 1.5792,  ..., 1.8773, 1.9909, 1.0221]],

          [[1.9169, 1.5060, 1.2363,  ..., 1.0604,...
          [[1.3287, 1.0196, 1.1083,  ..., 1.5206, 1.9129, 1.9595]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.9614, 1.9484, 1.5308,  ..., 1.0263, 1.6798, 1.1752],
           [1.3265, 1.7974, 1.2480,  ..., 1.2264, 1...
           [1.8018, 1.3024, 1.2725,  ..., 1.3863, 1.5773, 1.6542]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([300, 1, 3, 7, 7])
SDDMM(message func: u_mul_e)
_____________________ test_sddmm[idtype0-mul-e-u-shp2-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme...dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'u', rhs_target = 'e', msg = 'mul'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb705c04480>
op = 'mul'
lhs_data = tensor([[[[1.8002, 1.2539, 1.9046],
          [1.3710, 1.6473, 1.1905],
          [1.3224, 1.1616, 1.8274]]],


      ...2],
          [1.3756, 1.0690, 1.4796],
          [1.6007, 1.0588, 1.6470]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.5640, 1.1620, 1.5567]],

         [[1.4942, 1.3134, 1.6760]],

         [[1.7928, 1.1505, 1.8410]],

    ...

         [[1.3293, 1.5288, 1.8205]],

         [[1.4055, 1.2419, 1.0280]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([100, 4, 1, 3])
SDDMM(message func: u_mul_e)
_____________________ test_sddmm[idtype0-mul-e-u-shp2-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'u', rhs_target = 'e', msg = 'mul'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43e975cf0>
op = 'mul'
lhs_data = tensor([[[[1.3004, 1.1327, 1.1910],
          [1.0949, 1.5986, 1.8204],
          [1.8070, 1.4255, 1.1531]]],


      ...6],
          [1.8762, 1.6867, 1.5626],
          [1.9683, 1.4182, 1.1769]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.1579, 1.1016, 1.6830]],

         [[1.6926, 1.8125, 1.5481]],

         [[1.1133, 1.7029, 1.4074]],

    ...

         [[1.0099, 1.1582, 1.8603]],

         [[1.4307, 1.7713, 1.5930]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([300, 4, 1, 3])
SDDMM(message func: u_mul_e)
_____________________ test_sddmm[idtype0-mul-e-u-shp3-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'u', rhs_target = 'e', msg = 'mul'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd3fc0>
op = 'mul'
lhs_data = tensor([[1.3724, 1.4591, 1.3614],
        [1.2558, 1.0979, 1.9516],
        [1.9603, 1.0510, 1.1388],
        [1.3860,... 1.8649],
        [1.3237, 1.0581, 1.6371],
        [1.3278, 1.9360, 1.3043]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.4358, 1.0279, 1.6155],
        [1.1395, 1.6929, 1.0278],
        [1.2639, 1.8530, 1.8346],
        [1.7112,... 1.5788],
        [1.6616, 1.5229, 1.1072],
        [1.5474, 1.9247, 1.6150]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([100, 3])
SDDMM(message func: u_mul_e)
_____________________ test_sddmm[idtype0-mul-e-u-shp3-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'u', rhs_target = 'e', msg = 'mul'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72bd505a0>
op = 'mul'
lhs_data = tensor([[1.5201, 1.5445, 1.6694],
        [1.0258, 1.2446, 1.9489],
        [1.4834, 1.7409, 1.0631],
        [1.3784,... 1.9114],
        [1.3358, 1.3556, 1.3065],
        [1.2282, 1.9135, 1.0827]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.9937, 1.5702, 1.7991],
        [1.8722, 1.6941, 1.3649],
        [1.9039, 1.7871, 1.2303],
        [1.8948,... 1.1412],
        [1.3887, 1.1521, 1.7998],
        [1.5805, 1.9042, 1.4065]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([300, 3])
SDDMM(message func: u_mul_e)
_____________________ test_sddmm[idtype0-mul-e-u-shp4-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'u', rhs_target = 'e', msg = 'mul'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43e961048>
op = 'mul'
lhs_data = tensor([[1.1642],
        [1.2375],
        [1.5035],
        [1.0389],
        [1.8760],
        [1.2762],
        [1...997],
        [1.4861],
        [1.1064],
        [1.1989],
        [1.1036]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.6195],
        [1.7053],
        [1.7031],
        [1.3672],
        [1.8734],
        [1.1097],
        [1...467],
        [1.0339],
        [1.9913],
        [1.5524],
        [1.1420]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([100, 1])
SDDMM(message func: u_mul_e)
_____________________ test_sddmm[idtype0-mul-e-u-shp4-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'u', rhs_target = 'e', msg = 'mul'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72c051f78>
op = 'mul'
lhs_data = tensor([[1.0768],
        [1.0236],
        [1.0092],
        [1.7489],
        [1.5153],
        [1.1569],
        [1...862],
        [1.7320],
        [1.6302],
        [1.4124],
        [1.3658]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.2593],
        [1.4204],
        [1.6841],
        [1.1386],
        [1.5207],
        [1.7619],
        [1...137],
        [1.8948],
        [1.0163],
        [1.8506],
        [1.4189]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([300, 1])
SDDMM(message func: u_mul_e)
_____________________ test_sddmm[idtype0-mul-e-v-shp0-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': ...torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'v', rhs_target = 'e'
msg = 'mul', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb61f174bd0>
op = 'mul'
lhs_data = tensor([[[[[[1.1658],
            [1.2879],
            [1.3308]]],


          [[[1.8009],
            [1.5493],
    ...3]]],


          [[[1.5743],
            [1.8128],
            [1.1094]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.2792]],

           [[1.4643]],

           [[1.9769]]]],



         [[[[1.7496]],

           [[1.514...



         [[[[1.5876]],

           [[1.4089]],

           [[1.2501]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([100, 4, 1, 3, 1, 1])
SDDMM(message func: v_mul_e)
_____________________ test_sddmm[idtype0-mul-e-v-shp0-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'v', rhs_target = 'e'
msg = 'mul', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72aa9e630>
op = 'mul'
lhs_data = tensor([[[[[[1.3075],
            [1.9083],
            [1.9296]]],


          [[[1.0925],
            [1.8390],
    ...4]]],


          [[[1.5433],
            [1.8838],
            [1.3052]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.7958]],

           [[1.2582]],

           [[1.8026]]]],



         [[[[1.7004]],

           [[1.165...



         [[[[1.3630]],

           [[1.6917]],

           [[1.6379]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([40, 1, 2, 1, 3, 1]), rhs shape: torch.Size([300, 4, 1, 3, 1, 1])
SDDMM(message func: v_mul_e)
_____________________ test_sddmm[idtype0-mul-e-v-shp1-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Sch...pe=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'v', rhs_target = 'e'
msg = 'mul', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb4476b1240>
op = 'mul'
lhs_data = tensor([[[[[1.4085, 1.0010, 1.5031,  ..., 1.4730, 1.7961, 1.6322]],

          [[1.4783, 1.5227, 1.9458,  ..., 1.5199,...
          [[1.5619, 1.1602, 1.3188,  ..., 1.7716, 1.7537, 1.9728]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.2902, 1.9714, 1.4188,  ..., 1.2555, 1.6444, 1.7722],
           [1.9236, 1.5605, 1.7273,  ..., 1.9219, 1...
           [1.0559, 1.1506, 1.8783,  ..., 1.5147, 1.1438, 1.1013]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([100, 1, 3, 7, 7])
SDDMM(message func: v_mul_e)
_____________________ test_sddmm[idtype0-mul-e-v-shp1-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'v', rhs_target = 'e'
msg = 'mul', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb70622f870>
op = 'mul'
lhs_data = tensor([[[[[1.2531, 1.2894, 1.9953,  ..., 1.6765, 1.1281, 1.5220]],

          [[1.8461, 1.2460, 1.2754,  ..., 1.4522,...
          [[1.6731, 1.3153, 1.3616,  ..., 1.5880, 1.8054, 1.0122]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.3276, 1.4905, 1.1411,  ..., 1.6558, 1.5075, 1.2057],
           [1.4234, 1.8357, 1.7990,  ..., 1.7897, 1...
           [1.8371, 1.7149, 1.0030,  ..., 1.5398, 1.2794, 1.8114]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([40, 5, 3, 1, 7]), rhs shape: torch.Size([300, 1, 3, 7, 7])
SDDMM(message func: v_mul_e)
_____________________ test_sddmm[idtype0-mul-e-v-shp2-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme...dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'v', rhs_target = 'e', msg = 'mul'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72bd76708>
op = 'mul'
lhs_data = tensor([[[[1.7589, 1.6840, 1.0621],
          [1.7250, 1.6501, 1.1323],
          [1.4652, 1.1460, 1.2079]]],


      ...1],
          [1.5370, 1.8680, 1.5698],
          [1.3432, 1.2846, 1.6770]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.2712, 1.9779, 1.2866]],

         [[1.0847, 1.0979, 1.1385]],

         [[1.9986, 1.6884, 1.6886]],

    ...

         [[1.5847, 1.3770, 1.0644]],

         [[1.3734, 1.8261, 1.9138]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([100, 4, 1, 3])
SDDMM(message func: v_mul_e)
_____________________ test_sddmm[idtype0-mul-e-v-shp2-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'v', rhs_target = 'e', msg = 'mul'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72c060168>
op = 'mul'
lhs_data = tensor([[[[1.3089, 1.8696, 1.6580],
          [1.0443, 1.9462, 1.2462],
          [1.4845, 1.4021, 1.7593]]],


      ...8],
          [1.2344, 1.3851, 1.8836],
          [1.6786, 1.1050, 1.1435]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.6970, 1.1049, 1.7136]],

         [[1.2967, 1.0001, 1.5127]],

         [[1.3305, 1.8323, 1.7352]],

    ...

         [[1.8162, 1.0407, 1.6928]],

         [[1.4028, 1.7819, 1.0586]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([40, 1, 3, 3]), rhs shape: torch.Size([300, 4, 1, 3])
SDDMM(message func: v_mul_e)
_____________________ test_sddmm[idtype0-mul-e-v-shp3-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'v', rhs_target = 'e', msg = 'mul'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb457b61558>
op = 'mul'
lhs_data = tensor([[1.2224, 1.0345, 1.7816],
        [1.6028, 1.2182, 1.8710],
        [1.0095, 1.4133, 1.4941],
        [1.9115,... 1.0833],
        [1.1394, 1.7798, 1.9864],
        [1.2735, 1.0530, 1.9912]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.6021, 1.0316, 1.1274],
        [1.5554, 1.1671, 1.0547],
        [1.5038, 1.9284, 1.1104],
        [1.5078,... 1.3026],
        [1.1184, 1.9742, 1.7479],
        [1.1817, 1.8012, 1.2169]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([100, 3])
SDDMM(message func: v_mul_e)
_____________________ test_sddmm[idtype0-mul-e-v-shp3-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'v', rhs_target = 'e', msg = 'mul'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72aa9e168>
op = 'mul'
lhs_data = tensor([[1.3028, 1.1225, 1.3290],
        [1.1508, 1.6815, 1.9715],
        [1.5951, 1.3763, 1.1440],
        [1.4134,... 1.3548],
        [1.3867, 1.4942, 1.3122],
        [1.6474, 1.6325, 1.3318]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.8203, 1.0336, 1.4618],
        [1.0705, 1.8626, 1.9366],
        [1.4521, 1.1114, 1.4761],
        [1.7873,... 1.5453],
        [1.4926, 1.3622, 1.4683],
        [1.3458, 1.6815, 1.5852]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([40, 3]), rhs shape: torch.Size([300, 3])
SDDMM(message func: v_mul_e)
_____________________ test_sddmm[idtype0-mul-e-v-shp4-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'v', rhs_target = 'e', msg = 'mul'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acf3558>
op = 'mul'
lhs_data = tensor([[1.0660],
        [1.6974],
        [1.7842],
        [1.3994],
        [1.8322],
        [1.8055],
        [1...288],
        [1.7974],
        [1.2332],
        [1.3429],
        [1.7296]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.6045],
        [1.7298],
        [1.3808],
        [1.6797],
        [1.2018],
        [1.8308],
        [1...550],
        [1.1615],
        [1.5830],
        [1.4743],
        [1.1479]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([100, 1])
SDDMM(message func: v_mul_e)
_____________________ test_sddmm[idtype0-mul-e-v-shp4-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'v', rhs_target = 'e', msg = 'mul'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72aaba750>
op = 'mul'
lhs_data = tensor([[1.1243],
        [1.8712],
        [1.6370],
        [1.2621],
        [1.0715],
        [1.9880],
        [1...660],
        [1.3321],
        [1.2851],
        [1.6220],
        [1.2653]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.2569],
        [1.4489],
        [1.2709],
        [1.6247],
        [1.6473],
        [1.5295],
        [1...933],
        [1.4150],
        [1.7316],
        [1.3196],
        [1.0196]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([40, 1]), rhs shape: torch.Size([300, 1])
SDDMM(message func: v_mul_e)
_____________________ test_sddmm[idtype0-div-u-v-shp0-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': ...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'v', rhs_target = 'u'
msg = 'div', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72bd4ed80>
op = 'mul'
lhs_data = tensor([[[[[[1.9866],
            [1.5480],
            [1.5779]]],


          [[[1.1447],
            [1.6811],
    ...9]]],


          [[[1.0643],
            [1.6239],
            [1.5745]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[0.6377]],

           [[0.5649]],

           [[0.9094]]]],



         [[[[0.5400]],

           [[0.610...         [[[[0.6701]],

           [[0.9450]],

           [[0.5892]]]]]], dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: v_div_u)
_____________________ test_sddmm[idtype0-div-u-v-shp0-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'v', rhs_target = 'u'
msg = 'div', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72c064438>
op = 'mul'
lhs_data = tensor([[[[[[1.9099],
            [1.5367],
            [1.8763]]],


          [[[1.2631],
            [1.9042],
    ...1]]],


          [[[1.2370],
            [1.4687],
            [1.6097]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[0.7625]],

           [[0.9214]],

           [[0.6954]]]],



         [[[[0.5010]],

           [[0.862...         [[[[0.6420]],

           [[0.5795]],

           [[0.6126]]]]]], dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([40, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: v_div_u)
_____________________ test_sddmm[idtype0-div-u-v-shp1-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Sch...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'v', rhs_target = 'u'
msg = 'div', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb447687870>
op = 'mul'
lhs_data = tensor([[[[[1.8555, 1.9164, 1.5967,  ..., 1.1662, 1.4512, 1.1681]],

          [[1.6505, 1.0522, 1.7902,  ..., 1.6397,...
          [[1.7846, 1.9718, 1.8551,  ..., 1.3848, 1.4153, 1.6421]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[0.5024, 0.6521, 0.6525,  ..., 0.8856, 0.7492, 0.8938],
           [0.6349, 0.7362, 0.6234,  ..., 0.8438, 0...        [0.5620, 0.7750, 0.7954,  ..., 0.9466, 0.5349, 0.9389]]]]],
       dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: v_div_u)
_____________________ test_sddmm[idtype0-div-u-v-shp1-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'v', rhs_target = 'u'
msg = 'div', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72949ce58>
op = 'mul'
lhs_data = tensor([[[[[1.5673, 1.2480, 1.2600,  ..., 1.8784, 1.4879, 1.8699]],

          [[1.3166, 1.8324, 1.4255,  ..., 1.6636,...
          [[1.1049, 1.7645, 1.3913,  ..., 1.4622, 1.6118, 1.6491]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[0.5356, 0.5359, 0.5737,  ..., 0.9172, 0.8597, 0.8310],
           [0.7214, 0.7031, 0.7753,  ..., 0.6371, 0...        [0.6766, 0.5138, 0.6828,  ..., 0.7288, 0.5456, 0.7602]]]]],
       dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([40, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: v_div_u)
_____________________ test_sddmm[idtype0-div-u-v-shp2-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'v', rhs_target = 'u', msg = 'div'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb705836af8>
op = 'mul'
lhs_data = tensor([[[[1.8282, 1.0857, 1.0673],
          [1.0458, 1.4466, 1.3795],
          [1.5198, 1.5664, 1.1441]]],


      ...7],
          [1.2545, 1.9440, 1.0973],
          [1.2079, 1.4020, 1.5501]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[0.6112, 0.5844, 0.7674]],

         [[0.6056, 0.7481, 0.7537]],

         [[0.5407, 0.7531, 0.9137]],

    ...[[0.8544, 0.8876, 0.8377]],

         [[0.8571, 0.8380, 0.6504]]]], dtype=torch.float64,
       grad_fn=<MulBackward0>)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: v_div_u)
_____________________ test_sddmm[idtype0-div-u-v-shp2-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'v', rhs_target = 'u', msg = 'div'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43e6c1048>
op = 'mul'
lhs_data = tensor([[[[1.6433, 1.9790, 1.9988],
          [1.9050, 1.5134, 1.2155],
          [1.0781, 1.8728, 1.4729]]],


      ...9],
          [1.6063, 1.1504, 1.6423],
          [1.2266, 1.9064, 1.9879]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[0.5855, 0.5222, 0.6530]],

         [[0.8128, 0.5014, 0.6389]],

         [[0.7149, 0.5459, 0.7416]],

    ...[[0.7247, 0.7189, 0.6988]],

         [[0.8799, 0.5583, 0.5074]]]], dtype=torch.float64,
       grad_fn=<MulBackward0>)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([40, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: v_div_u)
_____________________ test_sddmm[idtype0-div-u-v-shp3-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'v', rhs_target = 'u', msg = 'div'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72c06d0d8>
op = 'mul'
lhs_data = tensor([[1.1064, 1.0093, 1.4689],
        [1.8075, 1.7615, 1.9978],
        [1.1142, 1.8379, 1.9771],
        [1.3419,... 1.8806],
        [1.3655, 1.2353, 1.2232],
        [1.7699, 1.5047, 1.1978]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.7378, 0.5708, 0.8614],
        [0.7521, 0.5038, 0.8533],
        [0.5430, 0.5513, 0.5756],
        [0.5486,...609],
        [0.6852, 0.5333, 0.5554],
        [0.6338, 0.7280, 0.9556]], dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: v_div_u)
_____________________ test_sddmm[idtype0-div-u-v-shp3-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'v', rhs_target = 'u', msg = 'div'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72c06d3a8>
op = 'mul'
lhs_data = tensor([[1.0621, 1.7863, 1.2616],
        [1.4445, 1.4513, 1.8602],
        [1.0796, 1.2316, 1.6199],
        [1.3527,... 1.6357],
        [1.8549, 1.0398, 1.3873],
        [1.0191, 1.6684, 1.3556]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.8502, 0.5658, 0.8037],
        [0.5082, 0.7623, 0.5355],
        [0.7728, 0.8042, 0.8028],
        [0.5732,...935],
        [0.8172, 0.6740, 0.5931],
        [0.5612, 0.8446, 0.8454]], dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([40, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: v_div_u)
_____________________ test_sddmm[idtype0-div-u-v-shp4-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'v', rhs_target = 'u', msg = 'div'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72bd4eaf8>
op = 'mul'
lhs_data = tensor([[1.2632],
        [1.3147],
        [1.9382],
        [1.5844],
        [1.4701],
        [1.4112],
        [1...571],
        [1.8783],
        [1.2501],
        [1.1138],
        [1.1380]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.5201],
        [0.7468],
        [0.7605],
        [0.7948],
        [0.5997],
        [0.5214],
        [0...,
        [0.6644],
        [0.9245],
        [0.8001],
        [0.8966]], dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: v_div_u)
_____________________ test_sddmm[idtype0-div-u-v-shp4-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'v', rhs_target = 'u', msg = 'div'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72bd4e798>
op = 'mul'
lhs_data = tensor([[1.2896],
        [1.0925],
        [1.4635],
        [1.7739],
        [1.6278],
        [1.8449],
        [1...515],
        [1.6816],
        [1.4621],
        [1.8584],
        [1.6362]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.6330],
        [0.8562],
        [0.9199],
        [0.5616],
        [0.7836],
        [0.6908],
        [0...,
        [0.7314],
        [0.5124],
        [0.8069],
        [0.8822]], dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([40, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: v_div_u)
_____________________ test_sddmm[idtype0-div-u-e-shp0-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...torch.float64), 'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'e', rhs_target = 'u'
msg = 'div', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72ae55c18>
op = 'mul'
lhs_data = tensor([[[[[[1.3762],
            [1.5574],
            [1.0369]]],


          [[[1.3514],
            [1.4346],
    ...4]]],


          [[[1.4585],
            [1.3787],
            [1.1078]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[0.5401]],

           [[0.8938]],

           [[0.6074]]]],



         [[[[0.9024]],

           [[0.544...         [[[[0.5276]],

           [[0.9780]],

           [[0.6039]]]]]], dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([100, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: e_div_u)
_____________________ test_sddmm[idtype0-div-u-e-shp0-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'e', rhs_target = 'u'
msg = 'div', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb70623f510>
op = 'mul'
lhs_data = tensor([[[[[[1.0216],
            [1.4480],
            [1.2476]]],


          [[[1.4477],
            [1.7414],
    ...4]]],


          [[[1.9746],
            [1.5940],
            [1.8647]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[0.6363]],

           [[0.9044]],

           [[0.6203]]]],



         [[[[0.5992]],

           [[0.601...         [[[[0.7952]],

           [[0.7904]],

           [[0.6823]]]]]], dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([300, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: e_div_u)
_____________________ test_sddmm[idtype0-div-u-e-shp1-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...pe=torch.float64), 'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'e', rhs_target = 'u'
msg = 'div', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72c06e870>
op = 'mul'
lhs_data = tensor([[[[[1.5019, 1.3927, 1.9067,  ..., 1.2266, 1.6784, 1.1655]],

          [[1.0651, 1.1120, 1.9446,  ..., 1.7489,...
          [[1.2487, 1.6271, 1.5927,  ..., 1.5852, 1.5059, 1.4011]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[0.5185, 0.5625, 0.7030,  ..., 0.6489, 0.8936, 0.9950],
           [0.5459, 0.6319, 0.9114,  ..., 0.7848, 0...        [0.6733, 0.6172, 0.9113,  ..., 0.7580, 0.7124, 0.8746]]]]],
       dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([100, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: e_div_u)
_____________________ test_sddmm[idtype0-div-u-e-shp1-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'e', rhs_target = 'u'
msg = 'div', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72c065288>
op = 'mul'
lhs_data = tensor([[[[[1.6714, 1.8480, 1.5386,  ..., 1.9696, 1.0237, 1.0453]],

          [[1.8062, 1.3812, 1.6890,  ..., 1.4616,...
          [[1.8427, 1.1109, 1.6216,  ..., 1.9619, 1.9209, 1.6633]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[0.5655, 0.7564, 0.6766,  ..., 0.8273, 0.6345, 0.9724],
           [0.6518, 0.8497, 0.5451,  ..., 0.8419, 0...        [0.7330, 0.6397, 0.7462,  ..., 0.5521, 0.7679, 0.5304]]]]],
       dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([300, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: e_div_u)
_____________________ test_sddmm[idtype0-div-u-e-shp2-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...dtype=torch.float64), 'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'e', rhs_target = 'u', msg = 'div'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb705b2ba20>
op = 'mul'
lhs_data = tensor([[[[1.7174, 1.5160, 1.8857],
          [1.8141, 1.6808, 1.9435],
          [1.7215, 1.5266, 1.0058]]],


      ...2],
          [2.0000, 1.7846, 1.1775],
          [1.5179, 1.2272, 1.2968]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[0.8168, 0.5050, 0.6534]],

         [[0.6053, 0.5178, 0.6719]],

         [[0.5432, 0.5030, 0.5614]],

    ...[[0.5908, 0.5323, 0.6337]],

         [[0.7600, 0.6696, 0.5185]]]], dtype=torch.float64,
       grad_fn=<MulBackward0>)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([100, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: e_div_u)
_____________________ test_sddmm[idtype0-div-u-e-shp2-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'e', rhs_target = 'u', msg = 'div'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb729529b40>
op = 'mul'
lhs_data = tensor([[[[1.1597, 1.0411, 1.3715],
          [1.3070, 1.8977, 1.6456],
          [1.0594, 1.4268, 1.6009]]],


      ...8],
          [1.9542, 1.2592, 1.8294],
          [1.9938, 1.8743, 1.6894]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[0.8155, 0.5636, 0.7172]],

         [[0.5441, 0.7508, 0.6817]],

         [[0.5434, 0.8700, 0.5293]],

    ...[[0.5494, 0.5600, 0.6420]],

         [[0.5538, 0.8677, 0.9131]]]], dtype=torch.float64,
       grad_fn=<MulBackward0>)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([300, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: e_div_u)
_____________________ test_sddmm[idtype0-div-u-e-shp3-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'e', rhs_target = 'u', msg = 'div'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb706237288>
op = 'mul'
lhs_data = tensor([[1.7400, 1.9278, 1.2637],
        [1.4755, 1.5598, 1.1050],
        [1.0878, 1.8436, 1.6716],
        [1.3789,... 1.8671],
        [1.6431, 1.9599, 1.9209],
        [1.8181, 1.4325, 1.0683]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.7030, 0.6346, 0.5024],
        [0.7149, 0.6148, 0.6046],
        [0.7011, 0.9979, 0.8029],
        [0.5918,...299],
        [0.6400, 0.5402, 0.7872],
        [0.8700, 0.5546, 0.5222]], dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([100, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: e_div_u)
_____________________ test_sddmm[idtype0-div-u-e-shp3-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'e', rhs_target = 'u', msg = 'div'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb61cf94a20>
op = 'mul'
lhs_data = tensor([[1.4369, 1.2047, 1.7964],
        [1.2622, 1.6497, 1.7312],
        [1.2008, 1.1699, 1.6222],
        [1.9506,... 1.4150],
        [1.7175, 1.2342, 1.8687],
        [1.4474, 1.0995, 1.3317]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.8518, 0.9181, 0.5242],
        [0.7975, 0.8318, 0.5009],
        [0.6452, 0.6103, 0.5148],
        [0.7509,...651],
        [0.5015, 0.5095, 0.5001],
        [0.7272, 0.6613, 0.9556]], dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([300, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: e_div_u)
_____________________ test_sddmm[idtype0-div-u-e-shp4-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'e', rhs_target = 'u', msg = 'div'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72c065f78>
op = 'mul'
lhs_data = tensor([[1.0654],
        [1.3684],
        [1.4169],
        [1.0510],
        [1.6884],
        [1.2486],
        [1...710],
        [1.5973],
        [1.7156],
        [1.5491],
        [1.0922]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.8495],
        [0.8427],
        [0.6854],
        [0.6014],
        [0.7531],
        [0.6234],
        [0...,
        [0.7437],
        [0.5955],
        [0.8215],
        [0.6168]], dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([100, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: e_div_u)
_____________________ test_sddmm[idtype0-div-u-e-shp4-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'e', rhs_target = 'u', msg = 'div'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72c07a288>
op = 'mul'
lhs_data = tensor([[1.7709],
        [1.7391],
        [1.4799],
        [1.1784],
        [1.7542],
        [1.5258],
        [1...994],
        [1.0645],
        [1.2962],
        [1.2424],
        [1.0197]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.8719],
        [0.5305],
        [0.5559],
        [0.6178],
        [0.6773],
        [0.6215],
        [0...,
        [0.5410],
        [0.6135],
        [0.5899],
        [0.6733]], dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([300, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: e_div_u)
_____________________ test_sddmm[idtype0-div-v-u-shp0-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': ...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'u', rhs_target = 'v'
msg = 'div', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb729529d80>
op = 'mul'
lhs_data = tensor([[[[[[1.2215],
            [1.8061],
            [1.6979]]],


          [[[1.2019],
            [1.1889],
    ...8]]],


          [[[1.3645],
            [1.5436],
            [1.5398]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[0.6664]],

           [[0.7925]],

           [[0.9062]]]],



         [[[[0.5621]],

           [[0.546...         [[[[0.5843]],

           [[0.7033]],

           [[0.6909]]]]]], dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: u_div_v)
_____________________ test_sddmm[idtype0-div-v-u-shp0-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'u', rhs_target = 'v'
msg = 'div', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72aaa0fc0>
op = 'mul'
lhs_data = tensor([[[[[[1.3546],
            [1.7960],
            [1.3730]]],


          [[[1.3833],
            [1.9315],
    ...8]]],


          [[[1.0959],
            [1.9987],
            [1.6210]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[0.7196]],

           [[0.7577]],

           [[0.6283]]]],



         [[[[0.8342]],

           [[0.627...         [[[[0.7039]],

           [[0.8008]],

           [[0.6043]]]]]], dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([40, 4, 1, 3, 1, 1])
SDDMM(message func: u_div_v)
_____________________ test_sddmm[idtype0-div-v-u-shp1-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Sch...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'u', rhs_target = 'v'
msg = 'div', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72abb5438>
op = 'mul'
lhs_data = tensor([[[[[1.3603, 1.5240, 1.3988,  ..., 1.4541, 1.8781, 1.5246]],

          [[1.3830, 1.3463, 1.2383,  ..., 1.2599,...
          [[1.3167, 1.5792, 1.0858,  ..., 1.8480, 1.8872, 1.2833]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[0.8812, 0.5572, 0.6619,  ..., 0.8246, 0.7454, 0.5387],
           [0.6458, 0.7358, 0.6061,  ..., 0.7216, 0...        [0.6367, 0.5411, 0.9808,  ..., 0.8092, 0.6121, 0.7229]]]]],
       dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: u_div_v)
_____________________ test_sddmm[idtype0-div-v-u-shp1-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'u', rhs_target = 'v'
msg = 'div', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43e930168>
op = 'mul'
lhs_data = tensor([[[[[1.9313, 1.4095, 1.1829,  ..., 1.4359, 1.8346, 1.9978]],

          [[1.2666, 1.0177, 1.4454,  ..., 1.3349,...
          [[1.0964, 1.9068, 1.1921,  ..., 1.6533, 1.8000, 1.1092]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[0.6297, 0.6729, 0.6590,  ..., 0.8269, 0.5950, 0.6062],
           [0.9605, 0.5766, 0.7982,  ..., 0.6771, 0...        [0.9617, 0.5305, 0.5736,  ..., 0.7481, 0.5414, 0.5454]]]]],
       dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([40, 1, 3, 7, 7])
SDDMM(message func: u_div_v)
_____________________ test_sddmm[idtype0-div-v-u-shp2-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'u', rhs_target = 'v', msg = 'div'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb61cfa4f30>
op = 'mul'
lhs_data = tensor([[[[1.5871, 1.6335, 1.0872],
          [1.1995, 1.8031, 1.3624],
          [1.9752, 1.1532, 1.2971]]],


      ...1],
          [1.7133, 1.0417, 1.3432],
          [1.1160, 1.7318, 1.1000]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[0.6293, 0.8364, 0.8648]],

         [[0.6618, 0.6059, 0.7023]],

         [[0.6813, 0.8854, 0.5106]],

    ...[[0.6516, 0.8252, 0.5535]],

         [[0.5345, 0.6634, 0.5462]]]], dtype=torch.float64,
       grad_fn=<MulBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: u_div_v)
_____________________ test_sddmm[idtype0-div-v-u-shp2-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'u', rhs_target = 'v', msg = 'div'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72c07e1b0>
op = 'mul'
lhs_data = tensor([[[[1.8991, 1.9952, 1.8711],
          [1.9800, 1.8389, 1.2116],
          [1.4986, 1.1650, 1.5708]]],


      ...9],
          [1.9299, 1.0354, 1.1391],
          [1.0341, 1.7410, 1.1279]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[0.5616, 0.6487, 0.5987]],

         [[0.7017, 0.6440, 0.5593]],

         [[0.5563, 0.5081, 0.7718]],

    ...[[0.5521, 0.5048, 0.5962]],

         [[0.7934, 0.9548, 0.5188]]]], dtype=torch.float64,
       grad_fn=<MulBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([40, 4, 1, 3])
SDDMM(message func: u_div_v)
_____________________ test_sddmm[idtype0-div-v-u-shp3-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'u', rhs_target = 'v', msg = 'div'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72c07e4c8>
op = 'mul'
lhs_data = tensor([[1.0316, 1.9402, 1.6978],
        [1.2887, 1.8011, 1.2617],
        [1.4274, 1.4769, 1.5826],
        [1.9380,... 1.6632],
        [1.6315, 1.2173, 1.6048],
        [1.9609, 1.4796, 1.6373]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.5020, 0.6843, 0.7144],
        [0.5284, 0.5928, 0.8069],
        [0.8257, 0.8655, 0.5545],
        [0.6584,...889],
        [0.7758, 0.7542, 0.7164],
        [0.5042, 0.6271, 0.9211]], dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: u_div_v)
_____________________ test_sddmm[idtype0-div-v-u-shp3-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'u', rhs_target = 'v', msg = 'div'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72c064cf0>
op = 'mul'
lhs_data = tensor([[1.2567, 1.5468, 1.3266],
        [1.9735, 1.7894, 1.9910],
        [1.4111, 1.5437, 1.6004],
        [1.8934,... 1.3243],
        [1.9359, 1.5764, 1.2190],
        [1.7758, 1.1517, 1.0635]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.9383, 0.6443, 0.8234],
        [0.6487, 0.7619, 0.5480],
        [0.5462, 0.6679, 0.7851],
        [0.9898,...374],
        [0.5627, 0.5197, 0.5443],
        [0.7696, 0.9537, 0.9585]], dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([40, 3])
SDDMM(message func: u_div_v)
_____________________ test_sddmm[idtype0-div-v-u-shp4-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'u', rhs_target = 'v', msg = 'div'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72bd76798>
op = 'mul'
lhs_data = tensor([[1.6704],
        [1.8181],
        [1.8854],
        [1.9441],
        [1.9652],
        [1.2066],
        [1...823],
        [1.4391],
        [1.8008],
        [1.2599],
        [1.1812]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.5291],
        [0.6635],
        [0.5235],
        [0.5431],
        [0.7508],
        [0.6023],
        [0...,
        [0.5581],
        [0.7853],
        [0.6989],
        [0.5768]], dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: u_div_v)
_____________________ test_sddmm[idtype0-div-v-u-shp4-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'u', rhs_target = 'v', msg = 'div'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb7058360d8>
op = 'mul'
lhs_data = tensor([[1.5089],
        [1.5938],
        [1.5492],
        [1.8543],
        [1.1548],
        [1.3603],
        [1...870],
        [1.1616],
        [1.1569],
        [1.9738],
        [1.9309]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.5219],
        [0.7416],
        [0.5028],
        [0.5731],
        [0.6776],
        [0.5432],
        [0...,
        [0.6186],
        [0.5777],
        [0.5058],
        [0.6025]], dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([40, 1])
SDDMM(message func: u_div_v)
_____________________ test_sddmm[idtype0-div-v-e-shp0-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...torch.float64), 'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'e', rhs_target = 'v'
msg = 'div', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72bd808b8>
op = 'mul'
lhs_data = tensor([[[[[[1.6568],
            [1.5851],
            [1.8434]]],


          [[[1.6286],
            [1.0904],
    ...3]]],


          [[[1.7986],
            [1.7755],
            [1.8013]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[0.6731]],

           [[0.7254]],

           [[0.6020]]]],



         [[[[0.7011]],

           [[0.709...         [[[[0.7159]],

           [[0.6563]],

           [[0.5240]]]]]], dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([100, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: e_div_v)
_____________________ test_sddmm[idtype0-div-v-e-shp0-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'e', rhs_target = 'v'
msg = 'div', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72c0d78b8>
op = 'mul'
lhs_data = tensor([[[[[[1.8599],
            [1.0659],
            [1.4443]]],


          [[[1.8542],
            [1.5634],
    ...0]]],


          [[[1.1381],
            [1.5678],
            [1.7528]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[0.5210]],

           [[0.5748]],

           [[0.9797]]]],



         [[[[0.9153]],

           [[0.956...         [[[[0.5219]],

           [[0.5352]],

           [[0.7394]]]]]], dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([300, 1, 2, 1, 3, 1]), rhs shape: torch.Size([40, 4, 1, 3, 1, 1])
SDDMM(message func: e_div_v)
_____________________ test_sddmm[idtype0-div-v-e-shp1-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...pe=torch.float64), 'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'e', rhs_target = 'v'
msg = 'div', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb7294bba68>
op = 'mul'
lhs_data = tensor([[[[[1.7659, 1.7164, 1.9858,  ..., 1.4154, 1.8837, 1.0904]],

          [[1.0740, 1.4350, 1.5604,  ..., 1.6351,...
          [[1.1507, 1.2047, 1.7526,  ..., 1.5836, 1.3597, 1.7299]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[0.6831, 0.7068, 0.8253,  ..., 0.6304, 0.5234, 0.6293],
           [0.5021, 0.5775, 0.7580,  ..., 0.5969, 0...        [0.5793, 0.8495, 0.6622,  ..., 0.5722, 0.5267, 0.7271]]]]],
       dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([100, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: e_div_v)
_____________________ test_sddmm[idtype0-div-v-e-shp1-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'e', rhs_target = 'v'
msg = 'div', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb7294bb168>
op = 'mul'
lhs_data = tensor([[[[[1.6653, 1.3021, 1.5754,  ..., 1.6931, 1.5621, 1.8680]],

          [[1.6887, 1.1986, 1.2692,  ..., 1.6482,...
          [[1.4100, 1.8557, 1.8667,  ..., 1.6440, 1.4655, 1.1896]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[0.7525, 0.9506, 0.8584,  ..., 0.5325, 0.7431, 0.6060],
           [0.5889, 0.9232, 0.8268,  ..., 0.5011, 0...        [0.7712, 0.8617, 0.7013,  ..., 0.5929, 0.5483, 0.5154]]]]],
       dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([300, 5, 3, 1, 7]), rhs shape: torch.Size([40, 1, 3, 7, 7])
SDDMM(message func: e_div_v)
_____________________ test_sddmm[idtype0-div-v-e-shp2-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...dtype=torch.float64), 'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'e', rhs_target = 'v', msg = 'div'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72c0ed6c0>
op = 'mul'
lhs_data = tensor([[[[1.7397, 1.0739, 1.7391],
          [1.7849, 1.4351, 1.7063],
          [1.4272, 1.9307, 1.5587]]],


      ...7],
          [1.1692, 1.8509, 1.2351],
          [1.9321, 1.1859, 1.6659]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[0.7143, 0.7792, 0.7648]],

         [[0.5008, 0.6617, 0.7071]],

         [[0.5709, 0.5683, 0.5215]],

    ...[[0.5232, 0.9002, 0.6652]],

         [[0.6246, 0.8847, 0.9173]]]], dtype=torch.float64,
       grad_fn=<MulBackward0>)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([100, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: e_div_v)
_____________________ test_sddmm[idtype0-div-v-e-shp2-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'e', rhs_target = 'v', msg = 'div'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72c0ed168>
op = 'mul'
lhs_data = tensor([[[[1.7885, 1.1409, 1.7997],
          [1.8075, 1.1268, 1.5178],
          [1.3090, 1.1804, 1.7897]]],


      ...1],
          [1.1748, 1.0072, 1.1882],
          [1.7232, 1.7574, 1.8170]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[0.9799, 0.7367, 0.5104]],

         [[0.8863, 0.6030, 0.5758]],

         [[0.5237, 0.5927, 0.6795]],

    ...[[0.9040, 0.5456, 0.6813]],

         [[0.7692, 0.8078, 0.6068]]]], dtype=torch.float64,
       grad_fn=<MulBackward0>)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([300, 1, 3, 3]), rhs shape: torch.Size([40, 4, 1, 3])
SDDMM(message func: e_div_v)
_____________________ test_sddmm[idtype0-div-v-e-shp3-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'e', rhs_target = 'v', msg = 'div'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72b8255e8>
op = 'mul'
lhs_data = tensor([[1.0972, 1.5571, 1.4325],
        [1.8840, 1.5881, 1.7085],
        [1.0623, 1.8818, 1.4494],
        [1.2662,... 1.7415],
        [1.6961, 1.1335, 1.8053],
        [1.2424, 1.0903, 1.4787]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.7345, 0.6697, 0.7467],
        [0.7335, 0.7964, 0.6312],
        [0.6986, 0.8320, 0.5171],
        [0.7699,...673],
        [0.5534, 0.5604, 0.6779],
        [0.6414, 0.8302, 0.6151]], dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([100, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: e_div_v)
_____________________ test_sddmm[idtype0-div-v-e-shp3-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'e', rhs_target = 'v', msg = 'div'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72bd76798>
op = 'mul'
lhs_data = tensor([[1.2308, 1.2708, 1.7950],
        [1.0666, 1.9308, 1.8350],
        [1.8648, 1.0786, 1.2477],
        [1.4555,... 1.9389],
        [1.3623, 1.2261, 1.4412],
        [1.1856, 1.2648, 1.9553]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.5794, 0.5872, 0.6625],
        [0.8219, 0.5103, 0.7505],
        [0.5330, 0.8525, 0.7770],
        [0.5034,...356],
        [0.6047, 0.5008, 0.5238],
        [0.8515, 0.6995, 0.7585]], dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([300, 3]), rhs shape: torch.Size([40, 3])
SDDMM(message func: e_div_v)
_____________________ test_sddmm[idtype0-div-v-e-shp4-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'e', rhs_target = 'v', msg = 'div'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb457b46ea0>
op = 'mul'
lhs_data = tensor([[1.2722],
        [1.6575],
        [1.1107],
        [1.4778],
        [1.8017],
        [1.8190],
        [1...473],
        [1.7082],
        [1.4082],
        [1.1997],
        [1.4178]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.6307],
        [0.5626],
        [0.9464],
        [0.6598],
        [0.9720],
        [0.6039],
        [0...,
        [0.9013],
        [0.7623],
        [0.6689],
        [0.6388]], dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([100, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: e_div_v)
_____________________ test_sddmm[idtype0-div-v-e-shp4-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'e', rhs_target = 'v', msg = 'div'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb61cfb5480>
op = 'mul'
lhs_data = tensor([[1.5059],
        [1.9850],
        [1.8420],
        [1.9422],
        [1.4088],
        [1.6924],
        [1...652],
        [1.6762],
        [1.7371],
        [1.3864],
        [1.9007]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.6646],
        [0.8538],
        [0.7749],
        [0.5928],
        [0.5449],
        [0.6743],
        [0...,
        [0.7296],
        [0.5079],
        [0.5033],
        [0.5919]], dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([300, 1]), rhs shape: torch.Size([40, 1])
SDDMM(message func: e_div_v)
_____________________ test_sddmm[idtype0-div-e-u-shp0-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': ...torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'u', rhs_target = 'e'
msg = 'div', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72c0c3240>
op = 'mul'
lhs_data = tensor([[[[[[1.6206],
            [1.8776],
            [1.0995]]],


          [[[1.2913],
            [1.9432],
    ...9]]],


          [[[1.5954],
            [1.0246],
            [1.2421]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[0.5347]],

           [[0.6582]],

           [[0.7037]]]],



         [[[[0.5457]],

           [[0.936...         [[[[0.5350]],

           [[0.7860]],

           [[0.5306]]]]]], dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([100, 4, 1, 3, 1, 1])
SDDMM(message func: u_div_e)
_____________________ test_sddmm[idtype0-div-e-u-shp0-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'u', rhs_target = 'e'
msg = 'div', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72c0c34c8>
op = 'mul'
lhs_data = tensor([[[[[[1.3870],
            [1.9628],
            [1.8338]]],


          [[[1.5869],
            [1.2319],
    ...6]]],


          [[[1.3799],
            [1.7796],
            [1.4120]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[0.9419]],

           [[0.7452]],

           [[0.6260]]]],



         [[[[0.7923]],

           [[0.783...         [[[[0.8148]],

           [[0.7065]],

           [[0.5465]]]]]], dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([300, 4, 1, 3, 1, 1])
SDDMM(message func: u_div_e)
_____________________ test_sddmm[idtype0-div-e-u-shp1-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Sch...pe=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'u', rhs_target = 'e'
msg = 'div', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72c060510>
op = 'mul'
lhs_data = tensor([[[[[1.6726, 1.2244, 1.9223,  ..., 1.2894, 1.3351, 1.1244]],

          [[1.5337, 1.9583, 1.7329,  ..., 1.3652,...
          [[1.4669, 1.3121, 1.4975,  ..., 1.7890, 1.0135, 1.1988]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[0.8912, 0.5686, 0.9188,  ..., 0.5845, 0.7493, 0.5541],
           [0.8410, 0.8880, 0.5166,  ..., 0.9294, 0...        [0.9067, 0.5436, 0.6268,  ..., 0.5448, 0.6572, 0.7631]]]]],
       dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([100, 1, 3, 7, 7])
SDDMM(message func: u_div_e)
_____________________ test_sddmm[idtype0-div-e-u-shp1-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'u', rhs_target = 'e'
msg = 'div', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43e8598b8>
op = 'mul'
lhs_data = tensor([[[[[1.2700, 1.6312, 1.7687,  ..., 1.4082, 1.6589, 1.8668]],

          [[1.7440, 1.3188, 1.9603,  ..., 1.7350,...
          [[1.1321, 1.2129, 1.5962,  ..., 1.0399, 1.3332, 1.9615]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[0.7367, 0.5115, 0.5947,  ..., 0.5184, 0.5641, 0.8891],
           [0.6241, 0.7131, 0.5841,  ..., 0.5486, 0...        [0.5342, 0.6197, 0.7269,  ..., 0.5177, 0.6437, 0.6298]]]]],
       dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([300, 1, 3, 7, 7])
SDDMM(message func: u_div_e)
_____________________ test_sddmm[idtype0-div-e-u-shp2-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme...dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'u', rhs_target = 'e', msg = 'div'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb7294c0558>
op = 'mul'
lhs_data = tensor([[[[1.6071, 1.4537, 1.1727],
          [1.1467, 1.2386, 1.3501],
          [1.8026, 1.8286, 1.5050]]],


      ...7],
          [1.2569, 1.9382, 1.3645],
          [1.0923, 1.4607, 1.8795]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[0.9701, 0.6393, 0.7174]],

         [[0.6485, 0.8209, 0.5691]],

         [[0.5969, 0.7840, 0.6058]],

    ...[[0.5749, 0.5963, 0.5713]],

         [[0.9082, 0.5799, 0.7885]]]], dtype=torch.float64,
       grad_fn=<MulBackward0>)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([100, 4, 1, 3])
SDDMM(message func: u_div_e)
_____________________ test_sddmm[idtype0-div-e-u-shp2-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'u', rhs_target = 'e', msg = 'div'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb4476b1ca8>
op = 'mul'
lhs_data = tensor([[[[1.5415, 1.1223, 1.4392],
          [1.8694, 1.6635, 1.6717],
          [1.3310, 1.6013, 1.4946]]],


      ...3],
          [1.6461, 1.3887, 1.8944],
          [1.1987, 1.0023, 1.3954]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[0.7118, 0.5675, 0.7694]],

         [[0.8795, 0.9166, 0.6337]],

         [[0.5088, 0.5661, 0.7242]],

    ...[[0.8291, 0.7136, 0.6220]],

         [[0.7039, 0.5710, 0.5498]]]], dtype=torch.float64,
       grad_fn=<MulBackward0>)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([300, 4, 1, 3])
SDDMM(message func: u_div_e)
_____________________ test_sddmm[idtype0-div-e-u-shp3-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'u', rhs_target = 'e', msg = 'div'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72c0f81b0>
op = 'mul'
lhs_data = tensor([[1.6260, 1.4132, 1.2763],
        [1.5482, 1.3077, 1.1432],
        [1.0900, 1.4776, 1.3477],
        [1.4207,... 1.1349],
        [1.8939, 1.6285, 1.0954],
        [1.7755, 1.9335, 1.5141]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.8567, 0.5250, 0.5352],
        [0.8538, 0.9636, 0.6001],
        [0.6660, 0.6220, 0.6609],
        [0.5910,...040],
        [0.6915, 0.5684, 0.8010],
        [0.5657, 0.5285, 0.5073]], dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([100, 3])
SDDMM(message func: u_div_e)
_____________________ test_sddmm[idtype0-div-e-u-shp3-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'u', rhs_target = 'e', msg = 'div'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72c0f83f0>
op = 'mul'
lhs_data = tensor([[1.5862, 1.1097, 1.8181],
        [1.8607, 1.8071, 1.7281],
        [1.6437, 1.4726, 1.5934],
        [1.8774,... 1.8782],
        [1.3578, 1.2334, 1.1907],
        [1.5748, 1.4363, 1.1877]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.7169, 0.5854, 0.7248],
        [0.5016, 0.6372, 0.9046],
        [0.6954, 0.7691, 0.9356],
        [0.5954,...704],
        [0.6248, 0.6162, 0.6571],
        [0.7472, 0.7193, 0.6151]], dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([300, 3])
SDDMM(message func: u_div_e)
_____________________ test_sddmm[idtype0-div-e-u-shp4-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'u', rhs_target = 'e', msg = 'div'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72c0f86c0>
op = 'mul'
lhs_data = tensor([[1.4571],
        [1.5717],
        [1.9055],
        [1.4150],
        [1.2286],
        [1.3118],
        [1...137],
        [1.2637],
        [1.1959],
        [1.4617],
        [1.3788]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.5216],
        [0.7501],
        [0.9509],
        [0.5260],
        [0.8139],
        [0.6904],
        [0...,
        [0.5582],
        [0.6625],
        [0.5007],
        [0.8873]], dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([100, 1])
SDDMM(message func: u_div_e)
_____________________ test_sddmm[idtype0-div-e-u-shp4-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'u', rhs_target = 'e', msg = 'div'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72c0f8948>
op = 'mul'
lhs_data = tensor([[1.5004],
        [1.2141],
        [1.0318],
        [1.8099],
        [1.0613],
        [1.0878],
        [1...157],
        [1.8194],
        [1.4410],
        [1.6976],
        [1.3485]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.7750],
        [0.5197],
        [0.6319],
        [0.5978],
        [0.9161],
        [0.9469],
        [0...,
        [0.6271],
        [0.7250],
        [0.6871],
        [0.7623]], dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([300, 1])
SDDMM(message func: u_div_e)
_____________________ test_sddmm[idtype0-div-e-v-shp0-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': ...torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'v', rhs_target = 'e'
msg = 'div', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72bd4b1f8>
op = 'mul'
lhs_data = tensor([[[[[[1.2722],
            [1.6776],
            [1.7314]]],


          [[[1.2549],
            [1.8267],
    ...1]]],


          [[[1.8519],
            [1.1652],
            [1.9334]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[0.6344]],

           [[0.8009]],

           [[0.6518]]]],



         [[[[0.8968]],

           [[0.636...         [[[[0.9324]],

           [[0.8493]],

           [[0.8163]]]]]], dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([100, 4, 1, 3, 1, 1])
SDDMM(message func: v_div_e)
_____________________ test_sddmm[idtype0-div-e-v-shp0-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'v', rhs_target = 'e'
msg = 'div', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb61421cc60>
op = 'mul'
lhs_data = tensor([[[[[[1.8473],
            [1.7931],
            [1.0033]]],


          [[[1.8619],
            [1.6976],
    ...5]]],


          [[[1.8853],
            [1.6152],
            [1.6836]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[0.6308]],

           [[0.5494]],

           [[0.7307]]]],



         [[[[0.5101]],

           [[0.699...         [[[[0.6638]],

           [[0.6828]],

           [[0.9017]]]]]], dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([40, 1, 2, 1, 3, 1]), rhs shape: torch.Size([300, 4, 1, 3, 1, 1])
SDDMM(message func: v_div_e)
_____________________ test_sddmm[idtype0-div-e-v-shp1-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Sch...pe=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'v', rhs_target = 'e'
msg = 'div', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72c0c9bd0>
op = 'mul'
lhs_data = tensor([[[[[1.1727, 1.5533, 1.8044,  ..., 1.0666, 1.7240, 1.5950]],

          [[1.9833, 1.7288, 1.3037,  ..., 1.3577,...
          [[1.0575, 1.9967, 1.7336,  ..., 1.9503, 1.8868, 1.5685]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[0.6180, 0.6880, 0.7843,  ..., 0.6692, 0.5606, 0.6481],
           [0.8177, 0.5315, 0.5247,  ..., 0.5926, 0...        [0.5920, 0.5170, 0.8276,  ..., 0.6055, 0.6568, 0.5574]]]]],
       dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([100, 1, 3, 7, 7])
SDDMM(message func: v_div_e)
_____________________ test_sddmm[idtype0-div-e-v-shp1-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'v', rhs_target = 'e'
msg = 'div', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72c0c9a20>
op = 'mul'
lhs_data = tensor([[[[[1.7076, 1.4480, 1.5398,  ..., 1.6947, 1.5130, 1.5359]],

          [[1.1456, 1.7426, 1.8378,  ..., 1.3600,...
          [[1.5781, 1.8695, 1.5133,  ..., 1.3046, 1.7949, 1.4840]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[0.6705, 0.5309, 0.9444,  ..., 0.9081, 0.6442, 0.5133],
           [0.5192, 0.9570, 0.9195,  ..., 0.7528, 0...        [0.8271, 0.7153, 0.5235,  ..., 0.6383, 0.5065, 0.6299]]]]],
       dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([40, 5, 3, 1, 7]), rhs shape: torch.Size([300, 1, 3, 7, 7])
SDDMM(message func: v_div_e)
_____________________ test_sddmm[idtype0-div-e-v-shp2-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme...dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'v', rhs_target = 'e', msg = 'div'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acf84c8>
op = 'mul'
lhs_data = tensor([[[[1.3344, 1.0130, 1.8957],
          [1.2831, 1.3497, 1.2426],
          [1.5127, 1.6926, 1.9407]]],


      ...7],
          [1.2858, 1.5756, 1.0830],
          [1.8427, 1.6452, 1.8694]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[0.7007, 0.5016, 0.8777]],

         [[0.6689, 0.6578, 0.7337]],

         [[0.8645, 0.7660, 0.7256]],

    ...[[0.7032, 0.5886, 0.7901]],

         [[0.7547, 0.7369, 0.5439]]]], dtype=torch.float64,
       grad_fn=<MulBackward0>)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([100, 4, 1, 3])
SDDMM(message func: v_div_e)
_____________________ test_sddmm[idtype0-div-e-v-shp2-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'v', rhs_target = 'e', msg = 'div'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72bd5c120>
op = 'mul'
lhs_data = tensor([[[[1.7777, 1.4325, 1.6425],
          [1.9743, 1.8445, 1.8093],
          [1.2732, 1.6159, 1.6446]]],


      ...4],
          [1.4588, 1.7225, 1.2866],
          [1.0509, 1.8525, 1.2671]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[0.7632, 0.7313, 0.7049]],

         [[0.6520, 0.7494, 0.5385]],

         [[0.7590, 0.6425, 0.8068]],

    ...[[0.6496, 0.6918, 0.5038]],

         [[0.9643, 0.8171, 0.8920]]]], dtype=torch.float64,
       grad_fn=<MulBackward0>)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([40, 1, 3, 3]), rhs shape: torch.Size([300, 4, 1, 3])
SDDMM(message func: v_div_e)
_____________________ test_sddmm[idtype0-div-e-v-shp3-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'v', rhs_target = 'e', msg = 'div'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72aaba0d8>
op = 'mul'
lhs_data = tensor([[1.6270, 1.2741, 1.0054],
        [1.3940, 1.0466, 1.1304],
        [1.5308, 1.6190, 1.2075],
        [1.7753,... 1.2067],
        [1.0556, 1.6129, 1.8009],
        [1.1916, 1.5847, 1.4399]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.5157, 0.8907, 0.9111],
        [0.9636, 0.5470, 0.5099],
        [0.8135, 0.6030, 0.6263],
        [0.8674,...953],
        [0.7117, 0.7941, 0.5017],
        [0.5431, 0.5893, 0.5903]], dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([100, 3])
SDDMM(message func: v_div_e)
_____________________ test_sddmm[idtype0-div-e-v-shp3-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'v', rhs_target = 'e', msg = 'div'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72b806318>
op = 'mul'
lhs_data = tensor([[1.2904, 1.2141, 1.4320],
        [1.0562, 1.1468, 1.7100],
        [1.4894, 1.6175, 1.0858],
        [1.4676,... 1.8934],
        [1.4175, 1.9799, 1.0570],
        [1.4635, 1.7422, 1.0403]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.7132, 0.9347, 0.5530],
        [0.6678, 0.9866, 0.6251],
        [0.7381, 0.5473, 0.6246],
        [0.5041,...323],
        [0.7625, 0.7837, 0.5197],
        [0.5911, 0.5100, 0.6218]], dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([40, 3]), rhs shape: torch.Size([300, 3])
SDDMM(message func: v_div_e)
_____________________ test_sddmm[idtype0-div-e-v-shp4-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'v', rhs_target = 'e', msg = 'div'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb61cfa4f30>
op = 'mul'
lhs_data = tensor([[1.7245],
        [1.3808],
        [1.8198],
        [1.3664],
        [1.1577],
        [1.6889],
        [1...737],
        [1.5811],
        [1.8362],
        [1.8943],
        [1.2729]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.6221],
        [0.6365],
        [0.6702],
        [0.6034],
        [0.5004],
        [0.9095],
        [0...,
        [0.6408],
        [0.8868],
        [0.5499],
        [0.9686]], dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([100, 1])
SDDMM(message func: v_div_e)
_____________________ test_sddmm[idtype0-div-e-v-shp4-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'v', rhs_target = 'e', msg = 'div'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72c0f9288>
op = 'mul'
lhs_data = tensor([[1.9236],
        [1.0332],
        [1.6238],
        [1.6577],
        [1.8218],
        [1.0086],
        [1...247],
        [1.1291],
        [1.2933],
        [1.8250],
        [1.3890]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.9922],
        [0.6440],
        [0.8057],
        [0.8979],
        [0.5470],
        [0.7213],
        [0...,
        [0.7244],
        [0.6782],
        [0.5619],
        [0.5717]], dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([40, 1]), rhs shape: torch.Size([300, 1])
SDDMM(message func: v_div_e)
_____________________ test_sddmm[idtype0-dot-u-v-shp0-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': ...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'v', rhs_target = 'u'
msg = 'dot', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72ab8fa68>
op = 'dot'
lhs_data = tensor([[[[[[1.7278],
            [1.1985],
            [1.4547]]],


          [[[1.1806],
            [1.9264],
    ...2]]],


          [[[1.7886],
            [1.4008],
            [1.3732]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.3155]],

           [[1.1620]],

           [[1.8878]]]],



         [[[[1.9444]],

           [[1.459...



         [[[[1.9426]],

           [[1.5915]],

           [[1.5745]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: v_dot_u)
_____________________ test_sddmm[idtype0-dot-u-v-shp0-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'v', rhs_target = 'u'
msg = 'dot', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72c0f4b40>
op = 'dot'
lhs_data = tensor([[[[[[1.8450],
            [1.9472],
            [1.7472]]],


          [[[1.0065],
            [1.5989],
    ...2]]],


          [[[1.5534],
            [1.3135],
            [1.5327]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.5235]],

           [[1.7083]],

           [[1.2252]]]],



         [[[[1.4583]],

           [[1.368...



         [[[[1.9812]],

           [[1.7117]],

           [[1.7927]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([40, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: v_dot_u)
_____________________ test_sddmm[idtype0-dot-u-v-shp1-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Sch...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'v', rhs_target = 'u'
msg = 'dot', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72c0641f8>
op = 'dot'
lhs_data = tensor([[[[[1.3269, 1.9457, 1.6463,  ..., 1.1097, 1.8729, 1.4492]],

          [[1.7134, 1.6167, 1.2678,  ..., 1.1323,...
          [[1.9678, 1.7023, 1.3024,  ..., 1.3135, 1.4632, 1.6250]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.6288, 1.7727, 1.0406,  ..., 1.7703, 1.9869, 1.7428],
           [1.9916, 1.7811, 1.7144,  ..., 1.7314, 1...
           [1.0223, 1.5172, 1.2116,  ..., 1.5314, 1.9098, 1.6673]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: v_dot_u)
_____________________ test_sddmm[idtype0-dot-u-v-shp1-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'v', rhs_target = 'u'
msg = 'dot', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72c0e80d8>
op = 'dot'
lhs_data = tensor([[[[[1.7104, 1.0629, 1.4039,  ..., 1.6261, 1.9966, 1.0484]],

          [[1.1519, 1.9346, 1.9526,  ..., 1.0984,...
          [[1.7235, 1.6870, 1.7802,  ..., 1.6384, 1.5729, 1.7144]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.9308, 1.3973, 1.7323,  ..., 1.5114, 1.1363, 1.0586],
           [1.0815, 1.1904, 1.0445,  ..., 1.9064, 1...
           [1.4518, 1.7539, 1.6333,  ..., 1.1671, 1.0651, 1.6037]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([40, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: v_dot_u)
_____________________ test_sddmm[idtype0-dot-u-v-shp2-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'v', rhs_target = 'u', msg = 'dot'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72aaa0ab0>
op = 'dot'
lhs_data = tensor([[[[1.9980, 1.9698, 1.4218],
          [1.3878, 1.5999, 1.7772],
          [1.9821, 1.2891, 1.2214]]],


      ...0],
          [1.8265, 1.6840, 1.0647],
          [1.6200, 1.0238, 1.2511]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.0800, 1.2335, 1.5837]],

         [[1.6338, 1.1110, 1.5362]],

         [[1.5167, 1.9957, 1.8914]],

    ...

         [[1.6116, 1.7307, 1.1742]],

         [[1.4410, 1.6614, 1.2691]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: v_dot_u)
_____________________ test_sddmm[idtype0-dot-u-v-shp2-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'v', rhs_target = 'u', msg = 'dot'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb706246750>
op = 'dot'
lhs_data = tensor([[[[1.0958, 1.9950, 1.7911],
          [1.8908, 1.5230, 1.5956],
          [1.0566, 1.2962, 1.2489]]],


      ...9],
          [1.6994, 1.5538, 1.9586],
          [1.6110, 1.1991, 1.4193]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.7085, 1.1098, 1.4944]],

         [[1.2858, 1.0117, 1.4911]],

         [[1.3249, 1.4092, 1.4915]],

    ...

         [[1.8935, 1.8007, 1.2951]],

         [[1.1115, 1.7013, 1.6205]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([40, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: v_dot_u)
_____________________ test_sddmm[idtype0-dot-u-v-shp3-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'v', rhs_target = 'u', msg = 'dot'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb61f174120>
op = 'dot'
lhs_data = tensor([[1.8094, 1.2590, 1.5608],
        [1.6526, 1.0523, 1.6153],
        [1.2969, 1.4602, 1.7874],
        [1.2094,... 1.7101],
        [1.7920, 1.1720, 1.2337],
        [1.1319, 1.8006, 1.1801]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.5378, 1.6713, 1.8225],
        [1.9674, 1.5754, 1.6601],
        [1.7436, 1.4322, 1.8114],
        [1.9136,... 1.9137],
        [1.1383, 1.4573, 1.3215],
        [1.8538, 1.1787, 1.6324]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: v_dot_u)
_____________________ test_sddmm[idtype0-dot-u-v-shp3-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'v', rhs_target = 'u', msg = 'dot'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb6146dcc18>
op = 'dot'
lhs_data = tensor([[1.6962, 1.2493, 1.4998],
        [1.0568, 1.4107, 1.7544],
        [1.4615, 1.8419, 1.3595],
        [1.7149,... 1.5138],
        [1.8457, 1.2552, 1.4752],
        [1.6774, 1.2239, 1.6404]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.3354, 1.5995, 1.3639],
        [1.0890, 1.3111, 1.7708],
        [1.9060, 1.8498, 1.9762],
        [1.9665,... 1.6832],
        [1.4952, 1.5407, 1.9573],
        [1.7965, 1.8944, 1.0314]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([40, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: v_dot_u)
_____________________ test_sddmm[idtype0-dot-u-v-shp4-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'v', rhs_target = 'u', msg = 'dot'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb61f1591f8>
op = 'dot'
lhs_data = tensor([[1.3889],
        [1.3115],
        [1.9554],
        [1.2426],
        [1.4425],
        [1.5445],
        [1...592],
        [1.5307],
        [1.5875],
        [1.4478],
        [1.1645]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.3573],
        [1.3823],
        [1.8837],
        [1.7541],
        [1.5926],
        [1.3008],
        [1...891],
        [1.0147],
        [1.8930],
        [1.4756],
        [1.2862]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: v_dot_u)
_____________________ test_sddmm[idtype0-dot-u-v-shp4-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'v', rhs_target = 'u', msg = 'dot'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72bd44e58>
op = 'dot'
lhs_data = tensor([[1.0768],
        [1.5686],
        [1.9474],
        [1.8881],
        [1.2692],
        [1.6137],
        [1...851],
        [1.3342],
        [1.7911],
        [1.1184],
        [1.0091]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.3113],
        [1.6323],
        [1.6015],
        [1.9089],
        [1.3061],
        [1.5841],
        [1...809],
        [1.5951],
        [1.3903],
        [1.6520],
        [1.5876]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([40, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: v_dot_u)
_____________________ test_sddmm[idtype0-dot-u-e-shp0-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...torch.float64), 'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'e', rhs_target = 'u'
msg = 'dot', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72c0d17e0>
op = 'dot'
lhs_data = tensor([[[[[[1.3424],
            [1.1974],
            [1.0986]]],


          [[[1.7665],
            [1.9831],
    ...2]]],


          [[[1.3818],
            [1.2792],
            [1.8089]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.8950]],

           [[1.8161]],

           [[1.3831]]]],



         [[[[1.6360]],

           [[1.508...



         [[[[1.3446]],

           [[1.8808]],

           [[1.1389]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([100, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: e_dot_u)
_____________________ test_sddmm[idtype0-dot-u-e-shp0-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'e', rhs_target = 'u'
msg = 'dot', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72c0d1c60>
op = 'dot'
lhs_data = tensor([[[[[[1.5573],
            [1.5611],
            [1.0516]]],


          [[[1.3478],
            [1.5715],
    ...1]]],


          [[[1.1002],
            [1.0944],
            [1.7895]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.5572]],

           [[1.0034]],

           [[1.6113]]]],



         [[[[1.2217]],

           [[1.441...



         [[[[1.8581]],

           [[1.0069]],

           [[1.6291]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([300, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: e_dot_u)
_____________________ test_sddmm[idtype0-dot-u-e-shp1-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...pe=torch.float64), 'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'e', rhs_target = 'u'
msg = 'dot', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43e859fc0>
op = 'dot'
lhs_data = tensor([[[[[1.7458, 1.2057, 1.5004,  ..., 1.6580, 1.7712, 1.7219]],

          [[1.6799, 1.1269, 1.5448,  ..., 1.5339,...
          [[1.4828, 1.1068, 1.5707,  ..., 1.4584, 1.6281, 1.6505]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.3122, 1.4337, 1.9610,  ..., 1.2124, 1.1923, 1.3032],
           [1.8750, 1.1824, 1.4674,  ..., 1.6545, 1...
           [1.9398, 1.4387, 1.9524,  ..., 1.9143, 1.9939, 1.3241]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([100, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: e_dot_u)
_____________________ test_sddmm[idtype0-dot-u-e-shp1-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'e', rhs_target = 'u'
msg = 'dot', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb6146e4750>
op = 'dot'
lhs_data = tensor([[[[[1.9985, 1.8489, 1.9636,  ..., 1.3808, 1.4074, 1.1565]],

          [[1.0604, 1.7185, 1.6332,  ..., 1.5370,...
          [[1.0872, 1.9419, 1.2708,  ..., 1.4661, 1.7932, 1.1252]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.8900, 1.7806, 1.5279,  ..., 1.4191, 1.6736, 1.5135],
           [1.8508, 1.6566, 1.9525,  ..., 1.2742, 1...
           [1.8302, 1.5017, 1.4871,  ..., 1.4715, 1.2909, 1.0376]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([300, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: e_dot_u)
_____________________ test_sddmm[idtype0-dot-u-e-shp2-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...dtype=torch.float64), 'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'e', rhs_target = 'u', msg = 'dot'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb7060dd1b0>
op = 'dot'
lhs_data = tensor([[[[1.3420, 1.1752, 1.1662],
          [1.7904, 1.6750, 1.5941],
          [1.6743, 1.4719, 1.0267]]],


      ...9],
          [1.6820, 1.6346, 1.5777],
          [1.3533, 1.2059, 1.1186]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.7790, 1.0603, 1.7372]],

         [[1.1621, 1.5469, 1.5287]],

         [[1.3225, 1.1023, 1.9930]],

    ...

         [[1.9960, 1.3793, 1.4153]],

         [[1.7340, 1.0550, 1.8098]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([100, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: e_dot_u)
_____________________ test_sddmm[idtype0-dot-u-e-shp2-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'e', rhs_target = 'u', msg = 'dot'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb4476b1c60>
op = 'dot'
lhs_data = tensor([[[[1.4788, 1.8808, 1.1482],
          [1.1905, 1.3476, 1.1404],
          [1.7147, 1.5428, 1.4756]]],


      ...7],
          [1.8597, 1.5397, 1.7700],
          [1.1870, 1.7201, 1.0745]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.2938, 1.9230, 1.1467]],

         [[1.3676, 1.7013, 1.5962]],

         [[1.6257, 1.1120, 1.7499]],

    ...

         [[1.0245, 1.5601, 1.3486]],

         [[1.9226, 1.2918, 1.4364]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([300, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: e_dot_u)
_____________________ test_sddmm[idtype0-dot-u-e-shp3-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'e', rhs_target = 'u', msg = 'dot'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72c0fa828>
op = 'dot'
lhs_data = tensor([[1.3515, 1.4390, 1.2811],
        [1.2326, 1.5047, 1.4297],
        [1.1645, 1.3024, 1.3400],
        [1.2793,... 1.4568],
        [1.3032, 1.1475, 1.4317],
        [1.4033, 1.3853, 1.9879]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.2148, 1.0063, 1.0089],
        [1.8091, 1.7303, 1.0463],
        [1.7356, 1.7635, 1.4693],
        [1.7969,... 1.1498],
        [1.2894, 1.0025, 1.1320],
        [1.6561, 1.2177, 1.5457]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([100, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: e_dot_u)
_____________________ test_sddmm[idtype0-dot-u-e-shp3-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'e', rhs_target = 'u', msg = 'dot'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72c0faee8>
op = 'dot'
lhs_data = tensor([[1.3429, 1.6739, 1.4585],
        [1.9876, 1.3148, 1.3027],
        [1.7958, 1.8014, 1.6535],
        [1.8406,... 1.0911],
        [1.3882, 1.8029, 1.7549],
        [1.3162, 1.6224, 1.3530]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.4360, 1.2884, 1.0850],
        [1.1064, 1.6056, 1.5823],
        [1.2088, 1.1523, 1.6573],
        [1.6138,... 1.5895],
        [1.5702, 1.9729, 1.7977],
        [1.0026, 1.6325, 1.0583]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([300, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: e_dot_u)
_____________________ test_sddmm[idtype0-dot-u-e-shp4-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'e', rhs_target = 'u', msg = 'dot'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72c0fa438>
op = 'dot'
lhs_data = tensor([[1.3873],
        [1.5865],
        [1.4412],
        [1.1003],
        [1.8325],
        [1.9172],
        [1...667],
        [1.5525],
        [1.3188],
        [1.1146],
        [1.3527]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.4984],
        [1.6425],
        [1.9342],
        [1.4939],
        [1.2566],
        [1.9242],
        [1...853],
        [1.0968],
        [1.6101],
        [1.0907],
        [1.4630]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([100, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: e_dot_u)
_____________________ test_sddmm[idtype0-dot-u-e-shp4-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'e', rhs_target = 'u', msg = 'dot'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72c0fa798>
op = 'dot'
lhs_data = tensor([[1.4265],
        [1.7735],
        [1.8362],
        [1.5035],
        [1.7469],
        [1.0346],
        [1...356],
        [1.7983],
        [1.1107],
        [1.2246],
        [1.9512]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.4933],
        [1.5307],
        [1.1377],
        [1.1852],
        [1.1667],
        [1.1581],
        [1...415],
        [1.1753],
        [1.0798],
        [1.8934],
        [1.9413]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([300, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: e_dot_u)
_____________________ test_sddmm[idtype0-dot-v-u-shp0-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': ...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'u', rhs_target = 'v'
msg = 'dot', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb61cfa8c18>
op = 'dot'
lhs_data = tensor([[[[[[1.8109],
            [1.3848],
            [1.3433]]],


          [[[1.1955],
            [1.7171],
    ...4]]],


          [[[1.2934],
            [1.1054],
            [1.7330]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.1829]],

           [[1.0825]],

           [[1.1472]]]],



         [[[[1.4529]],

           [[1.543...



         [[[[1.2161]],

           [[1.4200]],

           [[1.4233]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: u_dot_v)
_____________________ test_sddmm[idtype0-dot-v-u-shp0-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'u', rhs_target = 'v'
msg = 'dot', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72bd448b8>
op = 'dot'
lhs_data = tensor([[[[[[1.0744],
            [1.6940],
            [1.5025]]],


          [[[1.6765],
            [1.0223],
    ...0]]],


          [[[1.0804],
            [1.5469],
            [1.4333]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.2273]],

           [[1.0022]],

           [[1.0556]]]],



         [[[[1.5493]],

           [[1.694...



         [[[[1.8804]],

           [[1.8516]],

           [[1.6190]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([40, 4, 1, 3, 1, 1])
SDDMM(message func: u_dot_v)
_____________________ test_sddmm[idtype0-dot-v-u-shp1-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Sch...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'u', rhs_target = 'v'
msg = 'dot', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72c0fd090>
op = 'dot'
lhs_data = tensor([[[[[1.0196, 1.5999, 1.5410,  ..., 1.0230, 1.4280, 1.4549]],

          [[1.2693, 1.3840, 1.0877,  ..., 1.0765,...
          [[1.3520, 1.3783, 1.1334,  ..., 1.7427, 1.9147, 1.8652]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.0574, 1.1435, 1.7739,  ..., 1.8774, 1.2497, 1.4348],
           [1.2355, 1.5981, 1.1415,  ..., 1.6352, 1...
           [1.8673, 1.3996, 1.2863,  ..., 1.9306, 1.0224, 1.1401]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: u_dot_v)
_____________________ test_sddmm[idtype0-dot-v-u-shp1-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'u', rhs_target = 'v'
msg = 'dot', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72c0fd5e8>
op = 'dot'
lhs_data = tensor([[[[[1.7391, 1.9368, 1.5089,  ..., 1.3556, 1.8797, 1.0403]],

          [[1.2135, 1.3251, 1.4423,  ..., 1.0180,...
          [[1.7048, 1.4052, 1.3870,  ..., 1.2703, 1.8691, 1.3775]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.8110, 1.2446, 1.7583,  ..., 1.2364, 1.9474, 1.6141],
           [1.1459, 1.2360, 1.8159,  ..., 1.0437, 1...
           [1.2712, 1.3398, 1.7647,  ..., 1.7920, 1.2481, 1.3812]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([40, 1, 3, 7, 7])
SDDMM(message func: u_dot_v)
_____________________ test_sddmm[idtype0-dot-v-u-shp2-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'u', rhs_target = 'v', msg = 'dot'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb61f174c18>
op = 'dot'
lhs_data = tensor([[[[1.5451, 1.1690, 1.6552],
          [1.4909, 1.5211, 1.0802],
          [1.4443, 1.3110, 1.4815]]],


      ...4],
          [1.5925, 1.7659, 1.6498],
          [1.9981, 1.0332, 1.0156]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.9011, 1.8685, 1.7323]],

         [[1.9146, 1.2509, 1.8612]],

         [[1.4872, 1.3564, 1.6758]],

    ...

         [[1.0811, 1.0718, 1.0699]],

         [[1.1585, 1.6600, 1.7030]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: u_dot_v)
_____________________ test_sddmm[idtype0-dot-v-u-shp2-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'u', rhs_target = 'v', msg = 'dot'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72c0c93f0>
op = 'dot'
lhs_data = tensor([[[[1.3306, 1.0925, 1.3724],
          [1.3995, 1.7009, 1.3975],
          [1.8031, 1.9298, 1.5359]]],


      ...3],
          [1.8219, 1.0953, 1.2929],
          [1.6640, 1.4438, 1.5922]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.3954, 1.9743, 1.3307]],

         [[1.8522, 1.5207, 1.0362]],

         [[1.4170, 1.1948, 1.3659]],

    ...

         [[1.9476, 1.1282, 1.5113]],

         [[1.3597, 1.1992, 1.5282]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([40, 4, 1, 3])
SDDMM(message func: u_dot_v)
_____________________ test_sddmm[idtype0-dot-v-u-shp3-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'u', rhs_target = 'v', msg = 'dot'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb4476b4948>
op = 'dot'
lhs_data = tensor([[1.5862, 1.0177, 1.0326],
        [1.1628, 1.6147, 1.7862],
        [1.8375, 1.0922, 1.9771],
        [1.7571,... 1.5679],
        [1.4846, 1.0449, 1.7485],
        [1.7227, 1.6293, 1.2641]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.5624, 1.5971, 1.2074],
        [1.9373, 1.3221, 1.5196],
        [1.0126, 1.4923, 1.6577],
        [1.0095,... 1.0373],
        [1.7764, 1.4500, 1.2482],
        [1.9403, 1.6589, 1.6614]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: u_dot_v)
_____________________ test_sddmm[idtype0-dot-v-u-shp3-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'u', rhs_target = 'v', msg = 'dot'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb61cfa2af8>
op = 'dot'
lhs_data = tensor([[1.2632, 1.7337, 1.4518],
        [1.2237, 1.4936, 1.7172],
        [1.0458, 1.9375, 1.3490],
        [1.7155,... 1.6768],
        [1.3613, 1.6168, 1.5686],
        [1.9657, 1.2841, 1.3926]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.2517, 1.2217, 1.1005],
        [1.3074, 1.5364, 1.2411],
        [1.3278, 1.9146, 1.6942],
        [1.1741,... 1.5831],
        [1.0867, 1.5465, 1.4849],
        [1.5101, 1.8180, 1.7073]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([40, 3])
SDDMM(message func: u_dot_v)
_____________________ test_sddmm[idtype0-dot-v-u-shp4-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'u', rhs_target = 'v', msg = 'dot'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72c0dd120>
op = 'dot'
lhs_data = tensor([[1.0286],
        [1.9090],
        [1.1628],
        [1.1690],
        [1.6778],
        [1.9784],
        [1...649],
        [1.5639],
        [1.5329],
        [1.3372],
        [1.7075]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.3884],
        [1.3421],
        [1.2618],
        [1.8970],
        [1.5341],
        [1.5839],
        [1...955],
        [1.1700],
        [1.4103],
        [1.5564],
        [1.5462]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: u_dot_v)
_____________________ test_sddmm[idtype0-dot-v-u-shp4-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'u', rhs_target = 'v', msg = 'dot'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72c0dd318>
op = 'dot'
lhs_data = tensor([[1.7271],
        [1.3178],
        [1.7741],
        [1.3383],
        [1.7430],
        [1.3627],
        [1...814],
        [1.6940],
        [1.7332],
        [1.0119],
        [1.1733]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.4248],
        [1.0926],
        [1.1182],
        [1.0449],
        [1.2632],
        [1.4493],
        [1...714],
        [1.2177],
        [1.6047],
        [1.7722],
        [1.5263]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([40, 1])
SDDMM(message func: u_dot_v)
_____________________ test_sddmm[idtype0-dot-v-e-shp0-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...torch.float64), 'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'e', rhs_target = 'v'
msg = 'dot', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb6146e5750>
op = 'dot'
lhs_data = tensor([[[[[[1.9488],
            [1.9768],
            [1.8255]]],


          [[[1.3562],
            [1.0053],
    ...8]]],


          [[[1.9845],
            [1.3404],
            [1.0893]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.3428]],

           [[1.8871]],

           [[1.9130]]]],



         [[[[1.5133]],

           [[1.239...



         [[[[1.8175]],

           [[1.8037]],

           [[1.2324]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([100, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: e_dot_v)
_____________________ test_sddmm[idtype0-dot-v-e-shp0-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'e', rhs_target = 'v'
msg = 'dot', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72be224c8>
op = 'dot'
lhs_data = tensor([[[[[[1.2498],
            [1.2921],
            [1.0754]]],


          [[[1.9804],
            [1.9950],
    ...6]]],


          [[[1.3106],
            [1.5043],
            [1.2632]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.0263]],

           [[1.9990]],

           [[1.2802]]]],



         [[[[1.5246]],

           [[1.446...



         [[[[1.6736]],

           [[1.7816]],

           [[1.9877]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([300, 1, 2, 1, 3, 1]), rhs shape: torch.Size([40, 4, 1, 3, 1, 1])
SDDMM(message func: e_dot_v)
_____________________ test_sddmm[idtype0-dot-v-e-shp1-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...pe=torch.float64), 'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'e', rhs_target = 'v'
msg = 'dot', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72c052cf0>
op = 'dot'
lhs_data = tensor([[[[[1.7327, 1.9326, 1.5172,  ..., 1.5516, 1.3688, 1.6956]],

          [[1.9141, 1.3900, 1.9849,  ..., 1.6035,...
          [[1.3084, 1.7472, 1.3623,  ..., 1.3107, 1.7680, 1.1735]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.7594, 1.8808, 1.2798,  ..., 1.1024, 1.9745, 1.7784],
           [1.4895, 1.9859, 1.2314,  ..., 1.6507, 1...
           [1.9636, 1.5904, 1.3618,  ..., 1.9647, 1.3271, 1.7885]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([100, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: e_dot_v)
_____________________ test_sddmm[idtype0-dot-v-e-shp1-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'e', rhs_target = 'v'
msg = 'dot', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb705e270d8>
op = 'dot'
lhs_data = tensor([[[[[1.0037, 1.3297, 1.7971,  ..., 1.6631, 1.8631, 1.1139]],

          [[1.5094, 1.5169, 1.1299,  ..., 1.3116,...
          [[1.8861, 1.5064, 1.6445,  ..., 1.5289, 1.8510, 1.5506]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.7431, 1.2575, 1.3621,  ..., 1.4314, 1.3507, 1.3479],
           [1.1549, 1.0357, 1.0044,  ..., 1.8730, 1...
           [1.2416, 1.8925, 1.5381,  ..., 1.0700, 1.6217, 1.1910]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([300, 5, 3, 1, 7]), rhs shape: torch.Size([40, 1, 3, 7, 7])
SDDMM(message func: e_dot_v)
_____________________ test_sddmm[idtype0-dot-v-e-shp2-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...dtype=torch.float64), 'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'e', rhs_target = 'v', msg = 'dot'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72949eaf8>
op = 'dot'
lhs_data = tensor([[[[1.1999, 1.4421, 1.3385],
          [1.4866, 1.9242, 1.6945],
          [1.3434, 1.1979, 1.0174]]],


      ...6],
          [1.0286, 1.4423, 1.3252],
          [1.0747, 1.8696, 1.9083]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.1038, 1.4274, 1.5728]],

         [[1.7560, 1.8833, 1.5776]],

         [[1.2007, 1.4668, 1.7539]],

    ...

         [[1.4502, 1.6168, 1.9132]],

         [[1.4319, 1.5830, 1.1139]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([100, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: e_dot_v)
_____________________ test_sddmm[idtype0-dot-v-e-shp2-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'e', rhs_target = 'v', msg = 'dot'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb457b61ab0>
op = 'dot'
lhs_data = tensor([[[[1.1193, 1.7977, 1.7612],
          [1.3379, 1.5550, 1.8287],
          [1.7177, 1.7294, 1.6707]]],


      ...7],
          [1.8527, 1.4426, 1.1987],
          [1.5353, 1.1975, 1.5085]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.1456, 1.4849, 1.2408]],

         [[1.5894, 1.6952, 1.7773]],

         [[1.9092, 1.5718, 1.2226]],

    ...

         [[1.7487, 1.9186, 1.2503]],

         [[1.2604, 1.5374, 1.1958]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([300, 1, 3, 3]), rhs shape: torch.Size([40, 4, 1, 3])
SDDMM(message func: e_dot_v)
_____________________ test_sddmm[idtype0-dot-v-e-shp3-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'e', rhs_target = 'v', msg = 'dot'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43e849630>
op = 'dot'
lhs_data = tensor([[1.0948, 1.3034, 1.1468],
        [1.3004, 1.1811, 1.8804],
        [1.2389, 1.6896, 1.0807],
        [1.4372,... 1.1261],
        [1.2962, 1.4877, 1.2900],
        [1.4576, 1.2524, 1.6044]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.8235, 1.6558, 1.8373],
        [1.1642, 1.4153, 1.9821],
        [1.3554, 1.2323, 1.1133],
        [1.7685,... 1.2692],
        [1.8040, 1.6894, 1.3214],
        [1.4725, 1.4926, 1.4970]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([100, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: e_dot_v)
_____________________ test_sddmm[idtype0-dot-v-e-shp3-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'e', rhs_target = 'v', msg = 'dot'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72be1d1f8>
op = 'dot'
lhs_data = tensor([[1.9848, 1.6372, 1.8036],
        [1.5857, 1.8258, 1.9176],
        [1.0746, 1.5040, 1.3337],
        [1.5147,... 1.6223],
        [1.3714, 1.9019, 1.7524],
        [1.8473, 1.4799, 1.0451]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.7553, 1.7868, 1.2566],
        [1.5894, 1.5958, 1.2966],
        [1.8389, 1.7501, 1.0210],
        [1.7327,... 1.3532],
        [1.1860, 1.1258, 1.3869],
        [1.1217, 1.5307, 1.3850]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([300, 3]), rhs shape: torch.Size([40, 3])
SDDMM(message func: e_dot_v)
_____________________ test_sddmm[idtype0-dot-v-e-shp4-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'e', rhs_target = 'v', msg = 'dot'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72c05ea68>
op = 'dot'
lhs_data = tensor([[1.3833],
        [1.8132],
        [1.5787],
        [1.7619],
        [1.8556],
        [1.6800],
        [1...146],
        [1.9682],
        [1.9957],
        [1.1647],
        [1.9897]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.4136],
        [1.9394],
        [1.5522],
        [1.9671],
        [1.4397],
        [1.2569],
        [1...591],
        [1.6123],
        [1.8394],
        [1.8628],
        [1.9947]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([100, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: e_dot_v)
_____________________ test_sddmm[idtype0-dot-v-e-shp4-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'e', rhs_target = 'v', msg = 'dot'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb70610b438>
op = 'dot'
lhs_data = tensor([[1.4366],
        [1.5863],
        [1.0533],
        [1.0964],
        [1.6898],
        [1.1214],
        [1...559],
        [1.9026],
        [1.0101],
        [1.5271],
        [1.7303]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.8719],
        [1.8686],
        [1.0219],
        [1.0043],
        [1.0426],
        [1.8554],
        [1...961],
        [1.3028],
        [1.3177],
        [1.2702],
        [1.4461]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([300, 1]), rhs shape: torch.Size([40, 1])
SDDMM(message func: e_dot_v)
_____________________ test_sddmm[idtype0-dot-e-u-shp0-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': ...torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'u', rhs_target = 'e'
msg = 'dot', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72ab8f0d8>
op = 'dot'
lhs_data = tensor([[[[[[1.3497],
            [1.0358],
            [1.3673]]],


          [[[1.0300],
            [1.3829],
    ...5]]],


          [[[1.6346],
            [1.6897],
            [1.3109]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.8032]],

           [[1.2017]],

           [[1.7373]]]],



         [[[[1.6433]],

           [[1.833...



         [[[[1.3125]],

           [[1.4887]],

           [[1.0799]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([100, 4, 1, 3, 1, 1])
SDDMM(message func: u_dot_e)
_____________________ test_sddmm[idtype0-dot-e-u-shp0-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'u', rhs_target = 'e'
msg = 'dot', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72c06d828>
op = 'dot'
lhs_data = tensor([[[[[[1.0037],
            [1.3167],
            [1.6415]]],


          [[[1.1616],
            [1.2638],
    ...0]]],


          [[[1.0044],
            [1.9117],
            [1.3147]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.4892]],

           [[1.0454]],

           [[1.8983]]]],



         [[[[1.8547]],

           [[1.168...



         [[[[1.9343]],

           [[1.2524]],

           [[1.5178]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([300, 4, 1, 3, 1, 1])
SDDMM(message func: u_dot_e)
_____________________ test_sddmm[idtype0-dot-e-u-shp1-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Sch...pe=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'u', rhs_target = 'e'
msg = 'dot', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb447687f78>
op = 'dot'
lhs_data = tensor([[[[[1.8857, 1.2635, 1.2527,  ..., 1.3352, 1.4465, 1.6037]],

          [[1.5398, 1.3520, 1.4138,  ..., 1.1585,...
          [[1.5978, 1.3087, 1.0873,  ..., 1.5006, 1.4738, 1.5287]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.8324, 1.7187, 1.0281,  ..., 1.1795, 1.6230, 1.7044],
           [1.9014, 1.2540, 1.7843,  ..., 1.4041, 1...
           [1.5253, 1.4248, 1.1731,  ..., 1.2509, 1.2427, 1.0446]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([100, 1, 3, 7, 7])
SDDMM(message func: u_dot_e)
_____________________ test_sddmm[idtype0-dot-e-u-shp1-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'u', rhs_target = 'e'
msg = 'dot', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72be3b120>
op = 'dot'
lhs_data = tensor([[[[[1.7978, 1.9728, 1.3555,  ..., 1.5504, 1.2930, 1.3085]],

          [[1.3872, 1.1795, 1.9597,  ..., 1.0021,...
          [[1.5571, 1.5482, 1.5802,  ..., 1.8416, 1.7093, 1.8052]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.0112, 1.8115, 1.7208,  ..., 1.4334, 1.3071, 1.3689],
           [1.3941, 1.4290, 1.9038,  ..., 1.3913, 1...
           [1.4471, 1.7047, 1.3690,  ..., 1.4293, 1.8871, 1.8069]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([300, 1, 3, 7, 7])
SDDMM(message func: u_dot_e)
_____________________ test_sddmm[idtype0-dot-e-u-shp2-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme...dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'u', rhs_target = 'e', msg = 'dot'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72be3b3f0>
op = 'dot'
lhs_data = tensor([[[[1.6828, 1.0115, 1.3443],
          [1.5374, 1.7546, 1.6232],
          [1.8566, 1.7754, 1.9381]]],


      ...0],
          [1.7184, 1.0384, 1.8327],
          [1.4218, 1.7238, 1.2853]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.8228, 1.5467, 1.1307]],

         [[1.9668, 1.2751, 1.2712]],

         [[1.2554, 1.3990, 1.3126]],

    ...

         [[1.8016, 1.5962, 1.1440]],

         [[1.9392, 1.6864, 1.0653]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([100, 4, 1, 3])
SDDMM(message func: u_dot_e)
_____________________ test_sddmm[idtype0-dot-e-u-shp2-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'u', rhs_target = 'e', msg = 'dot'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72be3b678>
op = 'dot'
lhs_data = tensor([[[[1.2359, 1.3371, 1.0879],
          [1.7087, 1.0806, 1.2024],
          [1.6166, 1.4858, 1.6442]]],


      ...4],
          [1.3460, 1.6309, 1.4314],
          [1.9948, 1.1113, 1.3897]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.0672, 1.6130, 1.1200]],

         [[1.8595, 1.6259, 1.5739]],

         [[1.3548, 1.7564, 1.2176]],

    ...

         [[1.2075, 1.1170, 1.1196]],

         [[1.1035, 1.1918, 1.7411]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([300, 4, 1, 3])
SDDMM(message func: u_dot_e)
_____________________ test_sddmm[idtype0-dot-e-u-shp3-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'u', rhs_target = 'e', msg = 'dot'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43e873048>
op = 'dot'
lhs_data = tensor([[1.2835, 1.4809, 1.5264],
        [1.3371, 1.8595, 1.5893],
        [1.0765, 1.9563, 1.2598],
        [1.8766,... 1.9698],
        [1.9266, 1.3125, 1.9791],
        [1.2007, 1.0085, 1.2449]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.0456, 1.0221, 1.4322],
        [1.2929, 1.1925, 1.0634],
        [1.1630, 1.8086, 1.6596],
        [1.1510,... 1.1729],
        [1.8895, 1.9106, 1.4158],
        [1.1063, 1.7025, 1.3849]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([100, 3])
SDDMM(message func: u_dot_e)
_____________________ test_sddmm[idtype0-dot-e-u-shp3-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'u', rhs_target = 'e', msg = 'dot'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb44765e318>
op = 'dot'
lhs_data = tensor([[1.3024, 1.4333, 1.2108],
        [1.4319, 1.5613, 1.3974],
        [1.5019, 1.2707, 1.7073],
        [1.0162,... 1.5543],
        [1.5352, 1.9108, 1.1811],
        [1.9080, 1.8532, 1.2630]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.1151, 1.8777, 1.3367],
        [1.8807, 1.0349, 1.1493],
        [1.1105, 1.5979, 1.6812],
        [1.1899,... 1.0426],
        [1.1841, 1.0641, 1.1777],
        [1.5110, 1.5020, 1.0812]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([300, 3])
SDDMM(message func: u_dot_e)
_____________________ test_sddmm[idtype0-dot-e-u-shp4-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'u', rhs_target = 'e', msg = 'dot'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb705b51a20>
op = 'dot'
lhs_data = tensor([[1.3374],
        [1.5961],
        [1.5253],
        [1.2463],
        [1.7980],
        [1.5652],
        [1...807],
        [1.9599],
        [1.6073],
        [1.0407],
        [1.4675]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.1983],
        [1.9196],
        [1.5056],
        [1.1994],
        [1.9450],
        [1.1486],
        [1...584],
        [1.8539],
        [1.5825],
        [1.1240],
        [1.0340]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([100, 1])
SDDMM(message func: u_dot_e)
_____________________ test_sddmm[idtype0-dot-e-u-shp4-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'u', rhs_target = 'e', msg = 'dot'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72be04090>
op = 'dot'
lhs_data = tensor([[1.6480],
        [1.9762],
        [1.0556],
        [1.8116],
        [1.7008],
        [1.4151],
        [1...874],
        [1.2959],
        [1.6804],
        [1.4906],
        [1.1482]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.2825],
        [1.5755],
        [1.6002],
        [1.3573],
        [1.6168],
        [1.9283],
        [1...037],
        [1.6585],
        [1.6862],
        [1.2711],
        [1.7556]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([300, 1])
SDDMM(message func: u_dot_e)
_____________________ test_sddmm[idtype0-dot-e-v-shp0-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': ...torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'v', rhs_target = 'e'
msg = 'dot', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72be04360>
op = 'dot'
lhs_data = tensor([[[[[[1.6965],
            [1.7574],
            [1.0968]]],


          [[[1.5211],
            [1.0152],
    ...3]]],


          [[[1.2442],
            [1.2443],
            [1.5455]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.7027]],

           [[1.1727]],

           [[1.1275]]]],



         [[[[1.7270]],

           [[1.824...



         [[[[1.8747]],

           [[1.0336]],

           [[1.2489]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([100, 4, 1, 3, 1, 1])
SDDMM(message func: v_dot_e)
_____________________ test_sddmm[idtype0-dot-e-v-shp0-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'v', rhs_target = 'e'
msg = 'dot', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72be045e8>
op = 'dot'
lhs_data = tensor([[[[[[1.5104],
            [1.7606],
            [1.2982]]],


          [[[1.9852],
            [1.8114],
    ...9]]],


          [[[1.8397],
            [1.4997],
            [1.8997]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.4596]],

           [[1.1299]],

           [[1.8077]]]],



         [[[[1.6963]],

           [[1.495...



         [[[[1.2681]],

           [[1.7367]],

           [[1.2025]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([40, 1, 2, 1, 3, 1]), rhs shape: torch.Size([300, 4, 1, 3, 1, 1])
SDDMM(message func: v_dot_e)
_____________________ test_sddmm[idtype0-dot-e-v-shp1-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Sch...pe=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'v', rhs_target = 'e'
msg = 'dot', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43e9274c8>
op = 'dot'
lhs_data = tensor([[[[[1.2936, 1.3064, 1.6907,  ..., 1.2948, 1.8366, 1.6012]],

          [[1.1755, 1.8378, 1.7990,  ..., 1.7403,...
          [[1.0287, 1.8056, 1.1677,  ..., 1.7228, 1.5361, 1.1875]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.7375, 1.4850, 1.5103,  ..., 1.7238, 1.8159, 1.4189],
           [1.0428, 1.8035, 1.4621,  ..., 1.6600, 1...
           [1.1664, 1.5345, 1.2418,  ..., 1.4719, 1.6058, 1.1171]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([100, 1, 3, 7, 7])
SDDMM(message func: v_dot_e)
_____________________ test_sddmm[idtype0-dot-e-v-shp1-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'v', rhs_target = 'e'
msg = 'dot', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72ba32708>
op = 'dot'
lhs_data = tensor([[[[[1.7529, 1.8238, 1.0054,  ..., 1.4422, 1.7124, 1.1491]],

          [[1.9242, 1.2637, 1.1755,  ..., 1.6490,...
          [[1.7556, 1.1254, 1.1926,  ..., 1.9430, 1.5380, 1.2210]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.7215, 1.6786, 1.3138,  ..., 1.4638, 1.1647, 1.1155],
           [1.3375, 1.5924, 1.7125,  ..., 1.2052, 1...
           [1.5085, 1.8157, 1.3332,  ..., 1.9548, 1.8612, 1.7909]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([40, 5, 3, 1, 7]), rhs shape: torch.Size([300, 1, 3, 7, 7])
SDDMM(message func: v_dot_e)
_____________________ test_sddmm[idtype0-dot-e-v-shp2-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme...dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'v', rhs_target = 'e', msg = 'dot'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43e927798>
op = 'dot'
lhs_data = tensor([[[[1.9815, 1.2409, 1.4790],
          [1.0806, 1.4785, 1.7506],
          [1.7950, 1.3362, 1.8083]]],


      ...8],
          [1.8406, 1.4357, 1.8108],
          [1.0398, 1.5274, 1.4073]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.4680, 1.3425, 1.4757]],

         [[1.7407, 1.6556, 1.6782]],

         [[1.4782, 1.5052, 1.8812]],

    ...

         [[1.1037, 1.6966, 1.0433]],

         [[1.5215, 1.6795, 1.7228]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([100, 4, 1, 3])
SDDMM(message func: v_dot_e)
_____________________ test_sddmm[idtype0-dot-e-v-shp2-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'v', rhs_target = 'e', msg = 'dot'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb705e411b0>
op = 'dot'
lhs_data = tensor([[[[1.2509, 1.1028, 1.2931],
          [1.5070, 1.0918, 1.9627],
          [1.7776, 1.3294, 1.7558]]],


      ...1],
          [1.8012, 1.0851, 1.1174],
          [1.2318, 1.9634, 1.7486]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.8209, 1.5774, 1.9352]],

         [[1.5307, 1.4403, 1.6198]],

         [[1.9999, 1.7632, 1.8059]],

    ...

         [[1.0668, 1.1966, 1.9087]],

         [[1.0942, 1.2105, 1.8923]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([40, 1, 3, 3]), rhs shape: torch.Size([300, 4, 1, 3])
SDDMM(message func: v_dot_e)
_____________________ test_sddmm[idtype0-dot-e-v-shp3-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'v', rhs_target = 'e', msg = 'dot'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72be171f8>
op = 'dot'
lhs_data = tensor([[1.4302, 1.8201, 1.3457],
        [1.1042, 1.3760, 1.1904],
        [1.3413, 1.0499, 1.6852],
        [1.1099,... 1.4921],
        [1.1428, 1.7793, 1.6214],
        [1.6927, 1.0195, 1.6632]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.3658, 1.2280, 1.1978],
        [1.7691, 1.9099, 1.2474],
        [1.5229, 1.3515, 1.4848],
        [1.4617,... 1.1493],
        [1.3757, 1.3038, 1.9684],
        [1.1026, 1.1893, 1.0029]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([100, 3])
SDDMM(message func: v_dot_e)
_____________________ test_sddmm[idtype0-dot-e-v-shp3-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'v', rhs_target = 'e', msg = 'dot'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72be17438>
op = 'dot'
lhs_data = tensor([[1.0952, 1.5003, 1.1581],
        [1.0287, 1.6598, 1.8738],
        [1.9734, 1.4945, 1.7509],
        [1.3548,... 1.2096],
        [1.9662, 1.1699, 1.6117],
        [1.7480, 1.6916, 1.3036]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.8872, 1.5368, 1.1589],
        [1.2342, 1.1644, 1.7984],
        [1.0335, 1.5830, 1.2710],
        [1.7462,... 1.8702],
        [1.6280, 1.3304, 1.2558],
        [1.9104, 1.0512, 1.7903]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([40, 3]), rhs shape: torch.Size([300, 3])
SDDMM(message func: v_dot_e)
_____________________ test_sddmm[idtype0-dot-e-v-shp4-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'v', rhs_target = 'e', msg = 'dot'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43e9220d8>
op = 'dot'
lhs_data = tensor([[1.1832],
        [1.6929],
        [1.0749],
        [1.6664],
        [1.6261],
        [1.3676],
        [1...853],
        [1.8530],
        [1.1264],
        [1.2744],
        [1.4285]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.1722],
        [1.2304],
        [1.2738],
        [1.7734],
        [1.9383],
        [1.8808],
        [1...309],
        [1.9389],
        [1.8951],
        [1.3045],
        [1.9704]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([100, 1])
SDDMM(message func: v_dot_e)
_____________________ test_sddmm[idtype0-dot-e-v-shp4-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'v', rhs_target = 'e', msg = 'dot'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43e849630>
op = 'dot'
lhs_data = tensor([[1.0956],
        [1.2342],
        [1.1932],
        [1.9459],
        [1.9533],
        [1.4235],
        [1...210],
        [1.4134],
        [1.9273],
        [1.4785],
        [1.4462]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.5075],
        [1.2227],
        [1.1115],
        [1.9942],
        [1.3835],
        [1.4978],
        [1...511],
        [1.5345],
        [1.0993],
        [1.2742],
        [1.9286]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([40, 1]), rhs shape: torch.Size([300, 1])
SDDMM(message func: v_dot_e)
___________________ test_sddmm[idtype0-copy_lhs-u-v-shp0-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': ...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'v', rhs_target = 'u'
msg = 'copy_lhs', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb6143984c8>
op = 'copy_lhs'
lhs_data = tensor([[[[[[1.6837],
            [1.2413],
            [1.6359]]],


          [[[1.3636],
            [1.7270],
    ...1]]],


          [[[1.4177],
            [1.6080],
            [1.5037]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.3789]],

           [[1.5412]],

           [[1.7525]]]],



         [[[[1.3604]],

           [[1.853...



         [[[[1.1888]],

           [[1.3556]],

           [[1.4434]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: v_copy_lhs_u)
___________________ test_sddmm[idtype0-copy_lhs-u-v-shp0-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'v', rhs_target = 'u'
msg = 'copy_lhs', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72be27fc0>
op = 'copy_lhs'
lhs_data = tensor([[[[[[1.8100],
            [1.2427],
            [1.9228]]],


          [[[1.9665],
            [1.2009],
    ...8]]],


          [[[1.7426],
            [1.8462],
            [1.4277]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.3752]],

           [[1.7058]],

           [[1.4571]]]],



         [[[[1.2555]],

           [[1.822...



         [[[[1.7266]],

           [[1.5753]],

           [[1.4111]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([40, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: v_copy_lhs_u)
___________________ test_sddmm[idtype0-copy_lhs-u-v-shp1-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Sch...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'v', rhs_target = 'u'
msg = 'copy_lhs', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72be38288>
op = 'copy_lhs'
lhs_data = tensor([[[[[1.8199, 1.4694, 1.3741,  ..., 1.4988, 1.1039, 1.5017]],

          [[1.5580, 1.6459, 1.6290,  ..., 1.6362,...
          [[1.1442, 1.6178, 1.1569,  ..., 1.8643, 1.8336, 1.9806]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.2514, 1.2732, 1.7811,  ..., 1.8257, 1.2770, 1.8944],
           [1.6552, 1.3703, 1.2357,  ..., 1.9007, 1...
           [1.3363, 1.3523, 1.8399,  ..., 1.8144, 1.3443, 1.9490]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: v_copy_lhs_u)
___________________ test_sddmm[idtype0-copy_lhs-u-v-shp1-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'v', rhs_target = 'u'
msg = 'copy_lhs', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72be38480>
op = 'copy_lhs'
lhs_data = tensor([[[[[1.0565, 1.0760, 1.0084,  ..., 1.4590, 1.9797, 1.9070]],

          [[1.4584, 1.6912, 1.1529,  ..., 1.7272,...
          [[1.0768, 1.3847, 1.7455,  ..., 1.9571, 1.6459, 1.3916]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.9511, 1.7976, 1.1552,  ..., 1.1093, 1.3686, 1.0337],
           [1.4535, 1.9341, 1.0622,  ..., 1.7464, 1...
           [1.1255, 1.1114, 1.9432,  ..., 1.8531, 1.6606, 1.7971]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([40, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: v_copy_lhs_u)
___________________ test_sddmm[idtype0-copy_lhs-u-v-shp2-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'v', rhs_target = 'u'
msg = 'copy_lhs', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb705bf7ee8>
op = 'copy_lhs'
lhs_data = tensor([[[[1.4888, 1.9649, 1.9775],
          [1.6580, 1.8419, 1.2609],
          [1.2937, 1.4206, 1.3056]]],


      ...2],
          [1.0335, 1.0575, 1.2596],
          [1.9655, 1.2726, 1.4701]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.9892, 1.1318, 1.4347]],

         [[1.6237, 1.7112, 1.2158]],

         [[1.8485, 1.2237, 1.4326]],

    ...

         [[1.1814, 1.8891, 1.4270]],

         [[1.0771, 1.4917, 1.3531]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: v_copy_lhs_u)
___________________ test_sddmm[idtype0-copy_lhs-u-v-shp2-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'v', rhs_target = 'u'
msg = 'copy_lhs', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72bd600d8>
op = 'copy_lhs'
lhs_data = tensor([[[[1.9010, 1.5195, 1.0210],
          [1.8675, 1.3542, 1.0856],
          [1.8470, 1.2689, 1.2086]]],


      ...8],
          [1.5452, 1.4068, 1.6212],
          [1.3844, 1.3942, 1.5557]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.8741, 1.4867, 1.3565]],

         [[1.1622, 1.2473, 1.5471]],

         [[1.1489, 1.2734, 1.2641]],

    ...

         [[1.9079, 1.4841, 1.4489]],

         [[1.7003, 1.2821, 1.4313]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([40, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: v_copy_lhs_u)
___________________ test_sddmm[idtype0-copy_lhs-u-v-shp3-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'v', rhs_target = 'u', msg = 'copy_lhs'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb70624a3f0>
op = 'copy_lhs'
lhs_data = tensor([[1.1494, 1.8695, 1.2874],
        [1.8493, 1.3184, 1.0751],
        [1.4021, 1.2256, 1.2521],
        [1.4795,... 1.7497],
        [1.1192, 1.1459, 1.2280],
        [1.6896, 1.0224, 1.7988]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.6667, 1.6617, 1.2855],
        [1.3824, 1.0709, 1.3413],
        [1.6041, 1.0791, 1.9446],
        [1.4008,... 1.2198],
        [1.1328, 1.8523, 1.2828],
        [1.1939, 1.1404, 1.8454]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: v_copy_lhs_u)
___________________ test_sddmm[idtype0-copy_lhs-u-v-shp3-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'v', rhs_target = 'u', msg = 'copy_lhs'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72c07eb88>
op = 'copy_lhs'
lhs_data = tensor([[1.6231, 1.0379, 1.3574],
        [1.4450, 1.4854, 1.6868],
        [1.8645, 1.5591, 1.9867],
        [1.5653,... 1.6840],
        [1.6107, 1.0934, 1.4785],
        [1.5249, 1.2208, 1.5546]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.2063, 1.5090, 1.5880],
        [1.0566, 1.0652, 1.0894],
        [1.3934, 1.5257, 1.5699],
        [1.8262,... 1.4531],
        [1.7918, 1.5581, 1.9753],
        [1.4369, 1.6758, 1.4597]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([40, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: v_copy_lhs_u)
___________________ test_sddmm[idtype0-copy_lhs-u-v-shp4-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'v', rhs_target = 'u', msg = 'copy_lhs'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72b81a630>
op = 'copy_lhs'
lhs_data = tensor([[1.7799],
        [1.1450],
        [1.0195],
        [1.6964],
        [1.1447],
        [1.8296],
        [1...421],
        [1.3083],
        [1.9093],
        [1.2553],
        [1.5223]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.9003],
        [1.2734],
        [1.2600],
        [1.1282],
        [1.8839],
        [1.6089],
        [1...346],
        [1.8222],
        [1.8346],
        [1.4710],
        [1.7776]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: v_copy_lhs_u)
___________________ test_sddmm[idtype0-copy_lhs-u-v-shp4-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'v', rhs_target = 'u', msg = 'copy_lhs'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72be1a1f8>
op = 'copy_lhs'
lhs_data = tensor([[1.8817],
        [1.8024],
        [1.2411],
        [1.9283],
        [1.2407],
        [1.5111],
        [1...676],
        [1.7573],
        [1.5950],
        [1.8302],
        [1.3915]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.2841],
        [1.1942],
        [1.8219],
        [1.4976],
        [1.2506],
        [1.8497],
        [1...077],
        [1.9910],
        [1.8363],
        [1.8365],
        [1.8555]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([40, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: v_copy_lhs_u)
___________________ test_sddmm[idtype0-copy_lhs-u-e-shp0-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...torch.float64), 'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'e', rhs_target = 'u'
msg = 'copy_lhs', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72be1a480>
op = 'copy_lhs'
lhs_data = tensor([[[[[[1.9682],
            [1.6224],
            [1.4288]]],


          [[[1.5988],
            [1.8274],
    ...6]]],


          [[[1.6448],
            [1.9657],
            [1.1459]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.7639]],

           [[1.4653]],

           [[1.9670]]]],



         [[[[1.2659]],

           [[1.471...



         [[[[1.9569]],

           [[1.2241]],

           [[1.2063]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([100, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: e_copy_lhs_u)
___________________ test_sddmm[idtype0-copy_lhs-u-e-shp0-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'e', rhs_target = 'u'
msg = 'copy_lhs', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72be1a6c0>
op = 'copy_lhs'
lhs_data = tensor([[[[[[1.9123],
            [1.7268],
            [1.0714]]],


          [[[1.7865],
            [1.1983],
    ...0]]],


          [[[1.6869],
            [1.2344],
            [1.6618]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.5713]],

           [[1.6621]],

           [[1.3917]]]],



         [[[[1.2631]],

           [[1.988...



         [[[[1.8247]],

           [[1.0518]],

           [[1.9517]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([300, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: e_copy_lhs_u)
___________________ test_sddmm[idtype0-copy_lhs-u-e-shp1-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...pe=torch.float64), 'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'e', rhs_target = 'u'
msg = 'copy_lhs', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72949eaf8>
op = 'copy_lhs'
lhs_data = tensor([[[[[1.4102, 1.8487, 1.3722,  ..., 1.8695, 1.6304, 1.1701]],

          [[1.5535, 1.4272, 1.6620,  ..., 1.1921,...
          [[1.1499, 1.2955, 1.4665,  ..., 1.3395, 1.0404, 1.8818]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.7767, 1.8110, 1.6131,  ..., 1.4587, 1.5498, 1.5281],
           [1.4960, 1.3019, 1.0330,  ..., 1.9285, 1...
           [1.1037, 1.9946, 1.8857,  ..., 1.2944, 1.8166, 1.0929]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([100, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: e_copy_lhs_u)
___________________ test_sddmm[idtype0-copy_lhs-u-e-shp1-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'e', rhs_target = 'u'
msg = 'copy_lhs', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72bd5c120>
op = 'copy_lhs'
lhs_data = tensor([[[[[1.7629, 1.8786, 1.1493,  ..., 1.7969, 1.3521, 1.2002]],

          [[1.1412, 1.9673, 1.3361,  ..., 1.4476,...
          [[1.1321, 1.9901, 1.2797,  ..., 1.9628, 1.7525, 1.3087]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.2523, 1.1940, 1.1055,  ..., 1.5263, 1.0610, 1.1300],
           [1.4071, 1.2589, 1.7225,  ..., 1.7323, 1...
           [1.3553, 1.1138, 1.4057,  ..., 1.2675, 1.6391, 1.3590]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([300, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: e_copy_lhs_u)
___________________ test_sddmm[idtype0-copy_lhs-u-e-shp2-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...dtype=torch.float64), 'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'e', rhs_target = 'u'
msg = 'copy_lhs', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb705e25f30>
op = 'copy_lhs'
lhs_data = tensor([[[[1.3652, 1.1935, 1.0402],
          [1.7181, 1.9722, 1.5105],
          [1.2255, 1.3652, 1.5758]]],


      ...7],
          [1.5738, 1.6092, 1.8986],
          [1.4508, 1.8393, 1.8014]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.6595, 1.1773, 1.0730]],

         [[1.8726, 1.8121, 1.4642]],

         [[1.8301, 1.9626, 1.6352]],

    ...

         [[1.7945, 1.0533, 1.1469]],

         [[1.1097, 1.9927, 1.9860]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([100, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: e_copy_lhs_u)
___________________ test_sddmm[idtype0-copy_lhs-u-e-shp2-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'e', rhs_target = 'u'
msg = 'copy_lhs', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb614221ea0>
op = 'copy_lhs'
lhs_data = tensor([[[[1.7496, 1.1281, 1.2810],
          [1.8622, 1.9471, 1.3201],
          [1.4043, 1.0645, 1.8991]]],


      ...1],
          [1.4503, 1.4060, 1.5070],
          [1.7688, 1.3931, 1.0724]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.5379, 1.0749, 1.6175]],

         [[1.6706, 1.1107, 1.2519]],

         [[1.8316, 1.3205, 1.7963]],

    ...

         [[1.5038, 1.1221, 1.0400]],

         [[1.6048, 1.7368, 1.4570]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([300, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: e_copy_lhs_u)
___________________ test_sddmm[idtype0-copy_lhs-u-e-shp3-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'e', rhs_target = 'u', msg = 'copy_lhs'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72be06240>
op = 'copy_lhs'
lhs_data = tensor([[1.0988, 1.6893, 1.6191],
        [1.4013, 1.8048, 1.0220],
        [1.5435, 1.8145, 1.2006],
        [1.6132,... 1.9164],
        [1.8424, 1.3039, 1.4599],
        [1.8205, 1.3453, 1.8872]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.7278, 1.0361, 1.2738],
        [1.3780, 1.2438, 1.2102],
        [1.6145, 1.5063, 1.1464],
        [1.9012,... 1.5646],
        [1.3034, 1.6760, 1.5864],
        [1.1936, 1.5320, 1.6277]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([100, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: e_copy_lhs_u)
___________________ test_sddmm[idtype0-copy_lhs-u-e-shp3-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'e', rhs_target = 'u', msg = 'copy_lhs'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72be063f0>
op = 'copy_lhs'
lhs_data = tensor([[1.1588, 1.6701, 1.2237],
        [1.1366, 1.8710, 1.0451],
        [1.4581, 1.5400, 1.1282],
        [1.1267,... 1.9130],
        [1.3253, 1.5419, 1.3939],
        [1.3117, 1.8844, 1.0485]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.3793, 1.4263, 1.3680],
        [1.9029, 1.9187, 1.5125],
        [1.3538, 1.1392, 1.7200],
        [1.6496,... 1.9748],
        [1.3449, 1.4139, 1.0515],
        [1.0485, 1.6476, 1.2642]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([300, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: e_copy_lhs_u)
___________________ test_sddmm[idtype0-copy_lhs-u-e-shp4-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'e', rhs_target = 'u', msg = 'copy_lhs'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43e6deea0>
op = 'copy_lhs'
lhs_data = tensor([[1.0022],
        [1.4591],
        [1.2928],
        [1.2827],
        [1.4062],
        [1.8494],
        [1...160],
        [1.4533],
        [1.0655],
        [1.2637],
        [1.9610]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.5533],
        [1.1482],
        [1.0041],
        [1.1086],
        [1.0223],
        [1.4067],
        [1...115],
        [1.2296],
        [1.0890],
        [1.4027],
        [1.3794]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([100, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: e_copy_lhs_u)
___________________ test_sddmm[idtype0-copy_lhs-u-e-shp4-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'e', rhs_target = 'u', msg = 'copy_lhs'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb705e404c8>
op = 'copy_lhs'
lhs_data = tensor([[1.8608],
        [1.6209],
        [1.4177],
        [1.3730],
        [1.6429],
        [1.0036],
        [1...235],
        [1.5812],
        [1.4332],
        [1.4787],
        [1.2173]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.4708],
        [1.9216],
        [1.1388],
        [1.4444],
        [1.2153],
        [1.4296],
        [1...408],
        [1.6456],
        [1.2390],
        [1.8241],
        [1.9180]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([300, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: e_copy_lhs_u)
___________________ test_sddmm[idtype0-copy_lhs-v-u-shp0-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': ...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'u', rhs_target = 'v'
msg = 'copy_lhs', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb7060f2a68>
op = 'copy_lhs'
lhs_data = tensor([[[[[[1.3786],
            [1.4524],
            [1.1172]]],


          [[[1.1668],
            [1.4282],
    ...6]]],


          [[[1.1066],
            [1.2924],
            [1.4381]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.8587]],

           [[1.6105]],

           [[1.8337]]]],



         [[[[1.6670]],

           [[1.445...



         [[[[1.3177]],

           [[1.5683]],

           [[1.7306]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: u_copy_lhs_v)
___________________ test_sddmm[idtype0-copy_lhs-v-u-shp0-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'u', rhs_target = 'v'
msg = 'copy_lhs', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43e869d80>
op = 'copy_lhs'
lhs_data = tensor([[[[[[1.4804],
            [1.1267],
            [1.3953]]],


          [[[1.0879],
            [1.1035],
    ...5]]],


          [[[1.5195],
            [1.8432],
            [1.9167]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.3713]],

           [[1.9562]],

           [[1.8053]]]],



         [[[[1.6539]],

           [[1.031...



         [[[[1.5591]],

           [[1.4335]],

           [[1.0168]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([40, 4, 1, 3, 1, 1])
SDDMM(message func: u_copy_lhs_v)
___________________ test_sddmm[idtype0-copy_lhs-v-u-shp1-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Sch...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'u', rhs_target = 'v'
msg = 'copy_lhs', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72bd5c120>
op = 'copy_lhs'
lhs_data = tensor([[[[[1.2176, 1.3132, 1.6770,  ..., 1.0096, 1.6829, 1.9301]],

          [[1.4289, 1.1128, 1.8556,  ..., 1.2058,...
          [[1.1627, 1.7093, 1.1144,  ..., 1.9393, 1.9049, 1.0027]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.8082, 1.7791, 1.6426,  ..., 1.3908, 1.1652, 1.4104],
           [1.2721, 1.3060, 1.1005,  ..., 1.0297, 1...
           [1.6216, 1.6494, 1.0018,  ..., 1.8436, 1.1035, 1.1439]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: u_copy_lhs_v)
___________________ test_sddmm[idtype0-copy_lhs-v-u-shp1-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'u', rhs_target = 'v'
msg = 'copy_lhs', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72be16240>
op = 'copy_lhs'
lhs_data = tensor([[[[[1.3290, 1.4561, 1.8590,  ..., 1.9571, 1.8631, 1.7343]],

          [[1.4349, 1.3760, 1.4088,  ..., 1.6353,...
          [[1.3656, 1.3801, 1.3830,  ..., 1.0806, 1.0763, 1.2390]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.9090, 1.6785, 1.4356,  ..., 1.9332, 1.6074, 1.6073],
           [1.4716, 1.1377, 1.7245,  ..., 1.0166, 1...
           [1.3842, 1.4930, 1.2233,  ..., 1.1574, 1.9010, 1.9264]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([40, 1, 3, 7, 7])
SDDMM(message func: u_copy_lhs_v)
___________________ test_sddmm[idtype0-copy_lhs-v-u-shp2-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'u', rhs_target = 'v'
msg = 'copy_lhs', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72b810d80>
op = 'copy_lhs'
lhs_data = tensor([[[[1.1377, 1.6523, 1.3953],
          [1.3434, 1.7607, 1.4619],
          [1.6065, 1.3434, 1.1614]]],


      ...6],
          [1.0325, 1.8798, 1.1702],
          [1.6490, 1.5688, 1.5046]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.1660, 1.2248, 1.8803]],

         [[1.4899, 1.4458, 1.9441]],

         [[1.5039, 1.9143, 1.2282]],

    ...

         [[1.9571, 1.1592, 1.7972]],

         [[1.6322, 1.5090, 1.4939]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: u_copy_lhs_v)
___________________ test_sddmm[idtype0-copy_lhs-v-u-shp2-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'u', rhs_target = 'v'
msg = 'copy_lhs', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb614388240>
op = 'copy_lhs'
lhs_data = tensor([[[[1.1685, 1.7899, 1.8570],
          [1.3209, 1.0738, 1.4541],
          [1.4122, 1.7498, 1.3746]]],


      ...8],
          [1.2739, 1.1693, 1.2341],
          [1.1157, 1.5638, 1.0153]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.1320, 1.1644, 1.6405]],

         [[1.7243, 1.3386, 1.7535]],

         [[1.6047, 1.4968, 1.6567]],

    ...

         [[1.4299, 1.7993, 1.8521]],

         [[1.1229, 1.9830, 1.3181]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([40, 4, 1, 3])
SDDMM(message func: u_copy_lhs_v)
___________________ test_sddmm[idtype0-copy_lhs-v-u-shp3-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'u', rhs_target = 'v', msg = 'copy_lhs'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72c0dd3a8>
op = 'copy_lhs'
lhs_data = tensor([[1.6487, 1.6244, 1.5160],
        [1.8860, 1.9315, 1.4237],
        [1.3475, 1.9309, 1.2374],
        [1.9431,... 1.7923],
        [1.7486, 1.7614, 1.2973],
        [1.5628, 1.2704, 1.7439]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.7023, 1.3979, 1.7616],
        [1.7647, 1.5601, 1.7006],
        [1.3475, 1.9373, 1.0216],
        [1.7517,... 1.2401],
        [1.7019, 1.0516, 1.2988],
        [1.7480, 1.3891, 1.8313]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: u_copy_lhs_v)
___________________ test_sddmm[idtype0-copy_lhs-v-u-shp3-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'u', rhs_target = 'v', msg = 'copy_lhs'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb7060f28b8>
op = 'copy_lhs'
lhs_data = tensor([[1.3368, 1.5981, 1.1393],
        [1.7793, 1.5230, 1.4397],
        [1.7101, 1.1227, 1.5458],
        [1.7474,... 1.8731],
        [1.0056, 1.3044, 1.3595],
        [1.5012, 1.6378, 1.6813]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.2448, 1.1398, 1.6595],
        [1.6795, 1.5192, 1.7598],
        [1.2773, 1.7914, 1.0893],
        [1.5313,... 1.6717],
        [1.0208, 1.4857, 1.2192],
        [1.2322, 1.3404, 1.0274]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([40, 3])
SDDMM(message func: u_copy_lhs_v)
___________________ test_sddmm[idtype0-copy_lhs-v-u-shp4-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'u', rhs_target = 'v', msg = 'copy_lhs'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb6146f7630>
op = 'copy_lhs'
lhs_data = tensor([[1.1949],
        [1.5569],
        [1.1484],
        [1.1911],
        [1.8957],
        [1.3945],
        [1...388],
        [1.3129],
        [1.1058],
        [1.1074],
        [1.0527]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.4248],
        [1.3613],
        [1.3656],
        [1.4104],
        [1.9613],
        [1.4764],
        [1...645],
        [1.3018],
        [1.5059],
        [1.4383],
        [1.5089]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: u_copy_lhs_v)
___________________ test_sddmm[idtype0-copy_lhs-v-u-shp4-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'u', rhs_target = 'v', msg = 'copy_lhs'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb7293a3750>
op = 'copy_lhs'
lhs_data = tensor([[1.5810],
        [1.6652],
        [1.9289],
        [1.5272],
        [1.8532],
        [1.0978],
        [1...934],
        [1.9446],
        [1.9406],
        [1.2659],
        [1.6002]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.7947],
        [1.4581],
        [1.2923],
        [1.9881],
        [1.5787],
        [1.1047],
        [1...401],
        [1.1680],
        [1.5947],
        [1.6449],
        [1.6118]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([40, 1])
SDDMM(message func: u_copy_lhs_v)
___________________ test_sddmm[idtype0-copy_lhs-v-e-shp0-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...torch.float64), 'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'e', rhs_target = 'v'
msg = 'copy_lhs', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72ba3b948>
op = 'copy_lhs'
lhs_data = tensor([[[[[[1.4934],
            [1.8750],
            [1.1428]]],


          [[[1.0847],
            [1.6281],
    ...3]]],


          [[[1.0370],
            [1.4299],
            [1.8161]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.3847]],

           [[1.9502]],

           [[1.3862]]]],



         [[[[1.3226]],

           [[1.967...



         [[[[1.0026]],

           [[1.1400]],

           [[1.4025]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([100, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: e_copy_lhs_v)
___________________ test_sddmm[idtype0-copy_lhs-v-e-shp0-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'e', rhs_target = 'v'
msg = 'copy_lhs', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72bfc7a68>
op = 'copy_lhs'
lhs_data = tensor([[[[[[1.4032],
            [1.5536],
            [1.1636]]],


          [[[1.7309],
            [1.3589],
    ...2]]],


          [[[1.3282],
            [1.6093],
            [1.6524]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.0027]],

           [[1.7724]],

           [[1.5652]]]],



         [[[[1.4435]],

           [[1.519...



         [[[[1.0300]],

           [[1.8676]],

           [[1.1499]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([300, 1, 2, 1, 3, 1]), rhs shape: torch.Size([40, 4, 1, 3, 1, 1])
SDDMM(message func: e_copy_lhs_v)
___________________ test_sddmm[idtype0-copy_lhs-v-e-shp1-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...pe=torch.float64), 'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'e', rhs_target = 'v'
msg = 'copy_lhs', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb70624e168>
op = 'copy_lhs'
lhs_data = tensor([[[[[1.2908, 1.2382, 1.5887,  ..., 1.5564, 1.4113, 1.7163]],

          [[1.5149, 1.0488, 1.6798,  ..., 1.5238,...
          [[1.7193, 1.2198, 1.5572,  ..., 1.5195, 1.9197, 1.8070]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.8127, 1.2324, 1.6161,  ..., 1.1316, 1.9000, 1.9988],
           [1.1104, 1.9191, 1.4277,  ..., 1.5859, 1...
           [1.0821, 1.7470, 1.3086,  ..., 1.8793, 1.2015, 1.1289]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([100, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: e_copy_lhs_v)
___________________ test_sddmm[idtype0-copy_lhs-v-e-shp1-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'e', rhs_target = 'v'
msg = 'copy_lhs', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb6142285a0>
op = 'copy_lhs'
lhs_data = tensor([[[[[1.8500, 1.7437, 1.6374,  ..., 1.1963, 1.3857, 1.1868]],

          [[1.8597, 1.2737, 1.9956,  ..., 1.8287,...
          [[1.3530, 1.9746, 1.0798,  ..., 1.5055, 1.0626, 1.7793]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.4532, 1.1247, 1.9847,  ..., 1.4455, 1.0533, 1.4875],
           [1.0787, 1.3570, 1.1898,  ..., 1.4821, 1...
           [1.1944, 1.3382, 1.9883,  ..., 1.7304, 1.5516, 1.5562]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([300, 5, 3, 1, 7]), rhs shape: torch.Size([40, 1, 3, 7, 7])
SDDMM(message func: e_copy_lhs_v)
___________________ test_sddmm[idtype0-copy_lhs-v-e-shp2-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...dtype=torch.float64), 'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'e', rhs_target = 'v'
msg = 'copy_lhs', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72bfcb240>
op = 'copy_lhs'
lhs_data = tensor([[[[1.3318, 1.7373, 1.1638],
          [1.7744, 1.7960, 1.4251],
          [1.5133, 1.1480, 1.6928]]],


      ...6],
          [1.4074, 1.0011, 1.2896],
          [1.8300, 1.9250, 1.4728]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.7218, 1.3230, 1.0578]],

         [[1.0663, 1.4559, 1.9180]],

         [[1.4005, 1.6701, 1.9725]],

    ...

         [[1.0376, 1.4479, 1.3919]],

         [[1.0709, 1.2769, 1.1581]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([100, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: e_copy_lhs_v)
___________________ test_sddmm[idtype0-copy_lhs-v-e-shp2-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'e', rhs_target = 'v'
msg = 'copy_lhs', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72bfcb480>
op = 'copy_lhs'
lhs_data = tensor([[[[1.9837, 1.3453, 1.1451],
          [1.0393, 1.7102, 1.1231],
          [1.9667, 1.3427, 1.9188]]],


      ...3],
          [1.9856, 1.6677, 1.3436],
          [1.7302, 1.8782, 1.7244]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.1536, 1.1899, 1.0132]],

         [[1.5045, 1.0409, 1.8340]],

         [[1.3792, 1.3472, 1.6824]],

    ...

         [[1.9552, 1.6238, 1.0284]],

         [[1.1406, 1.4779, 1.4915]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([300, 1, 3, 3]), rhs shape: torch.Size([40, 4, 1, 3])
SDDMM(message func: e_copy_lhs_v)
___________________ test_sddmm[idtype0-copy_lhs-v-e-shp3-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'e', rhs_target = 'v', msg = 'copy_lhs'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb614223900>
op = 'copy_lhs'
lhs_data = tensor([[1.7252, 1.0840, 1.9640],
        [1.1687, 1.9203, 1.1945],
        [1.6023, 1.0339, 1.6254],
        [1.4512,... 1.1999],
        [1.6758, 1.4689, 1.2436],
        [1.8462, 1.3855, 1.1263]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.9029, 1.9144, 1.8875],
        [1.6326, 1.5043, 1.6219],
        [1.4152, 1.8648, 1.1577],
        [1.3812,... 1.8584],
        [1.4997, 1.2877, 1.1594],
        [1.1848, 1.9083, 1.3447]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([100, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: e_copy_lhs_v)
___________________ test_sddmm[idtype0-copy_lhs-v-e-shp3-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'e', rhs_target = 'v', msg = 'copy_lhs'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72c07e168>
op = 'copy_lhs'
lhs_data = tensor([[1.8749, 1.8069, 1.6576],
        [1.6872, 1.8014, 1.5111],
        [1.5775, 1.5092, 1.2959],
        [1.3529,... 1.8311],
        [1.9945, 1.5891, 1.8284],
        [1.3741, 1.1008, 1.3976]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.3056, 1.0249, 1.2464],
        [1.4173, 1.6492, 1.4299],
        [1.0583, 1.1717, 1.8360],
        [1.1449,... 1.9553],
        [1.1812, 1.2312, 1.5149],
        [1.7617, 1.6760, 1.5054]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([300, 3]), rhs shape: torch.Size([40, 3])
SDDMM(message func: e_copy_lhs_v)
___________________ test_sddmm[idtype0-copy_lhs-v-e-shp4-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'e', rhs_target = 'v', msg = 'copy_lhs'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72ba3d5e8>
op = 'copy_lhs'
lhs_data = tensor([[1.0691],
        [1.2853],
        [1.8684],
        [1.8701],
        [1.4684],
        [1.5600],
        [1...370],
        [1.5768],
        [1.0738],
        [1.9004],
        [1.9864]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.9871],
        [1.9690],
        [1.5126],
        [1.4504],
        [1.5510],
        [1.8117],
        [1...491],
        [1.4943],
        [1.1473],
        [1.1924],
        [1.0624]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([100, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: e_copy_lhs_v)
___________________ test_sddmm[idtype0-copy_lhs-v-e-shp4-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'e', rhs_target = 'v', msg = 'copy_lhs'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb457b7f750>
op = 'copy_lhs'
lhs_data = tensor([[1.0670],
        [1.8762],
        [1.2532],
        [1.4050],
        [1.8574],
        [1.3349],
        [1...663],
        [1.2824],
        [1.3146],
        [1.0117],
        [1.8680]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.0001],
        [1.1081],
        [1.3189],
        [1.9353],
        [1.0028],
        [1.4692],
        [1...740],
        [1.1331],
        [1.7331],
        [1.9915],
        [1.9095]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([300, 1]), rhs shape: torch.Size([40, 1])
SDDMM(message func: e_copy_lhs_v)
___________________ test_sddmm[idtype0-copy_lhs-e-u-shp0-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': ...torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'u', rhs_target = 'e'
msg = 'copy_lhs', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72ad143a8>
op = 'copy_lhs'
lhs_data = tensor([[[[[[1.0624],
            [1.7483],
            [1.9800]]],


          [[[1.2382],
            [1.5803],
    ...3]]],


          [[[1.4555],
            [1.2134],
            [1.6336]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.8494]],

           [[1.2623]],

           [[1.4378]]]],



         [[[[1.1479]],

           [[1.565...



         [[[[1.7823]],

           [[1.6677]],

           [[1.5693]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([100, 4, 1, 3, 1, 1])
SDDMM(message func: u_copy_lhs_e)
___________________ test_sddmm[idtype0-copy_lhs-e-u-shp0-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'u', rhs_target = 'e'
msg = 'copy_lhs', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72bfc9240>
op = 'copy_lhs'
lhs_data = tensor([[[[[[1.1961],
            [1.3003],
            [1.4553]]],


          [[[1.4166],
            [1.5199],
    ...8]]],


          [[[1.0564],
            [1.2833],
            [1.8066]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.7015]],

           [[1.0030]],

           [[1.7208]]]],



         [[[[1.0459]],

           [[1.087...



         [[[[1.2934]],

           [[1.6561]],

           [[1.9137]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([300, 4, 1, 3, 1, 1])
SDDMM(message func: u_copy_lhs_e)
___________________ test_sddmm[idtype0-copy_lhs-e-u-shp1-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Sch...pe=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'u', rhs_target = 'e'
msg = 'copy_lhs', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72bd608b8>
op = 'copy_lhs'
lhs_data = tensor([[[[[1.1791, 1.2826, 1.6741,  ..., 1.1083, 1.8207, 1.6090]],

          [[1.0745, 1.4603, 1.0722,  ..., 1.4101,...
          [[1.3109, 1.5710, 1.0713,  ..., 1.9303, 1.9478, 1.1405]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.2938, 1.3529, 1.6591,  ..., 1.0975, 1.9537, 1.4835],
           [1.0146, 1.4225, 1.2202,  ..., 1.6568, 1...
           [1.7539, 1.9383, 1.1131,  ..., 1.4198, 1.6792, 1.6433]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([100, 1, 3, 7, 7])
SDDMM(message func: u_copy_lhs_e)
___________________ test_sddmm[idtype0-copy_lhs-e-u-shp1-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'u', rhs_target = 'e'
msg = 'copy_lhs', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72c0fa798>
op = 'copy_lhs'
lhs_data = tensor([[[[[1.9904, 1.2548, 1.1574,  ..., 1.8041, 1.5464, 1.4524]],

          [[1.5448, 1.1787, 1.4344,  ..., 1.6082,...
          [[1.8325, 1.0489, 1.8912,  ..., 1.7816, 1.8473, 1.8800]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.4894, 1.3893, 1.1910,  ..., 1.8879, 1.0056, 1.6098],
           [1.7727, 1.6522, 1.0417,  ..., 1.3371, 1...
           [1.3661, 1.9908, 1.0858,  ..., 1.5143, 1.5450, 1.7196]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([300, 1, 3, 7, 7])
SDDMM(message func: u_copy_lhs_e)
___________________ test_sddmm[idtype0-copy_lhs-e-u-shp2-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme...dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'u', rhs_target = 'e'
msg = 'copy_lhs', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb6146f7ab0>
op = 'copy_lhs'
lhs_data = tensor([[[[1.5324, 1.5485, 1.2139],
          [1.9242, 1.7839, 1.3781],
          [1.2258, 1.1341, 1.9913]]],


      ...6],
          [1.0235, 1.8113, 1.9620],
          [1.9312, 1.8824, 1.7906]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.3669, 1.6057, 1.8378]],

         [[1.6629, 1.8381, 1.9616]],

         [[1.2158, 1.1751, 1.2575]],

    ...

         [[1.2645, 1.9198, 1.9612]],

         [[1.6951, 1.0434, 1.7765]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([100, 4, 1, 3])
SDDMM(message func: u_copy_lhs_e)
___________________ test_sddmm[idtype0-copy_lhs-e-u-shp2-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'u', rhs_target = 'e'
msg = 'copy_lhs', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb7293b0ab0>
op = 'copy_lhs'
lhs_data = tensor([[[[1.2801, 1.2769, 1.9722],
          [1.4143, 1.2311, 1.8854],
          [1.7968, 1.8388, 1.6983]]],


      ...7],
          [1.6830, 1.3585, 1.7203],
          [1.6298, 1.2057, 1.2798]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.9371, 1.4806, 1.9215]],

         [[1.0175, 1.1972, 1.6357]],

         [[1.8293, 1.3152, 1.5706]],

    ...

         [[1.7327, 1.4094, 1.8264]],

         [[1.8283, 1.5313, 1.4920]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([300, 4, 1, 3])
SDDMM(message func: u_copy_lhs_e)
___________________ test_sddmm[idtype0-copy_lhs-e-u-shp3-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'u', rhs_target = 'e', msg = 'copy_lhs'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb61cf93090>
op = 'copy_lhs'
lhs_data = tensor([[1.6249, 1.8908, 1.3320],
        [1.9618, 1.4523, 1.2901],
        [1.1383, 1.6876, 1.1540],
        [1.4797,... 1.7376],
        [1.4533, 1.0877, 1.5004],
        [1.5255, 1.5416, 1.9287]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.6792, 1.0753, 1.0662],
        [1.1235, 1.6646, 1.3864],
        [1.7688, 1.8523, 1.8335],
        [1.8866,... 1.1437],
        [1.1715, 1.8722, 1.6905],
        [1.6542, 1.8193, 1.1585]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([100, 3])
SDDMM(message func: u_copy_lhs_e)
___________________ test_sddmm[idtype0-copy_lhs-e-u-shp3-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'u', rhs_target = 'e', msg = 'copy_lhs'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb61cf93558>
op = 'copy_lhs'
lhs_data = tensor([[1.4485, 1.6414, 1.8405],
        [1.7595, 1.9777, 1.5570],
        [1.9310, 1.5722, 1.3900],
        [1.9861,... 1.1130],
        [1.1870, 1.2380, 1.6251],
        [1.6945, 1.7095, 1.3758]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.1330, 1.2972, 1.8120],
        [1.2711, 1.7327, 1.0769],
        [1.5340, 1.1607, 1.4522],
        [1.7734,... 1.1227],
        [1.0456, 1.7129, 1.0157],
        [1.3049, 1.8204, 1.7845]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([300, 3])
SDDMM(message func: u_copy_lhs_e)
___________________ test_sddmm[idtype0-copy_lhs-e-u-shp4-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'u', rhs_target = 'e', msg = 'copy_lhs'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72bfdf288>
op = 'copy_lhs'
lhs_data = tensor([[1.9563],
        [1.7849],
        [1.5919],
        [1.8425],
        [1.9007],
        [1.6927],
        [1...778],
        [1.5990],
        [1.1428],
        [1.8271],
        [1.8616]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.4231],
        [1.9177],
        [1.0442],
        [1.1806],
        [1.5982],
        [1.4427],
        [1...434],
        [1.8581],
        [1.1193],
        [1.4374],
        [1.7619]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([100, 1])
SDDMM(message func: u_copy_lhs_e)
___________________ test_sddmm[idtype0-copy_lhs-e-u-shp4-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'u', rhs_target = 'e', msg = 'copy_lhs'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72bfdf480>
op = 'copy_lhs'
lhs_data = tensor([[1.3281],
        [1.1975],
        [1.4361],
        [1.8411],
        [1.3519],
        [1.1455],
        [1...536],
        [1.6161],
        [1.4053],
        [1.7588],
        [1.1310]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.5576],
        [1.9454],
        [1.8877],
        [1.2804],
        [1.6172],
        [1.6647],
        [1...801],
        [1.7077],
        [1.7798],
        [1.9691],
        [1.2515]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([300, 1])
SDDMM(message func: u_copy_lhs_e)
___________________ test_sddmm[idtype0-copy_lhs-e-v-shp0-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': ...torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'v', rhs_target = 'e'
msg = 'copy_lhs', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb6141580d8>
op = 'copy_lhs'
lhs_data = tensor([[[[[[1.8220],
            [1.3753],
            [1.2649]]],


          [[[1.0853],
            [1.2128],
    ...9]]],


          [[[1.9071],
            [1.9477],
            [1.2897]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.3913]],

           [[1.2853]],

           [[1.1213]]]],



         [[[[1.6471]],

           [[1.260...



         [[[[1.3632]],

           [[1.9769]],

           [[1.3342]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([100, 4, 1, 3, 1, 1])
SDDMM(message func: v_copy_lhs_e)
___________________ test_sddmm[idtype0-copy_lhs-e-v-shp0-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'v', rhs_target = 'e'
msg = 'copy_lhs', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb614235798>
op = 'copy_lhs'
lhs_data = tensor([[[[[[1.3445],
            [1.2226],
            [1.0950]]],


          [[[1.7620],
            [1.6099],
    ...4]]],


          [[[1.7803],
            [1.3603],
            [1.2819]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.2847]],

           [[1.3140]],

           [[1.3418]]]],



         [[[[1.6582]],

           [[1.781...



         [[[[1.9570]],

           [[1.6524]],

           [[1.9967]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([40, 1, 2, 1, 3, 1]), rhs shape: torch.Size([300, 4, 1, 3, 1, 1])
SDDMM(message func: v_copy_lhs_e)
___________________ test_sddmm[idtype0-copy_lhs-e-v-shp1-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Sch...pe=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'v', rhs_target = 'e'
msg = 'copy_lhs', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72aabc0d8>
op = 'copy_lhs'
lhs_data = tensor([[[[[1.7822, 1.9130, 1.6696,  ..., 1.3939, 1.3665, 1.9430]],

          [[1.4002, 1.2091, 1.9487,  ..., 1.8657,...
          [[1.3533, 1.7981, 1.0315,  ..., 1.1557, 1.6062, 1.9215]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.8344, 1.1288, 1.5732,  ..., 1.4341, 1.4124, 1.8225],
           [1.1024, 1.2506, 1.2484,  ..., 1.8420, 1...
           [1.2144, 1.3644, 1.0972,  ..., 1.2569, 1.5531, 1.4400]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([100, 1, 3, 7, 7])
SDDMM(message func: v_copy_lhs_e)
___________________ test_sddmm[idtype0-copy_lhs-e-v-shp1-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'v', rhs_target = 'e'
msg = 'copy_lhs', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72ab9d048>
op = 'copy_lhs'
lhs_data = tensor([[[[[1.1121, 1.5525, 1.3798,  ..., 1.2629, 1.0121, 1.5539]],

          [[1.1104, 1.7773, 1.3643,  ..., 1.1210,...
          [[1.1705, 1.7079, 1.8208,  ..., 1.2935, 1.6274, 1.2704]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.1581, 1.4077, 1.8030,  ..., 1.9711, 1.4882, 1.6110],
           [1.7837, 1.6248, 1.8485,  ..., 1.0428, 1...
           [1.6650, 1.1476, 1.6530,  ..., 1.2402, 1.5438, 1.7100]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([40, 5, 3, 1, 7]), rhs shape: torch.Size([300, 1, 3, 7, 7])
SDDMM(message func: v_copy_lhs_e)
___________________ test_sddmm[idtype0-copy_lhs-e-v-shp2-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme...dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'v', rhs_target = 'e'
msg = 'copy_lhs', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72b83c948>
op = 'copy_lhs'
lhs_data = tensor([[[[1.6647, 1.7811, 1.2207],
          [1.5470, 1.7230, 1.3829],
          [1.1701, 1.9185, 1.6224]]],


      ...5],
          [1.4713, 1.6801, 1.9440],
          [1.0748, 1.5458, 1.6901]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.0815, 1.4871, 1.1951]],

         [[1.8406, 1.9330, 1.4237]],

         [[1.4545, 1.1599, 1.0733]],

    ...

         [[1.9461, 1.3212, 1.6651]],

         [[1.7065, 1.8056, 1.7461]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([100, 4, 1, 3])
SDDMM(message func: v_copy_lhs_e)
___________________ test_sddmm[idtype0-copy_lhs-e-v-shp2-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'v', rhs_target = 'e'
msg = 'copy_lhs', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72bfdd090>
op = 'copy_lhs'
lhs_data = tensor([[[[1.7740, 1.2320, 1.4729],
          [1.5181, 1.7876, 1.5622],
          [1.2581, 1.6931, 1.7259]]],


      ...0],
          [1.4796, 1.8378, 1.1058],
          [1.1063, 1.6498, 1.4591]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.4985, 1.1149, 1.1649]],

         [[1.3539, 1.0677, 1.7883]],

         [[1.9023, 1.4284, 1.7446]],

    ...

         [[1.7310, 1.5453, 1.7114]],

         [[1.8966, 1.3660, 1.0089]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([40, 1, 3, 3]), rhs shape: torch.Size([300, 4, 1, 3])
SDDMM(message func: v_copy_lhs_e)
___________________ test_sddmm[idtype0-copy_lhs-e-v-shp3-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'v', rhs_target = 'e', msg = 'copy_lhs'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb705e4a870>
op = 'copy_lhs'
lhs_data = tensor([[1.8804, 1.6431, 1.5271],
        [1.9561, 1.7598, 1.3268],
        [1.1503, 1.2076, 1.1251],
        [1.3837,... 1.6185],
        [1.1053, 1.3762, 1.5574],
        [1.6754, 1.0531, 1.8900]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.1472, 1.3937, 1.7249],
        [1.5837, 1.6301, 1.4014],
        [1.0803, 1.8863, 1.4258],
        [1.2031,... 1.0856],
        [1.4861, 1.0540, 1.9632],
        [1.0737, 1.3892, 1.4466]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([100, 3])
SDDMM(message func: v_copy_lhs_e)
___________________ test_sddmm[idtype0-copy_lhs-e-v-shp3-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'v', rhs_target = 'e', msg = 'copy_lhs'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72bd443a8>
op = 'copy_lhs'
lhs_data = tensor([[1.1300, 1.3493, 1.9652],
        [1.1681, 1.6250, 1.1064],
        [1.3460, 1.1028, 1.4489],
        [1.0944,... 1.0216],
        [1.6589, 1.3593, 1.8940],
        [1.3918, 1.0341, 1.1565]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.6579, 1.4902, 1.2919],
        [1.8761, 1.7128, 1.5741],
        [1.8100, 1.9962, 1.1444],
        [1.8325,... 1.7029],
        [1.8884, 1.5881, 1.2160],
        [1.0769, 1.6529, 1.2880]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([40, 3]), rhs shape: torch.Size([300, 3])
SDDMM(message func: v_copy_lhs_e)
___________________ test_sddmm[idtype0-copy_lhs-e-v-shp4-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'v', rhs_target = 'e', msg = 'copy_lhs'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb6143a1ca8>
op = 'copy_lhs'
lhs_data = tensor([[1.3459],
        [1.9917],
        [1.8277],
        [1.0269],
        [1.8051],
        [1.1978],
        [1...096],
        [1.2044],
        [1.0104],
        [1.7447],
        [1.7289]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.6287],
        [1.5357],
        [1.4072],
        [1.6709],
        [1.2096],
        [1.3287],
        [1...495],
        [1.9947],
        [1.0067],
        [1.7133],
        [1.7937]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([100, 1])
SDDMM(message func: v_copy_lhs_e)
___________________ test_sddmm[idtype0-copy_lhs-e-v-shp4-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'v', rhs_target = 'e', msg = 'copy_lhs'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb61cfa4ca8>
op = 'copy_lhs'
lhs_data = tensor([[1.7052],
        [1.7762],
        [1.0055],
        [1.6723],
        [1.8877],
        [1.9144],
        [1...072],
        [1.0797],
        [1.3217],
        [1.5767],
        [1.1520]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.9021],
        [1.4244],
        [1.2568],
        [1.6457],
        [1.4845],
        [1.0372],
        [1...200],
        [1.4393],
        [1.2553],
        [1.5010],
        [1.6209]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([40, 1]), rhs shape: torch.Size([300, 1])
SDDMM(message func: v_copy_lhs_e)
___________________ test_sddmm[idtype0-copy_rhs-u-v-shp0-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': ...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'v', rhs_target = 'u'
msg = 'copy_rhs', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43e645630>
op = 'copy_rhs'
lhs_data = tensor([[[[[[1.7310],
            [1.0657],
            [1.2240]]],


          [[[1.3965],
            [1.0811],
    ...2]]],


          [[[1.0436],
            [1.3816],
            [1.2828]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.2331]],

           [[1.6895]],

           [[1.1257]]]],



         [[[[1.2496]],

           [[1.325...



         [[[[1.3718]],

           [[1.0752]],

           [[1.5236]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: v_copy_rhs_u)
___________________ test_sddmm[idtype0-copy_rhs-u-v-shp0-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'v', rhs_target = 'u'
msg = 'copy_rhs', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72bfead80>
op = 'copy_rhs'
lhs_data = tensor([[[[[[1.7265],
            [1.9321],
            [1.3837]]],


          [[[1.7754],
            [1.4933],
    ...6]]],


          [[[1.4535],
            [1.1395],
            [1.5508]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.9610]],

           [[1.3691]],

           [[1.9871]]]],



         [[[[1.9613]],

           [[1.140...



         [[[[1.2451]],

           [[1.3396]],

           [[1.7465]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([40, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: v_copy_rhs_u)
___________________ test_sddmm[idtype0-copy_rhs-u-v-shp1-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Sch...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'v', rhs_target = 'u'
msg = 'copy_rhs', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72b806a68>
op = 'copy_rhs'
lhs_data = tensor([[[[[1.4896, 1.2096, 1.4753,  ..., 1.4225, 1.6737, 1.8242]],

          [[1.0695, 1.6274, 1.2593,  ..., 1.5674,...
          [[1.1998, 1.0530, 1.7705,  ..., 1.3930, 1.9870, 1.2553]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.0018, 1.7173, 1.4749,  ..., 1.2842, 1.3746, 1.2321],
           [1.6216, 1.4613, 1.7651,  ..., 1.8785, 1...
           [1.7726, 1.6455, 1.4912,  ..., 1.2032, 1.8312, 1.3027]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: v_copy_rhs_u)
___________________ test_sddmm[idtype0-copy_rhs-u-v-shp1-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'v', rhs_target = 'u'
msg = 'copy_rhs', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72bff0240>
op = 'copy_rhs'
lhs_data = tensor([[[[[1.7576, 1.6508, 1.1042,  ..., 1.6452, 1.8217, 1.6348]],

          [[1.7744, 1.9975, 1.8305,  ..., 1.4077,...
          [[1.3021, 1.7786, 1.4320,  ..., 1.5246, 1.4244, 1.3116]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.9162, 1.9451, 1.6158,  ..., 1.3440, 1.0455, 1.4317],
           [1.5410, 1.7673, 1.2578,  ..., 1.6912, 1...
           [1.8155, 1.7264, 1.2810,  ..., 1.1311, 1.7462, 1.4883]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([40, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: v_copy_rhs_u)
___________________ test_sddmm[idtype0-copy_rhs-u-v-shp2-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'v', rhs_target = 'u'
msg = 'copy_rhs', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb6142280d8>
op = 'copy_rhs'
lhs_data = tensor([[[[1.5606, 1.1362, 1.3621],
          [1.8850, 1.3933, 1.4851],
          [1.5085, 1.4587, 1.9302]]],


      ...7],
          [1.0583, 1.2494, 1.5154],
          [1.3666, 1.0713, 1.8109]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.5447, 1.7274, 1.6373]],

         [[1.7256, 1.8393, 1.2100]],

         [[1.9990, 1.7601, 1.3192]],

    ...

         [[1.1742, 1.6945, 1.2994]],

         [[1.6312, 1.6662, 1.8689]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: v_copy_rhs_u)
___________________ test_sddmm[idtype0-copy_rhs-u-v-shp2-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'v', rhs_target = 'u'
msg = 'copy_rhs', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43e65be58>
op = 'copy_rhs'
lhs_data = tensor([[[[1.5928, 1.8326, 1.5958],
          [1.2508, 1.9936, 1.7170],
          [1.5960, 1.7105, 1.6005]]],


      ...6],
          [1.5671, 1.1695, 1.6745],
          [1.2064, 1.7666, 1.6059]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.1500, 1.5606, 1.7945]],

         [[1.5205, 1.3016, 1.3024]],

         [[1.7308, 1.6442, 1.6868]],

    ...

         [[1.3492, 1.1115, 1.4304]],

         [[1.2105, 1.7126, 1.3066]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([40, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: v_copy_rhs_u)
___________________ test_sddmm[idtype0-copy_rhs-u-v-shp3-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'v', rhs_target = 'u', msg = 'copy_rhs'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb705e15630>
op = 'copy_rhs'
lhs_data = tensor([[1.6741, 1.2751, 1.3530],
        [1.3269, 1.0884, 1.9431],
        [1.0558, 1.8911, 1.8067],
        [1.4280,... 1.1530],
        [1.1834, 1.5423, 1.8515],
        [1.0019, 1.4416, 1.7749]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.1244, 1.7613, 1.5178],
        [1.6354, 1.3367, 1.6800],
        [1.4833, 1.4244, 1.3942],
        [1.4640,... 1.2713],
        [1.9793, 1.9462, 1.2070],
        [1.0194, 1.4198, 1.7450]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: v_copy_rhs_u)
___________________ test_sddmm[idtype0-copy_rhs-u-v-shp3-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'v', rhs_target = 'u', msg = 'copy_rhs'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb7293b9558>
op = 'copy_rhs'
lhs_data = tensor([[1.5230, 1.9641, 1.1275],
        [1.2700, 1.0789, 1.0602],
        [1.3816, 1.2927, 1.4873],
        [1.6465,... 1.1571],
        [1.1102, 1.2344, 1.6782],
        [1.0800, 1.4928, 1.8559]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.2137, 1.7112, 1.7870],
        [1.8556, 1.4908, 1.4403],
        [1.2102, 1.0208, 1.3160],
        [1.7901,... 1.4717],
        [1.6060, 1.7150, 1.1373],
        [1.7181, 1.6127, 1.0733]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([40, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: v_copy_rhs_u)
___________________ test_sddmm[idtype0-copy_rhs-u-v-shp4-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'v', rhs_target = 'u', msg = 'copy_rhs'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb7293826c0>
op = 'copy_rhs'
lhs_data = tensor([[1.9142],
        [1.0099],
        [1.1392],
        [1.8648],
        [1.3030],
        [1.2651],
        [1...262],
        [1.2682],
        [1.9401],
        [1.0999],
        [1.2500]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.2527],
        [1.2116],
        [1.4314],
        [1.9657],
        [1.6812],
        [1.7747],
        [1...255],
        [1.0744],
        [1.0306],
        [1.7682],
        [1.5977]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: v_copy_rhs_u)
___________________ test_sddmm[idtype0-copy_rhs-u-v-shp4-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'v', rhs_target = 'u', msg = 'copy_rhs'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb6146e5ca8>
op = 'copy_rhs'
lhs_data = tensor([[1.2950],
        [1.4792],
        [1.3251],
        [1.6102],
        [1.3907],
        [1.2258],
        [1...712],
        [1.1865],
        [1.8193],
        [1.5823],
        [1.0183]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.2780],
        [1.9811],
        [1.3152],
        [1.2188],
        [1.1036],
        [1.6533],
        [1...467],
        [1.1017],
        [1.3800],
        [1.5287],
        [1.9413]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([40, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: v_copy_rhs_u)
___________________ test_sddmm[idtype0-copy_rhs-u-e-shp0-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...torch.float64), 'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'e', rhs_target = 'u'
msg = 'copy_rhs', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72bfee288>
op = 'copy_rhs'
lhs_data = tensor([[[[[[1.8239],
            [1.1878],
            [1.3932]]],


          [[[1.3601],
            [1.3395],
    ...8]]],


          [[[1.5734],
            [1.4460],
            [1.2006]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.3186]],

           [[1.2395]],

           [[1.6592]]]],



         [[[[1.9979]],

           [[1.528...



         [[[[1.9309]],

           [[1.2179]],

           [[1.7182]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([100, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: e_copy_rhs_u)
___________________ test_sddmm[idtype0-copy_rhs-u-e-shp0-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'e', rhs_target = 'u'
msg = 'copy_rhs', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72bfee480>
op = 'copy_rhs'
lhs_data = tensor([[[[[[1.5589],
            [1.7977],
            [1.8103]]],


          [[[1.0118],
            [1.1558],
    ...7]]],


          [[[1.0052],
            [1.8154],
            [1.8506]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.1023]],

           [[1.6913]],

           [[1.6532]]]],



         [[[[1.1780]],

           [[1.562...



         [[[[1.0236]],

           [[1.1743]],

           [[1.2616]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([300, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: e_copy_rhs_u)
___________________ test_sddmm[idtype0-copy_rhs-u-e-shp1-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...pe=torch.float64), 'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'e', rhs_target = 'u'
msg = 'copy_rhs', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43e660a20>
op = 'copy_rhs'
lhs_data = tensor([[[[[1.0153, 1.1131, 1.1263,  ..., 1.9282, 1.6426, 1.9983]],

          [[1.2138, 1.8278, 1.7283,  ..., 1.3025,...
          [[1.2058, 1.3507, 1.6091,  ..., 1.1750, 1.8748, 1.3622]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.1054, 1.0579, 1.9309,  ..., 1.6054, 1.8784, 1.5365],
           [1.2375, 1.7143, 1.4317,  ..., 1.7698, 1...
           [1.1678, 1.2134, 1.7144,  ..., 1.2311, 1.8233, 1.7984]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([100, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: e_copy_rhs_u)
___________________ test_sddmm[idtype0-copy_rhs-u-e-shp1-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'e', rhs_target = 'u'
msg = 'copy_rhs', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb614220948>
op = 'copy_rhs'
lhs_data = tensor([[[[[1.3883, 1.4904, 1.4914,  ..., 1.3795, 1.7730, 1.0199]],

          [[1.5048, 1.3743, 1.7862,  ..., 1.8226,...
          [[1.6075, 1.0309, 1.9650,  ..., 1.6061, 1.7148, 1.5194]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.7487, 1.0307, 1.6264,  ..., 1.8150, 1.5558, 1.0851],
           [1.8447, 1.5967, 1.5395,  ..., 1.2910, 1...
           [1.4391, 1.6337, 1.1303,  ..., 1.5440, 1.9170, 1.8048]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([300, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: e_copy_rhs_u)
___________________ test_sddmm[idtype0-copy_rhs-u-e-shp2-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...dtype=torch.float64), 'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'e', rhs_target = 'u'
msg = 'copy_rhs', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72b81aaf8>
op = 'copy_rhs'
lhs_data = tensor([[[[1.7178, 1.1627, 1.8853],
          [1.8031, 1.8804, 1.8261],
          [1.5270, 1.4083, 1.5505]]],


      ...6],
          [1.0290, 1.2867, 1.4429],
          [1.3839, 1.0793, 1.0329]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.8848, 1.9976, 1.3741]],

         [[1.4514, 1.0485, 1.7106]],

         [[1.2675, 1.2326, 1.3463]],

    ...

         [[1.0648, 1.0843, 1.4034]],

         [[1.8632, 1.4432, 1.1407]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([100, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: e_copy_rhs_u)
___________________ test_sddmm[idtype0-copy_rhs-u-e-shp2-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'e', rhs_target = 'u'
msg = 'copy_rhs', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72bd608b8>
op = 'copy_rhs'
lhs_data = tensor([[[[1.3162, 1.3183, 1.1196],
          [1.8659, 1.2817, 1.2870],
          [1.2618, 1.1015, 1.1060]]],


      ...3],
          [1.7485, 1.4346, 1.7201],
          [1.6311, 1.8335, 1.5046]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.6275, 1.6177, 1.3027]],

         [[1.1447, 1.2823, 1.9677]],

         [[1.7510, 1.9881, 1.9009]],

    ...

         [[1.4359, 1.7138, 1.1797]],

         [[1.9307, 1.3845, 1.9195]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([300, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: e_copy_rhs_u)
___________________ test_sddmm[idtype0-copy_rhs-u-e-shp3-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'e', rhs_target = 'u', msg = 'copy_rhs'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72ad0abd0>
op = 'copy_rhs'
lhs_data = tensor([[1.6509, 1.0600, 1.5303],
        [1.7395, 1.0640, 1.1099],
        [1.0273, 1.8745, 1.3179],
        [1.2197,... 1.3883],
        [1.3881, 1.2809, 1.9196],
        [1.0514, 1.4639, 1.9897]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.9767, 1.2870, 1.3829],
        [1.8647, 1.9612, 1.7779],
        [1.5523, 1.8938, 1.6575],
        [1.7283,... 1.8652],
        [1.6729, 1.1796, 1.3347],
        [1.1576, 1.0978, 1.6740]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([100, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: e_copy_rhs_u)
___________________ test_sddmm[idtype0-copy_rhs-u-e-shp3-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'e', rhs_target = 'u', msg = 'copy_rhs'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72bffa240>
op = 'copy_rhs'
lhs_data = tensor([[1.2753, 1.7513, 1.1706],
        [1.3801, 1.3064, 1.8480],
        [1.8835, 1.1395, 1.8706],
        [1.9023,... 1.6417],
        [1.3733, 1.3856, 1.9368],
        [1.1690, 1.9252, 1.0803]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.4652, 1.3532, 1.1365],
        [1.7989, 1.8226, 1.9654],
        [1.6318, 1.7333, 1.3551],
        [1.7348,... 1.7347],
        [1.0039, 1.5503, 1.1060],
        [1.9282, 1.1276, 1.1299]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([300, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: e_copy_rhs_u)
___________________ test_sddmm[idtype0-copy_rhs-u-e-shp4-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'e', rhs_target = 'u', msg = 'copy_rhs'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72be3b510>
op = 'copy_rhs'
lhs_data = tensor([[1.4492],
        [1.6716],
        [1.9717],
        [1.7780],
        [1.4089],
        [1.9271],
        [1...393],
        [1.4673],
        [1.5461],
        [1.5897],
        [1.4662]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.7396],
        [1.6818],
        [1.0858],
        [1.2446],
        [1.2372],
        [1.4521],
        [1...548],
        [1.6513],
        [1.2347],
        [1.6637],
        [1.2551]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([100, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: e_copy_rhs_u)
___________________ test_sddmm[idtype0-copy_rhs-u-e-shp4-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'e', rhs_target = 'u', msg = 'copy_rhs'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72be3b8b8>
op = 'copy_rhs'
lhs_data = tensor([[1.7309],
        [1.0474],
        [1.6319],
        [1.7930],
        [1.2076],
        [1.2347],
        [1...548],
        [1.6283],
        [1.4615],
        [1.2977],
        [1.4035]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.2796],
        [1.7705],
        [1.1507],
        [1.2663],
        [1.5487],
        [1.9430],
        [1...344],
        [1.2539],
        [1.6720],
        [1.4927],
        [1.9496]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([300, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: e_copy_rhs_u)
___________________ test_sddmm[idtype0-copy_rhs-v-u-shp0-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': ...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'u', rhs_target = 'v'
msg = 'copy_rhs', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72c0d15a0>
op = 'copy_rhs'
lhs_data = tensor([[[[[[1.2655],
            [1.3285],
            [1.5473]]],


          [[[1.0432],
            [1.1595],
    ...9]]],


          [[[1.1816],
            [1.6608],
            [1.0395]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.9755]],

           [[1.9226]],

           [[1.5961]]]],



         [[[[1.6848]],

           [[1.931...



         [[[[1.2434]],

           [[1.9723]],

           [[1.0561]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: u_copy_rhs_v)
___________________ test_sddmm[idtype0-copy_rhs-v-u-shp0-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'u', rhs_target = 'v'
msg = 'copy_rhs', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb6143b9c60>
op = 'copy_rhs'
lhs_data = tensor([[[[[[1.3446],
            [1.1067],
            [1.8217]]],


          [[[1.0326],
            [1.4345],
    ...1]]],


          [[[1.8608],
            [1.3628],
            [1.0924]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.8326]],

           [[1.4894]],

           [[1.4799]]]],



         [[[[1.6410]],

           [[1.766...



         [[[[1.0691]],

           [[1.9373]],

           [[1.6154]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([40, 4, 1, 3, 1, 1])
SDDMM(message func: u_copy_rhs_v)
___________________ test_sddmm[idtype0-copy_rhs-v-u-shp1-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Sch...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'u', rhs_target = 'v'
msg = 'copy_rhs', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb6146dc990>
op = 'copy_rhs'
lhs_data = tensor([[[[[1.4008, 1.1111, 1.6505,  ..., 1.7863, 1.7594, 1.0311]],

          [[1.7125, 1.8667, 1.1753,  ..., 1.2737,...
          [[1.9052, 1.6000, 1.8308,  ..., 1.5747, 1.1232, 1.5374]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.7962, 1.9209, 1.8046,  ..., 1.9273, 1.5640, 1.6341],
           [1.9088, 1.7584, 1.1204,  ..., 1.0156, 1...
           [1.7442, 1.6289, 1.2363,  ..., 1.8700, 1.9038, 1.6044]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: u_copy_rhs_v)
___________________ test_sddmm[idtype0-copy_rhs-v-u-shp1-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'u', rhs_target = 'v'
msg = 'copy_rhs', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43e661318>
op = 'copy_rhs'
lhs_data = tensor([[[[[1.7605, 1.1117, 1.7899,  ..., 1.6046, 1.0771, 1.3212]],

          [[1.3860, 1.4661, 1.4485,  ..., 1.0035,...
          [[1.4297, 1.1885, 1.6412,  ..., 1.4832, 1.2505, 1.1683]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.9720, 1.6377, 1.7898,  ..., 1.5301, 1.8621, 1.0592],
           [1.1479, 1.1074, 1.8492,  ..., 1.0102, 1...
           [1.2968, 1.7510, 1.1970,  ..., 1.3612, 1.1404, 1.4187]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([40, 1, 3, 7, 7])
SDDMM(message func: u_copy_rhs_v)
___________________ test_sddmm[idtype0-copy_rhs-v-u-shp2-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'u', rhs_target = 'v'
msg = 'copy_rhs', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72bffd240>
op = 'copy_rhs'
lhs_data = tensor([[[[1.7930, 1.9441, 1.0280],
          [1.0159, 1.4705, 1.5682],
          [1.3870, 1.7904, 1.6208]]],


      ...1],
          [1.1109, 1.9330, 1.1112],
          [1.3300, 1.3678, 1.0521]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.6198, 1.5653, 1.9837]],

         [[1.4937, 1.9903, 1.6151]],

         [[1.9125, 1.7720, 1.5975]],

    ...

         [[1.0913, 1.0988, 1.8851]],

         [[1.4452, 1.3879, 1.0482]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: u_copy_rhs_v)
___________________ test_sddmm[idtype0-copy_rhs-v-u-shp2-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'u', rhs_target = 'v'
msg = 'copy_rhs', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72bffd438>
op = 'copy_rhs'
lhs_data = tensor([[[[1.2999, 1.8719, 1.9759],
          [1.3454, 1.2843, 1.2404],
          [1.2455, 1.5270, 1.5517]]],


      ...4],
          [1.9227, 1.5401, 1.9482],
          [1.6184, 1.9407, 1.7119]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.1608, 1.3534, 1.4531]],

         [[1.1858, 1.3279, 1.5701]],

         [[1.9603, 1.0991, 1.3843]],

    ...

         [[1.0894, 1.8011, 1.2018]],

         [[1.1663, 1.3692, 1.2265]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([40, 4, 1, 3])
SDDMM(message func: u_copy_rhs_v)
___________________ test_sddmm[idtype0-copy_rhs-v-u-shp3-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'u', rhs_target = 'v', msg = 'copy_rhs'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72bffd6c0>
op = 'copy_rhs'
lhs_data = tensor([[1.9639, 1.0239, 1.2571],
        [1.7790, 1.3941, 1.0523],
        [1.6039, 1.4590, 1.0021],
        [1.0027,... 1.4497],
        [1.7665, 1.0622, 1.0741],
        [1.3066, 1.6977, 1.1001]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.5712, 1.9363, 1.3104],
        [1.1894, 1.8547, 1.5371],
        [1.9450, 1.9212, 1.5301],
        [1.1920,... 1.4457],
        [1.8432, 1.9448, 1.4763],
        [1.5597, 1.4313, 1.3372]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: u_copy_rhs_v)
___________________ test_sddmm[idtype0-copy_rhs-v-u-shp3-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'u', rhs_target = 'v', msg = 'copy_rhs'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72bffd900>
op = 'copy_rhs'
lhs_data = tensor([[1.6690, 1.2531, 1.1182],
        [1.8775, 1.2700, 1.7103],
        [1.9945, 1.8193, 1.3662],
        [1.5818,... 1.2935],
        [1.4385, 1.3757, 1.6666],
        [1.1990, 1.7968, 1.7922]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.0893, 1.3089, 1.9598],
        [1.9804, 1.5762, 1.6910],
        [1.2411, 1.5835, 1.2766],
        [1.9662,... 1.3477],
        [1.3588, 1.7665, 1.3898],
        [1.2864, 1.0537, 1.5962]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([40, 3])
SDDMM(message func: u_copy_rhs_v)
___________________ test_sddmm[idtype0-copy_rhs-v-u-shp4-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'u', rhs_target = 'v', msg = 'copy_rhs'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb614223438>
op = 'copy_rhs'
lhs_data = tensor([[1.1380],
        [1.8415],
        [1.9156],
        [1.0732],
        [1.8895],
        [1.7877],
        [1...362],
        [1.7241],
        [1.4113],
        [1.5579],
        [1.0428]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.1863],
        [1.7497],
        [1.4809],
        [1.6636],
        [1.7999],
        [1.0074],
        [1...259],
        [1.4150],
        [1.8555],
        [1.6792],
        [1.7574]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: u_copy_rhs_v)
___________________ test_sddmm[idtype0-copy_rhs-v-u-shp4-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'u', rhs_target = 'v', msg = 'copy_rhs'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43e6508b8>
op = 'copy_rhs'
lhs_data = tensor([[1.9175],
        [1.4520],
        [1.3419],
        [1.8285],
        [1.0947],
        [1.0753],
        [1...549],
        [1.9096],
        [1.0978],
        [1.8677],
        [1.7710]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.1751],
        [1.2674],
        [1.9035],
        [1.7487],
        [1.0413],
        [1.1356],
        [1...761],
        [1.3874],
        [1.1438],
        [1.9298],
        [1.5716]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([40, 1])
SDDMM(message func: u_copy_rhs_v)
___________________ test_sddmm[idtype0-copy_rhs-v-e-shp0-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...torch.float64), 'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'e', rhs_target = 'v'
msg = 'copy_rhs', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72aabc0d8>
op = 'copy_rhs'
lhs_data = tensor([[[[[[1.7053],
            [1.6369],
            [1.0544]]],


          [[[1.6400],
            [1.8655],
    ...7]]],


          [[[1.0662],
            [1.1424],
            [1.0509]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.0262]],

           [[1.9492]],

           [[1.0875]]]],



         [[[[1.9354]],

           [[1.814...



         [[[[1.6761]],

           [[1.0841]],

           [[1.9212]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([100, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: e_copy_rhs_v)
___________________ test_sddmm[idtype0-copy_rhs-v-e-shp0-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'e', rhs_target = 'v'
msg = 'copy_rhs', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72b50a900>
op = 'copy_rhs'
lhs_data = tensor([[[[[[1.3431],
            [1.0254],
            [1.0715]]],


          [[[1.0848],
            [1.6962],
    ...9]]],


          [[[1.4190],
            [1.9976],
            [1.6751]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.2238]],

           [[1.9705]],

           [[1.4545]]]],



         [[[[1.1495]],

           [[1.170...



         [[[[1.7975]],

           [[1.2041]],

           [[1.8494]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([300, 1, 2, 1, 3, 1]), rhs shape: torch.Size([40, 4, 1, 3, 1, 1])
SDDMM(message func: e_copy_rhs_v)
___________________ test_sddmm[idtype0-copy_rhs-v-e-shp1-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...pe=torch.float64), 'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'e', rhs_target = 'v'
msg = 'copy_rhs', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72b806b40>
op = 'copy_rhs'
lhs_data = tensor([[[[[1.8245, 1.6041, 1.6158,  ..., 1.0767, 1.6997, 1.5001]],

          [[1.8021, 1.0192, 1.7696,  ..., 1.3858,...
          [[1.9196, 1.0044, 1.1139,  ..., 1.4880, 1.5116, 1.3238]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.2399, 1.4345, 1.6368,  ..., 1.2760, 1.9195, 1.2968],
           [1.3433, 1.6024, 1.0472,  ..., 1.4862, 1...
           [1.3562, 1.8342, 1.4409,  ..., 1.5425, 1.0097, 1.0578]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([100, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: e_copy_rhs_v)
___________________ test_sddmm[idtype0-copy_rhs-v-e-shp1-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'e', rhs_target = 'v'
msg = 'copy_rhs', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43e65bc18>
op = 'copy_rhs'
lhs_data = tensor([[[[[1.6144, 1.8796, 1.7662,  ..., 1.8507, 1.6427, 1.2019]],

          [[1.1855, 1.4936, 1.0348,  ..., 1.3364,...
          [[1.1395, 1.2236, 1.4971,  ..., 1.0583, 1.1720, 1.1739]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.6714, 1.0168, 1.5697,  ..., 1.0543, 1.1245, 1.3481],
           [1.9941, 1.0869, 1.5788,  ..., 1.6151, 1...
           [1.2195, 1.8003, 1.2144,  ..., 1.8822, 1.0519, 1.1884]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([300, 5, 3, 1, 7]), rhs shape: torch.Size([40, 1, 3, 7, 7])
SDDMM(message func: e_copy_rhs_v)
___________________ test_sddmm[idtype0-copy_rhs-v-e-shp2-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...dtype=torch.float64), 'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'e', rhs_target = 'v'
msg = 'copy_rhs', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72be3b8b8>
op = 'copy_rhs'
lhs_data = tensor([[[[1.2909, 1.9960, 1.6195],
          [1.1515, 1.6244, 1.6965],
          [1.4397, 1.8664, 1.5906]]],


      ...0],
          [1.2854, 1.2643, 1.3980],
          [1.4757, 1.6669, 1.6785]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.7555, 1.1718, 1.2527]],

         [[1.0958, 1.5202, 1.2048]],

         [[1.5449, 1.1647, 1.1193]],

    ...

         [[1.5235, 1.2580, 1.4749]],

         [[1.5464, 1.6145, 1.7354]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([100, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: e_copy_rhs_v)
___________________ test_sddmm[idtype0-copy_rhs-v-e-shp2-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'e', rhs_target = 'v'
msg = 'copy_rhs', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72b539240>
op = 'copy_rhs'
lhs_data = tensor([[[[1.8593, 1.7708, 1.9548],
          [1.6215, 1.3200, 1.0841],
          [1.8254, 1.8314, 1.1558]]],


      ...0],
          [1.6998, 1.0275, 1.1913],
          [1.3190, 1.0625, 1.7348]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.9105, 1.4618, 1.4655]],

         [[1.8862, 1.0712, 1.0487]],

         [[1.5455, 1.5366, 1.5574]],

    ...

         [[1.6548, 1.3397, 1.3083]],

         [[1.5714, 1.8381, 1.9192]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([300, 1, 3, 3]), rhs shape: torch.Size([40, 4, 1, 3])
SDDMM(message func: e_copy_rhs_v)
___________________ test_sddmm[idtype0-copy_rhs-v-e-shp3-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'e', rhs_target = 'v', msg = 'copy_rhs'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72be3b8b8>
op = 'copy_rhs'
lhs_data = tensor([[1.9426, 1.8016, 1.2469],
        [1.9191, 1.4651, 1.0822],
        [1.9949, 1.2217, 1.5662],
        [1.8663,... 1.3076],
        [1.7460, 1.1072, 1.4484],
        [1.5081, 1.0971, 1.1520]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.6770, 1.1039, 1.3808],
        [1.3925, 1.0672, 1.7036],
        [1.7111, 1.7083, 1.8968],
        [1.1813,... 1.1263],
        [1.7653, 1.1075, 1.8210],
        [1.1354, 1.8905, 1.8205]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([100, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: e_copy_rhs_v)
___________________ test_sddmm[idtype0-copy_rhs-v-e-shp3-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'e', rhs_target = 'v', msg = 'copy_rhs'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72b800fc0>
op = 'copy_rhs'
lhs_data = tensor([[1.1290, 1.7602, 1.2775],
        [1.4426, 1.7664, 1.2397],
        [1.2406, 1.1369, 1.4241],
        [1.3834,... 1.1420],
        [1.3346, 1.0864, 1.1065],
        [1.7656, 1.0885, 1.5233]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.1855, 1.9841, 1.4062],
        [1.2306, 1.9463, 1.5475],
        [1.9273, 1.3656, 1.4623],
        [1.7881,... 1.2964],
        [1.7121, 1.9983, 1.8245],
        [1.8169, 1.7527, 1.9264]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([300, 3]), rhs shape: torch.Size([40, 3])
SDDMM(message func: e_copy_rhs_v)
___________________ test_sddmm[idtype0-copy_rhs-v-e-shp4-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'e', rhs_target = 'v', msg = 'copy_rhs'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb7293ad3a8>
op = 'copy_rhs'
lhs_data = tensor([[1.2243],
        [1.9902],
        [1.9627],
        [1.6156],
        [1.8780],
        [1.1658],
        [1...756],
        [1.2187],
        [1.1564],
        [1.0408],
        [1.7726]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.2954],
        [1.1140],
        [1.9638],
        [1.7642],
        [1.1212],
        [1.6473],
        [1...196],
        [1.8439],
        [1.0766],
        [1.9331],
        [1.0589]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([100, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: e_copy_rhs_v)
___________________ test_sddmm[idtype0-copy_rhs-v-e-shp4-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'e', rhs_target = 'v', msg = 'copy_rhs'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72bffaaf8>
op = 'copy_rhs'
lhs_data = tensor([[1.5586],
        [1.7996],
        [1.1192],
        [1.0297],
        [1.1971],
        [1.7611],
        [1...407],
        [1.8142],
        [1.4639],
        [1.2579],
        [1.0431]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.6626],
        [1.4837],
        [1.3303],
        [1.4495],
        [1.6324],
        [1.8705],
        [1...601],
        [1.5088],
        [1.5048],
        [1.0666],
        [1.0236]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([300, 1]), rhs shape: torch.Size([40, 1])
SDDMM(message func: e_copy_rhs_v)
___________________ test_sddmm[idtype0-copy_rhs-e-u-shp0-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': ...torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'u', rhs_target = 'e'
msg = 'copy_rhs', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72ad0b828>
op = 'copy_rhs'
lhs_data = tensor([[[[[[1.2755],
            [1.3966],
            [1.4058]]],


          [[[1.0125],
            [1.8757],
    ...5]]],


          [[[1.7288],
            [1.8508],
            [1.7447]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.0562]],

           [[1.8950]],

           [[1.3700]]]],



         [[[[1.4632]],

           [[1.477...



         [[[[1.3227]],

           [[1.3506]],

           [[1.7976]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([100, 4, 1, 3, 1, 1])
SDDMM(message func: u_copy_rhs_e)
___________________ test_sddmm[idtype0-copy_rhs-e-u-shp0-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'u', rhs_target = 'e'
msg = 'copy_rhs', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72b7cd5a0>
op = 'copy_rhs'
lhs_data = tensor([[[[[[1.9466],
            [1.8593],
            [1.9934]]],


          [[[1.6803],
            [1.7303],
    ...4]]],


          [[[1.9303],
            [1.5316],
            [1.8979]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.1880]],

           [[1.6289]],

           [[1.8945]]]],



         [[[[1.3420]],

           [[1.523...



         [[[[1.6175]],

           [[1.6101]],

           [[1.6973]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([300, 4, 1, 3, 1, 1])
SDDMM(message func: u_copy_rhs_e)
___________________ test_sddmm[idtype0-copy_rhs-e-u-shp1-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Sch...pe=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'u', rhs_target = 'e'
msg = 'copy_rhs', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72b53d510>
op = 'copy_rhs'
lhs_data = tensor([[[[[1.6012, 1.8648, 1.7628,  ..., 1.8378, 1.8980, 1.1626]],

          [[1.3475, 1.1848, 1.9790,  ..., 1.1982,...
          [[1.8999, 1.2714, 1.0447,  ..., 1.9901, 1.3769, 1.3870]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.3802, 1.4667, 1.5615,  ..., 1.9304, 1.5874, 1.4009],
           [1.7752, 1.8350, 1.7122,  ..., 1.9160, 1...
           [1.7238, 1.8195, 1.0161,  ..., 1.3555, 1.0746, 1.6444]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([100, 1, 3, 7, 7])
SDDMM(message func: u_copy_rhs_e)
___________________ test_sddmm[idtype0-copy_rhs-e-u-shp1-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'u', rhs_target = 'e'
msg = 'copy_rhs', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72b53da20>
op = 'copy_rhs'
lhs_data = tensor([[[[[1.0980, 1.4826, 1.6943,  ..., 1.7534, 1.1136, 1.9065]],

          [[1.3097, 1.5189, 1.0925,  ..., 1.8635,...
          [[1.5106, 1.7584, 1.0918,  ..., 1.8648, 1.8079, 1.4356]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.3221, 1.2463, 1.5476,  ..., 1.9612, 1.5852, 1.7656],
           [1.2738, 1.9087, 1.9461,  ..., 1.6003, 1...
           [1.6153, 1.7940, 1.8366,  ..., 1.3046, 1.1975, 1.3411]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([300, 1, 3, 7, 7])
SDDMM(message func: u_copy_rhs_e)
___________________ test_sddmm[idtype0-copy_rhs-e-u-shp2-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme...dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'u', rhs_target = 'e'
msg = 'copy_rhs', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb6146f85e8>
op = 'copy_rhs'
lhs_data = tensor([[[[1.5096, 1.8434, 1.0132],
          [1.7790, 1.9061, 1.3367],
          [1.8423, 1.8252, 1.5294]]],


      ...4],
          [1.5003, 1.7620, 1.9793],
          [1.0718, 1.4073, 1.5609]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.5961, 1.0334, 1.1779]],

         [[1.6756, 1.2701, 1.3071]],

         [[1.1770, 1.9389, 1.9999]],

    ...

         [[1.0079, 1.1828, 1.6996]],

         [[1.2599, 1.1376, 1.9977]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([100, 4, 1, 3])
SDDMM(message func: u_copy_rhs_e)
___________________ test_sddmm[idtype0-copy_rhs-e-u-shp2-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'u', rhs_target = 'e'
msg = 'copy_rhs', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72ad173a8>
op = 'copy_rhs'
lhs_data = tensor([[[[1.2530, 1.9977, 1.1483],
          [1.2165, 1.3894, 1.9098],
          [1.1395, 1.6693, 1.3714]]],


      ...1],
          [1.0164, 1.7095, 1.7373],
          [1.2419, 1.5735, 1.9993]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.5349, 1.8008, 1.0834]],

         [[1.6861, 1.2888, 1.6929]],

         [[1.2201, 1.6257, 1.3611]],

    ...

         [[1.5543, 1.0197, 1.9412]],

         [[1.8641, 1.6017, 1.7317]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([300, 4, 1, 3])
SDDMM(message func: u_copy_rhs_e)
___________________ test_sddmm[idtype0-copy_rhs-e-u-shp3-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'u', rhs_target = 'e', msg = 'copy_rhs'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb705b518b8>
op = 'copy_rhs'
lhs_data = tensor([[1.2006, 1.5496, 1.0844],
        [1.5684, 1.3320, 1.9077],
        [1.5733, 1.1821, 1.6386],
        [1.5835,... 1.1900],
        [1.2296, 1.6496, 1.9405],
        [1.5529, 1.8905, 1.1272]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.7064, 1.9659, 1.0650],
        [1.6080, 1.1417, 1.1463],
        [1.5827, 1.8075, 1.5448],
        [1.0643,... 1.5789],
        [1.2039, 1.0200, 1.7040],
        [1.2967, 1.6156, 1.9675]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([100, 3])
SDDMM(message func: u_copy_rhs_e)
___________________ test_sddmm[idtype0-copy_rhs-e-u-shp3-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'u', rhs_target = 'e', msg = 'copy_rhs'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb7060f2240>
op = 'copy_rhs'
lhs_data = tensor([[1.6834, 1.7368, 1.3248],
        [1.0454, 1.5676, 1.1463],
        [1.2917, 1.3883, 1.3478],
        [1.7298,... 1.7546],
        [1.1306, 1.2643, 1.9414],
        [1.8721, 1.7117, 1.2132]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.9712, 1.4887, 1.7183],
        [1.5152, 1.1987, 1.4142],
        [1.9554, 1.7151, 1.3356],
        [1.9554,... 1.0403],
        [1.6688, 1.9425, 1.2210],
        [1.7642, 1.0171, 1.9990]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([300, 3])
SDDMM(message func: u_copy_rhs_e)
___________________ test_sddmm[idtype0-copy_rhs-e-u-shp4-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'u', rhs_target = 'e', msg = 'copy_rhs'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb457b8cf78>
op = 'copy_rhs'
lhs_data = tensor([[1.8136],
        [1.0292],
        [1.3249],
        [1.7070],
        [1.3617],
        [1.4978],
        [1...648],
        [1.1955],
        [1.9195],
        [1.5061],
        [1.0085]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.7459],
        [1.6050],
        [1.4002],
        [1.1707],
        [1.2867],
        [1.4445],
        [1...594],
        [1.5399],
        [1.9134],
        [1.7133],
        [1.8590]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([100, 1])
SDDMM(message func: u_copy_rhs_e)
___________________ test_sddmm[idtype0-copy_rhs-e-u-shp4-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'u', rhs_target = 'e', msg = 'copy_rhs'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72b528240>
op = 'copy_rhs'
lhs_data = tensor([[1.5960],
        [1.2116],
        [1.0118],
        [1.3420],
        [1.7355],
        [1.2537],
        [1...321],
        [1.3436],
        [1.7690],
        [1.3263],
        [1.2679]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.0193],
        [1.8553],
        [1.5332],
        [1.2373],
        [1.7843],
        [1.7702],
        [1...953],
        [1.1753],
        [1.9254],
        [1.4417],
        [1.9366]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([300, 1])
SDDMM(message func: u_copy_rhs_e)
___________________ test_sddmm[idtype0-copy_rhs-e-v-shp0-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': ...torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'v', rhs_target = 'e'
msg = 'copy_rhs', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb457b8c3f0>
op = 'copy_rhs'
lhs_data = tensor([[[[[[1.0257],
            [1.5564],
            [1.6735]]],


          [[[1.7574],
            [1.1896],
    ...6]]],


          [[[1.1870],
            [1.6611],
            [1.5146]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.9962]],

           [[1.6530]],

           [[1.2550]]]],



         [[[[1.6330]],

           [[1.668...



         [[[[1.6257]],

           [[1.0043]],

           [[1.6858]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([100, 4, 1, 3, 1, 1])
SDDMM(message func: v_copy_rhs_e)
___________________ test_sddmm[idtype0-copy_rhs-e-v-shp0-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'v', rhs_target = 'e'
msg = 'copy_rhs', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb614383d38>
op = 'copy_rhs'
lhs_data = tensor([[[[[[1.2068],
            [1.7089],
            [1.6667]]],


          [[[1.9400],
            [1.5986],
    ...6]]],


          [[[1.6033],
            [1.6855],
            [1.1678]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.1027]],

           [[1.8473]],

           [[1.1386]]]],



         [[[[1.5080]],

           [[1.411...



         [[[[1.2864]],

           [[1.7288]],

           [[1.9489]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([40, 1, 2, 1, 3, 1]), rhs shape: torch.Size([300, 4, 1, 3, 1, 1])
SDDMM(message func: v_copy_rhs_e)
___________________ test_sddmm[idtype0-copy_rhs-e-v-shp1-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Sch...pe=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'v', rhs_target = 'e'
msg = 'copy_rhs', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb705b51948>
op = 'copy_rhs'
lhs_data = tensor([[[[[1.6091, 1.6454, 1.4716,  ..., 1.8616, 1.6843, 1.9600]],

          [[1.5139, 1.3314, 1.0543,  ..., 1.2685,...
          [[1.2620, 1.6248, 1.2271,  ..., 1.6489, 1.3613, 1.5916]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.8039, 1.7798, 1.4873,  ..., 1.7127, 1.1045, 1.6116],
           [1.9957, 1.0704, 1.0743,  ..., 1.1543, 1...
           [1.3298, 1.7809, 1.7526,  ..., 1.5528, 1.0602, 1.7391]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([100, 1, 3, 7, 7])
SDDMM(message func: v_copy_rhs_e)
___________________ test_sddmm[idtype0-copy_rhs-e-v-shp1-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'v', rhs_target = 'e'
msg = 'copy_rhs', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb614171ca8>
op = 'copy_rhs'
lhs_data = tensor([[[[[1.4636, 1.8160, 1.7627,  ..., 1.4711, 1.8838, 1.0761]],

          [[1.3038, 1.3331, 1.5151,  ..., 1.9039,...
          [[1.0296, 1.4721, 1.8574,  ..., 1.2210, 1.4098, 1.0417]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.9172, 1.9872, 1.5898,  ..., 1.8814, 1.4536, 1.0514],
           [1.0999, 1.0333, 1.8741,  ..., 1.4705, 1...
           [1.4220, 1.5655, 1.3559,  ..., 1.9869, 1.7374, 1.9699]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([40, 5, 3, 1, 7]), rhs shape: torch.Size([300, 1, 3, 7, 7])
SDDMM(message func: v_copy_rhs_e)
___________________ test_sddmm[idtype0-copy_rhs-e-v-shp2-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme...dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'v', rhs_target = 'e'
msg = 'copy_rhs', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72ad14bd0>
op = 'copy_rhs'
lhs_data = tensor([[[[1.2458, 1.3479, 1.8402],
          [1.3437, 1.6396, 1.3057],
          [1.5878, 1.5778, 1.4103]]],


      ...0],
          [1.6906, 1.3375, 1.2168],
          [1.2763, 1.5508, 1.4283]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.6273, 1.7621, 1.2358]],

         [[1.5897, 1.5186, 1.0311]],

         [[1.0716, 1.3854, 1.5894]],

    ...

         [[1.1514, 1.1048, 1.0881]],

         [[1.4964, 1.8147, 1.8714]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([100, 4, 1, 3])
SDDMM(message func: v_copy_rhs_e)
___________________ test_sddmm[idtype0-copy_rhs-e-v-shp2-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'v', rhs_target = 'e'
msg = 'copy_rhs', idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb6146f5d38>
op = 'copy_rhs'
lhs_data = tensor([[[[1.1152, 1.4917, 1.7607],
          [1.4037, 1.0754, 1.5023],
          [1.8363, 1.8436, 1.2381]]],


      ...7],
          [1.6994, 1.9483, 1.8202],
          [1.4134, 1.2867, 1.4337]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.8554, 1.9332, 1.9851]],

         [[1.0861, 1.9409, 1.9667]],

         [[1.2474, 1.6816, 1.0463]],

    ...

         [[1.4845, 1.2568, 1.5520]],

         [[1.5971, 1.9989, 1.9337]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([40, 1, 3, 3]), rhs shape: torch.Size([300, 4, 1, 3])
SDDMM(message func: v_copy_rhs_e)
___________________ test_sddmm[idtype0-copy_rhs-e-v-shp3-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'v', rhs_target = 'e', msg = 'copy_rhs'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72b526240>
op = 'copy_rhs'
lhs_data = tensor([[1.5360, 1.4495, 1.7952],
        [1.9690, 1.5946, 1.2217],
        [1.1777, 1.3323, 1.7600],
        [1.1426,... 1.9804],
        [1.3878, 1.7394, 1.1338],
        [1.7572, 1.2068, 1.9435]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.6338, 1.6342, 1.1176],
        [1.4077, 1.7846, 1.2054],
        [1.4001, 1.0529, 1.6590],
        [1.4023,... 1.9762],
        [1.5035, 1.7848, 1.5263],
        [1.0739, 1.1247, 1.6939]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([100, 3])
SDDMM(message func: v_copy_rhs_e)
___________________ test_sddmm[idtype0-copy_rhs-e-v-shp3-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'v', rhs_target = 'e', msg = 'copy_rhs'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72b526480>
op = 'copy_rhs'
lhs_data = tensor([[1.0289, 1.1402, 1.4048],
        [1.7235, 1.5314, 1.9122],
        [1.5922, 1.1451, 1.2672],
        [1.6200,... 1.3440],
        [1.1027, 1.8939, 1.0087],
        [1.4126, 1.9978, 1.3046]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.6102, 1.1071, 1.7786],
        [1.7705, 1.8336, 1.0563],
        [1.0936, 1.6735, 1.8456],
        [1.1112,... 1.3709],
        [1.3424, 1.6053, 1.8940],
        [1.4987, 1.4609, 1.6209]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([40, 3]), rhs shape: torch.Size([300, 3])
SDDMM(message func: v_copy_rhs_e)
___________________ test_sddmm[idtype0-copy_rhs-e-v-shp4-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'v', rhs_target = 'e', msg = 'copy_rhs'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb457bbed80>
op = 'copy_rhs'
lhs_data = tensor([[1.0366],
        [1.6394],
        [1.2535],
        [1.0064],
        [1.8423],
        [1.1767],
        [1...192],
        [1.8957],
        [1.4372],
        [1.5105],
        [1.8043]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.6449],
        [1.6147],
        [1.7855],
        [1.5663],
        [1.8864],
        [1.0178],
        [1...278],
        [1.5245],
        [1.1148],
        [1.5242],
        [1.5755]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)})
torch.int32
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([100, 1])
SDDMM(message func: v_copy_rhs_e)
___________________ test_sddmm[idtype0-copy_rhs-e-v-shp4-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'v', rhs_target = 'e', msg = 'copy_rhs'
idtype = torch.int32

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43e66a1b0>
op = 'copy_rhs'
lhs_data = tensor([[1.4155],
        [1.8122],
        [1.1344],
        [1.8515],
        [1.6163],
        [1.4961],
        [1...667],
        [1.7594],
        [1.1312],
        [1.9622],
        [1.7887]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.1207],
        [1.4096],
        [1.2097],
        [1.5371],
        [1.9686],
        [1.5994],
        [1...753],
        [1.5654],
        [1.0716],
        [1.3416],
        [1.5813]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int32
lhs shape: torch.Size([40, 1]), rhs shape: torch.Size([300, 1])
SDDMM(message func: v_copy_rhs_e)
_____________________ test_sddmm[idtype1-add-u-v-shp0-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': ...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'v', rhs_target = 'u'
msg = 'add', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add'
lhs_data = tensor([[[[[[1.0538],
            [1.0238],
            [1.8920]]],


          [[[1.5129],
            [1.6785],
    ...4]]],


          [[[1.7665],
            [1.6139],
            [1.8177]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.2399]],

           [[1.7928]],

           [[1.9129]]]],



         [[[[1.5293]],

           [[1.198...



         [[[[1.0765]],

           [[1.2754]],

           [[1.9940]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: v_add_u)
_____________________ test_sddmm[idtype1-add-u-v-shp0-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'v', rhs_target = 'u'
msg = 'add', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add'
lhs_data = tensor([[[[[[1.3823],
            [1.7771],
            [1.3588]]],


          [[[1.0440],
            [1.9118],
    ...9]]],


          [[[1.0188],
            [1.0956],
            [1.3980]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.2014]],

           [[1.5210]],

           [[1.2375]]]],



         [[[[1.4609]],

           [[1.822...



         [[[[1.5467]],

           [[1.1544]],

           [[1.6318]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([40, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: v_add_u)
_____________________ test_sddmm[idtype1-add-u-v-shp1-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Sch...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'v', rhs_target = 'u'
msg = 'add', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add'
lhs_data = tensor([[[[[1.6195, 1.4338, 1.8764,  ..., 1.5938, 1.6521, 1.3412]],

          [[1.1294, 1.1809, 1.4592,  ..., 1.0061,...
          [[1.1553, 1.3660, 1.1181,  ..., 1.1456, 1.3594, 1.2059]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.0602, 1.7022, 1.4194,  ..., 1.3384, 1.2940, 1.6516],
           [1.7840, 1.7304, 1.4310,  ..., 1.7326, 1...
           [1.1816, 1.7416, 1.0626,  ..., 1.7666, 1.6093, 1.6890]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: v_add_u)
_____________________ test_sddmm[idtype1-add-u-v-shp1-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'v', rhs_target = 'u'
msg = 'add', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add'
lhs_data = tensor([[[[[1.1600, 1.3202, 1.9316,  ..., 1.5596, 1.7538, 1.3835]],

          [[1.7017, 1.7116, 1.0206,  ..., 1.0603,...
          [[1.5541, 1.4394, 1.2179,  ..., 1.5918, 1.1283, 1.7483]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.1217, 1.5164, 1.1071,  ..., 1.1877, 1.2064, 1.0435],
           [1.8529, 1.0346, 1.2597,  ..., 1.4338, 1...
           [1.8697, 1.9658, 1.0314,  ..., 1.6742, 1.9173, 1.7853]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([40, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: v_add_u)
_____________________ test_sddmm[idtype1-add-u-v-shp2-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'v', rhs_target = 'u', msg = 'add'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add'
lhs_data = tensor([[[[1.7166, 1.1163, 1.0126],
          [1.7361, 1.6406, 1.4116],
          [1.3560, 1.6306, 1.2666]]],


      ...0],
          [1.2746, 1.5040, 1.0010],
          [1.9613, 1.5678, 1.9850]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.1499, 1.1163, 1.8378]],

         [[1.6988, 1.1679, 1.3395]],

         [[1.6171, 1.5862, 1.1393]],

    ...

         [[1.2030, 1.5457, 1.2894]],

         [[1.0679, 1.1847, 1.3448]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: v_add_u)
_____________________ test_sddmm[idtype1-add-u-v-shp2-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'v', rhs_target = 'u', msg = 'add'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add'
lhs_data = tensor([[[[1.5322, 1.6398, 1.2337],
          [1.4101, 1.3125, 1.8689],
          [1.2051, 1.7237, 1.0797]]],


      ...8],
          [1.9485, 1.2556, 1.4809],
          [1.7045, 1.9069, 1.8979]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.8330, 1.8448, 1.8374]],

         [[1.3421, 1.9555, 1.2751]],

         [[1.2130, 1.4765, 1.0832]],

    ...

         [[1.8049, 1.0599, 1.7277]],

         [[1.9871, 1.0453, 1.2419]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([40, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: v_add_u)
_____________________ test_sddmm[idtype1-add-u-v-shp3-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'v', rhs_target = 'u', msg = 'add'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add'
lhs_data = tensor([[1.4041, 1.2498, 1.9160],
        [1.7599, 1.7851, 1.8746],
        [1.1167, 1.7918, 1.3392],
        [1.0029,... 1.7866],
        [1.6009, 1.9302, 1.0046],
        [1.4175, 1.7351, 1.0155]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.7312, 1.0880, 1.6387],
        [1.5006, 1.7762, 1.7168],
        [1.6555, 1.3096, 1.8716],
        [1.3026,... 1.0506],
        [1.6509, 1.7399, 1.1844],
        [1.7028, 1.7442, 1.2610]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: v_add_u)
_____________________ test_sddmm[idtype1-add-u-v-shp3-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'v', rhs_target = 'u', msg = 'add'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add'
lhs_data = tensor([[1.8832, 1.6151, 1.3374],
        [1.8579, 1.2571, 1.0948],
        [1.8246, 1.8354, 1.3758],
        [1.3532,... 1.2839],
        [1.7480, 1.5352, 1.9231],
        [1.7127, 1.9099, 1.7350]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.7571, 1.0550, 1.9447],
        [1.8379, 1.8479, 1.4755],
        [1.0972, 1.9695, 1.9028],
        [1.2951,... 1.4620],
        [1.4834, 1.3287, 1.9388],
        [1.6115, 1.1575, 1.4562]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([40, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: v_add_u)
_____________________ test_sddmm[idtype1-add-u-v-shp4-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'v', rhs_target = 'u', msg = 'add'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add'
lhs_data = tensor([[1.8462],
        [1.4939],
        [1.4392],
        [1.3176],
        [1.1358],
        [1.0058],
        [1...643],
        [1.8248],
        [1.9922],
        [1.2181],
        [1.4111]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.8180],
        [1.0905],
        [1.3669],
        [1.3680],
        [1.7346],
        [1.2131],
        [1...768],
        [1.0198],
        [1.1849],
        [1.1895],
        [1.7659]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: v_add_u)
_____________________ test_sddmm[idtype1-add-u-v-shp4-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'v', rhs_target = 'u', msg = 'add'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add'
lhs_data = tensor([[1.6881],
        [1.7776],
        [1.4317],
        [1.4669],
        [1.5663],
        [1.7057],
        [1...930],
        [1.5205],
        [1.3667],
        [1.6035],
        [1.3891]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.5490],
        [1.0649],
        [1.0893],
        [1.9682],
        [1.9800],
        [1.8304],
        [1...873],
        [1.1750],
        [1.2546],
        [1.4728],
        [1.5281]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([40, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: v_add_u)
_____________________ test_sddmm[idtype1-add-u-e-shp0-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...torch.float64), 'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'e', rhs_target = 'u'
msg = 'add', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add'
lhs_data = tensor([[[[[[1.6496],
            [1.8398],
            [1.0663]]],


          [[[1.8504],
            [1.2146],
    ...3]]],


          [[[1.8121],
            [1.6003],
            [1.2655]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.3107]],

           [[1.5943]],

           [[1.7119]]]],



         [[[[1.0288]],

           [[1.702...



         [[[[1.2305]],

           [[1.6976]],

           [[1.6779]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([100, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: e_add_u)
_____________________ test_sddmm[idtype1-add-u-e-shp0-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'e', rhs_target = 'u'
msg = 'add', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add'
lhs_data = tensor([[[[[[1.2931],
            [1.6969],
            [1.5546]]],


          [[[1.5194],
            [1.4346],
    ...4]]],


          [[[1.6681],
            [1.6534],
            [1.1137]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.1053]],

           [[1.8056]],

           [[1.1024]]]],



         [[[[1.5440]],

           [[1.824...



         [[[[1.8759]],

           [[1.0446]],

           [[1.4352]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([300, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: e_add_u)
_____________________ test_sddmm[idtype1-add-u-e-shp1-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...pe=torch.float64), 'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'e', rhs_target = 'u'
msg = 'add', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add'
lhs_data = tensor([[[[[1.7853, 1.3318, 1.8380,  ..., 1.9571, 1.9454, 1.9520]],

          [[1.5460, 1.3612, 1.5763,  ..., 1.2955,...
          [[1.7257, 1.9436, 1.3274,  ..., 1.4012, 1.4958, 1.9732]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.6728, 1.1462, 1.1225,  ..., 1.0103, 1.7624, 1.2612],
           [1.3764, 1.4709, 1.1657,  ..., 1.8701, 1...
           [1.4300, 1.3555, 1.2131,  ..., 1.9032, 1.2487, 1.7011]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([100, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: e_add_u)
_____________________ test_sddmm[idtype1-add-u-e-shp1-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'e', rhs_target = 'u'
msg = 'add', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add'
lhs_data = tensor([[[[[1.3147, 1.3936, 1.4539,  ..., 1.9280, 1.5986, 1.1598]],

          [[1.8093, 1.8279, 1.1898,  ..., 1.6844,...
          [[1.1054, 1.1837, 1.2861,  ..., 1.1124, 1.3360, 1.2909]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.5682, 1.7346, 1.3543,  ..., 1.0799, 1.5630, 1.0068],
           [1.3162, 1.0977, 1.5009,  ..., 1.5082, 1...
           [1.9870, 1.7976, 1.0759,  ..., 1.8762, 1.9021, 1.8766]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([300, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: e_add_u)
_____________________ test_sddmm[idtype1-add-u-e-shp2-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...dtype=torch.float64), 'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'e', rhs_target = 'u', msg = 'add'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add'
lhs_data = tensor([[[[1.9039, 1.2088, 1.8604],
          [1.2746, 1.2420, 1.3948],
          [1.0041, 1.1837, 1.9837]]],


      ...8],
          [1.6212, 1.2659, 1.8469],
          [1.0423, 1.0793, 1.0338]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.0473, 1.8674, 1.8347]],

         [[1.1869, 1.5923, 1.1536]],

         [[1.4307, 1.6849, 1.7813]],

    ...

         [[1.2519, 1.4271, 1.1986]],

         [[1.5948, 1.4978, 1.0726]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([100, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: e_add_u)
_____________________ test_sddmm[idtype1-add-u-e-shp2-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'e', rhs_target = 'u', msg = 'add'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add'
lhs_data = tensor([[[[1.4492, 1.0472, 1.7897],
          [1.7059, 1.4095, 1.1656],
          [1.4365, 1.7435, 1.7607]]],


      ...3],
          [1.0161, 1.1641, 1.2765],
          [1.9876, 1.9702, 1.3058]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.4381, 1.5484, 1.5701]],

         [[1.6824, 1.3452, 1.8506]],

         [[1.5706, 1.9608, 1.0699]],

    ...

         [[1.4500, 1.2843, 1.7690]],

         [[1.5950, 1.6498, 1.5661]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([300, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: e_add_u)
_____________________ test_sddmm[idtype1-add-u-e-shp3-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'e', rhs_target = 'u', msg = 'add'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add'
lhs_data = tensor([[1.3227, 1.2291, 1.3393],
        [1.7868, 1.9504, 1.1512],
        [1.6298, 1.3822, 1.3393],
        [1.8321,... 1.4614],
        [1.2820, 1.3785, 1.1502],
        [1.1314, 1.6627, 1.4948]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.7733, 1.9940, 1.8581],
        [1.7789, 1.4331, 1.0407],
        [1.8122, 1.0511, 1.9251],
        [1.9997,... 1.8276],
        [1.6763, 1.0636, 1.0683],
        [1.5853, 1.3156, 1.3697]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([100, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: e_add_u)
_____________________ test_sddmm[idtype1-add-u-e-shp3-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'e', rhs_target = 'u', msg = 'add'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add'
lhs_data = tensor([[1.6364, 1.8343, 1.0332],
        [1.3506, 1.8756, 1.2626],
        [1.9546, 1.4373, 1.4615],
        [1.8129,... 1.9634],
        [1.5978, 1.3524, 1.6716],
        [1.9852, 1.5549, 1.0187]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.5000, 1.4774, 1.7995],
        [1.2240, 1.2854, 1.4324],
        [1.5105, 1.0868, 1.1815],
        [1.6374,... 1.0983],
        [1.1758, 1.8476, 1.6485],
        [1.4133, 1.9991, 1.5480]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([300, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: e_add_u)
_____________________ test_sddmm[idtype1-add-u-e-shp4-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'e', rhs_target = 'u', msg = 'add'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add'
lhs_data = tensor([[1.6351],
        [1.9366],
        [1.3573],
        [1.5476],
        [1.2673],
        [1.3948],
        [1...032],
        [1.6637],
        [1.9603],
        [1.8138],
        [1.8834]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.9794],
        [1.8160],
        [1.9009],
        [1.5220],
        [1.8801],
        [1.1373],
        [1...316],
        [1.4479],
        [1.9844],
        [1.6537],
        [1.2113]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([100, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: e_add_u)
_____________________ test_sddmm[idtype1-add-u-e-shp4-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'e', rhs_target = 'u', msg = 'add'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add'
lhs_data = tensor([[1.1618],
        [1.1798],
        [1.2742],
        [1.5499],
        [1.8895],
        [1.9858],
        [1...835],
        [1.4085],
        [1.8906],
        [1.9375],
        [1.3631]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.1687],
        [1.6901],
        [1.2331],
        [1.3742],
        [1.0971],
        [1.2780],
        [1...437],
        [1.6090],
        [1.7226],
        [1.9370],
        [1.7529]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([300, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: e_add_u)
_____________________ test_sddmm[idtype1-add-v-u-shp0-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': ...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'u', rhs_target = 'v'
msg = 'add', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add'
lhs_data = tensor([[[[[[1.2759],
            [1.3763],
            [1.8254]]],


          [[[1.1036],
            [1.8550],
    ...7]]],


          [[[1.2511],
            [1.6386],
            [1.6516]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.7965]],

           [[1.8408]],

           [[1.1849]]]],



         [[[[1.6544]],

           [[1.566...



         [[[[1.5783]],

           [[1.2274]],

           [[1.5214]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: u_add_v)
_____________________ test_sddmm[idtype1-add-v-u-shp0-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'u', rhs_target = 'v'
msg = 'add', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add'
lhs_data = tensor([[[[[[1.1234],
            [1.1129],
            [1.0583]]],


          [[[1.9270],
            [1.2250],
    ...0]]],


          [[[1.5501],
            [1.2429],
            [1.0326]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.2150]],

           [[1.2113]],

           [[1.0876]]]],



         [[[[1.6120]],

           [[1.754...



         [[[[1.8177]],

           [[1.9799]],

           [[1.9595]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([40, 4, 1, 3, 1, 1])
SDDMM(message func: u_add_v)
_____________________ test_sddmm[idtype1-add-v-u-shp1-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Sch...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'u', rhs_target = 'v'
msg = 'add', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add'
lhs_data = tensor([[[[[1.9616, 1.3250, 1.1694,  ..., 1.2429, 1.2902, 1.1506]],

          [[1.2153, 1.8921, 1.4752,  ..., 1.3573,...
          [[1.2122, 1.1336, 1.4438,  ..., 1.9343, 1.3701, 1.1420]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.3610, 1.5333, 1.8000,  ..., 1.0707, 1.4869, 1.3097],
           [1.6478, 1.8728, 1.2781,  ..., 1.5580, 1...
           [1.4602, 1.3757, 1.5316,  ..., 1.7127, 1.0436, 1.2323]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: u_add_v)
_____________________ test_sddmm[idtype1-add-v-u-shp1-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'u', rhs_target = 'v'
msg = 'add', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add'
lhs_data = tensor([[[[[1.5992, 1.7324, 1.7361,  ..., 1.2902, 1.6432, 1.4127]],

          [[1.8964, 1.7048, 1.9325,  ..., 1.4847,...
          [[1.4973, 1.8664, 1.9865,  ..., 1.2186, 1.5518, 1.6210]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.1072, 1.2897, 1.6140,  ..., 1.8711, 1.7574, 1.8624],
           [1.1230, 1.1707, 1.3297,  ..., 1.6470, 1...
           [1.2288, 1.4526, 1.1731,  ..., 1.3688, 1.7685, 1.0778]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([40, 1, 3, 7, 7])
SDDMM(message func: u_add_v)
_____________________ test_sddmm[idtype1-add-v-u-shp2-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'u', rhs_target = 'v', msg = 'add'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add'
lhs_data = tensor([[[[1.8757, 1.3527, 1.8474],
          [1.8032, 1.9593, 1.2079],
          [1.1150, 1.1603, 1.2374]]],


      ...0],
          [1.1163, 1.0629, 1.2497],
          [1.0586, 1.7478, 1.3502]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.1893, 1.1923, 1.2180]],

         [[1.2348, 1.1854, 1.4189]],

         [[1.9971, 1.3223, 1.5105]],

    ...

         [[1.4583, 1.2554, 1.4599]],

         [[1.3967, 1.5470, 1.4219]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: u_add_v)
_____________________ test_sddmm[idtype1-add-v-u-shp2-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'u', rhs_target = 'v', msg = 'add'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add'
lhs_data = tensor([[[[1.2782, 1.6104, 1.2671],
          [1.4370, 1.1082, 1.7245],
          [1.9933, 1.9310, 1.5772]]],


      ...7],
          [1.1351, 1.3879, 1.7876],
          [1.5961, 1.6663, 1.2265]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.8985, 1.9589, 1.4105]],

         [[1.5008, 1.3416, 1.6445]],

         [[1.3215, 1.5816, 1.4917]],

    ...

         [[1.0602, 1.2194, 1.3892]],

         [[1.7730, 1.2385, 1.0433]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([40, 4, 1, 3])
SDDMM(message func: u_add_v)
_____________________ test_sddmm[idtype1-add-v-u-shp3-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'u', rhs_target = 'v', msg = 'add'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add'
lhs_data = tensor([[1.2645, 1.1337, 1.7420],
        [1.1715, 1.6651, 1.4458],
        [1.4852, 1.0119, 1.0427],
        [1.3778,... 1.9359],
        [1.1104, 1.1656, 1.7260],
        [1.2002, 1.0001, 1.5048]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.6242, 1.0726, 1.6203],
        [1.9408, 1.3224, 1.7036],
        [1.0054, 1.3940, 1.4129],
        [1.2666,... 1.9846],
        [1.9936, 1.6940, 1.9017],
        [1.5016, 1.4799, 1.8082]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: u_add_v)
_____________________ test_sddmm[idtype1-add-v-u-shp3-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'u', rhs_target = 'v', msg = 'add'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add'
lhs_data = tensor([[1.5833, 1.1756, 1.0194],
        [1.9846, 1.0712, 1.2492],
        [1.3490, 1.7131, 1.2302],
        [1.7925,... 1.2327],
        [1.5662, 1.2526, 1.6551],
        [1.4188, 1.6150, 1.7798]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.1308, 1.3051, 1.1554],
        [1.2795, 1.1820, 1.4405],
        [1.0247, 1.1935, 1.4956],
        [1.3573,... 1.9886],
        [1.8133, 1.4897, 1.1417],
        [1.4957, 1.5050, 1.0287]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([40, 3])
SDDMM(message func: u_add_v)
_____________________ test_sddmm[idtype1-add-v-u-shp4-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'u', rhs_target = 'v', msg = 'add'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add'
lhs_data = tensor([[1.6561],
        [1.7191],
        [1.2062],
        [1.9252],
        [1.7600],
        [1.5344],
        [1...061],
        [1.3335],
        [1.2665],
        [1.5747],
        [1.5915]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.1429],
        [1.6412],
        [1.7583],
        [1.7666],
        [1.3087],
        [1.1981],
        [1...910],
        [1.1681],
        [1.1331],
        [1.9722],
        [1.6215]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: u_add_v)
_____________________ test_sddmm[idtype1-add-v-u-shp4-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'u', rhs_target = 'v', msg = 'add'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add'
lhs_data = tensor([[1.9532],
        [1.0102],
        [1.6925],
        [1.1873],
        [1.8773],
        [1.1254],
        [1...892],
        [1.9468],
        [1.8549],
        [1.3866],
        [1.3802]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.4181],
        [1.0448],
        [1.2489],
        [1.8844],
        [1.4039],
        [1.7845],
        [1...211],
        [1.6263],
        [1.1343],
        [1.3712],
        [1.5317]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([40, 1])
SDDMM(message func: u_add_v)
_____________________ test_sddmm[idtype1-add-v-e-shp0-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...torch.float64), 'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'e', rhs_target = 'v'
msg = 'add', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add'
lhs_data = tensor([[[[[[1.1146],
            [1.7684],
            [1.8266]]],


          [[[1.3946],
            [1.8078],
    ...7]]],


          [[[1.5341],
            [1.6418],
            [1.1312]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.7498]],

           [[1.2046]],

           [[1.6757]]]],



         [[[[1.7208]],

           [[1.224...



         [[[[1.4508]],

           [[1.7058]],

           [[1.8671]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([100, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: e_add_v)
_____________________ test_sddmm[idtype1-add-v-e-shp0-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'e', rhs_target = 'v'
msg = 'add', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add'
lhs_data = tensor([[[[[[1.9730],
            [1.9797],
            [1.2280]]],


          [[[1.0796],
            [1.7329],
    ...8]]],


          [[[1.0448],
            [1.9750],
            [1.3335]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.5435]],

           [[1.7884]],

           [[1.7035]]]],



         [[[[1.4579]],

           [[1.503...



         [[[[1.3682]],

           [[1.7304]],

           [[1.8775]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([300, 1, 2, 1, 3, 1]), rhs shape: torch.Size([40, 4, 1, 3, 1, 1])
SDDMM(message func: e_add_v)
_____________________ test_sddmm[idtype1-add-v-e-shp1-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...pe=torch.float64), 'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'e', rhs_target = 'v'
msg = 'add', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add'
lhs_data = tensor([[[[[1.2434, 1.7485, 1.4575,  ..., 1.0595, 1.9408, 1.3017]],

          [[1.0811, 1.7514, 1.2345,  ..., 1.3628,...
          [[1.6862, 1.7738, 1.2545,  ..., 1.7720, 1.4763, 1.8924]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.1037, 1.8413, 1.3547,  ..., 1.0432, 1.5343, 1.8118],
           [1.9101, 1.7309, 1.2263,  ..., 1.4962, 1...
           [1.7990, 1.7387, 1.1957,  ..., 1.6894, 1.6975, 1.0813]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([100, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: e_add_v)
_____________________ test_sddmm[idtype1-add-v-e-shp1-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'e', rhs_target = 'v'
msg = 'add', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add'
lhs_data = tensor([[[[[1.1781, 1.8104, 1.4167,  ..., 1.7402, 1.7841, 1.2432]],

          [[1.2595, 1.5318, 1.4352,  ..., 1.5960,...
          [[1.6245, 1.1906, 1.1987,  ..., 1.2696, 1.0877, 1.7847]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.1033, 1.2277, 1.2138,  ..., 1.1396, 1.9349, 1.7216],
           [1.7964, 1.4656, 1.7572,  ..., 1.3913, 1...
           [1.3547, 1.3458, 1.1348,  ..., 1.7423, 1.4263, 1.1541]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([300, 5, 3, 1, 7]), rhs shape: torch.Size([40, 1, 3, 7, 7])
SDDMM(message func: e_add_v)
_____________________ test_sddmm[idtype1-add-v-e-shp2-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...dtype=torch.float64), 'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'e', rhs_target = 'v', msg = 'add'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add'
lhs_data = tensor([[[[1.0624, 1.4986, 1.4694],
          [1.3361, 1.7209, 1.1186],
          [1.8727, 1.8659, 1.2851]]],


      ...2],
          [1.8533, 1.2509, 1.8916],
          [1.1403, 1.3029, 1.7535]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.0850, 1.5078, 1.1561]],

         [[1.9792, 1.5955, 1.8601]],

         [[1.3758, 1.0178, 1.9089]],

    ...

         [[1.6209, 1.8861, 1.4962]],

         [[1.1680, 1.2349, 1.6488]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([100, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: e_add_v)
_____________________ test_sddmm[idtype1-add-v-e-shp2-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'e', rhs_target = 'v', msg = 'add'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add'
lhs_data = tensor([[[[1.9571, 1.3955, 1.5467],
          [1.9316, 1.3681, 1.2087],
          [1.8756, 1.5289, 1.7945]]],


      ...0],
          [1.2124, 1.2538, 1.0605],
          [1.9247, 1.8526, 1.2114]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.3582, 1.1320, 1.3279]],

         [[1.5667, 1.7028, 1.9052]],

         [[1.9322, 1.8559, 1.6137]],

    ...

         [[1.4866, 1.1908, 1.5167]],

         [[1.0797, 1.1349, 1.9022]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([300, 1, 3, 3]), rhs shape: torch.Size([40, 4, 1, 3])
SDDMM(message func: e_add_v)
_____________________ test_sddmm[idtype1-add-v-e-shp3-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'e', rhs_target = 'v', msg = 'add'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add'
lhs_data = tensor([[1.5191, 1.4811, 1.2883],
        [1.7521, 1.4338, 1.1398],
        [1.8837, 1.6209, 1.2522],
        [1.9714,... 1.9972],
        [1.0341, 1.6540, 1.1054],
        [1.7554, 1.9855, 1.9266]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.6207, 1.5643, 1.9972],
        [1.7160, 1.5222, 1.0550],
        [1.9263, 1.0072, 1.6982],
        [1.0538,... 1.7073],
        [1.3092, 1.9873, 1.7210],
        [1.5783, 1.1538, 1.1296]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([100, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: e_add_v)
_____________________ test_sddmm[idtype1-add-v-e-shp3-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'e', rhs_target = 'v', msg = 'add'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add'
lhs_data = tensor([[1.1654, 1.8071, 1.8092],
        [1.4466, 1.1657, 1.4091],
        [1.2519, 1.2495, 1.5696],
        [1.4298,... 1.0394],
        [1.6665, 1.5216, 1.6141],
        [1.5416, 1.1146, 1.6650]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.7269, 1.4707, 1.7936],
        [1.2871, 1.1603, 1.2706],
        [1.3134, 1.5496, 1.2597],
        [1.2935,... 1.0325],
        [1.1027, 1.4124, 1.8089],
        [1.3275, 1.7205, 1.4660]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([300, 3]), rhs shape: torch.Size([40, 3])
SDDMM(message func: e_add_v)
_____________________ test_sddmm[idtype1-add-v-e-shp4-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'e', rhs_target = 'v', msg = 'add'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add'
lhs_data = tensor([[1.7608],
        [1.2294],
        [1.8637],
        [1.6588],
        [1.1241],
        [1.3427],
        [1...554],
        [1.5407],
        [1.7656],
        [1.9060],
        [1.2256]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.1851],
        [1.2650],
        [1.1349],
        [1.9278],
        [1.5218],
        [1.2316],
        [1...978],
        [1.9687],
        [1.4469],
        [1.1233],
        [1.7872]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([100, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: e_add_v)
_____________________ test_sddmm[idtype1-add-v-e-shp4-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'e', rhs_target = 'v', msg = 'add'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add'
lhs_data = tensor([[1.3876],
        [1.9929],
        [1.3771],
        [1.4147],
        [1.6702],
        [1.2728],
        [1...280],
        [1.8772],
        [1.3917],
        [1.5684],
        [1.0805]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.1477],
        [1.0744],
        [1.9818],
        [1.5123],
        [1.8177],
        [1.4724],
        [1...269],
        [1.1013],
        [1.0969],
        [1.8492],
        [1.3295]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([300, 1]), rhs shape: torch.Size([40, 1])
SDDMM(message func: e_add_v)
_____________________ test_sddmm[idtype1-add-e-u-shp0-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': ...torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'u', rhs_target = 'e'
msg = 'add', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add'
lhs_data = tensor([[[[[[1.0053],
            [1.7804],
            [1.6319]]],


          [[[1.9662],
            [1.4485],
    ...7]]],


          [[[1.5787],
            [1.2895],
            [1.7345]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.5193]],

           [[1.4445]],

           [[1.0664]]]],



         [[[[1.2708]],

           [[1.306...



         [[[[1.3641]],

           [[1.4687]],

           [[1.6581]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([100, 4, 1, 3, 1, 1])
SDDMM(message func: u_add_e)
_____________________ test_sddmm[idtype1-add-e-u-shp0-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'u', rhs_target = 'e'
msg = 'add', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add'
lhs_data = tensor([[[[[[1.2054],
            [1.9421],
            [1.1976]]],


          [[[1.2240],
            [1.5792],
    ...4]]],


          [[[1.1710],
            [1.4424],
            [1.5743]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.3623]],

           [[1.4797]],

           [[1.7773]]]],



         [[[[1.4888]],

           [[1.624...



         [[[[1.9893]],

           [[1.8137]],

           [[1.3094]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([300, 4, 1, 3, 1, 1])
SDDMM(message func: u_add_e)
_____________________ test_sddmm[idtype1-add-e-u-shp1-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Sch...pe=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'u', rhs_target = 'e'
msg = 'add', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add'
lhs_data = tensor([[[[[1.8333, 1.8507, 1.6303,  ..., 1.0348, 1.3060, 1.2684]],

          [[1.2158, 1.9658, 1.8782,  ..., 1.9968,...
          [[1.5517, 1.3615, 1.5388,  ..., 1.7414, 1.5114, 1.0358]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.3245, 1.0786, 1.4657,  ..., 1.6911, 1.2880, 1.3843],
           [1.3764, 1.4646, 1.2317,  ..., 1.4859, 1...
           [1.1218, 1.2017, 1.1513,  ..., 1.8895, 1.8786, 1.7025]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([100, 1, 3, 7, 7])
SDDMM(message func: u_add_e)
_____________________ test_sddmm[idtype1-add-e-u-shp1-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'u', rhs_target = 'e'
msg = 'add', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add'
lhs_data = tensor([[[[[1.1426, 1.8587, 1.6395,  ..., 1.1051, 1.9792, 1.9669]],

          [[1.9704, 1.7791, 1.4883,  ..., 1.9561,...
          [[1.6509, 1.8679, 1.7501,  ..., 1.1128, 1.6782, 1.5815]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.8804, 1.6917, 1.7911,  ..., 1.6810, 1.4966, 1.1863],
           [1.1276, 1.6891, 1.9992,  ..., 1.2787, 1...
           [1.8118, 1.2794, 1.4323,  ..., 1.8533, 1.4293, 1.6536]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([300, 1, 3, 7, 7])
SDDMM(message func: u_add_e)
_____________________ test_sddmm[idtype1-add-e-u-shp2-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme...dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'u', rhs_target = 'e', msg = 'add'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add'
lhs_data = tensor([[[[1.3524, 1.6432, 1.9267],
          [1.8275, 1.4220, 1.2313],
          [1.3328, 1.3519, 1.2628]]],


      ...4],
          [1.3974, 1.7028, 1.4401],
          [1.5274, 1.5624, 1.1759]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.1679, 1.2677, 1.9612]],

         [[1.7909, 1.0585, 1.6631]],

         [[1.8006, 1.7809, 1.8786]],

    ...

         [[1.1210, 1.0931, 1.2646]],

         [[1.8463, 1.2050, 1.7938]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([100, 4, 1, 3])
SDDMM(message func: u_add_e)
_____________________ test_sddmm[idtype1-add-e-u-shp2-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'u', rhs_target = 'e', msg = 'add'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add'
lhs_data = tensor([[[[1.4916, 1.7195, 1.2323],
          [1.5563, 1.3816, 1.7122],
          [1.3958, 1.2980, 1.3259]]],


      ...4],
          [1.8233, 1.3874, 1.3650],
          [1.9273, 1.0656, 1.6052]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.2484, 1.7657, 1.9997]],

         [[1.8718, 1.2602, 1.9289]],

         [[1.1586, 1.6986, 1.1465]],

    ...

         [[1.8887, 1.7340, 1.5588]],

         [[1.1147, 1.8705, 1.7062]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([300, 4, 1, 3])
SDDMM(message func: u_add_e)
_____________________ test_sddmm[idtype1-add-e-u-shp3-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'u', rhs_target = 'e', msg = 'add'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add'
lhs_data = tensor([[1.6416, 1.0958, 1.7409],
        [1.0392, 1.6610, 1.6641],
        [1.2580, 1.5950, 1.4496],
        [1.7508,... 1.6311],
        [1.2313, 1.3431, 1.1137],
        [1.3299, 1.9229, 1.8152]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.6670, 1.4466, 1.2826],
        [1.3144, 1.2591, 1.6553],
        [1.2251, 1.6580, 1.8085],
        [1.0381,... 1.7061],
        [1.6886, 1.9629, 1.0166],
        [1.4792, 1.7211, 1.0570]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([100, 3])
SDDMM(message func: u_add_e)
_____________________ test_sddmm[idtype1-add-e-u-shp3-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'u', rhs_target = 'e', msg = 'add'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add'
lhs_data = tensor([[1.4878, 1.5795, 1.1755],
        [1.7800, 1.8636, 1.7180],
        [1.0097, 1.1239, 1.4087],
        [1.0143,... 1.9595],
        [1.3204, 1.1145, 1.0414],
        [1.7340, 1.2666, 1.8919]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.5303, 1.1038, 1.6339],
        [1.6315, 1.5851, 1.0045],
        [1.1619, 1.3706, 1.8699],
        [1.4778,... 1.7528],
        [1.2638, 1.0320, 1.1046],
        [1.4972, 1.5621, 1.7183]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([300, 3])
SDDMM(message func: u_add_e)
_____________________ test_sddmm[idtype1-add-e-u-shp4-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'u', rhs_target = 'e', msg = 'add'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add'
lhs_data = tensor([[1.1547],
        [1.2152],
        [1.0983],
        [1.5987],
        [1.6596],
        [1.4047],
        [1...827],
        [1.7516],
        [1.5310],
        [1.0022],
        [1.6837]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.2879],
        [1.5099],
        [1.4739],
        [1.6035],
        [1.4394],
        [1.4727],
        [1...754],
        [1.7084],
        [1.0721],
        [1.1669],
        [1.5342]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([100, 1])
SDDMM(message func: u_add_e)
_____________________ test_sddmm[idtype1-add-e-u-shp4-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'u', rhs_target = 'e', msg = 'add'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add'
lhs_data = tensor([[1.0345],
        [1.5694],
        [1.4498],
        [1.1511],
        [1.9972],
        [1.7766],
        [1...832],
        [1.7728],
        [1.4430],
        [1.7514],
        [1.7721]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.6846],
        [1.8675],
        [1.5406],
        [1.1838],
        [1.0617],
        [1.5683],
        [1...033],
        [1.8840],
        [1.1895],
        [1.8599],
        [1.5083]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([300, 1])
SDDMM(message func: u_add_e)
_____________________ test_sddmm[idtype1-add-e-v-shp0-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': ...torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'v', rhs_target = 'e'
msg = 'add', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add'
lhs_data = tensor([[[[[[1.7166],
            [1.0125],
            [1.9982]]],


          [[[1.5105],
            [1.0188],
    ...5]]],


          [[[1.7876],
            [1.1723],
            [1.4712]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.5437]],

           [[1.1858]],

           [[1.3519]]]],



         [[[[1.4843]],

           [[1.323...



         [[[[1.9395]],

           [[1.8879]],

           [[1.9716]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([100, 4, 1, 3, 1, 1])
SDDMM(message func: v_add_e)
_____________________ test_sddmm[idtype1-add-e-v-shp0-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'v', rhs_target = 'e'
msg = 'add', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add'
lhs_data = tensor([[[[[[1.4652],
            [1.4014],
            [1.2784]]],


          [[[1.7718],
            [1.9749],
    ...2]]],


          [[[1.6112],
            [1.9120],
            [1.0811]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.7643]],

           [[1.5711]],

           [[1.5727]]]],



         [[[[1.0619]],

           [[1.848...



         [[[[1.4383]],

           [[1.6854]],

           [[1.7106]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([40, 1, 2, 1, 3, 1]), rhs shape: torch.Size([300, 4, 1, 3, 1, 1])
SDDMM(message func: v_add_e)
_____________________ test_sddmm[idtype1-add-e-v-shp1-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Sch...pe=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'v', rhs_target = 'e'
msg = 'add', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add'
lhs_data = tensor([[[[[1.9528, 1.9638, 1.1179,  ..., 1.0546, 1.3443, 1.2611]],

          [[1.7479, 1.8338, 1.1627,  ..., 1.2441,...
          [[1.4377, 1.6141, 1.5060,  ..., 1.2473, 1.0602, 1.6031]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.5676, 1.6662, 1.3185,  ..., 1.7996, 1.2741, 1.1561],
           [1.0577, 1.4773, 1.0189,  ..., 1.6398, 1...
           [1.3247, 1.7295, 1.9057,  ..., 1.4371, 1.7640, 1.3646]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([100, 1, 3, 7, 7])
SDDMM(message func: v_add_e)
_____________________ test_sddmm[idtype1-add-e-v-shp1-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'v', rhs_target = 'e'
msg = 'add', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add'
lhs_data = tensor([[[[[1.9412, 1.2174, 1.0288,  ..., 1.6425, 1.6423, 1.7061]],

          [[1.3762, 1.4356, 1.2509,  ..., 1.6601,...
          [[1.4885, 1.3210, 1.6336,  ..., 1.7804, 1.0740, 1.5944]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.9288, 1.9025, 1.0017,  ..., 1.8073, 1.2327, 1.5438],
           [1.6093, 1.9591, 1.4688,  ..., 1.1794, 1...
           [1.8248, 1.2181, 1.5041,  ..., 1.6413, 1.8539, 1.4402]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([40, 5, 3, 1, 7]), rhs shape: torch.Size([300, 1, 3, 7, 7])
SDDMM(message func: v_add_e)
_____________________ test_sddmm[idtype1-add-e-v-shp2-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme...dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'v', rhs_target = 'e', msg = 'add'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add'
lhs_data = tensor([[[[1.3057, 1.3358, 1.6410],
          [1.5787, 1.4799, 1.0859],
          [1.9504, 1.7084, 1.5230]]],


      ...1],
          [1.1970, 1.2981, 1.1247],
          [1.1059, 1.9311, 1.1091]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.9542, 1.3705, 1.0056]],

         [[1.1461, 1.3292, 1.3641]],

         [[1.9008, 1.9733, 1.9780]],

    ...

         [[1.7357, 1.6844, 1.1678]],

         [[1.3068, 1.6040, 1.8751]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([100, 4, 1, 3])
SDDMM(message func: v_add_e)
_____________________ test_sddmm[idtype1-add-e-v-shp2-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'v', rhs_target = 'e', msg = 'add'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add'
lhs_data = tensor([[[[1.4017, 1.7608, 1.5787],
          [1.6960, 1.5984, 1.5215],
          [1.9810, 1.8388, 1.5699]]],


      ...5],
          [1.4491, 1.4456, 1.5457],
          [1.4297, 1.2485, 1.9296]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.7159, 1.4648, 1.4866]],

         [[1.2840, 1.9872, 1.6424]],

         [[1.3253, 1.9569, 1.2142]],

    ...

         [[1.8874, 1.3427, 1.8343]],

         [[1.3518, 1.8188, 1.4617]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([40, 1, 3, 3]), rhs shape: torch.Size([300, 4, 1, 3])
SDDMM(message func: v_add_e)
_____________________ test_sddmm[idtype1-add-e-v-shp3-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'v', rhs_target = 'e', msg = 'add'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add'
lhs_data = tensor([[1.8753, 1.0464, 1.4332],
        [1.7454, 1.9349, 1.5659],
        [1.7439, 1.1389, 1.0809],
        [1.8452,... 1.3128],
        [1.8697, 1.6044, 1.8984],
        [1.9075, 1.7643, 1.4044]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.8388, 1.3841, 1.6436],
        [1.6505, 1.6263, 1.9125],
        [1.3973, 1.0063, 1.8325],
        [1.9237,... 1.6438],
        [1.9051, 1.6129, 1.0077],
        [1.9544, 1.0961, 1.0564]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([100, 3])
SDDMM(message func: v_add_e)
_____________________ test_sddmm[idtype1-add-e-v-shp3-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'v', rhs_target = 'e', msg = 'add'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add'
lhs_data = tensor([[1.5695, 1.8482, 1.4711],
        [1.2731, 1.3974, 1.4041],
        [1.0698, 1.6518, 1.4825],
        [1.8847,... 1.3981],
        [1.4921, 1.2392, 1.8893],
        [1.3390, 1.0323, 1.5885]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.3723, 1.6890, 1.8322],
        [1.5101, 1.8706, 1.7367],
        [1.4249, 1.6973, 1.0243],
        [1.4492,... 1.2356],
        [1.0756, 1.2438, 1.3293],
        [1.1815, 1.6957, 1.8840]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([40, 3]), rhs shape: torch.Size([300, 3])
SDDMM(message func: v_add_e)
_____________________ test_sddmm[idtype1-add-e-v-shp4-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'v', rhs_target = 'e', msg = 'add'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add'
lhs_data = tensor([[1.8148],
        [1.9720],
        [1.2225],
        [1.4484],
        [1.0708],
        [1.9389],
        [1...377],
        [1.2912],
        [1.9396],
        [1.1191],
        [1.6907]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.6715],
        [1.7463],
        [1.3158],
        [1.9026],
        [1.1690],
        [1.2882],
        [1...560],
        [1.2269],
        [1.6553],
        [1.6649],
        [1.8677]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([100, 1])
SDDMM(message func: v_add_e)
_____________________ test_sddmm[idtype1-add-e-v-shp4-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'v', rhs_target = 'e', msg = 'add'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add'
lhs_data = tensor([[1.2780],
        [1.8581],
        [1.4071],
        [1.0348],
        [1.1021],
        [1.9189],
        [1...021],
        [1.3575],
        [1.2928],
        [1.5549],
        [1.2882]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.4866],
        [1.9414],
        [1.8698],
        [1.5752],
        [1.6475],
        [1.7066],
        [1...327],
        [1.6885],
        [1.2074],
        [1.5713],
        [1.2050]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([40, 1]), rhs shape: torch.Size([300, 1])
SDDMM(message func: v_add_e)
_____________________ test_sddmm[idtype1-sub-u-v-shp0-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': ...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'v', rhs_target = 'u'
msg = 'sub', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add'
lhs_data = tensor([[[[[[1.1795],
            [1.2445],
            [1.3854]]],


          [[[1.4870],
            [1.4548],
    ...4]]],


          [[[1.6869],
            [1.5282],
            [1.4581]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[-1.7679]],

           [[-1.8316]],

           [[-1.1894]]]],



         [[[[-1.9721]],

           [[-...       [[[[-1.9645]],

           [[-1.5439]],

           [[-1.2592]]]]]], dtype=torch.float64, grad_fn=<NegBackward>)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: v_sub_u)
_____________________ test_sddmm[idtype1-sub-u-v-shp0-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'v', rhs_target = 'u'
msg = 'sub', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add'
lhs_data = tensor([[[[[[1.7745],
            [1.6942],
            [1.2201]]],


          [[[1.5010],
            [1.1492],
    ...0]]],


          [[[1.5129],
            [1.8245],
            [1.4573]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[-1.2053]],

           [[-1.7442]],

           [[-1.9906]]]],



         [[[[-1.1710]],

           [[-...       [[[[-1.6080]],

           [[-1.7713]],

           [[-1.1877]]]]]], dtype=torch.float64, grad_fn=<NegBackward>)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([40, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: v_sub_u)
_____________________ test_sddmm[idtype1-sub-u-v-shp1-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Sch...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'v', rhs_target = 'u'
msg = 'sub', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add'
lhs_data = tensor([[[[[1.2112, 1.6402, 1.8733,  ..., 1.1528, 1.1314, 1.7905]],

          [[1.5433, 1.4121, 1.1877,  ..., 1.2342,...
          [[1.2173, 1.0743, 1.3142,  ..., 1.9577, 1.3387, 1.3137]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[-1.1166, -1.2127, -1.5396,  ..., -1.1722, -1.5038, -1.4897],
           [-1.4282, -1.9482, -1.5794,  ..., ...   [-1.7182, -1.3709, -1.0238,  ..., -1.4181, -1.4262, -1.6308]]]]],
       dtype=torch.float64, grad_fn=<NegBackward>)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: v_sub_u)
_____________________ test_sddmm[idtype1-sub-u-v-shp1-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'v', rhs_target = 'u'
msg = 'sub', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add'
lhs_data = tensor([[[[[1.8975, 1.1037, 1.7579,  ..., 1.3435, 1.8576, 1.9476]],

          [[1.5955, 1.8504, 1.5567,  ..., 1.9212,...
          [[1.6284, 1.0826, 1.4157,  ..., 1.4065, 1.8299, 1.8006]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[-1.1811, -1.3984, -1.4694,  ..., -1.3201, -1.1436, -1.8778],
           [-1.2517, -1.2074, -1.7647,  ..., ...   [-1.2746, -1.8346, -1.9630,  ..., -1.8238, -1.7100, -1.4371]]]]],
       dtype=torch.float64, grad_fn=<NegBackward>)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([40, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: v_sub_u)
_____________________ test_sddmm[idtype1-sub-u-v-shp2-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'v', rhs_target = 'u', msg = 'sub'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add'
lhs_data = tensor([[[[1.8155, 1.9733, 1.9601],
          [1.1673, 1.6327, 1.6664],
          [1.9783, 1.1182, 1.4094]]],


      ...9],
          [1.4658, 1.2297, 1.8897],
          [1.9914, 1.6260, 1.3477]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[-1.2463, -1.1755, -1.2695]],

         [[-1.8347, -1.3264, -1.2830]],

         [[-1.2136, -1.1117, -1.8637...1240, -1.0793, -1.6351]],

         [[-1.2996, -1.6086, -1.6366]]]], dtype=torch.float64,
       grad_fn=<NegBackward>)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: v_sub_u)
_____________________ test_sddmm[idtype1-sub-u-v-shp2-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'v', rhs_target = 'u', msg = 'sub'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add'
lhs_data = tensor([[[[1.3369, 1.7815, 1.2619],
          [1.2740, 1.8385, 1.6761],
          [1.4766, 1.4518, 1.7626]]],


      ...6],
          [1.1029, 1.7361, 1.5116],
          [1.0777, 1.9530, 1.9462]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[-1.9952, -1.3456, -1.0052]],

         [[-1.3023, -1.6300, -1.0370]],

         [[-1.9797, -1.5283, -1.7989...7164, -1.3998, -1.9559]],

         [[-1.1417, -1.7327, -1.7614]]]], dtype=torch.float64,
       grad_fn=<NegBackward>)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([40, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: v_sub_u)
_____________________ test_sddmm[idtype1-sub-u-v-shp3-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'v', rhs_target = 'u', msg = 'sub'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add'
lhs_data = tensor([[1.3548, 1.6724, 1.2639],
        [1.2232, 1.1428, 1.5033],
        [1.1367, 1.7010, 1.0270],
        [1.3830,... 1.3659],
        [1.6541, 1.0976, 1.9485],
        [1.4686, 1.5412, 1.2845]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.9622, -1.0297, -1.7694],
        [-1.5773, -1.8425, -1.8534],
        [-1.9259, -1.5715, -1.2766],
       ...  [-1.8178, -1.4289, -1.1438],
        [-1.1341, -1.8094, -1.3500]], dtype=torch.float64,
       grad_fn=<NegBackward>)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: v_sub_u)
_____________________ test_sddmm[idtype1-sub-u-v-shp3-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'v', rhs_target = 'u', msg = 'sub'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add'
lhs_data = tensor([[1.2998, 1.2458, 1.1229],
        [1.9411, 1.4997, 1.4892],
        [1.2512, 1.8789, 1.1003],
        [1.8518,... 1.4803],
        [1.7580, 1.4317, 1.2687],
        [1.4405, 1.8234, 1.1340]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.6268, -1.6755, -1.3902],
        [-1.7552, -1.3351, -1.2151],
        [-1.9100, -1.4768, -1.2976],
       ...  [-1.0272, -1.9719, -1.6333],
        [-1.9035, -1.4353, -1.3766]], dtype=torch.float64,
       grad_fn=<NegBackward>)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([40, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: v_sub_u)
_____________________ test_sddmm[idtype1-sub-u-v-shp4-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'v', rhs_target = 'u', msg = 'sub'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add'
lhs_data = tensor([[1.3567],
        [1.7589],
        [1.2408],
        [1.9908],
        [1.8823],
        [1.2824],
        [1...274],
        [1.6649],
        [1.3238],
        [1.9560],
        [1.0259]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.8992],
        [-1.0755],
        [-1.3164],
        [-1.1441],
        [-1.6808],
        [-1.7479],
    ...       [-1.2042],
        [-1.5258],
        [-1.6347],
        [-1.1022]], dtype=torch.float64, grad_fn=<NegBackward>)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: v_sub_u)
_____________________ test_sddmm[idtype1-sub-u-v-shp4-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'v', rhs_target = 'u', msg = 'sub'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add'
lhs_data = tensor([[1.4615],
        [1.6912],
        [1.4956],
        [1.4110],
        [1.6425],
        [1.9516],
        [1...956],
        [1.0059],
        [1.1095],
        [1.5681],
        [1.0757]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.6238],
        [-1.6586],
        [-1.1293],
        [-1.3257],
        [-1.9223],
        [-1.4091],
    ...       [-1.7848],
        [-1.7967],
        [-1.3291],
        [-1.0773]], dtype=torch.float64, grad_fn=<NegBackward>)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([40, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: v_sub_u)
_____________________ test_sddmm[idtype1-sub-u-e-shp0-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...torch.float64), 'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'e', rhs_target = 'u'
msg = 'sub', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add'
lhs_data = tensor([[[[[[1.5645],
            [1.5317],
            [1.2214]]],


          [[[1.5989],
            [1.9585],
    ...1]]],


          [[[1.5152],
            [1.8388],
            [1.9746]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[-1.3473]],

           [[-1.0661]],

           [[-1.8359]]]],



         [[[[-1.9997]],

           [[-...       [[[[-1.9918]],

           [[-1.3307]],

           [[-1.7427]]]]]], dtype=torch.float64, grad_fn=<NegBackward>)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([100, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: e_sub_u)
_____________________ test_sddmm[idtype1-sub-u-e-shp0-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'e', rhs_target = 'u'
msg = 'sub', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add'
lhs_data = tensor([[[[[[1.1991],
            [1.0801],
            [1.5652]]],


          [[[1.1688],
            [1.3135],
    ...5]]],


          [[[1.5099],
            [1.2017],
            [1.2282]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[-1.5831]],

           [[-1.9279]],

           [[-1.0280]]]],



         [[[[-1.4511]],

           [[-...       [[[[-1.2122]],

           [[-1.2279]],

           [[-1.9157]]]]]], dtype=torch.float64, grad_fn=<NegBackward>)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([300, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: e_sub_u)
_____________________ test_sddmm[idtype1-sub-u-e-shp1-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...pe=torch.float64), 'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'e', rhs_target = 'u'
msg = 'sub', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add'
lhs_data = tensor([[[[[1.2415, 1.6276, 1.6331,  ..., 1.2767, 1.4382, 1.4394]],

          [[1.6485, 1.0773, 1.3831,  ..., 1.4612,...
          [[1.2428, 1.8404, 1.5045,  ..., 1.8584, 1.6209, 1.5052]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[-1.6642, -1.3548, -1.7054,  ..., -1.4920, -1.4596, -1.6399],
           [-1.4950, -1.1548, -1.9028,  ..., ...   [-1.8430, -1.1306, -1.7906,  ..., -1.0242, -1.7503, -1.8562]]]]],
       dtype=torch.float64, grad_fn=<NegBackward>)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([100, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: e_sub_u)
_____________________ test_sddmm[idtype1-sub-u-e-shp1-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'e', rhs_target = 'u'
msg = 'sub', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add'
lhs_data = tensor([[[[[1.5050, 1.5502, 1.0872,  ..., 1.6453, 1.5792, 1.4944]],

          [[1.4555, 1.8809, 1.4106,  ..., 1.2840,...
          [[1.2984, 1.8286, 1.0988,  ..., 1.9322, 1.8762, 1.4155]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[-1.2426, -1.7832, -1.8431,  ..., -1.4781, -1.0629, -1.6853],
           [-1.5422, -1.9730, -1.8806,  ..., ...   [-1.1050, -1.9451, -1.4447,  ..., -1.7610, -1.0656, -1.7471]]]]],
       dtype=torch.float64, grad_fn=<NegBackward>)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([300, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: e_sub_u)
_____________________ test_sddmm[idtype1-sub-u-e-shp2-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...dtype=torch.float64), 'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'e', rhs_target = 'u', msg = 'sub'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add'
lhs_data = tensor([[[[1.5235, 1.5122, 1.3541],
          [1.2758, 1.8799, 1.7244],
          [1.3875, 1.6798, 1.4945]]],


      ...1],
          [1.2899, 1.8673, 1.8388],
          [1.6174, 1.0995, 1.3016]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[-1.5328, -1.6652, -1.6618]],

         [[-1.9245, -1.7228, -1.1234]],

         [[-1.1985, -1.1065, -1.2348...2727, -1.8309, -1.6680]],

         [[-1.2651, -1.8905, -1.4274]]]], dtype=torch.float64,
       grad_fn=<NegBackward>)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([100, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: e_sub_u)
_____________________ test_sddmm[idtype1-sub-u-e-shp2-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'e', rhs_target = 'u', msg = 'sub'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add'
lhs_data = tensor([[[[1.3092, 1.0459, 1.6709],
          [1.8223, 1.0758, 1.8456],
          [1.7942, 1.3586, 1.2927]]],


      ...1],
          [1.6282, 1.2192, 1.7107],
          [1.3478, 1.8580, 1.4100]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[-1.2527, -1.8022, -1.9185]],

         [[-1.4580, -1.3525, -1.8126]],

         [[-1.9765, -1.6055, -1.0420...3707, -1.5319, -1.8480]],

         [[-1.8877, -1.6211, -1.8737]]]], dtype=torch.float64,
       grad_fn=<NegBackward>)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([300, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: e_sub_u)
_____________________ test_sddmm[idtype1-sub-u-e-shp3-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'e', rhs_target = 'u', msg = 'sub'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add'
lhs_data = tensor([[1.5291, 1.9632, 1.8953],
        [1.2442, 1.6081, 1.1752],
        [1.5645, 1.7444, 1.4179],
        [1.5383,... 1.9451],
        [1.0789, 1.9217, 1.2273],
        [1.5545, 1.0784, 1.3065]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.5602, -1.3264, -1.3103],
        [-1.0196, -1.5806, -1.8802],
        [-1.9131, -1.1947, -1.8721],
       ...  [-1.2449, -1.9987, -1.6683],
        [-1.0360, -1.8585, -1.6127]], dtype=torch.float64,
       grad_fn=<NegBackward>)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([100, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: e_sub_u)
_____________________ test_sddmm[idtype1-sub-u-e-shp3-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'e', rhs_target = 'u', msg = 'sub'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add'
lhs_data = tensor([[1.2152, 1.3666, 1.2668],
        [1.3625, 1.2008, 1.9899],
        [1.7475, 1.7672, 1.2782],
        [1.9977,... 1.9951],
        [1.5960, 1.6139, 1.9031],
        [1.5687, 1.9061, 1.9211]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.7142, -1.2973, -1.7765],
        [-1.2422, -1.1643, -1.2007],
        [-1.0732, -1.9983, -1.9500],
       ...  [-1.6947, -1.9732, -1.6444],
        [-1.2798, -1.4979, -1.2885]], dtype=torch.float64,
       grad_fn=<NegBackward>)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([300, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: e_sub_u)
_____________________ test_sddmm[idtype1-sub-u-e-shp4-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'e', rhs_target = 'u', msg = 'sub'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add'
lhs_data = tensor([[1.8600],
        [1.6807],
        [1.0031],
        [1.4744],
        [1.1900],
        [1.0911],
        [1...304],
        [1.7670],
        [1.0391],
        [1.4124],
        [1.4834]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.1282],
        [-1.0110],
        [-1.3812],
        [-1.9968],
        [-1.9724],
        [-1.6613],
    ...       [-1.0716],
        [-1.2790],
        [-1.7579],
        [-1.5015]], dtype=torch.float64, grad_fn=<NegBackward>)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([100, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: e_sub_u)
_____________________ test_sddmm[idtype1-sub-u-e-shp4-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'e', rhs_target = 'u', msg = 'sub'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add'
lhs_data = tensor([[1.2060],
        [1.7790],
        [1.3621],
        [1.4128],
        [1.6326],
        [1.0460],
        [1...888],
        [1.2157],
        [1.3723],
        [1.6692],
        [1.9861]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.1070],
        [-1.9880],
        [-1.9932],
        [-1.9291],
        [-1.9676],
        [-1.1745],
    ...       [-1.9045],
        [-1.9337],
        [-1.9633],
        [-1.8735]], dtype=torch.float64, grad_fn=<NegBackward>)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([300, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: e_sub_u)
_____________________ test_sddmm[idtype1-sub-v-u-shp0-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': ...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'u', rhs_target = 'v'
msg = 'sub', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add'
lhs_data = tensor([[[[[[1.5469],
            [1.1286],
            [1.7467]]],


          [[[1.9825],
            [1.4169],
    ...0]]],


          [[[1.1366],
            [1.3316],
            [1.9732]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[-1.0357]],

           [[-1.7126]],

           [[-1.9189]]]],



         [[[[-1.7302]],

           [[-...       [[[[-1.2794]],

           [[-1.1734]],

           [[-1.9419]]]]]], dtype=torch.float64, grad_fn=<NegBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: u_sub_v)
_____________________ test_sddmm[idtype1-sub-v-u-shp0-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'u', rhs_target = 'v'
msg = 'sub', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add'
lhs_data = tensor([[[[[[1.9626],
            [1.8952],
            [1.0755]]],


          [[[1.3119],
            [1.3008],
    ...0]]],


          [[[1.3230],
            [1.0573],
            [1.4648]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[-1.0666]],

           [[-1.0623]],

           [[-1.9948]]]],



         [[[[-1.7551]],

           [[-...       [[[[-1.4088]],

           [[-1.5472]],

           [[-1.6737]]]]]], dtype=torch.float64, grad_fn=<NegBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([40, 4, 1, 3, 1, 1])
SDDMM(message func: u_sub_v)
_____________________ test_sddmm[idtype1-sub-v-u-shp1-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Sch...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'u', rhs_target = 'v'
msg = 'sub', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add'
lhs_data = tensor([[[[[1.0916, 1.6394, 1.9499,  ..., 1.7324, 1.0636, 1.7753]],

          [[1.1012, 1.3743, 1.8074,  ..., 1.0851,...
          [[1.6344, 1.8453, 1.5683,  ..., 1.5978, 1.4301, 1.5283]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[-1.9733, -1.6405, -1.8963,  ..., -1.9427, -1.3973, -1.4310],
           [-1.1689, -1.3353, -1.0083,  ..., ...   [-1.0799, -1.5754, -1.5704,  ..., -1.2054, -1.0211, -1.8702]]]]],
       dtype=torch.float64, grad_fn=<NegBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: u_sub_v)
_____________________ test_sddmm[idtype1-sub-v-u-shp1-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'u', rhs_target = 'v'
msg = 'sub', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add'
lhs_data = tensor([[[[[1.6782, 1.7972, 1.3928,  ..., 1.3512, 1.3215, 1.7546]],

          [[1.4781, 1.7571, 1.7869,  ..., 1.6011,...
          [[1.2183, 1.6723, 1.3355,  ..., 1.1812, 1.2144, 1.4293]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[-1.3361, -1.0273, -1.8206,  ..., -1.1408, -1.4483, -1.3585],
           [-1.5306, -1.7053, -1.4878,  ..., ...   [-1.5916, -1.9594, -1.4334,  ..., -1.2051, -1.1727, -1.9359]]]]],
       dtype=torch.float64, grad_fn=<NegBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([40, 1, 3, 7, 7])
SDDMM(message func: u_sub_v)
_____________________ test_sddmm[idtype1-sub-v-u-shp2-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'u', rhs_target = 'v', msg = 'sub'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add'
lhs_data = tensor([[[[1.0145, 1.1180, 1.4647],
          [1.2802, 1.8468, 1.9549],
          [1.2540, 1.1403, 1.0721]]],


      ...1],
          [1.1646, 1.4387, 1.0399],
          [1.3295, 1.5346, 1.3773]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[-1.6538, -1.1548, -1.2536]],

         [[-1.9515, -1.3072, -1.4614]],

         [[-1.4512, -1.5941, -1.7223...7456, -1.4132, -1.4245]],

         [[-1.0464, -1.8621, -1.7405]]]], dtype=torch.float64,
       grad_fn=<NegBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: u_sub_v)
_____________________ test_sddmm[idtype1-sub-v-u-shp2-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'u', rhs_target = 'v', msg = 'sub'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add'
lhs_data = tensor([[[[1.9611, 1.6974, 1.1459],
          [1.6558, 1.9671, 1.4280],
          [1.6449, 1.6687, 1.7912]]],


      ...3],
          [1.0329, 1.7177, 1.0228],
          [1.0800, 1.2412, 1.5783]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[-1.0484, -1.3139, -1.2293]],

         [[-1.3943, -1.3584, -1.6992]],

         [[-1.0724, -1.4368, -1.6966...6049, -1.3551, -1.1730]],

         [[-1.9240, -1.2171, -1.6515]]]], dtype=torch.float64,
       grad_fn=<NegBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([40, 4, 1, 3])
SDDMM(message func: u_sub_v)
_____________________ test_sddmm[idtype1-sub-v-u-shp3-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'u', rhs_target = 'v', msg = 'sub'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add'
lhs_data = tensor([[1.6724, 1.4218, 1.6506],
        [1.5510, 1.7182, 1.9379],
        [1.6494, 1.7588, 1.8375],
        [1.5015,... 1.4739],
        [1.2903, 1.8275, 1.9033],
        [1.4139, 1.7342, 1.0533]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.7669, -1.4504, -1.7368],
        [-1.1404, -1.7953, -1.3462],
        [-1.1395, -1.8815, -1.5919],
       ...  [-1.5007, -1.6308, -1.5976],
        [-1.1261, -1.7114, -1.7212]], dtype=torch.float64,
       grad_fn=<NegBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: u_sub_v)
_____________________ test_sddmm[idtype1-sub-v-u-shp3-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'u', rhs_target = 'v', msg = 'sub'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add'
lhs_data = tensor([[1.3160, 1.5149, 1.6110],
        [1.4937, 1.9326, 1.2580],
        [1.0883, 1.4589, 1.1909],
        [1.4441,... 1.7348],
        [1.2015, 1.8781, 1.4183],
        [1.0987, 1.8815, 1.5452]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.2148, -1.0419, -1.2661],
        [-1.5703, -1.3522, -1.2725],
        [-1.1222, -1.6309, -1.8740],
       ...  [-1.9592, -1.8316, -1.0700],
        [-1.0258, -1.6824, -1.5050]], dtype=torch.float64,
       grad_fn=<NegBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([40, 3])
SDDMM(message func: u_sub_v)
_____________________ test_sddmm[idtype1-sub-v-u-shp4-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'u', rhs_target = 'v', msg = 'sub'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add'
lhs_data = tensor([[1.3684],
        [1.7203],
        [1.6244],
        [1.5457],
        [1.1011],
        [1.6327],
        [1...188],
        [1.1351],
        [1.9788],
        [1.8242],
        [1.5584]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.2757],
        [-1.3728],
        [-1.3119],
        [-1.1470],
        [-1.8669],
        [-1.5417],
    ...       [-1.3810],
        [-1.9403],
        [-1.8139],
        [-1.6507]], dtype=torch.float64, grad_fn=<NegBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: u_sub_v)
_____________________ test_sddmm[idtype1-sub-v-u-shp4-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'u', rhs_target = 'v', msg = 'sub'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add'
lhs_data = tensor([[1.9255],
        [1.7186],
        [1.4539],
        [1.9214],
        [1.6388],
        [1.8467],
        [1...861],
        [1.9113],
        [1.7273],
        [1.4800],
        [1.2690]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.7849],
        [-1.9287],
        [-1.1195],
        [-1.0549],
        [-1.1682],
        [-1.4818],
    ...       [-1.5995],
        [-1.7929],
        [-1.6663],
        [-1.6964]], dtype=torch.float64, grad_fn=<NegBackward>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([40, 1])
SDDMM(message func: u_sub_v)
_____________________ test_sddmm[idtype1-sub-v-e-shp0-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...torch.float64), 'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'e', rhs_target = 'v'
msg = 'sub', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add'
lhs_data = tensor([[[[[[1.2046],
            [1.8173],
            [1.3427]]],


          [[[1.2721],
            [1.7787],
    ...5]]],


          [[[1.3389],
            [1.2144],
            [1.8694]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[-1.0346]],

           [[-1.2330]],

           [[-1.6298]]]],



         [[[[-1.9067]],

           [[-...       [[[[-1.9807]],

           [[-1.4770]],

           [[-1.0401]]]]]], dtype=torch.float64, grad_fn=<NegBackward>)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([100, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: e_sub_v)
_____________________ test_sddmm[idtype1-sub-v-e-shp0-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'e', rhs_target = 'v'
msg = 'sub', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add'
lhs_data = tensor([[[[[[1.5669],
            [1.2279],
            [1.4558]]],


          [[[1.4621],
            [1.2711],
    ...6]]],


          [[[1.7846],
            [1.4215],
            [1.5848]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[-1.9462]],

           [[-1.9526]],

           [[-1.5376]]]],



         [[[[-1.1777]],

           [[-...       [[[[-1.6566]],

           [[-1.6652]],

           [[-1.1678]]]]]], dtype=torch.float64, grad_fn=<NegBackward>)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([300, 1, 2, 1, 3, 1]), rhs shape: torch.Size([40, 4, 1, 3, 1, 1])
SDDMM(message func: e_sub_v)
_____________________ test_sddmm[idtype1-sub-v-e-shp1-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...pe=torch.float64), 'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'e', rhs_target = 'v'
msg = 'sub', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add'
lhs_data = tensor([[[[[1.6682, 1.4507, 1.0613,  ..., 1.7389, 1.2368, 1.8955]],

          [[1.3401, 1.7821, 1.1081,  ..., 1.8836,...
          [[1.2306, 1.6741, 1.9921,  ..., 1.9746, 1.5470, 1.2583]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[-1.5709, -1.6898, -1.6861,  ..., -1.9487, -1.3783, -1.7146],
           [-1.8861, -1.0851, -1.6838,  ..., ...   [-1.4777, -1.9619, -1.9812,  ..., -1.5858, -1.3032, -1.9006]]]]],
       dtype=torch.float64, grad_fn=<NegBackward>)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([100, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: e_sub_v)
_____________________ test_sddmm[idtype1-sub-v-e-shp1-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'e', rhs_target = 'v'
msg = 'sub', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add'
lhs_data = tensor([[[[[1.3356, 1.1504, 1.5126,  ..., 1.6765, 1.2137, 1.3489]],

          [[1.6461, 1.9470, 1.8173,  ..., 1.8891,...
          [[1.8952, 1.7025, 1.1815,  ..., 1.5108, 1.3279, 1.9652]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[-1.2399, -1.1078, -1.0195,  ..., -1.9529, -1.3178, -1.7421],
           [-1.6048, -1.0769, -1.4265,  ..., ...   [-1.7443, -1.7579, -1.9766,  ..., -1.6016, -1.8456, -1.3396]]]]],
       dtype=torch.float64, grad_fn=<NegBackward>)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([300, 5, 3, 1, 7]), rhs shape: torch.Size([40, 1, 3, 7, 7])
SDDMM(message func: e_sub_v)
_____________________ test_sddmm[idtype1-sub-v-e-shp2-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...dtype=torch.float64), 'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'e', rhs_target = 'v', msg = 'sub'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add'
lhs_data = tensor([[[[1.0810, 1.0268, 1.5138],
          [1.1740, 1.2205, 1.1645],
          [1.0390, 1.7951, 1.4311]]],


      ...9],
          [1.9239, 1.2098, 1.2818],
          [1.5400, 1.6357, 1.3263]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[-1.5953, -1.5838, -1.1075]],

         [[-1.0709, -1.2633, -1.6062]],

         [[-1.6404, -1.8070, -1.7814...5158, -1.0803, -1.2692]],

         [[-1.6735, -1.3252, -1.8462]]]], dtype=torch.float64,
       grad_fn=<NegBackward>)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([100, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: e_sub_v)
_____________________ test_sddmm[idtype1-sub-v-e-shp2-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'e', rhs_target = 'v', msg = 'sub'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add'
lhs_data = tensor([[[[1.4402, 1.9725, 1.2595],
          [1.0845, 1.7408, 1.1711],
          [1.1753, 1.0306, 1.9043]]],


      ...3],
          [1.2083, 1.9528, 1.9350],
          [1.8106, 1.4528, 1.0789]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[-1.2812, -1.8389, -1.8139]],

         [[-1.7799, -1.6884, -1.7673]],

         [[-1.3090, -1.5301, -1.7467...7515, -1.1713, -1.6013]],

         [[-1.6686, -1.1697, -1.4757]]]], dtype=torch.float64,
       grad_fn=<NegBackward>)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([300, 1, 3, 3]), rhs shape: torch.Size([40, 4, 1, 3])
SDDMM(message func: e_sub_v)
_____________________ test_sddmm[idtype1-sub-v-e-shp3-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'e', rhs_target = 'v', msg = 'sub'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add'
lhs_data = tensor([[1.2177, 1.0173, 1.8710],
        [1.8871, 1.4363, 1.3116],
        [1.9540, 1.1527, 1.5144],
        [1.9934,... 1.7932],
        [1.0310, 1.3584, 1.7513],
        [1.9493, 1.3276, 1.7810]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.0269, -1.7930, -1.5813],
        [-1.0415, -1.3892, -1.4178],
        [-1.3731, -1.3610, -1.1991],
       ...  [-1.5330, -1.3855, -1.8701],
        [-1.1839, -1.5091, -1.7912]], dtype=torch.float64,
       grad_fn=<NegBackward>)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([100, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: e_sub_v)
_____________________ test_sddmm[idtype1-sub-v-e-shp3-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'e', rhs_target = 'v', msg = 'sub'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add'
lhs_data = tensor([[1.1562, 1.5643, 1.2543],
        [1.5006, 1.8305, 1.2538],
        [1.9485, 1.0129, 1.1352],
        [1.6267,... 1.1546],
        [1.6467, 1.0580, 1.0669],
        [1.7478, 1.6475, 1.9085]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.6627, -1.0032, -1.6440],
        [-1.6099, -1.2508, -1.6476],
        [-1.3672, -1.0637, -1.8628],
       ...  [-1.0166, -1.4397, -1.7026],
        [-1.1858, -1.3122, -1.7349]], dtype=torch.float64,
       grad_fn=<NegBackward>)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([300, 3]), rhs shape: torch.Size([40, 3])
SDDMM(message func: e_sub_v)
_____________________ test_sddmm[idtype1-sub-v-e-shp4-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'e', rhs_target = 'v', msg = 'sub'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add'
lhs_data = tensor([[1.2161],
        [1.2216],
        [1.7045],
        [1.7284],
        [1.3526],
        [1.7305],
        [1...586],
        [1.3848],
        [1.3075],
        [1.9723],
        [1.4654]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.8555],
        [-1.6936],
        [-1.5644],
        [-1.6055],
        [-1.4712],
        [-1.4152],
    ...       [-1.6496],
        [-1.2822],
        [-1.2718],
        [-1.4357]], dtype=torch.float64, grad_fn=<NegBackward>)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([100, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: e_sub_v)
_____________________ test_sddmm[idtype1-sub-v-e-shp4-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'e', rhs_target = 'v', msg = 'sub'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add'
lhs_data = tensor([[1.4931],
        [1.1264],
        [1.8713],
        [1.4827],
        [1.5839],
        [1.5511],
        [1...610],
        [1.3995],
        [1.5760],
        [1.2832],
        [1.1544]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.2032],
        [-1.5582],
        [-1.0949],
        [-1.1938],
        [-1.7044],
        [-1.4414],
    ...       [-1.0361],
        [-1.4262],
        [-1.5480],
        [-1.6794]], dtype=torch.float64, grad_fn=<NegBackward>)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([300, 1]), rhs shape: torch.Size([40, 1])
SDDMM(message func: e_sub_v)
_____________________ test_sddmm[idtype1-sub-e-u-shp0-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': ...torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'u', rhs_target = 'e'
msg = 'sub', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add'
lhs_data = tensor([[[[[[1.9526],
            [1.2947],
            [1.9650]]],


          [[[1.2917],
            [1.4245],
    ...8]]],


          [[[1.7272],
            [1.8195],
            [1.3735]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[-1.9536]],

           [[-1.2107]],

           [[-1.9437]]]],



         [[[[-1.8186]],

           [[-...       [[[[-1.7144]],

           [[-1.7417]],

           [[-1.1721]]]]]], dtype=torch.float64, grad_fn=<NegBackward>)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([100, 4, 1, 3, 1, 1])
SDDMM(message func: u_sub_e)
_____________________ test_sddmm[idtype1-sub-e-u-shp0-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'u', rhs_target = 'e'
msg = 'sub', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add'
lhs_data = tensor([[[[[[1.3293],
            [1.2561],
            [1.9354]]],


          [[[1.3691],
            [1.9942],
    ...6]]],


          [[[1.1086],
            [1.4032],
            [1.4583]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[-1.2943]],

           [[-1.9933]],

           [[-1.9813]]]],



         [[[[-1.1802]],

           [[-...       [[[[-1.4038]],

           [[-1.2828]],

           [[-1.2383]]]]]], dtype=torch.float64, grad_fn=<NegBackward>)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([300, 4, 1, 3, 1, 1])
SDDMM(message func: u_sub_e)
_____________________ test_sddmm[idtype1-sub-e-u-shp1-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Sch...pe=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'u', rhs_target = 'e'
msg = 'sub', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add'
lhs_data = tensor([[[[[1.8181, 1.1007, 1.6485,  ..., 1.1239, 1.7598, 1.3682]],

          [[1.9904, 1.9557, 1.4140,  ..., 1.7244,...
          [[1.4133, 1.9303, 1.9831,  ..., 1.8528, 1.4536, 1.4157]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[-1.7224, -1.8535, -1.9622,  ..., -1.3631, -1.9488, -1.4833],
           [-1.8803, -1.4743, -1.9308,  ..., ...   [-1.2151, -1.7222, -1.2751,  ..., -1.9015, -1.1622, -1.1543]]]]],
       dtype=torch.float64, grad_fn=<NegBackward>)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([100, 1, 3, 7, 7])
SDDMM(message func: u_sub_e)
_____________________ test_sddmm[idtype1-sub-e-u-shp1-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'u', rhs_target = 'e'
msg = 'sub', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add'
lhs_data = tensor([[[[[1.9747, 1.1306, 1.2680,  ..., 1.6113, 1.6816, 1.4729]],

          [[1.7334, 1.5350, 1.6061,  ..., 1.3686,...
          [[1.1151, 1.6032, 1.5638,  ..., 1.5715, 1.8865, 1.2726]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[-1.6307, -1.8897, -1.5502,  ..., -1.0348, -1.4697, -1.5671],
           [-1.6579, -1.3180, -1.4142,  ..., ...   [-1.0250, -1.3774, -1.0580,  ..., -1.0943, -1.2854, -1.9786]]]]],
       dtype=torch.float64, grad_fn=<NegBackward>)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([300, 1, 3, 7, 7])
SDDMM(message func: u_sub_e)
_____________________ test_sddmm[idtype1-sub-e-u-shp2-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme...dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'u', rhs_target = 'e', msg = 'sub'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add'
lhs_data = tensor([[[[1.6601, 1.4979, 1.5408],
          [1.1230, 1.0890, 1.8146],
          [1.1215, 1.1028, 1.1390]]],


      ...2],
          [1.7819, 1.8706, 1.8283],
          [1.1883, 1.9107, 1.0158]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[-1.7886, -1.5813, -1.3102]],

         [[-1.0811, -1.8758, -1.0925]],

         [[-1.8076, -1.3486, -1.9381...0266, -1.9579, -1.3061]],

         [[-1.6022, -1.4367, -1.1829]]]], dtype=torch.float64,
       grad_fn=<NegBackward>)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([100, 4, 1, 3])
SDDMM(message func: u_sub_e)
_____________________ test_sddmm[idtype1-sub-e-u-shp2-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'u', rhs_target = 'e', msg = 'sub'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add'
lhs_data = tensor([[[[1.9101, 1.2677, 1.8874],
          [1.6738, 1.3741, 1.2490],
          [1.5970, 1.2650, 1.1849]]],


      ...6],
          [1.8942, 1.3690, 1.9710],
          [1.3430, 1.2749, 1.3995]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[-1.6080, -1.5515, -1.4512]],

         [[-1.8503, -1.2060, -1.2013]],

         [[-1.3376, -1.0801, -1.9751...8564, -1.5686, -1.2084]],

         [[-1.1603, -1.6134, -1.7815]]]], dtype=torch.float64,
       grad_fn=<NegBackward>)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([300, 4, 1, 3])
SDDMM(message func: u_sub_e)
_____________________ test_sddmm[idtype1-sub-e-u-shp3-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'u', rhs_target = 'e', msg = 'sub'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add'
lhs_data = tensor([[1.5193, 1.2062, 1.5326],
        [1.0477, 1.5651, 1.7073],
        [1.1433, 1.8500, 1.3251],
        [1.8684,... 1.6824],
        [1.9680, 1.4964, 1.6993],
        [1.8219, 1.2446, 1.3764]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.6169, -1.2209, -1.4318],
        [-1.1836, -1.7932, -1.6829],
        [-1.0858, -1.7381, -1.9354],
       ...  [-1.3118, -1.6105, -1.2196],
        [-1.3870, -1.0330, -1.3073]], dtype=torch.float64,
       grad_fn=<NegBackward>)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([100, 3])
SDDMM(message func: u_sub_e)
_____________________ test_sddmm[idtype1-sub-e-u-shp3-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'u', rhs_target = 'e', msg = 'sub'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add'
lhs_data = tensor([[1.5219, 1.7176, 1.7205],
        [1.2934, 1.5280, 1.5746],
        [1.5783, 1.5302, 1.4059],
        [1.2328,... 1.4355],
        [1.5700, 1.9383, 1.9214],
        [1.0208, 1.8359, 1.0045]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.3384, -1.8749, -1.4861],
        [-1.5074, -1.0558, -1.8679],
        [-1.9392, -1.8479, -1.0750],
       ...  [-1.7422, -1.0498, -1.6112],
        [-1.7571, -1.8581, -1.0128]], dtype=torch.float64,
       grad_fn=<NegBackward>)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([300, 3])
SDDMM(message func: u_sub_e)
_____________________ test_sddmm[idtype1-sub-e-u-shp4-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'u', rhs_target = 'e', msg = 'sub'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add'
lhs_data = tensor([[1.1544],
        [1.4232],
        [1.1240],
        [1.6978],
        [1.6964],
        [1.3038],
        [1...638],
        [1.2853],
        [1.2546],
        [1.6617],
        [1.8876]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.4148],
        [-1.5413],
        [-1.0198],
        [-1.2588],
        [-1.8079],
        [-1.7418],
    ...       [-1.3995],
        [-1.4208],
        [-1.2292],
        [-1.5645]], dtype=torch.float64, grad_fn=<NegBackward>)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([100, 1])
SDDMM(message func: u_sub_e)
_____________________ test_sddmm[idtype1-sub-e-u-shp4-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'u', rhs_target = 'e', msg = 'sub'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add'
lhs_data = tensor([[1.3745],
        [1.4401],
        [1.1289],
        [1.6580],
        [1.0698],
        [1.3260],
        [1...450],
        [1.8924],
        [1.5668],
        [1.6685],
        [1.4259]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.5056],
        [-1.6474],
        [-1.6680],
        [-1.3557],
        [-1.0207],
        [-1.2172],
    ...       [-1.2641],
        [-1.5277],
        [-1.0542],
        [-1.9554]], dtype=torch.float64, grad_fn=<NegBackward>)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([300, 1])
SDDMM(message func: u_sub_e)
_____________________ test_sddmm[idtype1-sub-e-v-shp0-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': ...torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'v', rhs_target = 'e'
msg = 'sub', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add'
lhs_data = tensor([[[[[[1.4056],
            [1.5701],
            [1.1921]]],


          [[[1.3110],
            [1.0473],
    ...9]]],


          [[[1.5644],
            [1.8543],
            [1.9310]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[-1.4156]],

           [[-1.8976]],

           [[-1.1221]]]],



         [[[[-1.2603]],

           [[-...       [[[[-1.6568]],

           [[-1.6863]],

           [[-1.5126]]]]]], dtype=torch.float64, grad_fn=<NegBackward>)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([100, 4, 1, 3, 1, 1])
SDDMM(message func: v_sub_e)
_____________________ test_sddmm[idtype1-sub-e-v-shp0-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'v', rhs_target = 'e'
msg = 'sub', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add'
lhs_data = tensor([[[[[[1.6560],
            [1.8678],
            [1.9362]]],


          [[[1.0783],
            [1.7030],
    ...7]]],


          [[[1.6662],
            [1.0966],
            [1.2753]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[-1.7283]],

           [[-1.9641]],

           [[-1.2553]]]],



         [[[[-1.0931]],

           [[-...       [[[[-1.4369]],

           [[-1.7470]],

           [[-1.4667]]]]]], dtype=torch.float64, grad_fn=<NegBackward>)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([40, 1, 2, 1, 3, 1]), rhs shape: torch.Size([300, 4, 1, 3, 1, 1])
SDDMM(message func: v_sub_e)
_____________________ test_sddmm[idtype1-sub-e-v-shp1-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Sch...pe=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'v', rhs_target = 'e'
msg = 'sub', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add'
lhs_data = tensor([[[[[1.8771, 1.3065, 1.2376,  ..., 1.9291, 1.7646, 1.2378]],

          [[1.3224, 1.0322, 1.1939,  ..., 1.0556,...
          [[1.2776, 1.8818, 1.7425,  ..., 1.6335, 1.0581, 1.7346]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[-1.1554, -1.1808, -1.7767,  ..., -1.2359, -1.8785, -1.6293],
           [-1.0831, -1.7309, -1.0048,  ..., ...   [-1.8876, -1.4873, -1.9862,  ..., -1.0448, -1.3683, -1.2352]]]]],
       dtype=torch.float64, grad_fn=<NegBackward>)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([100, 1, 3, 7, 7])
SDDMM(message func: v_sub_e)
_____________________ test_sddmm[idtype1-sub-e-v-shp1-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'v', rhs_target = 'e'
msg = 'sub', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add'
lhs_data = tensor([[[[[1.7421, 1.9351, 1.2814,  ..., 1.8339, 1.9784, 1.6842]],

          [[1.8969, 1.6111, 1.2531,  ..., 1.5094,...
          [[1.9547, 1.7263, 1.5784,  ..., 1.0959, 1.8942, 1.1778]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[-1.8871, -1.0097, -1.8559,  ..., -1.9213, -1.5116, -1.6298],
           [-1.5807, -1.0560, -1.5736,  ..., ...   [-1.0887, -1.6643, -1.9875,  ..., -1.3280, -1.6337, -1.2682]]]]],
       dtype=torch.float64, grad_fn=<NegBackward>)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([40, 5, 3, 1, 7]), rhs shape: torch.Size([300, 1, 3, 7, 7])
SDDMM(message func: v_sub_e)
_____________________ test_sddmm[idtype1-sub-e-v-shp2-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme...dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'v', rhs_target = 'e', msg = 'sub'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add'
lhs_data = tensor([[[[1.4523, 1.2404, 1.1998],
          [1.3882, 1.6579, 1.0651],
          [1.6645, 1.1100, 1.3140]]],


      ...2],
          [1.7227, 1.6477, 1.3170],
          [1.2473, 1.3413, 1.3127]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[-1.5020, -1.6302, -1.1276]],

         [[-1.8809, -1.7202, -1.5407]],

         [[-1.9171, -1.5923, -1.0611...1141, -1.8167, -1.0931]],

         [[-1.8856, -1.0791, -1.8737]]]], dtype=torch.float64,
       grad_fn=<NegBackward>)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([100, 4, 1, 3])
SDDMM(message func: v_sub_e)
_____________________ test_sddmm[idtype1-sub-e-v-shp2-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'v', rhs_target = 'e', msg = 'sub'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add'
lhs_data = tensor([[[[1.6281, 1.9924, 1.6818],
          [1.1010, 1.7657, 1.3831],
          [1.6821, 1.5664, 1.9120]]],


      ...7],
          [1.9046, 1.4928, 1.8350],
          [1.7803, 1.1957, 1.4271]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[-1.3609, -1.0262, -1.5249]],

         [[-1.0272, -1.8908, -1.1984]],

         [[-1.0240, -1.9879, -1.8948...8123, -1.0174, -1.7080]],

         [[-1.3416, -1.2324, -1.3625]]]], dtype=torch.float64,
       grad_fn=<NegBackward>)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([40, 1, 3, 3]), rhs shape: torch.Size([300, 4, 1, 3])
SDDMM(message func: v_sub_e)
_____________________ test_sddmm[idtype1-sub-e-v-shp3-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'v', rhs_target = 'e', msg = 'sub'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add'
lhs_data = tensor([[1.6483, 1.4758, 1.7637],
        [1.4423, 1.4533, 1.6903],
        [1.8065, 1.1851, 1.8254],
        [1.6056,... 1.1103],
        [1.9560, 1.8169, 1.2261],
        [1.9788, 1.9089, 1.4076]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.7707, -1.0283, -1.0313],
        [-1.8328, -1.3360, -1.2937],
        [-1.5996, -1.8474, -1.1870],
       ...  [-1.9474, -1.8660, -1.1462],
        [-1.3228, -1.1862, -1.0976]], dtype=torch.float64,
       grad_fn=<NegBackward>)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([100, 3])
SDDMM(message func: v_sub_e)
_____________________ test_sddmm[idtype1-sub-e-v-shp3-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'v', rhs_target = 'e', msg = 'sub'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add'
lhs_data = tensor([[1.0714, 1.6349, 1.6321],
        [1.8641, 1.0605, 1.5391],
        [1.5364, 1.5679, 1.8272],
        [1.5644,... 1.4511],
        [1.9220, 1.9167, 1.7309],
        [1.3592, 1.1486, 1.9225]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.1548, -1.6136, -1.8671],
        [-1.2796, -1.3754, -1.7014],
        [-1.0747, -1.4451, -1.7041],
       ...  [-1.8445, -1.5779, -1.5459],
        [-1.8664, -1.0707, -1.5644]], dtype=torch.float64,
       grad_fn=<NegBackward>)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([40, 3]), rhs shape: torch.Size([300, 3])
SDDMM(message func: v_sub_e)
_____________________ test_sddmm[idtype1-sub-e-v-shp4-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'v', rhs_target = 'e', msg = 'sub'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'add'
lhs_data = tensor([[1.6762],
        [1.1178],
        [1.0401],
        [1.5047],
        [1.2785],
        [1.5639],
        [1...310],
        [1.7694],
        [1.2187],
        [1.0044],
        [1.0395]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.6668],
        [-1.9723],
        [-1.6517],
        [-1.2654],
        [-1.8268],
        [-1.4945],
    ...       [-1.7611],
        [-1.4391],
        [-1.4630],
        [-1.8685]], dtype=torch.float64, grad_fn=<NegBackward>)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([100, 1])
SDDMM(message func: v_sub_e)
_____________________ test_sddmm[idtype1-sub-e-v-shp4-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'v', rhs_target = 'e', msg = 'sub'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'add'
lhs_data = tensor([[1.7641],
        [1.6165],
        [1.8629],
        [1.7053],
        [1.8771],
        [1.8627],
        [1...323],
        [1.1062],
        [1.3414],
        [1.8614],
        [1.8025]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[-1.5352],
        [-1.8544],
        [-1.0516],
        [-1.7146],
        [-1.8005],
        [-1.0210],
    ...       [-1.0909],
        [-1.3832],
        [-1.6541],
        [-1.2847]], dtype=torch.float64, grad_fn=<NegBackward>)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([40, 1]), rhs shape: torch.Size([300, 1])
SDDMM(message func: v_sub_e)
_____________________ test_sddmm[idtype1-mul-u-v-shp0-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': ...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'v', rhs_target = 'u'
msg = 'mul', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul'
lhs_data = tensor([[[[[[1.7787],
            [1.4521],
            [1.2007]]],


          [[[1.5562],
            [1.3736],
    ...1]]],


          [[[1.7512],
            [1.9998],
            [1.5599]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.7476]],

           [[1.8222]],

           [[1.4793]]]],



         [[[[1.0911]],

           [[1.341...



         [[[[1.6953]],

           [[1.2448]],

           [[1.3325]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: v_mul_u)
_____________________ test_sddmm[idtype1-mul-u-v-shp0-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'v', rhs_target = 'u'
msg = 'mul', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul'
lhs_data = tensor([[[[[[1.5668],
            [1.3724],
            [1.5403]]],


          [[[1.9326],
            [1.6528],
    ...4]]],


          [[[1.5961],
            [1.5779],
            [1.7453]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.7751]],

           [[1.4951]],

           [[1.7505]]]],



         [[[[1.5637]],

           [[1.147...



         [[[[1.1727]],

           [[1.6577]],

           [[1.8926]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([40, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: v_mul_u)
_____________________ test_sddmm[idtype1-mul-u-v-shp1-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Sch...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'v', rhs_target = 'u'
msg = 'mul', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul'
lhs_data = tensor([[[[[1.6669, 1.4709, 1.3502,  ..., 1.6760, 1.9194, 1.7501]],

          [[1.7190, 1.7599, 1.5738,  ..., 1.7121,...
          [[1.9303, 1.2870, 1.0325,  ..., 1.3675, 1.8683, 1.1418]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.4817, 1.8160, 1.3749,  ..., 1.1790, 1.2399, 1.8836],
           [1.6524, 1.3047, 1.5435,  ..., 1.1600, 1...
           [1.6829, 1.0316, 1.2869,  ..., 1.1500, 1.4020, 1.4034]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: v_mul_u)
_____________________ test_sddmm[idtype1-mul-u-v-shp1-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'v', rhs_target = 'u'
msg = 'mul', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul'
lhs_data = tensor([[[[[1.5738, 1.1004, 1.4082,  ..., 1.5219, 1.2581, 1.5546]],

          [[1.5505, 1.2458, 1.2889,  ..., 1.8906,...
          [[1.0605, 1.9541, 1.7632,  ..., 1.3258, 1.6364, 1.8737]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.5424, 1.5585, 1.2291,  ..., 1.2602, 1.1737, 1.1702],
           [1.9105, 1.7044, 1.8732,  ..., 1.3988, 1...
           [1.7424, 1.0190, 1.0684,  ..., 1.9680, 1.7334, 1.1538]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([40, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: v_mul_u)
_____________________ test_sddmm[idtype1-mul-u-v-shp2-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'v', rhs_target = 'u', msg = 'mul'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul'
lhs_data = tensor([[[[1.4715, 1.0878, 1.4817],
          [1.3317, 1.2222, 1.8158],
          [1.1740, 1.6384, 1.7323]]],


      ...9],
          [1.9976, 1.8154, 1.3285],
          [1.2037, 1.0261, 1.6133]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.1280, 1.2938, 1.3190]],

         [[1.5742, 1.6153, 1.9799]],

         [[1.8381, 1.2140, 1.6567]],

    ...

         [[1.5114, 1.6171, 1.4929]],

         [[1.7910, 1.4763, 1.8406]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: v_mul_u)
_____________________ test_sddmm[idtype1-mul-u-v-shp2-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'v', rhs_target = 'u', msg = 'mul'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul'
lhs_data = tensor([[[[1.2785, 1.8520, 1.2977],
          [1.1036, 1.9371, 1.5487],
          [1.0514, 1.0439, 1.1750]]],


      ...6],
          [1.8539, 1.7685, 1.2290],
          [1.7869, 1.0038, 1.7770]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.9758, 1.2727, 1.7402]],

         [[1.8349, 1.7189, 1.1814]],

         [[1.9275, 1.5968, 1.7652]],

    ...

         [[1.6536, 1.4018, 1.5574]],

         [[1.0974, 1.7586, 1.1138]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([40, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: v_mul_u)
_____________________ test_sddmm[idtype1-mul-u-v-shp3-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'v', rhs_target = 'u', msg = 'mul'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul'
lhs_data = tensor([[1.0989, 1.7847, 1.6674],
        [1.2205, 1.1681, 1.8298],
        [1.3430, 1.8901, 1.4197],
        [1.6268,... 1.9569],
        [1.5157, 1.9466, 1.3382],
        [1.9357, 1.6787, 1.0558]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.1655, 1.2264, 1.1719],
        [1.7137, 1.6528, 1.5447],
        [1.4748, 1.3955, 1.5920],
        [1.7262,... 1.6938],
        [1.2340, 1.2767, 1.8034],
        [1.6512, 1.4512, 1.5102]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: v_mul_u)
_____________________ test_sddmm[idtype1-mul-u-v-shp3-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'v', rhs_target = 'u', msg = 'mul'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul'
lhs_data = tensor([[1.0091, 1.8343, 1.2350],
        [1.0361, 1.2128, 1.5819],
        [1.9829, 1.0141, 1.8584],
        [1.1760,... 1.3860],
        [1.9954, 1.8739, 1.2644],
        [1.0534, 1.6242, 1.8251]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.9991, 1.5945, 1.4981],
        [1.5311, 1.4832, 1.8950],
        [1.5726, 1.3015, 1.6665],
        [1.5619,... 1.0557],
        [1.4517, 1.7113, 1.2186],
        [1.8020, 1.6131, 1.5585]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([40, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: v_mul_u)
_____________________ test_sddmm[idtype1-mul-u-v-shp4-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'v', rhs_target = 'u', msg = 'mul'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul'
lhs_data = tensor([[1.8514],
        [1.9614],
        [1.0299],
        [1.7878],
        [1.7093],
        [1.5252],
        [1...304],
        [1.3203],
        [1.7522],
        [1.5695],
        [1.8629]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.7895],
        [1.4785],
        [1.1928],
        [1.0555],
        [1.4862],
        [1.5159],
        [1...902],
        [1.5637],
        [1.0352],
        [1.9255],
        [1.5763]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: v_mul_u)
_____________________ test_sddmm[idtype1-mul-u-v-shp4-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'v', rhs_target = 'u', msg = 'mul'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul'
lhs_data = tensor([[1.4124],
        [1.8754],
        [1.3747],
        [1.2632],
        [1.3599],
        [1.0120],
        [1...557],
        [1.8140],
        [1.1345],
        [1.3646],
        [1.1662]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.7591],
        [1.8211],
        [1.4248],
        [1.6079],
        [1.5766],
        [1.2623],
        [1...226],
        [1.1430],
        [1.3997],
        [1.2405],
        [1.1861]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([40, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: v_mul_u)
_____________________ test_sddmm[idtype1-mul-u-e-shp0-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...torch.float64), 'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'e', rhs_target = 'u'
msg = 'mul', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul'
lhs_data = tensor([[[[[[1.3854],
            [1.9519],
            [1.6396]]],


          [[[1.2518],
            [1.0485],
    ...2]]],


          [[[1.2581],
            [1.3244],
            [1.4813]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.2310]],

           [[1.0552]],

           [[1.2162]]]],



         [[[[1.1846]],

           [[1.090...



         [[[[1.0394]],

           [[1.8507]],

           [[1.5595]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([100, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: e_mul_u)
_____________________ test_sddmm[idtype1-mul-u-e-shp0-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'e', rhs_target = 'u'
msg = 'mul', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul'
lhs_data = tensor([[[[[[1.4642],
            [1.3755],
            [1.1536]]],


          [[[1.7864],
            [1.4307],
    ...3]]],


          [[[1.4835],
            [1.3961],
            [1.7106]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.2389]],

           [[1.1663]],

           [[1.9337]]]],



         [[[[1.3079]],

           [[1.808...



         [[[[1.8741]],

           [[1.4267]],

           [[1.9497]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([300, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: e_mul_u)
_____________________ test_sddmm[idtype1-mul-u-e-shp1-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...pe=torch.float64), 'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'e', rhs_target = 'u'
msg = 'mul', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul'
lhs_data = tensor([[[[[1.3991, 1.0989, 1.1444,  ..., 1.6576, 1.9765, 1.5259]],

          [[1.2263, 1.7218, 1.7377,  ..., 1.0352,...
          [[1.8267, 1.7683, 1.7574,  ..., 1.5849, 1.2964, 1.5571]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.2335, 1.4929, 1.1598,  ..., 1.9702, 1.1826, 1.5263],
           [1.9596, 1.9163, 1.3799,  ..., 1.3414, 1...
           [1.9004, 1.7020, 1.3166,  ..., 1.6751, 1.1374, 1.3156]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([100, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: e_mul_u)
_____________________ test_sddmm[idtype1-mul-u-e-shp1-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'e', rhs_target = 'u'
msg = 'mul', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul'
lhs_data = tensor([[[[[1.8714, 1.8575, 1.0125,  ..., 1.5100, 1.6216, 1.2561]],

          [[1.8130, 1.8588, 1.2249,  ..., 1.0622,...
          [[1.0455, 1.1080, 1.0686,  ..., 1.1299, 1.5899, 1.4423]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.6262, 1.0981, 1.1848,  ..., 1.8794, 1.7959, 1.2538],
           [1.4129, 1.3388, 1.9704,  ..., 1.7765, 1...
           [1.6579, 1.7841, 1.9555,  ..., 1.1068, 1.2123, 1.6847]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([300, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: e_mul_u)
_____________________ test_sddmm[idtype1-mul-u-e-shp2-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...dtype=torch.float64), 'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'e', rhs_target = 'u', msg = 'mul'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul'
lhs_data = tensor([[[[1.2268, 1.5672, 1.5478],
          [1.2674, 1.9871, 1.0027],
          [1.5210, 1.2163, 1.8255]]],


      ...8],
          [1.6756, 1.7368, 1.3513],
          [1.1281, 1.5607, 1.4968]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.9942, 1.4701, 1.7212]],

         [[1.0654, 1.4718, 1.0318]],

         [[1.7912, 1.0175, 1.9801]],

    ...

         [[1.5640, 1.6373, 1.1098]],

         [[1.2129, 1.5409, 1.6227]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([100, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: e_mul_u)
_____________________ test_sddmm[idtype1-mul-u-e-shp2-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'e', rhs_target = 'u', msg = 'mul'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul'
lhs_data = tensor([[[[1.8835, 1.5106, 1.8899],
          [1.4797, 1.7410, 1.6716],
          [1.9968, 1.8038, 1.7091]]],


      ...5],
          [1.9241, 1.9951, 1.3449],
          [1.7242, 1.8281, 1.0594]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.2986, 1.3706, 1.5430]],

         [[1.7057, 1.7337, 1.1026]],

         [[1.8655, 1.6313, 1.7918]],

    ...

         [[1.3972, 1.9239, 1.4840]],

         [[1.0105, 1.9309, 1.0577]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([300, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: e_mul_u)
_____________________ test_sddmm[idtype1-mul-u-e-shp3-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'e', rhs_target = 'u', msg = 'mul'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul'
lhs_data = tensor([[1.9786, 1.7374, 1.9428],
        [1.5016, 1.8362, 1.6301],
        [1.5201, 1.1490, 1.1774],
        [1.5926,... 1.9630],
        [1.0444, 1.1395, 1.7414],
        [1.1903, 1.3424, 1.0212]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.7885, 1.1472, 1.7614],
        [1.9835, 1.6607, 1.3032],
        [1.8130, 1.4084, 1.1771],
        [1.1351,... 1.6904],
        [1.2697, 1.9368, 1.9555],
        [1.1745, 1.2780, 1.5579]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([100, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: e_mul_u)
_____________________ test_sddmm[idtype1-mul-u-e-shp3-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'e', rhs_target = 'u', msg = 'mul'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul'
lhs_data = tensor([[1.6067, 1.6205, 1.6731],
        [1.4293, 1.2572, 1.5483],
        [1.9376, 1.0875, 1.8784],
        [1.4495,... 1.0222],
        [1.8783, 1.9627, 1.8379],
        [1.6664, 1.8075, 1.4426]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.6854, 1.3732, 1.9222],
        [1.3698, 1.5666, 1.7152],
        [1.0346, 1.6259, 1.4296],
        [1.1183,... 1.2803],
        [1.1913, 1.6895, 1.0111],
        [1.3378, 1.4486, 1.0581]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([300, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: e_mul_u)
_____________________ test_sddmm[idtype1-mul-u-e-shp4-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'e', rhs_target = 'u', msg = 'mul'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul'
lhs_data = tensor([[1.4782],
        [1.4866],
        [1.8678],
        [1.4174],
        [1.9703],
        [1.6101],
        [1...451],
        [1.9476],
        [1.5639],
        [1.4552],
        [1.6073]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.6875],
        [1.1658],
        [1.5206],
        [1.1277],
        [1.0370],
        [1.4307],
        [1...986],
        [1.2311],
        [1.1092],
        [1.4202],
        [1.3762]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([100, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: e_mul_u)
_____________________ test_sddmm[idtype1-mul-u-e-shp4-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'e', rhs_target = 'u', msg = 'mul'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul'
lhs_data = tensor([[1.0902],
        [1.6350],
        [1.1043],
        [1.8910],
        [1.0808],
        [1.4964],
        [1...753],
        [1.9690],
        [1.2696],
        [1.9771],
        [1.1422]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.6844],
        [1.9944],
        [1.8537],
        [1.8263],
        [1.1202],
        [1.2829],
        [1...249],
        [1.4571],
        [1.6680],
        [1.0194],
        [1.9979]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([300, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: e_mul_u)
_____________________ test_sddmm[idtype1-mul-v-u-shp0-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': ...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'u', rhs_target = 'v'
msg = 'mul', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul'
lhs_data = tensor([[[[[[1.8668],
            [1.4754],
            [1.9082]]],


          [[[1.6112],
            [1.1442],
    ...2]]],


          [[[1.3700],
            [1.1320],
            [1.2164]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.4848]],

           [[1.5027]],

           [[1.4367]]]],



         [[[[1.9293]],

           [[1.392...



         [[[[1.2344]],

           [[1.1072]],

           [[1.8930]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: u_mul_v)
_____________________ test_sddmm[idtype1-mul-v-u-shp0-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'u', rhs_target = 'v'
msg = 'mul', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul'
lhs_data = tensor([[[[[[1.8950],
            [1.2756],
            [1.3521]]],


          [[[1.9388],
            [1.6291],
    ...8]]],


          [[[1.4780],
            [1.8980],
            [1.9817]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.4746]],

           [[1.4285]],

           [[1.8713]]]],



         [[[[1.2079]],

           [[1.639...



         [[[[1.8496]],

           [[1.2931]],

           [[1.9952]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([40, 4, 1, 3, 1, 1])
SDDMM(message func: u_mul_v)
_____________________ test_sddmm[idtype1-mul-v-u-shp1-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Sch...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'u', rhs_target = 'v'
msg = 'mul', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul'
lhs_data = tensor([[[[[1.9236, 1.0743, 1.4444,  ..., 1.3077, 1.6208, 1.6680]],

          [[1.4552, 1.7461, 1.9188,  ..., 1.3765,...
          [[1.6586, 1.0983, 1.8119,  ..., 1.1361, 1.6990, 1.0830]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.5830, 1.0039, 1.1217,  ..., 1.9640, 1.4453, 1.0192],
           [1.5297, 1.0367, 1.8214,  ..., 1.4523, 1...
           [1.2861, 1.4303, 1.7729,  ..., 1.7188, 1.0989, 1.1608]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: u_mul_v)
_____________________ test_sddmm[idtype1-mul-v-u-shp1-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'u', rhs_target = 'v'
msg = 'mul', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul'
lhs_data = tensor([[[[[1.9709, 1.8779, 1.1597,  ..., 1.7810, 1.7763, 1.1718]],

          [[1.0849, 1.8135, 1.9027,  ..., 1.8702,...
          [[1.5438, 1.9210, 1.9572,  ..., 1.9573, 1.7276, 1.1259]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.0109, 1.1547, 1.2643,  ..., 1.2379, 1.7375, 1.7215],
           [1.3536, 1.0327, 1.9459,  ..., 1.1945, 1...
           [1.5211, 1.0722, 1.1555,  ..., 1.6975, 1.4597, 1.4347]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([40, 1, 3, 7, 7])
SDDMM(message func: u_mul_v)
_____________________ test_sddmm[idtype1-mul-v-u-shp2-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'u', rhs_target = 'v', msg = 'mul'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul'
lhs_data = tensor([[[[1.1271, 1.4449, 1.7561],
          [1.8270, 1.5435, 1.1492],
          [1.9740, 1.7289, 1.2275]]],


      ...2],
          [1.0359, 1.9441, 1.7003],
          [1.0783, 1.9920, 1.6230]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.5094, 1.5366, 1.3000]],

         [[1.4286, 1.6809, 1.4311]],

         [[1.2286, 1.3748, 1.8456]],

    ...

         [[1.6538, 1.5953, 1.2834]],

         [[1.8598, 1.7512, 1.0290]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: u_mul_v)
_____________________ test_sddmm[idtype1-mul-v-u-shp2-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'u', rhs_target = 'v', msg = 'mul'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul'
lhs_data = tensor([[[[1.6562, 1.9572, 1.6798],
          [1.8059, 1.6668, 1.3412],
          [1.7000, 1.0931, 1.2953]]],


      ...4],
          [1.1994, 1.8019, 1.1896],
          [1.6033, 1.5100, 1.3070]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.8781, 1.5960, 1.4969]],

         [[1.0821, 1.2513, 1.0302]],

         [[1.8768, 1.8967, 1.1073]],

    ...

         [[1.7913, 1.8275, 1.9413]],

         [[1.0196, 1.0055, 1.7099]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([40, 4, 1, 3])
SDDMM(message func: u_mul_v)
_____________________ test_sddmm[idtype1-mul-v-u-shp3-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'u', rhs_target = 'v', msg = 'mul'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul'
lhs_data = tensor([[1.0708, 1.4712, 1.0718],
        [1.4940, 1.1141, 1.3583],
        [1.1039, 1.6606, 1.5475],
        [1.0872,... 1.9240],
        [1.4698, 1.2142, 1.2348],
        [1.7228, 1.7563, 1.7187]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.6200, 1.0850, 1.2068],
        [1.1036, 1.6878, 1.7269],
        [1.8639, 1.5205, 1.4895],
        [1.5343,... 1.0668],
        [1.2700, 1.2286, 1.7354],
        [1.1518, 1.9806, 1.8310]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: u_mul_v)
_____________________ test_sddmm[idtype1-mul-v-u-shp3-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'u', rhs_target = 'v', msg = 'mul'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul'
lhs_data = tensor([[1.8526, 1.8628, 1.3842],
        [1.3539, 1.4571, 1.0369],
        [1.7153, 1.7374, 1.2340],
        [1.5772,... 1.9098],
        [1.0316, 1.5976, 1.6375],
        [1.1613, 1.3665, 1.3235]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.7204, 1.7122, 1.6627],
        [1.4033, 1.9065, 1.1243],
        [1.2757, 1.8863, 1.8481],
        [1.9902,... 1.4510],
        [1.2278, 1.5195, 1.3484],
        [1.4602, 1.5960, 1.1746]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([40, 3])
SDDMM(message func: u_mul_v)
_____________________ test_sddmm[idtype1-mul-v-u-shp4-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'u', rhs_target = 'v', msg = 'mul'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul'
lhs_data = tensor([[1.7447],
        [1.0977],
        [1.1605],
        [1.1699],
        [1.7280],
        [1.1955],
        [1...945],
        [1.0206],
        [1.8104],
        [1.9290],
        [1.2434]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.3151],
        [1.8780],
        [1.0511],
        [1.5927],
        [1.7643],
        [1.5577],
        [1...291],
        [1.7006],
        [1.0785],
        [1.4544],
        [1.6991]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: u_mul_v)
_____________________ test_sddmm[idtype1-mul-v-u-shp4-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'u', rhs_target = 'v', msg = 'mul'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul'
lhs_data = tensor([[1.4797],
        [1.5080],
        [1.3000],
        [1.2199],
        [1.6334],
        [1.1926],
        [1...438],
        [1.3532],
        [1.2832],
        [1.0975],
        [1.2944]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.6318],
        [1.9717],
        [1.4260],
        [1.4815],
        [1.2073],
        [1.8216],
        [1...118],
        [1.7150],
        [1.9249],
        [1.7056],
        [1.6486]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([40, 1])
SDDMM(message func: u_mul_v)
_____________________ test_sddmm[idtype1-mul-v-e-shp0-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...torch.float64), 'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'e', rhs_target = 'v'
msg = 'mul', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul'
lhs_data = tensor([[[[[[1.2870],
            [1.1709],
            [1.1443]]],


          [[[1.0512],
            [1.6898],
    ...3]]],


          [[[1.9058],
            [1.8941],
            [1.2694]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.8389]],

           [[1.8405]],

           [[1.5054]]]],



         [[[[1.5595]],

           [[1.642...



         [[[[1.4130]],

           [[1.8839]],

           [[1.2010]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([100, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: e_mul_v)
_____________________ test_sddmm[idtype1-mul-v-e-shp0-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'e', rhs_target = 'v'
msg = 'mul', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul'
lhs_data = tensor([[[[[[1.6886],
            [1.4546],
            [1.3302]]],


          [[[1.2906],
            [1.4352],
    ...5]]],


          [[[1.9333],
            [1.3627],
            [1.8411]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.6726]],

           [[1.3669]],

           [[1.5893]]]],



         [[[[1.4796]],

           [[1.220...



         [[[[1.5919]],

           [[1.2949]],

           [[1.8926]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([300, 1, 2, 1, 3, 1]), rhs shape: torch.Size([40, 4, 1, 3, 1, 1])
SDDMM(message func: e_mul_v)
_____________________ test_sddmm[idtype1-mul-v-e-shp1-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...pe=torch.float64), 'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'e', rhs_target = 'v'
msg = 'mul', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul'
lhs_data = tensor([[[[[1.6259, 1.7264, 1.7769,  ..., 1.4884, 1.3189, 1.6583]],

          [[1.5677, 1.4254, 1.7889,  ..., 1.4070,...
          [[1.8092, 1.3862, 1.8646,  ..., 1.1597, 1.9357, 1.5826]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.7689, 1.2032, 1.2270,  ..., 1.0566, 1.1873, 1.8577],
           [1.6220, 1.2643, 1.9575,  ..., 1.4785, 1...
           [1.9379, 1.8563, 1.9795,  ..., 1.8919, 1.5956, 1.1825]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([100, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: e_mul_v)
_____________________ test_sddmm[idtype1-mul-v-e-shp1-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'e', rhs_target = 'v'
msg = 'mul', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul'
lhs_data = tensor([[[[[1.3209, 1.3462, 1.2890,  ..., 1.6273, 1.4436, 1.4973]],

          [[1.5195, 1.3820, 1.1296,  ..., 1.0644,...
          [[1.1304, 1.9924, 1.5674,  ..., 1.0432, 1.4992, 1.2157]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.8544, 1.6506, 1.6443,  ..., 1.6427, 1.1242, 1.2461],
           [1.5557, 1.1699, 1.3803,  ..., 1.0661, 1...
           [1.2809, 1.6251, 1.6316,  ..., 1.9050, 1.5734, 1.2653]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([300, 5, 3, 1, 7]), rhs shape: torch.Size([40, 1, 3, 7, 7])
SDDMM(message func: e_mul_v)
_____________________ test_sddmm[idtype1-mul-v-e-shp2-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...dtype=torch.float64), 'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'e', rhs_target = 'v', msg = 'mul'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul'
lhs_data = tensor([[[[1.3258, 1.8935, 1.7949],
          [1.3176, 1.2671, 1.5264],
          [1.5445, 1.0973, 1.8597]]],


      ...6],
          [1.2539, 1.2207, 1.2083],
          [1.3331, 1.5108, 1.9461]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.5799, 1.0558, 1.9009]],

         [[1.4771, 1.2857, 1.6309]],

         [[1.0133, 1.8170, 1.1772]],

    ...

         [[1.2495, 1.2330, 1.0486]],

         [[1.1928, 1.3125, 1.7528]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([100, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: e_mul_v)
_____________________ test_sddmm[idtype1-mul-v-e-shp2-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'e', rhs_target = 'v', msg = 'mul'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul'
lhs_data = tensor([[[[1.0455, 1.2440, 1.4769],
          [1.3110, 1.9650, 1.1153],
          [1.5924, 1.5585, 1.9389]]],


      ...6],
          [1.8498, 1.3496, 1.1115],
          [1.8914, 1.9516, 1.4108]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.2030, 1.0832, 1.8257]],

         [[1.0471, 1.6081, 1.4972]],

         [[1.3391, 1.0673, 1.1552]],

    ...

         [[1.0737, 1.1273, 1.4478]],

         [[1.8735, 1.9425, 1.3258]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([300, 1, 3, 3]), rhs shape: torch.Size([40, 4, 1, 3])
SDDMM(message func: e_mul_v)
_____________________ test_sddmm[idtype1-mul-v-e-shp3-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'e', rhs_target = 'v', msg = 'mul'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul'
lhs_data = tensor([[1.8055, 1.7229, 1.4054],
        [1.8238, 1.9218, 1.1533],
        [1.6310, 1.6381, 1.2830],
        [1.7855,... 1.8201],
        [1.8971, 1.6831, 1.4362],
        [1.2436, 1.0087, 1.6565]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.2965, 1.7617, 1.7138],
        [1.1397, 1.8248, 1.5737],
        [1.3713, 1.6038, 1.5635],
        [1.1334,... 1.8840],
        [1.6291, 1.6371, 1.5674],
        [1.4461, 1.9196, 1.3486]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([100, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: e_mul_v)
_____________________ test_sddmm[idtype1-mul-v-e-shp3-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'e', rhs_target = 'v', msg = 'mul'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul'
lhs_data = tensor([[1.6056, 1.0287, 1.8608],
        [1.1483, 1.0951, 1.4827],
        [1.0872, 1.0822, 1.4508],
        [1.9490,... 1.6094],
        [1.2282, 1.6168, 1.2994],
        [1.9344, 1.0077, 1.2450]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.4011, 1.4621, 1.4492],
        [1.0278, 1.0604, 1.8991],
        [1.6385, 1.2406, 1.2096],
        [1.8295,... 1.0127],
        [1.5315, 1.1140, 1.2797],
        [1.3542, 1.6127, 1.5797]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([300, 3]), rhs shape: torch.Size([40, 3])
SDDMM(message func: e_mul_v)
_____________________ test_sddmm[idtype1-mul-v-e-shp4-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'e', rhs_target = 'v', msg = 'mul'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul'
lhs_data = tensor([[1.8010],
        [1.2553],
        [1.7059],
        [1.1071],
        [1.7445],
        [1.1315],
        [1...248],
        [1.6432],
        [1.7878],
        [1.1287],
        [1.1543]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.5986],
        [1.2796],
        [1.5926],
        [1.2006],
        [1.6274],
        [1.7098],
        [1...465],
        [1.8434],
        [1.9635],
        [1.0160],
        [1.2263]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([100, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: e_mul_v)
_____________________ test_sddmm[idtype1-mul-v-e-shp4-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'e', rhs_target = 'v', msg = 'mul'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul'
lhs_data = tensor([[1.2812],
        [1.2206],
        [1.9070],
        [1.5653],
        [1.5974],
        [1.3065],
        [1...364],
        [1.9225],
        [1.0521],
        [1.0072],
        [1.5875]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.4659],
        [1.7129],
        [1.8669],
        [1.2812],
        [1.2874],
        [1.9381],
        [1...273],
        [1.0436],
        [1.0531],
        [1.5536],
        [1.7149]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([300, 1]), rhs shape: torch.Size([40, 1])
SDDMM(message func: e_mul_v)
_____________________ test_sddmm[idtype1-mul-e-u-shp0-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': ...torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'u', rhs_target = 'e'
msg = 'mul', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul'
lhs_data = tensor([[[[[[1.2091],
            [1.6041],
            [1.4995]]],


          [[[1.0597],
            [1.8227],
    ...3]]],


          [[[1.4028],
            [1.3849],
            [1.6059]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.0087]],

           [[1.4885]],

           [[1.1595]]]],



         [[[[1.0937]],

           [[1.434...



         [[[[1.0382]],

           [[1.7318]],

           [[1.9591]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([100, 4, 1, 3, 1, 1])
SDDMM(message func: u_mul_e)Client [1855] waits on 172.17.0.3:58773
Machine (0) group (0) client (0) connect to server successfuly!
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
  NumNodes: 2708
  NumEdges: 10556
  NumFeats: 1433
  NumClasses: 7
  NumTrainingSamples: 140
  NumValidationSamples: 500
  NumTestSamples: 1000
Done loading data from cached files.
  NumNodes: 2708
  NumEdges: 10556
  NumFeats: 1433
  NumClasses: 7
  NumTrainingSamples: 140
  NumValidationSamples: 500
  NumTestSamples: 1000
Done loading data from cached files.
Client[0] in group[0] is exiting...

_____________________ test_sddmm[idtype1-mul-e-u-shp0-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'u', rhs_target = 'e'
msg = 'mul', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul'
lhs_data = tensor([[[[[[1.9871],
            [1.0719],
            [1.7900]]],


          [[[1.9728],
            [1.7734],
    ...7]]],


          [[[1.4477],
            [1.6713],
            [1.9757]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.7089]],

           [[1.6758]],

           [[1.9749]]]],



         [[[[1.9892]],

           [[1.794...



         [[[[1.0219]],

           [[1.5674]],

           [[1.9940]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([300, 4, 1, 3, 1, 1])
SDDMM(message func: u_mul_e)
_____________________ test_sddmm[idtype1-mul-e-u-shp1-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Sch...pe=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'u', rhs_target = 'e'
msg = 'mul', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul'
lhs_data = tensor([[[[[1.7937, 1.0480, 1.2819,  ..., 1.3761, 1.6345, 1.7707]],

          [[1.4677, 1.9887, 1.3785,  ..., 1.0646,...
          [[1.1023, 1.6034, 1.9840,  ..., 1.1519, 1.2516, 1.8233]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.7166, 1.6330, 1.7850,  ..., 1.8094, 1.3726, 1.9899],
           [1.2054, 1.5840, 1.9454,  ..., 1.2852, 1...
           [1.4929, 1.7977, 1.0449,  ..., 1.2169, 1.6463, 1.5163]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([100, 1, 3, 7, 7])
SDDMM(message func: u_mul_e)
_____________________ test_sddmm[idtype1-mul-e-u-shp1-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'u', rhs_target = 'e'
msg = 'mul', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul'
lhs_data = tensor([[[[[1.6966, 1.1328, 1.0874,  ..., 1.5797, 1.1374, 1.7391]],

          [[1.8028, 1.1476, 1.3888,  ..., 1.0179,...
          [[1.1818, 1.5013, 1.4979,  ..., 1.6002, 1.7826, 1.4266]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.4506, 1.4353, 1.0786,  ..., 1.7154, 1.2219, 1.4575],
           [1.4823, 1.5309, 1.4850,  ..., 1.0753, 1...
           [1.9302, 1.7808, 1.5517,  ..., 1.4241, 1.3809, 1.0710]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([300, 1, 3, 7, 7])
SDDMM(message func: u_mul_e)
_____________________ test_sddmm[idtype1-mul-e-u-shp2-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme...dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'u', rhs_target = 'e', msg = 'mul'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul'
lhs_data = tensor([[[[1.4534, 1.4258, 1.7369],
          [1.0604, 1.1058, 1.8237],
          [1.1913, 1.3278, 1.3169]]],


      ...9],
          [1.0812, 1.3206, 1.0860],
          [1.0877, 1.9454, 1.3630]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.4341, 1.0261, 1.9338]],

         [[1.4688, 1.0015, 1.9379]],

         [[1.3759, 1.8256, 1.2408]],

    ...

         [[1.5256, 1.1140, 1.4046]],

         [[1.2537, 1.9752, 1.6830]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([100, 4, 1, 3])
SDDMM(message func: u_mul_e)
_____________________ test_sddmm[idtype1-mul-e-u-shp2-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'u', rhs_target = 'e', msg = 'mul'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul'
lhs_data = tensor([[[[1.0188, 1.0452, 1.9309],
          [1.3120, 1.3906, 1.8953],
          [1.0657, 1.1934, 1.4368]]],


      ...5],
          [1.4418, 1.0814, 1.2674],
          [1.7025, 1.0648, 1.5503]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.1295, 1.0998, 1.8158]],

         [[1.5265, 1.3213, 1.4723]],

         [[1.2278, 1.0919, 1.3069]],

    ...

         [[1.6562, 1.1565, 1.6126]],

         [[1.3345, 1.1664, 1.8727]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([300, 4, 1, 3])
SDDMM(message func: u_mul_e)
_____________________ test_sddmm[idtype1-mul-e-u-shp3-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'u', rhs_target = 'e', msg = 'mul'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul'
lhs_data = tensor([[1.5889, 1.1640, 1.6539],
        [1.9048, 1.0412, 1.7000],
        [1.8646, 1.2115, 1.9822],
        [1.5114,... 1.8914],
        [1.1304, 1.8902, 1.9334],
        [1.6455, 1.5278, 1.9543]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.9642, 1.0704, 1.2753],
        [1.9871, 1.8235, 1.0788],
        [1.9160, 1.7704, 1.7115],
        [1.9959,... 1.6998],
        [1.3684, 1.5637, 1.9469],
        [1.6608, 1.2065, 1.9630]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([100, 3])
SDDMM(message func: u_mul_e)
_____________________ test_sddmm[idtype1-mul-e-u-shp3-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'u', rhs_target = 'e', msg = 'mul'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul'
lhs_data = tensor([[1.4713, 1.7156, 1.3728],
        [1.9633, 1.0473, 1.6408],
        [1.2661, 1.0489, 1.1762],
        [1.9925,... 1.7833],
        [1.2893, 1.3998, 1.0311],
        [1.2190, 1.8728, 1.3518]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.9493, 1.3654, 1.1701],
        [1.2903, 1.3318, 1.5707],
        [1.9637, 1.5268, 1.5765],
        [1.7259,... 1.1955],
        [1.0118, 1.6856, 1.4724],
        [1.7778, 1.2957, 1.1057]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([300, 3])
SDDMM(message func: u_mul_e)
_____________________ test_sddmm[idtype1-mul-e-u-shp4-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'u', rhs_target = 'e', msg = 'mul'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul'
lhs_data = tensor([[1.5438],
        [1.7096],
        [1.8274],
        [1.3256],
        [1.5151],
        [1.9734],
        [1...713],
        [1.3038],
        [1.1269],
        [1.4991],
        [1.6111]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.3976],
        [1.0791],
        [1.5968],
        [1.7456],
        [1.1439],
        [1.3560],
        [1...848],
        [1.6299],
        [1.0889],
        [1.7709],
        [1.0695]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([100, 1])
SDDMM(message func: u_mul_e)
_____________________ test_sddmm[idtype1-mul-e-u-shp4-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'u', rhs_target = 'e', msg = 'mul'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul'
lhs_data = tensor([[1.0825],
        [1.5248],
        [1.1247],
        [1.4179],
        [1.7251],
        [1.7557],
        [1...816],
        [1.6505],
        [1.6945],
        [1.1515],
        [1.4240]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.5619],
        [1.4914],
        [1.4919],
        [1.9564],
        [1.3221],
        [1.0574],
        [1...358],
        [1.8597],
        [1.0817],
        [1.4238],
        [1.0869]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([300, 1])
SDDMM(message func: u_mul_e)
_____________________ test_sddmm[idtype1-mul-e-v-shp0-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': ...torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'v', rhs_target = 'e'
msg = 'mul', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul'
lhs_data = tensor([[[[[[1.8584],
            [1.4760],
            [1.0941]]],


          [[[1.3584],
            [1.4544],
    ...0]]],


          [[[1.1913],
            [1.9811],
            [1.3405]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.7535]],

           [[1.7347]],

           [[1.7556]]]],



         [[[[1.9471]],

           [[1.410...



         [[[[1.2798]],

           [[1.4416]],

           [[1.6569]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([100, 4, 1, 3, 1, 1])
SDDMM(message func: v_mul_e)
_____________________ test_sddmm[idtype1-mul-e-v-shp0-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'v', rhs_target = 'e'
msg = 'mul', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul'
lhs_data = tensor([[[[[[1.4562],
            [1.8866],
            [1.1487]]],


          [[[1.4358],
            [1.9730],
    ...4]]],


          [[[1.4149],
            [1.3870],
            [1.4143]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.0029]],

           [[1.6133]],

           [[1.4935]]]],



         [[[[1.7785]],

           [[1.714...



         [[[[1.2694]],

           [[1.9672]],

           [[1.8345]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([40, 1, 2, 1, 3, 1]), rhs shape: torch.Size([300, 4, 1, 3, 1, 1])
SDDMM(message func: v_mul_e)
_____________________ test_sddmm[idtype1-mul-e-v-shp1-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Sch...pe=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'v', rhs_target = 'e'
msg = 'mul', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul'
lhs_data = tensor([[[[[1.2431, 1.6259, 1.3401,  ..., 1.6003, 1.3444, 1.9827]],

          [[1.1797, 1.0830, 1.5008,  ..., 1.3889,...
          [[1.9644, 1.8727, 1.7211,  ..., 1.4322, 1.2887, 1.2948]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.6684, 1.5015, 1.8387,  ..., 1.0918, 1.3181, 1.2764],
           [1.0682, 1.9373, 1.8768,  ..., 1.2729, 1...
           [1.8431, 1.0173, 1.9516,  ..., 1.3804, 1.1138, 1.2195]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([100, 1, 3, 7, 7])
SDDMM(message func: v_mul_e)
_____________________ test_sddmm[idtype1-mul-e-v-shp1-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'v', rhs_target = 'e'
msg = 'mul', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul'
lhs_data = tensor([[[[[1.1000, 1.9243, 1.5685,  ..., 1.8226, 1.1347, 1.5372]],

          [[1.5681, 1.9130, 1.1376,  ..., 1.4817,...
          [[1.3281, 1.7770, 1.5365,  ..., 1.3003, 1.3167, 1.3647]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.7397, 1.2011, 1.5537,  ..., 1.8968, 1.5075, 1.5036],
           [1.4125, 1.6760, 1.6734,  ..., 1.9716, 1...
           [1.8936, 1.8885, 1.6009,  ..., 1.8860, 1.7184, 1.5022]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([40, 5, 3, 1, 7]), rhs shape: torch.Size([300, 1, 3, 7, 7])
SDDMM(message func: v_mul_e)
_____________________ test_sddmm[idtype1-mul-e-v-shp2-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme...dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'v', rhs_target = 'e', msg = 'mul'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul'
lhs_data = tensor([[[[1.7272, 1.8152, 1.8586],
          [1.3600, 1.4937, 1.1282],
          [1.3696, 1.9875, 1.1373]]],


      ...2],
          [1.0375, 1.1773, 1.0199],
          [1.5049, 1.5845, 1.4399]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.1519, 1.4474, 1.8998]],

         [[1.7206, 1.5584, 1.6795]],

         [[1.5041, 1.4939, 1.4070]],

    ...

         [[1.6481, 1.8856, 1.1983]],

         [[1.2620, 1.1034, 1.6047]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([100, 4, 1, 3])
SDDMM(message func: v_mul_e)
_____________________ test_sddmm[idtype1-mul-e-v-shp2-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'v', rhs_target = 'e', msg = 'mul'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul'
lhs_data = tensor([[[[1.1652, 1.9152, 1.2894],
          [1.2694, 1.3219, 1.2988],
          [1.4443, 1.5157, 1.2813]]],


      ...2],
          [1.3920, 1.8013, 1.1823],
          [1.8857, 1.0490, 1.6126]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.3963, 1.3454, 1.0572]],

         [[1.1330, 1.6814, 1.2676]],

         [[1.5073, 1.0279, 1.5653]],

    ...

         [[1.9617, 1.2057, 1.0059]],

         [[1.6312, 1.4347, 1.6278]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([40, 1, 3, 3]), rhs shape: torch.Size([300, 4, 1, 3])
SDDMM(message func: v_mul_e)
_____________________ test_sddmm[idtype1-mul-e-v-shp3-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'v', rhs_target = 'e', msg = 'mul'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul'
lhs_data = tensor([[1.2438, 1.8076, 1.5272],
        [1.4223, 1.8100, 1.8955],
        [1.3991, 1.7024, 1.3753],
        [1.5545,... 1.0709],
        [1.8536, 1.0304, 1.2993],
        [1.6183, 1.7798, 1.4196]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.1269, 1.5455, 1.5097],
        [1.9961, 1.1527, 1.5716],
        [1.1510, 1.2879, 1.4670],
        [1.3711,... 1.0947],
        [1.1198, 1.9452, 1.2436],
        [1.0655, 1.0889, 1.9567]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([100, 3])
SDDMM(message func: v_mul_e)
_____________________ test_sddmm[idtype1-mul-e-v-shp3-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'v', rhs_target = 'e', msg = 'mul'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul'
lhs_data = tensor([[1.6105, 1.0824, 1.5420],
        [1.9821, 1.8779, 1.0206],
        [1.8394, 1.2962, 1.3744],
        [1.0116,... 1.4157],
        [1.3231, 1.0906, 1.5152],
        [1.2331, 1.3382, 1.2811]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.4688, 1.6436, 1.4887],
        [1.4417, 1.3062, 1.0146],
        [1.4402, 1.9420, 1.2522],
        [1.5830,... 1.9011],
        [1.2394, 1.1699, 1.2997],
        [1.1173, 1.2023, 1.9297]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([40, 3]), rhs shape: torch.Size([300, 3])
SDDMM(message func: v_mul_e)
_____________________ test_sddmm[idtype1-mul-e-v-shp4-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'v', rhs_target = 'e', msg = 'mul'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul'
lhs_data = tensor([[1.0819],
        [1.3501],
        [1.3166],
        [1.7673],
        [1.8311],
        [1.1960],
        [1...065],
        [1.5400],
        [1.6630],
        [1.2732],
        [1.9890]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.1304],
        [1.3613],
        [1.8186],
        [1.2268],
        [1.9539],
        [1.9699],
        [1...635],
        [1.2263],
        [1.6695],
        [1.5738],
        [1.1066]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([100, 1])
SDDMM(message func: v_mul_e)
_____________________ test_sddmm[idtype1-mul-e-v-shp4-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'v', rhs_target = 'e', msg = 'mul'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul'
lhs_data = tensor([[1.2050],
        [1.6784],
        [1.4032],
        [1.6232],
        [1.8185],
        [1.6023],
        [1...702],
        [1.6586],
        [1.8764],
        [1.5565],
        [1.1835]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.8703],
        [1.1009],
        [1.7755],
        [1.6587],
        [1.6815],
        [1.0922],
        [1...968],
        [1.3828],
        [1.9065],
        [1.0431],
        [1.3888]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([40, 1]), rhs shape: torch.Size([300, 1])
SDDMM(message func: v_mul_e)
_____________________ test_sddmm[idtype1-div-u-v-shp0-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': ...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'v', rhs_target = 'u'
msg = 'div', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul'
lhs_data = tensor([[[[[[1.3267],
            [1.5514],
            [1.6718]]],


          [[[1.8335],
            [1.4144],
    ...0]]],


          [[[1.6789],
            [1.6344],
            [1.5417]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[0.5170]],

           [[0.5800]],

           [[0.6923]]]],



         [[[[0.8538]],

           [[0.512...         [[[[0.7897]],

           [[0.6037]],

           [[0.8538]]]]]], dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: v_div_u)
_____________________ test_sddmm[idtype1-div-u-v-shp0-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'v', rhs_target = 'u'
msg = 'div', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul'
lhs_data = tensor([[[[[[1.8091],
            [1.4656],
            [1.7126]]],


          [[[1.6917],
            [1.6925],
    ...1]]],


          [[[1.2970],
            [1.3452],
            [1.4641]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[0.6777]],

           [[0.5402]],

           [[0.5131]]]],



         [[[[0.6018]],

           [[0.743...         [[[[0.5015]],

           [[0.5749]],

           [[0.5426]]]]]], dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([40, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: v_div_u)
_____________________ test_sddmm[idtype1-div-u-v-shp1-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Sch...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'v', rhs_target = 'u'
msg = 'div', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul'
lhs_data = tensor([[[[[1.0940, 1.8524, 1.3775,  ..., 1.0559, 1.3075, 1.8208]],

          [[1.8323, 1.6675, 1.1852,  ..., 1.0768,...
          [[1.7131, 1.8282, 1.9809,  ..., 1.9622, 1.7790, 1.7616]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[0.5675, 0.5541, 0.5607,  ..., 0.5650, 0.5374, 0.5964],
           [0.5795, 0.8476, 0.7291,  ..., 0.5087, 0...        [0.6079, 0.5399, 0.6167,  ..., 0.5313, 0.8087, 0.5988]]]]],
       dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: v_div_u)
_____________________ test_sddmm[idtype1-div-u-v-shp1-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'v', rhs_target = 'u'
msg = 'div', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul'
lhs_data = tensor([[[[[1.9849, 1.3669, 1.6298,  ..., 1.0795, 1.5194, 1.9437]],

          [[1.8941, 1.9953, 1.6785,  ..., 1.8819,...
          [[1.5846, 1.7233, 1.1018,  ..., 1.6152, 1.5017, 1.5194]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[0.5835, 0.8168, 0.5343,  ..., 0.6379, 0.5567, 0.8736],
           [0.8719, 0.5296, 0.6563,  ..., 0.6355, 0...        [0.6348, 0.8694, 0.8777,  ..., 0.9575, 0.7697, 0.6561]]]]],
       dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([40, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: v_div_u)
_____________________ test_sddmm[idtype1-div-u-v-shp2-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'v', rhs_target = 'u', msg = 'div'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul'
lhs_data = tensor([[[[1.1791, 1.7044, 1.8254],
          [1.7367, 1.1623, 1.1621],
          [1.1044, 1.6694, 1.3022]]],


      ...8],
          [1.9425, 1.0571, 1.7099],
          [1.1015, 1.9357, 1.1846]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[0.7003, 0.5946, 0.6661]],

         [[0.5057, 0.5971, 0.6115]],

         [[0.7977, 0.6525, 0.6440]],

    ...[[0.7840, 0.7627, 0.8502]],

         [[0.8777, 0.8470, 0.7649]]]], dtype=torch.float64,
       grad_fn=<MulBackward0>)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: v_div_u)
_____________________ test_sddmm[idtype1-div-u-v-shp2-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'v', rhs_target = 'u', msg = 'div'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul'
lhs_data = tensor([[[[1.9847, 1.9895, 1.3482],
          [1.7031, 1.3355, 1.2070],
          [1.0349, 1.6468, 1.5403]]],


      ...0],
          [1.0783, 1.7128, 1.1107],
          [1.3201, 1.1327, 1.3662]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[0.6698, 0.7466, 0.6540]],

         [[0.9664, 0.7073, 0.8249]],

         [[0.5920, 0.7863, 0.8928]],

    ...[[0.5459, 0.8889, 0.6265]],

         [[0.7125, 0.6286, 0.7322]]]], dtype=torch.float64,
       grad_fn=<MulBackward0>)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([40, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: v_div_u)
_____________________ test_sddmm[idtype1-div-u-v-shp3-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'v', rhs_target = 'u', msg = 'div'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul'
lhs_data = tensor([[1.7896, 1.8404, 1.3619],
        [1.3703, 1.4828, 1.0450],
        [1.3289, 1.4439, 1.2330],
        [1.6568,... 1.2989],
        [1.1626, 1.9586, 1.6591],
        [1.4603, 1.8263, 1.7479]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.6091, 0.7491, 0.5951],
        [0.7064, 0.5264, 0.5021],
        [0.9358, 0.5760, 0.6383],
        [0.5898,...539],
        [0.5285, 0.7217, 0.6500],
        [0.7819, 0.5251, 0.8974]], dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: v_div_u)
_____________________ test_sddmm[idtype1-div-u-v-shp3-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'v', rhs_target = 'u', msg = 'div'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul'
lhs_data = tensor([[1.7961, 1.6695, 1.6413],
        [1.8898, 1.4786, 1.4683],
        [1.9651, 1.3024, 1.5569],
        [1.7996,... 1.4638],
        [1.5506, 1.2230, 1.3609],
        [1.4023, 1.7919, 1.5597]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.5062, 0.9785, 0.6619],
        [0.6365, 0.5337, 0.5154],
        [0.5422, 0.5006, 0.5692],
        [0.9793,...242],
        [0.6024, 0.6666, 0.9948],
        [0.9206, 0.5391, 0.8683]], dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([40, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: v_div_u)
_____________________ test_sddmm[idtype1-div-u-v-shp4-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'v', rhs_target = 'u', msg = 'div'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul'
lhs_data = tensor([[1.0806],
        [1.3323],
        [1.9814],
        [1.5548],
        [1.5535],
        [1.2513],
        [1...733],
        [1.7836],
        [1.7014],
        [1.3369],
        [1.1266]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.8006],
        [0.8300],
        [0.5795],
        [0.8089],
        [0.6091],
        [0.9510],
        [0...,
        [0.6527],
        [0.5749],
        [0.5928],
        [0.8624]], dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: v_div_u)
_____________________ test_sddmm[idtype1-div-u-v-shp4-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'v', rhs_target = 'u', msg = 'div'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul'
lhs_data = tensor([[1.5869],
        [1.2869],
        [1.2853],
        [1.2111],
        [1.0599],
        [1.2761],
        [1...638],
        [1.8260],
        [1.6971],
        [1.1956],
        [1.6558]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.5080],
        [0.8520],
        [0.8038],
        [0.9543],
        [0.6219],
        [0.7422],
        [0...,
        [0.5403],
        [0.6427],
        [0.6079],
        [0.6008]], dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([40, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: v_div_u)
_____________________ test_sddmm[idtype1-div-u-e-shp0-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...torch.float64), 'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'e', rhs_target = 'u'
msg = 'div', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul'
lhs_data = tensor([[[[[[1.0336],
            [1.4484],
            [1.3914]]],


          [[[1.9198],
            [1.5298],
    ...4]]],


          [[[1.8633],
            [1.1969],
            [1.5978]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[0.7462]],

           [[0.9000]],

           [[0.6626]]]],



         [[[[0.9185]],

           [[0.599...         [[[[0.6779]],

           [[0.7199]],

           [[0.7764]]]]]], dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([100, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: e_div_u)
_____________________ test_sddmm[idtype1-div-u-e-shp0-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'e', rhs_target = 'u'
msg = 'div', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul'
lhs_data = tensor([[[[[[1.9036],
            [1.1184],
            [1.7869]]],


          [[[1.3016],
            [1.2753],
    ...7]]],


          [[[1.7089],
            [1.0625],
            [1.6879]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[0.9893]],

           [[0.6345]],

           [[0.5146]]]],



         [[[[0.5039]],

           [[0.682...         [[[[0.9139]],

           [[0.5092]],

           [[0.5010]]]]]], dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([300, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: e_div_u)
_____________________ test_sddmm[idtype1-div-u-e-shp1-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...pe=torch.float64), 'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'e', rhs_target = 'u'
msg = 'div', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul'
lhs_data = tensor([[[[[1.0564, 1.8778, 1.8352,  ..., 1.0682, 1.6198, 1.5083]],

          [[1.1241, 1.5413, 1.8917,  ..., 1.7206,...
          [[1.4000, 1.6573, 1.7527,  ..., 1.3895, 1.8649, 1.6804]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[0.8797, 0.5105, 0.6319,  ..., 0.7886, 0.5280, 0.9287],
           [0.6329, 0.7801, 0.6077,  ..., 0.7847, 0...        [0.5535, 0.5214, 0.7836,  ..., 0.5942, 0.7181, 0.7019]]]]],
       dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([100, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: e_div_u)
_____________________ test_sddmm[idtype1-div-u-e-shp1-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'e', rhs_target = 'u'
msg = 'div', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul'
lhs_data = tensor([[[[[1.0432, 1.0792, 1.7461,  ..., 1.0150, 1.0637, 1.0870]],

          [[1.6880, 1.5385, 1.3657,  ..., 1.3778,...
          [[1.3937, 1.6154, 1.9238,  ..., 1.2538, 1.9396, 1.7517]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[0.5775, 0.8969, 0.7804,  ..., 0.6310, 0.7377, 0.6524],
           [0.9355, 0.7800, 0.8923,  ..., 0.5583, 0...        [0.7168, 0.6027, 0.9568,  ..., 0.5766, 0.5374, 0.5515]]]]],
       dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([300, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: e_div_u)
_____________________ test_sddmm[idtype1-div-u-e-shp2-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...dtype=torch.float64), 'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'e', rhs_target = 'u', msg = 'div'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul'
lhs_data = tensor([[[[1.0824, 1.6020, 1.9945],
          [1.6854, 1.9621, 1.0959],
          [1.7682, 1.3474, 1.6673]]],


      ...7],
          [1.0023, 1.5537, 1.0459],
          [1.1140, 1.1316, 1.8250]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[0.5497, 0.5851, 0.8793]],

         [[0.6926, 0.5842, 0.6744]],

         [[0.5338, 0.7333, 0.6026]],

    ...[[0.7001, 0.6154, 0.7896]],

         [[0.6408, 0.5829, 0.5523]]]], dtype=torch.float64,
       grad_fn=<MulBackward0>)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([100, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: e_div_u)
_____________________ test_sddmm[idtype1-div-u-e-shp2-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'e', rhs_target = 'u', msg = 'div'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul'
lhs_data = tensor([[[[1.7166, 1.4005, 1.7477],
          [1.4573, 1.0533, 1.7876],
          [1.1963, 1.6245, 1.2920]]],


      ...2],
          [1.2111, 1.7123, 1.5131],
          [1.4187, 1.0807, 1.5301]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[0.7838, 0.8823, 0.6293]],

         [[0.8066, 0.5033, 0.5635]],

         [[0.5817, 0.5481, 0.6656]],

    ...[[0.8036, 0.5432, 0.8065]],

         [[0.7118, 0.5386, 0.8360]]]], dtype=torch.float64,
       grad_fn=<MulBackward0>)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([300, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: e_div_u)
_____________________ test_sddmm[idtype1-div-u-e-shp3-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'e', rhs_target = 'u', msg = 'div'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul'
lhs_data = tensor([[1.9580, 1.0852, 1.5475],
        [1.3932, 1.0797, 1.2081],
        [1.0054, 1.8459, 1.7408],
        [1.9593,... 1.6804],
        [1.5395, 1.7258, 1.1065],
        [1.6809, 1.5847, 1.7103]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.5001, 0.9066, 0.8645],
        [0.8584, 0.9838, 0.6409],
        [0.6401, 0.6440, 0.6612],
        [0.5162,...968],
        [0.6136, 0.7492, 0.9209],
        [0.8043, 0.6843, 0.5179]], dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([100, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: e_div_u)
_____________________ test_sddmm[idtype1-div-u-e-shp3-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'e', rhs_target = 'u', msg = 'div'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul'
lhs_data = tensor([[1.8987, 1.5644, 1.7559],
        [1.4956, 1.2446, 1.6133],
        [1.6692, 1.8703, 1.4445],
        [1.0109,... 1.4918],
        [1.4810, 1.0696, 1.9411],
        [1.2551, 1.9457, 1.9244]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.5587, 0.6647, 0.5768],
        [0.9532, 0.5342, 0.8432],
        [0.5054, 0.5882, 0.7498],
        [0.6976,...906],
        [0.5383, 0.6119, 0.6429],
        [0.6145, 0.6497, 0.6247]], dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([300, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: e_div_u)
_____________________ test_sddmm[idtype1-div-u-e-shp4-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'e', rhs_target = 'u', msg = 'div'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul'
lhs_data = tensor([[1.4006],
        [1.1180],
        [1.8767],
        [1.9691],
        [1.9281],
        [1.0048],
        [1...635],
        [1.4981],
        [1.7993],
        [1.4952],
        [1.4270]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.6370],
        [0.5495],
        [0.7585],
        [0.5051],
        [0.8274],
        [0.8750],
        [0...,
        [0.6250],
        [0.5060],
        [0.9236],
        [0.5274]], dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([100, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: e_div_u)
_____________________ test_sddmm[idtype1-div-u-e-shp4-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'e', rhs_target = 'u', msg = 'div'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul'
lhs_data = tensor([[1.0343],
        [1.5394],
        [1.7950],
        [1.7214],
        [1.7305],
        [1.4360],
        [1...108],
        [1.3343],
        [1.4553],
        [1.5640],
        [1.2713]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.5387],
        [0.9368],
        [0.5916],
        [0.7420],
        [0.5982],
        [0.5360],
        [0...,
        [0.9509],
        [0.5313],
        [0.5922],
        [0.5836]], dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([300, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: e_div_u)
_____________________ test_sddmm[idtype1-div-v-u-shp0-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': ...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'u', rhs_target = 'v'
msg = 'div', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul'
lhs_data = tensor([[[[[[1.4849],
            [1.9476],
            [1.2178]]],


          [[[1.9958],
            [1.5739],
    ...3]]],


          [[[1.2187],
            [1.9133],
            [1.6610]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[0.5071]],

           [[0.5178]],

           [[0.9643]]]],



         [[[[0.5922]],

           [[0.702...         [[[[0.8349]],

           [[0.7025]],

           [[0.6696]]]]]], dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: u_div_v)
_____________________ test_sddmm[idtype1-div-v-u-shp0-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'u', rhs_target = 'v'
msg = 'div', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul'
lhs_data = tensor([[[[[[1.5722],
            [1.7491],
            [1.3023]]],


          [[[1.0971],
            [1.1750],
    ...6]]],


          [[[1.7476],
            [1.2014],
            [1.7119]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[0.9891]],

           [[0.9913]],

           [[0.6793]]]],



         [[[[0.6734]],

           [[0.563...         [[[[0.5279]],

           [[0.5035]],

           [[0.9785]]]]]], dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([40, 4, 1, 3, 1, 1])
SDDMM(message func: u_div_v)
_____________________ test_sddmm[idtype1-div-v-u-shp1-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Sch...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'u', rhs_target = 'v'
msg = 'div', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul'
lhs_data = tensor([[[[[1.0296, 1.4276, 1.4322,  ..., 1.6284, 1.5918, 1.1676]],

          [[1.5146, 1.8067, 1.6551,  ..., 1.5064,...
          [[1.3988, 1.7330, 1.9913,  ..., 1.6052, 1.0411, 1.6197]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[0.5024, 0.5875, 0.5020,  ..., 0.5552, 0.5263, 0.7511],
           [0.5061, 0.6722, 0.5899,  ..., 0.7354, 0...        [0.9766, 0.6395, 0.6618,  ..., 0.5796, 0.6840, 0.6987]]]]],
       dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: u_div_v)
_____________________ test_sddmm[idtype1-div-v-u-shp1-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'u', rhs_target = 'v'
msg = 'div', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul'
lhs_data = tensor([[[[[1.7606, 1.2121, 1.7528,  ..., 1.0741, 1.0480, 1.2104]],

          [[1.0972, 1.8569, 1.2252,  ..., 1.0506,...
          [[1.9504, 1.2783, 1.2846,  ..., 1.3514, 1.1990, 1.4715]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[0.5854, 0.5309, 0.7232,  ..., 0.7326, 0.8136, 0.9944],
           [0.5903, 0.8760, 0.9333,  ..., 0.8738, 0...        [0.5336, 0.7239, 0.5823,  ..., 0.6226, 0.6955, 0.7124]]]]],
       dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([40, 1, 3, 7, 7])
SDDMM(message func: u_div_v)
_____________________ test_sddmm[idtype1-div-v-u-shp2-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'u', rhs_target = 'v', msg = 'div'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul'
lhs_data = tensor([[[[1.1747, 1.3128, 1.1715],
          [1.0669, 1.1022, 1.7639],
          [1.5306, 1.7145, 1.7450]]],


      ...0],
          [1.4901, 1.6857, 1.4852],
          [1.9524, 1.7607, 1.0696]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[0.7583, 0.5163, 0.5145]],

         [[0.8695, 0.9617, 0.5339]],

         [[0.7004, 0.8415, 0.6316]],

    ...[[0.5868, 0.5949, 0.9847]],

         [[0.7305, 0.6118, 0.8718]]]], dtype=torch.float64,
       grad_fn=<MulBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: u_div_v)
_____________________ test_sddmm[idtype1-div-v-u-shp2-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'u', rhs_target = 'v', msg = 'div'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul'
lhs_data = tensor([[[[1.8618, 1.3926, 1.1581],
          [1.5585, 1.5470, 1.6373],
          [1.1260, 1.9382, 1.3553]]],


      ...4],
          [1.0317, 1.0414, 1.6079],
          [1.8561, 1.8542, 1.6447]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[0.8702, 0.7191, 0.5482]],

         [[0.6401, 0.9669, 0.6721]],

         [[0.6168, 0.5266, 0.7429]],

    ...[[0.7474, 0.5841, 0.7410]],

         [[0.7015, 0.6405, 0.5733]]]], dtype=torch.float64,
       grad_fn=<MulBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([40, 4, 1, 3])
SDDMM(message func: u_div_v)
_____________________ test_sddmm[idtype1-div-v-u-shp3-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'u', rhs_target = 'v', msg = 'div'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul'
lhs_data = tensor([[1.1445, 1.5504, 1.9072],
        [1.3397, 1.0500, 1.4681],
        [1.5581, 1.3720, 1.1637],
        [1.6676,... 1.2990],
        [1.8399, 1.4033, 1.3475],
        [1.1070, 1.9276, 1.3255]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.7424, 0.9794, 0.6131],
        [0.9465, 0.6225, 0.8821],
        [0.7037, 0.6592, 0.5176],
        [0.6601,...052],
        [0.7306, 0.7130, 0.5719],
        [0.5871, 0.6352, 0.5707]], dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: u_div_v)
_____________________ test_sddmm[idtype1-div-v-u-shp3-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'u', rhs_target = 'v', msg = 'div'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul'
lhs_data = tensor([[1.3546, 1.6474, 1.2912],
        [1.3105, 1.0100, 1.7811],
        [1.2823, 1.1580, 1.6820],
        [1.3746,... 1.6846],
        [1.6060, 1.0721, 1.3378],
        [1.4544, 1.1346, 1.7591]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.9172, 0.5466, 0.7954],
        [0.5316, 0.6252, 0.5874],
        [0.7533, 0.5662, 0.9814],
        [0.5111,...078],
        [0.8703, 0.6106, 0.9892],
        [0.7870, 0.7476, 0.7541]], dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([40, 3])
SDDMM(message func: u_div_v)
_____________________ test_sddmm[idtype1-div-v-u-shp4-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'u', rhs_target = 'v', msg = 'div'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul'
lhs_data = tensor([[1.8756],
        [1.3721],
        [1.1484],
        [1.4799],
        [1.0891],
        [1.2399],
        [1...686],
        [1.9269],
        [1.8710],
        [1.6208],
        [1.1022]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.5975],
        [0.7085],
        [0.6019],
        [0.7789],
        [0.7155],
        [0.6698],
        [0...,
        [0.5358],
        [0.5477],
        [0.5694],
        [0.5290]], dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: u_div_v)
_____________________ test_sddmm[idtype1-div-v-u-shp4-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'u', rhs_target = 'v', msg = 'div'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul'
lhs_data = tensor([[1.7437],
        [1.0608],
        [1.2018],
        [1.2732],
        [1.1948],
        [1.2048],
        [1...511],
        [1.4895],
        [1.0911],
        [1.3330],
        [1.0628]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.7181],
        [0.8371],
        [0.8549],
        [0.9973],
        [0.6662],
        [0.9085],
        [0...,
        [0.7999],
        [0.8134],
        [0.8213],
        [0.7588]], dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([40, 1])
SDDMM(message func: u_div_v)
_____________________ test_sddmm[idtype1-div-v-e-shp0-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...torch.float64), 'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'e', rhs_target = 'v'
msg = 'div', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul'
lhs_data = tensor([[[[[[1.0842],
            [1.9470],
            [1.1406]]],


          [[[1.4130],
            [1.8032],
    ...4]]],


          [[[1.7742],
            [1.8163],
            [1.9835]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[0.6188]],

           [[0.9997]],

           [[0.5720]]]],



         [[[[0.8190]],

           [[0.565...         [[[[0.8070]],

           [[0.5386]],

           [[0.8603]]]]]], dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([100, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: e_div_v)
_____________________ test_sddmm[idtype1-div-v-e-shp0-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'e', rhs_target = 'v'
msg = 'div', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul'
lhs_data = tensor([[[[[[1.6496],
            [1.3593],
            [1.6356]]],


          [[[1.8758],
            [1.8274],
    ...1]]],


          [[[1.4382],
            [1.6970],
            [1.7394]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[0.5325]],

           [[0.7225]],

           [[0.5020]]]],



         [[[[0.5997]],

           [[0.939...         [[[[0.5678]],

           [[0.5997]],

           [[0.8961]]]]]], dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([300, 1, 2, 1, 3, 1]), rhs shape: torch.Size([40, 4, 1, 3, 1, 1])
SDDMM(message func: e_div_v)
_____________________ test_sddmm[idtype1-div-v-e-shp1-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...pe=torch.float64), 'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'e', rhs_target = 'v'
msg = 'div', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul'
lhs_data = tensor([[[[[1.7031, 1.4639, 1.7386,  ..., 1.9274, 1.9027, 1.9381]],

          [[1.8258, 1.7656, 1.4615,  ..., 1.2879,...
          [[1.5038, 1.6215, 1.2240,  ..., 1.1297, 1.0297, 1.4822]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[0.5852, 0.7411, 0.5133,  ..., 0.9405, 0.6973, 0.6454],
           [0.5190, 0.5984, 0.5570,  ..., 0.7665, 0...        [0.7409, 0.5313, 0.7410,  ..., 0.7530, 0.5074, 0.9129]]]]],
       dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([100, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: e_div_v)
_____________________ test_sddmm[idtype1-div-v-e-shp1-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'e', rhs_target = 'v'
msg = 'div', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul'
lhs_data = tensor([[[[[1.6833, 1.0193, 1.5016,  ..., 1.7930, 1.4734, 1.5235]],

          [[1.0592, 1.6965, 1.9324,  ..., 1.6573,...
          [[1.4558, 1.9102, 1.7252,  ..., 1.9135, 1.7847, 1.8627]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[0.6092, 0.5621, 0.8365,  ..., 0.8749, 0.8817, 0.5297],
           [0.5059, 0.7434, 0.8918,  ..., 0.7337, 0...        [0.8380, 0.5863, 0.5043,  ..., 0.5902, 0.5191, 0.6839]]]]],
       dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([300, 5, 3, 1, 7]), rhs shape: torch.Size([40, 1, 3, 7, 7])
SDDMM(message func: e_div_v)
_____________________ test_sddmm[idtype1-div-v-e-shp2-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...dtype=torch.float64), 'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'e', rhs_target = 'v', msg = 'div'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul'
lhs_data = tensor([[[[1.1267, 1.6303, 1.1637],
          [1.1344, 1.2095, 1.5798],
          [1.4941, 1.2798, 1.0491]]],


      ...1],
          [1.4829, 1.7115, 1.0869],
          [1.1247, 1.1216, 1.2771]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[0.9263, 0.5910, 0.5040]],

         [[0.5893, 0.5539, 0.5258]],

         [[0.5103, 0.6093, 0.9632]],

    ...[[0.6545, 0.6107, 0.9108]],

         [[0.8786, 0.5377, 0.7057]]]], dtype=torch.float64,
       grad_fn=<MulBackward0>)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([100, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: e_div_v)
_____________________ test_sddmm[idtype1-div-v-e-shp2-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'e', rhs_target = 'v', msg = 'div'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul'
lhs_data = tensor([[[[1.4500, 1.8919, 1.2470],
          [1.9012, 1.6501, 1.7750],
          [1.8768, 1.6705, 1.7534]]],


      ...6],
          [1.0325, 1.8154, 1.1440],
          [1.3809, 1.1682, 1.5104]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[0.7320, 0.6293, 0.5393]],

         [[0.6187, 0.5619, 0.6487]],

         [[0.6389, 0.5306, 0.8432]],

    ...[[0.5642, 0.6182, 0.8368]],

         [[0.7509, 0.5392, 0.5442]]]], dtype=torch.float64,
       grad_fn=<MulBackward0>)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([300, 1, 3, 3]), rhs shape: torch.Size([40, 4, 1, 3])
SDDMM(message func: e_div_v)
_____________________ test_sddmm[idtype1-div-v-e-shp3-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'e', rhs_target = 'v', msg = 'div'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul'
lhs_data = tensor([[1.9576, 1.2553, 1.9669],
        [1.3483, 1.1628, 1.6334],
        [1.1897, 1.5169, 1.1966],
        [1.9336,... 1.0504],
        [1.4019, 1.2244, 1.9385],
        [1.5849, 1.0510, 1.8974]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.5944, 0.9707, 0.6103],
        [0.8273, 0.8661, 0.6468],
        [0.7376, 0.7088, 0.5330],
        [0.5606,...821],
        [0.6281, 0.5489, 0.9211],
        [0.5118, 0.8092, 0.8763]], dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([100, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: e_div_v)
_____________________ test_sddmm[idtype1-div-v-e-shp3-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'e', rhs_target = 'v', msg = 'div'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul'
lhs_data = tensor([[1.5098, 1.0502, 1.0430],
        [1.8293, 1.2776, 1.0073],
        [1.9160, 1.4866, 1.6121],
        [1.4524,... 1.2236],
        [1.2868, 1.0066, 1.1087],
        [1.8112, 1.1180, 1.6217]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.7146, 0.5055, 0.6538],
        [0.5538, 0.8570, 0.7870],
        [0.5866, 0.5792, 0.6056],
        [0.8468,...540],
        [0.5201, 0.8855, 0.5933],
        [0.5745, 0.5619, 0.8353]], dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([300, 3]), rhs shape: torch.Size([40, 3])
SDDMM(message func: e_div_v)
_____________________ test_sddmm[idtype1-div-v-e-shp4-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'e', rhs_target = 'v', msg = 'div'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul'
lhs_data = tensor([[1.2091],
        [1.1438],
        [1.7597],
        [1.9748],
        [1.3687],
        [1.1905],
        [1...685],
        [1.2516],
        [1.4729],
        [1.0888],
        [1.6813]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.5726],
        [0.7527],
        [0.5347],
        [0.7207],
        [0.6294],
        [0.8150],
        [0...,
        [0.5507],
        [0.9374],
        [0.5007],
        [0.5648]], dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([100, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: e_div_v)
_____________________ test_sddmm[idtype1-div-v-e-shp4-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'e', rhs_target = 'v', msg = 'div'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul'
lhs_data = tensor([[1.6198],
        [1.4795],
        [1.9537],
        [1.0415],
        [1.3658],
        [1.7181],
        [1...870],
        [1.6126],
        [1.2706],
        [1.1624],
        [1.5998]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.6909],
        [0.6657],
        [0.8059],
        [0.6008],
        [0.6530],
        [0.7227],
        [0...,
        [0.5109],
        [0.6518],
        [0.8629],
        [0.8798]], dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([300, 1]), rhs shape: torch.Size([40, 1])
SDDMM(message func: e_div_v)
_____________________ test_sddmm[idtype1-div-e-u-shp0-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': ...torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'u', rhs_target = 'e'
msg = 'div', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul'
lhs_data = tensor([[[[[[1.5787],
            [1.9254],
            [1.4822]]],


          [[[1.1826],
            [1.1512],
    ...5]]],


          [[[1.6681],
            [1.9337],
            [1.5997]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[0.5372]],

           [[0.7708]],

           [[0.6228]]]],



         [[[[0.6108]],

           [[0.968...         [[[[0.8428]],

           [[0.6000]],

           [[0.6049]]]]]], dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([100, 4, 1, 3, 1, 1])
SDDMM(message func: u_div_e)
_____________________ test_sddmm[idtype1-div-e-u-shp0-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'u', rhs_target = 'e'
msg = 'div', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul'
lhs_data = tensor([[[[[[1.0497],
            [1.3980],
            [1.1248]]],


          [[[1.8696],
            [1.5180],
    ...3]]],


          [[[1.8786],
            [1.1540],
            [1.4359]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[0.6858]],

           [[0.6631]],

           [[0.6284]]]],



         [[[[0.5485]],

           [[0.512...         [[[[0.7975]],

           [[0.8801]],

           [[0.7680]]]]]], dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([300, 4, 1, 3, 1, 1])
SDDMM(message func: u_div_e)
_____________________ test_sddmm[idtype1-div-e-u-shp1-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Sch...pe=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'u', rhs_target = 'e'
msg = 'div', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul'
lhs_data = tensor([[[[[1.4389, 1.6032, 1.6155,  ..., 1.3564, 1.2222, 1.3152]],

          [[1.3897, 1.5175, 1.4667,  ..., 1.9636,...
          [[1.9316, 1.2442, 1.5770,  ..., 1.7335, 1.7845, 1.4579]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[0.6778, 0.9497, 0.6679,  ..., 0.6343, 0.6524, 0.5106],
           [0.7241, 0.7295, 0.6675,  ..., 0.5063, 0...        [0.8261, 0.9609, 0.5788,  ..., 0.9775, 0.6056, 0.5544]]]]],
       dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([100, 1, 3, 7, 7])
SDDMM(message func: u_div_e)
_____________________ test_sddmm[idtype1-div-e-u-shp1-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'u', rhs_target = 'e'
msg = 'div', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul'
lhs_data = tensor([[[[[1.4760, 1.9773, 1.2339,  ..., 1.5700, 1.2312, 1.4203]],

          [[1.0392, 1.0729, 1.2471,  ..., 1.1620,...
          [[1.8361, 1.5921, 1.2374,  ..., 1.6536, 1.1545, 1.4498]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[0.7756, 0.6699, 0.6004,  ..., 0.8856, 0.7880, 0.5103],
           [0.8095, 0.7432, 0.7333,  ..., 0.6132, 0...        [0.9311, 0.5521, 0.7497,  ..., 0.5195, 0.5313, 0.5483]]]]],
       dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([300, 1, 3, 7, 7])
SDDMM(message func: u_div_e)
_____________________ test_sddmm[idtype1-div-e-u-shp2-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme...dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'u', rhs_target = 'e', msg = 'div'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul'
lhs_data = tensor([[[[1.8924, 1.9017, 1.3467],
          [1.8844, 1.8657, 1.4196],
          [1.8300, 1.8229, 1.9475]]],


      ...5],
          [1.2248, 1.2928, 1.1088],
          [1.4213, 1.4955, 1.0926]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[0.5265, 0.6423, 0.7433]],

         [[0.5001, 0.6110, 0.7751]],

         [[0.6125, 0.9162, 0.5747]],

    ...[[0.8856, 0.6100, 0.5867]],

         [[0.8623, 0.5062, 0.5554]]]], dtype=torch.float64,
       grad_fn=<MulBackward0>)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([100, 4, 1, 3])
SDDMM(message func: u_div_e)
_____________________ test_sddmm[idtype1-div-e-u-shp2-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'u', rhs_target = 'e', msg = 'div'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul'
lhs_data = tensor([[[[1.0559, 1.7663, 1.5525],
          [1.3992, 1.8087, 1.6537],
          [1.3712, 1.0787, 1.0675]]],


      ...7],
          [1.5915, 1.6187, 1.4747],
          [1.7886, 1.2517, 1.5389]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[0.8218, 0.5791, 0.7331]],

         [[0.8182, 0.7430, 0.5559]],

         [[0.8706, 0.5964, 0.6225]],

    ...[[0.5669, 0.7311, 0.5241]],

         [[0.5663, 0.7395, 0.9309]]]], dtype=torch.float64,
       grad_fn=<MulBackward0>)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([300, 4, 1, 3])
SDDMM(message func: u_div_e)
_____________________ test_sddmm[idtype1-div-e-u-shp3-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'u', rhs_target = 'e', msg = 'div'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul'
lhs_data = tensor([[1.1938, 1.7497, 1.8304],
        [1.0364, 1.1151, 1.6600],
        [1.9304, 1.2881, 1.3306],
        [1.7613,... 1.7814],
        [1.4301, 1.7045, 1.3185],
        [1.1038, 1.5673, 1.8023]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.6159, 0.6329, 0.8159],
        [0.5124, 0.8333, 0.5030],
        [0.8938, 0.5063, 0.5986],
        [0.9832,...629],
        [0.5790, 0.7223, 0.8672],
        [0.5482, 0.5269, 0.7914]], dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([100, 3])
SDDMM(message func: u_div_e)
_____________________ test_sddmm[idtype1-div-e-u-shp3-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'u', rhs_target = 'e', msg = 'div'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul'
lhs_data = tensor([[1.4834, 1.2620, 1.4882],
        [1.9454, 1.6106, 1.7618],
        [1.2441, 1.6945, 1.3341],
        [1.5991,... 1.5447],
        [1.4072, 1.8494, 1.5500],
        [1.5442, 1.0979, 1.5081]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.6620, 0.7021, 0.7077],
        [0.6188, 0.7651, 0.6241],
        [0.9593, 0.5740, 0.5667],
        [0.5869,...675],
        [0.8086, 0.5203, 0.5293],
        [0.7263, 0.7347, 0.6045]], dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([300, 3])
SDDMM(message func: u_div_e)
_____________________ test_sddmm[idtype1-div-e-u-shp4-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'u', rhs_target = 'e', msg = 'div'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul'
lhs_data = tensor([[1.6643],
        [1.8286],
        [1.7351],
        [1.7639],
        [1.7416],
        [1.7143],
        [1...201],
        [1.4222],
        [1.6598],
        [1.4736],
        [1.9381]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.5728],
        [0.5221],
        [0.6619],
        [0.5392],
        [0.5273],
        [0.9652],
        [0...,
        [0.5448],
        [0.6262],
        [0.5994],
        [0.5492]], dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([100, 1])
SDDMM(message func: u_div_e)
_____________________ test_sddmm[idtype1-div-e-u-shp4-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'u', rhs_target = 'e', msg = 'div'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul'
lhs_data = tensor([[1.7038],
        [1.7336],
        [1.6811],
        [1.4660],
        [1.5304],
        [1.1648],
        [1...964],
        [1.7800],
        [1.4928],
        [1.8949],
        [1.6912]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.5038],
        [0.7487],
        [0.7345],
        [0.6095],
        [0.6039],
        [0.6319],
        [0...,
        [0.6771],
        [0.5248],
        [0.5708],
        [0.7657]], dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([300, 1])
SDDMM(message func: u_div_e)
_____________________ test_sddmm[idtype1-div-e-v-shp0-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': ...torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'v', rhs_target = 'e'
msg = 'div', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul'
lhs_data = tensor([[[[[[1.4364],
            [1.6382],
            [1.9810]]],


          [[[1.9048],
            [1.0283],
    ...4]]],


          [[[1.0579],
            [1.3374],
            [1.3963]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[0.6942]],

           [[0.5389]],

           [[0.7697]]]],



         [[[[0.8847]],

           [[0.575...         [[[[0.9564]],

           [[0.8027]],

           [[0.5766]]]]]], dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([100, 4, 1, 3, 1, 1])
SDDMM(message func: v_div_e)
_____________________ test_sddmm[idtype1-div-e-v-shp0-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'v', rhs_target = 'e'
msg = 'div', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul'
lhs_data = tensor([[[[[[1.0387],
            [1.5022],
            [1.1555]]],


          [[[1.6103],
            [1.8578],
    ...8]]],


          [[[1.0593],
            [1.2320],
            [1.1111]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[0.5928]],

           [[0.7341]],

           [[0.5511]]]],



         [[[[0.6236]],

           [[0.645...         [[[[0.7769]],

           [[0.5712]],

           [[0.5798]]]]]], dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([40, 1, 2, 1, 3, 1]), rhs shape: torch.Size([300, 4, 1, 3, 1, 1])
SDDMM(message func: v_div_e)
_____________________ test_sddmm[idtype1-div-e-v-shp1-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Sch...pe=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'v', rhs_target = 'e'
msg = 'div', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul'
lhs_data = tensor([[[[[1.2380, 1.7252, 1.2806,  ..., 1.3164, 1.1429, 1.7729]],

          [[1.1620, 1.5094, 1.5257,  ..., 1.4025,...
          [[1.8423, 1.1178, 1.9932,  ..., 1.5302, 1.3837, 1.5648]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[0.9336, 0.9392, 0.9603,  ..., 0.5776, 0.7538, 0.8107],
           [0.5363, 0.8434, 0.5953,  ..., 0.7883, 0...        [0.8423, 0.5362, 0.6850,  ..., 0.6196, 0.7930, 0.5256]]]]],
       dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([100, 1, 3, 7, 7])
SDDMM(message func: v_div_e)
_____________________ test_sddmm[idtype1-div-e-v-shp1-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'v', rhs_target = 'e'
msg = 'div', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul'
lhs_data = tensor([[[[[1.5372, 1.1394, 1.5685,  ..., 1.4138, 1.4269, 1.4358]],

          [[1.3510, 1.1368, 1.6741,  ..., 1.4809,...
          [[1.8473, 1.2733, 1.4840,  ..., 1.5750, 1.2731, 1.6663]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[0.5085, 0.6526, 0.9399,  ..., 0.9721, 0.9807, 0.7947],
           [0.5468, 0.8228, 0.5806,  ..., 0.5097, 0...        [0.5225, 0.6731, 0.8384,  ..., 0.5976, 0.8337, 0.7091]]]]],
       dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([40, 5, 3, 1, 7]), rhs shape: torch.Size([300, 1, 3, 7, 7])
SDDMM(message func: v_div_e)
_____________________ test_sddmm[idtype1-div-e-v-shp2-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme...dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'v', rhs_target = 'e', msg = 'div'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul'
lhs_data = tensor([[[[1.6324, 1.6124, 1.7800],
          [1.0068, 1.1822, 1.3675],
          [1.1928, 1.7627, 1.5024]]],


      ...6],
          [1.0510, 1.5573, 1.5522],
          [1.4307, 1.3714, 1.7270]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[0.7490, 0.9849, 0.5162]],

         [[0.8117, 0.9630, 0.6639]],

         [[0.7226, 0.7211, 0.5299]],

    ...[[0.6170, 0.6270, 0.9441]],

         [[0.5857, 0.9567, 0.6860]]]], dtype=torch.float64,
       grad_fn=<MulBackward0>)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([100, 4, 1, 3])
SDDMM(message func: v_div_e)
_____________________ test_sddmm[idtype1-div-e-v-shp2-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'v', rhs_target = 'e', msg = 'div'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul'
lhs_data = tensor([[[[1.9865, 1.8408, 1.8303],
          [1.9357, 1.2549, 1.4069],
          [1.8510, 1.9848, 1.3306]]],


      ...0],
          [1.2396, 1.8497, 1.4609],
          [1.9788, 1.1936, 1.3777]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[0.8203, 0.9037, 0.7620]],

         [[0.9624, 0.8962, 0.8656]],

         [[0.6471, 0.6209, 0.5313]],

    ...[[0.8333, 0.5884, 0.7613]],

         [[0.9339, 0.8862, 0.5511]]]], dtype=torch.float64,
       grad_fn=<MulBackward0>)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([40, 1, 3, 3]), rhs shape: torch.Size([300, 4, 1, 3])
SDDMM(message func: v_div_e)
_____________________ test_sddmm[idtype1-div-e-v-shp3-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'v', rhs_target = 'e', msg = 'div'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul'
lhs_data = tensor([[1.2187, 1.0082, 1.1930],
        [1.9563, 1.5278, 1.4786],
        [1.9947, 1.9068, 1.7273],
        [1.0381,... 1.6719],
        [1.9231, 1.1595, 1.0325],
        [1.0777, 1.8706, 1.7916]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.6129, 0.7580, 0.7973],
        [0.5116, 0.5225, 0.7652],
        [0.5080, 0.5693, 0.5776],
        [0.5886,...280],
        [0.6755, 0.5693, 0.6948],
        [0.9887, 0.9151, 0.5274]], dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([100, 3])
SDDMM(message func: v_div_e)
_____________________ test_sddmm[idtype1-div-e-v-shp3-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'v', rhs_target = 'e', msg = 'div'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul'
lhs_data = tensor([[1.4120, 1.0809, 1.5701],
        [1.4447, 1.8924, 1.3372],
        [1.0333, 1.8482, 1.3602],
        [1.7132,... 1.6523],
        [1.5197, 1.8825, 1.9653],
        [1.1747, 1.0930, 1.9432]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.9633, 0.9563, 0.6105],
        [0.7753, 0.8769, 0.5391],
        [0.7560, 0.7548, 0.5051],
        [0.6896,...319],
        [0.9452, 0.6187, 0.5087],
        [0.9431, 0.6061, 0.5799]], dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([40, 3]), rhs shape: torch.Size([300, 3])
SDDMM(message func: v_div_e)
_____________________ test_sddmm[idtype1-div-e-v-shp4-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'v', rhs_target = 'e', msg = 'div'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'mul'
lhs_data = tensor([[1.2674],
        [1.4155],
        [1.0721],
        [1.1219],
        [1.9644],
        [1.7487],
        [1...462],
        [1.6972],
        [1.7045],
        [1.8428],
        [1.7066]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.8901],
        [0.8067],
        [0.8978],
        [0.7434],
        [0.8903],
        [0.5139],
        [0...,
        [0.7003],
        [0.8689],
        [0.7335],
        [0.7119]], dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([100, 1])
SDDMM(message func: v_div_e)
_____________________ test_sddmm[idtype1-div-e-v-shp4-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'v', rhs_target = 'e', msg = 'div'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'mul'
lhs_data = tensor([[1.0628],
        [1.4392],
        [1.2294],
        [1.8668],
        [1.1214],
        [1.3314],
        [1...048],
        [1.9779],
        [1.7646],
        [1.2075],
        [1.5006]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[0.6647],
        [0.6761],
        [0.8345],
        [0.5349],
        [0.5143],
        [0.9442],
        [0...,
        [0.8277],
        [0.7046],
        [0.7166],
        [0.5978]], dtype=torch.float64, grad_fn=<MulBackward0>)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([40, 1]), rhs shape: torch.Size([300, 1])
SDDMM(message func: v_div_e)
_____________________ test_sddmm[idtype1-dot-u-v-shp0-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': ...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'v', rhs_target = 'u'
msg = 'dot', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'dot'
lhs_data = tensor([[[[[[1.8189],
            [1.3453],
            [1.8884]]],


          [[[1.7532],
            [1.0228],
    ...4]]],


          [[[1.4853],
            [1.6186],
            [1.9518]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.9099]],

           [[1.0278]],

           [[1.0772]]]],



         [[[[1.2251]],

           [[1.573...



         [[[[1.7330]],

           [[1.8190]],

           [[1.0781]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: v_dot_u)
_____________________ test_sddmm[idtype1-dot-u-v-shp0-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'v', rhs_target = 'u'
msg = 'dot', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'dot'
lhs_data = tensor([[[[[[1.7760],
            [1.2393],
            [1.5138]]],


          [[[1.0175],
            [1.3400],
    ...8]]],


          [[[1.4799],
            [1.6625],
            [1.1197]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.0297]],

           [[1.5895]],

           [[1.9178]]]],



         [[[[1.1162]],

           [[1.769...



         [[[[1.8005]],

           [[1.8173]],

           [[1.8691]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([40, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: v_dot_u)
_____________________ test_sddmm[idtype1-dot-u-v-shp1-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Sch...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'v', rhs_target = 'u'
msg = 'dot', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'dot'
lhs_data = tensor([[[[[1.9286, 1.0802, 1.0907,  ..., 1.5392, 1.7572, 1.0918]],

          [[1.7500, 1.4109, 1.9551,  ..., 1.4802,...
          [[1.3761, 1.6389, 1.9952,  ..., 1.0159, 1.1743, 1.4686]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.7081, 1.4755, 1.8697,  ..., 1.8235, 1.9327, 1.6759],
           [1.2111, 1.8813, 1.5229,  ..., 1.0063, 1...
           [1.6095, 1.5583, 1.1600,  ..., 1.7183, 1.2343, 1.7711]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: v_dot_u)
_____________________ test_sddmm[idtype1-dot-u-v-shp1-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'v', rhs_target = 'u'
msg = 'dot', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'dot'
lhs_data = tensor([[[[[1.1897, 1.5240, 1.1583,  ..., 1.6977, 1.9428, 1.2124]],

          [[1.2515, 1.0954, 1.5262,  ..., 1.7140,...
          [[1.0679, 1.7825, 1.4596,  ..., 1.3151, 1.4537, 1.4969]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.5634, 1.3394, 1.4448,  ..., 1.5168, 1.2215, 1.3454],
           [1.3187, 1.1328, 1.3904,  ..., 1.1254, 1...
           [1.0594, 1.6545, 1.9769,  ..., 1.8870, 1.1332, 1.6121]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([40, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: v_dot_u)
_____________________ test_sddmm[idtype1-dot-u-v-shp2-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'v', rhs_target = 'u', msg = 'dot'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'dot'
lhs_data = tensor([[[[1.7154, 1.9857, 1.4841],
          [1.1353, 1.9368, 1.9457],
          [1.4903, 1.4256, 1.1381]]],


      ...1],
          [1.5464, 1.7996, 1.6934],
          [1.1862, 1.7508, 1.7190]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.4748, 1.9218, 1.8959]],

         [[1.3226, 1.5075, 1.2893]],

         [[1.3137, 1.6486, 1.5460]],

    ...

         [[1.7130, 1.1586, 1.8846]],

         [[1.5466, 1.4934, 1.7203]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: v_dot_u)
_____________________ test_sddmm[idtype1-dot-u-v-shp2-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'v', rhs_target = 'u', msg = 'dot'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'dot'
lhs_data = tensor([[[[1.2984, 1.1937, 1.1768],
          [1.1830, 1.9557, 1.3281],
          [1.0873, 1.8844, 1.6834]]],


      ...4],
          [1.2345, 1.1369, 1.5728],
          [1.2106, 1.1660, 1.4732]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.1782, 1.5760, 1.4976]],

         [[1.7414, 1.7780, 1.3291]],

         [[1.7296, 1.9534, 1.2018]],

    ...

         [[1.5180, 1.2065, 1.5405]],

         [[1.2616, 1.1475, 1.6568]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([40, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: v_dot_u)
_____________________ test_sddmm[idtype1-dot-u-v-shp3-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'v', rhs_target = 'u', msg = 'dot'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'dot'
lhs_data = tensor([[1.5221, 1.7418, 1.6335],
        [1.8512, 1.8646, 1.1893],
        [1.4918, 1.0388, 1.9677],
        [1.0985,... 1.8488],
        [1.9915, 1.6781, 1.3986],
        [1.4906, 1.3865, 1.4576]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.6965, 1.2033, 1.6021],
        [1.9703, 1.8728, 1.7426],
        [1.6681, 1.4684, 1.6405],
        [1.7241,... 1.3571],
        [1.4551, 1.4522, 1.5083],
        [1.6603, 1.9156, 1.6470]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: v_dot_u)
_____________________ test_sddmm[idtype1-dot-u-v-shp3-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'v', rhs_target = 'u', msg = 'dot'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'dot'
lhs_data = tensor([[1.5969, 1.1467, 1.6152],
        [1.0161, 1.4809, 1.7099],
        [1.7721, 1.7740, 1.1722],
        [1.2299,... 1.5426],
        [1.3700, 1.0197, 1.4586],
        [1.0145, 1.9348, 1.2015]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.0919, 1.8144, 1.0953],
        [1.0632, 1.3552, 1.4137],
        [1.2022, 1.0138, 1.0983],
        [1.0527,... 1.9314],
        [1.2248, 1.5670, 1.2072],
        [1.5283, 1.5033, 1.4354]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([40, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: v_dot_u)
_____________________ test_sddmm[idtype1-dot-u-v-shp4-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'v', rhs_target = 'u', msg = 'dot'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'dot'
lhs_data = tensor([[1.4065],
        [1.6453],
        [1.5940],
        [1.8834],
        [1.5205],
        [1.8767],
        [1...887],
        [1.7349],
        [1.4353],
        [1.0336],
        [1.9258]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.1926],
        [1.7048],
        [1.0471],
        [1.1825],
        [1.9938],
        [1.8587],
        [1...590],
        [1.1581],
        [1.8913],
        [1.4191],
        [1.1389]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: v_dot_u)
_____________________ test_sddmm[idtype1-dot-u-v-shp4-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'v', rhs_target = 'u', msg = 'dot'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'dot'
lhs_data = tensor([[1.8913],
        [1.9761],
        [1.2691],
        [1.7124],
        [1.9891],
        [1.5234],
        [1...451],
        [1.4823],
        [1.4475],
        [1.2084],
        [1.0155]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.1799],
        [1.9573],
        [1.0769],
        [1.6050],
        [1.0366],
        [1.6334],
        [1...927],
        [1.7050],
        [1.5606],
        [1.3233],
        [1.1366]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([40, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: v_dot_u)
_____________________ test_sddmm[idtype1-dot-u-e-shp0-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...torch.float64), 'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'e', rhs_target = 'u'
msg = 'dot', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'dot'
lhs_data = tensor([[[[[[1.8929],
            [1.8172],
            [1.1081]]],


          [[[1.7914],
            [1.7172],
    ...5]]],


          [[[1.8290],
            [1.9116],
            [1.1230]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.9307]],

           [[1.4847]],

           [[1.1425]]]],



         [[[[1.7151]],

           [[1.944...



         [[[[1.7452]],

           [[1.1922]],

           [[1.4873]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([100, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: e_dot_u)
_____________________ test_sddmm[idtype1-dot-u-e-shp0-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'e', rhs_target = 'u'
msg = 'dot', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'dot'
lhs_data = tensor([[[[[[1.0015],
            [1.9556],
            [1.8495]]],


          [[[1.7520],
            [1.8096],
    ...7]]],


          [[[1.6085],
            [1.5341],
            [1.3644]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.8238]],

           [[1.3152]],

           [[1.9954]]]],



         [[[[1.9656]],

           [[1.693...



         [[[[1.2299]],

           [[1.7161]],

           [[1.4107]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([300, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: e_dot_u)
_____________________ test_sddmm[idtype1-dot-u-e-shp1-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...pe=torch.float64), 'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'e', rhs_target = 'u'
msg = 'dot', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'dot'
lhs_data = tensor([[[[[1.6087, 1.5758, 1.3744,  ..., 1.3427, 1.0609, 1.8380]],

          [[1.7976, 1.8480, 1.7812,  ..., 1.4050,...
          [[1.6688, 1.3049, 1.4868,  ..., 1.7376, 1.0302, 1.2928]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.9245, 1.3635, 1.0887,  ..., 1.8764, 1.8067, 1.0915],
           [1.4775, 1.8152, 1.5350,  ..., 1.3062, 1...
           [1.1121, 1.0329, 1.5102,  ..., 1.8420, 1.9678, 1.0070]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([100, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: e_dot_u)
_____________________ test_sddmm[idtype1-dot-u-e-shp1-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'e', rhs_target = 'u'
msg = 'dot', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'dot'
lhs_data = tensor([[[[[1.2297, 1.5828, 1.4560,  ..., 1.9342, 1.5930, 1.9337]],

          [[1.2020, 1.1646, 1.0314,  ..., 1.3889,...
          [[1.7323, 1.2561, 1.0115,  ..., 1.0068, 1.8932, 1.9372]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.1379, 1.5543, 1.0774,  ..., 1.0734, 1.5426, 1.0490],
           [1.6841, 1.2757, 1.2800,  ..., 1.3195, 1...
           [1.3295, 1.9932, 1.9135,  ..., 1.7510, 1.9703, 1.1179]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([300, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: e_dot_u)
_____________________ test_sddmm[idtype1-dot-u-e-shp2-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...dtype=torch.float64), 'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'e', rhs_target = 'u', msg = 'dot'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'dot'
lhs_data = tensor([[[[1.2564, 1.0165, 1.7853],
          [1.6427, 1.9880, 1.9496],
          [1.8283, 1.3121, 1.0973]]],


      ...9],
          [1.7624, 1.2640, 1.3450],
          [1.1071, 1.7139, 1.1210]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.0367, 1.7632, 1.7744]],

         [[1.6872, 1.5771, 1.4852]],

         [[1.1877, 1.9525, 1.1519]],

    ...

         [[1.3692, 1.5533, 1.1285]],

         [[1.2271, 1.2466, 1.8894]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([100, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: e_dot_u)
_____________________ test_sddmm[idtype1-dot-u-e-shp2-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'e', rhs_target = 'u', msg = 'dot'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'dot'
lhs_data = tensor([[[[1.8641, 1.0490, 1.0675],
          [1.7391, 1.0514, 1.2092],
          [1.1094, 1.3734, 1.8647]]],


      ...7],
          [1.4900, 1.9876, 1.6776],
          [1.3845, 1.1853, 1.7816]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.9559, 1.0588, 1.5548]],

         [[1.1503, 1.6057, 1.0936]],

         [[1.7977, 1.5481, 1.7440]],

    ...

         [[1.7704, 1.4147, 1.3119]],

         [[1.4699, 1.7053, 1.3757]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([300, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: e_dot_u)
_____________________ test_sddmm[idtype1-dot-u-e-shp3-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'e', rhs_target = 'u', msg = 'dot'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'dot'
lhs_data = tensor([[1.6878, 1.8380, 1.3412],
        [1.4477, 1.1997, 1.7449],
        [1.1725, 1.1558, 1.7330],
        [1.4799,... 1.7423],
        [1.3610, 1.5671, 1.9384],
        [1.4528, 1.0491, 1.1346]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.5589, 1.8930, 1.6787],
        [1.5708, 1.7006, 1.5178],
        [1.6503, 1.1264, 1.9338],
        [1.1213,... 1.9438],
        [1.1982, 1.1001, 1.4376],
        [1.8619, 1.9389, 1.7284]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([100, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: e_dot_u)
_____________________ test_sddmm[idtype1-dot-u-e-shp3-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'e', rhs_target = 'u', msg = 'dot'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'dot'
lhs_data = tensor([[1.1501, 1.2027, 1.8192],
        [1.0310, 1.5309, 1.4071],
        [1.9518, 1.2560, 1.4114],
        [1.5758,... 1.5170],
        [1.6794, 1.9234, 1.3274],
        [1.0308, 1.9787, 1.4570]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.8581, 1.5765, 1.5019],
        [1.5453, 1.5669, 1.6639],
        [1.4623, 1.7714, 1.0711],
        [1.0080,... 1.5355],
        [1.2865, 1.5617, 1.5364],
        [1.4438, 1.2778, 1.5828]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([300, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: e_dot_u)
_____________________ test_sddmm[idtype1-dot-u-e-shp4-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'e', rhs_target = 'u', msg = 'dot'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'dot'
lhs_data = tensor([[1.7934],
        [1.7021],
        [1.7350],
        [1.1015],
        [1.3428],
        [1.3501],
        [1...227],
        [1.3342],
        [1.1890],
        [1.7169],
        [1.3256]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.1553],
        [1.5367],
        [1.3074],
        [1.4910],
        [1.6408],
        [1.2142],
        [1...186],
        [1.8776],
        [1.2527],
        [1.0650],
        [1.3499]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([100, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: e_dot_u)
_____________________ test_sddmm[idtype1-dot-u-e-shp4-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'e', rhs_target = 'u', msg = 'dot'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'dot'
lhs_data = tensor([[1.1368],
        [1.2121],
        [1.8893],
        [1.2679],
        [1.0351],
        [1.4249],
        [1...781],
        [1.9054],
        [1.1142],
        [1.9216],
        [1.1599]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.9453],
        [1.0233],
        [1.9063],
        [1.6982],
        [1.3402],
        [1.7719],
        [1...106],
        [1.4667],
        [1.9137],
        [1.0967],
        [1.4981]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([300, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: e_dot_u)
_____________________ test_sddmm[idtype1-dot-v-u-shp0-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': ...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'u', rhs_target = 'v'
msg = 'dot', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'dot'
lhs_data = tensor([[[[[[1.0467],
            [1.4673],
            [1.8423]]],


          [[[1.7383],
            [1.4226],
    ...9]]],


          [[[1.7902],
            [1.1769],
            [1.9763]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.3873]],

           [[1.2994]],

           [[1.3116]]]],



         [[[[1.7125]],

           [[1.245...



         [[[[1.3221]],

           [[1.4569]],

           [[1.0036]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: u_dot_v)
_____________________ test_sddmm[idtype1-dot-v-u-shp0-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'u', rhs_target = 'v'
msg = 'dot', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'dot'
lhs_data = tensor([[[[[[1.3675],
            [1.7645],
            [1.6237]]],


          [[[1.0176],
            [1.6864],
    ...1]]],


          [[[1.0047],
            [1.1932],
            [1.5328]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.5098]],

           [[1.5806]],

           [[1.3653]]]],



         [[[[1.4074]],

           [[1.726...



         [[[[1.8302]],

           [[1.5568]],

           [[1.6229]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([40, 4, 1, 3, 1, 1])
SDDMM(message func: u_dot_v)
_____________________ test_sddmm[idtype1-dot-v-u-shp1-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Sch...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'u', rhs_target = 'v'
msg = 'dot', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'dot'
lhs_data = tensor([[[[[1.8915, 1.5989, 1.1340,  ..., 1.2125, 1.8818, 1.8904]],

          [[1.3225, 1.3147, 1.1275,  ..., 1.5766,...
          [[1.2852, 1.6357, 1.5347,  ..., 1.2923, 1.0561, 1.4420]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.4277, 1.8787, 1.1416,  ..., 1.8206, 1.9680, 1.0438],
           [1.5271, 1.1148, 1.6219,  ..., 1.0966, 1...
           [1.0304, 1.5528, 1.6306,  ..., 1.5635, 1.8186, 1.4631]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: u_dot_v)
_____________________ test_sddmm[idtype1-dot-v-u-shp1-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'u', rhs_target = 'v'
msg = 'dot', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'dot'
lhs_data = tensor([[[[[1.9506, 1.5653, 1.0005,  ..., 1.3471, 1.1497, 1.2962]],

          [[1.4865, 1.8055, 1.7359,  ..., 1.5358,...
          [[1.9473, 1.9040, 1.1555,  ..., 1.6542, 1.7798, 1.3116]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.4132, 1.8718, 1.3931,  ..., 1.4979, 1.2243, 1.7997],
           [1.8917, 1.1275, 1.7960,  ..., 1.4312, 1...
           [1.5247, 1.3845, 1.5499,  ..., 1.5387, 1.7007, 1.1583]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([40, 1, 3, 7, 7])
SDDMM(message func: u_dot_v)
_____________________ test_sddmm[idtype1-dot-v-u-shp2-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'u', rhs_target = 'v', msg = 'dot'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'dot'
lhs_data = tensor([[[[1.0889, 1.9650, 1.9019],
          [1.3482, 1.7156, 1.9754],
          [1.6069, 1.2622, 1.0293]]],


      ...5],
          [1.0239, 1.9309, 1.2482],
          [1.9643, 1.9551, 1.2763]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.6221, 1.8249, 1.0321]],

         [[1.2691, 1.6156, 1.6902]],

         [[1.4451, 1.1199, 1.2279]],

    ...

         [[1.8761, 1.2515, 1.2776]],

         [[1.2714, 1.8511, 1.7830]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: u_dot_v)
_____________________ test_sddmm[idtype1-dot-v-u-shp2-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'u', rhs_target = 'v', msg = 'dot'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'dot'
lhs_data = tensor([[[[1.9406, 1.9268, 1.7475],
          [1.1313, 1.9823, 1.2510],
          [1.2104, 1.6287, 1.0968]]],


      ...8],
          [1.5879, 1.6819, 1.1967],
          [1.6055, 1.4173, 1.6767]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.8736, 1.1848, 1.9562]],

         [[1.9488, 1.1951, 1.0181]],

         [[1.6250, 1.4908, 1.5382]],

    ...

         [[1.5130, 1.5392, 1.5847]],

         [[1.7939, 1.6738, 1.0989]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([40, 4, 1, 3])
SDDMM(message func: u_dot_v)
_____________________ test_sddmm[idtype1-dot-v-u-shp3-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'u', rhs_target = 'v', msg = 'dot'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'dot'
lhs_data = tensor([[1.3113, 1.3529, 1.2787],
        [1.6873, 1.3606, 1.4951],
        [1.3553, 1.3898, 1.9577],
        [1.3161,... 1.8151],
        [1.5692, 1.1288, 1.3587],
        [1.2505, 1.7218, 1.1145]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.1791, 1.8427, 1.3442],
        [1.7211, 1.7867, 1.0867],
        [1.9562, 1.2629, 1.3715],
        [1.1341,... 1.1920],
        [1.2003, 1.7139, 1.4543],
        [1.1261, 1.9884, 1.2712]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: u_dot_v)
_____________________ test_sddmm[idtype1-dot-v-u-shp3-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'u', rhs_target = 'v', msg = 'dot'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'dot'
lhs_data = tensor([[1.3878, 1.6107, 1.8187],
        [1.8357, 1.8281, 1.8685],
        [1.2801, 1.7859, 1.6465],
        [1.9826,... 1.0161],
        [1.7712, 1.4592, 1.9401],
        [1.1659, 1.1624, 1.0600]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.0156, 1.7377, 1.8359],
        [1.4311, 1.5457, 1.5336],
        [1.0907, 1.4142, 1.3061],
        [1.7312,... 1.3949],
        [1.8827, 1.0167, 1.8519],
        [1.2620, 1.5207, 1.7247]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([40, 3])
SDDMM(message func: u_dot_v)
_____________________ test_sddmm[idtype1-dot-v-u-shp4-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'u', rhs_target = 'v', msg = 'dot'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'dot'
lhs_data = tensor([[1.2055],
        [1.9136],
        [1.4972],
        [1.7878],
        [1.5796],
        [1.9549],
        [1...501],
        [1.5069],
        [1.6796],
        [1.1126],
        [1.0139]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.6940],
        [1.7052],
        [1.8491],
        [1.3453],
        [1.4438],
        [1.4881],
        [1...459],
        [1.5810],
        [1.3541],
        [1.1650],
        [1.5332]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: u_dot_v)
_____________________ test_sddmm[idtype1-dot-v-u-shp4-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'u', rhs_target = 'v', msg = 'dot'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'dot'
lhs_data = tensor([[1.8671],
        [1.0599],
        [1.4775],
        [1.7413],
        [1.3570],
        [1.9611],
        [1...460],
        [1.0361],
        [1.8306],
        [1.8849],
        [1.6426]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.5792],
        [1.8845],
        [1.6502],
        [1.4301],
        [1.3093],
        [1.8026],
        [1...142],
        [1.8429],
        [1.8652],
        [1.9344],
        [1.0605]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([40, 1])
SDDMM(message func: u_dot_v)
_____________________ test_sddmm[idtype1-dot-v-e-shp0-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...torch.float64), 'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'e', rhs_target = 'v'
msg = 'dot', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'dot'
lhs_data = tensor([[[[[[1.3624],
            [1.9463],
            [1.8961]]],


          [[[1.3172],
            [1.1322],
    ...4]]],


          [[[1.6084],
            [1.2910],
            [1.8822]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.6585]],

           [[1.5509]],

           [[1.4005]]]],



         [[[[1.9506]],

           [[1.699...



         [[[[1.9920]],

           [[1.7987]],

           [[1.7522]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([100, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: e_dot_v)
_____________________ test_sddmm[idtype1-dot-v-e-shp0-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'e', rhs_target = 'v'
msg = 'dot', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'dot'
lhs_data = tensor([[[[[[1.5385],
            [1.7244],
            [1.7215]]],


          [[[1.7658],
            [1.1687],
    ...2]]],


          [[[1.7080],
            [1.6922],
            [1.4262]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.2511]],

           [[1.5675]],

           [[1.7078]]]],



         [[[[1.4187]],

           [[1.503...



         [[[[1.0737]],

           [[1.7920]],

           [[1.7068]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([300, 1, 2, 1, 3, 1]), rhs shape: torch.Size([40, 4, 1, 3, 1, 1])
SDDMM(message func: e_dot_v)
_____________________ test_sddmm[idtype1-dot-v-e-shp1-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...pe=torch.float64), 'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'e', rhs_target = 'v'
msg = 'dot', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'dot'
lhs_data = tensor([[[[[1.7451, 1.1100, 1.6950,  ..., 1.9527, 1.4761, 1.6414]],

          [[1.3741, 1.8333, 1.2986,  ..., 1.7411,...
          [[1.4123, 1.4143, 1.7928,  ..., 1.4411, 1.6050, 1.3844]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.5321, 1.1202, 1.0931,  ..., 1.8870, 1.9547, 1.9141],
           [1.4783, 1.9403, 1.1381,  ..., 1.9157, 1...
           [1.9914, 1.9464, 1.2886,  ..., 1.6032, 1.1797, 1.2098]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([100, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: e_dot_v)
_____________________ test_sddmm[idtype1-dot-v-e-shp1-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'e', rhs_target = 'v'
msg = 'dot', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'dot'
lhs_data = tensor([[[[[1.1774, 1.5003, 1.0867,  ..., 1.5996, 1.3963, 1.4214]],

          [[1.7921, 1.6893, 1.6427,  ..., 1.5893,...
          [[1.6904, 1.9930, 1.0986,  ..., 1.1158, 1.6970, 1.8260]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.5827, 1.2240, 1.1486,  ..., 1.4725, 1.2519, 1.4280],
           [1.8428, 1.3703, 1.0524,  ..., 1.3406, 1...
           [1.4292, 1.6483, 1.6035,  ..., 1.9275, 1.0854, 1.2038]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([300, 5, 3, 1, 7]), rhs shape: torch.Size([40, 1, 3, 7, 7])
SDDMM(message func: e_dot_v)
_____________________ test_sddmm[idtype1-dot-v-e-shp2-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...dtype=torch.float64), 'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'e', rhs_target = 'v', msg = 'dot'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'dot'
lhs_data = tensor([[[[1.3708, 1.2076, 1.4345],
          [1.3086, 1.0393, 1.0773],
          [1.3688, 1.2844, 1.8632]]],


      ...1],
          [1.6034, 1.1851, 1.2510],
          [1.0352, 1.0183, 1.2309]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.5150, 1.9627, 1.9241]],

         [[1.9398, 1.0143, 1.9653]],

         [[1.8158, 1.0311, 1.8225]],

    ...

         [[1.4878, 1.5913, 1.6467]],

         [[1.1007, 1.8505, 1.3870]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([100, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: e_dot_v)
_____________________ test_sddmm[idtype1-dot-v-e-shp2-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'e', rhs_target = 'v', msg = 'dot'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'dot'
lhs_data = tensor([[[[1.5956, 1.5682, 1.8542],
          [1.0390, 1.0541, 1.5640],
          [1.1749, 1.1664, 1.7720]]],


      ...9],
          [1.9764, 1.1421, 1.1666],
          [1.4514, 1.3994, 1.0554]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.8922, 1.5315, 1.3230]],

         [[1.7347, 1.3464, 1.5292]],

         [[1.5697, 1.0394, 1.2550]],

    ...

         [[1.6620, 1.0940, 1.0252]],

         [[1.0695, 1.9404, 1.1487]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([300, 1, 3, 3]), rhs shape: torch.Size([40, 4, 1, 3])
SDDMM(message func: e_dot_v)
_____________________ test_sddmm[idtype1-dot-v-e-shp3-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'e', rhs_target = 'v', msg = 'dot'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'dot'
lhs_data = tensor([[1.5077, 1.2882, 1.4224],
        [1.7700, 1.0649, 1.1696],
        [1.4030, 1.8990, 1.9476],
        [1.6898,... 1.2822],
        [1.1009, 1.5197, 1.4961],
        [1.8869, 1.3446, 1.3204]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.9543, 1.2781, 1.9834],
        [1.9718, 1.6689, 1.1482],
        [1.6806, 1.0704, 1.6283],
        [1.2114,... 1.7363],
        [1.0768, 1.1985, 1.6570],
        [1.5389, 1.5287, 1.8190]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([100, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: e_dot_v)
_____________________ test_sddmm[idtype1-dot-v-e-shp3-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'e', rhs_target = 'v', msg = 'dot'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'dot'
lhs_data = tensor([[1.8884, 1.1693, 1.5430],
        [1.9204, 1.6840, 1.1354],
        [1.6232, 1.6689, 1.0184],
        [1.5278,... 1.3114],
        [1.3214, 1.1851, 1.3446],
        [1.9926, 1.3957, 1.4255]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.9318, 1.7552, 1.1033],
        [1.2841, 1.3438, 1.9675],
        [1.8555, 1.1081, 1.6052],
        [1.2148,... 1.4506],
        [1.6662, 1.5969, 1.5379],
        [1.5267, 1.7346, 1.6486]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([300, 3]), rhs shape: torch.Size([40, 3])
SDDMM(message func: e_dot_v)
_____________________ test_sddmm[idtype1-dot-v-e-shp4-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'e', rhs_target = 'v', msg = 'dot'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'dot'
lhs_data = tensor([[1.1012],
        [1.1866],
        [1.8903],
        [1.8884],
        [1.4342],
        [1.9021],
        [1...694],
        [1.4694],
        [1.5698],
        [1.1110],
        [1.3046]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.3297],
        [1.3408],
        [1.2274],
        [1.5583],
        [1.8539],
        [1.6226],
        [1...680],
        [1.0438],
        [1.8081],
        [1.1909],
        [1.4280]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([100, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: e_dot_v)
_____________________ test_sddmm[idtype1-dot-v-e-shp4-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'e', rhs_target = 'v', msg = 'dot'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'dot'
lhs_data = tensor([[1.8790],
        [1.4774],
        [1.6299],
        [1.6946],
        [1.4382],
        [1.0692],
        [1...163],
        [1.9907],
        [1.1154],
        [1.2377],
        [1.4193]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.1208],
        [1.3626],
        [1.3239],
        [1.3982],
        [1.8120],
        [1.8535],
        [1...597],
        [1.0594],
        [1.8772],
        [1.2056],
        [1.7497]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([300, 1]), rhs shape: torch.Size([40, 1])
SDDMM(message func: e_dot_v)
_____________________ test_sddmm[idtype1-dot-e-u-shp0-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': ...torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'u', rhs_target = 'e'
msg = 'dot', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'dot'
lhs_data = tensor([[[[[[1.1945],
            [1.9400],
            [1.9952]]],


          [[[1.9620],
            [1.9267],
    ...4]]],


          [[[1.2246],
            [1.2643],
            [1.7787]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.2624]],

           [[1.8297]],

           [[1.5291]]]],



         [[[[1.8235]],

           [[1.853...



         [[[[1.4873]],

           [[1.7090]],

           [[1.6346]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([100, 4, 1, 3, 1, 1])
SDDMM(message func: u_dot_e)
_____________________ test_sddmm[idtype1-dot-e-u-shp0-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'u', rhs_target = 'e'
msg = 'dot', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'dot'
lhs_data = tensor([[[[[[1.5628],
            [1.1029],
            [1.0513]]],


          [[[1.6238],
            [1.3869],
    ...8]]],


          [[[1.2889],
            [1.8150],
            [1.5238]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.5960]],

           [[1.2302]],

           [[1.8312]]]],



         [[[[1.6356]],

           [[1.247...



         [[[[1.9315]],

           [[1.7624]],

           [[1.5983]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([300, 4, 1, 3, 1, 1])
SDDMM(message func: u_dot_e)
_____________________ test_sddmm[idtype1-dot-e-u-shp1-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Sch...pe=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'u', rhs_target = 'e'
msg = 'dot', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'dot'
lhs_data = tensor([[[[[1.0774, 1.5703, 1.2509,  ..., 1.3731, 1.3843, 1.2944]],

          [[1.8360, 1.5992, 1.6855,  ..., 1.7838,...
          [[1.2743, 1.0855, 1.4721,  ..., 1.6444, 1.1865, 1.9987]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.4922, 1.4791, 1.6518,  ..., 1.7840, 1.7998, 1.5239],
           [1.5792, 1.1244, 1.9514,  ..., 1.9750, 1...
           [1.8071, 1.2642, 1.6370,  ..., 1.8452, 1.5992, 1.5798]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([100, 1, 3, 7, 7])
SDDMM(message func: u_dot_e)
_____________________ test_sddmm[idtype1-dot-e-u-shp1-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'u', rhs_target = 'e'
msg = 'dot', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'dot'
lhs_data = tensor([[[[[1.0130, 1.5309, 1.8017,  ..., 1.4307, 1.5458, 1.2170]],

          [[1.3876, 1.4547, 1.2631,  ..., 1.6803,...
          [[1.6241, 1.8992, 1.3570,  ..., 1.0490, 1.1917, 1.9640]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.3935, 1.7812, 1.8697,  ..., 1.5745, 1.0253, 1.7465],
           [1.9389, 1.5610, 1.7873,  ..., 1.6695, 1...
           [1.2640, 1.2689, 1.9104,  ..., 1.2295, 1.8020, 1.2428]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([300, 1, 3, 7, 7])
SDDMM(message func: u_dot_e)
_____________________ test_sddmm[idtype1-dot-e-u-shp2-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme...dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'u', rhs_target = 'e', msg = 'dot'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'dot'
lhs_data = tensor([[[[1.9378, 1.6432, 1.9686],
          [1.8623, 1.6083, 1.9538],
          [1.1355, 1.5319, 1.6126]]],


      ...3],
          [1.6490, 1.8281, 1.9841],
          [1.8922, 1.2374, 1.1853]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.1048, 1.0971, 1.7056]],

         [[1.5415, 1.6967, 1.8207]],

         [[1.5947, 1.6955, 1.4455]],

    ...

         [[1.3166, 1.1456, 1.0038]],

         [[1.6948, 1.7281, 1.1908]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([100, 4, 1, 3])
SDDMM(message func: u_dot_e)
_____________________ test_sddmm[idtype1-dot-e-u-shp2-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'u', rhs_target = 'e', msg = 'dot'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'dot'
lhs_data = tensor([[[[1.3032, 1.7717, 1.9217],
          [1.6208, 1.6544, 1.8709],
          [1.4111, 1.5549, 1.7264]]],


      ...5],
          [1.4005, 1.9868, 1.7700],
          [1.8589, 1.4737, 1.8538]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.9765, 1.7628, 1.1244]],

         [[1.6103, 1.5165, 1.3714]],

         [[1.1755, 1.2812, 1.7415]],

    ...

         [[1.6052, 1.7177, 1.1160]],

         [[1.5347, 1.0729, 1.7811]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([300, 4, 1, 3])
SDDMM(message func: u_dot_e)
_____________________ test_sddmm[idtype1-dot-e-u-shp3-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'u', rhs_target = 'e', msg = 'dot'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'dot'
lhs_data = tensor([[1.1682, 1.8107, 1.9583],
        [1.9434, 1.4679, 1.1802],
        [1.0121, 1.4922, 1.0515],
        [1.5045,... 1.8074],
        [1.0643, 1.5140, 1.6761],
        [1.8237, 1.1094, 1.6819]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.6334, 1.7235, 1.4422],
        [1.7956, 1.0305, 1.1689],
        [1.2592, 1.6216, 1.7027],
        [1.6347,... 1.9869],
        [1.0558, 1.0621, 1.9528],
        [1.6053, 1.9537, 1.5277]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([100, 3])
SDDMM(message func: u_dot_e)
_____________________ test_sddmm[idtype1-dot-e-u-shp3-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'u', rhs_target = 'e', msg = 'dot'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'dot'
lhs_data = tensor([[1.2733, 1.8742, 1.0625],
        [1.7340, 1.0669, 1.9306],
        [1.7317, 1.6711, 1.0970],
        [1.7523,... 1.2587],
        [1.6483, 1.9892, 1.9828],
        [1.6099, 1.9064, 1.1331]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.9571, 1.1342, 1.2127],
        [1.7155, 1.6162, 1.3206],
        [1.9695, 1.4545, 1.8853],
        [1.0698,... 1.4339],
        [1.3925, 1.7291, 1.1365],
        [1.8440, 1.3762, 1.9897]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([300, 3])
SDDMM(message func: u_dot_e)
_____________________ test_sddmm[idtype1-dot-e-u-shp4-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'u', rhs_target = 'e', msg = 'dot'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'dot'
lhs_data = tensor([[1.2130],
        [1.5956],
        [1.6414],
        [1.6334],
        [1.7067],
        [1.9887],
        [1...762],
        [1.9131],
        [1.4939],
        [1.6461],
        [1.4341]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.0316],
        [1.4667],
        [1.7724],
        [1.5337],
        [1.6631],
        [1.7637],
        [1...855],
        [1.1573],
        [1.2662],
        [1.0914],
        [1.1998]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([100, 1])
SDDMM(message func: u_dot_e)
_____________________ test_sddmm[idtype1-dot-e-u-shp4-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'u', rhs_target = 'e', msg = 'dot'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'dot'
lhs_data = tensor([[1.5709],
        [1.4020],
        [1.2995],
        [1.5997],
        [1.9352],
        [1.2500],
        [1...320],
        [1.2015],
        [1.0277],
        [1.2282],
        [1.3244]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.8539],
        [1.1600],
        [1.3582],
        [1.1121],
        [1.5962],
        [1.2492],
        [1...202],
        [1.1825],
        [1.1682],
        [1.2401],
        [1.1557]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([300, 1])
SDDMM(message func: u_dot_e)
_____________________ test_sddmm[idtype1-dot-e-v-shp0-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': ...torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'v', rhs_target = 'e'
msg = 'dot', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'dot'
lhs_data = tensor([[[[[[1.9421],
            [1.4252],
            [1.6577]]],


          [[[1.6132],
            [1.5193],
    ...3]]],


          [[[1.5714],
            [1.4502],
            [1.3485]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.1767]],

           [[1.5334]],

           [[1.9602]]]],



         [[[[1.1972]],

           [[1.688...



         [[[[1.9981]],

           [[1.5878]],

           [[1.0421]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([100, 4, 1, 3, 1, 1])
SDDMM(message func: v_dot_e)
_____________________ test_sddmm[idtype1-dot-e-v-shp0-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'v', rhs_target = 'e'
msg = 'dot', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'dot'
lhs_data = tensor([[[[[[1.4707],
            [1.9657],
            [1.6542]]],


          [[[1.4865],
            [1.7533],
    ...5]]],


          [[[1.9119],
            [1.9567],
            [1.7391]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.6920]],

           [[1.3153]],

           [[1.6453]]]],



         [[[[1.2596]],

           [[1.517...



         [[[[1.2392]],

           [[1.8777]],

           [[1.3116]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([40, 1, 2, 1, 3, 1]), rhs shape: torch.Size([300, 4, 1, 3, 1, 1])
SDDMM(message func: v_dot_e)
_____________________ test_sddmm[idtype1-dot-e-v-shp1-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Sch...pe=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'v', rhs_target = 'e'
msg = 'dot', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'dot'
lhs_data = tensor([[[[[1.3262, 1.6335, 1.0949,  ..., 1.0140, 1.2495, 1.7573]],

          [[1.7617, 1.9038, 1.6343,  ..., 1.7327,...
          [[1.2469, 1.0776, 1.0710,  ..., 1.6969, 1.9579, 1.3821]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.6950, 1.1490, 1.0880,  ..., 1.4891, 1.2473, 1.1150],
           [1.4769, 1.0638, 1.7498,  ..., 1.6114, 1...
           [1.8052, 1.9947, 1.8003,  ..., 1.5738, 1.3147, 1.5358]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([100, 1, 3, 7, 7])
SDDMM(message func: v_dot_e)
_____________________ test_sddmm[idtype1-dot-e-v-shp1-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'v', rhs_target = 'e'
msg = 'dot', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'dot'
lhs_data = tensor([[[[[1.1927, 1.7452, 1.9939,  ..., 1.4078, 1.1441, 1.5442]],

          [[1.7687, 1.3602, 1.6473,  ..., 1.2877,...
          [[1.2351, 1.5235, 1.2418,  ..., 1.3609, 1.8298, 1.6583]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.3881, 1.0503, 1.3004,  ..., 1.7390, 1.2464, 1.4040],
           [1.1079, 1.4754, 1.1690,  ..., 1.0385, 1...
           [1.8119, 1.7097, 1.9615,  ..., 1.3602, 1.7249, 1.7553]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([40, 5, 3, 1, 7]), rhs shape: torch.Size([300, 1, 3, 7, 7])
SDDMM(message func: v_dot_e)
_____________________ test_sddmm[idtype1-dot-e-v-shp2-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme...dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'v', rhs_target = 'e', msg = 'dot'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'dot'
lhs_data = tensor([[[[1.5362, 1.1293, 1.1563],
          [1.3976, 1.7856, 1.9868],
          [1.6447, 1.2823, 1.9782]]],


      ...9],
          [1.6567, 1.7843, 1.4148],
          [1.7014, 1.9047, 1.7100]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.5737, 1.4187, 1.9404]],

         [[1.8145, 1.2146, 1.2003]],

         [[1.6665, 1.0760, 1.4772]],

    ...

         [[1.9016, 1.6986, 1.8525]],

         [[1.4131, 1.4121, 1.0188]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([100, 4, 1, 3])
SDDMM(message func: v_dot_e)
_____________________ test_sddmm[idtype1-dot-e-v-shp2-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'v', rhs_target = 'e', msg = 'dot'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'dot'
lhs_data = tensor([[[[1.7355, 1.2201, 1.8749],
          [1.0209, 1.2152, 1.2575],
          [1.0469, 1.8115, 1.9880]]],


      ...4],
          [1.2193, 1.1487, 1.4605],
          [1.8169, 1.1896, 1.9021]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.0821, 1.4975, 1.0069]],

         [[1.8373, 1.1774, 1.5682]],

         [[1.1486, 1.7594, 1.7199]],

    ...

         [[1.6236, 1.8134, 1.6707]],

         [[1.9847, 1.4068, 1.7313]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([40, 1, 3, 3]), rhs shape: torch.Size([300, 4, 1, 3])
SDDMM(message func: v_dot_e)
_____________________ test_sddmm[idtype1-dot-e-v-shp3-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'v', rhs_target = 'e', msg = 'dot'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'dot'
lhs_data = tensor([[1.7478, 1.4414, 1.3539],
        [1.2022, 1.4407, 1.1001],
        [1.1745, 1.4530, 1.7985],
        [1.3743,... 1.3778],
        [1.6764, 1.2867, 1.8189],
        [1.6522, 1.9301, 1.9552]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.3251, 1.5877, 1.1435],
        [1.0050, 1.6677, 1.4602],
        [1.7983, 1.3158, 1.0337],
        [1.5934,... 1.5654],
        [1.4431, 1.7522, 1.9819],
        [1.3436, 1.9348, 1.1541]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([100, 3])
SDDMM(message func: v_dot_e)
_____________________ test_sddmm[idtype1-dot-e-v-shp3-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'v', rhs_target = 'e', msg = 'dot'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'dot'
lhs_data = tensor([[1.6681, 1.7116, 1.9496],
        [1.6292, 1.6256, 1.2970],
        [1.2133, 1.3908, 1.3690],
        [1.3701,... 1.1282],
        [1.5681, 1.2565, 1.3712],
        [1.3950, 1.1303, 1.4192]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.5167, 1.8730, 1.4248],
        [1.8278, 1.7106, 1.3435],
        [1.8316, 1.4778, 1.1982],
        [1.1976,... 1.6698],
        [1.0373, 1.9943, 1.1253],
        [1.8003, 1.0035, 1.2439]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([40, 3]), rhs shape: torch.Size([300, 3])
SDDMM(message func: v_dot_e)
_____________________ test_sddmm[idtype1-dot-e-v-shp4-g0] ______________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'v', rhs_target = 'e', msg = 'dot'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'dot'
lhs_data = tensor([[1.9119],
        [1.3743],
        [1.3638],
        [1.5771],
        [1.1736],
        [1.2490],
        [1...657],
        [1.0048],
        [1.1303],
        [1.9523],
        [1.1776]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.7678],
        [1.7468],
        [1.4362],
        [1.0762],
        [1.5290],
        [1.4939],
        [1...708],
        [1.0839],
        [1.9640],
        [1.3522],
        [1.1006]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([100, 1])
SDDMM(message func: v_dot_e)
_____________________ test_sddmm[idtype1-dot-e-v-shp4-g1] ______________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'v', rhs_target = 'e', msg = 'dot'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'dot'
lhs_data = tensor([[1.1687],
        [1.7145],
        [1.1558],
        [1.2259],
        [1.2953],
        [1.5087],
        [1...701],
        [1.0115],
        [1.6604],
        [1.5836],
        [1.0106]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.6716],
        [1.6043],
        [1.4566],
        [1.8221],
        [1.0443],
        [1.4419],
        [1...654],
        [1.2677],
        [1.1911],
        [1.0997],
        [1.5732]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([40, 1]), rhs shape: torch.Size([300, 1])
SDDMM(message func: v_dot_e)
___________________ test_sddmm[idtype1-copy_lhs-u-v-shp0-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': ...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'v', rhs_target = 'u'
msg = 'copy_lhs', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_lhs'
lhs_data = tensor([[[[[[1.8675],
            [1.0206],
            [1.2845]]],


          [[[1.8718],
            [1.7337],
    ...2]]],


          [[[1.1574],
            [1.4678],
            [1.4400]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.8690]],

           [[1.4323]],

           [[1.6786]]]],



         [[[[1.8232]],

           [[1.504...



         [[[[1.8948]],

           [[1.0844]],

           [[1.4069]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: v_copy_lhs_u)
___________________ test_sddmm[idtype1-copy_lhs-u-v-shp0-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'v', rhs_target = 'u'
msg = 'copy_lhs', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_lhs'
lhs_data = tensor([[[[[[1.2479],
            [1.1303],
            [1.4148]]],


          [[[1.2507],
            [1.6994],
    ...8]]],


          [[[1.6696],
            [1.8992],
            [1.6719]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.2308]],

           [[1.6940]],

           [[1.7230]]]],



         [[[[1.4613]],

           [[1.310...



         [[[[1.4941]],

           [[1.3921]],

           [[1.8305]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([40, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: v_copy_lhs_u)
___________________ test_sddmm[idtype1-copy_lhs-u-v-shp1-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Sch...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'v', rhs_target = 'u'
msg = 'copy_lhs', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_lhs'
lhs_data = tensor([[[[[1.4776, 1.9842, 1.4483,  ..., 1.7004, 1.7314, 1.9909]],

          [[1.7723, 1.5797, 1.6623,  ..., 1.1283,...
          [[1.3820, 1.7274, 1.9855,  ..., 1.9915, 1.9667, 1.1931]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.2706, 1.9567, 1.9919,  ..., 1.0507, 1.1191, 1.8330],
           [1.0977, 1.9548, 1.8454,  ..., 1.3631, 1...
           [1.8445, 1.5661, 1.3754,  ..., 1.7456, 1.0591, 1.2400]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: v_copy_lhs_u)
___________________ test_sddmm[idtype1-copy_lhs-u-v-shp1-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'v', rhs_target = 'u'
msg = 'copy_lhs', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_lhs'
lhs_data = tensor([[[[[1.7395, 1.0039, 1.3932,  ..., 1.1242, 1.0033, 1.2328]],

          [[1.9298, 1.3646, 1.2446,  ..., 1.7217,...
          [[1.8530, 1.6091, 1.9977,  ..., 1.1662, 1.4236, 1.0533]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.4937, 1.4621, 1.8748,  ..., 1.0403, 1.5010, 1.8164],
           [1.2791, 1.0800, 1.5329,  ..., 1.0259, 1...
           [1.3815, 1.5010, 1.7820,  ..., 1.4184, 1.1208, 1.5196]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([40, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: v_copy_lhs_u)
___________________ test_sddmm[idtype1-copy_lhs-u-v-shp2-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'v', rhs_target = 'u'
msg = 'copy_lhs', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_lhs'
lhs_data = tensor([[[[1.9954, 1.4527, 1.8213],
          [1.0003, 1.3334, 1.1451],
          [1.8654, 1.1197, 1.9362]]],


      ...5],
          [1.7221, 1.2098, 1.7111],
          [1.0525, 1.4768, 1.9315]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.1021, 1.2765, 1.0858]],

         [[1.5933, 1.8655, 1.3214]],

         [[1.9619, 1.9870, 1.1616]],

    ...

         [[1.0829, 1.9557, 1.7908]],

         [[1.4774, 1.2727, 1.0532]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: v_copy_lhs_u)
___________________ test_sddmm[idtype1-copy_lhs-u-v-shp2-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'v', rhs_target = 'u'
msg = 'copy_lhs', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_lhs'
lhs_data = tensor([[[[1.9108, 1.5720, 1.0175],
          [1.4092, 1.0124, 1.4880],
          [1.5475, 1.9800, 1.5418]]],


      ...5],
          [1.5868, 1.4542, 1.9297],
          [1.9646, 1.1813, 1.6155]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.2904, 1.8206, 1.4082]],

         [[1.4669, 1.1782, 1.6900]],

         [[1.3639, 1.4182, 1.4173]],

    ...

         [[1.4758, 1.3729, 1.1856]],

         [[1.7855, 1.2546, 1.7272]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([40, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: v_copy_lhs_u)
___________________ test_sddmm[idtype1-copy_lhs-u-v-shp3-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'v', rhs_target = 'u', msg = 'copy_lhs'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_lhs'
lhs_data = tensor([[1.0380, 1.5109, 1.2718],
        [1.0118, 1.0570, 1.9879],
        [1.4740, 1.3289, 1.8788],
        [1.8905,... 1.9536],
        [1.3018, 1.9381, 1.0982],
        [1.2108, 1.9424, 1.3818]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.5467, 1.7791, 1.2854],
        [1.9377, 1.3534, 1.3228],
        [1.4860, 1.6616, 1.8030],
        [1.1258,... 1.0048],
        [1.9627, 1.1595, 1.4121],
        [1.1290, 1.4941, 1.4106]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: v_copy_lhs_u)
___________________ test_sddmm[idtype1-copy_lhs-u-v-shp3-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'v', rhs_target = 'u', msg = 'copy_lhs'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_lhs'
lhs_data = tensor([[1.1638, 1.2192, 1.3485],
        [1.6900, 1.6789, 1.3356],
        [1.7039, 1.7088, 1.6117],
        [1.7337,... 1.7583],
        [1.4732, 1.7152, 1.8431],
        [1.4816, 1.2754, 1.4495]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.7489, 1.8365, 1.0498],
        [1.6201, 1.1686, 1.5363],
        [1.3387, 1.1068, 1.8025],
        [1.4558,... 1.9082],
        [1.8495, 1.6834, 1.8283],
        [1.5437, 1.7082, 1.0204]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([40, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: v_copy_lhs_u)
___________________ test_sddmm[idtype1-copy_lhs-u-v-shp4-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'v', rhs_target = 'u', msg = 'copy_lhs'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_lhs'
lhs_data = tensor([[1.2179],
        [1.4187],
        [1.7139],
        [1.4933],
        [1.3050],
        [1.1470],
        [1...210],
        [1.9160],
        [1.4267],
        [1.1241],
        [1.8079]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.0348],
        [1.1660],
        [1.1915],
        [1.9635],
        [1.8000],
        [1.2538],
        [1...124],
        [1.9605],
        [1.4172],
        [1.0912],
        [1.2544]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: v_copy_lhs_u)
___________________ test_sddmm[idtype1-copy_lhs-u-v-shp4-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'v', rhs_target = 'u', msg = 'copy_lhs'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_lhs'
lhs_data = tensor([[1.8500],
        [1.3071],
        [1.8243],
        [1.4282],
        [1.0176],
        [1.9596],
        [1...173],
        [1.0747],
        [1.3580],
        [1.6767],
        [1.3225]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.1051],
        [1.9993],
        [1.9146],
        [1.1581],
        [1.9463],
        [1.2304],
        [1...232],
        [1.4077],
        [1.7607],
        [1.2180],
        [1.7963]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([40, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: v_copy_lhs_u)
___________________ test_sddmm[idtype1-copy_lhs-u-e-shp0-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...torch.float64), 'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'e', rhs_target = 'u'
msg = 'copy_lhs', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_lhs'
lhs_data = tensor([[[[[[1.8004],
            [1.7387],
            [1.1754]]],


          [[[1.8799],
            [1.8597],
    ...3]]],


          [[[1.2546],
            [1.1127],
            [1.2069]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.8697]],

           [[1.4444]],

           [[1.8412]]]],



         [[[[1.8775]],

           [[1.732...



         [[[[1.4692]],

           [[1.1575]],

           [[1.8933]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([100, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: e_copy_lhs_u)
___________________ test_sddmm[idtype1-copy_lhs-u-e-shp0-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'e', rhs_target = 'u'
msg = 'copy_lhs', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_lhs'
lhs_data = tensor([[[[[[1.9791],
            [1.2185],
            [1.0215]]],


          [[[1.8914],
            [1.0920],
    ...0]]],


          [[[1.8214],
            [1.4672],
            [1.1450]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.5282]],

           [[1.9131]],

           [[1.1447]]]],



         [[[[1.1783]],

           [[1.953...



         [[[[1.6610]],

           [[1.7401]],

           [[1.4187]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([300, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: e_copy_lhs_u)
___________________ test_sddmm[idtype1-copy_lhs-u-e-shp1-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...pe=torch.float64), 'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'e', rhs_target = 'u'
msg = 'copy_lhs', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_lhs'
lhs_data = tensor([[[[[1.0841, 1.6331, 1.0541,  ..., 1.1931, 1.2801, 1.1540]],

          [[1.7310, 1.2420, 1.3768,  ..., 1.0252,...
          [[1.1757, 1.1763, 1.9841,  ..., 1.5540, 1.9391, 1.9718]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.6860, 1.6989, 1.8702,  ..., 1.9080, 1.8519, 1.4444],
           [1.1979, 1.9445, 1.2948,  ..., 1.7514, 1...
           [1.4639, 1.7800, 1.5167,  ..., 1.9648, 1.0010, 1.4364]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([100, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: e_copy_lhs_u)
___________________ test_sddmm[idtype1-copy_lhs-u-e-shp1-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'e', rhs_target = 'u'
msg = 'copy_lhs', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_lhs'
lhs_data = tensor([[[[[1.4043, 1.8206, 1.3378,  ..., 1.8676, 1.2239, 1.9980]],

          [[1.2413, 1.4278, 1.9326,  ..., 1.3758,...
          [[1.3958, 1.7872, 1.5683,  ..., 1.8476, 1.8194, 1.5296]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.6126, 1.8564, 1.4484,  ..., 1.5429, 1.5396, 1.4027],
           [1.6241, 1.6291, 1.3392,  ..., 1.4397, 1...
           [1.5192, 1.4930, 1.5353,  ..., 1.7218, 1.6692, 1.0060]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([300, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: e_copy_lhs_u)
___________________ test_sddmm[idtype1-copy_lhs-u-e-shp2-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...dtype=torch.float64), 'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'e', rhs_target = 'u'
msg = 'copy_lhs', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_lhs'
lhs_data = tensor([[[[1.3765, 1.3587, 1.8016],
          [1.7584, 1.9209, 1.3575],
          [1.4821, 1.1172, 1.4550]]],


      ...6],
          [1.6346, 1.6075, 1.5385],
          [1.2873, 1.5116, 1.3330]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.0976, 1.5019, 1.1483]],

         [[1.1734, 1.4702, 1.1699]],

         [[1.4126, 1.4572, 1.7934]],

    ...

         [[1.1007, 1.8222, 1.2812]],

         [[1.4515, 1.7894, 1.2175]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([100, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: e_copy_lhs_u)
___________________ test_sddmm[idtype1-copy_lhs-u-e-shp2-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'e', rhs_target = 'u'
msg = 'copy_lhs', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_lhs'
lhs_data = tensor([[[[1.9454, 1.2228, 1.8152],
          [1.0698, 1.9326, 1.0942],
          [1.8140, 1.0939, 1.1496]]],


      ...5],
          [1.7336, 1.7107, 1.0348],
          [1.6234, 1.4649, 1.3446]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.9643, 1.1033, 1.9716]],

         [[1.5143, 1.4862, 1.8411]],

         [[1.8734, 1.1087, 1.5753]],

    ...

         [[1.0258, 1.2825, 1.9649]],

         [[1.2877, 1.6651, 1.9773]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([300, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: e_copy_lhs_u)
___________________ test_sddmm[idtype1-copy_lhs-u-e-shp3-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'e', rhs_target = 'u', msg = 'copy_lhs'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_lhs'
lhs_data = tensor([[1.8907, 1.5974, 1.6311],
        [1.5089, 1.5788, 1.9862],
        [1.3157, 1.1306, 1.7316],
        [1.2479,... 1.1356],
        [1.1336, 1.6385, 1.8058],
        [1.1578, 1.1647, 1.4084]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.3833, 1.2045, 1.9940],
        [1.8560, 1.2968, 1.7646],
        [1.9188, 1.5058, 1.8088],
        [1.5003,... 1.1127],
        [1.7024, 1.8444, 1.9861],
        [1.1373, 1.0344, 1.3795]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([100, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: e_copy_lhs_u)
___________________ test_sddmm[idtype1-copy_lhs-u-e-shp3-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'e', rhs_target = 'u', msg = 'copy_lhs'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_lhs'
lhs_data = tensor([[1.1981, 1.5950, 1.1628],
        [1.4137, 1.4414, 1.1546],
        [1.4394, 1.3592, 1.2561],
        [1.1672,... 1.3905],
        [1.5206, 1.5852, 1.9453],
        [1.4618, 1.6912, 1.1424]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.3319, 1.6451, 1.6144],
        [1.0310, 1.2091, 1.0171],
        [1.0004, 1.5154, 1.1175],
        [1.7756,... 1.2391],
        [1.2468, 1.9850, 1.9108],
        [1.1753, 1.7900, 1.3673]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([300, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: e_copy_lhs_u)
___________________ test_sddmm[idtype1-copy_lhs-u-e-shp4-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'e', rhs_target = 'u', msg = 'copy_lhs'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_lhs'
lhs_data = tensor([[1.9316],
        [1.7104],
        [1.2784],
        [1.1190],
        [1.7336],
        [1.1146],
        [1...722],
        [1.4331],
        [1.3561],
        [1.6965],
        [1.1979]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.2909],
        [1.7442],
        [1.3241],
        [1.9588],
        [1.6560],
        [1.4196],
        [1...992],
        [1.7830],
        [1.1639],
        [1.6030],
        [1.7170]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([100, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: e_copy_lhs_u)
___________________ test_sddmm[idtype1-copy_lhs-u-e-shp4-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'e', rhs_target = 'u', msg = 'copy_lhs'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_lhs'
lhs_data = tensor([[1.3541],
        [1.7342],
        [1.5055],
        [1.4516],
        [1.9896],
        [1.2369],
        [1...843],
        [1.6608],
        [1.3701],
        [1.3041],
        [1.8732]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.3701],
        [1.3462],
        [1.8756],
        [1.9082],
        [1.2503],
        [1.7995],
        [1...924],
        [1.5982],
        [1.7592],
        [1.5850],
        [1.2001]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([300, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: e_copy_lhs_u)
___________________ test_sddmm[idtype1-copy_lhs-v-u-shp0-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': ...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'u', rhs_target = 'v'
msg = 'copy_lhs', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_lhs'
lhs_data = tensor([[[[[[1.1819],
            [1.0431],
            [1.8339]]],


          [[[1.1863],
            [1.6217],
    ...4]]],


          [[[1.3009],
            [1.0252],
            [1.2716]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.7874]],

           [[1.6619]],

           [[1.3176]]]],



         [[[[1.0016]],

           [[1.829...



         [[[[1.7675]],

           [[1.7694]],

           [[1.4944]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: u_copy_lhs_v)
___________________ test_sddmm[idtype1-copy_lhs-v-u-shp0-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'u', rhs_target = 'v'
msg = 'copy_lhs', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_lhs'
lhs_data = tensor([[[[[[1.3526],
            [1.1789],
            [1.6073]]],


          [[[1.8965],
            [1.0257],
    ...4]]],


          [[[1.5432],
            [1.9902],
            [1.2742]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.3220]],

           [[1.6849]],

           [[1.4223]]]],



         [[[[1.8699]],

           [[1.471...



         [[[[1.4108]],

           [[1.8084]],

           [[1.6405]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([40, 4, 1, 3, 1, 1])
SDDMM(message func: u_copy_lhs_v)
___________________ test_sddmm[idtype1-copy_lhs-v-u-shp1-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Sch...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'u', rhs_target = 'v'
msg = 'copy_lhs', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_lhs'
lhs_data = tensor([[[[[1.5561, 1.8871, 1.0184,  ..., 1.5064, 1.3270, 1.9694]],

          [[1.2264, 1.0025, 1.3948,  ..., 1.1065,...
          [[1.0992, 1.9297, 1.4461,  ..., 1.9264, 1.1843, 1.7846]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.8211, 1.9612, 1.9631,  ..., 1.0152, 1.4112, 1.3692],
           [1.2476, 1.0526, 1.3748,  ..., 1.6339, 1...
           [1.1154, 1.6598, 1.7627,  ..., 1.5784, 1.8155, 1.4158]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: u_copy_lhs_v)
___________________ test_sddmm[idtype1-copy_lhs-v-u-shp1-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'u', rhs_target = 'v'
msg = 'copy_lhs', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_lhs'
lhs_data = tensor([[[[[1.1794, 1.6164, 1.8565,  ..., 1.5151, 1.5993, 1.0030]],

          [[1.1231, 1.0350, 1.8215,  ..., 1.7676,...
          [[1.9976, 1.8567, 1.8239,  ..., 1.7237, 1.9598, 1.0943]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.2856, 1.8023, 1.6034,  ..., 1.9224, 1.8205, 1.6218],
           [1.2171, 1.1685, 1.4702,  ..., 1.1953, 1...
           [1.7701, 1.1971, 1.4810,  ..., 1.3332, 1.4286, 1.5929]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([40, 1, 3, 7, 7])
SDDMM(message func: u_copy_lhs_v)
___________________ test_sddmm[idtype1-copy_lhs-v-u-shp2-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'u', rhs_target = 'v'
msg = 'copy_lhs', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_lhs'
lhs_data = tensor([[[[1.8541, 1.8513, 1.8445],
          [1.7373, 1.2497, 1.4329],
          [1.6225, 1.4784, 1.0719]]],


      ...9],
          [1.3459, 1.2811, 1.3829],
          [1.8964, 1.0527, 1.1006]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.3944, 1.7932, 1.0765]],

         [[1.0277, 1.8146, 1.0207]],

         [[1.5182, 1.8331, 1.5898]],

    ...

         [[1.9572, 1.6994, 1.4964]],

         [[1.9161, 1.6408, 1.2235]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: u_copy_lhs_v)
___________________ test_sddmm[idtype1-copy_lhs-v-u-shp2-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'u', rhs_target = 'v'
msg = 'copy_lhs', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_lhs'
lhs_data = tensor([[[[1.1912, 1.9733, 1.1927],
          [1.3544, 1.1651, 1.1696],
          [1.0857, 1.9177, 1.1933]]],


      ...1],
          [1.9690, 1.2502, 1.4173],
          [1.5183, 1.1739, 1.9657]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.5289, 1.4814, 1.9917]],

         [[1.4739, 1.7412, 1.4337]],

         [[1.8769, 1.8290, 1.8942]],

    ...

         [[1.7712, 1.7710, 1.3239]],

         [[1.6718, 1.2734, 1.4105]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([40, 4, 1, 3])
SDDMM(message func: u_copy_lhs_v)
___________________ test_sddmm[idtype1-copy_lhs-v-u-shp3-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'u', rhs_target = 'v', msg = 'copy_lhs'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_lhs'
lhs_data = tensor([[1.7595, 1.4911, 1.8291],
        [1.6259, 1.0092, 1.5723],
        [1.0257, 1.5861, 1.7728],
        [1.7821,... 1.8016],
        [1.1909, 1.6391, 1.2835],
        [1.6429, 1.7901, 1.0500]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.1631, 1.6660, 1.7548],
        [1.7207, 1.7261, 1.7475],
        [1.5909, 1.9028, 1.6724],
        [1.7011,... 1.7503],
        [1.5954, 1.8909, 1.9868],
        [1.7760, 1.9593, 1.8097]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: u_copy_lhs_v)
___________________ test_sddmm[idtype1-copy_lhs-v-u-shp3-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'u', rhs_target = 'v', msg = 'copy_lhs'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_lhs'
lhs_data = tensor([[1.4060, 1.9452, 1.3289],
        [1.2762, 1.3225, 1.8319],
        [1.6641, 1.6746, 1.9498],
        [1.0168,... 1.5622],
        [1.2169, 1.9802, 1.1071],
        [1.3679, 1.1979, 1.1032]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.8127, 1.9901, 1.2777],
        [1.6354, 1.1500, 1.5413],
        [1.6792, 1.2982, 1.5686],
        [1.9665,... 1.5302],
        [1.0600, 1.6068, 1.9847],
        [1.4871, 1.6399, 1.1527]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([40, 3])
SDDMM(message func: u_copy_lhs_v)
___________________ test_sddmm[idtype1-copy_lhs-v-u-shp4-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'u', rhs_target = 'v', msg = 'copy_lhs'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_lhs'
lhs_data = tensor([[1.2681],
        [1.9923],
        [1.9238],
        [1.1621],
        [1.0512],
        [1.3139],
        [1...832],
        [1.5573],
        [1.2067],
        [1.0624],
        [1.3485]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.4228],
        [1.0737],
        [1.7576],
        [1.0290],
        [1.5293],
        [1.4278],
        [1...166],
        [1.7090],
        [1.8847],
        [1.6914],
        [1.8191]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: u_copy_lhs_v)
___________________ test_sddmm[idtype1-copy_lhs-v-u-shp4-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'u', rhs_target = 'v', msg = 'copy_lhs'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_lhs'
lhs_data = tensor([[1.5129],
        [1.6158],
        [1.0755],
        [1.3483],
        [1.0305],
        [1.2443],
        [1...713],
        [1.6458],
        [1.3623],
        [1.7179],
        [1.0628]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.8016],
        [1.6342],
        [1.4983],
        [1.3553],
        [1.4266],
        [1.7006],
        [1...910],
        [1.5633],
        [1.5668],
        [1.5028],
        [1.3914]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([40, 1])
SDDMM(message func: u_copy_lhs_v)
___________________ test_sddmm[idtype1-copy_lhs-v-e-shp0-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...torch.float64), 'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'e', rhs_target = 'v'
msg = 'copy_lhs', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_lhs'
lhs_data = tensor([[[[[[1.5734],
            [1.1405],
            [1.7113]]],


          [[[1.2978],
            [1.0789],
    ...7]]],


          [[[1.2074],
            [1.0350],
            [1.1639]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.1872]],

           [[1.1673]],

           [[1.9346]]]],



         [[[[1.5809]],

           [[1.991...



         [[[[1.9763]],

           [[1.9414]],

           [[1.3441]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([100, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: e_copy_lhs_v)
___________________ test_sddmm[idtype1-copy_lhs-v-e-shp0-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'e', rhs_target = 'v'
msg = 'copy_lhs', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_lhs'
lhs_data = tensor([[[[[[1.5317],
            [1.4458],
            [1.5616]]],


          [[[1.9567],
            [1.4268],
    ...2]]],


          [[[1.9130],
            [1.0495],
            [1.0960]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.1073]],

           [[1.6466]],

           [[1.8807]]]],



         [[[[1.1860]],

           [[1.682...



         [[[[1.2196]],

           [[1.0422]],

           [[1.4604]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([300, 1, 2, 1, 3, 1]), rhs shape: torch.Size([40, 4, 1, 3, 1, 1])
SDDMM(message func: e_copy_lhs_v)
___________________ test_sddmm[idtype1-copy_lhs-v-e-shp1-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...pe=torch.float64), 'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'e', rhs_target = 'v'
msg = 'copy_lhs', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_lhs'
lhs_data = tensor([[[[[1.0753, 1.1372, 1.7311,  ..., 1.6447, 1.0433, 1.9471]],

          [[1.5203, 1.9014, 1.5135,  ..., 1.8196,...
          [[1.6356, 1.8445, 1.6899,  ..., 1.7706, 1.9419, 1.1745]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.8705, 1.0115, 1.5437,  ..., 1.0341, 1.2846, 1.8375],
           [1.1253, 1.0789, 1.1565,  ..., 1.2553, 1...
           [1.9245, 1.7686, 1.2121,  ..., 1.0896, 1.6521, 1.5982]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([100, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: e_copy_lhs_v)
___________________ test_sddmm[idtype1-copy_lhs-v-e-shp1-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'e', rhs_target = 'v'
msg = 'copy_lhs', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_lhs'
lhs_data = tensor([[[[[1.6210, 1.4089, 1.8352,  ..., 1.3706, 1.7449, 1.8426]],

          [[1.1158, 1.3159, 1.9025,  ..., 1.3879,...
          [[1.4771, 1.8165, 1.0119,  ..., 1.8106, 1.9181, 1.8695]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.9510, 1.4526, 1.3147,  ..., 1.5819, 1.0340, 1.9429],
           [1.1381, 1.5083, 1.2966,  ..., 1.2857, 1...
           [1.1451, 1.6690, 1.2771,  ..., 1.7415, 1.1328, 1.7266]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([300, 5, 3, 1, 7]), rhs shape: torch.Size([40, 1, 3, 7, 7])
SDDMM(message func: e_copy_lhs_v)
___________________ test_sddmm[idtype1-copy_lhs-v-e-shp2-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...dtype=torch.float64), 'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'e', rhs_target = 'v'
msg = 'copy_lhs', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_lhs'
lhs_data = tensor([[[[1.5295, 1.9029, 1.2710],
          [1.9051, 1.7200, 1.2496],
          [1.1494, 1.1254, 1.8200]]],


      ...2],
          [1.9319, 1.0570, 1.7514],
          [1.2434, 1.0833, 1.1355]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.6901, 1.1123, 1.5896]],

         [[1.0123, 1.1778, 1.1965]],

         [[1.0898, 1.8625, 1.9610]],

    ...

         [[1.7705, 1.1424, 1.6507]],

         [[1.8691, 1.1874, 1.8027]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([100, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: e_copy_lhs_v)
___________________ test_sddmm[idtype1-copy_lhs-v-e-shp2-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'e', rhs_target = 'v'
msg = 'copy_lhs', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_lhs'
lhs_data = tensor([[[[1.5354, 1.2174, 1.0828],
          [1.3556, 1.3016, 1.1139],
          [1.9361, 1.0697, 1.4871]]],


      ...0],
          [1.4658, 1.1459, 1.5662],
          [1.5611, 1.2196, 1.5398]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.7081, 1.9562, 1.4367]],

         [[1.3434, 1.5381, 1.3691]],

         [[1.3118, 1.3399, 1.0414]],

    ...

         [[1.7419, 1.6202, 1.6748]],

         [[1.3423, 1.7560, 1.1994]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([300, 1, 3, 3]), rhs shape: torch.Size([40, 4, 1, 3])
SDDMM(message func: e_copy_lhs_v)
___________________ test_sddmm[idtype1-copy_lhs-v-e-shp3-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'e', rhs_target = 'v', msg = 'copy_lhs'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_lhs'
lhs_data = tensor([[1.0050, 1.5193, 1.1750],
        [1.5025, 1.6315, 1.0934],
        [1.2715, 1.8252, 1.2383],
        [1.6853,... 1.4193],
        [1.7477, 1.0726, 1.1497],
        [1.4576, 1.2901, 1.1567]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.7333, 1.0006, 1.1735],
        [1.1952, 1.6252, 1.0293],
        [1.0436, 1.2068, 1.6157],
        [1.0427,... 1.4087],
        [1.5819, 1.6604, 1.5157],
        [1.2078, 1.8540, 1.2035]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([100, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: e_copy_lhs_v)
___________________ test_sddmm[idtype1-copy_lhs-v-e-shp3-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'e', rhs_target = 'v', msg = 'copy_lhs'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_lhs'
lhs_data = tensor([[1.7094, 1.4927, 1.7456],
        [1.0073, 1.7838, 1.1406],
        [1.7166, 1.3654, 1.0492],
        [1.9667,... 1.8427],
        [1.5711, 1.7431, 1.1064],
        [1.4164, 1.5577, 1.5394]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.5906, 1.8498, 1.8729],
        [1.5606, 1.2851, 1.7560],
        [1.3134, 1.1908, 1.6316],
        [1.6748,... 1.4669],
        [1.8417, 1.8713, 1.8511],
        [1.1412, 1.1232, 1.0036]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([300, 3]), rhs shape: torch.Size([40, 3])
SDDMM(message func: e_copy_lhs_v)
___________________ test_sddmm[idtype1-copy_lhs-v-e-shp4-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'e', rhs_target = 'v', msg = 'copy_lhs'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_lhs'
lhs_data = tensor([[1.4950],
        [1.5639],
        [1.0186],
        [1.3247],
        [1.0631],
        [1.0082],
        [1...885],
        [1.5755],
        [1.9805],
        [1.8463],
        [1.1116]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.8780],
        [1.9256],
        [1.9051],
        [1.5508],
        [1.4576],
        [1.7243],
        [1...428],
        [1.1386],
        [1.1359],
        [1.5470],
        [1.5669]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([100, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: e_copy_lhs_v)
___________________ test_sddmm[idtype1-copy_lhs-v-e-shp4-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'e', rhs_target = 'v', msg = 'copy_lhs'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_lhs'
lhs_data = tensor([[1.0860],
        [1.2440],
        [1.2445],
        [1.0440],
        [1.3505],
        [1.8235],
        [1...023],
        [1.9755],
        [1.1235],
        [1.9354],
        [1.6193]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.9962],
        [1.5200],
        [1.6211],
        [1.5661],
        [1.2273],
        [1.4411],
        [1...761],
        [1.1355],
        [1.2812],
        [1.6534],
        [1.4744]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([300, 1]), rhs shape: torch.Size([40, 1])
SDDMM(message func: e_copy_lhs_v)
___________________ test_sddmm[idtype1-copy_lhs-e-u-shp0-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': ...torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'u', rhs_target = 'e'
msg = 'copy_lhs', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_lhs'
lhs_data = tensor([[[[[[1.7601],
            [1.7939],
            [1.4167]]],


          [[[1.7733],
            [1.0478],
    ...9]]],


          [[[1.5769],
            [1.4548],
            [1.9332]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.7977]],

           [[1.6251]],

           [[1.2796]]]],



         [[[[1.2260]],

           [[1.767...



         [[[[1.9805]],

           [[1.9779]],

           [[1.4258]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([100, 4, 1, 3, 1, 1])
SDDMM(message func: u_copy_lhs_e)
___________________ test_sddmm[idtype1-copy_lhs-e-u-shp0-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'u', rhs_target = 'e'
msg = 'copy_lhs', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_lhs'
lhs_data = tensor([[[[[[1.9763],
            [1.4426],
            [1.1327]]],


          [[[1.8984],
            [1.8083],
    ...9]]],


          [[[1.1778],
            [1.8545],
            [1.3696]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.2201]],

           [[1.8154]],

           [[1.1607]]]],



         [[[[1.0270]],

           [[1.089...



         [[[[1.4671]],

           [[1.6281]],

           [[1.6014]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([300, 4, 1, 3, 1, 1])
SDDMM(message func: u_copy_lhs_e)
___________________ test_sddmm[idtype1-copy_lhs-e-u-shp1-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Sch...pe=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'u', rhs_target = 'e'
msg = 'copy_lhs', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_lhs'
lhs_data = tensor([[[[[1.5867, 1.1550, 1.9071,  ..., 1.9460, 1.7336, 1.3658]],

          [[1.2041, 1.0029, 1.1502,  ..., 1.6352,...
          [[1.3715, 1.7795, 1.6505,  ..., 1.4812, 1.7781, 1.2727]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.3842, 1.1454, 1.8414,  ..., 1.7993, 1.1740, 1.0609],
           [1.1740, 1.9631, 1.5938,  ..., 1.1665, 1...
           [1.5190, 1.7778, 1.7393,  ..., 1.7129, 1.9409, 1.0582]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([100, 1, 3, 7, 7])
SDDMM(message func: u_copy_lhs_e)
___________________ test_sddmm[idtype1-copy_lhs-e-u-shp1-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'u', rhs_target = 'e'
msg = 'copy_lhs', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_lhs'
lhs_data = tensor([[[[[1.9701, 1.7894, 1.3187,  ..., 1.6066, 1.3179, 1.3830]],

          [[1.5272, 1.0412, 1.1507,  ..., 1.4315,...
          [[1.2352, 1.6113, 1.9964,  ..., 1.9928, 1.2116, 1.4515]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.1616, 1.5474, 1.2346,  ..., 1.9076, 1.9521, 1.1645],
           [1.6608, 1.5260, 1.9431,  ..., 1.0269, 1...
           [1.4955, 1.1688, 1.9228,  ..., 1.3397, 1.4086, 1.8147]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([300, 1, 3, 7, 7])
SDDMM(message func: u_copy_lhs_e)
___________________ test_sddmm[idtype1-copy_lhs-e-u-shp2-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme...dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'u', rhs_target = 'e'
msg = 'copy_lhs', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_lhs'
lhs_data = tensor([[[[1.1581, 1.5576, 1.1531],
          [1.4989, 1.8287, 1.0739],
          [1.6781, 1.2461, 1.4105]]],


      ...7],
          [1.9582, 1.4221, 1.0420],
          [1.5072, 1.0601, 1.9200]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.2365, 1.9025, 1.6727]],

         [[1.7989, 1.4046, 1.2815]],

         [[1.7886, 1.7697, 1.6580]],

    ...

         [[1.7329, 1.1433, 1.4419]],

         [[1.6142, 1.8687, 1.0488]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([100, 4, 1, 3])
SDDMM(message func: u_copy_lhs_e)
___________________ test_sddmm[idtype1-copy_lhs-e-u-shp2-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'u', rhs_target = 'e'
msg = 'copy_lhs', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_lhs'
lhs_data = tensor([[[[1.5994, 1.3044, 1.2601],
          [1.1898, 1.6585, 1.5091],
          [1.1382, 1.0728, 1.7889]]],


      ...1],
          [1.1673, 1.6966, 1.9452],
          [1.8129, 1.7478, 1.2764]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.6374, 1.1347, 1.0506]],

         [[1.0259, 1.2595, 1.3427]],

         [[1.9415, 1.1672, 1.7548]],

    ...

         [[1.1249, 1.1933, 1.2638]],

         [[1.7123, 1.8035, 1.4671]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([300, 4, 1, 3])
SDDMM(message func: u_copy_lhs_e)
___________________ test_sddmm[idtype1-copy_lhs-e-u-shp3-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'u', rhs_target = 'e', msg = 'copy_lhs'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_lhs'
lhs_data = tensor([[1.1524, 1.7764, 1.2455],
        [1.8400, 1.5934, 1.5055],
        [1.6096, 1.3194, 1.1589],
        [1.6952,... 1.1816],
        [1.5138, 1.5859, 1.4181],
        [1.0022, 1.2448, 1.7641]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.3579, 1.5256, 1.1661],
        [1.1946, 1.4070, 1.8829],
        [1.6965, 1.3512, 1.8653],
        [1.8575,... 1.5090],
        [1.5951, 1.8295, 1.0938],
        [1.9126, 1.7937, 1.2962]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([100, 3])
SDDMM(message func: u_copy_lhs_e)
___________________ test_sddmm[idtype1-copy_lhs-e-u-shp3-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'u', rhs_target = 'e', msg = 'copy_lhs'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_lhs'
lhs_data = tensor([[1.8771, 1.7487, 1.8288],
        [1.9405, 1.3970, 1.9049],
        [1.6163, 1.3027, 1.9998],
        [1.2615,... 1.1422],
        [1.1329, 1.8300, 1.6169],
        [1.7184, 1.8721, 1.2175]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.1439, 1.6657, 1.2080],
        [1.5668, 1.8912, 1.9410],
        [1.6489, 1.0878, 1.6488],
        [1.6279,... 1.1961],
        [1.7581, 1.9509, 1.4490],
        [1.8712, 1.4019, 1.5435]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([300, 3])
SDDMM(message func: u_copy_lhs_e)
___________________ test_sddmm[idtype1-copy_lhs-e-u-shp4-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'u', rhs_target = 'e', msg = 'copy_lhs'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_lhs'
lhs_data = tensor([[1.4236],
        [1.3263],
        [1.6429],
        [1.4935],
        [1.7650],
        [1.9343],
        [1...080],
        [1.3809],
        [1.7107],
        [1.6608],
        [1.6767]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.6225],
        [1.7148],
        [1.4774],
        [1.3209],
        [1.1273],
        [1.8710],
        [1...986],
        [1.5104],
        [1.0052],
        [1.9681],
        [1.6338]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([100, 1])
SDDMM(message func: u_copy_lhs_e)
___________________ test_sddmm[idtype1-copy_lhs-e-u-shp4-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'u', rhs_target = 'e', msg = 'copy_lhs'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_lhs'
lhs_data = tensor([[1.9557],
        [1.8414],
        [1.7188],
        [1.6424],
        [1.5514],
        [1.6066],
        [1...559],
        [1.9973],
        [1.1468],
        [1.3689],
        [1.3196]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.0706],
        [1.0113],
        [1.9389],
        [1.2934],
        [1.0140],
        [1.0298],
        [1...700],
        [1.0576],
        [1.9075],
        [1.6596],
        [1.8565]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([300, 1])
SDDMM(message func: u_copy_lhs_e)
___________________ test_sddmm[idtype1-copy_lhs-e-v-shp0-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': ...torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'v', rhs_target = 'e'
msg = 'copy_lhs', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_lhs'
lhs_data = tensor([[[[[[1.7965],
            [1.4236],
            [1.0504]]],


          [[[1.7718],
            [1.3539],
    ...0]]],


          [[[1.3637],
            [1.1902],
            [1.2667]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.1633]],

           [[1.5660]],

           [[1.3113]]]],



         [[[[1.9385]],

           [[1.011...



         [[[[1.7276]],

           [[1.4577]],

           [[1.3539]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([100, 4, 1, 3, 1, 1])
SDDMM(message func: v_copy_lhs_e)
___________________ test_sddmm[idtype1-copy_lhs-e-v-shp0-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'v', rhs_target = 'e'
msg = 'copy_lhs', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_lhs'
lhs_data = tensor([[[[[[1.2883],
            [1.7912],
            [1.2649]]],


          [[[1.3669],
            [1.1166],
    ...7]]],


          [[[1.4534],
            [1.0721],
            [1.3245]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.9550]],

           [[1.7226]],

           [[1.0386]]]],



         [[[[1.5542]],

           [[1.078...



         [[[[1.7417]],

           [[1.0968]],

           [[1.0021]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([40, 1, 2, 1, 3, 1]), rhs shape: torch.Size([300, 4, 1, 3, 1, 1])
SDDMM(message func: v_copy_lhs_e)
___________________ test_sddmm[idtype1-copy_lhs-e-v-shp1-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Sch...pe=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'v', rhs_target = 'e'
msg = 'copy_lhs', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_lhs'
lhs_data = tensor([[[[[1.1367, 1.1999, 1.5531,  ..., 1.8294, 1.4571, 1.8476]],

          [[1.8770, 1.2740, 1.4006,  ..., 1.0029,...
          [[1.2366, 1.2043, 1.4790,  ..., 1.7014, 1.3797, 1.0508]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.8340, 1.2619, 1.8488,  ..., 1.0034, 1.5965, 1.1817],
           [1.0523, 1.3089, 1.2682,  ..., 1.1485, 1...
           [1.9518, 1.2787, 1.2554,  ..., 1.3111, 1.3383, 1.1257]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([100, 1, 3, 7, 7])
SDDMM(message func: v_copy_lhs_e)
___________________ test_sddmm[idtype1-copy_lhs-e-v-shp1-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'v', rhs_target = 'e'
msg = 'copy_lhs', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_lhs'
lhs_data = tensor([[[[[1.7329, 1.6879, 1.0965,  ..., 1.4957, 1.1479, 1.5070]],

          [[1.5472, 1.4661, 1.0626,  ..., 1.1828,...
          [[1.1696, 1.5439, 1.2435,  ..., 1.7212, 1.5104, 1.0815]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.0825, 1.6407, 1.5275,  ..., 1.3233, 1.3831, 1.6069],
           [1.8063, 1.7358, 1.1506,  ..., 1.5304, 1...
           [1.7249, 1.3259, 1.2372,  ..., 1.6936, 1.5356, 1.0344]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([40, 5, 3, 1, 7]), rhs shape: torch.Size([300, 1, 3, 7, 7])
SDDMM(message func: v_copy_lhs_e)
___________________ test_sddmm[idtype1-copy_lhs-e-v-shp2-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme...dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'v', rhs_target = 'e'
msg = 'copy_lhs', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_lhs'
lhs_data = tensor([[[[1.2454, 1.9277, 1.9022],
          [1.5053, 1.7084, 1.2193],
          [1.5668, 1.4877, 1.1230]]],


      ...8],
          [1.8198, 1.1848, 1.8085],
          [1.7506, 1.5056, 1.8104]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.3308, 1.6315, 1.6962]],

         [[1.6126, 1.2834, 1.5749]],

         [[1.3179, 1.8458, 1.9977]],

    ...

         [[1.1543, 1.1017, 1.1460]],

         [[1.1239, 1.0624, 1.3871]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([100, 4, 1, 3])
SDDMM(message func: v_copy_lhs_e)
___________________ test_sddmm[idtype1-copy_lhs-e-v-shp2-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'v', rhs_target = 'e'
msg = 'copy_lhs', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_lhs'
lhs_data = tensor([[[[1.8221, 1.6875, 1.5640],
          [1.8905, 1.4823, 1.9607],
          [1.3243, 1.2018, 1.6402]]],


      ...9],
          [1.4759, 1.5234, 1.2140],
          [1.2721, 1.8308, 1.3807]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.5962, 1.1077, 1.1539]],

         [[1.0634, 1.6628, 1.4460]],

         [[1.1336, 1.2426, 1.5354]],

    ...

         [[1.2232, 1.1218, 1.6699]],

         [[1.5097, 1.0936, 1.9467]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([40, 1, 3, 3]), rhs shape: torch.Size([300, 4, 1, 3])
SDDMM(message func: v_copy_lhs_e)
___________________ test_sddmm[idtype1-copy_lhs-e-v-shp3-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'v', rhs_target = 'e', msg = 'copy_lhs'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_lhs'
lhs_data = tensor([[1.1907, 1.9696, 1.6135],
        [1.9904, 1.5014, 1.1565],
        [1.8698, 1.6235, 1.4486],
        [1.7127,... 1.3882],
        [1.4177, 1.6879, 1.9970],
        [1.7496, 1.6167, 1.4577]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.9638, 1.1755, 1.0783],
        [1.2550, 1.8097, 1.8491],
        [1.0469, 1.2364, 1.4575],
        [1.2483,... 1.4466],
        [1.6013, 1.5042, 1.6380],
        [1.4088, 1.6600, 1.0450]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([100, 3])
SDDMM(message func: v_copy_lhs_e)
___________________ test_sddmm[idtype1-copy_lhs-e-v-shp3-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'v', rhs_target = 'e', msg = 'copy_lhs'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_lhs'
lhs_data = tensor([[1.9336, 1.7381, 1.4603],
        [1.6139, 1.6232, 1.0338],
        [1.9312, 1.9130, 1.9724],
        [1.4374,... 1.9102],
        [1.7352, 1.3562, 1.3169],
        [1.9285, 1.8843, 1.6642]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.2354, 1.9116, 1.5873],
        [1.3373, 1.8557, 1.6164],
        [1.0514, 1.5079, 1.7633],
        [1.6622,... 1.2034],
        [1.3018, 1.6270, 1.7017],
        [1.8403, 1.6896, 1.3867]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([40, 3]), rhs shape: torch.Size([300, 3])
SDDMM(message func: v_copy_lhs_e)
___________________ test_sddmm[idtype1-copy_lhs-e-v-shp4-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'v', rhs_target = 'e', msg = 'copy_lhs'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_lhs'
lhs_data = tensor([[1.5008],
        [1.4250],
        [1.8372],
        [1.4599],
        [1.6957],
        [1.1407],
        [1...703],
        [1.0152],
        [1.3272],
        [1.1778],
        [1.4062]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.7431],
        [1.5568],
        [1.9677],
        [1.5446],
        [1.1946],
        [1.0412],
        [1...940],
        [1.0818],
        [1.6499],
        [1.2651],
        [1.0512]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([100, 1])
SDDMM(message func: v_copy_lhs_e)
___________________ test_sddmm[idtype1-copy_lhs-e-v-shp4-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'v', rhs_target = 'e', msg = 'copy_lhs'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_lhs'
lhs_data = tensor([[1.4615],
        [1.5584],
        [1.4473],
        [1.9378],
        [1.8598],
        [1.4749],
        [1...129],
        [1.0186],
        [1.7965],
        [1.4086],
        [1.4539]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.2790],
        [1.3208],
        [1.1413],
        [1.1194],
        [1.4245],
        [1.9019],
        [1...215],
        [1.3027],
        [1.7433],
        [1.3230],
        [1.3154]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([40, 1]), rhs shape: torch.Size([300, 1])
SDDMM(message func: v_copy_lhs_e)
___________________ test_sddmm[idtype1-copy_rhs-u-v-shp0-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': ...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'v', rhs_target = 'u'
msg = 'copy_rhs', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_rhs'
lhs_data = tensor([[[[[[1.6291],
            [1.1225],
            [1.7415]]],


          [[[1.5790],
            [1.3822],
    ...1]]],


          [[[1.7486],
            [1.1184],
            [1.2849]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.9585]],

           [[1.9334]],

           [[1.1448]]]],



         [[[[1.3890]],

           [[1.918...



         [[[[1.2672]],

           [[1.3649]],

           [[1.5683]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: v_copy_rhs_u)
___________________ test_sddmm[idtype1-copy_rhs-u-v-shp0-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'v', rhs_target = 'u'
msg = 'copy_rhs', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_rhs'
lhs_data = tensor([[[[[[1.4937],
            [1.4818],
            [1.8651]]],


          [[[1.3721],
            [1.2252],
    ...2]]],


          [[[1.3301],
            [1.1518],
            [1.6225]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.6242]],

           [[1.3513]],

           [[1.9170]]]],



         [[[[1.2673]],

           [[1.434...



         [[[[1.9827]],

           [[1.5036]],

           [[1.1952]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([40, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: v_copy_rhs_u)
___________________ test_sddmm[idtype1-copy_rhs-u-v-shp1-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Sch...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'v', rhs_target = 'u'
msg = 'copy_rhs', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_rhs'
lhs_data = tensor([[[[[1.2061, 1.2313, 1.5480,  ..., 1.0157, 1.6696, 1.1659]],

          [[1.1176, 1.5177, 1.6735,  ..., 1.5570,...
          [[1.1135, 1.2751, 1.2031,  ..., 1.6622, 1.3587, 1.2122]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.2331, 1.6460, 1.3804,  ..., 1.8505, 1.9845, 1.0986],
           [1.4978, 1.7674, 1.0306,  ..., 1.6187, 1...
           [1.0995, 1.3293, 1.5871,  ..., 1.5373, 1.1107, 1.1755]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: v_copy_rhs_u)
___________________ test_sddmm[idtype1-copy_rhs-u-v-shp1-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'v', rhs_target = 'u'
msg = 'copy_rhs', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_rhs'
lhs_data = tensor([[[[[1.8677, 1.3000, 1.3058,  ..., 1.3390, 1.9030, 1.8391]],

          [[1.1638, 1.9529, 1.2580,  ..., 1.8058,...
          [[1.8923, 1.0498, 1.1947,  ..., 1.9103, 1.1152, 1.8329]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.0625, 1.7962, 1.5774,  ..., 1.1361, 1.4652, 1.6187],
           [1.3761, 1.3514, 1.2769,  ..., 1.3487, 1...
           [1.9117, 1.2729, 1.4320,  ..., 1.4265, 1.6270, 1.8220]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([40, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: v_copy_rhs_u)
___________________ test_sddmm[idtype1-copy_rhs-u-v-shp2-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'v', rhs_target = 'u'
msg = 'copy_rhs', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_rhs'
lhs_data = tensor([[[[1.4188, 1.9769, 1.7908],
          [1.6082, 1.7355, 1.1373],
          [1.6330, 1.1088, 1.7386]]],


      ...7],
          [1.6349, 1.6788, 1.2188],
          [1.6931, 1.6234, 1.9384]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.4415, 1.6879, 1.8654]],

         [[1.1386, 1.8167, 1.7708]],

         [[1.1599, 1.1233, 1.0628]],

    ...

         [[1.7573, 1.6637, 1.2333]],

         [[1.3928, 1.0783, 1.6524]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: v_copy_rhs_u)
___________________ test_sddmm[idtype1-copy_rhs-u-v-shp2-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'v', rhs_target = 'u'
msg = 'copy_rhs', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_rhs'
lhs_data = tensor([[[[1.2492, 1.3507, 1.0544],
          [1.5834, 1.4312, 1.0431],
          [1.0876, 1.9952, 1.5027]]],


      ...6],
          [1.3787, 1.0537, 1.1970],
          [1.9258, 1.6100, 1.1759]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.6347, 1.8895, 1.2946]],

         [[1.8939, 1.0732, 1.0343]],

         [[1.0201, 1.4766, 1.2948]],

    ...

         [[1.0581, 1.6616, 1.1952]],

         [[1.5389, 1.4259, 1.0294]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([40, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: v_copy_rhs_u)
___________________ test_sddmm[idtype1-copy_rhs-u-v-shp3-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'v', rhs_target = 'u', msg = 'copy_rhs'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_rhs'
lhs_data = tensor([[1.2420, 1.3425, 1.0938],
        [1.9123, 1.7822, 1.6531],
        [1.7166, 1.9094, 1.3112],
        [1.9534,... 1.3844],
        [1.1347, 1.6533, 1.2640],
        [1.6527, 1.5157, 1.5305]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.0392, 1.5213, 1.7219],
        [1.4467, 1.7178, 1.0755],
        [1.2275, 1.2936, 1.5568],
        [1.1621,... 1.0614],
        [1.5292, 1.3066, 1.1765],
        [1.8768, 1.1567, 1.3539]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: v_copy_rhs_u)
___________________ test_sddmm[idtype1-copy_rhs-u-v-shp3-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'v', rhs_target = 'u', msg = 'copy_rhs'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_rhs'
lhs_data = tensor([[1.9762, 1.9818, 1.0645],
        [1.5738, 1.2004, 1.2534],
        [1.5774, 1.5576, 1.7900],
        [1.5631,... 1.9511],
        [1.6534, 1.7226, 1.8922],
        [1.3786, 1.5588, 1.6608]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.3335, 1.5122, 1.6639],
        [1.8889, 1.7903, 1.4100],
        [1.5278, 1.4712, 1.1353],
        [1.4736,... 1.5093],
        [1.0424, 1.9540, 1.7089],
        [1.6041, 1.1928, 1.0700]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([40, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: v_copy_rhs_u)
___________________ test_sddmm[idtype1-copy_rhs-u-v-shp4-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'v', rhs_target = 'u', msg = 'copy_rhs'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_rhs'
lhs_data = tensor([[1.9662],
        [1.2567],
        [1.4651],
        [1.7389],
        [1.1952],
        [1.8174],
        [1...777],
        [1.7647],
        [1.6692],
        [1.0981],
        [1.7161]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.7650],
        [1.8312],
        [1.2242],
        [1.4089],
        [1.3459],
        [1.5577],
        [1...709],
        [1.2251],
        [1.9662],
        [1.3730],
        [1.9566]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: v_copy_rhs_u)
___________________ test_sddmm[idtype1-copy_rhs-u-v-shp4-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'v', rhs_target = 'u', msg = 'copy_rhs'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_rhs'
lhs_data = tensor([[1.1264],
        [1.5515],
        [1.0168],
        [1.9571],
        [1.5646],
        [1.2618],
        [1...017],
        [1.8337],
        [1.8917],
        [1.0919],
        [1.0326]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.6986],
        [1.4126],
        [1.4957],
        [1.0176],
        [1.5817],
        [1.0533],
        [1...679],
        [1.5279],
        [1.4464],
        [1.9217],
        [1.7157]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([40, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: v_copy_rhs_u)
___________________ test_sddmm[idtype1-copy_rhs-u-e-shp0-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...torch.float64), 'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'e', rhs_target = 'u'
msg = 'copy_rhs', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_rhs'
lhs_data = tensor([[[[[[1.0456],
            [1.8746],
            [1.8577]]],


          [[[1.9211],
            [1.4567],
    ...3]]],


          [[[1.2162],
            [1.8553],
            [1.4634]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.0702]],

           [[1.7371]],

           [[1.9267]]]],



         [[[[1.0578]],

           [[1.370...



         [[[[1.3504]],

           [[1.7387]],

           [[1.6076]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([100, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: e_copy_rhs_u)
___________________ test_sddmm[idtype1-copy_rhs-u-e-shp0-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'e', rhs_target = 'u'
msg = 'copy_rhs', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_rhs'
lhs_data = tensor([[[[[[1.6132],
            [1.3854],
            [1.4197]]],


          [[[1.4399],
            [1.0520],
    ...4]]],


          [[[1.2280],
            [1.5006],
            [1.5257]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.5059]],

           [[1.2807]],

           [[1.5405]]]],



         [[[[1.0886]],

           [[1.208...



         [[[[1.1911]],

           [[1.0602]],

           [[1.6928]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([300, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: e_copy_rhs_u)
___________________ test_sddmm[idtype1-copy_rhs-u-e-shp1-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...pe=torch.float64), 'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'e', rhs_target = 'u'
msg = 'copy_rhs', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_rhs'
lhs_data = tensor([[[[[1.7366, 1.9760, 1.6664,  ..., 1.3234, 1.0130, 1.3156]],

          [[1.6481, 1.9281, 1.2488,  ..., 1.8182,...
          [[1.7611, 1.2102, 1.9551,  ..., 1.0094, 1.9745, 1.7921]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.9706, 1.9971, 1.7967,  ..., 1.0984, 1.7114, 1.8928],
           [1.7321, 1.2868, 1.9791,  ..., 1.6976, 1...
           [1.1580, 1.7554, 1.6277,  ..., 1.0015, 1.0374, 1.7349]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([100, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: e_copy_rhs_u)
___________________ test_sddmm[idtype1-copy_rhs-u-e-shp1-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'e', rhs_target = 'u'
msg = 'copy_rhs', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_rhs'
lhs_data = tensor([[[[[1.7653, 1.8709, 1.5487,  ..., 1.4052, 1.5651, 1.7352]],

          [[1.0579, 1.1672, 1.9320,  ..., 1.7057,...
          [[1.4762, 1.5008, 1.5274,  ..., 1.2248, 1.5698, 1.9883]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.0299, 1.0493, 1.0160,  ..., 1.5696, 1.4054, 1.4249],
           [1.7669, 1.0288, 1.3058,  ..., 1.5749, 1...
           [1.5038, 1.8184, 1.8289,  ..., 1.1258, 1.1972, 1.2629]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([300, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: e_copy_rhs_u)
___________________ test_sddmm[idtype1-copy_rhs-u-e-shp2-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...dtype=torch.float64), 'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'e', rhs_target = 'u'
msg = 'copy_rhs', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_rhs'
lhs_data = tensor([[[[1.2053, 1.0080, 1.2298],
          [1.3946, 1.9386, 1.5017],
          [1.1396, 1.4608, 1.3740]]],


      ...8],
          [1.9695, 1.0130, 1.7822],
          [1.0094, 1.3753, 1.9390]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.4886, 1.7947, 1.4601]],

         [[1.7778, 1.1963, 1.0723]],

         [[1.6671, 1.3425, 1.0912]],

    ...

         [[1.0365, 1.8839, 1.5855]],

         [[1.4497, 1.7258, 1.9009]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([100, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: e_copy_rhs_u)
___________________ test_sddmm[idtype1-copy_rhs-u-e-shp2-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'e', rhs_target = 'u'
msg = 'copy_rhs', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_rhs'
lhs_data = tensor([[[[1.0450, 1.1668, 1.4157],
          [1.5253, 1.6875, 1.9128],
          [1.7759, 1.2874, 1.3063]]],


      ...1],
          [1.5048, 1.4824, 1.1163],
          [1.3741, 1.7316, 1.2224]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.2775, 1.0533, 1.8606]],

         [[1.8647, 1.0778, 1.1814]],

         [[1.1673, 1.0298, 1.4962]],

    ...

         [[1.7116, 1.8171, 1.4476]],

         [[1.7066, 1.1091, 1.4305]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([300, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: e_copy_rhs_u)
___________________ test_sddmm[idtype1-copy_rhs-u-e-shp3-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'e', rhs_target = 'u', msg = 'copy_rhs'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_rhs'
lhs_data = tensor([[1.6106, 1.6705, 1.1544],
        [1.8285, 1.3085, 1.0105],
        [1.2865, 1.3508, 1.8545],
        [1.5860,... 1.1928],
        [1.3755, 1.1253, 1.1409],
        [1.1397, 1.8812, 1.9931]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.9261, 1.5791, 1.0972],
        [1.1572, 1.6486, 1.7997],
        [1.8080, 1.7699, 1.0784],
        [1.3790,... 1.5344],
        [1.5284, 1.6683, 1.6730],
        [1.4040, 1.2984, 1.7466]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([100, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: e_copy_rhs_u)
___________________ test_sddmm[idtype1-copy_rhs-u-e-shp3-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'e', rhs_target = 'u', msg = 'copy_rhs'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_rhs'
lhs_data = tensor([[1.6096, 1.7304, 1.1662],
        [1.4496, 1.7845, 1.1277],
        [1.1843, 1.3712, 1.9589],
        [1.7669,... 1.4617],
        [1.1178, 1.0133, 1.6632],
        [1.9586, 1.5956, 1.4733]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.0732, 1.5918, 1.6487],
        [1.9801, 1.2260, 1.1374],
        [1.3195, 1.8637, 1.0720],
        [1.9616,... 1.3366],
        [1.0032, 1.2965, 1.8382],
        [1.9083, 1.2041, 1.1688]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([300, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: e_copy_rhs_u)
___________________ test_sddmm[idtype1-copy_rhs-u-e-shp4-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'e', rhs_target = 'u', msg = 'copy_rhs'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_rhs'
lhs_data = tensor([[1.0068],
        [1.9424],
        [1.0747],
        [1.9272],
        [1.4238],
        [1.7208],
        [1...447],
        [1.7090],
        [1.3542],
        [1.9789],
        [1.1186]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.3103],
        [1.6103],
        [1.2788],
        [1.0737],
        [1.1961],
        [1.6237],
        [1...994],
        [1.4367],
        [1.9222],
        [1.8798],
        [1.9240]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([100, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: e_copy_rhs_u)
___________________ test_sddmm[idtype1-copy_rhs-u-e-shp4-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'e', rhs_target = 'u', msg = 'copy_rhs'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_rhs'
lhs_data = tensor([[1.1983],
        [1.1172],
        [1.3279],
        [1.8725],
        [1.2942],
        [1.1378],
        [1...779],
        [1.0489],
        [1.3661],
        [1.9353],
        [1.0367]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.9565],
        [1.8461],
        [1.1652],
        [1.5473],
        [1.8192],
        [1.5600],
        [1...258],
        [1.7318],
        [1.7198],
        [1.2356],
        [1.2440]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'u'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([300, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: e_copy_rhs_u)
___________________ test_sddmm[idtype1-copy_rhs-v-u-shp0-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': ...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'u', rhs_target = 'v'
msg = 'copy_rhs', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_rhs'
lhs_data = tensor([[[[[[1.6794],
            [1.6223],
            [1.7873]]],


          [[[1.3737],
            [1.4448],
    ...5]]],


          [[[1.4546],
            [1.6278],
            [1.1623]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.9937]],

           [[1.2781]],

           [[1.6332]]]],



         [[[[1.7344]],

           [[1.534...



         [[[[1.0389]],

           [[1.5966]],

           [[1.7849]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: u_copy_rhs_v)
___________________ test_sddmm[idtype1-copy_rhs-v-u-shp0-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'u', rhs_target = 'v'
msg = 'copy_rhs', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_rhs'
lhs_data = tensor([[[[[[1.7566],
            [1.6998],
            [1.4486]]],


          [[[1.8060],
            [1.0930],
    ...5]]],


          [[[1.0224],
            [1.1479],
            [1.7822]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.4029]],

           [[1.4715]],

           [[1.7374]]]],



         [[[[1.3103]],

           [[1.124...



         [[[[1.6265]],

           [[1.2343]],

           [[1.4707]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([40, 4, 1, 3, 1, 1])
SDDMM(message func: u_copy_rhs_v)
___________________ test_sddmm[idtype1-copy_rhs-v-u-shp1-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Sch...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'u', rhs_target = 'v'
msg = 'copy_rhs', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_rhs'
lhs_data = tensor([[[[[1.4042, 1.1688, 1.1855,  ..., 1.4124, 1.7078, 1.2472]],

          [[1.8047, 1.9794, 1.6942,  ..., 1.7401,...
          [[1.2402, 1.6104, 1.5274,  ..., 1.2434, 1.9983, 1.1747]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.8149, 1.8766, 1.0050,  ..., 1.5803, 1.3096, 1.8311],
           [1.6920, 1.3454, 1.2707,  ..., 1.8658, 1...
           [1.9819, 1.3907, 1.0912,  ..., 1.2610, 1.0254, 1.7970]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: u_copy_rhs_v)
___________________ test_sddmm[idtype1-copy_rhs-v-u-shp1-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'u', rhs_target = 'v'
msg = 'copy_rhs', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_rhs'
lhs_data = tensor([[[[[1.5741, 1.3235, 1.9688,  ..., 1.3610, 1.0009, 1.4182]],

          [[1.4844, 1.3355, 1.1535,  ..., 1.9405,...
          [[1.8138, 1.6719, 1.1254,  ..., 1.2182, 1.3649, 1.1402]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.2152, 1.4514, 1.3310,  ..., 1.7310, 1.8854, 1.1400],
           [1.6516, 1.0507, 1.6635,  ..., 1.9981, 1...
           [1.8831, 1.5544, 1.6347,  ..., 1.2754, 1.9061, 1.6695]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([40, 1, 3, 7, 7])
SDDMM(message func: u_copy_rhs_v)
___________________ test_sddmm[idtype1-copy_rhs-v-u-shp2-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'u', rhs_target = 'v'
msg = 'copy_rhs', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_rhs'
lhs_data = tensor([[[[1.9125, 1.9425, 1.1205],
          [1.8770, 1.7751, 1.9139],
          [1.4841, 1.8045, 1.3477]]],


      ...6],
          [1.2569, 1.9663, 1.8199],
          [1.7656, 1.7314, 1.6401]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.7372, 1.3161, 1.4013]],

         [[1.6525, 1.3166, 1.1082]],

         [[1.6602, 1.3911, 1.1412]],

    ...

         [[1.0177, 1.2549, 1.8863]],

         [[1.1960, 1.1550, 1.5574]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: u_copy_rhs_v)
___________________ test_sddmm[idtype1-copy_rhs-v-u-shp2-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'u', rhs_target = 'v'
msg = 'copy_rhs', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_rhs'
lhs_data = tensor([[[[1.6892, 1.4718, 1.5288],
          [1.1747, 1.1061, 1.1128],
          [1.2894, 1.6658, 1.0061]]],


      ...7],
          [1.2705, 1.4962, 1.2117],
          [1.5349, 1.6612, 1.7823]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.8535, 1.8309, 1.0402]],

         [[1.3840, 1.5707, 1.6884]],

         [[1.8716, 1.6700, 1.9982]],

    ...

         [[1.8960, 1.4371, 1.8981]],

         [[1.4773, 1.6815, 1.6405]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([40, 4, 1, 3])
SDDMM(message func: u_copy_rhs_v)
___________________ test_sddmm[idtype1-copy_rhs-v-u-shp3-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'u', rhs_target = 'v', msg = 'copy_rhs'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_rhs'
lhs_data = tensor([[1.8630, 1.9060, 1.1776],
        [1.0902, 1.3517, 1.5325],
        [1.8940, 1.4555, 1.6522],
        [1.4899,... 1.7218],
        [1.5857, 1.6441, 1.2994],
        [1.4839, 1.6116, 1.1497]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.1736, 1.2842, 1.0662],
        [1.5156, 1.3894, 1.2688],
        [1.3504, 1.4003, 1.1837],
        [1.0969,... 1.4538],
        [1.6112, 1.0936, 1.7959],
        [1.9215, 1.5730, 1.0588]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: u_copy_rhs_v)
___________________ test_sddmm[idtype1-copy_rhs-v-u-shp3-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'u', rhs_target = 'v', msg = 'copy_rhs'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_rhs'
lhs_data = tensor([[1.6159, 1.4771, 1.6227],
        [1.0444, 1.1715, 1.0147],
        [1.7458, 1.2922, 1.0138],
        [1.7999,... 1.4271],
        [1.7738, 1.6324, 1.1667],
        [1.5142, 1.9934, 1.6457]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.8607, 1.2119, 1.4634],
        [1.2545, 1.9394, 1.2924],
        [1.1723, 1.4647, 1.6411],
        [1.7368,... 1.7276],
        [1.3771, 1.6292, 1.2366],
        [1.3783, 1.7941, 1.1480]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([40, 3])
SDDMM(message func: u_copy_rhs_v)
___________________ test_sddmm[idtype1-copy_rhs-v-u-shp4-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'u', rhs_target = 'v', msg = 'copy_rhs'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_rhs'
lhs_data = tensor([[1.7830],
        [1.7972],
        [1.7725],
        [1.7153],
        [1.6229],
        [1.2132],
        [1...722],
        [1.8942],
        [1.9986],
        [1.1903],
        [1.8871]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.1100],
        [1.8672],
        [1.3104],
        [1.5136],
        [1.3420],
        [1.4550],
        [1...951],
        [1.9523],
        [1.9164],
        [1.6898],
        [1.6262]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: u_copy_rhs_v)
___________________ test_sddmm[idtype1-copy_rhs-v-u-shp4-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'u', rhs_target = 'v', msg = 'copy_rhs'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_rhs'
lhs_data = tensor([[1.1395],
        [1.4913],
        [1.5789],
        [1.0858],
        [1.8546],
        [1.1825],
        [1...521],
        [1.3482],
        [1.1771],
        [1.4772],
        [1.7653]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.8269],
        [1.5564],
        [1.2909],
        [1.8875],
        [1.5540],
        [1.9768],
        [1...350],
        [1.7076],
        [1.7867],
        [1.4643],
        [1.8863]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([40, 1])
SDDMM(message func: u_copy_rhs_v)
___________________ test_sddmm[idtype1-copy_rhs-v-e-shp0-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...torch.float64), 'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'e', rhs_target = 'v'
msg = 'copy_rhs', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_rhs'
lhs_data = tensor([[[[[[1.6238],
            [1.6203],
            [1.2871]]],


          [[[1.9735],
            [1.8839],
    ...2]]],


          [[[1.5462],
            [1.7423],
            [1.9263]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.9123]],

           [[1.4183]],

           [[1.5502]]]],



         [[[[1.4487]],

           [[1.308...



         [[[[1.7153]],

           [[1.7501]],

           [[1.9170]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([100, 1, 2, 1, 3, 1]), rhs shape: torch.Size([30, 4, 1, 3, 1, 1])
SDDMM(message func: e_copy_rhs_v)
___________________ test_sddmm[idtype1-copy_rhs-v-e-shp0-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'e', rhs_target = 'v'
msg = 'copy_rhs', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_rhs'
lhs_data = tensor([[[[[[1.6770],
            [1.7819],
            [1.0424]]],


          [[[1.7217],
            [1.8366],
    ...0]]],


          [[[1.2316],
            [1.4700],
            [1.5471]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.8948]],

           [[1.3100]],

           [[1.2132]]]],



         [[[[1.7865]],

           [[1.435...



         [[[[1.2657]],

           [[1.9176]],

           [[1.3631]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([300, 1, 2, 1, 3, 1]), rhs shape: torch.Size([40, 4, 1, 3, 1, 1])
SDDMM(message func: e_copy_rhs_v)
___________________ test_sddmm[idtype1-copy_rhs-v-e-shp1-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...pe=torch.float64), 'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'e', rhs_target = 'v'
msg = 'copy_rhs', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_rhs'
lhs_data = tensor([[[[[1.9641, 1.9815, 1.4448,  ..., 1.6561, 1.8218, 1.3373]],

          [[1.3055, 1.9929, 1.0163,  ..., 1.9257,...
          [[1.3281, 1.4286, 1.5223,  ..., 1.2736, 1.8459, 1.7464]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.9849, 1.4282, 1.0569,  ..., 1.7170, 1.0159, 1.1048],
           [1.2116, 1.8215, 1.9905,  ..., 1.0697, 1...
           [1.5769, 1.6737, 1.0088,  ..., 1.7638, 1.5744, 1.7185]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([100, 5, 3, 1, 7]), rhs shape: torch.Size([30, 1, 3, 7, 7])
SDDMM(message func: e_copy_rhs_v)
___________________ test_sddmm[idtype1-copy_rhs-v-e-shp1-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'e', rhs_target = 'v'
msg = 'copy_rhs', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_rhs'
lhs_data = tensor([[[[[1.8546, 1.3536, 1.4759,  ..., 1.0847, 1.8553, 1.2774]],

          [[1.7191, 1.6016, 1.7964,  ..., 1.5747,...
          [[1.6939, 1.1369, 1.6721,  ..., 1.8769, 1.6617, 1.0097]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.0663, 1.2857, 1.4270,  ..., 1.5127, 1.4912, 1.8075],
           [1.7535, 1.1541, 1.9257,  ..., 1.4877, 1...
           [1.7343, 1.0070, 1.0867,  ..., 1.4413, 1.8222, 1.4980]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([300, 5, 3, 1, 7]), rhs shape: torch.Size([40, 1, 3, 7, 7])
SDDMM(message func: e_copy_rhs_v)
___________________ test_sddmm[idtype1-copy_rhs-v-e-shp2-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...dtype=torch.float64), 'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'e', rhs_target = 'v'
msg = 'copy_rhs', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_rhs'
lhs_data = tensor([[[[1.1118, 1.4788, 1.2563],
          [1.8143, 1.9665, 1.9869],
          [1.4352, 1.1784, 1.5436]]],


      ...6],
          [1.6767, 1.5017, 1.4281],
          [1.4779, 1.6798, 1.8858]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.8794, 1.6636, 1.0376]],

         [[1.7517, 1.6194, 1.3183]],

         [[1.5614, 1.3730, 1.5258]],

    ...

         [[1.9802, 1.9723, 1.8544]],

         [[1.8060, 1.6154, 1.5927]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([100, 1, 3, 3]), rhs shape: torch.Size([30, 4, 1, 3])
SDDMM(message func: e_copy_rhs_v)
___________________ test_sddmm[idtype1-copy_rhs-v-e-shp2-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'e', rhs_target = 'v'
msg = 'copy_rhs', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_rhs'
lhs_data = tensor([[[[1.8964, 1.4154, 1.5482],
          [1.1941, 1.4415, 1.1312],
          [1.1268, 1.0640, 1.5958]]],


      ...2],
          [1.7960, 1.4057, 1.8019],
          [1.1760, 1.1674, 1.7909]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.3069, 1.1007, 1.0777]],

         [[1.5394, 1.9101, 1.6580]],

         [[1.6451, 1.8117, 1.0866]],

    ...

         [[1.3082, 1.4998, 1.8956]],

         [[1.2093, 1.8587, 1.6293]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([300, 1, 3, 3]), rhs shape: torch.Size([40, 4, 1, 3])
SDDMM(message func: e_copy_rhs_v)
___________________ test_sddmm[idtype1-copy_rhs-v-e-shp3-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'e', rhs_target = 'v', msg = 'copy_rhs'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_rhs'
lhs_data = tensor([[1.6281, 1.2408, 1.1825],
        [1.4080, 1.5573, 1.3649],
        [1.7294, 1.1051, 1.8609],
        [1.0601,... 1.7188],
        [1.0737, 1.6160, 1.0153],
        [1.5974, 1.6757, 1.7323]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.7611, 1.7558, 1.8339],
        [1.5734, 1.6849, 1.8636],
        [1.0387, 1.6301, 1.5041],
        [1.3624,... 1.8514],
        [1.3114, 1.1233, 1.8182],
        [1.6630, 1.4949, 1.5909]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([100, 3]), rhs shape: torch.Size([30, 3])
SDDMM(message func: e_copy_rhs_v)
___________________ test_sddmm[idtype1-copy_rhs-v-e-shp3-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'e', rhs_target = 'v', msg = 'copy_rhs'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_rhs'
lhs_data = tensor([[1.9612, 1.8720, 1.6527],
        [1.9495, 1.5509, 1.3765],
        [1.2584, 1.3356, 1.5273],
        [1.7433,... 1.0302],
        [1.2033, 1.2182, 1.0469],
        [1.5828, 1.4369, 1.4399]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.6722, 1.8646, 1.6738],
        [1.6234, 1.2459, 1.3900],
        [1.4756, 1.8122, 1.4645],
        [1.2687,... 1.0888],
        [1.5728, 1.2693, 1.0705],
        [1.7029, 1.2034, 1.3514]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([300, 3]), rhs shape: torch.Size([40, 3])
SDDMM(message func: e_copy_rhs_v)
___________________ test_sddmm[idtype1-copy_rhs-v-e-shp4-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'e', rhs_target = 'v', msg = 'copy_rhs'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_rhs'
lhs_data = tensor([[1.3955],
        [1.0309],
        [1.4836],
        [1.8744],
        [1.0316],
        [1.6302],
        [1...165],
        [1.9590],
        [1.7205],
        [1.7072],
        [1.1169]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.7460],
        [1.9300],
        [1.4345],
        [1.7185],
        [1.9642],
        [1.4376],
        [1...618],
        [1.5139],
        [1.3723],
        [1.5475],
        [1.3429]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([100, 1]), rhs shape: torch.Size([30, 1])
SDDMM(message func: e_copy_rhs_v)
___________________ test_sddmm[idtype1-copy_rhs-v-e-shp4-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'e', rhs_target = 'v', msg = 'copy_rhs'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_rhs'
lhs_data = tensor([[1.3825],
        [1.5744],
        [1.1796],
        [1.2382],
        [1.5182],
        [1.6538],
        [1...041],
        [1.2699],
        [1.0968],
        [1.8226],
        [1.7192]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.8729],
        [1.5064],
        [1.0722],
        [1.6177],
        [1.9078],
        [1.9900],
        [1...665],
        [1.1091],
        [1.6736],
        [1.1175],
        [1.2328]], dtype=torch.float64, requires_grad=True)
lhs_target = 'e', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([300, 1]), rhs shape: torch.Size([40, 1])
SDDMM(message func: e_copy_rhs_v)
___________________ test_sddmm[idtype1-copy_rhs-e-u-shp0-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': ...torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'u', rhs_target = 'e'
msg = 'copy_rhs', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_rhs'
lhs_data = tensor([[[[[[1.5461],
            [1.1134],
            [1.9148]]],


          [[[1.0466],
            [1.5476],
    ...8]]],


          [[[1.6569],
            [1.7629],
            [1.3791]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.2359]],

           [[1.7339]],

           [[1.8069]]]],



         [[[[1.1511]],

           [[1.271...



         [[[[1.9445]],

           [[1.7840]],

           [[1.5601]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([100, 4, 1, 3, 1, 1])
SDDMM(message func: u_copy_rhs_e)
___________________ test_sddmm[idtype1-copy_rhs-e-u-shp0-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'u', rhs_target = 'e'
msg = 'copy_rhs', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_rhs'
lhs_data = tensor([[[[[[1.4922],
            [1.3346],
            [1.7629]]],


          [[[1.8040],
            [1.1739],
    ...8]]],


          [[[1.7703],
            [1.7001],
            [1.9953]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.8304]],

           [[1.3147]],

           [[1.3362]]]],



         [[[[1.0286]],

           [[1.303...



         [[[[1.8334]],

           [[1.1880]],

           [[1.3397]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([300, 4, 1, 3, 1, 1])
SDDMM(message func: u_copy_rhs_e)
___________________ test_sddmm[idtype1-copy_rhs-e-u-shp1-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Sch...pe=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'u', rhs_target = 'e'
msg = 'copy_rhs', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_rhs'
lhs_data = tensor([[[[[1.8624, 1.6415, 1.3106,  ..., 1.2416, 1.2662, 1.3377]],

          [[1.8340, 1.7144, 1.1928,  ..., 1.6145,...
          [[1.1213, 1.4219, 1.6994,  ..., 1.5997, 1.5291, 1.4684]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.5920, 1.5072, 1.7589,  ..., 1.0577, 1.4105, 1.7537],
           [1.1820, 1.5643, 1.9879,  ..., 1.0542, 1...
           [1.4105, 1.7766, 1.5906,  ..., 1.9838, 1.2326, 1.1804]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([100, 1, 3, 7, 7])
SDDMM(message func: u_copy_rhs_e)
___________________ test_sddmm[idtype1-copy_rhs-e-u-shp1-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'u', rhs_target = 'e'
msg = 'copy_rhs', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_rhs'
lhs_data = tensor([[[[[1.2140, 1.2734, 1.5587,  ..., 1.8210, 1.2328, 1.7916]],

          [[1.9830, 1.9815, 1.6637,  ..., 1.9394,...
          [[1.3839, 1.5979, 1.5713,  ..., 1.8910, 1.1787, 1.3532]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.2148, 1.1031, 1.0556,  ..., 1.3033, 1.7380, 1.3854],
           [1.4534, 1.4142, 1.8542,  ..., 1.7547, 1...
           [1.5417, 1.3774, 1.1831,  ..., 1.0114, 1.3820, 1.8008]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([300, 1, 3, 7, 7])
SDDMM(message func: u_copy_rhs_e)
___________________ test_sddmm[idtype1-copy_rhs-e-u-shp2-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme...dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'u', rhs_target = 'e'
msg = 'copy_rhs', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_rhs'
lhs_data = tensor([[[[1.0862, 1.5947, 1.9391],
          [1.9296, 1.8309, 1.4733],
          [1.9755, 1.1658, 1.9348]]],


      ...2],
          [1.1396, 1.2926, 1.4690],
          [1.8883, 1.0083, 1.5965]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.1449, 1.9681, 1.3970]],

         [[1.9209, 1.8942, 1.8088]],

         [[1.5332, 1.9072, 1.4435]],

    ...

         [[1.3262, 1.6057, 1.3069]],

         [[1.0330, 1.0424, 1.1016]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([100, 4, 1, 3])
SDDMM(message func: u_copy_rhs_e)
___________________ test_sddmm[idtype1-copy_rhs-e-u-shp2-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'u', rhs_target = 'e'
msg = 'copy_rhs', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_rhs'
lhs_data = tensor([[[[1.9025, 1.4085, 1.8118],
          [1.6018, 1.2329, 1.9827],
          [1.1081, 1.6972, 1.9571]]],


      ...7],
          [1.3857, 1.1395, 1.5252],
          [1.4621, 1.2023, 1.2420]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.5660, 1.7580, 1.6765]],

         [[1.5567, 1.2891, 1.3040]],

         [[1.4994, 1.0405, 1.9391]],

    ...

         [[1.8702, 1.4696, 1.1758]],

         [[1.2114, 1.6079, 1.2159]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([300, 4, 1, 3])
SDDMM(message func: u_copy_rhs_e)
___________________ test_sddmm[idtype1-copy_rhs-e-u-shp3-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'u', rhs_target = 'e', msg = 'copy_rhs'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_rhs'
lhs_data = tensor([[1.3189, 1.1416, 1.6804],
        [1.0114, 1.8601, 1.0906],
        [1.6158, 1.2644, 1.9614],
        [1.5814,... 1.5004],
        [1.6899, 1.3111, 1.1417],
        [1.9739, 1.8635, 1.4802]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.8650, 1.0127, 1.2948],
        [1.9417, 1.5314, 1.9632],
        [1.5745, 1.7152, 1.2097],
        [1.2700,... 1.5164],
        [1.2597, 1.4262, 1.8704],
        [1.9379, 1.8416, 1.3589]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([100, 3])
SDDMM(message func: u_copy_rhs_e)
___________________ test_sddmm[idtype1-copy_rhs-e-u-shp3-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'u', rhs_target = 'e', msg = 'copy_rhs'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_rhs'
lhs_data = tensor([[1.4449, 1.5294, 1.3899],
        [1.0322, 1.9397, 1.0844],
        [1.6004, 1.5091, 1.6544],
        [1.5420,... 1.3438],
        [1.9588, 1.4960, 1.2632],
        [1.6547, 1.2876, 1.2955]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.7808, 1.9129, 1.8497],
        [1.4977, 1.7720, 1.8779],
        [1.1004, 1.5152, 1.1619],
        [1.0339,... 1.1049],
        [1.5561, 1.3637, 1.1545],
        [1.7710, 1.3495, 1.3663]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([300, 3])
SDDMM(message func: u_copy_rhs_e)
___________________ test_sddmm[idtype1-copy_rhs-e-u-shp4-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'u', rhs_target = 'e', msg = 'copy_rhs'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_rhs'
lhs_data = tensor([[1.4225],
        [1.7946],
        [1.1119],
        [1.1458],
        [1.8793],
        [1.5906],
        [1...385],
        [1.5711],
        [1.1216],
        [1.9969],
        [1.2157]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.0313],
        [1.9611],
        [1.5375],
        [1.5047],
        [1.0283],
        [1.8949],
        [1...917],
        [1.5854],
        [1.4629],
        [1.5031],
        [1.6580]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([100, 1])
SDDMM(message func: u_copy_rhs_e)
___________________ test_sddmm[idtype1-copy_rhs-e-u-shp4-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'u', rhs_target = 'e', msg = 'copy_rhs'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_rhs'
lhs_data = tensor([[1.8480],
        [1.3502],
        [1.1393],
        [1.0587],
        [1.9366],
        [1.8910],
        [1...100],
        [1.0277],
        [1.9124],
        [1.2885],
        [1.6785]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.3557],
        [1.5511],
        [1.9906],
        [1.6949],
        [1.6239],
        [1.6512],
        [1...219],
        [1.7536],
        [1.0172],
        [1.3477],
        [1.5440]], dtype=torch.float64, requires_grad=True)
lhs_target = 'u', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([300, 1])
SDDMM(message func: u_copy_rhs_e)
___________________ test_sddmm[idtype1-copy_rhs-e-v-shp0-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': ...torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'v', rhs_target = 'e'
msg = 'copy_rhs', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_rhs'
lhs_data = tensor([[[[[[1.4433],
            [1.9211],
            [1.2812]]],


          [[[1.3511],
            [1.8356],
    ...2]]],


          [[[1.7621],
            [1.9022],
            [1.5878]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.7974]],

           [[1.6621]],

           [[1.4124]]]],



         [[[[1.2558]],

           [[1.650...



         [[[[1.7451]],

           [[1.5954]],

           [[1.2952]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1, 2, 1, 3, 1]), rhs shape: torch.Size([100, 4, 1, 3, 1, 1])
SDDMM(message func: v_copy_rhs_e)
___________________ test_sddmm[idtype1-copy_rhs-e-v-shp0-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 2, 1, 3, 1), (4, 1, 3, 1, 1)), lhs_target = 'v', rhs_target = 'e'
msg = 'copy_rhs', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_rhs'
lhs_data = tensor([[[[[[1.8300],
            [1.3979],
            [1.1178]]],


          [[[1.2207],
            [1.7805],
    ...9]]],


          [[[1.1486],
            [1.1917],
            [1.5695]]]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[[1.8957]],

           [[1.5145]],

           [[1.3299]]]],



         [[[[1.7190]],

           [[1.777...



         [[[[1.8994]],

           [[1.8122]],

           [[1.6836]]]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([40, 1, 2, 1, 3, 1]), rhs shape: torch.Size([300, 4, 1, 3, 1, 1])
SDDMM(message func: v_copy_rhs_e)
___________________ test_sddmm[idtype1-copy_rhs-e-v-shp1-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Sch...pe=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)})
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'v', rhs_target = 'e'
msg = 'copy_rhs', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_rhs'
lhs_data = tensor([[[[[1.5327, 1.9880, 1.3726,  ..., 1.9888, 1.8943, 1.5037]],

          [[1.5743, 1.1922, 1.5087,  ..., 1.6350,...
          [[1.6179, 1.1439, 1.7235,  ..., 1.9702, 1.5892, 1.3810]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.0775, 1.3257, 1.6654,  ..., 1.3174, 1.2627, 1.9638],
           [1.0101, 1.9184, 1.3939,  ..., 1.5252, 1...
           [1.7335, 1.8750, 1.7117,  ..., 1.7456, 1.2397, 1.7010]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 2, 1, 3, 1), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3, 1, 1), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 5, 3, 1, 7]), rhs shape: torch.Size([100, 1, 3, 7, 7])
SDDMM(message func: v_copy_rhs_e)
___________________ test_sddmm[idtype1-copy_rhs-e-v-shp1-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((5, 3, 1, 7), (1, 3, 7, 7)), lhs_target = 'v', rhs_target = 'e'
msg = 'copy_rhs', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_rhs'
lhs_data = tensor([[[[[1.9374, 1.9616, 1.1744,  ..., 1.4169, 1.3948, 1.7673]],

          [[1.0227, 1.5676, 1.2092,  ..., 1.4458,...
          [[1.4234, 1.0770, 1.3745,  ..., 1.3638, 1.7544, 1.5819]]]]],
       dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[[1.1519, 1.3228, 1.3466,  ..., 1.6335, 1.1371, 1.0469],
           [1.1579, 1.0918, 1.7194,  ..., 1.3101, 1...
           [1.1575, 1.0498, 1.6135,  ..., 1.0010, 1.4324, 1.4437]]]]],
       dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([40, 5, 3, 1, 7]), rhs shape: torch.Size([300, 1, 3, 7, 7])
SDDMM(message func: v_copy_rhs_e)
___________________ test_sddmm[idtype1-copy_rhs-e-v-shp2-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme...dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)})
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'v', rhs_target = 'e'
msg = 'copy_rhs', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_rhs'
lhs_data = tensor([[[[1.8103, 1.4579, 1.2261],
          [1.7490, 1.7044, 1.5434],
          [1.8802, 1.2330, 1.7154]]],


      ...0],
          [1.2581, 1.0961, 1.5376],
          [1.1816, 1.5752, 1.9188]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.3789, 1.5061, 1.3105]],

         [[1.9299, 1.0755, 1.9331]],

         [[1.4375, 1.8366, 1.2557]],

    ...

         [[1.0134, 1.8976, 1.7609]],

         [[1.0335, 1.5703, 1.0897]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(5, 3, 1, 7), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1, 3, 7, 7), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1, 3, 3]), rhs shape: torch.Size([100, 4, 1, 3])
SDDMM(message func: v_copy_rhs_e)
___________________ test_sddmm[idtype1-copy_rhs-e-v-shp2-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1, 3, 3), (4, 1, 3)), lhs_target = 'v', rhs_target = 'e'
msg = 'copy_rhs', idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_rhs'
lhs_data = tensor([[[[1.8413, 1.4241, 1.1501],
          [1.3339, 1.9850, 1.1396],
          [1.3381, 1.3118, 1.9018]]],


      ...2],
          [1.1101, 1.6047, 1.3562],
          [1.6221, 1.4148, 1.9604]]]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[[[1.6795, 1.9798, 1.0570]],

         [[1.3804, 1.4328, 1.8370]],

         [[1.3010, 1.6446, 1.6242]],

    ...

         [[1.9671, 1.3054, 1.7572]],

         [[1.5429, 1.3686, 1.1859]]]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([40, 1, 3, 3]), rhs shape: torch.Size([300, 4, 1, 3])
SDDMM(message func: v_copy_rhs_e)
___________________ test_sddmm[idtype1-copy_rhs-e-v-shp3-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)})
shp = ((3,), (3,)), lhs_target = 'v', rhs_target = 'e', msg = 'copy_rhs'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_rhs'
lhs_data = tensor([[1.7812, 1.8957, 1.2217],
        [1.8328, 1.3363, 1.7436],
        [1.6988, 1.1592, 1.6486],
        [1.4804,... 1.5552],
        [1.9987, 1.9963, 1.4032],
        [1.5138, 1.9576, 1.9523]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.8989, 1.8609, 1.7037],
        [1.7855, 1.0936, 1.8359],
        [1.9846, 1.1933, 1.6315],
        [1.5431,... 1.0257],
        [1.1509, 1.9224, 1.1748],
        [1.7452, 1.2552, 1.5176]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1, 3, 3), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(4, 1, 3), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 3]), rhs shape: torch.Size([100, 3])
SDDMM(message func: v_copy_rhs_e)
___________________ test_sddmm[idtype1-copy_rhs-e-v-shp3-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((3,), (3,)), lhs_target = 'v', rhs_target = 'e', msg = 'copy_rhs'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_rhs'
lhs_data = tensor([[1.4051, 1.2752, 1.0507],
        [1.1362, 1.5415, 1.0611],
        [1.8419, 1.1748, 1.3601],
        [1.6928,... 1.9858],
        [1.0749, 1.7624, 1.3769],
        [1.2872, 1.7262, 1.1155]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.7797, 1.0596, 1.9202],
        [1.3271, 1.9990, 1.8196],
        [1.0188, 1.9803, 1.8149],
        [1.6953,... 1.1374],
        [1.6073, 1.3805, 1.9170],
        [1.5963, 1.8923, 1.0916]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([40, 3]), rhs shape: torch.Size([300, 3])
SDDMM(message func: v_copy_rhs_e)
___________________ test_sddmm[idtype1-copy_rhs-e-v-shp4-g0] ___________________

g = Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shap...=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)})
shp = ((1,), (1,)), lhs_target = 'v', rhs_target = 'e', msg = 'copy_rhs'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4798>
op = 'copy_rhs'
lhs_data = tensor([[1.5791],
        [1.3981],
        [1.5732],
        [1.9094],
        [1.7909],
        [1.1953],
        [1...656],
        [1.7553],
        [1.5537],
        [1.4973],
        [1.6416]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.8651],
        [1.5651],
        [1.2820],
        [1.6269],
        [1.1079],
        [1.3281],
        [1...865],
        [1.9662],
        [1.8719],
        [1.2008],
        [1.8400]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes=30, num_edges=100,
      ndata_schemes={'x': Scheme(shape=(3,), dtype=torch.float64), 'y': Scheme(shape=(1,), dtype=torch.float64)}
      edata_schemes={'w': Scheme(shape=(), dtype=torch.float64), 'x': Scheme(shape=(1,), dtype=torch.float64), 'y': Scheme(shape=(3,), dtype=torch.float64)})
torch.int64
lhs shape: torch.Size([30, 1]), rhs shape: torch.Size([100, 1])
SDDMM(message func: v_copy_rhs_e)
___________________ test_sddmm[idtype1-copy_rhs-e-v-shp4-g1] ___________________

g = Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
shp = ((1,), (1,)), lhs_target = 'v', rhs_target = 'e', msg = 'copy_rhs'
idtype = torch.int64

    @pytest.mark.parametrize('g', graphs)
    @pytest.mark.parametrize('shp', sddmm_shapes)
    @pytest.mark.parametrize('lhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('rhs_target', ['u', 'v', 'e'])
    @pytest.mark.parametrize('msg', ['add', 'sub', 'mul', 'div', 'dot', 'copy_lhs', 'copy_rhs'])
    @parametrize_idtype
    def test_sddmm(g, shp, lhs_target, rhs_target, msg, idtype):
        if lhs_target == rhs_target:
            return
        g = g.astype(idtype).to(F.ctx())
        if dgl.backend.backend_name == 'mxnet' and g.number_of_edges() == 0:
            pytest.skip()   # mxnet do not support zero shape tensor
        print(g)
        print(g.idtype)
    
        len_lhs = select(
            lhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        lhs_shp = (len_lhs,) + shp[0]
        len_rhs = select(
            rhs_target,
            g.number_of_src_nodes(),
            g.number_of_edges(),
            g.number_of_dst_nodes())
        rhs_shp = (len_rhs,) + shp[1]
        feat_lhs = F.tensor(np.random.rand(*lhs_shp) + 1)
        feat_rhs = F.tensor(np.random.rand(*rhs_shp) + 1)
        print('lhs shape: {}, rhs shape: {}'.format(F.shape(feat_lhs), F.shape(feat_rhs)))
    
        lhs_frame = select(
            lhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        rhs_frame = select(
            rhs_target,
            g.srcdata,
            g.edata,
            g.dstdata)
        lhs_frame['x'] = F.attach_grad(F.clone(feat_lhs))
        rhs_frame['y'] = F.attach_grad(F.clone(feat_rhs))
        msg_func = lhs_target + '_' + msg + '_' + rhs_target
        print('SDDMM(message func: {})'.format(msg_func))
    
        lhs = F.attach_grad(F.clone(feat_lhs))
        rhs = F.attach_grad(F.clone(feat_rhs))
        with F.record_grad():
>           e = gsddmm(g, msg, lhs, rhs, lhs_target=lhs_target, rhs_target=rhs_target)

tests/compute/test_sparse.py:207: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72acd4c60>
op = 'copy_rhs'
lhs_data = tensor([[1.3214],
        [1.7330],
        [1.2684],
        [1.6351],
        [1.7854],
        [1.0336],
        [1...414],
        [1.4728],
        [1.4876],
        [1.7324],
        [1.9385]], dtype=torch.float64, requires_grad=True)
rhs_data = tensor([[1.3953],
        [1.6293],
        [1.0898],
        [1.0729],
        [1.0595],
        [1.8069],
        [1...907],
        [1.4576],
        [1.2714],
        [1.5366],
        [1.3968]], dtype=torch.float64, requires_grad=True)
lhs_target = 'v', rhs_target = 'e'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
----------------------------- Captured stdout call -----------------------------
Graph(num_nodes={'_U': 30, '_V': 40},
      num_edges={('_U', '_E', '_V'): 300},
      metagraph=[('_U', '_V', '_E')])
torch.int64
lhs shape: torch.Size([40, 1]), rhs shape: torch.Size([300, 1])
SDDMM(message func: v_copy_rhs_e)
____________________ test_edge_softmax[idtype0-shp0-src-g0] ____________________

g = Graph(num_nodes=3, num_edges=9,
      ndata_schemes={}
      edata_schemes={})
norm_by = 'src', shp = (1,), idtype = torch.int32

    @pytest.mark.parametrize('g', get_cases(['clique']))
    @pytest.mark.parametrize('norm_by', ['src', 'dst'])
    @pytest.mark.parametrize('shp', edge_softmax_shapes)
    @parametrize_idtype
    def test_edge_softmax(g, norm_by, shp, idtype):
        g = g.astype(idtype).to(F.ctx())
        edata = F.tensor(np.random.rand(g.number_of_edges(), *shp))
        e1 = F.attach_grad(F.clone(edata))
    
        with F.record_grad():
>           score1 = edge_softmax(g, e1, norm_by=norm_by)

tests/compute/test_sparse.py:240: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/edge_softmax.py:135: in edge_softmax
    eids=eids, norm_by=norm_by)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72b93c8b8>
logits = tensor([[0.0753],
        [0.8055],
        [0.8982],
        [0.0914],
        [0.3260],
        [0.6156],
        [0.2587],
        [0.9070],
        [0.5179]], dtype=torch.float64, requires_grad=True)
eids = '__ALL__', norm_by = 'src'

    def edge_softmax(gidx, logits, eids=ALL, norm_by='dst'):
        args = _cast_if_autocast_enabled(gidx, logits, eids, norm_by)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:770: TypeError
____________________ test_edge_softmax[idtype0-shp0-dst-g0] ____________________

g = Graph(num_nodes=3, num_edges=9,
      ndata_schemes={}
      edata_schemes={})
norm_by = 'dst', shp = (1,), idtype = torch.int32

    @pytest.mark.parametrize('g', get_cases(['clique']))
    @pytest.mark.parametrize('norm_by', ['src', 'dst'])
    @pytest.mark.parametrize('shp', edge_softmax_shapes)
    @parametrize_idtype
    def test_edge_softmax(g, norm_by, shp, idtype):
        g = g.astype(idtype).to(F.ctx())
        edata = F.tensor(np.random.rand(g.number_of_edges(), *shp))
        e1 = F.attach_grad(F.clone(edata))
    
        with F.record_grad():
>           score1 = edge_softmax(g, e1, norm_by=norm_by)

tests/compute/test_sparse.py:240: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/edge_softmax.py:135: in edge_softmax
    eids=eids, norm_by=norm_by)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72b580798>
logits = tensor([[0.5736],
        [0.3749],
        [0.6673],
        [0.4789],
        [0.1116],
        [0.1633],
        [0.0682],
        [0.7571],
        [0.5223]], dtype=torch.float64, requires_grad=True)
eids = '__ALL__', norm_by = 'dst'

    def edge_softmax(gidx, logits, eids=ALL, norm_by='dst'):
        args = _cast_if_autocast_enabled(gidx, logits, eids, norm_by)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:770: TypeError
____________________ test_edge_softmax[idtype0-shp1-src-g0] ____________________

g = Graph(num_nodes=3, num_edges=9,
      ndata_schemes={}
      edata_schemes={})
norm_by = 'src', shp = (1, 3), idtype = torch.int32

    @pytest.mark.parametrize('g', get_cases(['clique']))
    @pytest.mark.parametrize('norm_by', ['src', 'dst'])
    @pytest.mark.parametrize('shp', edge_softmax_shapes)
    @parametrize_idtype
    def test_edge_softmax(g, norm_by, shp, idtype):
        g = g.astype(idtype).to(F.ctx())
        edata = F.tensor(np.random.rand(g.number_of_edges(), *shp))
        e1 = F.attach_grad(F.clone(edata))
    
        with F.record_grad():
>           score1 = edge_softmax(g, e1, norm_by=norm_by)

tests/compute/test_sparse.py:240: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/edge_softmax.py:135: in edge_softmax
    eids=eids, norm_by=norm_by)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb61f07ea68>
logits = tensor([[[0.6629, 0.3666, 0.2642]],

        [[0.6724, 0.8719, 0.7598]],

        [[0.7094, 0.3214, 0.3994]],

       ...]],

        [[0.5871, 0.8898, 0.1532]],

        [[0.2140, 0.0463, 0.5087]]], dtype=torch.float64, requires_grad=True)
eids = '__ALL__', norm_by = 'src'

    def edge_softmax(gidx, logits, eids=ALL, norm_by='dst'):
        args = _cast_if_autocast_enabled(gidx, logits, eids, norm_by)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:770: TypeError
____________________ test_edge_softmax[idtype0-shp1-dst-g0] ____________________

g = Graph(num_nodes=3, num_edges=9,
      ndata_schemes={}
      edata_schemes={})
norm_by = 'dst', shp = (1, 3), idtype = torch.int32

    @pytest.mark.parametrize('g', get_cases(['clique']))
    @pytest.mark.parametrize('norm_by', ['src', 'dst'])
    @pytest.mark.parametrize('shp', edge_softmax_shapes)
    @parametrize_idtype
    def test_edge_softmax(g, norm_by, shp, idtype):
        g = g.astype(idtype).to(F.ctx())
        edata = F.tensor(np.random.rand(g.number_of_edges(), *shp))
        e1 = F.attach_grad(F.clone(edata))
    
        with F.record_grad():
>           score1 = edge_softmax(g, e1, norm_by=norm_by)

tests/compute/test_sparse.py:240: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/edge_softmax.py:135: in edge_softmax
    eids=eids, norm_by=norm_by)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72b580c18>
logits = tensor([[[0.9109, 0.4017, 0.8143]],

        [[0.1241, 0.5949, 0.0951]],

        [[0.4755, 0.2490, 0.2331]],

       ...]],

        [[0.4047, 0.4939, 0.9638]],

        [[0.8521, 0.0064, 0.9972]]], dtype=torch.float64, requires_grad=True)
eids = '__ALL__', norm_by = 'dst'

    def edge_softmax(gidx, logits, eids=ALL, norm_by='dst'):
        args = _cast_if_autocast_enabled(gidx, logits, eids, norm_by)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:770: TypeError
____________________ test_edge_softmax[idtype0-shp2-src-g0] ____________________

g = Graph(num_nodes=3, num_edges=9,
      ndata_schemes={}
      edata_schemes={})
norm_by = 'src', shp = (3, 4, 5), idtype = torch.int32

    @pytest.mark.parametrize('g', get_cases(['clique']))
    @pytest.mark.parametrize('norm_by', ['src', 'dst'])
    @pytest.mark.parametrize('shp', edge_softmax_shapes)
    @parametrize_idtype
    def test_edge_softmax(g, norm_by, shp, idtype):
        g = g.astype(idtype).to(F.ctx())
        edata = F.tensor(np.random.rand(g.number_of_edges(), *shp))
        e1 = F.attach_grad(F.clone(edata))
    
        with F.record_grad():
>           score1 = edge_softmax(g, e1, norm_by=norm_by)

tests/compute/test_sparse.py:240: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/edge_softmax.py:135: in edge_softmax
    eids=eids, norm_by=norm_by)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb706151438>
logits = tensor([[[[2.3568e-01, 3.1214e-01, 6.2320e-01, 9.0590e-02, 1.1290e-01],
          [5.4334e-01, 5.5248e-01, 8.9387e-01,...       [2.0251e-01, 3.4976e-01, 1.0666e-01, 5.7245e-01, 8.2016e-01]]]],
       dtype=torch.float64, requires_grad=True)
eids = '__ALL__', norm_by = 'src'

    def edge_softmax(gidx, logits, eids=ALL, norm_by='dst'):
        args = _cast_if_autocast_enabled(gidx, logits, eids, norm_by)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:770: TypeError
____________________ test_edge_softmax[idtype0-shp2-dst-g0] ____________________

g = Graph(num_nodes=3, num_edges=9,
      ndata_schemes={}
      edata_schemes={})
norm_by = 'dst', shp = (3, 4, 5), idtype = torch.int32

    @pytest.mark.parametrize('g', get_cases(['clique']))
    @pytest.mark.parametrize('norm_by', ['src', 'dst'])
    @pytest.mark.parametrize('shp', edge_softmax_shapes)
    @parametrize_idtype
    def test_edge_softmax(g, norm_by, shp, idtype):
        g = g.astype(idtype).to(F.ctx())
        edata = F.tensor(np.random.rand(g.number_of_edges(), *shp))
        e1 = F.attach_grad(F.clone(edata))
    
        with F.record_grad():
>           score1 = edge_softmax(g, e1, norm_by=norm_by)

tests/compute/test_sparse.py:240: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/edge_softmax.py:135: in edge_softmax
    eids=eids, norm_by=norm_by)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72b545048>
logits = tensor([[[[0.8553, 0.1891, 0.1369, 0.2803, 0.0448],
          [0.4629, 0.1602, 0.5088, 0.5059, 0.6377],
          [0.8...0.6237, 0.5539],
          [0.8131, 0.7173, 0.0011, 0.5368, 0.5062]]]], dtype=torch.float64,
       requires_grad=True)
eids = '__ALL__', norm_by = 'dst'

    def edge_softmax(gidx, logits, eids=ALL, norm_by='dst'):
        args = _cast_if_autocast_enabled(gidx, logits, eids, norm_by)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:770: TypeError
____________________ test_edge_softmax[idtype1-shp0-src-g0] ____________________

g = Graph(num_nodes=3, num_edges=9,
      ndata_schemes={}
      edata_schemes={})
norm_by = 'src', shp = (1,), idtype = torch.int64

    @pytest.mark.parametrize('g', get_cases(['clique']))
    @pytest.mark.parametrize('norm_by', ['src', 'dst'])
    @pytest.mark.parametrize('shp', edge_softmax_shapes)
    @parametrize_idtype
    def test_edge_softmax(g, norm_by, shp, idtype):
        g = g.astype(idtype).to(F.ctx())
        edata = F.tensor(np.random.rand(g.number_of_edges(), *shp))
        e1 = F.attach_grad(F.clone(edata))
    
        with F.record_grad():
>           score1 = edge_softmax(g, e1, norm_by=norm_by)

tests/compute/test_sparse.py:240: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/edge_softmax.py:135: in edge_softmax
    eids=eids, norm_by=norm_by)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72b1b8948>
logits = tensor([[0.8406],
        [0.1745],
        [0.3345],
        [0.3646],
        [0.8309],
        [0.6845],
        [0.4655],
        [0.4576],
        [0.9292]], dtype=torch.float64, requires_grad=True)
eids = '__ALL__', norm_by = 'src'

    def edge_softmax(gidx, logits, eids=ALL, norm_by='dst'):
        args = _cast_if_autocast_enabled(gidx, logits, eids, norm_by)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:770: TypeError
____________________ test_edge_softmax[idtype1-shp0-dst-g0] ____________________

g = Graph(num_nodes=3, num_edges=9,
      ndata_schemes={}
      edata_schemes={})
norm_by = 'dst', shp = (1,), idtype = torch.int64

    @pytest.mark.parametrize('g', get_cases(['clique']))
    @pytest.mark.parametrize('norm_by', ['src', 'dst'])
    @pytest.mark.parametrize('shp', edge_softmax_shapes)
    @parametrize_idtype
    def test_edge_softmax(g, norm_by, shp, idtype):
        g = g.astype(idtype).to(F.ctx())
        edata = F.tensor(np.random.rand(g.number_of_edges(), *shp))
        e1 = F.attach_grad(F.clone(edata))
    
        with F.record_grad():
>           score1 = edge_softmax(g, e1, norm_by=norm_by)

tests/compute/test_sparse.py:240: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/edge_softmax.py:135: in edge_softmax
    eids=eids, norm_by=norm_by)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72b1b8948>
logits = tensor([[0.8463],
        [0.7901],
        [0.9339],
        [0.8135],
        [0.9437],
        [0.2803],
        [0.2323],
        [0.1038],
        [0.3317]], dtype=torch.float64, requires_grad=True)
eids = '__ALL__', norm_by = 'dst'

    def edge_softmax(gidx, logits, eids=ALL, norm_by='dst'):
        args = _cast_if_autocast_enabled(gidx, logits, eids, norm_by)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:770: TypeError
____________________ test_edge_softmax[idtype1-shp1-src-g0] ____________________

g = Graph(num_nodes=3, num_edges=9,
      ndata_schemes={}
      edata_schemes={})
norm_by = 'src', shp = (1, 3), idtype = torch.int64

    @pytest.mark.parametrize('g', get_cases(['clique']))
    @pytest.mark.parametrize('norm_by', ['src', 'dst'])
    @pytest.mark.parametrize('shp', edge_softmax_shapes)
    @parametrize_idtype
    def test_edge_softmax(g, norm_by, shp, idtype):
        g = g.astype(idtype).to(F.ctx())
        edata = F.tensor(np.random.rand(g.number_of_edges(), *shp))
        e1 = F.attach_grad(F.clone(edata))
    
        with F.record_grad():
>           score1 = edge_softmax(g, e1, norm_by=norm_by)

tests/compute/test_sparse.py:240: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/edge_softmax.py:135: in edge_softmax
    eids=eids, norm_by=norm_by)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72b1b8948>
logits = tensor([[[0.6063, 0.7308, 0.0134]],

        [[0.5140, 0.2919, 0.0165]],

        [[0.1430, 0.6988, 0.0289]],

       ...]],

        [[0.6512, 0.7367, 0.2710]],

        [[0.7152, 0.4692, 0.2998]]], dtype=torch.float64, requires_grad=True)
eids = '__ALL__', norm_by = 'src'

    def edge_softmax(gidx, logits, eids=ALL, norm_by='dst'):
        args = _cast_if_autocast_enabled(gidx, logits, eids, norm_by)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:770: TypeError
____________________ test_edge_softmax[idtype1-shp1-dst-g0] ____________________

g = Graph(num_nodes=3, num_edges=9,
      ndata_schemes={}
      edata_schemes={})
norm_by = 'dst', shp = (1, 3), idtype = torch.int64

    @pytest.mark.parametrize('g', get_cases(['clique']))
    @pytest.mark.parametrize('norm_by', ['src', 'dst'])
    @pytest.mark.parametrize('shp', edge_softmax_shapes)
    @parametrize_idtype
    def test_edge_softmax(g, norm_by, shp, idtype):
        g = g.astype(idtype).to(F.ctx())
        edata = F.tensor(np.random.rand(g.number_of_edges(), *shp))
        e1 = F.attach_grad(F.clone(edata))
    
        with F.record_grad():
>           score1 = edge_softmax(g, e1, norm_by=norm_by)

tests/compute/test_sparse.py:240: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/edge_softmax.py:135: in edge_softmax
    eids=eids, norm_by=norm_by)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72b1b8948>
logits = tensor([[[0.8018, 0.2317, 0.2448]],

        [[0.5826, 0.9925, 0.3086]],

        [[0.9497, 0.0042, 0.9720]],

       ...]],

        [[0.3193, 0.1606, 0.5287]],

        [[0.3239, 0.7866, 0.1637]]], dtype=torch.float64, requires_grad=True)
eids = '__ALL__', norm_by = 'dst'

    def edge_softmax(gidx, logits, eids=ALL, norm_by='dst'):
        args = _cast_if_autocast_enabled(gidx, logits, eids, norm_by)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:770: TypeError
____________________ test_edge_softmax[idtype1-shp2-src-g0] ____________________

g = Graph(num_nodes=3, num_edges=9,
      ndata_schemes={}
      edata_schemes={})
norm_by = 'src', shp = (3, 4, 5), idtype = torch.int64

    @pytest.mark.parametrize('g', get_cases(['clique']))
    @pytest.mark.parametrize('norm_by', ['src', 'dst'])
    @pytest.mark.parametrize('shp', edge_softmax_shapes)
    @parametrize_idtype
    def test_edge_softmax(g, norm_by, shp, idtype):
        g = g.astype(idtype).to(F.ctx())
        edata = F.tensor(np.random.rand(g.number_of_edges(), *shp))
        e1 = F.attach_grad(F.clone(edata))
    
        with F.record_grad():
>           score1 = edge_softmax(g, e1, norm_by=norm_by)

tests/compute/test_sparse.py:240: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/edge_softmax.py:135: in edge_softmax
    eids=eids, norm_by=norm_by)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72b1b8948>
logits = tensor([[[[0.7391, 0.7704, 0.6141, 0.9523, 0.4700],
          [0.7583, 0.1576, 0.5015, 0.5538, 0.9407],
          [0.2...0.3769, 0.2626],
          [0.6082, 0.1104, 0.3483, 0.0534, 0.1117]]]], dtype=torch.float64,
       requires_grad=True)
eids = '__ALL__', norm_by = 'src'

    def edge_softmax(gidx, logits, eids=ALL, norm_by='dst'):
        args = _cast_if_autocast_enabled(gidx, logits, eids, norm_by)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:770: TypeError
____________________ test_edge_softmax[idtype1-shp2-dst-g0] ____________________

g = Graph(num_nodes=3, num_edges=9,
      ndata_schemes={}
      edata_schemes={})
norm_by = 'dst', shp = (3, 4, 5), idtype = torch.int64

    @pytest.mark.parametrize('g', get_cases(['clique']))
    @pytest.mark.parametrize('norm_by', ['src', 'dst'])
    @pytest.mark.parametrize('shp', edge_softmax_shapes)
    @parametrize_idtype
    def test_edge_softmax(g, norm_by, shp, idtype):
        g = g.astype(idtype).to(F.ctx())
        edata = F.tensor(np.random.rand(g.number_of_edges(), *shp))
        e1 = F.attach_grad(F.clone(edata))
    
        with F.record_grad():
>           score1 = edge_softmax(g, e1, norm_by=norm_by)

tests/compute/test_sparse.py:240: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/edge_softmax.py:135: in edge_softmax
    eids=eids, norm_by=norm_by)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72b1b8948>
logits = tensor([[[[6.6720e-01, 7.0214e-01, 7.0333e-01, 1.2660e-01, 6.5713e-03],
          [1.4280e-01, 7.9931e-01, 7.0660e-01,...       [7.5633e-01, 4.3664e-01, 5.7141e-01, 4.3226e-01, 2.6665e-01]]]],
       dtype=torch.float64, requires_grad=True)
eids = '__ALL__', norm_by = 'dst'

    def edge_softmax(gidx, logits, eids=ALL, norm_by='dst'):
        args = _cast_if_autocast_enabled(gidx, logits, eids, norm_by)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:770: TypeError
___________________________ test_segment_reduce[sum] ___________________________

reducer = 'sum'

    @pytest.mark.parametrize('reducer', ['sum', 'max', 'min', 'mean'])
    def test_segment_reduce(reducer):
        ctx = F.ctx()
        value = F.tensor(np.random.rand(10, 5))
        v1 = F.attach_grad(F.clone(value))
        v2 = F.attach_grad(F.clone(value))
        seglen = F.tensor([2, 3, 0, 4, 1, 0, 0])
        u = F.copy_to(F.arange(0, F.shape(value)[0], F.int32), ctx)
        v = F.repeat(F.copy_to(F.arange(0, len(seglen), F.int32), ctx),
                     seglen, dim=0)
    
        num_nodes = {'_U': len(u), '_V': len(seglen)}
        g = dgl.convert.heterograph({('_U', '_E', '_V'): (u, v)}, num_nodes_dict=num_nodes)
        with F.record_grad():
>           rst1 = gspmm(g, 'copy_lhs', reducer, v1, None)

tests/compute/test_sparse.py:275: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43e7d6af8>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0.5684, 0.6808, 0.9339, 0.2656, 0.1624],
        [0.2807, 0.1330, 0.1766, 0.5875, 0.8666],
        [0.3529, 0...17, 0.7085, 0.0792],
        [0.4422, 0.3290, 0.8093, 0.7278, 0.1863]], dtype=torch.float64,
       requires_grad=True)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
___________________________ test_segment_reduce[max] ___________________________

reducer = 'max'

    @pytest.mark.parametrize('reducer', ['sum', 'max', 'min', 'mean'])
    def test_segment_reduce(reducer):
        ctx = F.ctx()
        value = F.tensor(np.random.rand(10, 5))
        v1 = F.attach_grad(F.clone(value))
        v2 = F.attach_grad(F.clone(value))
        seglen = F.tensor([2, 3, 0, 4, 1, 0, 0])
        u = F.copy_to(F.arange(0, F.shape(value)[0], F.int32), ctx)
        v = F.repeat(F.copy_to(F.arange(0, len(seglen), F.int32), ctx),
                     seglen, dim=0)
    
        num_nodes = {'_U': len(u), '_V': len(seglen)}
        g = dgl.convert.heterograph({('_U', '_E', '_V'): (u, v)}, num_nodes_dict=num_nodes)
        with F.record_grad():
>           rst1 = gspmm(g, 'copy_lhs', reducer, v1, None)

tests/compute/test_sparse.py:275: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72b54b1b0>
op = 'copy_lhs', reduce_op = 'max'
lhs_data = tensor([[0.3728, 0.7506, 0.0820, 0.9710, 0.0291],
        [0.6357, 0.9240, 0.7497, 0.3817, 0.5144],
        [0.1001, 0...81, 0.4114, 0.6433],
        [0.5135, 0.2694, 0.9685, 0.4559, 0.6366]], dtype=torch.float64,
       requires_grad=True)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
___________________________ test_segment_reduce[min] ___________________________

reducer = 'min'

    @pytest.mark.parametrize('reducer', ['sum', 'max', 'min', 'mean'])
    def test_segment_reduce(reducer):
        ctx = F.ctx()
        value = F.tensor(np.random.rand(10, 5))
        v1 = F.attach_grad(F.clone(value))
        v2 = F.attach_grad(F.clone(value))
        seglen = F.tensor([2, 3, 0, 4, 1, 0, 0])
        u = F.copy_to(F.arange(0, F.shape(value)[0], F.int32), ctx)
        v = F.repeat(F.copy_to(F.arange(0, len(seglen), F.int32), ctx),
                     seglen, dim=0)
    
        num_nodes = {'_U': len(u), '_V': len(seglen)}
        g = dgl.convert.heterograph({('_U', '_E', '_V'): (u, v)}, num_nodes_dict=num_nodes)
        with F.record_grad():
>           rst1 = gspmm(g, 'copy_lhs', reducer, v1, None)

tests/compute/test_sparse.py:275: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb705f937e0>
op = 'copy_lhs', reduce_op = 'min'
lhs_data = tensor([[0.4160, 0.4735, 0.8111, 0.7301, 0.0301],
        [0.7698, 0.4966, 0.0686, 0.6620, 0.5250],
        [0.6728, 0...60, 0.4410, 0.1940],
        [0.3650, 0.2974, 0.4912, 0.9652, 0.7168]], dtype=torch.float64,
       requires_grad=True)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
__________________________ test_segment_reduce[mean] ___________________________

reducer = 'mean'

    @pytest.mark.parametrize('reducer', ['sum', 'max', 'min', 'mean'])
    def test_segment_reduce(reducer):
        ctx = F.ctx()
        value = F.tensor(np.random.rand(10, 5))
        v1 = F.attach_grad(F.clone(value))
        v2 = F.attach_grad(F.clone(value))
        seglen = F.tensor([2, 3, 0, 4, 1, 0, 0])
        u = F.copy_to(F.arange(0, F.shape(value)[0], F.int32), ctx)
        v = F.repeat(F.copy_to(F.arange(0, len(seglen), F.int32), ctx),
                     seglen, dim=0)
    
        num_nodes = {'_U': len(u), '_V': len(seglen)}
        g = dgl.convert.heterograph({('_U', '_E', '_V'): (u, v)}, num_nodes_dict=num_nodes)
        with F.record_grad():
>           rst1 = gspmm(g, 'copy_lhs', reducer, v1, None)

tests/compute/test_sparse.py:275: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43e681e58>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[0.1120, 0.9807, 0.7658, 0.3284, 0.4239],
        [0.3536, 0.1679, 0.5856, 0.1316, 0.8034],
        [0.9506, 0...98, 0.4188, 0.7377],
        [0.5520, 0.7393, 0.2410, 0.9761, 0.4062]], dtype=torch.float64,
       requires_grad=True)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
___________________ test_segment_mm[dtype2-0.003-1-idtype0] ____________________

idtype = torch.int32, feat_size = 1, dtype = torch.float32, tol = 0.003

    @unittest.skipIf(dgl.backend.backend_name != 'pytorch', reason='Only support PyTorch for now')
    @parametrize_idtype
    @pytest.mark.parametrize('feat_size', [1, 8, 16, 64, 256])
    @pytest.mark.parametrize('dtype, tol', [(torch.float16, 1e-2), (torch.bfloat16, 1e-2),
        (torch.float32, 3e-3), (torch.float64, 1e-4)])
    def test_segment_mm(idtype, feat_size, dtype, tol):
        if F._default_context_str == 'cpu' and dtype in (torch.float16, torch.bfloat16):
            pytest.skip("Only support float32 and float64 on CPU.")
>       if LooseVersion(torch.version.cuda) < LooseVersion("11.0") \
            and dtype == torch.bfloat16:

tests/compute/test_sparse.py:299: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/setuptools/_distutils/version.py:72: in __lt__
    c = self._cmp(other)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

self = <[AttributeError("'LooseVersion' object has no attribute 'vstring'") raised in repr()] LooseVersion object at 0x7fb704909048>
other = LooseVersion ('11.0')

    def _cmp (self, other):
        if isinstance(other, str):
            other = LooseVersion(other)
        elif not isinstance(other, LooseVersion):
            return NotImplemented
    
>       if self.version == other.version:
E       AttributeError: 'LooseVersion' object has no attribute 'version'

/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/setuptools/_distutils/version.py:355: AttributeError
___________________ test_segment_mm[dtype2-0.003-1-idtype1] ____________________

idtype = torch.int64, feat_size = 1, dtype = torch.float32, tol = 0.003

    @unittest.skipIf(dgl.backend.backend_name != 'pytorch', reason='Only support PyTorch for now')
    @parametrize_idtype
    @pytest.mark.parametrize('feat_size', [1, 8, 16, 64, 256])
    @pytest.mark.parametrize('dtype, tol', [(torch.float16, 1e-2), (torch.bfloat16, 1e-2),
        (torch.float32, 3e-3), (torch.float64, 1e-4)])
    def test_segment_mm(idtype, feat_size, dtype, tol):
        if F._default_context_str == 'cpu' and dtype in (torch.float16, torch.bfloat16):
            pytest.skip("Only support float32 and float64 on CPU.")
>       if LooseVersion(torch.version.cuda) < LooseVersion("11.0") \
            and dtype == torch.bfloat16:

tests/compute/test_sparse.py:299: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/setuptools/_distutils/version.py:72: in __lt__
    c = self._cmp(other)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

self = <[AttributeError("'LooseVersion' object has no attribute 'vstring'") raised in repr()] LooseVersion object at 0x7fb457b40a20>
other = LooseVersion ('11.0')

    def _cmp (self, other):
        if isinstance(other, str):
            other = LooseVersion(other)
        elif not isinstance(other, LooseVersion):
            return NotImplemented
    
>       if self.version == other.version:
E       AttributeError: 'LooseVersion' object has no attribute 'version'

/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/setuptools/_distutils/version.py:355: AttributeError
___________________ test_segment_mm[dtype2-0.003-8-idtype0] ____________________

idtype = torch.int32, feat_size = 8, dtype = torch.float32, tol = 0.003

    @unittest.skipIf(dgl.backend.backend_name != 'pytorch', reason='Only support PyTorch for now')
    @parametrize_idtype
    @pytest.mark.parametrize('feat_size', [1, 8, 16, 64, 256])
    @pytest.mark.parametrize('dtype, tol', [(torch.float16, 1e-2), (torch.bfloat16, 1e-2),
        (torch.float32, 3e-3), (torch.float64, 1e-4)])
    def test_segment_mm(idtype, feat_size, dtype, tol):
        if F._default_context_str == 'cpu' and dtype in (torch.float16, torch.bfloat16):
            pytest.skip("Only support float32 and float64 on CPU.")
>       if LooseVersion(torch.version.cuda) < LooseVersion("11.0") \
            and dtype == torch.bfloat16:

tests/compute/test_sparse.py:299: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/setuptools/_distutils/version.py:72: in __lt__
    c = self._cmp(other)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

self = <[AttributeError("'LooseVersion' object has no attribute 'vstring'") raised in repr()] LooseVersion object at 0x7fb43e431278>
other = LooseVersion ('11.0')

    def _cmp (self, other):
        if isinstance(other, str):
            other = LooseVersion(other)
        elif not isinstance(other, LooseVersion):
            return NotImplemented
    
>       if self.version == other.version:
E       AttributeError: 'LooseVersion' object has no attribute 'version'

/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/setuptools/_distutils/version.py:355: AttributeError
___________________ test_segment_mm[dtype2-0.003-8-idtype1] ____________________

idtype = torch.int64, feat_size = 8, dtype = torch.float32, tol = 0.003

    @unittest.skipIf(dgl.backend.backend_name != 'pytorch', reason='Only support PyTorch for now')
    @parametrize_idtype
    @pytest.mark.parametrize('feat_size', [1, 8, 16, 64, 256])
    @pytest.mark.parametrize('dtype, tol', [(torch.float16, 1e-2), (torch.bfloat16, 1e-2),
        (torch.float32, 3e-3), (torch.float64, 1e-4)])
    def test_segment_mm(idtype, feat_size, dtype, tol):
        if F._default_context_str == 'cpu' and dtype in (torch.float16, torch.bfloat16):
            pytest.skip("Only support float32 and float64 on CPU.")
>       if LooseVersion(torch.version.cuda) < LooseVersion("11.0") \
            and dtype == torch.bfloat16:

tests/compute/test_sparse.py:299: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/setuptools/_distutils/version.py:72: in __lt__
    c = self._cmp(other)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

self = <[AttributeError("'LooseVersion' object has no attribute 'vstring'") raised in repr()] LooseVersion object at 0x7fb43e6664e0>
other = LooseVersion ('11.0')

    def _cmp (self, other):
        if isinstance(other, str):
            other = LooseVersion(other)
        elif not isinstance(other, LooseVersion):
            return NotImplemented
    
>       if self.version == other.version:
E       AttributeError: 'LooseVersion' object has no attribute 'version'

/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/setuptools/_distutils/version.py:355: AttributeError
___________________ test_segment_mm[dtype2-0.003-16-idtype0] ___________________

idtype = torch.int32, feat_size = 16, dtype = torch.float32, tol = 0.003

    @unittest.skipIf(dgl.backend.backend_name != 'pytorch', reason='Only support PyTorch for now')
    @parametrize_idtype
    @pytest.mark.parametrize('feat_size', [1, 8, 16, 64, 256])
    @pytest.mark.parametrize('dtype, tol', [(torch.float16, 1e-2), (torch.bfloat16, 1e-2),
        (torch.float32, 3e-3), (torch.float64, 1e-4)])
    def test_segment_mm(idtype, feat_size, dtype, tol):
        if F._default_context_str == 'cpu' and dtype in (torch.float16, torch.bfloat16):
            pytest.skip("Only support float32 and float64 on CPU.")
>       if LooseVersion(torch.version.cuda) < LooseVersion("11.0") \
            and dtype == torch.bfloat16:

tests/compute/test_sparse.py:299: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/setuptools/_distutils/version.py:72: in __lt__
    c = self._cmp(other)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

self = <[AttributeError("'LooseVersion' object has no attribute 'vstring'") raised in repr()] LooseVersion object at 0x7fb72ad11748>
other = LooseVersion ('11.0')

    def _cmp (self, other):
        if isinstance(other, str):
            other = LooseVersion(other)
        elif not isinstance(other, LooseVersion):
            return NotImplemented
    
>       if self.version == other.version:
E       AttributeError: 'LooseVersion' object has no attribute 'version'

/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/setuptools/_distutils/version.py:355: AttributeError
___________________ test_segment_mm[dtype2-0.003-16-idtype1] ___________________

idtype = torch.int64, feat_size = 16, dtype = torch.float32, tol = 0.003

    @unittest.skipIf(dgl.backend.backend_name != 'pytorch', reason='Only support PyTorch for now')
    @parametrize_idtype
    @pytest.mark.parametrize('feat_size', [1, 8, 16, 64, 256])
    @pytest.mark.parametrize('dtype, tol', [(torch.float16, 1e-2), (torch.bfloat16, 1e-2),
        (torch.float32, 3e-3), (torch.float64, 1e-4)])
    def test_segment_mm(idtype, feat_size, dtype, tol):
        if F._default_context_str == 'cpu' and dtype in (torch.float16, torch.bfloat16):
            pytest.skip("Only support float32 and float64 on CPU.")
>       if LooseVersion(torch.version.cuda) < LooseVersion("11.0") \
            and dtype == torch.bfloat16:

tests/compute/test_sparse.py:299: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/setuptools/_distutils/version.py:72: in __lt__
    c = self._cmp(other)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

self = <[AttributeError("'LooseVersion' object has no attribute 'vstring'") raised in repr()] LooseVersion object at 0x7fb6152cd6a0>
other = LooseVersion ('11.0')

    def _cmp (self, other):
        if isinstance(other, str):
            other = LooseVersion(other)
        elif not isinstance(other, LooseVersion):
            return NotImplemented
    
>       if self.version == other.version:
E       AttributeError: 'LooseVersion' object has no attribute 'version'

/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/setuptools/_distutils/version.py:355: AttributeError
___________________ test_segment_mm[dtype2-0.003-64-idtype0] ___________________

idtype = torch.int32, feat_size = 64, dtype = torch.float32, tol = 0.003

    @unittest.skipIf(dgl.backend.backend_name != 'pytorch', reason='Only support PyTorch for now')
    @parametrize_idtype
    @pytest.mark.parametrize('feat_size', [1, 8, 16, 64, 256])
    @pytest.mark.parametrize('dtype, tol', [(torch.float16, 1e-2), (torch.bfloat16, 1e-2),
        (torch.float32, 3e-3), (torch.float64, 1e-4)])
    def test_segment_mm(idtype, feat_size, dtype, tol):
        if F._default_context_str == 'cpu' and dtype in (torch.float16, torch.bfloat16):
            pytest.skip("Only support float32 and float64 on CPU.")
>       if LooseVersion(torch.version.cuda) < LooseVersion("11.0") \
            and dtype == torch.bfloat16:

tests/compute/test_sparse.py:299: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/setuptools/_distutils/version.py:72: in __lt__
    c = self._cmp(other)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

self = <[AttributeError("'LooseVersion' object has no attribute 'vstring'") raised in repr()] LooseVersion object at 0x7fb72b848fd0>
other = LooseVersion ('11.0')

    def _cmp (self, other):
        if isinstance(other, str):
            other = LooseVersion(other)
        elif not isinstance(other, LooseVersion):
            return NotImplemented
    
>       if self.version == other.version:
E       AttributeError: 'LooseVersion' object has no attribute 'version'

/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/setuptools/_distutils/version.py:355: AttributeError
___________________ test_segment_mm[dtype2-0.003-64-idtype1] ___________________

idtype = torch.int64, feat_size = 64, dtype = torch.float32, tol = 0.003

    @unittest.skipIf(dgl.backend.backend_name != 'pytorch', reason='Only support PyTorch for now')
    @parametrize_idtype
    @pytest.mark.parametrize('feat_size', [1, 8, 16, 64, 256])
    @pytest.mark.parametrize('dtype, tol', [(torch.float16, 1e-2), (torch.bfloat16, 1e-2),
        (torch.float32, 3e-3), (torch.float64, 1e-4)])
    def test_segment_mm(idtype, feat_size, dtype, tol):
        if F._default_context_str == 'cpu' and dtype in (torch.float16, torch.bfloat16):
            pytest.skip("Only support float32 and float64 on CPU.")
>       if LooseVersion(torch.version.cuda) < LooseVersion("11.0") \
            and dtype == torch.bfloat16:

tests/compute/test_sparse.py:299: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/setuptools/_distutils/version.py:72: in __lt__
    c = self._cmp(other)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

self = <[AttributeError("'LooseVersion' object has no attribute 'vstring'") raised in repr()] LooseVersion object at 0x7fb7060e8ba8>
other = LooseVersion ('11.0')

    def _cmp (self, other):
        if isinstance(other, str):
            other = LooseVersion(other)
        elif not isinstance(other, LooseVersion):
            return NotImplemented
    
>       if self.version == other.version:
E       AttributeError: 'LooseVersion' object has no attribute 'version'

/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/setuptools/_distutils/version.py:355: AttributeError
__________________ test_segment_mm[dtype2-0.003-256-idtype0] ___________________

idtype = torch.int32, feat_size = 256, dtype = torch.float32, tol = 0.003

    @unittest.skipIf(dgl.backend.backend_name != 'pytorch', reason='Only support PyTorch for now')
    @parametrize_idtype
    @pytest.mark.parametrize('feat_size', [1, 8, 16, 64, 256])
    @pytest.mark.parametrize('dtype, tol', [(torch.float16, 1e-2), (torch.bfloat16, 1e-2),
        (torch.float32, 3e-3), (torch.float64, 1e-4)])
    def test_segment_mm(idtype, feat_size, dtype, tol):
        if F._default_context_str == 'cpu' and dtype in (torch.float16, torch.bfloat16):
            pytest.skip("Only support float32 and float64 on CPU.")
>       if LooseVersion(torch.version.cuda) < LooseVersion("11.0") \
            and dtype == torch.bfloat16:

tests/compute/test_sparse.py:299: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/setuptools/_distutils/version.py:72: in __lt__
    c = self._cmp(other)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

self = <[AttributeError("'LooseVersion' object has no attribute 'vstring'") raised in repr()] LooseVersion object at 0x7fb43e7eaba8>
other = LooseVersion ('11.0')

    def _cmp (self, other):
        if isinstance(other, str):
            other = LooseVersion(other)
        elif not isinstance(other, LooseVersion):
            return NotImplemented
    
>       if self.version == other.version:
E       AttributeError: 'LooseVersion' object has no attribute 'version'

/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/setuptools/_distutils/version.py:355: AttributeError
__________________ test_segment_mm[dtype2-0.003-256-idtype1] ___________________

idtype = torch.int64, feat_size = 256, dtype = torch.float32, tol = 0.003

    @unittest.skipIf(dgl.backend.backend_name != 'pytorch', reason='Only support PyTorch for now')
    @parametrize_idtype
    @pytest.mark.parametrize('feat_size', [1, 8, 16, 64, 256])
    @pytest.mark.parametrize('dtype, tol', [(torch.float16, 1e-2), (torch.bfloat16, 1e-2),
        (torch.float32, 3e-3), (torch.float64, 1e-4)])
    def test_segment_mm(idtype, feat_size, dtype, tol):
        if F._default_context_str == 'cpu' and dtype in (torch.float16, torch.bfloat16):
            pytest.skip("Only support float32 and float64 on CPU.")
>       if LooseVersion(torch.version.cuda) < LooseVersion("11.0") \
            and dtype == torch.bfloat16:

tests/compute/test_sparse.py:299: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/setuptools/_distutils/version.py:72: in __lt__
    c = self._cmp(other)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

self = <[AttributeError("'LooseVersion' object has no attribute 'vstring'") raised in repr()] LooseVersion object at 0x7fb72b90a8d0>
other = LooseVersion ('11.0')

    def _cmp (self, other):
        if isinstance(other, str):
            other = LooseVersion(other)
        elif not isinstance(other, LooseVersion):
            return NotImplemented
    
>       if self.version == other.version:
E       AttributeError: 'LooseVersion' object has no attribute 'version'

/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/setuptools/_distutils/version.py:355: AttributeError
___________________ test_segment_mm[dtype3-0.0001-1-idtype0] ___________________

idtype = torch.int32, feat_size = 1, dtype = torch.float64, tol = 0.0001

    @unittest.skipIf(dgl.backend.backend_name != 'pytorch', reason='Only support PyTorch for now')
    @parametrize_idtype
    @pytest.mark.parametrize('feat_size', [1, 8, 16, 64, 256])
    @pytest.mark.parametrize('dtype, tol', [(torch.float16, 1e-2), (torch.bfloat16, 1e-2),
        (torch.float32, 3e-3), (torch.float64, 1e-4)])
    def test_segment_mm(idtype, feat_size, dtype, tol):
        if F._default_context_str == 'cpu' and dtype in (torch.float16, torch.bfloat16):
            pytest.skip("Only support float32 and float64 on CPU.")
>       if LooseVersion(torch.version.cuda) < LooseVersion("11.0") \
            and dtype == torch.bfloat16:

tests/compute/test_sparse.py:299: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/setuptools/_distutils/version.py:72: in __lt__
    c = self._cmp(other)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

self = <[AttributeError("'LooseVersion' object has no attribute 'vstring'") raised in repr()] LooseVersion object at 0x7fb43e683438>
other = LooseVersion ('11.0')

    def _cmp (self, other):
        if isinstance(other, str):
            other = LooseVersion(other)
        elif not isinstance(other, LooseVersion):
            return NotImplemented
    
>       if self.version == other.version:
E       AttributeError: 'LooseVersion' object has no attribute 'version'

/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/setuptools/_distutils/version.py:355: AttributeError
___________________ test_segment_mm[dtype3-0.0001-1-idtype1] ___________________

idtype = torch.int64, feat_size = 1, dtype = torch.float64, tol = 0.0001

    @unittest.skipIf(dgl.backend.backend_name != 'pytorch', reason='Only support PyTorch for now')
    @parametrize_idtype
    @pytest.mark.parametrize('feat_size', [1, 8, 16, 64, 256])
    @pytest.mark.parametrize('dtype, tol', [(torch.float16, 1e-2), (torch.bfloat16, 1e-2),
        (torch.float32, 3e-3), (torch.float64, 1e-4)])
    def test_segment_mm(idtype, feat_size, dtype, tol):
        if F._default_context_str == 'cpu' and dtype in (torch.float16, torch.bfloat16):
            pytest.skip("Only support float32 and float64 on CPU.")
>       if LooseVersion(torch.version.cuda) < LooseVersion("11.0") \
            and dtype == torch.bfloat16:

tests/compute/test_sparse.py:299: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/setuptools/_distutils/version.py:72: in __lt__
    c = self._cmp(other)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

self = <[AttributeError("'LooseVersion' object has no attribute 'vstring'") raised in repr()] LooseVersion object at 0x7fb7297275f8>
other = LooseVersion ('11.0')

    def _cmp (self, other):
        if isinstance(other, str):
            other = LooseVersion(other)
        elif not isinstance(other, LooseVersion):
            return NotImplemented
    
>       if self.version == other.version:
E       AttributeError: 'LooseVersion' object has no attribute 'version'

/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/setuptools/_distutils/version.py:355: AttributeError
___________________ test_segment_mm[dtype3-0.0001-8-idtype0] ___________________

idtype = torch.int32, feat_size = 8, dtype = torch.float64, tol = 0.0001

    @unittest.skipIf(dgl.backend.backend_name != 'pytorch', reason='Only support PyTorch for now')
    @parametrize_idtype
    @pytest.mark.parametrize('feat_size', [1, 8, 16, 64, 256])
    @pytest.mark.parametrize('dtype, tol', [(torch.float16, 1e-2), (torch.bfloat16, 1e-2),
        (torch.float32, 3e-3), (torch.float64, 1e-4)])
    def test_segment_mm(idtype, feat_size, dtype, tol):
        if F._default_context_str == 'cpu' and dtype in (torch.float16, torch.bfloat16):
            pytest.skip("Only support float32 and float64 on CPU.")
>       if LooseVersion(torch.version.cuda) < LooseVersion("11.0") \
            and dtype == torch.bfloat16:

tests/compute/test_sparse.py:299: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/setuptools/_distutils/version.py:72: in __lt__
    c = self._cmp(other)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

self = <[AttributeError("'LooseVersion' object has no attribute 'vstring'") raised in repr()] LooseVersion object at 0x7fb61fa534e0>
other = LooseVersion ('11.0')

    def _cmp (self, other):
        if isinstance(other, str):
            other = LooseVersion(other)
        elif not isinstance(other, LooseVersion):
            return NotImplemented
    
>       if self.version == other.version:
E       AttributeError: 'LooseVersion' object has no attribute 'version'

/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/setuptools/_distutils/version.py:355: AttributeError
___________________ test_segment_mm[dtype3-0.0001-8-idtype1] ___________________

idtype = torch.int64, feat_size = 8, dtype = torch.float64, tol = 0.0001

    @unittest.skipIf(dgl.backend.backend_name != 'pytorch', reason='Only support PyTorch for now')
    @parametrize_idtype
    @pytest.mark.parametrize('feat_size', [1, 8, 16, 64, 256])
    @pytest.mark.parametrize('dtype, tol', [(torch.float16, 1e-2), (torch.bfloat16, 1e-2),
        (torch.float32, 3e-3), (torch.float64, 1e-4)])
    def test_segment_mm(idtype, feat_size, dtype, tol):
        if F._default_context_str == 'cpu' and dtype in (torch.float16, torch.bfloat16):
            pytest.skip("Only support float32 and float64 on CPU.")
>       if LooseVersion(torch.version.cuda) < LooseVersion("11.0") \
            and dtype == torch.bfloat16:

tests/compute/test_sparse.py:299: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/setuptools/_distutils/version.py:72: in __lt__
    c = self._cmp(other)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

self = <[AttributeError("'LooseVersion' object has no attribute 'vstring'") raised in repr()] LooseVersion object at 0x7fb706123e48>
other = LooseVersion ('11.0')

    def _cmp (self, other):
        if isinstance(other, str):
            other = LooseVersion(other)
        elif not isinstance(other, LooseVersion):
            return NotImplemented
    
>       if self.version == other.version:
E       AttributeError: 'LooseVersion' object has no attribute 'version'

/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/setuptools/_distutils/version.py:355: AttributeError
__________________ test_segment_mm[dtype3-0.0001-16-idtype0] ___________________

idtype = torch.int32, feat_size = 16, dtype = torch.float64, tol = 0.0001

    @unittest.skipIf(dgl.backend.backend_name != 'pytorch', reason='Only support PyTorch for now')
    @parametrize_idtype
    @pytest.mark.parametrize('feat_size', [1, 8, 16, 64, 256])
    @pytest.mark.parametrize('dtype, tol', [(torch.float16, 1e-2), (torch.bfloat16, 1e-2),
        (torch.float32, 3e-3), (torch.float64, 1e-4)])
    def test_segment_mm(idtype, feat_size, dtype, tol):
        if F._default_context_str == 'cpu' and dtype in (torch.float16, torch.bfloat16):
            pytest.skip("Only support float32 and float64 on CPU.")
>       if LooseVersion(torch.version.cuda) < LooseVersion("11.0") \
            and dtype == torch.bfloat16:

tests/compute/test_sparse.py:299: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/setuptools/_distutils/version.py:72: in __lt__
    c = self._cmp(other)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

self = <[AttributeError("'LooseVersion' object has no attribute 'vstring'") raised in repr()] LooseVersion object at 0x7fb43e7ea588>
other = LooseVersion ('11.0')

    def _cmp (self, other):
        if isinstance(other, str):
            other = LooseVersion(other)
        elif not isinstance(other, LooseVersion):
            return NotImplemented
    
>       if self.version == other.version:
E       AttributeError: 'LooseVersion' object has no attribute 'version'

/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/setuptools/_distutils/version.py:355: AttributeError
__________________ test_segment_mm[dtype3-0.0001-16-idtype1] ___________________

idtype = torch.int64, feat_size = 16, dtype = torch.float64, tol = 0.0001

    @unittest.skipIf(dgl.backend.backend_name != 'pytorch', reason='Only support PyTorch for now')
    @parametrize_idtype
    @pytest.mark.parametrize('feat_size', [1, 8, 16, 64, 256])
    @pytest.mark.parametrize('dtype, tol', [(torch.float16, 1e-2), (torch.bfloat16, 1e-2),
        (torch.float32, 3e-3), (torch.float64, 1e-4)])
    def test_segment_mm(idtype, feat_size, dtype, tol):
        if F._default_context_str == 'cpu' and dtype in (torch.float16, torch.bfloat16):
            pytest.skip("Only support float32 and float64 on CPU.")
>       if LooseVersion(torch.version.cuda) < LooseVersion("11.0") \
            and dtype == torch.bfloat16:

tests/compute/test_sparse.py:299: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/setuptools/_distutils/version.py:72: in __lt__
    c = self._cmp(other)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

self = <[AttributeError("'LooseVersion' object has no attribute 'vstring'") raised in repr()] LooseVersion object at 0x7fb72a9bfbe0>
other = LooseVersion ('11.0')

    def _cmp (self, other):
        if isinstance(other, str):
            other = LooseVersion(other)
        elif not isinstance(other, LooseVersion):
            return NotImplemented
    
>       if self.version == other.version:
E       AttributeError: 'LooseVersion' object has no attribute 'version'

/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/setuptools/_distutils/version.py:355: AttributeError
__________________ test_segment_mm[dtype3-0.0001-64-idtype0] ___________________

idtype = torch.int32, feat_size = 64, dtype = torch.float64, tol = 0.0001

    @unittest.skipIf(dgl.backend.backend_name != 'pytorch', reason='Only support PyTorch for now')
    @parametrize_idtype
    @pytest.mark.parametrize('feat_size', [1, 8, 16, 64, 256])
    @pytest.mark.parametrize('dtype, tol', [(torch.float16, 1e-2), (torch.bfloat16, 1e-2),
        (torch.float32, 3e-3), (torch.float64, 1e-4)])
    def test_segment_mm(idtype, feat_size, dtype, tol):
        if F._default_context_str == 'cpu' and dtype in (torch.float16, torch.bfloat16):
            pytest.skip("Only support float32 and float64 on CPU.")
>       if LooseVersion(torch.version.cuda) < LooseVersion("11.0") \
            and dtype == torch.bfloat16:

tests/compute/test_sparse.py:299: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/setuptools/_distutils/version.py:72: in __lt__
    c = self._cmp(other)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

self = <[AttributeError("'LooseVersion' object has no attribute 'vstring'") raised in repr()] LooseVersion object at 0x7fb6146d2780>
other = LooseVersion ('11.0')

    def _cmp (self, other):
        if isinstance(other, str):
            other = LooseVersion(other)
        elif not isinstance(other, LooseVersion):
            return NotImplemented
    
>       if self.version == other.version:
E       AttributeError: 'LooseVersion' object has no attribute 'version'

/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/setuptools/_distutils/version.py:355: AttributeError
__________________ test_segment_mm[dtype3-0.0001-64-idtype1] ___________________

idtype = torch.int64, feat_size = 64, dtype = torch.float64, tol = 0.0001

    @unittest.skipIf(dgl.backend.backend_name != 'pytorch', reason='Only support PyTorch for now')
    @parametrize_idtype
    @pytest.mark.parametrize('feat_size', [1, 8, 16, 64, 256])
    @pytest.mark.parametrize('dtype, tol', [(torch.float16, 1e-2), (torch.bfloat16, 1e-2),
        (torch.float32, 3e-3), (torch.float64, 1e-4)])
    def test_segment_mm(idtype, feat_size, dtype, tol):
        if F._default_context_str == 'cpu' and dtype in (torch.float16, torch.bfloat16):
            pytest.skip("Only support float32 and float64 on CPU.")
>       if LooseVersion(torch.version.cuda) < LooseVersion("11.0") \
            and dtype == torch.bfloat16:

tests/compute/test_sparse.py:299: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/setuptools/_distutils/version.py:72: in __lt__
    c = self._cmp(other)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

self = <[AttributeError("'LooseVersion' object has no attribute 'vstring'") raised in repr()] LooseVersion object at 0x7fb43eaf64a8>
other = LooseVersion ('11.0')

    def _cmp (self, other):
        if isinstance(other, str):
            other = LooseVersion(other)
        elif not isinstance(other, LooseVersion):
            return NotImplemented
    
>       if self.version == other.version:
E       AttributeError: 'LooseVersion' object has no attribute 'version'

/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/setuptools/_distutils/version.py:355: AttributeError
__________________ test_segment_mm[dtype3-0.0001-256-idtype0] __________________

idtype = torch.int32, feat_size = 256, dtype = torch.float64, tol = 0.0001

    @unittest.skipIf(dgl.backend.backend_name != 'pytorch', reason='Only support PyTorch for now')
    @parametrize_idtype
    @pytest.mark.parametrize('feat_size', [1, 8, 16, 64, 256])
    @pytest.mark.parametrize('dtype, tol', [(torch.float16, 1e-2), (torch.bfloat16, 1e-2),
        (torch.float32, 3e-3), (torch.float64, 1e-4)])
    def test_segment_mm(idtype, feat_size, dtype, tol):
        if F._default_context_str == 'cpu' and dtype in (torch.float16, torch.bfloat16):
            pytest.skip("Only support float32 and float64 on CPU.")
>       if LooseVersion(torch.version.cuda) < LooseVersion("11.0") \
            and dtype == torch.bfloat16:

tests/compute/test_sparse.py:299: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/setuptools/_distutils/version.py:72: in __lt__
    c = self._cmp(other)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

self = <[AttributeError("'LooseVersion' object has no attribute 'vstring'") raised in repr()] LooseVersion object at 0x7fb72b7fccc0>
other = LooseVersion ('11.0')

    def _cmp (self, other):
        if isinstance(other, str):
            other = LooseVersion(other)
        elif not isinstance(other, LooseVersion):
            return NotImplemented
    
>       if self.version == other.version:
E       AttributeError: 'LooseVersion' object has no attribute 'version'

/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/setuptools/_distutils/version.py:355: AttributeError
__________________ test_segment_mm[dtype3-0.0001-256-idtype1] __________________

idtype = torch.int64, feat_size = 256, dtype = torch.float64, tol = 0.0001

    @unittest.skipIf(dgl.backend.backend_name != 'pytorch', reason='Only support PyTorch for now')
    @parametrize_idtype
    @pytest.mark.parametrize('feat_size', [1, 8, 16, 64, 256])
    @pytest.mark.parametrize('dtype, tol', [(torch.float16, 1e-2), (torch.bfloat16, 1e-2),
        (torch.float32, 3e-3), (torch.float64, 1e-4)])
    def test_segment_mm(idtype, feat_size, dtype, tol):
        if F._default_context_str == 'cpu' and dtype in (torch.float16, torch.bfloat16):
            pytest.skip("Only support float32 and float64 on CPU.")
>       if LooseVersion(torch.version.cuda) < LooseVersion("11.0") \
            and dtype == torch.bfloat16:

tests/compute/test_sparse.py:299: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/setuptools/_distutils/version.py:72: in __lt__
    c = self._cmp(other)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

self = <[AttributeError("'LooseVersion' object has no attribute 'vstring'") raised in repr()] LooseVersion object at 0x7fb705d057b8>
other = LooseVersion ('11.0')

    def _cmp (self, other):
        if isinstance(other, str):
            other = LooseVersion(other)
        elif not isinstance(other, LooseVersion):
            return NotImplemented
    
>       if self.version == other.version:
E       AttributeError: 'LooseVersion' object has no attribute 'version'

/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/setuptools/_distutils/version.py:355: AttributeError
_____________________ test_gather_mm_idx_b[dtype2-0.003-1] _____________________

feat_size = 1, dtype = torch.float32, tol = 0.003

    @unittest.skipIf(dgl.backend.backend_name != 'pytorch', reason='Only support PyTorch for now')
    @pytest.mark.parametrize('feat_size', [1, 8, 16, 64, 256])
    @pytest.mark.parametrize('dtype, tol', [(torch.float16, 1e-2), (torch.bfloat16, 2e-2),
        (torch.float32, 3e-3), (torch.float64, 1e-4)])
    def test_gather_mm_idx_b(feat_size, dtype, tol):
        if F._default_context_str == 'cpu' and dtype in (torch.float16, torch.bfloat16):
            pytest.skip("Only support float32 and float64 on CPU.")
>       if LooseVersion(torch.version.cuda) < LooseVersion("11.0") \
            and dtype == torch.bfloat16:

tests/compute/test_sparse.py:339: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/setuptools/_distutils/version.py:72: in __lt__
    c = self._cmp(other)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

self = <[AttributeError("'LooseVersion' object has no attribute 'vstring'") raised in repr()] LooseVersion object at 0x7fb7291ef4e0>
other = LooseVersion ('11.0')

    def _cmp (self, other):
        if isinstance(other, str):
            other = LooseVersion(other)
        elif not isinstance(other, LooseVersion):
            return NotImplemented
    
>       if self.version == other.version:
E       AttributeError: 'LooseVersion' object has no attribute 'version'

/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/setuptools/_distutils/version.py:355: AttributeError
_____________________ test_gather_mm_idx_b[dtype2-0.003-8] _____________________

feat_size = 8, dtype = torch.float32, tol = 0.003

    @unittest.skipIf(dgl.backend.backend_name != 'pytorch', reason='Only support PyTorch for now')
    @pytest.mark.parametrize('feat_size', [1, 8, 16, 64, 256])
    @pytest.mark.parametrize('dtype, tol', [(torch.float16, 1e-2), (torch.bfloat16, 2e-2),
        (torch.float32, 3e-3), (torch.float64, 1e-4)])
    def test_gather_mm_idx_b(feat_size, dtype, tol):
        if F._default_context_str == 'cpu' and dtype in (torch.float16, torch.bfloat16):
            pytest.skip("Only support float32 and float64 on CPU.")
>       if LooseVersion(torch.version.cuda) < LooseVersion("11.0") \
            and dtype == torch.bfloat16:

tests/compute/test_sparse.py:339: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/setuptools/_distutils/version.py:72: in __lt__
    c = self._cmp(other)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

self = <[AttributeError("'LooseVersion' object has no attribute 'vstring'") raised in repr()] LooseVersion object at 0x7fb43e422f98>
other = LooseVersion ('11.0')

    def _cmp (self, other):
        if isinstance(other, str):
            other = LooseVersion(other)
        elif not isinstance(other, LooseVersion):
            return NotImplemented
    
>       if self.version == other.version:
E       AttributeError: 'LooseVersion' object has no attribute 'version'

/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/setuptools/_distutils/version.py:355: AttributeError
____________________ test_gather_mm_idx_b[dtype2-0.003-16] _____________________

feat_size = 16, dtype = torch.float32, tol = 0.003

    @unittest.skipIf(dgl.backend.backend_name != 'pytorch', reason='Only support PyTorch for now')
    @pytest.mark.parametrize('feat_size', [1, 8, 16, 64, 256])
    @pytest.mark.parametrize('dtype, tol', [(torch.float16, 1e-2), (torch.bfloat16, 2e-2),
        (torch.float32, 3e-3), (torch.float64, 1e-4)])
    def test_gather_mm_idx_b(feat_size, dtype, tol):
        if F._default_context_str == 'cpu' and dtype in (torch.float16, torch.bfloat16):
            pytest.skip("Only support float32 and float64 on CPU.")
>       if LooseVersion(torch.version.cuda) < LooseVersion("11.0") \
            and dtype == torch.bfloat16:

tests/compute/test_sparse.py:339: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/setuptools/_distutils/version.py:72: in __lt__
    c = self._cmp(other)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

self = <[AttributeError("'LooseVersion' object has no attribute 'vstring'") raised in repr()] LooseVersion object at 0x7fb7062b07f0>
other = LooseVersion ('11.0')

    def _cmp (self, other):
        if isinstance(other, str):
            other = LooseVersion(other)
        elif not isinstance(other, LooseVersion):
            return NotImplemented
    
>       if self.version == other.version:
E       AttributeError: 'LooseVersion' object has no attribute 'version'

/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/setuptools/_distutils/version.py:355: AttributeError
____________________ test_gather_mm_idx_b[dtype2-0.003-64] _____________________

feat_size = 64, dtype = torch.float32, tol = 0.003

    @unittest.skipIf(dgl.backend.backend_name != 'pytorch', reason='Only support PyTorch for now')
    @pytest.mark.parametrize('feat_size', [1, 8, 16, 64, 256])
    @pytest.mark.parametrize('dtype, tol', [(torch.float16, 1e-2), (torch.bfloat16, 2e-2),
        (torch.float32, 3e-3), (torch.float64, 1e-4)])
    def test_gather_mm_idx_b(feat_size, dtype, tol):
        if F._default_context_str == 'cpu' and dtype in (torch.float16, torch.bfloat16):
            pytest.skip("Only support float32 and float64 on CPU.")
>       if LooseVersion(torch.version.cuda) < LooseVersion("11.0") \
            and dtype == torch.bfloat16:

tests/compute/test_sparse.py:339: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/setuptools/_distutils/version.py:72: in __lt__
    c = self._cmp(other)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

self = <[AttributeError("'LooseVersion' object has no attribute 'vstring'") raised in repr()] LooseVersion object at 0x7fb72aba1e10>
other = LooseVersion ('11.0')

    def _cmp (self, other):
        if isinstance(other, str):
            other = LooseVersion(other)
        elif not isinstance(other, LooseVersion):
            return NotImplemented
    
>       if self.version == other.version:
E       AttributeError: 'LooseVersion' object has no attribute 'version'

/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/setuptools/_distutils/version.py:355: AttributeError
____________________ test_gather_mm_idx_b[dtype2-0.003-256] ____________________

feat_size = 256, dtype = torch.float32, tol = 0.003

    @unittest.skipIf(dgl.backend.backend_name != 'pytorch', reason='Only support PyTorch for now')
    @pytest.mark.parametrize('feat_size', [1, 8, 16, 64, 256])
    @pytest.mark.parametrize('dtype, tol', [(torch.float16, 1e-2), (torch.bfloat16, 2e-2),
        (torch.float32, 3e-3), (torch.float64, 1e-4)])
    def test_gather_mm_idx_b(feat_size, dtype, tol):
        if F._default_context_str == 'cpu' and dtype in (torch.float16, torch.bfloat16):
            pytest.skip("Only support float32 and float64 on CPU.")
>       if LooseVersion(torch.version.cuda) < LooseVersion("11.0") \
            and dtype == torch.bfloat16:

tests/compute/test_sparse.py:339: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/setuptools/_distutils/version.py:72: in __lt__
    c = self._cmp(other)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

self = <[AttributeError("'LooseVersion' object has no attribute 'vstring'") raised in repr()] LooseVersion object at 0x7fb7064f00f0>
other = LooseVersion ('11.0')

    def _cmp (self, other):
        if isinstance(other, str):
            other = LooseVersion(other)
        elif not isinstance(other, LooseVersion):
            return NotImplemented
    
>       if self.version == other.version:
E       AttributeError: 'LooseVersion' object has no attribute 'version'

/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/setuptools/_distutils/version.py:355: AttributeError
____________________ test_gather_mm_idx_b[dtype3-0.0001-1] _____________________

feat_size = 1, dtype = torch.float64, tol = 0.0001

    @unittest.skipIf(dgl.backend.backend_name != 'pytorch', reason='Only support PyTorch for now')
    @pytest.mark.parametrize('feat_size', [1, 8, 16, 64, 256])
    @pytest.mark.parametrize('dtype, tol', [(torch.float16, 1e-2), (torch.bfloat16, 2e-2),
        (torch.float32, 3e-3), (torch.float64, 1e-4)])
    def test_gather_mm_idx_b(feat_size, dtype, tol):
        if F._default_context_str == 'cpu' and dtype in (torch.float16, torch.bfloat16):
            pytest.skip("Only support float32 and float64 on CPU.")
>       if LooseVersion(torch.version.cuda) < LooseVersion("11.0") \
            and dtype == torch.bfloat16:

tests/compute/test_sparse.py:339: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/setuptools/_distutils/version.py:72: in __lt__
    c = self._cmp(other)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

self = <[AttributeError("'LooseVersion' object has no attribute 'vstring'") raised in repr()] LooseVersion object at 0x7fb43e9cb668>
other = LooseVersion ('11.0')

    def _cmp (self, other):
        if isinstance(other, str):
            other = LooseVersion(other)
        elif not isinstance(other, LooseVersion):
            return NotImplemented
    
>       if self.version == other.version:
E       AttributeError: 'LooseVersion' object has no attribute 'version'

/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/setuptools/_distutils/version.py:355: AttributeError
____________________ test_gather_mm_idx_b[dtype3-0.0001-8] _____________________

feat_size = 8, dtype = torch.float64, tol = 0.0001

    @unittest.skipIf(dgl.backend.backend_name != 'pytorch', reason='Only support PyTorch for now')
    @pytest.mark.parametrize('feat_size', [1, 8, 16, 64, 256])
    @pytest.mark.parametrize('dtype, tol', [(torch.float16, 1e-2), (torch.bfloat16, 2e-2),
        (torch.float32, 3e-3), (torch.float64, 1e-4)])
    def test_gather_mm_idx_b(feat_size, dtype, tol):
        if F._default_context_str == 'cpu' and dtype in (torch.float16, torch.bfloat16):
            pytest.skip("Only support float32 and float64 on CPU.")
>       if LooseVersion(torch.version.cuda) < LooseVersion("11.0") \
            and dtype == torch.bfloat16:

tests/compute/test_sparse.py:339: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/setuptools/_distutils/version.py:72: in __lt__
    c = self._cmp(other)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

self = <[AttributeError("'LooseVersion' object has no attribute 'vstring'") raised in repr()] LooseVersion object at 0x7fb705fba128>
other = LooseVersion ('11.0')

    def _cmp (self, other):
        if isinstance(other, str):
            other = LooseVersion(other)
        elif not isinstance(other, LooseVersion):
            return NotImplemented
    
>       if self.version == other.version:
E       AttributeError: 'LooseVersion' object has no attribute 'version'

/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/setuptools/_distutils/version.py:355: AttributeError
____________________ test_gather_mm_idx_b[dtype3-0.0001-16] ____________________

feat_size = 16, dtype = torch.float64, tol = 0.0001

    @unittest.skipIf(dgl.backend.backend_name != 'pytorch', reason='Only support PyTorch for now')
    @pytest.mark.parametrize('feat_size', [1, 8, 16, 64, 256])
    @pytest.mark.parametrize('dtype, tol', [(torch.float16, 1e-2), (torch.bfloat16, 2e-2),
        (torch.float32, 3e-3), (torch.float64, 1e-4)])
    def test_gather_mm_idx_b(feat_size, dtype, tol):
        if F._default_context_str == 'cpu' and dtype in (torch.float16, torch.bfloat16):
            pytest.skip("Only support float32 and float64 on CPU.")
>       if LooseVersion(torch.version.cuda) < LooseVersion("11.0") \
            and dtype == torch.bfloat16:

tests/compute/test_sparse.py:339: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/setuptools/_distutils/version.py:72: in __lt__
    c = self._cmp(other)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

self = <[AttributeError("'LooseVersion' object has no attribute 'vstring'") raised in repr()] LooseVersion object at 0x7fb7061235f8>
other = LooseVersion ('11.0')

    def _cmp (self, other):
        if isinstance(other, str):
            other = LooseVersion(other)
        elif not isinstance(other, LooseVersion):
            return NotImplemented
    
>       if self.version == other.version:
E       AttributeError: 'LooseVersion' object has no attribute 'version'

/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/setuptools/_distutils/version.py:355: AttributeError
____________________ test_gather_mm_idx_b[dtype3-0.0001-64] ____________________

feat_size = 64, dtype = torch.float64, tol = 0.0001

    @unittest.skipIf(dgl.backend.backend_name != 'pytorch', reason='Only support PyTorch for now')
    @pytest.mark.parametrize('feat_size', [1, 8, 16, 64, 256])
    @pytest.mark.parametrize('dtype, tol', [(torch.float16, 1e-2), (torch.bfloat16, 2e-2),
        (torch.float32, 3e-3), (torch.float64, 1e-4)])
    def test_gather_mm_idx_b(feat_size, dtype, tol):
        if F._default_context_str == 'cpu' and dtype in (torch.float16, torch.bfloat16):
            pytest.skip("Only support float32 and float64 on CPU.")
>       if LooseVersion(torch.version.cuda) < LooseVersion("11.0") \
            and dtype == torch.bfloat16:

tests/compute/test_sparse.py:339: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/setuptools/_distutils/version.py:72: in __lt__
    c = self._cmp(other)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

self = <[AttributeError("'LooseVersion' object has no attribute 'vstring'") raised in repr()] LooseVersion object at 0x7fb729703748>
other = LooseVersion ('11.0')

    def _cmp (self, other):
        if isinstance(other, str):
            other = LooseVersion(other)
        elif not isinstance(other, LooseVersion):
            return NotImplemented
    
>       if self.version == other.version:
E       AttributeError: 'LooseVersion' object has no attribute 'version'

/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/setuptools/_distutils/version.py:355: AttributeError
___________________ test_gather_mm_idx_b[dtype3-0.0001-256] ____________________

feat_size = 256, dtype = torch.float64, tol = 0.0001

    @unittest.skipIf(dgl.backend.backend_name != 'pytorch', reason='Only support PyTorch for now')
    @pytest.mark.parametrize('feat_size', [1, 8, 16, 64, 256])
    @pytest.mark.parametrize('dtype, tol', [(torch.float16, 1e-2), (torch.bfloat16, 2e-2),
        (torch.float32, 3e-3), (torch.float64, 1e-4)])
    def test_gather_mm_idx_b(feat_size, dtype, tol):
        if F._default_context_str == 'cpu' and dtype in (torch.float16, torch.bfloat16):
            pytest.skip("Only support float32 and float64 on CPU.")
>       if LooseVersion(torch.version.cuda) < LooseVersion("11.0") \
            and dtype == torch.bfloat16:

tests/compute/test_sparse.py:339: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/setuptools/_distutils/version.py:72: in __lt__
    c = self._cmp(other)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

self = <[AttributeError("'LooseVersion' object has no attribute 'vstring'") raised in repr()] LooseVersion object at 0x7fb72baf5550>
other = LooseVersion ('11.0')

    def _cmp (self, other):
        if isinstance(other, str):
            other = LooseVersion(other)
        elif not isinstance(other, LooseVersion):
            return NotImplemented
    
>       if self.version == other.version:
E       AttributeError: 'LooseVersion' object has no attribute 'version'

/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/setuptools/_distutils/version.py:355: AttributeError
___________________________ test_use_libxsmm_switch ____________________________

    @unittest.skipIf(dgl.backend.backend_name != 'pytorch', reason='Only support PyTorch for now')
    @unittest.skipIf(F._default_context_str == 'gpu', reason="Libxsmm only fit in CPU.")
    def test_use_libxsmm_switch():
        import torch
        g = dgl.graph(([0, 0, 0, 1, 1, 2], [0, 1, 2, 1, 2, 2]))
        x = torch.ones(3, 2, requires_grad=True)
        y = torch.arange(1, 13).float().view(6, 2).requires_grad_()
    
        assert dgl.is_libxsmm_enabled()
>       dgl.ops.u_mul_e_sum(g, x, y)

tests/compute/test_sparse.py:407: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/ops/spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb705cfe120>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[1., 1.],
        [1., 1.],
        [1., 1.]], requires_grad=True)
rhs_data = tensor([[ 1.,  2.],
        [ 3.,  4.],
        [ 5.,  6.],
        [ 7.,  8.],
        [ 9., 10.],
        [11., 12.]], requires_grad=True)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
_________________________ test_v2v_update_all[idtype0] _________________________

idtype = torch.int32

    @parametrize_idtype
    def test_v2v_update_all(idtype):
        def _test(fld):
            def message_func(edges):
                return {'m' : edges.src[fld]}
    
            def message_func_edge(edges):
                if len(edges.src[fld].shape) == 1:
                    return {'m' : edges.src[fld] * edges.data['e1']}
                else:
                    return {'m' : edges.src[fld] * edges.data['e2']}
    
            def reduce_func(nodes):
                return {fld : F.sum(nodes.mailbox['m'], 1)}
    
            def apply_func(nodes):
                return {fld : 2 * nodes.data[fld]}
            g = generate_graph(idtype)
            # update all
            v1 = g.ndata[fld]
            g.update_all(fn.copy_src(src=fld, out='m'), fn.sum(msg='m', out=fld), apply_func)
            v2 = g.ndata[fld]
            g.ndata.update({fld : v1})
            g.update_all(message_func, reduce_func, apply_func)
            v3 = g.ndata[fld]
            assert F.allclose(v2, v3)
            # update all with edge weights
            v1 = g.ndata[fld]
            g.update_all(fn.src_mul_edge(src=fld, edge='e1', out='m'),
                    fn.sum(msg='m', out=fld), apply_func)
            v2 = g.ndata[fld]
            g.ndata.update({fld : v1})
            g.update_all(message_func_edge, reduce_func, apply_func)
            v4 = g.ndata[fld]
            assert F.allclose(v2, v4)
        # test 1d node features
>       _test('f1')

tests/compute/test_specialization.py:61: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
tests/compute/test_specialization.py:45: in _test
    g.update_all(fn.copy_src(src=fld, out='m'), fn.sum(msg='m', out=fld), apply_func)
python/dgl/heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python/dgl/core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python/dgl/core.py:332: in invoke_gspmm
    z = op(graph, x)
python/dgl/ops/spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43eacc360>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([ 1.7772, -0.6740, -0.8050,  0.1309, -0.4080, -1.8522, -0.3026,  0.1197,
        -1.4796,  0.8724])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
_________________________ test_v2v_update_all[idtype1] _________________________

idtype = torch.int64

    @parametrize_idtype
    def test_v2v_update_all(idtype):
        def _test(fld):
            def message_func(edges):
                return {'m' : edges.src[fld]}
    
            def message_func_edge(edges):
                if len(edges.src[fld].shape) == 1:
                    return {'m' : edges.src[fld] * edges.data['e1']}
                else:
                    return {'m' : edges.src[fld] * edges.data['e2']}
    
            def reduce_func(nodes):
                return {fld : F.sum(nodes.mailbox['m'], 1)}
    
            def apply_func(nodes):
                return {fld : 2 * nodes.data[fld]}
            g = generate_graph(idtype)
            # update all
            v1 = g.ndata[fld]
            g.update_all(fn.copy_src(src=fld, out='m'), fn.sum(msg='m', out=fld), apply_func)
            v2 = g.ndata[fld]
            g.ndata.update({fld : v1})
            g.update_all(message_func, reduce_func, apply_func)
            v3 = g.ndata[fld]
            assert F.allclose(v2, v3)
            # update all with edge weights
            v1 = g.ndata[fld]
            g.update_all(fn.src_mul_edge(src=fld, edge='e1', out='m'),
                    fn.sum(msg='m', out=fld), apply_func)
            v2 = g.ndata[fld]
            g.ndata.update({fld : v1})
            g.update_all(message_func_edge, reduce_func, apply_func)
            v4 = g.ndata[fld]
            assert F.allclose(v2, v4)
        # test 1d node features
>       _test('f1')

tests/compute/test_specialization.py:61: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
tests/compute/test_specialization.py:45: in _test
    g.update_all(fn.copy_src(src=fld, out='m'), fn.sum(msg='m', out=fld), apply_func)
python/dgl/heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python/dgl/core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python/dgl/core.py:332: in invoke_gspmm
    z = op(graph, x)
python/dgl/ops/spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb61c1f3948>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([-2.1621,  0.5268, -0.1093, -0.0476,  0.5928, -0.7815,  0.8331, -0.5041,
        -0.0937, -0.1564])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
____________________________ test_v2v_snr[idtype0] _____________________________

idtype = torch.int32

    @parametrize_idtype
    def test_v2v_snr(idtype):
        u = F.tensor([0, 0, 0, 3, 4, 9], idtype)
        v = F.tensor([1, 2, 3, 9, 9, 0], idtype)
        def _test(fld):
            def message_func(edges):
                return {'m' : edges.src[fld]}
    
            def message_func_edge(edges):
                if len(edges.src[fld].shape) == 1:
                    return {'m' : edges.src[fld] * edges.data['e1']}
                else:
                    return {'m' : edges.src[fld] * edges.data['e2']}
    
            def reduce_func(nodes):
                return {fld : F.sum(nodes.mailbox['m'], 1)}
    
            def apply_func(nodes):
                return {fld : 2 * nodes.data[fld]}
            g = generate_graph(idtype)
            # send and recv
            v1 = g.ndata[fld]
            g.send_and_recv((u, v), fn.copy_src(src=fld, out='m'),
                    fn.sum(msg='m', out=fld), apply_func)
            v2 = g.ndata[fld]
            g.ndata.update({fld : v1})
            g.send_and_recv((u, v), message_func, reduce_func, apply_func)
            v3 = g.ndata[fld]
            assert F.allclose(v2, v3)
            # send and recv with edge weights
            v1 = g.ndata[fld]
            g.send_and_recv((u, v), fn.src_mul_edge(src=fld, edge='e1', out='m'),
                    fn.sum(msg='m', out=fld), apply_func)
            v2 = g.ndata[fld]
            g.ndata.update({fld : v1})
            g.send_and_recv((u, v), message_func_edge, reduce_func, apply_func)
            v4 = g.ndata[fld]
            assert F.allclose(v2, v4)
        # test 1d node features
>       _test('f1')

tests/compute/test_specialization.py:104: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
tests/compute/test_specialization.py:88: in _test
    fn.sum(msg='m', out=fld), apply_func)
python/dgl/heterograph.py:4607: in send_and_recv
    compute_graph, message_func, reduce_func, apply_node_func)
python/dgl/core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python/dgl/core.py:332: in invoke_gspmm
    z = op(graph, x)
python/dgl/ops/spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43ead3360>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([-0.9186, -1.2810,  0.3374,  1.1025]), rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
____________________________ test_v2v_snr[idtype1] _____________________________

idtype = torch.int64

    @parametrize_idtype
    def test_v2v_snr(idtype):
        u = F.tensor([0, 0, 0, 3, 4, 9], idtype)
        v = F.tensor([1, 2, 3, 9, 9, 0], idtype)
        def _test(fld):
            def message_func(edges):
                return {'m' : edges.src[fld]}
    
            def message_func_edge(edges):
                if len(edges.src[fld].shape) == 1:
                    return {'m' : edges.src[fld] * edges.data['e1']}
                else:
                    return {'m' : edges.src[fld] * edges.data['e2']}
    
            def reduce_func(nodes):
                return {fld : F.sum(nodes.mailbox['m'], 1)}
    
            def apply_func(nodes):
                return {fld : 2 * nodes.data[fld]}
            g = generate_graph(idtype)
            # send and recv
            v1 = g.ndata[fld]
            g.send_and_recv((u, v), fn.copy_src(src=fld, out='m'),
                    fn.sum(msg='m', out=fld), apply_func)
            v2 = g.ndata[fld]
            g.ndata.update({fld : v1})
            g.send_and_recv((u, v), message_func, reduce_func, apply_func)
            v3 = g.ndata[fld]
            assert F.allclose(v2, v3)
            # send and recv with edge weights
            v1 = g.ndata[fld]
            g.send_and_recv((u, v), fn.src_mul_edge(src=fld, edge='e1', out='m'),
                    fn.sum(msg='m', out=fld), apply_func)
            v2 = g.ndata[fld]
            g.ndata.update({fld : v1})
            g.send_and_recv((u, v), message_func_edge, reduce_func, apply_func)
            v4 = g.ndata[fld]
            assert F.allclose(v2, v4)
        # test 1d node features
>       _test('f1')

tests/compute/test_specialization.py:104: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
tests/compute/test_specialization.py:88: in _test
    fn.sum(msg='m', out=fld), apply_func)
python/dgl/heterograph.py:4607: in send_and_recv
    compute_graph, message_func, reduce_func, apply_node_func)
python/dgl/core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python/dgl/core.py:332: in invoke_gspmm
    z = op(graph, x)
python/dgl/ops/spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43e7da240>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([-0.8413,  0.4031,  2.2376,  0.3881]), rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
____________________________ test_v2v_pull[idtype0] ____________________________

idtype = torch.int32

    @parametrize_idtype
    def test_v2v_pull(idtype):
        nodes = F.tensor([1, 2, 3, 9], idtype)
        def _test(fld):
            def message_func(edges):
                return {'m' : edges.src[fld]}
    
            def message_func_edge(edges):
                if len(edges.src[fld].shape) == 1:
                    return {'m' : edges.src[fld] * edges.data['e1']}
                else:
                    return {'m' : edges.src[fld] * edges.data['e2']}
    
            def reduce_func(nodes):
                return {fld : F.sum(nodes.mailbox['m'], 1)}
    
            def apply_func(nodes):
                return {fld : 2 * nodes.data[fld]}
            g = generate_graph(idtype)
            # send and recv
            v1 = g.ndata[fld]
            g.pull(nodes, fn.copy_src(src=fld, out='m'), fn.sum(msg='m', out=fld), apply_func)
            v2 = g.ndata[fld]
            g.ndata[fld] = v1
            g.pull(nodes, message_func, reduce_func, apply_func)
            v3 = g.ndata[fld]
            assert F.allclose(v2, v3)
            # send and recv with edge weights
            v1 = g.ndata[fld]
            g.pull(nodes, fn.src_mul_edge(src=fld, edge='e1', out='m'),
                    fn.sum(msg='m', out=fld), apply_func)
            v2 = g.ndata[fld]
            g.ndata[fld] = v1
            g.pull(nodes, message_func_edge, reduce_func, apply_func)
            v4 = g.ndata[fld]
            assert F.allclose(v2, v4)
        # test 1d node features
>       _test('f1')

tests/compute/test_specialization.py:146: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
tests/compute/test_specialization.py:130: in _test
    g.pull(nodes, fn.copy_src(src=fld, out='m'), fn.sum(msg='m', out=fld), apply_func)
python/dgl/heterograph.py:4711: in pull
    compute_graph, message_func, reduce_func, apply_node_func)
python/dgl/core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python/dgl/core.py:332: in invoke_gspmm
    z = op(graph, x)
python/dgl/ops/spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb4547f3e10>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([ 0.7355, -0.4884,  0.4938, -0.2935, -0.2445,  1.1038, -2.9475, -0.7513,
         0.4680])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
____________________________ test_v2v_pull[idtype1] ____________________________

idtype = torch.int64

    @parametrize_idtype
    def test_v2v_pull(idtype):
        nodes = F.tensor([1, 2, 3, 9], idtype)
        def _test(fld):
            def message_func(edges):
                return {'m' : edges.src[fld]}
    
            def message_func_edge(edges):
                if len(edges.src[fld].shape) == 1:
                    return {'m' : edges.src[fld] * edges.data['e1']}
                else:
                    return {'m' : edges.src[fld] * edges.data['e2']}
    
            def reduce_func(nodes):
                return {fld : F.sum(nodes.mailbox['m'], 1)}
    
            def apply_func(nodes):
                return {fld : 2 * nodes.data[fld]}
            g = generate_graph(idtype)
            # send and recv
            v1 = g.ndata[fld]
            g.pull(nodes, fn.copy_src(src=fld, out='m'), fn.sum(msg='m', out=fld), apply_func)
            v2 = g.ndata[fld]
            g.ndata[fld] = v1
            g.pull(nodes, message_func, reduce_func, apply_func)
            v3 = g.ndata[fld]
            assert F.allclose(v2, v3)
            # send and recv with edge weights
            v1 = g.ndata[fld]
            g.pull(nodes, fn.src_mul_edge(src=fld, edge='e1', out='m'),
                    fn.sum(msg='m', out=fld), apply_func)
            v2 = g.ndata[fld]
            g.ndata[fld] = v1
            g.pull(nodes, message_func_edge, reduce_func, apply_func)
            v4 = g.ndata[fld]
            assert F.allclose(v2, v4)
        # test 1d node features
>       _test('f1')

tests/compute/test_specialization.py:146: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
tests/compute/test_specialization.py:130: in _test
    g.pull(nodes, fn.copy_src(src=fld, out='m'), fn.sum(msg='m', out=fld), apply_func)
python/dgl/heterograph.py:4711: in pull
    compute_graph, message_func, reduce_func, apply_node_func)
python/dgl/core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python/dgl/core.py:332: in invoke_gspmm
    z = op(graph, x)
python/dgl/ops/spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb705b6d048>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([ 0.9869, -0.2625, -0.7306, -0.1077,  1.2263,  1.3180, -0.8481,  0.3126,
        -0.8374])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
___________________ test_update_all_multi_fallback[idtype0] ____________________

idtype = torch.int32

    @parametrize_idtype
    def test_update_all_multi_fallback(idtype):
        # create a graph with zero in degree nodes
        g = dgl.DGLGraph()
        g = g.astype(idtype).to(F.ctx())
        g.add_nodes(10)
        for i in range(1, 9):
            g.add_edge(0, i)
            g.add_edge(i, 9)
        g.ndata['h'] = F.randn((10, D))
        g.edata['w1'] = F.randn((16,))
        g.edata['w2'] = F.randn((16, D))
        def _mfunc_hxw1(edges):
            return {'m1' : edges.src['h'] * F.unsqueeze(edges.data['w1'], 1)}
        def _mfunc_hxw2(edges):
            return {'m2' : edges.src['h'] * edges.data['w2']}
        def _rfunc_m1(nodes):
            return {'o1' : F.sum(nodes.mailbox['m1'], 1)}
        def _rfunc_m2(nodes):
            return {'o2' : F.sum(nodes.mailbox['m2'], 1)}
        def _rfunc_m1max(nodes):
            return {'o3' : F.max(nodes.mailbox['m1'], 1)}
        def _afunc(nodes):
            ret = {}
            for k, v in nodes.data.items():
                if k.startswith('o'):
                    ret[k] = 2 * v
            return ret
        # compute ground truth
        g.update_all(_mfunc_hxw1, _rfunc_m1, _afunc)
        o1 = g.ndata.pop('o1')
        g.update_all(_mfunc_hxw2, _rfunc_m2, _afunc)
        o2 = g.ndata.pop('o2')
        g.update_all(_mfunc_hxw1, _rfunc_m1max, _afunc)
        o3 = g.ndata.pop('o3')
        # v2v spmv
        g.update_all(fn.src_mul_edge(src='h', edge='w1', out='m1'),
                     fn.sum(msg='m1', out='o1'),
>                    _afunc)

tests/compute/test_specialization.py:188: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python/dgl/core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python/dgl/core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python/dgl/ops/spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb614695630>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[ 1.2555, -0.5233, -0.6529, -0.9409, -0.7616],
        [ 0.8602, -0.1180, -0.1393, -0.6949,  1.6017],
        ...0.2302],
        [ 0.4938,  0.6535, -0.3516, -0.2553,  0.4336],
        [-2.0572,  0.5819, -0.4576,  0.9890,  1.5942]])
rhs_data = tensor([[-1.4983],
        [-1.2475],
        [-0.6912],
        [ 0.6846],
        [ 0.2962],
        [-0.4096],
    ...92],
        [-0.4065],
        [-0.1517],
        [ 0.2591],
        [ 0.9079],
        [ 0.3428],
        [ 0.5044]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
___________________ test_update_all_multi_fallback[idtype1] ____________________

idtype = torch.int64

    @parametrize_idtype
    def test_update_all_multi_fallback(idtype):
        # create a graph with zero in degree nodes
        g = dgl.DGLGraph()
        g = g.astype(idtype).to(F.ctx())
        g.add_nodes(10)
        for i in range(1, 9):
            g.add_edge(0, i)
            g.add_edge(i, 9)
        g.ndata['h'] = F.randn((10, D))
        g.edata['w1'] = F.randn((16,))
        g.edata['w2'] = F.randn((16, D))
        def _mfunc_hxw1(edges):
            return {'m1' : edges.src['h'] * F.unsqueeze(edges.data['w1'], 1)}
        def _mfunc_hxw2(edges):
            return {'m2' : edges.src['h'] * edges.data['w2']}
        def _rfunc_m1(nodes):
            return {'o1' : F.sum(nodes.mailbox['m1'], 1)}
        def _rfunc_m2(nodes):
            return {'o2' : F.sum(nodes.mailbox['m2'], 1)}
        def _rfunc_m1max(nodes):
            return {'o3' : F.max(nodes.mailbox['m1'], 1)}
        def _afunc(nodes):
            ret = {}
            for k, v in nodes.data.items():
                if k.startswith('o'):
                    ret[k] = 2 * v
            return ret
        # compute ground truth
        g.update_all(_mfunc_hxw1, _rfunc_m1, _afunc)
        o1 = g.ndata.pop('o1')
        g.update_all(_mfunc_hxw2, _rfunc_m2, _afunc)
        o2 = g.ndata.pop('o2')
        g.update_all(_mfunc_hxw1, _rfunc_m1max, _afunc)
        o3 = g.ndata.pop('o3')
        # v2v spmv
        g.update_all(fn.src_mul_edge(src='h', edge='w1', out='m1'),
                     fn.sum(msg='m1', out='o1'),
>                    _afunc)

tests/compute/test_specialization.py:188: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python/dgl/core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python/dgl/core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python/dgl/ops/spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43e692480>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[-1.2040, -1.8783,  1.2041, -1.0065, -0.0673],
        [ 0.0639, -1.7493,  0.0023, -0.1434, -0.2575],
        ...0.6172],
        [-1.1544,  1.9917, -1.1734,  0.8495, -0.3927],
        [ 0.0521,  0.0979,  0.4685, -1.2702,  1.0220]])
rhs_data = tensor([[ 1.4022],
        [-0.5181],
        [-0.6281],
        [-1.5431],
        [-0.3130],
        [ 2.7795],
    ...09],
        [ 1.6756],
        [ 0.0592],
        [-0.7749],
        [-0.1535],
        [ 0.4783],
        [-0.4258]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
______________________ test_pull_multi_fallback[idtype0] _______________________

idtype = torch.int32

    @parametrize_idtype
    def test_pull_multi_fallback(idtype):
        # create a graph with zero in degree nodes
        g = dgl.DGLGraph()
        g = g.astype(idtype).to(F.ctx())
        g.add_nodes(10)
        for i in range(1, 9):
            g.add_edge(0, i)
            g.add_edge(i, 9)
        g.ndata['h'] = F.randn((10, D))
        g.edata['w1'] = F.randn((16,))
        g.edata['w2'] = F.randn((16, D))
        def _mfunc_hxw1(edges):
            return {'m1' : edges.src['h'] * F.unsqueeze(edges.data['w1'], 1)}
        def _mfunc_hxw2(edges):
            return {'m2' : edges.src['h'] * edges.data['w2']}
        def _rfunc_m1(nodes):
            return {'o1' : F.sum(nodes.mailbox['m1'], 1)}
        def _rfunc_m2(nodes):
            return {'o2' : F.sum(nodes.mailbox['m2'], 1)}
        def _rfunc_m1max(nodes):
            return {'o3' : F.max(nodes.mailbox['m1'], 1)}
        def _afunc(nodes):
            ret = {}
            for k, v in nodes.data.items():
                if k.startswith('o'):
                    ret[k] = 2 * v
            return ret
        # nodes to pull
        def _pull_nodes(nodes):
            # compute ground truth
            g.pull(nodes, _mfunc_hxw1, _rfunc_m1, _afunc)
            o1 = g.ndata.pop('o1')
            g.pull(nodes, _mfunc_hxw2, _rfunc_m2, _afunc)
            o2 = g.ndata.pop('o2')
            g.pull(nodes, _mfunc_hxw1, _rfunc_m1max, _afunc)
            o3 = g.ndata.pop('o3')
            # v2v spmv
            g.pull(nodes, fn.src_mul_edge(src='h', edge='w1', out='m1'),
                         fn.sum(msg='m1', out='o1'),
                         _afunc)
            assert F.allclose(o1, g.ndata.pop('o1'))
            # v2v fallback to e2v
            g.pull(nodes, fn.src_mul_edge(src='h', edge='w2', out='m2'),
                         fn.sum(msg='m2', out='o2'),
                         _afunc)
            assert F.allclose(o2, g.ndata.pop('o2'))
        # test#1: non-0deg nodes
        nodes = [1, 2, 9]
>       _pull_nodes(nodes)

tests/compute/test_specialization.py:245: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
tests/compute/test_specialization.py:236: in _pull_nodes
    _afunc)
python/dgl/heterograph.py:4711: in pull
    compute_graph, message_func, reduce_func, apply_node_func)
python/dgl/core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python/dgl/core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python/dgl/ops/spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb4547cb0d8>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[-0.4455,  0.0341, -0.4392, -1.0152,  0.2764],
        [-0.2255, -1.1024,  0.8154,  1.9210,  0.4685],
        ...1.3489],
        [-0.8711, -0.2574,  0.3529, -0.6929, -0.1747],
        [-0.0171, -0.0439, -0.5086, -1.2868, -1.2972]])
rhs_data = tensor([[-0.4035],
        [ 0.0707],
        [-0.5416],
        [-0.2029],
        [-0.3885],
        [ 0.5734],
        [ 0.5644],
        [-0.1903],
        [-0.3218],
        [ 1.1178]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
______________________ test_pull_multi_fallback[idtype1] _______________________

idtype = torch.int64

    @parametrize_idtype
    def test_pull_multi_fallback(idtype):
        # create a graph with zero in degree nodes
        g = dgl.DGLGraph()
        g = g.astype(idtype).to(F.ctx())
        g.add_nodes(10)
        for i in range(1, 9):
            g.add_edge(0, i)
            g.add_edge(i, 9)
        g.ndata['h'] = F.randn((10, D))
        g.edata['w1'] = F.randn((16,))
        g.edata['w2'] = F.randn((16, D))
        def _mfunc_hxw1(edges):
            return {'m1' : edges.src['h'] * F.unsqueeze(edges.data['w1'], 1)}
        def _mfunc_hxw2(edges):
            return {'m2' : edges.src['h'] * edges.data['w2']}
        def _rfunc_m1(nodes):
            return {'o1' : F.sum(nodes.mailbox['m1'], 1)}
        def _rfunc_m2(nodes):
            return {'o2' : F.sum(nodes.mailbox['m2'], 1)}
        def _rfunc_m1max(nodes):
            return {'o3' : F.max(nodes.mailbox['m1'], 1)}
        def _afunc(nodes):
            ret = {}
            for k, v in nodes.data.items():
                if k.startswith('o'):
                    ret[k] = 2 * v
            return ret
        # nodes to pull
        def _pull_nodes(nodes):
            # compute ground truth
            g.pull(nodes, _mfunc_hxw1, _rfunc_m1, _afunc)
            o1 = g.ndata.pop('o1')
            g.pull(nodes, _mfunc_hxw2, _rfunc_m2, _afunc)
            o2 = g.ndata.pop('o2')
            g.pull(nodes, _mfunc_hxw1, _rfunc_m1max, _afunc)
            o3 = g.ndata.pop('o3')
            # v2v spmv
            g.pull(nodes, fn.src_mul_edge(src='h', edge='w1', out='m1'),
                         fn.sum(msg='m1', out='o1'),
                         _afunc)
            assert F.allclose(o1, g.ndata.pop('o1'))
            # v2v fallback to e2v
            g.pull(nodes, fn.src_mul_edge(src='h', edge='w2', out='m2'),
                         fn.sum(msg='m2', out='o2'),
                         _afunc)
            assert F.allclose(o2, g.ndata.pop('o2'))
        # test#1: non-0deg nodes
        nodes = [1, 2, 9]
>       _pull_nodes(nodes)

tests/compute/test_specialization.py:245: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
tests/compute/test_specialization.py:236: in _pull_nodes
    _afunc)
python/dgl/heterograph.py:4711: in pull
    compute_graph, message_func, reduce_func, apply_node_func)
python/dgl/core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python/dgl/core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python/dgl/ops/spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb705b6e708>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[ 0.0952,  0.6272, -0.6850, -0.2446, -0.5820],
        [ 0.0647, -0.5618, -0.1020,  1.3919, -0.5272],
        ...0.3953],
        [-0.0307,  1.1244, -0.8526,  0.3079,  0.8971],
        [ 0.1836,  2.9280, -1.3431,  0.3906,  0.9234]])
rhs_data = tensor([[-1.0083],
        [ 2.0015],
        [-0.4615],
        [ 0.0321],
        [ 0.2798],
        [ 0.1885],
        [-0.6768],
        [-1.3942],
        [-0.8055],
        [-0.9279]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
__________________________ test_spmv_3d_feat[idtype0] __________________________

idtype = torch.int32

    @parametrize_idtype
    def test_spmv_3d_feat(idtype):
        def src_mul_edge_udf(edges):
            return {'sum': edges.src['h'] * F.unsqueeze(F.unsqueeze(edges.data['h'], 1), 1)}
    
        def sum_udf(nodes):
            return {'h': F.sum(nodes.mailbox['sum'], 1)}
    
        n = 100
        p = 0.1
        a = sp.random(n, n, p, data_rvs=lambda n: np.ones(n))
        g = dgl.DGLGraph(a)
        g = g.astype(idtype).to(F.ctx())
        m = g.number_of_edges()
    
        # test#1: v2v with adj data
        h = F.randn((n, 5, 5))
        e = F.randn((m,))
    
        g.ndata['h'] = h
        g.edata['h'] = e
>       g.update_all(message_func=fn.src_mul_edge('h', 'h', 'sum'), reduce_func=fn.sum('sum', 'h')) # 1

tests/compute/test_specialization.py:271: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python/dgl/core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python/dgl/core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python/dgl/ops/spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43e692360>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[[ 0.1781, -1.7383,  0.5895, -0.7099, -0.3568],
         [ 0.6212,  0.1393, -0.3065, -1.9176, -1.1655],
      ...793],
         [ 0.1927,  0.7801,  0.6507,  0.0136,  0.2582],
         [-0.5435,  0.6627,  0.4242,  1.0154,  0.3597]]])
rhs_data = tensor([[[ 1.4667e+00]],

        [[-1.7203e+00]],

        [[-4.8379e-01]],

        [[-7.1619e-01]],

        [[-9.4... 1.0453e+00]],

        [[ 1.0157e+00]],

        [[-1.0315e+00]],

        [[ 9.9752e-01]],

        [[-4.0651e-02]]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
__________________________ test_spmv_3d_feat[idtype1] __________________________

idtype = torch.int64

    @parametrize_idtype
    def test_spmv_3d_feat(idtype):
        def src_mul_edge_udf(edges):
            return {'sum': edges.src['h'] * F.unsqueeze(F.unsqueeze(edges.data['h'], 1), 1)}
    
        def sum_udf(nodes):
            return {'h': F.sum(nodes.mailbox['sum'], 1)}
    
        n = 100
        p = 0.1
        a = sp.random(n, n, p, data_rvs=lambda n: np.ones(n))
        g = dgl.DGLGraph(a)
        g = g.astype(idtype).to(F.ctx())
        m = g.number_of_edges()
    
        # test#1: v2v with adj data
        h = F.randn((n, 5, 5))
        e = F.randn((m,))
    
        g.ndata['h'] = h
        g.edata['h'] = e
>       g.update_all(message_func=fn.src_mul_edge('h', 'h', 'sum'), reduce_func=fn.sum('sum', 'h')) # 1

tests/compute/test_specialization.py:271: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python/dgl/core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python/dgl/core.py:323: in invoke_gspmm
    z = op(graph, x, y)
python/dgl/ops/spmm.py:147: in func
    return gspmm(g, binary_op, reduce_op, x, y)
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb43e7da288>
op = 'mul', reduce_op = 'sum'
lhs_data = tensor([[[-1.2946,  2.1732,  0.7860,  0.4174, -0.6091],
         [ 1.0391, -0.0897,  0.3572, -0.1215,  1.1465],
      ...113],
         [ 0.0366,  1.7481, -0.2313, -0.8954,  0.1273],
         [ 0.0860, -0.1194, -0.4341, -1.7336, -0.7365]]])
rhs_data = tensor([[[-1.0758e+00]],

        [[-1.0779e+00]],

        [[-1.2231e+00]],

        [[-2.1482e+00]],

        [[ 3.0...-9.5957e-01]],

        [[ 2.0114e+00]],

        [[-1.7067e+00]],

        [[ 1.0071e+00]],

        [[-7.6480e-01]]])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
_______________________________ test_khop_graph ________________________________

    @unittest.skipIf(F._default_context_str == 'gpu', reason="GPU not implemented")
    def test_khop_graph():
        N = 20
        feat = F.randn((N, 5))
    
        def _test(g):
            for k in range(4):
                g_k = dgl.khop_graph(g, k)
                # use original graph to do message passing for k times.
                g.ndata['h'] = feat
                for _ in range(k):
                    g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
                h_0 = g.ndata.pop('h')
                # use k-hop graph to do message passing for one time.
                g_k.ndata['h'] = feat
                g_k.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
                h_1 = g_k.ndata.pop('h')
                assert F.allclose(h_0, h_1, rtol=1e-3, atol=1e-3)
    
        # Test for random undirected graphs
        g = dgl.DGLGraph(nx.erdos_renyi_graph(N, 0.3))
>       _test(g)

tests/compute/test_transform.py:471: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
tests/compute/test_transform.py:465: in _test
    g_k.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))
python/dgl/heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python/dgl/core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python/dgl/core.py:332: in invoke_gspmm
    z = op(graph, x)
python/dgl/ops/spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72bd9fe58>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-0.5472,  0.0936,  0.8332,  0.2144,  1.4115],
        [ 1.5455, -0.3460, -1.0035, -0.6472, -0.4461],
        ...0.9718],
        [-1.0126, -0.3500,  1.0720, -0.9389,  1.1804],
        [-0.1025,  0.7345, -0.1483,  1.0094, -0.1806]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
________________________________ test_khop_adj _________________________________

    @unittest.skipIf(F._default_context_str == 'gpu', reason="GPU not implemented")
    def test_khop_adj():
        N = 20
        feat = F.randn((N, 5))
        g = dgl.DGLGraph(nx.erdos_renyi_graph(N, 0.3))
        for k in range(3):
            adj = F.tensor(F.swapaxes(dgl.khop_adj(g, k), 0, 1))
            # use original graph to do message passing for k times.
            g.ndata['h'] = feat
            for _ in range(k):
>               g.update_all(fn.copy_u('h', 'm'), fn.sum('m', 'h'))

tests/compute/test_transform.py:486: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python/dgl/core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python/dgl/core.py:332: in invoke_gspmm
    z = op(graph, x)
python/dgl/ops/spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb61c1cb828>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 2.7372e+00, -8.0551e-02, -1.7735e+00, -1.2046e+00, -8.0728e-02],
        [-1.0605e+00, -3.3512e-01, -8.3698e...01, -1.0084e+00, -9.4934e-01,  9.5661e-01],
        [-1.5257e+00,  4.6575e-02,  1.4175e+00,  9.3071e-01,  7.3970e-01]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
___________________________ test_to_simple[idtype0] ____________________________

idtype = torch.int32

    @unittest.skipIf(F._default_context_str == 'gpu', reason="GPU to simple not implemented")
    @parametrize_idtype
    def test_to_simple(idtype):
        # homogeneous graph
        g = dgl.graph((F.tensor([0, 1, 2, 1]), F.tensor([1, 2, 0, 2])))
        g.ndata['h'] = F.tensor([[0.], [1.], [2.]])
        g.edata['h'] = F.tensor([[3.], [4.], [5.], [6.]])
        sg, wb = dgl.to_simple(g, writeback_mapping=True)
        u, v = g.all_edges(form='uv', order='eid')
        u = F.asnumpy(u).tolist()
        v = F.asnumpy(v).tolist()
        uv = list(zip(u, v))
        eid_map = F.asnumpy(wb)
    
        su, sv = sg.all_edges(form='uv', order='eid')
        su = F.asnumpy(su).tolist()
        sv = F.asnumpy(sv).tolist()
        suv = list(zip(su, sv))
        sc = F.asnumpy(sg.edata['count'])
        assert set(uv) == set(suv)
        for i, e in enumerate(suv):
            assert sc[i] == sum(e == _e for _e in uv)
        for i, e in enumerate(uv):
            assert eid_map[i] == suv.index(e)
        # shared ndata
        assert F.array_equal(sg.ndata['h'], g.ndata['h'])
        assert 'h' not in sg.edata
        # new ndata to sg
        sg.ndata['hh'] = F.tensor([[0.], [1.], [2.]])
        assert 'hh' not in g.ndata
    
        sg = dgl.to_simple(g, writeback_mapping=False, copy_ndata=False)
        assert 'h' not in sg.ndata
        assert 'h' not in sg.edata
    
        # test coalesce edge feature
        sg = dgl.to_simple(g, copy_edata=True, aggregator='arbitrary')
        assert F.allclose(sg.edata['h'][1], F.tensor([4.]))
>       sg = dgl.to_simple(g, copy_edata=True, aggregator='sum')

tests/compute/test_transform.py:816: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/transforms/functional.py:2625: in to_simple
    new_edge_frames = _coalesce_edge_frame(g, edge_maps, counts, aggregator)
python/dgl/transforms/functional.py:2453: in _coalesce_edge_frame
    new_data = F.scatter_add(data, feat_idx, _num_rows)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

x = tensor([[3.],
        [4.],
        [5.],
        [6.]])
idx = tensor([0, 1, 2, 1]), m = 3

    def scatter_add(x, idx, m):
        args = _cast_if_autocast_enabled(x, idx, m)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:785: TypeError
___________________________ test_to_simple[idtype1] ____________________________

idtype = torch.int64

    @unittest.skipIf(F._default_context_str == 'gpu', reason="GPU to simple not implemented")
    @parametrize_idtype
    def test_to_simple(idtype):
        # homogeneous graph
        g = dgl.graph((F.tensor([0, 1, 2, 1]), F.tensor([1, 2, 0, 2])))
        g.ndata['h'] = F.tensor([[0.], [1.], [2.]])
        g.edata['h'] = F.tensor([[3.], [4.], [5.], [6.]])
        sg, wb = dgl.to_simple(g, writeback_mapping=True)
        u, v = g.all_edges(form='uv', order='eid')
        u = F.asnumpy(u).tolist()
        v = F.asnumpy(v).tolist()
        uv = list(zip(u, v))
        eid_map = F.asnumpy(wb)
    
        su, sv = sg.all_edges(form='uv', order='eid')
        su = F.asnumpy(su).tolist()
        sv = F.asnumpy(sv).tolist()
        suv = list(zip(su, sv))
        sc = F.asnumpy(sg.edata['count'])
        assert set(uv) == set(suv)
        for i, e in enumerate(suv):
            assert sc[i] == sum(e == _e for _e in uv)
        for i, e in enumerate(uv):
            assert eid_map[i] == suv.index(e)
        # shared ndata
        assert F.array_equal(sg.ndata['h'], g.ndata['h'])
        assert 'h' not in sg.edata
        # new ndata to sg
        sg.ndata['hh'] = F.tensor([[0.], [1.], [2.]])
        assert 'hh' not in g.ndata
    
        sg = dgl.to_simple(g, writeback_mapping=False, copy_ndata=False)
        assert 'h' not in sg.ndata
        assert 'h' not in sg.edata
    
        # test coalesce edge feature
        sg = dgl.to_simple(g, copy_edata=True, aggregator='arbitrary')
        assert F.allclose(sg.edata['h'][1], F.tensor([4.]))
>       sg = dgl.to_simple(g, copy_edata=True, aggregator='sum')

tests/compute/test_transform.py:816: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/transforms/functional.py:2625: in to_simple
    new_edge_frames = _coalesce_edge_frame(g, edge_maps, counts, aggregator)
python/dgl/transforms/functional.py:2453: in _coalesce_edge_frame
    new_data = F.scatter_add(data, feat_idx, _num_rows)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

x = tensor([[3.],
        [4.],
        [5.],
        [6.]])
idx = tensor([0, 1, 2, 1]), m = 3

    def scatter_add(x, idx, m):
        args = _cast_if_autocast_enabled(x, idx, m)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:785: TypeError
__________________________ test_remove_edges[idtype0] __________________________

idtype = torch.int32

    @parametrize_idtype
    def test_remove_edges(idtype):
        # homogeneous Graphs
        g = dgl.graph(([0, 1], [1, 2]), idtype=idtype, device=F.ctx())
        e = 0
        g = dgl.remove_edges(g, e)
        assert g.number_of_edges() == 1
        u, v = g.edges(form='uv', order='eid')
        assert F.array_equal(u, F.tensor([1], dtype=idtype))
        assert F.array_equal(v, F.tensor([2], dtype=idtype))
        g = dgl.graph(([0, 1], [1, 2]), idtype=idtype, device=F.ctx())
        e = [0]
        g = dgl.remove_edges(g, e)
        assert g.number_of_edges() == 1
        u, v = g.edges(form='uv', order='eid')
        assert F.array_equal(u, F.tensor([1], dtype=idtype))
        assert F.array_equal(v, F.tensor([2], dtype=idtype))
        e = F.tensor([0], dtype=idtype)
        g = dgl.remove_edges(g, e)
        assert g.number_of_edges() == 0
    
        # has node data
        g = dgl.graph(([0, 1], [1, 2]), idtype=idtype, device=F.ctx())
        g.ndata['h'] = F.copy_to(F.tensor([1, 2, 3], dtype=idtype), ctx=F.ctx())
        g = dgl.remove_edges(g, 1)
        assert g.number_of_edges() == 1
        assert F.array_equal(g.ndata['h'], F.tensor([1, 2, 3], dtype=idtype))
    
        # has edge data
        g = dgl.graph(([0, 1], [1, 2]), idtype=idtype, device=F.ctx())
        g.edata['h'] = F.copy_to(F.tensor([1, 2], dtype=idtype), ctx=F.ctx())
        g = dgl.remove_edges(g, 0)
        assert g.number_of_edges() == 1
        assert F.array_equal(g.edata['h'], F.tensor([2], dtype=idtype))
    
        # invalid eid
        assert_fail = False
        try:
            g = dgl.remove_edges(g, 1)
        except:
            assert_fail = True
        assert assert_fail
    
        # bipartite graph
        g = dgl.heterograph(
            {('user', 'plays', 'game'): ([0, 1], [1, 2])}, idtype=idtype, device=F.ctx())
        e = 0
        g = dgl.remove_edges(g, e)
        assert g.number_of_edges() == 1
        u, v = g.edges(form='uv', order='eid')
        assert F.array_equal(u, F.tensor([1], dtype=idtype))
        assert F.array_equal(v, F.tensor([2], dtype=idtype))
        g = dgl.heterograph(
            {('user', 'plays', 'game'): ([0, 1], [1, 2])}, idtype=idtype, device=F.ctx())
        e = [0]
        g = dgl.remove_edges(g, e)
        assert g.number_of_edges() == 1
        u, v = g.edges(form='uv', order='eid')
        assert F.array_equal(u, F.tensor([1], dtype=idtype))
        assert F.array_equal(v, F.tensor([2], dtype=idtype))
        e = F.tensor([0], dtype=idtype)
        g = dgl.remove_edges(g, e)
        assert g.number_of_edges() == 0
    
        # has data
        g = dgl.heterograph(
            {('user', 'plays', 'game'): ([0, 1], [1, 2])}, idtype=idtype, device=F.ctx())
        g.nodes['user'].data['h'] = F.copy_to(F.tensor([1, 1], dtype=idtype), ctx=F.ctx())
        g.nodes['game'].data['h'] = F.copy_to(F.tensor([2, 2, 2], dtype=idtype), ctx=F.ctx())
        g.edata['h'] = F.copy_to(F.tensor([1, 2], dtype=idtype), ctx=F.ctx())
        g = dgl.remove_edges(g, 1)
        assert g.number_of_edges() == 1
        assert F.array_equal(g.nodes['user'].data['h'], F.tensor([1, 1], dtype=idtype))
        assert F.array_equal(g.nodes['game'].data['h'], F.tensor([2, 2, 2], dtype=idtype))
        assert F.array_equal(g.edata['h'], F.tensor([1], dtype=idtype))
    
        # heterogeneous graph
        g = create_test_heterograph3(idtype)
        g.edges['plays'].data['h'] = F.copy_to(F.tensor([1, 2, 3, 4], dtype=idtype), ctx=F.ctx())
        g = dgl.remove_edges(g, 1, etype='plays')
        assert g.number_of_edges('plays') == 3
        u, v = g.edges(form='uv', order='eid', etype='plays')
        assert F.array_equal(u, F.tensor([0, 1, 2], dtype=idtype))
        assert F.array_equal(v, F.tensor([0, 1, 1], dtype=idtype))
        assert F.array_equal(g.edges['plays'].data['h'], F.tensor([1, 3, 4], dtype=idtype))
        # remove all edges of 'develops'
        g = dgl.remove_edges(g, [0, 1], etype='develops')
        assert g.number_of_edges('develops') == 0
        assert F.array_equal(g.nodes['user'].data['h'], F.tensor([1, 1, 1], dtype=idtype))
        assert F.array_equal(g.nodes['game'].data['h'], F.tensor([2, 2], dtype=idtype))
        assert F.array_equal(g.nodes['developer'].data['h'], F.tensor([3, 3], dtype=idtype))
    
        # batched graph
        ctx = F.ctx()
        g1 = dgl.graph(([0, 1], [1, 2]), num_nodes=5, idtype=idtype, device=ctx)
        g2 = dgl.graph(([], []), idtype=idtype, device=ctx)
        g3 = dgl.graph(([2, 3, 4], [3, 2, 1]), idtype=idtype, device=ctx)
        bg = dgl.batch([g1, g2, g3])
>       bg_r = dgl.remove_edges(bg, 2)

tests/compute/test_transform.py:1389: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/transforms/functional.py:1755: in remove_edges
    g.remove_edges(eids, etype=etype, store_ids=store_ids)
python/dgl/heterograph.py:648: in remove_edges
    one_hot_removed_edges, reducer='sum')
python/dgl/ops/segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'sum', x = tensor([0., 0., 1., 0., 0.]), offsets = tensor([0, 2, 2, 5])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
__________________________ test_remove_edges[idtype1] __________________________

idtype = torch.int64

    @parametrize_idtype
    def test_remove_edges(idtype):
        # homogeneous Graphs
        g = dgl.graph(([0, 1], [1, 2]), idtype=idtype, device=F.ctx())
        e = 0
        g = dgl.remove_edges(g, e)
        assert g.number_of_edges() == 1
        u, v = g.edges(form='uv', order='eid')
        assert F.array_equal(u, F.tensor([1], dtype=idtype))
        assert F.array_equal(v, F.tensor([2], dtype=idtype))
        g = dgl.graph(([0, 1], [1, 2]), idtype=idtype, device=F.ctx())
        e = [0]
        g = dgl.remove_edges(g, e)
        assert g.number_of_edges() == 1
        u, v = g.edges(form='uv', order='eid')
        assert F.array_equal(u, F.tensor([1], dtype=idtype))
        assert F.array_equal(v, F.tensor([2], dtype=idtype))
        e = F.tensor([0], dtype=idtype)
        g = dgl.remove_edges(g, e)
        assert g.number_of_edges() == 0
    
        # has node data
        g = dgl.graph(([0, 1], [1, 2]), idtype=idtype, device=F.ctx())
        g.ndata['h'] = F.copy_to(F.tensor([1, 2, 3], dtype=idtype), ctx=F.ctx())
        g = dgl.remove_edges(g, 1)
        assert g.number_of_edges() == 1
        assert F.array_equal(g.ndata['h'], F.tensor([1, 2, 3], dtype=idtype))
    
        # has edge data
        g = dgl.graph(([0, 1], [1, 2]), idtype=idtype, device=F.ctx())
        g.edata['h'] = F.copy_to(F.tensor([1, 2], dtype=idtype), ctx=F.ctx())
        g = dgl.remove_edges(g, 0)
        assert g.number_of_edges() == 1
        assert F.array_equal(g.edata['h'], F.tensor([2], dtype=idtype))
    
        # invalid eid
        assert_fail = False
        try:
            g = dgl.remove_edges(g, 1)
        except:
            assert_fail = True
        assert assert_fail
    
        # bipartite graph
        g = dgl.heterograph(
            {('user', 'plays', 'game'): ([0, 1], [1, 2])}, idtype=idtype, device=F.ctx())
        e = 0
        g = dgl.remove_edges(g, e)
        assert g.number_of_edges() == 1
        u, v = g.edges(form='uv', order='eid')
        assert F.array_equal(u, F.tensor([1], dtype=idtype))
        assert F.array_equal(v, F.tensor([2], dtype=idtype))
        g = dgl.heterograph(
            {('user', 'plays', 'game'): ([0, 1], [1, 2])}, idtype=idtype, device=F.ctx())
        e = [0]
        g = dgl.remove_edges(g, e)
        assert g.number_of_edges() == 1
        u, v = g.edges(form='uv', order='eid')
        assert F.array_equal(u, F.tensor([1], dtype=idtype))
        assert F.array_equal(v, F.tensor([2], dtype=idtype))
        e = F.tensor([0], dtype=idtype)
        g = dgl.remove_edges(g, e)
        assert g.number_of_edges() == 0
    
        # has data
        g = dgl.heterograph(
            {('user', 'plays', 'game'): ([0, 1], [1, 2])}, idtype=idtype, device=F.ctx())
        g.nodes['user'].data['h'] = F.copy_to(F.tensor([1, 1], dtype=idtype), ctx=F.ctx())
        g.nodes['game'].data['h'] = F.copy_to(F.tensor([2, 2, 2], dtype=idtype), ctx=F.ctx())
        g.edata['h'] = F.copy_to(F.tensor([1, 2], dtype=idtype), ctx=F.ctx())
        g = dgl.remove_edges(g, 1)
        assert g.number_of_edges() == 1
        assert F.array_equal(g.nodes['user'].data['h'], F.tensor([1, 1], dtype=idtype))
        assert F.array_equal(g.nodes['game'].data['h'], F.tensor([2, 2, 2], dtype=idtype))
        assert F.array_equal(g.edata['h'], F.tensor([1], dtype=idtype))
    
        # heterogeneous graph
        g = create_test_heterograph3(idtype)
        g.edges['plays'].data['h'] = F.copy_to(F.tensor([1, 2, 3, 4], dtype=idtype), ctx=F.ctx())
        g = dgl.remove_edges(g, 1, etype='plays')
        assert g.number_of_edges('plays') == 3
        u, v = g.edges(form='uv', order='eid', etype='plays')
        assert F.array_equal(u, F.tensor([0, 1, 2], dtype=idtype))
        assert F.array_equal(v, F.tensor([0, 1, 1], dtype=idtype))
        assert F.array_equal(g.edges['plays'].data['h'], F.tensor([1, 3, 4], dtype=idtype))
        # remove all edges of 'develops'
        g = dgl.remove_edges(g, [0, 1], etype='develops')
        assert g.number_of_edges('develops') == 0
        assert F.array_equal(g.nodes['user'].data['h'], F.tensor([1, 1, 1], dtype=idtype))
        assert F.array_equal(g.nodes['game'].data['h'], F.tensor([2, 2], dtype=idtype))
        assert F.array_equal(g.nodes['developer'].data['h'], F.tensor([3, 3], dtype=idtype))
    
        # batched graph
        ctx = F.ctx()
        g1 = dgl.graph(([0, 1], [1, 2]), num_nodes=5, idtype=idtype, device=ctx)
        g2 = dgl.graph(([], []), idtype=idtype, device=ctx)
        g3 = dgl.graph(([2, 3, 4], [3, 2, 1]), idtype=idtype, device=ctx)
        bg = dgl.batch([g1, g2, g3])
>       bg_r = dgl.remove_edges(bg, 2)

tests/compute/test_transform.py:1389: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/transforms/functional.py:1755: in remove_edges
    g.remove_edges(eids, etype=etype, store_ids=store_ids)
python/dgl/heterograph.py:648: in remove_edges
    one_hot_removed_edges, reducer='sum')
python/dgl/ops/segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'sum', x = tensor([0., 0., 1., 0., 0.]), offsets = tensor([0, 2, 2, 5])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
__________________________ test_remove_nodes[idtype0] __________________________

idtype = torch.int32

    @parametrize_idtype
    def test_remove_nodes(idtype):
        # homogeneous Graphs
        g = dgl.graph(([0, 1], [1, 2]), idtype=idtype, device=F.ctx())
        n = 0
        g = dgl.remove_nodes(g, n)
        assert g.number_of_nodes() == 2
        assert g.number_of_edges() == 1
        u, v = g.edges(form='uv', order='eid')
        assert F.array_equal(u, F.tensor([0], dtype=idtype))
        assert F.array_equal(v, F.tensor([1], dtype=idtype))
        g = dgl.graph(([0, 1], [1, 2]), idtype=idtype, device=F.ctx())
        n = [1]
        g = dgl.remove_nodes(g, n)
        assert g.number_of_nodes() == 2
        assert g.number_of_edges() == 0
        g = dgl.graph(([0, 1], [1, 2]), idtype=idtype, device=F.ctx())
        n = F.tensor([2], dtype=idtype)
        g = dgl.remove_nodes(g, n)
        assert g.number_of_nodes() == 2
        assert g.number_of_edges() == 1
        u, v = g.edges(form='uv', order='eid')
        assert F.array_equal(u, F.tensor([0], dtype=idtype))
        assert F.array_equal(v, F.tensor([1], dtype=idtype))
    
        # invalid nid
        assert_fail = False
        try:
            g.remove_nodes(3)
        except:
            assert_fail = True
        assert assert_fail
    
        # has node and edge data
        g = dgl.graph(([0, 0, 2], [0, 1, 2]), idtype=idtype, device=F.ctx())
        g.ndata['hv'] = F.copy_to(F.tensor([1, 2, 3], dtype=idtype), ctx=F.ctx())
        g.edata['he'] = F.copy_to(F.tensor([1, 2, 3], dtype=idtype), ctx=F.ctx())
        g = dgl.remove_nodes(g, F.tensor([0], dtype=idtype))
        assert g.number_of_nodes() == 2
        assert g.number_of_edges() == 1
        u, v = g.edges(form='uv', order='eid')
        assert F.array_equal(u, F.tensor([1], dtype=idtype))
        assert F.array_equal(v, F.tensor([1], dtype=idtype))
        assert F.array_equal(g.ndata['hv'], F.tensor([2, 3], dtype=idtype))
        assert F.array_equal(g.edata['he'], F.tensor([3], dtype=idtype))
    
        # node id larger than current max node id
        g = dgl.heterograph(
            {('user', 'plays', 'game'): ([0, 1], [1, 2])}, idtype=idtype, device=F.ctx())
        n = 0
        g = dgl.remove_nodes(g, n, ntype='user')
        assert g.number_of_nodes('user') == 1
        assert g.number_of_nodes('game') == 3
        assert g.number_of_edges() == 1
        u, v = g.edges(form='uv', order='eid')
        assert F.array_equal(u, F.tensor([0], dtype=idtype))
        assert F.array_equal(v, F.tensor([2], dtype=idtype))
        g = dgl.heterograph(
            {('user', 'plays', 'game'): ([0, 1], [1, 2])}, idtype=idtype, device=F.ctx())
        n = [1]
        g = dgl.remove_nodes(g, n, ntype='user')
        assert g.number_of_nodes('user') == 1
        assert g.number_of_nodes('game') == 3
        assert g.number_of_edges() == 1
        u, v = g.edges(form='uv', order='eid')
        assert F.array_equal(u, F.tensor([0], dtype=idtype))
        assert F.array_equal(v, F.tensor([1], dtype=idtype))
        g = dgl.heterograph(
            {('user', 'plays', 'game'): ([0, 1], [1, 2])}, idtype=idtype, device=F.ctx())
        n = F.tensor([0], dtype=idtype)
        g = dgl.remove_nodes(g, n, ntype='game')
        assert g.number_of_nodes('user') == 2
        assert g.number_of_nodes('game') == 2
        assert g.number_of_edges() == 2
        u, v = g.edges(form='uv', order='eid')
        assert F.array_equal(u, F.tensor([0, 1], dtype=idtype))
        assert F.array_equal(v, F.tensor([0 ,1], dtype=idtype))
    
        # heterogeneous graph
        g = create_test_heterograph3(idtype)
        g.edges['plays'].data['h'] = F.copy_to(F.tensor([1, 2, 3, 4], dtype=idtype), ctx=F.ctx())
        g = dgl.remove_nodes(g, 0, ntype='game')
        assert g.number_of_nodes('user') == 3
        assert g.number_of_nodes('game') == 1
        assert g.number_of_nodes('developer') == 2
        assert g.number_of_edges('plays') == 2
        assert g.number_of_edges('develops') == 1
        assert F.array_equal(g.nodes['user'].data['h'], F.tensor([1, 1, 1], dtype=idtype))
        assert F.array_equal(g.nodes['game'].data['h'], F.tensor([2], dtype=idtype))
        assert F.array_equal(g.nodes['developer'].data['h'], F.tensor([3, 3], dtype=idtype))
        u, v = g.edges(form='uv', order='eid', etype='plays')
        assert F.array_equal(u, F.tensor([1, 2], dtype=idtype))
        assert F.array_equal(v, F.tensor([0, 0], dtype=idtype))
        assert F.array_equal(g.edges['plays'].data['h'], F.tensor([3, 4], dtype=idtype))
        u, v = g.edges(form='uv', order='eid', etype='develops')
        assert F.array_equal(u, F.tensor([1], dtype=idtype))
        assert F.array_equal(v, F.tensor([0], dtype=idtype))
    
        # batched graph
        ctx = F.ctx()
        g1 = dgl.graph(([0, 1], [1, 2]), num_nodes=5, idtype=idtype, device=ctx)
        g2 = dgl.graph(([], []), idtype=idtype, device=ctx)
        g3 = dgl.graph(([2, 3, 4], [3, 2, 1]), idtype=idtype, device=ctx)
        bg = dgl.batch([g1, g2, g3])
>       bg_r = dgl.remove_nodes(bg, 1)

tests/compute/test_transform.py:1565: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/transforms/functional.py:1835: in remove_nodes
    g.remove_nodes(nids, ntype=ntype, store_ids=store_ids)
python/dgl/heterograph.py:774: in remove_nodes
    c_ntype_batch_num_nodes, one_hot_removed_nodes, reducer='sum')
python/dgl/ops/segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'sum', x = tensor([0., 1., 0., 0., 0., 0., 0., 0., 0., 0.])
offsets = tensor([ 0,  5,  5, 10])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
__________________________ test_remove_nodes[idtype1] __________________________

idtype = torch.int64

    @parametrize_idtype
    def test_remove_nodes(idtype):
        # homogeneous Graphs
        g = dgl.graph(([0, 1], [1, 2]), idtype=idtype, device=F.ctx())
        n = 0
        g = dgl.remove_nodes(g, n)
        assert g.number_of_nodes() == 2
        assert g.number_of_edges() == 1
        u, v = g.edges(form='uv', order='eid')
        assert F.array_equal(u, F.tensor([0], dtype=idtype))
        assert F.array_equal(v, F.tensor([1], dtype=idtype))
        g = dgl.graph(([0, 1], [1, 2]), idtype=idtype, device=F.ctx())
        n = [1]
        g = dgl.remove_nodes(g, n)
        assert g.number_of_nodes() == 2
        assert g.number_of_edges() == 0
        g = dgl.graph(([0, 1], [1, 2]), idtype=idtype, device=F.ctx())
        n = F.tensor([2], dtype=idtype)
        g = dgl.remove_nodes(g, n)
        assert g.number_of_nodes() == 2
        assert g.number_of_edges() == 1
        u, v = g.edges(form='uv', order='eid')
        assert F.array_equal(u, F.tensor([0], dtype=idtype))
        assert F.array_equal(v, F.tensor([1], dtype=idtype))
    
        # invalid nid
        assert_fail = False
        try:
            g.remove_nodes(3)
        except:
            assert_fail = True
        assert assert_fail
    
        # has node and edge data
        g = dgl.graph(([0, 0, 2], [0, 1, 2]), idtype=idtype, device=F.ctx())
        g.ndata['hv'] = F.copy_to(F.tensor([1, 2, 3], dtype=idtype), ctx=F.ctx())
        g.edata['he'] = F.copy_to(F.tensor([1, 2, 3], dtype=idtype), ctx=F.ctx())
        g = dgl.remove_nodes(g, F.tensor([0], dtype=idtype))
        assert g.number_of_nodes() == 2
        assert g.number_of_edges() == 1
        u, v = g.edges(form='uv', order='eid')
        assert F.array_equal(u, F.tensor([1], dtype=idtype))
        assert F.array_equal(v, F.tensor([1], dtype=idtype))
        assert F.array_equal(g.ndata['hv'], F.tensor([2, 3], dtype=idtype))
        assert F.array_equal(g.edata['he'], F.tensor([3], dtype=idtype))
    
        # node id larger than current max node id
        g = dgl.heterograph(
            {('user', 'plays', 'game'): ([0, 1], [1, 2])}, idtype=idtype, device=F.ctx())
        n = 0
        g = dgl.remove_nodes(g, n, ntype='user')
        assert g.number_of_nodes('user') == 1
        assert g.number_of_nodes('game') == 3
        assert g.number_of_edges() == 1
        u, v = g.edges(form='uv', order='eid')
        assert F.array_equal(u, F.tensor([0], dtype=idtype))
        assert F.array_equal(v, F.tensor([2], dtype=idtype))
        g = dgl.heterograph(
            {('user', 'plays', 'game'): ([0, 1], [1, 2])}, idtype=idtype, device=F.ctx())
        n = [1]
        g = dgl.remove_nodes(g, n, ntype='user')
        assert g.number_of_nodes('user') == 1
        assert g.number_of_nodes('game') == 3
        assert g.number_of_edges() == 1
        u, v = g.edges(form='uv', order='eid')
        assert F.array_equal(u, F.tensor([0], dtype=idtype))
        assert F.array_equal(v, F.tensor([1], dtype=idtype))
        g = dgl.heterograph(
            {('user', 'plays', 'game'): ([0, 1], [1, 2])}, idtype=idtype, device=F.ctx())
        n = F.tensor([0], dtype=idtype)
        g = dgl.remove_nodes(g, n, ntype='game')
        assert g.number_of_nodes('user') == 2
        assert g.number_of_nodes('game') == 2
        assert g.number_of_edges() == 2
        u, v = g.edges(form='uv', order='eid')
        assert F.array_equal(u, F.tensor([0, 1], dtype=idtype))
        assert F.array_equal(v, F.tensor([0 ,1], dtype=idtype))
    
        # heterogeneous graph
        g = create_test_heterograph3(idtype)
        g.edges['plays'].data['h'] = F.copy_to(F.tensor([1, 2, 3, 4], dtype=idtype), ctx=F.ctx())
        g = dgl.remove_nodes(g, 0, ntype='game')
        assert g.number_of_nodes('user') == 3
        assert g.number_of_nodes('game') == 1
        assert g.number_of_nodes('developer') == 2
        assert g.number_of_edges('plays') == 2
        assert g.number_of_edges('develops') == 1
        assert F.array_equal(g.nodes['user'].data['h'], F.tensor([1, 1, 1], dtype=idtype))
        assert F.array_equal(g.nodes['game'].data['h'], F.tensor([2], dtype=idtype))
        assert F.array_equal(g.nodes['developer'].data['h'], F.tensor([3, 3], dtype=idtype))
        u, v = g.edges(form='uv', order='eid', etype='plays')
        assert F.array_equal(u, F.tensor([1, 2], dtype=idtype))
        assert F.array_equal(v, F.tensor([0, 0], dtype=idtype))
        assert F.array_equal(g.edges['plays'].data['h'], F.tensor([3, 4], dtype=idtype))
        u, v = g.edges(form='uv', order='eid', etype='develops')
        assert F.array_equal(u, F.tensor([1], dtype=idtype))
        assert F.array_equal(v, F.tensor([0], dtype=idtype))
    
        # batched graph
        ctx = F.ctx()
        g1 = dgl.graph(([0, 1], [1, 2]), num_nodes=5, idtype=idtype, device=ctx)
        g2 = dgl.graph(([], []), idtype=idtype, device=ctx)
        g3 = dgl.graph(([2, 3, 4], [3, 2, 1]), idtype=idtype, device=ctx)
        bg = dgl.batch([g1, g2, g3])
>       bg_r = dgl.remove_nodes(bg, 1)

tests/compute/test_transform.py:1565: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/transforms/functional.py:1835: in remove_nodes
    g.remove_nodes(nids, ntype=ntype, store_ids=store_ids)
python/dgl/heterograph.py:774: in remove_nodes
    c_ntype_batch_num_nodes, one_hot_removed_nodes, reducer='sum')
python/dgl/ops/segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'sum', x = tensor([0., 1., 0., 0., 0., 0., 0., 0., 0., 0.])
offsets = tensor([ 0,  5,  5, 10])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
__________________________ test_add_selfloop[idtype0] __________________________

idtype = torch.int32

    @parametrize_idtype
    def test_add_selfloop(idtype):
        # homogeneous graph
    
        # test for fill_data is float
        g = dgl.graph(([0, 0, 2], [2, 1, 0]), idtype=idtype, device=F.ctx())
        g.edata['he'] = F.copy_to(F.tensor([1, 2, 3], dtype=idtype), ctx=F.ctx())
        g.edata['he1'] = F.copy_to(F.tensor([[0., 1.], [2., 3.], [4., 5.]]), ctx=F.ctx())
        g.ndata['hn'] = F.copy_to(F.tensor([1, 2, 3], dtype=idtype), ctx=F.ctx())
        g = dgl.add_self_loop(g)
        assert g.number_of_nodes() == 3
        assert g.number_of_edges() == 6
        u, v = g.edges(form='uv', order='eid')
        assert F.array_equal(u, F.tensor([0, 0, 2, 0, 1, 2], dtype=idtype))
        assert F.array_equal(v, F.tensor([2, 1, 0, 0, 1, 2], dtype=idtype))
        assert F.array_equal(g.edata['he'], F.tensor([1, 2, 3, 1, 1, 1], dtype=idtype))
        assert F.array_equal(g.edata['he1'], F.tensor([[0., 1.], [2., 3.], [4., 5.],
                                                       [1., 1.], [1., 1.], [1., 1.]]))
    
        # test for fill_data is int
        g = dgl.graph(([0, 0, 2], [2, 1, 0]), idtype=idtype, device=F.ctx())
        g.edata['he'] = F.copy_to(F.tensor([1, 2, 3], dtype=idtype), ctx=F.ctx())
        g.edata['he1'] = F.copy_to(F.tensor([[0, 1], [2, 3], [4, 5]], dtype=idtype), ctx=F.ctx())
        g.ndata['hn'] = F.copy_to(F.tensor([1, 2, 3], dtype=idtype), ctx=F.ctx())
        g = dgl.add_self_loop(g, fill_data=1)
        assert g.number_of_nodes() == 3
        assert g.number_of_edges() == 6
        u, v = g.edges(form='uv', order='eid')
        assert F.array_equal(u, F.tensor([0, 0, 2, 0, 1, 2], dtype=idtype))
        assert F.array_equal(v, F.tensor([2, 1, 0, 0, 1, 2], dtype=idtype))
        assert F.array_equal(g.edata['he'], F.tensor([1, 2, 3, 1, 1, 1], dtype=idtype))
        assert F.array_equal(g.edata['he1'], F.tensor([[0, 1], [2, 3], [4, 5],
                                                       [1, 1], [1, 1], [1, 1]], dtype=idtype))
    
        # test for fill_data is str
        g = dgl.graph(([0, 0, 2], [2, 1, 0]), idtype=idtype, device=F.ctx())
        g.edata['he'] = F.copy_to(F.tensor([1., 2., 3.]), ctx=F.ctx())
        g.edata['he1'] = F.copy_to(F.tensor([[0., 1.], [2., 3.], [4., 5.]]), ctx=F.ctx())
        g.ndata['hn'] = F.copy_to(F.tensor([1, 2, 3], dtype=idtype), ctx=F.ctx())
>       g = dgl.add_self_loop(g, fill_data='sum')

tests/compute/test_transform.py:1675: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/transforms/functional.py:1950: in add_self_loop
    g.update_all(function.copy_e(feat_name, "h"), reducer('h', 'h'), etype=etype)
python/dgl/heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python/dgl/core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python/dgl/core.py:332: in invoke_gspmm
    z = op(graph, x)
python/dgl/ops/spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb705fab630>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([1., 2., 3.])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
__________________________ test_add_selfloop[idtype1] __________________________

idtype = torch.int64

    @parametrize_idtype
    def test_add_selfloop(idtype):
        # homogeneous graph
    
        # test for fill_data is float
        g = dgl.graph(([0, 0, 2], [2, 1, 0]), idtype=idtype, device=F.ctx())
        g.edata['he'] = F.copy_to(F.tensor([1, 2, 3], dtype=idtype), ctx=F.ctx())
        g.edata['he1'] = F.copy_to(F.tensor([[0., 1.], [2., 3.], [4., 5.]]), ctx=F.ctx())
        g.ndata['hn'] = F.copy_to(F.tensor([1, 2, 3], dtype=idtype), ctx=F.ctx())
        g = dgl.add_self_loop(g)
        assert g.number_of_nodes() == 3
        assert g.number_of_edges() == 6
        u, v = g.edges(form='uv', order='eid')
        assert F.array_equal(u, F.tensor([0, 0, 2, 0, 1, 2], dtype=idtype))
        assert F.array_equal(v, F.tensor([2, 1, 0, 0, 1, 2], dtype=idtype))
        assert F.array_equal(g.edata['he'], F.tensor([1, 2, 3, 1, 1, 1], dtype=idtype))
        assert F.array_equal(g.edata['he1'], F.tensor([[0., 1.], [2., 3.], [4., 5.],
                                                       [1., 1.], [1., 1.], [1., 1.]]))
    
        # test for fill_data is int
        g = dgl.graph(([0, 0, 2], [2, 1, 0]), idtype=idtype, device=F.ctx())
        g.edata['he'] = F.copy_to(F.tensor([1, 2, 3], dtype=idtype), ctx=F.ctx())
        g.edata['he1'] = F.copy_to(F.tensor([[0, 1], [2, 3], [4, 5]], dtype=idtype), ctx=F.ctx())
        g.ndata['hn'] = F.copy_to(F.tensor([1, 2, 3], dtype=idtype), ctx=F.ctx())
        g = dgl.add_self_loop(g, fill_data=1)
        assert g.number_of_nodes() == 3
        assert g.number_of_edges() == 6
        u, v = g.edges(form='uv', order='eid')
        assert F.array_equal(u, F.tensor([0, 0, 2, 0, 1, 2], dtype=idtype))
        assert F.array_equal(v, F.tensor([2, 1, 0, 0, 1, 2], dtype=idtype))
        assert F.array_equal(g.edata['he'], F.tensor([1, 2, 3, 1, 1, 1], dtype=idtype))
        assert F.array_equal(g.edata['he1'], F.tensor([[0, 1], [2, 3], [4, 5],
                                                       [1, 1], [1, 1], [1, 1]], dtype=idtype))
    
        # test for fill_data is str
        g = dgl.graph(([0, 0, 2], [2, 1, 0]), idtype=idtype, device=F.ctx())
        g.edata['he'] = F.copy_to(F.tensor([1., 2., 3.]), ctx=F.ctx())
        g.edata['he1'] = F.copy_to(F.tensor([[0., 1.], [2., 3.], [4., 5.]]), ctx=F.ctx())
        g.ndata['hn'] = F.copy_to(F.tensor([1, 2, 3], dtype=idtype), ctx=F.ctx())
>       g = dgl.add_self_loop(g, fill_data='sum')

tests/compute/test_transform.py:1675: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/transforms/functional.py:1950: in add_self_loop
    g.update_all(function.copy_e(feat_name, "h"), reducer('h', 'h'), etype=etype)
python/dgl/heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python/dgl/core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python/dgl/core.py:332: in invoke_gspmm
    z = op(graph, x)
python/dgl/ops/spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72b32b0d8>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([1., 2., 3.])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
________________________ test_remove_selfloop[idtype0] _________________________

idtype = torch.int32

    @parametrize_idtype
    def test_remove_selfloop(idtype):
        # homogeneous graph
        g = dgl.graph(([0, 0, 0, 1], [1, 0, 0, 2]), idtype=idtype, device=F.ctx())
        g.edata['he'] = F.copy_to(F.tensor([1, 2, 3, 4], dtype=idtype), ctx=F.ctx())
        g = dgl.remove_self_loop(g)
        assert g.number_of_nodes() == 3
        assert g.number_of_edges() == 2
        assert F.array_equal(g.edata['he'], F.tensor([1, 4], dtype=idtype))
    
        # bipartite graph
        g = dgl.heterograph(
            {('user', 'plays', 'game'): ([0, 1, 2], [1, 2, 2])}, idtype=idtype, device=F.ctx())
        # nothing will happend
        raise_error = False
        try:
            g = dgl.remove_self_loop(g, etype='plays')
        except:
            raise_error = True
        assert raise_error
    
        g = create_test_heterograph4(idtype)
        g = dgl.remove_self_loop(g, etype='follows')
        assert g.number_of_nodes('user') == 3
        assert g.number_of_nodes('game') == 2
        assert g.number_of_edges('follows') == 2
        assert g.number_of_edges('plays') == 2
        u, v = g.edges(form='uv', order='eid', etype='follows')
        assert F.array_equal(u, F.tensor([1, 2], dtype=idtype))
        assert F.array_equal(v, F.tensor([0, 1], dtype=idtype))
        assert F.array_equal(g.edges['follows'].data['h'], F.tensor([2, 4], dtype=idtype))
        assert F.array_equal(g.edges['plays'].data['h'], F.tensor([1, 2], dtype=idtype))
    
        raise_error = False
        try:
            g = dgl.remove_self_loop(g, etype='plays')
        except:
            raise_error = True
        assert raise_error
    
        # batch information
        g = dgl.graph(([0, 0, 0, 1, 3, 3, 4], [1, 0, 0, 2, 3, 4, 4]), idtype=idtype, device=F.ctx())
        g.set_batch_num_nodes(F.tensor([3, 2], dtype=F.int64))
        g.set_batch_num_edges(F.tensor([4, 3], dtype=F.int64))
>       g = dgl.remove_self_loop(g)

tests/compute/test_transform.py:1804: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/transforms/functional.py:2032: in remove_self_loop
    new_g = remove_edges(g, self_loop_eids, etype=etype)
python/dgl/transforms/functional.py:1755: in remove_edges
    g.remove_edges(eids, etype=etype, store_ids=store_ids)
python/dgl/heterograph.py:648: in remove_edges
    one_hot_removed_edges, reducer='sum')
python/dgl/ops/segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'sum', x = tensor([0., 1., 1., 0., 1., 0., 1.])
offsets = tensor([0, 4, 7])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
________________________ test_remove_selfloop[idtype1] _________________________

idtype = torch.int64

    @parametrize_idtype
    def test_remove_selfloop(idtype):
        # homogeneous graph
        g = dgl.graph(([0, 0, 0, 1], [1, 0, 0, 2]), idtype=idtype, device=F.ctx())
        g.edata['he'] = F.copy_to(F.tensor([1, 2, 3, 4], dtype=idtype), ctx=F.ctx())
        g = dgl.remove_self_loop(g)
        assert g.number_of_nodes() == 3
        assert g.number_of_edges() == 2
        assert F.array_equal(g.edata['he'], F.tensor([1, 4], dtype=idtype))
    
        # bipartite graph
        g = dgl.heterograph(
            {('user', 'plays', 'game'): ([0, 1, 2], [1, 2, 2])}, idtype=idtype, device=F.ctx())
        # nothing will happend
        raise_error = False
        try:
            g = dgl.remove_self_loop(g, etype='plays')
        except:
            raise_error = True
        assert raise_error
    
        g = create_test_heterograph4(idtype)
        g = dgl.remove_self_loop(g, etype='follows')
        assert g.number_of_nodes('user') == 3
        assert g.number_of_nodes('game') == 2
        assert g.number_of_edges('follows') == 2
        assert g.number_of_edges('plays') == 2
        u, v = g.edges(form='uv', order='eid', etype='follows')
        assert F.array_equal(u, F.tensor([1, 2], dtype=idtype))
        assert F.array_equal(v, F.tensor([0, 1], dtype=idtype))
        assert F.array_equal(g.edges['follows'].data['h'], F.tensor([2, 4], dtype=idtype))
        assert F.array_equal(g.edges['plays'].data['h'], F.tensor([1, 2], dtype=idtype))
    
        raise_error = False
        try:
            g = dgl.remove_self_loop(g, etype='plays')
        except:
            raise_error = True
        assert raise_error
    
        # batch information
        g = dgl.graph(([0, 0, 0, 1, 3, 3, 4], [1, 0, 0, 2, 3, 4, 4]), idtype=idtype, device=F.ctx())
        g.set_batch_num_nodes(F.tensor([3, 2], dtype=F.int64))
        g.set_batch_num_edges(F.tensor([4, 3], dtype=F.int64))
>       g = dgl.remove_self_loop(g)

tests/compute/test_transform.py:1804: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/transforms/functional.py:2032: in remove_self_loop
    new_g = remove_edges(g, self_loop_eids, etype=etype)
python/dgl/transforms/functional.py:1755: in remove_edges
    g.remove_edges(eids, etype=etype, store_ids=store_ids)
python/dgl/heterograph.py:648: in remove_edges
    one_hot_removed_edges, reducer='sum')
python/dgl/ops/segment.py:52: in segment_reduce
    rst = F.segment_reduce(reducer, value, offsets)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

op = 'sum', x = tensor([0., 1., 1., 0., 1., 0., 1.])
offsets = tensor([0, 4, 7])

    def segment_reduce(op, x, offsets):
        args = _cast_if_autocast_enabled(op, x, offsets)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:780: TypeError
_________________________ test_module_gcnnorm[idtype0] _________________________

idtype = torch.int32

    @parametrize_idtype
    def test_module_gcnnorm(idtype):
        g = dgl.heterograph({
            ('A', 'r1', 'A'): ([0, 1, 2], [0, 0, 1]),
            ('A', 'r2', 'B'): ([0, 0], [1, 1]),
            ('B', 'r3', 'B'): ([0, 1, 2], [0, 0, 1])
        }, idtype=idtype, device=F.ctx())
        g.edges['r3'].data['w'] = F.tensor([0.1, 0.2, 0.3])
        transform = dgl.GCNNorm()
>       new_g = transform(g)

tests/compute/test_transform.py:2305: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/transforms/module.py:1043: in __call__
    result[c_etype] = self.calc_etype(c_etype, g)
python/dgl/transforms/module.py:1025: in calc_etype
    g.update_all(fn.copy_e(self.eweight_name, 'm'), fn.sum('m', 'deg'), etype=c_etype)
python/dgl/heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python/dgl/core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python/dgl/core.py:332: in invoke_gspmm
    z = op(graph, x)
python/dgl/ops/spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb61dcdc3a8>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([0.1000, 0.2000, 0.3000])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
_________________________ test_module_gcnnorm[idtype1] _________________________

idtype = torch.int64

    @parametrize_idtype
    def test_module_gcnnorm(idtype):
        g = dgl.heterograph({
            ('A', 'r1', 'A'): ([0, 1, 2], [0, 0, 1]),
            ('A', 'r2', 'B'): ([0, 0], [1, 1]),
            ('B', 'r3', 'B'): ([0, 1, 2], [0, 0, 1])
        }, idtype=idtype, device=F.ctx())
        g.edges['r3'].data['w'] = F.tensor([0.1, 0.2, 0.3])
        transform = dgl.GCNNorm()
>       new_g = transform(g)

tests/compute/test_transform.py:2305: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/transforms/module.py:1043: in __call__
    result[c_etype] = self.calc_etype(c_etype, g)
python/dgl/transforms/module.py:1025: in calc_etype
    g.update_all(fn.copy_e(self.eweight_name, 'm'), fn.sum('m', 'deg'), etype=c_etype)
python/dgl/heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python/dgl/core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python/dgl/core.py:332: in invoke_gspmm
    z = op(graph, x)
python/dgl/ops/spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb72ba4e9d8>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([0.1000, 0.2000, 0.3000])

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
_____________________________ test_module_sign[g0] _____________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})

    @unittest.skipIf(dgl.backend.backend_name != 'pytorch', reason='Only support PyTorch for now')
    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature']))
    def test_module_sign(g):
        import torch
    
        atol = 1e-06
    
        ctx = F.ctx()
        g = g.to(ctx)
        adj = g.adj(transpose=True, scipy_fmt='coo').todense()
        adj = torch.tensor(adj).float().to(ctx)
    
        weight_adj = g.adj(transpose=True, scipy_fmt='coo').astype(float).todense()
        weight_adj = torch.tensor(weight_adj).float().to(ctx)
        src, dst = g.edges()
        src, dst = src.long(), dst.long()
        weight_adj[dst, src] = g.edata['scalar_w']
    
        # raw
        transform = dgl.SIGNDiffusion(k=1, in_feat_name='h', diffuse_op='raw')
>       g = transform(g)

tests/compute/test_transform.py:2476: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
python/dgl/transforms/module.py:1593: in __call__
    feat_list = self.diffuse(g)
python/dgl/transforms/module.py:1611: in raw
    g.update_all(message_func, fn.sum('m', self.in_feat_name))
python/dgl/heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python/dgl/core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python/dgl/core.py:332: in invoke_gspmm
    z = op(graph, x)
python/dgl/ops/spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fb705e87120>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.5296, -0.6817],
        [ 0.9662,  0.1646],
        [ 0.9910,  1.7198],
        [-0.8319, -1.1789],
      ...3282],
        [-0.8515, -0.5536],
        [ 0.4702,  1.1942],
        [ 0.8706,  0.7244],
        [ 1.5997, -0.1258]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
=============================== warnings summary ===============================
python/dgl/backend/backend.py:1741
  /root/jenkins/workspace/dgl_PR-4648@3/python/dgl/backend/backend.py:1741: DeprecationWarning: invalid escape sequence \P
    """

python/dgl/backend/pytorch/tensor.py:16
python/dgl/backend/pytorch/tensor.py:16
  /root/jenkins/workspace/dgl_PR-4648@3/python/dgl/backend/pytorch/tensor.py:16: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    if LooseVersion(th.__version__) < LooseVersion("1.9.0"):

python/dgl/backend/pytorch/tensor.py:340
python/dgl/backend/pytorch/tensor.py:340
  /root/jenkins/workspace/dgl_PR-4648@3/python/dgl/backend/pytorch/tensor.py:340: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    if LooseVersion(th.__version__) >= LooseVersion("1.10.0"):

python/dgl/dataloading/dataloader.py:33
  /root/jenkins/workspace/dgl_PR-4648@3/python/dgl/dataloading/dataloader.py:33: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    PYTORCH_VER = LooseVersion(torch.__version__)

python/dgl/_dataloading/pytorch/dataloader.py:23
  /root/jenkins/workspace/dgl_PR-4648@3/python/dgl/_dataloading/pytorch/dataloader.py:23: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    PYTORCH_VER = LooseVersion(th.__version__)

python/dgl/_dataloading/pytorch/dataloader.py:24
  /root/jenkins/workspace/dgl_PR-4648@3/python/dgl/_dataloading/pytorch/dataloader.py:24: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    PYTORCH_16 = PYTORCH_VER >= LooseVersion("1.6.0")

python/dgl/_dataloading/pytorch/dataloader.py:25
  /root/jenkins/workspace/dgl_PR-4648@3/python/dgl/_dataloading/pytorch/dataloader.py:25: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    PYTORCH_17 = PYTORCH_VER >= LooseVersion("1.7.0")

../../../../opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:10
  /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:10: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    _nlv = LooseVersion(_np_version)

../../../../opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:11
  /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:11: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    _np_version_under1p16 = _nlv < LooseVersion("1.16")

../../../../opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:12
  /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:12: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    _np_version_under1p17 = _nlv < LooseVersion("1.17")

../../../../opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:13
  /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:13: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    _np_version_under1p18 = _nlv < LooseVersion("1.18")

../../../../opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:14
  /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:14: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    _np_version_under1p19 = _nlv < LooseVersion("1.19")

../../../../opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:15
  /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:15: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    _np_version_under1p20 = _nlv < LooseVersion("1.20")

../../../../opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/setuptools/_distutils/version.py:351
  /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/setuptools/_distutils/version.py:351: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    other = LooseVersion(other)

../../../../opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/pandas/compat/numpy/function.py:125
../../../../opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/pandas/compat/numpy/function.py:125
  /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/pandas/compat/numpy/function.py:125: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    if LooseVersion(_np_version) >= LooseVersion("1.17.0"):

tests/compute/test_basics.py: 2 warnings
tests/compute/test_filter.py: 1 warning
tests/compute/test_graph.py: 9 warnings
tests/compute/test_kernel.py: 3 warnings
tests/compute/test_propagate.py: 2 warnings
tests/compute/test_removal.py: 16 warnings
tests/compute/test_serialize.py: 3 warnings
tests/compute/test_specialization.py: 12 warnings
tests/compute/test_subgraph.py: 2 warnings
tests/compute/test_transform.py: 10 warnings
tests/compute/test_traversal.py: 2 warnings
  /root/jenkins/workspace/dgl_PR-4648@3/python/dgl/heterograph.py:72: DGLWarning: Recommend creating graphs by `dgl.graph(data)` instead of `dgl.DGLGraph(data)`.
    dgl_warning('Recommend creating graphs by `dgl.graph(data)`'

tests/compute/test_basics.py: 2 warnings
tests/compute/test_batched_graph.py: 10 warnings
tests/compute/test_graph.py: 2 warnings
tests/compute/test_kernel.py: 1 warning
tests/compute/test_propagate.py: 2 warnings
tests/compute/test_removal.py: 10 warnings
tests/compute/test_specialization.py: 10 warnings
tests/compute/test_subgraph.py: 2 warnings
  /root/jenkins/workspace/dgl_PR-4648@3/python/dgl/heterograph.py:354: DGLWarning: DGLGraph.add_edge is deprecated. Please use DGLGraph.add_edges
    dgl_warning("DGLGraph.add_edge is deprecated. Please use DGLGraph.add_edges")

tests/compute/test_basics.py::test_update_all_0deg[idtype0]
tests/compute/test_basics.py::test_update_all_0deg[idtype1]
tests/compute/test_basics.py::test_pull_0deg[idtype0]
tests/compute/test_basics.py::test_pull_0deg[idtype1]
tests/compute/test_propagate.py::test_prop_nodes_topo[idtype0]
tests/compute/test_propagate.py::test_prop_nodes_topo[idtype1]
  /root/jenkins/workspace/dgl_PR-4648@3/python/dgl/core.py:79: DGLWarning: The input graph for the user-defined edge function does not contain valid edges
    dgl_warning('The input graph for the user-defined edge function ' \

tests/compute/test_batched_graph.py::test_batched_edge_ordering[idtype0]
tests/compute/test_batched_graph.py::test_batched_edge_ordering[idtype1]
tests/compute/test_graph.py::test_query
tests/compute/test_graph.py::test_hypersparse_query
tests/compute/test_transform.py::test_no_backtracking
tests/compute/test_transform.py::test_reverse[idtype0]
tests/compute/test_transform.py::test_reverse[idtype1]
  /root/jenkins/workspace/dgl_PR-4648@3/python/dgl/heterograph.py:2978: DGLWarning: DGLGraph.edge_id is deprecated. Please use DGLGraph.edge_ids.
    dgl_warning("DGLGraph.edge_id is deprecated. Please use DGLGraph.edge_ids.")

tests/compute/test_batched_heterograph.py::test_features[idtype0]
tests/compute/test_batched_heterograph.py::test_features[idtype1]
  /root/jenkins/workspace/dgl_PR-4648@3/python/dgl/batch.py:159: DGLWarning: Arguments edge_attrs has been deprecated. Please use edata instead.
    dgl_warning('Arguments edge_attrs has been deprecated. Please use'

tests/compute/test_csrmm.py::test_csrmm[dtype0-idtype0]
tests/compute/test_csrmm.py::test_csrmm[dtype0-idtype1]
tests/compute/test_csrmm.py::test_csrmm[dtype1-idtype0]
tests/compute/test_csrmm.py::test_csrmm[dtype1-idtype1]
tests/compute/test_csrmm.py::test_csrsum[dtype0-idtype0]
tests/compute/test_csrmm.py::test_csrsum[dtype0-idtype1]
tests/compute/test_csrmm.py::test_csrsum[dtype1-idtype0]
tests/compute/test_csrmm.py::test_csrsum[dtype1-idtype1]
  /root/jenkins/workspace/dgl_PR-4648@3/python/dgl/heterograph_index.py:797: FutureWarning: Adjacency matrix by default currently returns edge IDs.  As a result there is one 0 entry which is not eliminated.  In the next release it will return 1s by default, and 0 will be eliminated otherwise.
    FutureWarning)

tests/compute/test_data.py::test_minigc
  /root/jenkins/workspace/dgl_PR-4648@3/python/dgl/data/minigc.py:159: DeprecationWarning: `np.int` is a deprecated alias for the builtin `int`. To silence this warning, use `int` by itself. Doing this will not modify any behavior and is safe. When replacing `np.int`, you may wish to use e.g. `np.int64` or `np.int32` to specify the precision. If you wish to review your current use, check the release note link for additional information.
  Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations
    self.labels = F.tensor(np.array(self.labels).astype(np.int))

tests/compute/test_data.py::test_fraud
tests/compute/test_data.py::test_fraud
  /root/jenkins/workspace/dgl_PR-4648@3/python/dgl/data/fraud.py:206: DeprecationWarning: `np.bool` is a deprecated alias for the builtin `bool`. To silence this warning, use `bool` by itself. Doing this will not modify any behavior and is safe. If you specifically wanted the numpy scalar type, use `np.bool_` here.
  Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations
    train_mask = np.zeros(N, dtype=np.bool)

tests/compute/test_data.py::test_fraud
tests/compute/test_data.py::test_fraud
  /root/jenkins/workspace/dgl_PR-4648@3/python/dgl/data/fraud.py:207: DeprecationWarning: `np.bool` is a deprecated alias for the builtin `bool`. To silence this warning, use `bool` by itself. Doing this will not modify any behavior and is safe. If you specifically wanted the numpy scalar type, use `np.bool_` here.
  Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations
    val_mask = np.zeros(N, dtype=np.bool)

tests/compute/test_data.py::test_fraud
tests/compute/test_data.py::test_fraud
  /root/jenkins/workspace/dgl_PR-4648@3/python/dgl/data/fraud.py:208: DeprecationWarning: `np.bool` is a deprecated alias for the builtin `bool`. To silence this warning, use `bool` by itself. Doing this will not modify any behavior and is safe. If you specifically wanted the numpy scalar type, use `np.bool_` here.
  Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations
    test_mask = np.zeros(N, dtype=np.bool)

tests/compute/test_data.py::test_fakenews
tests/compute/test_data.py::test_fakenews
  /root/jenkins/workspace/dgl_PR-4648@3/python/dgl/data/fakenews.py:154: DeprecationWarning: `np.bool` is a deprecated alias for the builtin `bool`. To silence this warning, use `bool` by itself. Doing this will not modify any behavior and is safe. If you specifically wanted the numpy scalar type, use `np.bool_` here.
  Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations
    train_mask = np.zeros(num_graphs, dtype=np.bool)

tests/compute/test_data.py::test_fakenews
tests/compute/test_data.py::test_fakenews
  /root/jenkins/workspace/dgl_PR-4648@3/python/dgl/data/fakenews.py:155: DeprecationWarning: `np.bool` is a deprecated alias for the builtin `bool`. To silence this warning, use `bool` by itself. Doing this will not modify any behavior and is safe. If you specifically wanted the numpy scalar type, use `np.bool_` here.
  Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations
    val_mask = np.zeros(num_graphs, dtype=np.bool)

tests/compute/test_data.py::test_fakenews
tests/compute/test_data.py::test_fakenews
  /root/jenkins/workspace/dgl_PR-4648@3/python/dgl/data/fakenews.py:156: DeprecationWarning: `np.bool` is a deprecated alias for the builtin `bool`. To silence this warning, use `bool` by itself. Doing this will not modify any behavior and is safe. If you specifically wanted the numpy scalar type, use `np.bool_` here.
  Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations
    test_mask = np.zeros(num_graphs, dtype=np.bool)

tests/compute/test_data.py::test_citation_graph
  /root/jenkins/workspace/dgl_PR-4648@3/python/dgl/data/citation_graph.py:287: RuntimeWarning: divide by zero encountered in power
    r_inv = np.power(rowsum, -1).flatten()

tests/compute/test_data.py::test_csvdataset
  /root/jenkins/workspace/dgl_PR-4648@3/tests/compute/test_data.py:1194: DeprecationWarning: `np.float` is a deprecated alias for the builtin `float`. To silence this warning, use `float` by itself. Doing this will not modify any behavior and is safe. If you specifically wanted the numpy scalar type, use `np.float64` here.
  Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations
    node_ids = np.arange(num_nodes, dtype=np.float)

tests/compute/test_data.py::test_csvdataset
  /root/jenkins/workspace/dgl_PR-4648@3/tests/compute/test_data.py:1231: DeprecationWarning: `np.float` is a deprecated alias for the builtin `float`. To silence this warning, use `float` by itself. Doing this will not modify any behavior and is safe. If you specifically wanted the numpy scalar type, use `np.float64` here.
  Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations
    src_ids = np.random.randint(num_nodes, size=num_edges).astype(np.float)

tests/compute/test_data.py::test_csvdataset
  /root/jenkins/workspace/dgl_PR-4648@3/tests/compute/test_data.py:1232: DeprecationWarning: `np.float` is a deprecated alias for the builtin `float`. To silence this warning, use `float` by itself. Doing this will not modify any behavior and is safe. If you specifically wanted the numpy scalar type, use `np.float64` here.
  Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations
    dst_ids = np.random.randint(num_nodes, size=num_edges).astype(np.float)

tests/compute/test_data.py::test_csvdataset
  /root/jenkins/workspace/dgl_PR-4648@3/tests/compute/test_data.py:1262: DeprecationWarning: `np.float` is a deprecated alias for the builtin `float`. To silence this warning, use `float` by itself. Doing this will not modify any behavior and is safe. If you specifically wanted the numpy scalar type, use `np.float64` here.
  Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations
    graph_ids = np.arange(num_graphs).astype(np.float)

tests/compute/test_data.py::test_csvdataset
  /root/jenkins/workspace/dgl_PR-4648@3/tests/compute/test_data.py:513: DeprecationWarning: `np.int` is a deprecated alias for the builtin `int`. To silence this warning, use `int` by itself. Doing this will not modify any behavior and is safe. When replacing `np.int`, you may wish to use e.g. `np.int64` or `np.int32` to specify the precision. If you wish to review your current use, check the release note link for additional information.
  Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations
    node_ids = np.array([], dtype=np.int)

tests/compute/test_data.py::test_csvdataset
  /root/jenkins/workspace/dgl_PR-4648@3/tests/compute/test_data.py:514: DeprecationWarning: `np.int` is a deprecated alias for the builtin `int`. To silence this warning, use `int` by itself. Doing this will not modify any behavior and is safe. When replacing `np.int`, you may wish to use e.g. `np.int64` or `np.int32` to specify the precision. If you wish to review your current use, check the release note link for additional information.
  Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations
    src_ids = np.array([], dtype=np.int)

tests/compute/test_data.py::test_csvdataset
  /root/jenkins/workspace/dgl_PR-4648@3/tests/compute/test_data.py:515: DeprecationWarning: `np.int` is a deprecated alias for the builtin `int`. To silence this warning, use `int` by itself. Doing this will not modify any behavior and is safe. When replacing `np.int`, you may wish to use e.g. `np.int64` or `np.int32` to specify the precision. If you wish to review your current use, check the release note link for additional information.
  Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations
    dst_ids = np.array([], dtype=np.int)

tests/compute/test_data.py::test_csvdataset
  /root/jenkins/workspace/dgl_PR-4648@3/tests/compute/test_data.py:516: DeprecationWarning: `np.int` is a deprecated alias for the builtin `int`. To silence this warning, use `int` by itself. Doing this will not modify any behavior and is safe. When replacing `np.int`, you may wish to use e.g. `np.int64` or `np.int32` to specify the precision. If you wish to review your current use, check the release note link for additional information.
  Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations
    ngraph_ids = np.array([], dtype=np.int)

tests/compute/test_data.py::test_csvdataset
  /root/jenkins/workspace/dgl_PR-4648@3/tests/compute/test_data.py:517: DeprecationWarning: `np.int` is a deprecated alias for the builtin `int`. To silence this warning, use `int` by itself. Doing this will not modify any behavior and is safe. When replacing `np.int`, you may wish to use e.g. `np.int64` or `np.int32` to specify the precision. If you wish to review your current use, check the release note link for additional information.
  Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations
    egraph_ids = np.array([], dtype=np.int)

tests/compute/test_data.py::test_csvdataset
  /root/jenkins/workspace/dgl_PR-4648@3/tests/compute/test_data.py:518: DeprecationWarning: `np.int` is a deprecated alias for the builtin `int`. To silence this warning, use `int` by itself. Doing this will not modify any behavior and is safe. When replacing `np.int`, you may wish to use e.g. `np.int64` or `np.int32` to specify the precision. If you wish to review your current use, check the release note link for additional information.
  Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations
    u_indices = np.array([], dtype=np.int)

tests/compute/test_data.py::test_csvdataset
  /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/numpy/lib/arraysetops.py:604: FutureWarning: elementwise comparison failed; returning scalar instead, but in the future will perform elementwise comparison
    mask &= (ar1 != a)

tests/compute/test_data.py::test_csvdataset
  /root/jenkins/workspace/dgl_PR-4648@3/python/dgl/data/csv_dataset_base.py:298: DGLWarning: Unamed column is found. Ignored...
    dgl_warning("Unamed column is found. Ignored...")

tests/compute/test_graph.py::test_query
tests/compute/test_graph.py::test_hypersparse_query
tests/compute/test_heterograph.py::test_query[idtype0]
tests/compute/test_heterograph.py::test_query[idtype1]
  /root/jenkins/workspace/dgl_PR-4648@3/python/dgl/heterograph.py:2753: DGLWarning: DGLGraph.has_node is deprecated. Please use DGLGraph.has_nodes
    dgl_warning("DGLGraph.has_node is deprecated. Please use DGLGraph.has_nodes")

tests/compute/test_graph.py::test_query
tests/compute/test_graph.py::test_hypersparse_query
  /root/jenkins/workspace/dgl_PR-4648@3/python/dgl/heterograph.py:2687: DGLWarning: DGLGraph.__contains__ is deprecated. Please directly call has_nodes.
    dgl_warning('DGLGraph.__contains__ is deprecated.'

tests/compute/test_graph.py::test_query
tests/compute/test_sampling.py::test_non_uniform_random_walk[False]
tests/compute/test_sampling.py::test_uniform_random_walk[False]
tests/compute/test_sampling.py::test_node2vec
tests/compute/test_transform.py::test_no_backtracking
  /root/jenkins/workspace/dgl_PR-4648@3/python/dgl/heterograph.py:2851: DGLWarning: DGLGraph.has_edge_between is deprecated. Please use DGLGraph.has_edges_between
    dgl_warning("DGLGraph.has_edge_between is deprecated. "

tests/compute/test_graph.py::test_query
tests/compute/test_graph.py::test_hypersparse_query
  /root/jenkins/workspace/dgl_PR-4648@3/python/dgl/heterograph.py:3432: DGLWarning: DGLGraph.in_degree is deprecated. Please use DGLGraph.in_degrees
    dgl_warning("DGLGraph.in_degree is deprecated. Please use DGLGraph.in_degrees")

tests/compute/test_graph.py::test_query
tests/compute/test_graph.py::test_hypersparse_query
  /root/jenkins/workspace/dgl_PR-4648@3/python/dgl/heterograph.py:3516: DGLWarning: DGLGraph.out_degree is deprecated. Please use DGLGraph.out_degrees
    dgl_warning("DGLGraph.out_degree is deprecated. Please use DGLGraph.out_degrees")

tests/compute/test_graph.py::test_query
  /root/jenkins/workspace/dgl_PR-4648@3/python/dgl/heterograph.py:84: DGLWarning: Keyword arguments ['readonly', 'sort_csr'] are deprecated in v0.5, and can be safely removed in all cases.
    ' removed in all cases.'.format(list(deprecate_kwargs.keys())))

tests/compute/test_heterograph.py: 20 warnings
  /root/jenkins/workspace/dgl_PR-4648@3/tests/compute/test_heterograph.py:1128: DeprecationWarning: np.asscalar(a) is deprecated since NumPy v1.16, use a.item() instead
    assert np.asscalar(F.asnumpy(src_i)) == nid[src[i]]

tests/compute/test_heterograph.py: 20 warnings
  /root/jenkins/workspace/dgl_PR-4648@3/tests/compute/test_heterograph.py:1129: DeprecationWarning: np.asscalar(a) is deprecated since NumPy v1.16, use a.item() instead
    assert np.asscalar(F.asnumpy(dst_i)) == nid[dst[i]]

tests/compute/test_heterograph.py::test_invertible_conversion[idtype0]
tests/compute/test_heterograph.py::test_invertible_conversion[idtype1]
tests/compute/test_shared_mem.py::test_single_process[idtype0]
tests/compute/test_shared_mem.py::test_single_process[idtype1]
  /root/jenkins/workspace/dgl_PR-4648@3/python/dgl/heterograph.py:2635: DGLWarning: DGLGraph.is_readonly is deprecated in v0.5.
  DGLGraph now always supports mutable operations like add_nodes and add_edges.
    dgl_warning('DGLGraph.is_readonly is deprecated in v0.5.\n'

tests/compute/test_heterograph.py::test_types_in_function[idtype0]
tests/compute/test_heterograph.py::test_types_in_function[idtype1]
  /root/jenkins/workspace/dgl_PR-4648@3/python/dgl/backend/pytorch/tensor.py:277: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).
    mask = th.tensor(mask, dtype=th.bool)

tests/compute/test_pickle.py::test_pickling_batched_heterograph
  /root/jenkins/workspace/dgl_PR-4648@3/python/dgl/batch.py:511: DGLWarning: From v0.5, DGLHeteroGraph is merged into DGLGraph. You can safely replace dgl.batch_hetero with dgl.batch
    dgl_warning('From v0.5, DGLHeteroGraph is merged into DGLGraph. You can safely'

tests/compute/test_sampler.py::test_create_full
tests/compute/test_sampler.py::test_1neighbor_sampler_all
tests/compute/test_sampler.py::test_1neighbor_sampler
tests/compute/test_sampler.py::test_prefetch_neighbor_sampler
tests/compute/test_sampler.py::test_10neighbor_sampler_all
tests/compute/test_sampler.py::test_10neighbor_sampler
tests/compute/test_sampler.py::test_nonuniform_neighbor_sampler
tests/compute/test_sampler.py::test_setseed
  /root/jenkins/workspace/dgl_PR-4648@3/python/dgl/contrib/sampling/sampler.py:317: DGLWarning: dgl.contrib.sampling.NeighborSampler is deprecated starting from v0.5. Please read our guide<link> for how to use the new sampling APIs.
    dgl_warning('dgl.contrib.sampling.NeighborSampler is deprecated starting from v0.5.'

tests/compute/test_sampler.py::test_create_full
tests/compute/test_sampler.py::test_1neighbor_sampler_all
tests/compute/test_sampler.py::test_1neighbor_sampler
tests/compute/test_sampler.py::test_prefetch_neighbor_sampler
tests/compute/test_sampler.py::test_10neighbor_sampler_all
tests/compute/test_sampler.py::test_10neighbor_sampler
tests/compute/test_sampler.py::test_layer_sampler
tests/compute/test_sampler.py::test_nonuniform_neighbor_sampler
tests/compute/test_sampler.py::test_setseed
  /root/jenkins/workspace/dgl_PR-4648@3/python/dgl/_deprecate/nodeflow.py:99: DGLWarning: NodeFlow APIs are deprecated starting from v0.5. Please read our guide<link> for how to use the new sampling APIs.
    dgl_warning('NodeFlow APIs are deprecated starting from v0.5. Please read our'

tests/compute/test_sampling.py::test_global_uniform_negative_sampling[int32]
tests/compute/test_sampling.py::test_global_uniform_negative_sampling[int64]
  /root/jenkins/workspace/dgl_PR-4648@3/python/dgl/sampling/negative.py:102: ComplexWarning: Casting complex values to real discards the imaginary part
    g._graph, etype_id, num_samples, 3, exclude_self_loops, replace, redundancy)

tests/compute/test_serialize.py::test_graph_serialize_with_feature[False]
tests/compute/test_serialize.py::test_graph_serialize_without_feature[False]
tests/compute/test_serialize.py::test_graph_serialize_with_labels[False]
tests/compute/test_transform.py::test_simple_graph
  /root/jenkins/workspace/dgl_PR-4648@3/python/dgl/heterograph.py:84: DGLWarning: Keyword arguments ['readonly'] are deprecated in v0.5, and can be safely removed in all cases.
    ' removed in all cases.'.format(list(deprecate_kwargs.keys())))

tests/compute/test_serialize.py::test_load_old_files1
tests/compute/test_serialize.py::test_load_old_files2
  /root/jenkins/workspace/dgl_PR-4648@3/python/dgl/data/graph_serialize.py:179: DGLWarning: You are loading a graph file saved by old version of dgl.              Please consider saving it again with the current format.
    Please consider saving it again with the current format.")

tests/compute/test_sparse.py: 40 warnings
  /root/jenkins/workspace/dgl_PR-4648@3/tests/compute/test_sparse.py:299: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    if LooseVersion(torch.version.cuda) < LooseVersion("11.0") \

tests/compute/test_sparse.py: 20 warnings
  /root/jenkins/workspace/dgl_PR-4648@3/tests/compute/test_sparse.py:339: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    if LooseVersion(torch.version.cuda) < LooseVersion("11.0") \

tests/compute/test_transform.py::test_reverse_shared_frames[idtype0]
tests/compute/test_transform.py::test_reverse_shared_frames[idtype1]
  /root/jenkins/workspace/dgl_PR-4648@3/python/dgl/transforms/functional.py:1267: DGLWarning: share_ndata argument has been renamed to copy_ndata.
    dgl_warning('share_ndata argument has been renamed to copy_ndata.')

tests/compute/test_transform.py::test_reverse_shared_frames[idtype0]
tests/compute/test_transform.py::test_reverse_shared_frames[idtype1]
  /root/jenkins/workspace/dgl_PR-4648@3/python/dgl/transforms/functional.py:1270: DGLWarning: share_edata argument has been renamed to copy_edata.
    dgl_warning('share_edata argument has been renamed to copy_edata.')

tests/compute/test_transform.py::test_simple_graph
  /root/jenkins/workspace/dgl_PR-4648@3/python/dgl/transforms/functional.py:1319: DGLWarning: dgl.to_simple_graph is renamed to dgl.to_simple in v0.5.
    dgl_warning('dgl.to_simple_graph is renamed to dgl.to_simple in v0.5.')

tests/compute/test_transform.py::test_module_random_walk_pe[idtype0]
  /root/jenkins/workspace/dgl_PR-4648@3/python/dgl/backend/pytorch/tensor.py:44: UserWarning: The given NumPy array is not writeable, and PyTorch does not support non-writeable tensors. This means you can write to the underlying (supposedly non-writeable) NumPy array using the tensor. You may want to copy the array to protect its data or make it writeable before converting it to a tensor. This type of warning will be suppressed for the rest of this program. (Triggered internally at  /pytorch/torch/csrc/utils/tensor_numpy.cpp:180.)
    return th.as_tensor(data, dtype=dtype)

-- Docs: https://docs.pytest.org/en/stable/how-to/capture-warnings.html
- generated xml file: /root/jenkins/workspace/dgl_PR-4648@3/pytest_compute.xml -
============================ slowest 100 durations =============================
148.52s call     tests/compute/test_data.py::test_reddit
104.10s call     tests/compute/test_data.py::test_fakenews
26.63s call     tests/compute/test_data.py::test_as_graphpred
25.53s call     tests/compute/test_data.py::test_as_graphpred_ogb
23.57s call     tests/compute/test_data.py::test_gin
22.26s call     tests/compute/test_sampler.py::test_negative_sampler
12.38s call     tests/compute/test_data.py::test_as_linkpred_ogb
10.04s call     tests/compute/test_data.py::test_tudataset_regression
8.16s call     tests/compute/test_data.py::test_gnn_benchmark
7.70s call     tests/compute/test_data.py::test_as_graphpred_reprocess
7.25s call     tests/compute/test_data.py::test_as_nodepred_ogb
7.08s call     tests/compute/test_data.py::test_wiki_cs
4.96s call     tests/compute/test_data.py::test_fraud
4.18s call     tests/compute/test_graph.py::test_query
3.28s call     tests/compute/test_data.py::test_citation_graph
3.19s call     tests/compute/test_data.py::test_add_nodepred_split
3.06s call     tests/compute/test_sampling.py::test_sample_neighbors_outedge
3.00s call     tests/compute/test_data.py::test_flickr
2.79s call     tests/compute/test_heterograph.py::test_query[idtype1]
2.10s call     tests/compute/test_transform.py::test_metis_partition[idtype1]
1.73s call     tests/compute/test_heterograph.py::test_query[idtype0]
1.70s call     tests/compute/test_sampling.py::test_sample_neighbors_noprob
1.66s call     tests/compute/test_sampling.py::test_sample_neighbors_prob
1.39s call     tests/compute/test_heterograph.py::test_level2[idtype0]
1.38s call     tests/compute/test_data.py::test_explain_syn
1.13s call     tests/compute/test_heterograph.py::test_level2[idtype1]
1.09s call     tests/compute/test_heterograph.py::test_forking_pickler
0.91s call     tests/compute/test_heterograph.py::test_view1[idtype0]
0.90s call     tests/compute/test_heterograph.py::test_view1[idtype1]
0.82s call     tests/compute/test_basics.py::test_update_routines[idtype0]
0.80s call     tests/compute/test_basics.py::test_update_routines[idtype1]
0.66s call     tests/compute/test_transform.py::test_reorder_nodes
0.61s call     tests/compute/test_sampler.py::test_10neighbor_sampler_all
0.54s call     tests/compute/test_data.py::test_csvdataset
0.44s call     tests/compute/test_propagate.py::test_prop_edges_dfs[idtype0]
0.43s call     tests/compute/test_propagate.py::test_prop_edges_dfs[idtype1]
0.39s call     tests/compute/test_csrmm.py::test_csrmm[dtype0-idtype0]
0.39s call     tests/compute/test_csrmm.py::test_csrmm[dtype0-idtype1]
0.39s call     tests/compute/test_subgraph.py::test_in_subgraph[idtype0]
0.39s call     tests/compute/test_csrmm.py::test_csrmm_backward[1-dtype0-idtype1]
0.39s call     tests/compute/test_csrmm.py::test_csrsum[dtype0-idtype1]
0.38s call     tests/compute/test_transform.py::test_remove_nodes[idtype1]
0.37s call     tests/compute/test_csrmm.py::test_csrsum[dtype1-idtype0]
0.36s call     tests/compute/test_heterograph.py::test_types_in_function[idtype0]
0.36s call     tests/compute/test_subgraph.py::test_out_subgraph[idtype1]
0.35s call     tests/compute/test_csrmm.py::test_csrmm_backward[2-dtype0-idtype1]
0.35s call     tests/compute/test_csrmm.py::test_csrmm_backward[1-dtype1-idtype0]
0.34s call     tests/compute/test_batched_graph.py::test_batch_propagate[idtype0]
0.34s call     tests/compute/test_sampling.py::test_sample_neighbors_topk
0.34s call     tests/compute/test_batched_graph.py::test_batch_propagate[idtype1]
0.33s call     tests/compute/test_csrmm.py::test_csrmm[dtype1-idtype0]
0.33s call     tests/compute/test_csrmm.py::test_csrmm_backward[1-dtype1-idtype1]
0.33s call     tests/compute/test_csrmm.py::test_csrmm_backward[2-dtype0-idtype0]
0.32s call     tests/compute/test_sampling.py::test_uniform_random_walk[False]
0.32s call     tests/compute/test_csrmm.py::test_csrmm[dtype1-idtype1]
0.32s call     tests/compute/test_csrmm.py::test_csrmm_backward[1-dtype0-idtype0]
0.31s call     tests/compute/test_csrmm.py::test_csrmm_backward[2-dtype1-idtype1]
0.30s call     tests/compute/test_basics.py::test_pull_0deg[idtype1]
0.29s call     tests/compute/test_sampling.py::test_non_uniform_random_walk[False]
0.29s call     tests/compute/test_data.py::test_minigc
0.27s call     tests/compute/test_specialization.py::test_pull_multi_fallback[idtype0]
0.27s call     tests/compute/test_specialization.py::test_pull_multi_fallback[idtype1]
0.27s call     tests/compute/test_data.py::test_as_nodepred2
0.26s call     tests/compute/test_subgraph.py::test_in_subgraph[idtype1]
0.26s call     tests/compute/test_sampling.py::test_sample_neighbors_biased_bipartite
0.25s call     tests/compute/test_specialization.py::test_update_all_multi_fallback[idtype0]
0.25s call     tests/compute/test_csrmm.py::test_csrmm_backward[2-dtype1-idtype0]
0.24s call     tests/compute/test_heterograph.py::test_empty_query[idtype1]
0.23s call     tests/compute/test_batched_heterograph.py::test_slice_batch[idtype1]
0.21s call     tests/compute/test_batched_heterograph.py::test_slice_batch[idtype0]
0.21s call     tests/compute/test_csrmm.py::test_csrsum[dtype0-idtype0]
0.21s call     tests/compute/test_basics.py::test_pull_0deg[idtype0]
0.20s call     tests/compute/test_basics.py::test_send_multigraph[idtype0]
0.20s call     tests/compute/test_subgraph.py::test_out_subgraph[idtype0]
0.20s call     tests/compute/test_sampling.py::test_pinsage_sampling[False]
0.20s call     tests/compute/test_batched_graph.py::test_batch_send_and_recv[idtype0]
0.19s call     tests/compute/test_subgraph.py::test_khop_out_subgraph[idtype1]
0.19s call     tests/compute/test_sampling.py::test_sample_neighbors_biased_homogeneous
0.19s call     tests/compute/test_transform.py::test_reorder_graph[idtype0]
0.18s call     tests/compute/test_batched_graph.py::test_set_batch_info[idtype0]
0.18s call     tests/compute/test_batched_graph.py::test_batch_send_and_recv[idtype1]
0.18s call     tests/compute/test_batched_graph.py::test_set_batch_info[idtype1]
0.18s call     tests/compute/test_heterograph.py::test_format[idtype0]
0.18s call     tests/compute/test_subgraph.py::test_khop_in_subgraph[idtype1]
0.18s call     tests/compute/test_csrmm.py::test_csrmask_backward[dtype1-idtype0]
0.17s call     tests/compute/test_specialization.py::test_update_all_multi_fallback[idtype1]
0.17s call     tests/compute/test_csrmm.py::test_csrsum_backward[2-dtype1-idtype1]
0.17s call     tests/compute/test_csrmm.py::test_csrmask_backward[dtype0-idtype1]
0.17s call     tests/compute/test_csrmm.py::test_csrmask_backward[dtype0-idtype0]
0.17s call     tests/compute/test_csrmm.py::test_csrsum[dtype1-idtype1]
0.16s call     tests/compute/test_heterograph.py::test_ismultigraph[idtype0]
0.16s call     tests/compute/test_basics.py::test_degree_bucket_edge_ordering[idtype1]
0.16s call     tests/compute/test_sort.py::test_sort_with_tag[idtype0]
0.16s call     tests/compute/test_csrmm.py::test_csrsum_backward[2-dtype0-idtype0]
0.16s call     tests/compute/test_csrmm.py::test_csrsum_backward[2-dtype1-idtype0]
0.16s call     tests/compute/test_propagate.py::test_prop_nodes_topo[idtype0]
0.15s call     tests/compute/test_csrmm.py::test_csrsum_backward[2-dtype0-idtype1]
0.15s call     tests/compute/test_subgraph.py::test_khop_in_subgraph[idtype0]
0.15s call     tests/compute/test_basics.py::test_update_all_0deg[idtype0]
0.15s call     tests/compute/test_batched_graph.py::test_batched_edge_ordering[idtype1]
=========================== short test summary info ============================
FAILED tests/compute/test_apply_edges_hetero.py::test_unary_copy_u[idtype0]
FAILED tests/compute/test_apply_edges_hetero.py::test_unary_copy_u[idtype1]
FAILED tests/compute/test_apply_edges_hetero.py::test_binary_op[idtype0] - Ty...
FAILED tests/compute/test_apply_edges_hetero.py::test_binary_op[idtype1] - Ty...
FAILED tests/compute/test_basics.py::test_issue_1088[idtype0] - TypeError: em...
FAILED tests/compute/test_basics.py::test_issue_1088[idtype1] - TypeError: em...
FAILED tests/compute/test_basics.py::test_issue_2484[idtype0] - TypeError: em...
FAILED tests/compute/test_basics.py::test_issue_2484[idtype1] - TypeError: em...
FAILED tests/compute/test_edge_softmax_hetero.py::test_edge_softmax_unidirectional
FAILED tests/compute/test_edge_softmax_hetero.py::test_edge_softmax[idtype0-src-g0]
FAILED tests/compute/test_edge_softmax_hetero.py::test_edge_softmax[idtype0-dst-g0]
FAILED tests/compute/test_edge_softmax_hetero.py::test_edge_softmax[idtype1-src-g0]
FAILED tests/compute/test_edge_softmax_hetero.py::test_edge_softmax[idtype1-dst-g0]
FAILED tests/compute/test_heterograph.py::test_updates[idtype0] - TypeError: ...
FAILED tests/compute/test_heterograph.py::test_updates[idtype1] - TypeError: ...
FAILED tests/compute/test_heterograph.py::test_backward[idtype0] - TypeError:...
FAILED tests/compute/test_heterograph.py::test_backward[idtype1] - TypeError:...
FAILED tests/compute/test_kernel.py::test_copy_src_reduce - TypeError: empty_...
FAILED tests/compute/test_kernel.py::test_copy_edge_reduce - TypeError: empty...
FAILED tests/compute/test_kernel.py::test_all_binary_builtins - TypeError: em...
FAILED tests/compute/test_kernel.py::test_mean_zero_degree[g0-idtype0] - Type...
FAILED tests/compute/test_kernel.py::test_mean_zero_degree[g0-idtype1] - Type...
FAILED tests/compute/test_new_update_all_hetero.py::test_unary_copy_u[idtype0]
FAILED tests/compute/test_new_update_all_hetero.py::test_unary_copy_u[idtype1]
FAILED tests/compute/test_new_update_all_hetero.py::test_unary_copy_e[idtype0]
FAILED tests/compute/test_new_update_all_hetero.py::test_unary_copy_e[idtype1]
FAILED tests/compute/test_new_update_all_hetero.py::test_binary_op[idtype0]
FAILED tests/compute/test_new_update_all_hetero.py::test_binary_op[idtype1]
FAILED tests/compute/test_readout.py::test_sum_case1[idtype0] - TypeError: em...
FAILED tests/compute/test_readout.py::test_sum_case1[idtype1] - TypeError: em...
FAILED tests/compute/test_readout.py::test_reduce_readout[sum-g0-idtype0] - T...
FAILED tests/compute/test_readout.py::test_reduce_readout[sum-g0-idtype1] - T...
FAILED tests/compute/test_readout.py::test_reduce_readout[sum-g1-idtype0] - T...
FAILED tests/compute/test_readout.py::test_reduce_readout[sum-g1-idtype1] - T...
FAILED tests/compute/test_readout.py::test_reduce_readout[sum-g2-idtype0] - T...
FAILED tests/compute/test_readout.py::test_reduce_readout[sum-g2-idtype1] - T...
FAILED tests/compute/test_readout.py::test_reduce_readout[sum-g3-idtype0] - T...
FAILED tests/compute/test_readout.py::test_reduce_readout[sum-g3-idtype1] - T...
FAILED tests/compute/test_readout.py::test_reduce_readout[sum-g4-idtype0] - T...
FAILED tests/compute/test_readout.py::test_reduce_readout[sum-g4-idtype1] - T...
FAILED tests/compute/test_readout.py::test_reduce_readout[sum-g5-idtype0] - T...
FAILED tests/compute/test_readout.py::test_reduce_readout[sum-g5-idtype1] - T...
FAILED tests/compute/test_readout.py::test_reduce_readout[sum-g6-idtype0] - T...
FAILED tests/compute/test_readout.py::test_reduce_readout[sum-g6-idtype1] - T...
FAILED tests/compute/test_readout.py::test_reduce_readout[max-g0-idtype0] - T...
FAILED tests/compute/test_readout.py::test_reduce_readout[max-g0-idtype1] - T...
FAILED tests/compute/test_readout.py::test_reduce_readout[max-g1-idtype0] - T...
FAILED tests/compute/test_readout.py::test_reduce_readout[max-g1-idtype1] - T...
FAILED tests/compute/test_readout.py::test_reduce_readout[max-g2-idtype0] - T...
FAILED tests/compute/test_readout.py::test_reduce_readout[max-g2-idtype1] - T...
FAILED tests/compute/test_readout.py::test_reduce_readout[max-g3-idtype0] - T...
FAILED tests/compute/test_readout.py::test_reduce_readout[max-g3-idtype1] - T...
FAILED tests/compute/test_readout.py::test_reduce_readout[max-g4-idtype0] - T...
FAILED tests/compute/test_readout.py::test_reduce_readout[max-g4-idtype1] - T...
FAILED tests/compute/test_readout.py::test_reduce_readout[max-g5-idtype0] - T...
FAILED tests/compute/test_readout.py::test_reduce_readout[max-g5-idtype1] - T...
FAILED tests/compute/test_readout.py::test_reduce_readout[max-g6-idtype0] - T...
FAILED tests/compute/test_readout.py::test_reduce_readout[max-g6-idtype1] - T...
FAILED tests/compute/test_readout.py::test_reduce_readout[mean-g0-idtype0] - ...
FAILED tests/compute/test_readout.py::test_reduce_readout[mean-g0-idtype1] - ...
FAILED tests/compute/test_readout.py::test_reduce_readout[mean-g1-idtype0] - ...
FAILED tests/compute/test_readout.py::test_reduce_readout[mean-g1-idtype1] - ...
FAILED tests/compute/test_readout.py::test_reduce_readout[mean-g2-idtype0] - ...
FAILED tests/compute/test_readout.py::test_reduce_readout[mean-g2-idtype1] - ...
FAILED tests/compute/test_readout.py::test_reduce_readout[mean-g3-idtype0] - ...
FAILED tests/compute/test_readout.py::test_reduce_readout[mean-g3-idtype1] - ...
FAILED tests/compute/test_readout.py::test_reduce_readout[mean-g4-idtype0] - ...
FAILED tests/compute/test_readout.py::test_reduce_readout[mean-g4-idtype1] - ...
FAILED tests/compute/test_readout.py::test_reduce_readout[mean-g5-idtype0] - ...
FAILED tests/compute/test_readout.py::test_reduce_readout[mean-g5-idtype1] - ...
FAILED tests/compute/test_readout.py::test_reduce_readout[mean-g6-idtype0] - ...
FAILED tests/compute/test_readout.py::test_reduce_readout[mean-g6-idtype1] - ...
FAILED tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g0-idtype0]
FAILED tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g0-idtype1]
FAILED tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g1-idtype0]
FAILED tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g1-idtype1]
FAILED tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g2-idtype0]
FAILED tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g2-idtype1]
FAILED tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g3-idtype0]
FAILED tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g3-idtype1]
FAILED tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g4-idtype0]
FAILED tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g4-idtype1]
FAILED tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g5-idtype0]
FAILED tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g5-idtype1]
FAILED tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g6-idtype0]
FAILED tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g6-idtype1]
FAILED tests/compute/test_readout.py::test_weighted_reduce_readout[max-g0-idtype0]
FAILED tests/compute/test_readout.py::test_weighted_reduce_readout[max-g0-idtype1]
FAILED tests/compute/test_readout.py::test_weighted_reduce_readout[max-g1-idtype0]
FAILED tests/compute/test_readout.py::test_weighted_reduce_readout[max-g1-idtype1]
FAILED tests/compute/test_readout.py::test_weighted_reduce_readout[max-g2-idtype0]
FAILED tests/compute/test_readout.py::test_weighted_reduce_readout[max-g2-idtype1]
FAILED tests/compute/test_readout.py::test_weighted_reduce_readout[max-g3-idtype0]
FAILED tests/compute/test_readout.py::test_weighted_reduce_readout[max-g3-idtype1]
FAILED tests/compute/test_readout.py::test_weighted_reduce_readout[max-g4-idtype0]
FAILED tests/compute/test_readout.py::test_weighted_reduce_readout[max-g4-idtype1]
FAILED tests/compute/test_readout.py::test_weighted_reduce_readout[max-g5-idtype0]
FAILED tests/compute/test_readout.py::test_weighted_reduce_readout[max-g5-idtype1]
FAILED tests/compute/test_readout.py::test_weighted_reduce_readout[max-g6-idtype0]
FAILED tests/compute/test_readout.py::test_weighted_reduce_readout[max-g6-idtype1]
FAILED tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g0-idtype0]
FAILED tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g0-idtype1]
FAILED tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g1-idtype0]
FAILED tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g1-idtype1]
FAILED tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g2-idtype0]
FAILED tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g2-idtype1]
FAILED tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g3-idtype0]
FAILED tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g3-idtype1]
FAILED tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g4-idtype0]
FAILED tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g4-idtype1]
FAILED tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g5-idtype0]
FAILED tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g5-idtype1]
FAILED tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g6-idtype0]
FAILED tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g6-idtype1]
FAILED tests/compute/test_readout.py::test_softmax[g0-idtype0] - TypeError: e...
FAILED tests/compute/test_readout.py::test_softmax[g0-idtype1] - TypeError: e...
FAILED tests/compute/test_readout.py::test_softmax[g1-idtype0] - TypeError: e...
FAILED tests/compute/test_readout.py::test_softmax[g1-idtype1] - TypeError: e...
FAILED tests/compute/test_readout.py::test_softmax[g2-idtype0] - TypeError: e...
FAILED tests/compute/test_readout.py::test_softmax[g2-idtype1] - TypeError: e...
FAILED tests/compute/test_readout.py::test_softmax[g3-idtype0] - TypeError: e...
FAILED tests/compute/test_readout.py::test_softmax[g3-idtype1] - TypeError: e...
FAILED tests/compute/test_readout.py::test_softmax[g4-idtype0] - TypeError: e...
FAILED tests/compute/test_readout.py::test_softmax[g4-idtype1] - TypeError: e...
FAILED tests/compute/test_readout.py::test_softmax[g5-idtype0] - TypeError: e...
FAILED tests/compute/test_readout.py::test_softmax[g5-idtype1] - TypeError: e...
FAILED tests/compute/test_readout.py::test_softmax[g6-idtype0] - TypeError: e...
FAILED tests/compute/test_readout.py::test_softmax[g6-idtype1] - TypeError: e...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-sum-add-shp0-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-sum-add-shp0-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-sum-add-shp1-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-sum-add-shp1-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-sum-add-shp2-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-sum-add-shp2-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-sum-add-shp3-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-sum-add-shp3-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-sum-add-shp4-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-sum-add-shp4-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-sum-add-shp5-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-sum-add-shp5-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-sum-sub-shp0-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-sum-sub-shp0-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-sum-sub-shp1-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-sum-sub-shp1-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-sum-sub-shp2-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-sum-sub-shp2-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-sum-sub-shp3-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-sum-sub-shp3-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-sum-sub-shp4-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-sum-sub-shp4-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-sum-sub-shp5-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-sum-sub-shp5-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-sum-mul-shp0-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-sum-mul-shp0-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-sum-mul-shp1-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-sum-mul-shp1-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-sum-mul-shp2-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-sum-mul-shp2-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-sum-mul-shp3-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-sum-mul-shp3-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-sum-mul-shp4-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-sum-mul-shp4-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-sum-mul-shp5-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-sum-mul-shp5-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-sum-div-shp0-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-sum-div-shp0-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-sum-div-shp1-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-sum-div-shp1-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-sum-div-shp2-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-sum-div-shp2-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-sum-div-shp3-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-sum-div-shp3-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-sum-div-shp4-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-sum-div-shp4-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-sum-div-shp5-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-sum-div-shp5-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_lhs-shp0-g0]
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_lhs-shp0-g1]
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_lhs-shp1-g0]
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_lhs-shp1-g1]
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_lhs-shp2-g0]
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_lhs-shp2-g1]
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_lhs-shp3-g0]
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_lhs-shp3-g1]
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_lhs-shp4-g0]
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_lhs-shp4-g1]
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_lhs-shp5-g0]
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_lhs-shp5-g1]
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_rhs-shp0-g0]
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_rhs-shp0-g1]
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_rhs-shp1-g0]
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_rhs-shp1-g1]
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_rhs-shp2-g0]
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_rhs-shp2-g1]
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_rhs-shp3-g0]
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_rhs-shp3-g1]
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_rhs-shp4-g0]
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_rhs-shp4-g1]
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_rhs-shp5-g0]
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_rhs-shp5-g1]
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-min-add-shp0-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-min-add-shp0-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-min-add-shp1-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-min-add-shp1-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-min-add-shp2-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-min-add-shp2-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-min-add-shp3-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-min-add-shp3-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-min-add-shp4-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-min-add-shp4-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-min-add-shp5-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-min-add-shp5-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-min-sub-shp0-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-min-sub-shp0-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-min-sub-shp1-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-min-sub-shp1-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-min-sub-shp2-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-min-sub-shp2-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-min-sub-shp3-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-min-sub-shp3-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-min-sub-shp4-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-min-sub-shp4-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-min-sub-shp5-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-min-sub-shp5-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-min-mul-shp0-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-min-mul-shp0-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-min-mul-shp1-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-min-mul-shp1-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-min-mul-shp2-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-min-mul-shp2-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-min-mul-shp3-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-min-mul-shp3-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-min-mul-shp4-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-min-mul-shp4-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-min-mul-shp5-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-min-mul-shp5-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-min-div-shp0-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-min-div-shp0-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-min-div-shp1-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-min-div-shp1-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-min-div-shp2-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-min-div-shp2-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-min-div-shp3-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-min-div-shp3-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-min-div-shp4-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-min-div-shp4-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-min-div-shp5-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-min-div-shp5-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_lhs-shp0-g0]
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_lhs-shp0-g1]
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_lhs-shp1-g0]
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_lhs-shp1-g1]
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_lhs-shp2-g0]
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_lhs-shp2-g1]
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_lhs-shp3-g0]
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_lhs-shp3-g1]
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_lhs-shp4-g0]
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_lhs-shp4-g1]
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_lhs-shp5-g0]
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_lhs-shp5-g1]
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_rhs-shp0-g0]
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_rhs-shp0-g1]
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_rhs-shp1-g0]
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_rhs-shp1-g1]
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_rhs-shp2-g0]
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_rhs-shp2-g1]
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_rhs-shp3-g0]
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_rhs-shp3-g1]
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_rhs-shp4-g0]
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_rhs-shp4-g1]
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_rhs-shp5-g0]
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_rhs-shp5-g1]
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-max-add-shp0-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-max-add-shp0-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-max-add-shp1-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-max-add-shp1-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-max-add-shp2-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-max-add-shp2-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-max-add-shp3-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-max-add-shp3-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-max-add-shp4-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-max-add-shp4-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-max-add-shp5-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-max-add-shp5-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-max-sub-shp0-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-max-sub-shp0-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-max-sub-shp1-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-max-sub-shp1-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-max-sub-shp2-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-max-sub-shp2-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-max-sub-shp3-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-max-sub-shp3-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-max-sub-shp4-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-max-sub-shp4-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-max-sub-shp5-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-max-sub-shp5-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-max-mul-shp0-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-max-mul-shp0-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-max-mul-shp1-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-max-mul-shp1-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-max-mul-shp2-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-max-mul-shp2-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-max-mul-shp3-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-max-mul-shp3-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-max-mul-shp4-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-max-mul-shp4-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-max-mul-shp5-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-max-mul-shp5-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-max-div-shp0-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-max-div-shp0-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-max-div-shp1-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-max-div-shp1-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-max-div-shp2-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-max-div-shp2-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-max-div-shp3-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-max-div-shp3-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-max-div-shp4-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-max-div-shp4-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-max-div-shp5-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-max-div-shp5-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_lhs-shp0-g0]
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_lhs-shp0-g1]
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_lhs-shp1-g0]
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_lhs-shp1-g1]
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_lhs-shp2-g0]
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_lhs-shp2-g1]
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_lhs-shp3-g0]
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_lhs-shp3-g1]
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_lhs-shp4-g0]
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_lhs-shp4-g1]
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_lhs-shp5-g0]
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_lhs-shp5-g1]
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_rhs-shp0-g0]
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_rhs-shp0-g1]
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_rhs-shp1-g0]
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_rhs-shp1-g1]
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_rhs-shp2-g0]
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_rhs-shp2-g1]
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_rhs-shp3-g0]
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_rhs-shp3-g1]
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_rhs-shp4-g0]
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_rhs-shp4-g1]
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_rhs-shp5-g0]
FAILED tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_rhs-shp5-g1]
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-sum-add-shp0-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-sum-add-shp0-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-sum-add-shp1-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-sum-add-shp1-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-sum-add-shp2-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-sum-add-shp2-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-sum-add-shp3-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-sum-add-shp3-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-sum-add-shp4-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-sum-add-shp4-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-sum-add-shp5-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-sum-add-shp5-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-sum-sub-shp0-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-sum-sub-shp0-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-sum-sub-shp1-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-sum-sub-shp1-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-sum-sub-shp2-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-sum-sub-shp2-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-sum-sub-shp3-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-sum-sub-shp3-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-sum-sub-shp4-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-sum-sub-shp4-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-sum-sub-shp5-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-sum-sub-shp5-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-sum-mul-shp0-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-sum-mul-shp0-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-sum-mul-shp1-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-sum-mul-shp1-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-sum-mul-shp2-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-sum-mul-shp2-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-sum-mul-shp3-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-sum-mul-shp3-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-sum-mul-shp4-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-sum-mul-shp4-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-sum-mul-shp5-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-sum-mul-shp5-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-sum-div-shp0-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-sum-div-shp0-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-sum-div-shp1-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-sum-div-shp1-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-sum-div-shp2-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-sum-div-shp2-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-sum-div-shp3-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-sum-div-shp3-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-sum-div-shp4-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-sum-div-shp4-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-sum-div-shp5-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-sum-div-shp5-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_lhs-shp0-g0]
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_lhs-shp0-g1]
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_lhs-shp1-g0]
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_lhs-shp1-g1]
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_lhs-shp2-g0]
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_lhs-shp2-g1]
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_lhs-shp3-g0]
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_lhs-shp3-g1]
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_lhs-shp4-g0]
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_lhs-shp4-g1]
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_lhs-shp5-g0]
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_lhs-shp5-g1]
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_rhs-shp0-g0]
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_rhs-shp0-g1]
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_rhs-shp1-g0]
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_rhs-shp1-g1]
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_rhs-shp2-g0]
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_rhs-shp2-g1]
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_rhs-shp3-g0]
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_rhs-shp3-g1]
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_rhs-shp4-g0]
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_rhs-shp4-g1]
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_rhs-shp5-g0]
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_rhs-shp5-g1]
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-min-add-shp0-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-min-add-shp0-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-min-add-shp1-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-min-add-shp1-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-min-add-shp2-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-min-add-shp2-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-min-add-shp3-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-min-add-shp3-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-min-add-shp4-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-min-add-shp4-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-min-add-shp5-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-min-add-shp5-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-min-sub-shp0-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-min-sub-shp0-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-min-sub-shp1-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-min-sub-shp1-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-min-sub-shp2-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-min-sub-shp2-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-min-sub-shp3-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-min-sub-shp3-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-min-sub-shp4-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-min-sub-shp4-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-min-sub-shp5-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-min-sub-shp5-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-min-mul-shp0-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-min-mul-shp0-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-min-mul-shp1-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-min-mul-shp1-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-min-mul-shp2-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-min-mul-shp2-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-min-mul-shp3-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-min-mul-shp3-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-min-mul-shp4-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-min-mul-shp4-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-min-mul-shp5-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-min-mul-shp5-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-min-div-shp0-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-min-div-shp0-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-min-div-shp1-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-min-div-shp1-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-min-div-shp2-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-min-div-shp2-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-min-div-shp3-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-min-div-shp3-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-min-div-shp4-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-min-div-shp4-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-min-div-shp5-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-min-div-shp5-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_lhs-shp0-g0]
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_lhs-shp0-g1]
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_lhs-shp1-g0]
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_lhs-shp1-g1]
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_lhs-shp2-g0]
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_lhs-shp2-g1]
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_lhs-shp3-g0]
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_lhs-shp3-g1]
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_lhs-shp4-g0]
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_lhs-shp4-g1]
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_lhs-shp5-g0]
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_lhs-shp5-g1]
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_rhs-shp0-g0]
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_rhs-shp0-g1]
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_rhs-shp1-g0]
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_rhs-shp1-g1]
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_rhs-shp2-g0]
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_rhs-shp2-g1]
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_rhs-shp3-g0]
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_rhs-shp3-g1]
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_rhs-shp4-g0]
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_rhs-shp4-g1]
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_rhs-shp5-g0]
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_rhs-shp5-g1]
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-max-add-shp0-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-max-add-shp0-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-max-add-shp1-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-max-add-shp1-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-max-add-shp2-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-max-add-shp2-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-max-add-shp3-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-max-add-shp3-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-max-add-shp4-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-max-add-shp4-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-max-add-shp5-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-max-add-shp5-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-max-sub-shp0-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-max-sub-shp0-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-max-sub-shp1-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-max-sub-shp1-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-max-sub-shp2-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-max-sub-shp2-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-max-sub-shp3-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-max-sub-shp3-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-max-sub-shp4-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-max-sub-shp4-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-max-sub-shp5-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-max-sub-shp5-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-max-mul-shp0-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-max-mul-shp0-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-max-mul-shp1-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-max-mul-shp1-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-max-mul-shp2-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-max-mul-shp2-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-max-mul-shp3-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-max-mul-shp3-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-max-mul-shp4-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-max-mul-shp4-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-max-mul-shp5-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-max-mul-shp5-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-max-div-shp0-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-max-div-shp0-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-max-div-shp1-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-max-div-shp1-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-max-div-shp2-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-max-div-shp2-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-max-div-shp3-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-max-div-shp3-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-max-div-shp4-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-max-div-shp4-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-max-div-shp5-g0] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-max-div-shp5-g1] - Typ...
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_lhs-shp0-g0]
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_lhs-shp0-g1]
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_lhs-shp1-g0]
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_lhs-shp1-g1]
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_lhs-shp2-g0]
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_lhs-shp2-g1]
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_lhs-shp3-g0]
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_lhs-shp3-g1]
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_lhs-shp4-g0]
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_lhs-shp4-g1]
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_lhs-shp5-g0]
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_lhs-shp5-g1]
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_rhs-shp0-g0]
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_rhs-shp0-g1]
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_rhs-shp1-g0]
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_rhs-shp1-g1]
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_rhs-shp2-g0]
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_rhs-shp2-g1]
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_rhs-shp3-g0]
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_rhs-shp3-g1]
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_rhs-shp4-g0]
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_rhs-shp4-g1]
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_rhs-shp5-g0]
FAILED tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_rhs-shp5-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-v-shp0-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-v-shp0-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-v-shp1-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-v-shp1-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-v-shp2-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-v-shp2-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-v-shp3-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-v-shp3-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-v-shp4-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-v-shp4-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-e-shp0-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-e-shp0-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-e-shp1-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-e-shp1-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-e-shp2-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-e-shp2-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-e-shp3-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-e-shp3-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-e-shp4-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-e-shp4-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-u-shp0-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-u-shp0-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-u-shp1-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-u-shp1-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-u-shp2-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-u-shp2-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-u-shp3-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-u-shp3-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-u-shp4-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-u-shp4-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-e-shp0-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-e-shp0-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-e-shp1-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-e-shp1-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-e-shp2-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-e-shp2-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-e-shp3-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-e-shp3-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-e-shp4-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-e-shp4-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-u-shp0-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-u-shp0-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-u-shp1-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-u-shp1-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-u-shp2-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-u-shp2-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-u-shp3-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-u-shp3-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-u-shp4-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-u-shp4-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-v-shp0-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-v-shp0-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-v-shp1-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-v-shp1-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-v-shp2-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-v-shp2-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-v-shp3-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-v-shp3-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-v-shp4-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-v-shp4-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-v-shp0-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-v-shp0-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-v-shp1-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-v-shp1-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-v-shp2-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-v-shp2-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-v-shp3-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-v-shp3-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-v-shp4-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-v-shp4-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-e-shp0-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-e-shp0-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-e-shp1-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-e-shp1-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-e-shp2-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-e-shp2-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-e-shp3-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-e-shp3-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-e-shp4-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-e-shp4-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-u-shp0-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-u-shp0-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-u-shp1-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-u-shp1-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-u-shp2-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-u-shp2-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-u-shp3-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-u-shp3-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-u-shp4-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-u-shp4-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-e-shp0-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-e-shp0-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-e-shp1-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-e-shp1-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-e-shp2-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-e-shp2-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-e-shp3-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-e-shp3-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-e-shp4-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-e-shp4-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-u-shp0-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-u-shp0-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-u-shp1-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-u-shp1-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-u-shp2-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-u-shp2-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-u-shp3-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-u-shp3-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-u-shp4-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-u-shp4-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-v-shp0-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-v-shp0-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-v-shp1-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-v-shp1-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-v-shp2-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-v-shp2-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-v-shp3-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-v-shp3-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-v-shp4-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-v-shp4-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-v-shp0-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-v-shp0-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-v-shp1-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-v-shp1-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-v-shp2-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-v-shp2-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-v-shp3-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-v-shp3-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-v-shp4-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-v-shp4-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-e-shp0-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-e-shp0-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-e-shp1-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-e-shp1-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-e-shp2-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-e-shp2-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-e-shp3-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-e-shp3-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-e-shp4-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-e-shp4-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-u-shp0-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-u-shp0-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-u-shp1-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-u-shp1-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-u-shp2-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-u-shp2-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-u-shp3-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-u-shp3-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-u-shp4-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-u-shp4-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-e-shp0-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-e-shp0-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-e-shp1-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-e-shp1-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-e-shp2-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-e-shp2-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-e-shp3-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-e-shp3-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-e-shp4-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-e-shp4-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-u-shp0-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-u-shp0-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-u-shp1-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-u-shp1-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-u-shp2-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-u-shp2-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-u-shp3-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-u-shp3-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-u-shp4-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-u-shp4-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-v-shp0-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-v-shp0-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-v-shp1-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-v-shp1-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-v-shp2-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-v-shp2-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-v-shp3-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-v-shp3-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-v-shp4-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-v-shp4-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-v-shp0-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-v-shp0-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-v-shp1-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-v-shp1-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-v-shp2-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-v-shp2-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-v-shp3-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-v-shp3-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-v-shp4-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-v-shp4-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-e-shp0-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-e-shp0-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-e-shp1-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-e-shp1-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-e-shp2-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-e-shp2-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-e-shp3-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-e-shp3-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-e-shp4-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-e-shp4-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-u-shp0-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-u-shp0-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-u-shp1-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-u-shp1-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-u-shp2-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-u-shp2-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-u-shp3-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-u-shp3-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-u-shp4-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-u-shp4-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-e-shp0-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-e-shp0-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-e-shp1-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-e-shp1-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-e-shp2-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-e-shp2-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-e-shp3-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-e-shp3-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-e-shp4-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-e-shp4-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-u-shp0-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-u-shp0-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-u-shp1-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-u-shp1-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-u-shp2-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-u-shp2-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-u-shp3-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-u-shp3-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-u-shp4-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-u-shp4-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-v-shp0-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-v-shp0-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-v-shp1-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-v-shp1-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-v-shp2-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-v-shp2-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-v-shp3-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-v-shp3-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-v-shp4-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-v-shp4-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-v-shp0-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-v-shp0-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-v-shp1-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-v-shp1-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-v-shp2-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-v-shp2-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-v-shp3-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-v-shp3-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-v-shp4-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-v-shp4-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-e-shp0-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-e-shp0-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-e-shp1-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-e-shp1-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-e-shp2-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-e-shp2-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-e-shp3-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-e-shp3-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-e-shp4-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-e-shp4-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-u-shp0-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-u-shp0-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-u-shp1-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-u-shp1-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-u-shp2-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-u-shp2-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-u-shp3-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-u-shp3-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-u-shp4-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-u-shp4-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-e-shp0-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-e-shp0-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-e-shp1-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-e-shp1-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-e-shp2-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-e-shp2-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-e-shp3-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-e-shp3-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-e-shp4-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-e-shp4-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-u-shp0-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-u-shp0-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-u-shp1-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-u-shp1-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-u-shp2-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-u-shp2-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-u-shp3-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-u-shp3-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-u-shp4-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-u-shp4-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-v-shp0-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-v-shp0-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-v-shp1-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-v-shp1-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-v-shp2-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-v-shp2-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-v-shp3-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-v-shp3-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-v-shp4-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-v-shp4-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-v-shp0-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-v-shp0-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-v-shp1-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-v-shp1-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-v-shp2-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-v-shp2-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-v-shp3-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-v-shp3-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-v-shp4-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-v-shp4-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-e-shp0-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-e-shp0-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-e-shp1-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-e-shp1-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-e-shp2-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-e-shp2-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-e-shp3-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-e-shp3-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-e-shp4-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-e-shp4-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-u-shp0-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-u-shp0-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-u-shp1-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-u-shp1-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-u-shp2-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-u-shp2-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-u-shp3-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-u-shp3-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-u-shp4-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-u-shp4-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-e-shp0-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-e-shp0-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-e-shp1-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-e-shp1-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-e-shp2-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-e-shp2-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-e-shp3-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-e-shp3-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-e-shp4-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-e-shp4-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-u-shp0-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-u-shp0-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-u-shp1-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-u-shp1-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-u-shp2-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-u-shp2-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-u-shp3-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-u-shp3-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-u-shp4-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-u-shp4-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-v-shp0-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-v-shp0-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-v-shp1-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-v-shp1-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-v-shp2-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-v-shp2-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-v-shp3-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-v-shp3-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-v-shp4-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-v-shp4-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-v-shp0-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-v-shp0-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-v-shp1-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-v-shp1-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-v-shp2-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-v-shp2-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-v-shp3-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-v-shp3-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-v-shp4-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-v-shp4-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-e-shp0-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-e-shp0-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-e-shp1-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-e-shp1-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-e-shp2-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-e-shp2-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-e-shp3-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-e-shp3-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-e-shp4-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-e-shp4-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-u-shp0-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-u-shp0-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-u-shp1-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-u-shp1-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-u-shp2-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-u-shp2-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-u-shp3-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-u-shp3-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-u-shp4-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-u-shp4-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-e-shp0-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-e-shp0-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-e-shp1-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-e-shp1-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-e-shp2-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-e-shp2-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-e-shp3-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-e-shp3-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-e-shp4-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-e-shp4-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-u-shp0-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-u-shp0-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-u-shp1-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-u-shp1-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-u-shp2-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-u-shp2-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-u-shp3-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-u-shp3-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-u-shp4-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-u-shp4-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-v-shp0-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-v-shp0-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-v-shp1-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-v-shp1-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-v-shp2-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-v-shp2-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-v-shp3-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-v-shp3-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-v-shp4-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-v-shp4-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-v-shp0-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-v-shp0-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-v-shp1-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-v-shp1-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-v-shp2-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-v-shp2-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-v-shp3-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-v-shp3-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-v-shp4-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-v-shp4-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-e-shp0-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-e-shp0-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-e-shp1-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-e-shp1-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-e-shp2-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-e-shp2-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-e-shp3-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-e-shp3-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-e-shp4-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-e-shp4-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-u-shp0-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-u-shp0-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-u-shp1-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-u-shp1-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-u-shp2-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-u-shp2-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-u-shp3-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-u-shp3-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-u-shp4-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-u-shp4-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-e-shp0-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-e-shp0-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-e-shp1-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-e-shp1-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-e-shp2-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-e-shp2-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-e-shp3-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-e-shp3-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-e-shp4-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-e-shp4-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-u-shp0-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-u-shp0-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-u-shp1-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-u-shp1-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-u-shp2-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-u-shp2-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-u-shp3-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-u-shp3-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-u-shp4-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-u-shp4-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-v-shp0-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-v-shp0-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-v-shp1-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-v-shp1-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-v-shp2-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-v-shp2-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-v-shp3-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-v-shp3-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-v-shp4-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-v-shp4-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-v-shp0-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-v-shp0-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-v-shp1-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-v-shp1-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-v-shp2-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-v-shp2-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-v-shp3-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-v-shp3-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-v-shp4-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-v-shp4-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-e-shp0-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-e-shp0-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-e-shp1-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-e-shp1-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-e-shp2-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-e-shp2-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-e-shp3-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-e-shp3-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-e-shp4-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-e-shp4-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-u-shp0-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-u-shp0-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-u-shp1-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-u-shp1-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-u-shp2-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-u-shp2-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-u-shp3-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-u-shp3-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-u-shp4-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-u-shp4-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-e-shp0-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-e-shp0-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-e-shp1-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-e-shp1-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-e-shp2-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-e-shp2-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-e-shp3-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-e-shp3-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-e-shp4-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-e-shp4-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-u-shp0-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-u-shp0-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-u-shp1-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-u-shp1-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-u-shp2-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-u-shp2-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-u-shp3-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-u-shp3-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-u-shp4-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-u-shp4-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-v-shp0-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-v-shp0-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-v-shp1-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-v-shp1-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-v-shp2-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-v-shp2-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-v-shp3-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-v-shp3-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-v-shp4-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-v-shp4-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-v-shp0-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-v-shp0-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-v-shp1-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-v-shp1-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-v-shp2-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-v-shp2-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-v-shp3-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-v-shp3-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-v-shp4-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-v-shp4-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-e-shp0-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-e-shp0-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-e-shp1-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-e-shp1-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-e-shp2-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-e-shp2-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-e-shp3-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-e-shp3-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-e-shp4-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-e-shp4-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-u-shp0-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-u-shp0-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-u-shp1-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-u-shp1-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-u-shp2-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-u-shp2-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-u-shp3-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-u-shp3-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-u-shp4-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-u-shp4-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-e-shp0-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-e-shp0-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-e-shp1-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-e-shp1-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-e-shp2-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-e-shp2-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-e-shp3-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-e-shp3-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-e-shp4-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-e-shp4-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-u-shp0-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-u-shp0-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-u-shp1-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-u-shp1-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-u-shp2-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-u-shp2-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-u-shp3-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-u-shp3-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-u-shp4-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-u-shp4-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-v-shp0-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-v-shp0-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-v-shp1-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-v-shp1-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-v-shp2-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-v-shp2-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-v-shp3-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-v-shp3-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-v-shp4-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-v-shp4-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-v-shp0-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-v-shp0-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-v-shp1-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-v-shp1-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-v-shp2-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-v-shp2-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-v-shp3-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-v-shp3-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-v-shp4-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-v-shp4-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-e-shp0-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-e-shp0-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-e-shp1-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-e-shp1-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-e-shp2-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-e-shp2-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-e-shp3-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-e-shp3-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-e-shp4-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-e-shp4-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-u-shp0-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-u-shp0-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-u-shp1-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-u-shp1-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-u-shp2-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-u-shp2-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-u-shp3-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-u-shp3-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-u-shp4-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-u-shp4-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-e-shp0-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-e-shp0-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-e-shp1-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-e-shp1-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-e-shp2-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-e-shp2-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-e-shp3-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-e-shp3-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-e-shp4-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-e-shp4-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-u-shp0-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-u-shp0-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-u-shp1-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-u-shp1-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-u-shp2-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-u-shp2-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-u-shp3-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-u-shp3-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-u-shp4-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-u-shp4-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-v-shp0-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-v-shp0-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-v-shp1-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-v-shp1-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-v-shp2-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-v-shp2-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-v-shp3-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-v-shp3-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-v-shp4-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-v-shp4-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-v-shp0-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-v-shp0-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-v-shp1-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-v-shp1-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-v-shp2-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-v-shp2-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-v-shp3-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-v-shp3-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-v-shp4-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-v-shp4-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-e-shp0-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-e-shp0-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-e-shp1-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-e-shp1-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-e-shp2-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-e-shp2-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-e-shp3-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-e-shp3-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-e-shp4-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-e-shp4-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-u-shp0-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-u-shp0-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-u-shp1-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-u-shp1-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-u-shp2-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-u-shp2-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-u-shp3-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-u-shp3-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-u-shp4-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-u-shp4-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-e-shp0-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-e-shp0-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-e-shp1-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-e-shp1-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-e-shp2-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-e-shp2-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-e-shp3-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-e-shp3-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-e-shp4-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-e-shp4-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-u-shp0-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-u-shp0-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-u-shp1-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-u-shp1-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-u-shp2-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-u-shp2-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-u-shp3-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-u-shp3-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-u-shp4-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-u-shp4-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-v-shp0-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-v-shp0-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-v-shp1-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-v-shp1-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-v-shp2-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-v-shp2-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-v-shp3-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-v-shp3-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-v-shp4-g0] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-v-shp4-g1] - Ty...
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-v-shp0-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-v-shp0-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-v-shp1-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-v-shp1-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-v-shp2-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-v-shp2-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-v-shp3-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-v-shp3-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-v-shp4-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-v-shp4-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-e-shp0-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-e-shp0-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-e-shp1-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-e-shp1-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-e-shp2-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-e-shp2-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-e-shp3-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-e-shp3-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-e-shp4-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-e-shp4-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-u-shp0-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-u-shp0-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-u-shp1-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-u-shp1-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-u-shp2-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-u-shp2-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-u-shp3-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-u-shp3-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-u-shp4-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-u-shp4-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-e-shp0-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-e-shp0-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-e-shp1-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-e-shp1-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-e-shp2-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-e-shp2-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-e-shp3-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-e-shp3-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-e-shp4-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-e-shp4-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-u-shp0-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-u-shp0-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-u-shp1-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-u-shp1-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-u-shp2-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-u-shp2-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-u-shp3-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-u-shp3-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-u-shp4-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-u-shp4-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-v-shp0-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-v-shp0-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-v-shp1-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-v-shp1-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-v-shp2-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-v-shp2-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-v-shp3-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-v-shp3-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-v-shp4-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-v-shp4-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-v-shp0-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-v-shp0-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-v-shp1-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-v-shp1-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-v-shp2-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-v-shp2-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-v-shp3-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-v-shp3-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-v-shp4-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-v-shp4-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-e-shp0-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-e-shp0-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-e-shp1-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-e-shp1-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-e-shp2-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-e-shp2-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-e-shp3-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-e-shp3-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-e-shp4-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-e-shp4-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-u-shp0-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-u-shp0-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-u-shp1-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-u-shp1-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-u-shp2-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-u-shp2-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-u-shp3-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-u-shp3-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-u-shp4-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-u-shp4-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-e-shp0-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-e-shp0-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-e-shp1-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-e-shp1-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-e-shp2-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-e-shp2-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-e-shp3-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-e-shp3-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-e-shp4-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-e-shp4-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-u-shp0-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-u-shp0-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-u-shp1-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-u-shp1-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-u-shp2-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-u-shp2-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-u-shp3-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-u-shp3-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-u-shp4-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-u-shp4-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-v-shp0-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-v-shp0-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-v-shp1-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-v-shp1-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-v-shp2-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-v-shp2-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-v-shp3-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-v-shp3-g1]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-v-shp4-g0]
FAILED tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-v-shp4-g1]
FAILED tests/compute/test_sparse.py::test_edge_softmax[idtype0-shp0-src-g0]
FAILED tests/compute/test_sparse.py::test_edge_softmax[idtype0-shp0-dst-g0]
FAILED tests/compute/test_sparse.py::test_edge_softmax[idtype0-shp1-src-g0]
FAILED tests/compute/test_sparse.py::test_edge_softmax[idtype0-shp1-dst-g0]
FAILED tests/compute/test_sparse.py::test_edge_softmax[idtype0-shp2-src-g0]
FAILED tests/compute/test_sparse.py::test_edge_softmax[idtype0-shp2-dst-g0]
FAILED tests/compute/test_sparse.py::test_edge_softmax[idtype1-shp0-src-g0]
FAILED tests/compute/test_sparse.py::test_edge_softmax[idtype1-shp0-dst-g0]
FAILED tests/compute/test_sparse.py::test_edge_softmax[idtype1-shp1-src-g0]
FAILED tests/compute/test_sparse.py::test_edge_softmax[idtype1-shp1-dst-g0]
FAILED tests/compute/test_sparse.py::test_edge_softmax[idtype1-shp2-src-g0]
FAILED tests/compute/test_sparse.py::test_edge_softmax[idtype1-shp2-dst-g0]
FAILED tests/compute/test_sparse.py::test_segment_reduce[sum] - TypeError: em...
FAILED tests/compute/test_sparse.py::test_segment_reduce[max] - TypeError: em...
FAILED tests/compute/test_sparse.py::test_segment_reduce[min] - TypeError: em...
FAILED tests/compute/test_sparse.py::test_segment_reduce[mean] - TypeError: e...
FAILED tests/compute/test_sparse.py::test_segment_mm[dtype2-0.003-1-idtype0]
FAILED tests/compute/test_sparse.py::test_segment_mm[dtype2-0.003-1-idtype1]
FAILED tests/compute/test_sparse.py::test_segment_mm[dtype2-0.003-8-idtype0]
FAILED tests/compute/test_sparse.py::test_segment_mm[dtype2-0.003-8-idtype1]
FAILED tests/compute/test_sparse.py::test_segment_mm[dtype2-0.003-16-idtype0]
FAILED tests/compute/test_sparse.py::test_segment_mm[dtype2-0.003-16-idtype1]
FAILED tests/compute/test_sparse.py::test_segment_mm[dtype2-0.003-64-idtype0]
FAILED tests/compute/test_sparse.py::test_segment_mm[dtype2-0.003-64-idtype1]
FAILED tests/compute/test_sparse.py::test_segment_mm[dtype2-0.003-256-idtype0]
FAILED tests/compute/test_sparse.py::test_segment_mm[dtype2-0.003-256-idtype1]
FAILED tests/compute/test_sparse.py::test_segment_mm[dtype3-0.0001-1-idtype0]
FAILED tests/compute/test_sparse.py::test_segment_mm[dtype3-0.0001-1-idtype1]
FAILED tests/compute/test_sparse.py::test_segment_mm[dtype3-0.0001-8-idtype0]
FAILED tests/compute/test_sparse.py::test_segment_mm[dtype3-0.0001-8-idtype1]
FAILED tests/compute/test_sparse.py::test_segment_mm[dtype3-0.0001-16-idtype0]
FAILED tests/compute/test_sparse.py::test_segment_mm[dtype3-0.0001-16-idtype1]
FAILED tests/compute/test_sparse.py::test_segment_mm[dtype3-0.0001-64-idtype0]
FAILED tests/compute/test_sparse.py::test_segment_mm[dtype3-0.0001-64-idtype1]
FAILED tests/compute/test_sparse.py::test_segment_mm[dtype3-0.0001-256-idtype0]
FAILED tests/compute/test_sparse.py::test_segment_mm[dtype3-0.0001-256-idtype1]
FAILED tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype2-0.003-1] - A...
FAILED tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype2-0.003-8] - A...
FAILED tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype2-0.003-16] - ...
FAILED tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype2-0.003-64] - ...
FAILED tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype2-0.003-256]
FAILED tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype3-0.0001-1] - ...
FAILED tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype3-0.0001-8] - ...
FAILED tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype3-0.0001-16]
FAILED tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype3-0.0001-64]
FAILED tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype3-0.0001-256]
FAILED tests/compute/test_sparse.py::test_use_libxsmm_switch - TypeError: emp...
FAILED tests/compute/test_specialization.py::test_v2v_update_all[idtype0] - T...
FAILED tests/compute/test_specialization.py::test_v2v_update_all[idtype1] - T...
FAILED tests/compute/test_specialization.py::test_v2v_snr[idtype0] - TypeErro...
FAILED tests/compute/test_specialization.py::test_v2v_snr[idtype1] - TypeErro...
FAILED tests/compute/test_specialization.py::test_v2v_pull[idtype0] - TypeErr...
FAILED tests/compute/test_specialization.py::test_v2v_pull[idtype1] - TypeErr...
FAILED tests/compute/test_specialization.py::test_update_all_multi_fallback[idtype0]
FAILED tests/compute/test_specialization.py::test_update_all_multi_fallback[idtype1]
FAILED tests/compute/test_specialization.py::test_pull_multi_fallback[idtype0]
FAILED tests/compute/test_specialization.py::test_pull_multi_fallback[idtype1]
FAILED tests/compute/test_specialization.py::test_spmv_3d_feat[idtype0] - Typ...
FAILED tests/compute/test_specialization.py::test_spmv_3d_feat[idtype1] - Typ...
FAILED tests/compute/test_transform.py::test_khop_graph - TypeError: empty_co...
FAILED tests/compute/test_transform.py::test_khop_adj - TypeError: empty_cont...
FAILED tests/compute/test_transform.py::test_to_simple[idtype0] - TypeError: ...
FAILED tests/compute/test_transform.py::test_to_simple[idtype1] - TypeError: ...
FAILED tests/compute/test_transform.py::test_remove_edges[idtype0] - TypeErro...
FAILED tests/compute/test_transform.py::test_remove_edges[idtype1] - TypeErro...
FAILED tests/compute/test_transform.py::test_remove_nodes[idtype0] - TypeErro...
FAILED tests/compute/test_transform.py::test_remove_nodes[idtype1] - TypeErro...
FAILED tests/compute/test_transform.py::test_add_selfloop[idtype0] - TypeErro...
FAILED tests/compute/test_transform.py::test_add_selfloop[idtype1] - TypeErro...
FAILED tests/compute/test_transform.py::test_remove_selfloop[idtype0] - TypeE...
FAILED tests/compute/test_transform.py::test_remove_selfloop[idtype1] - TypeE...
FAILED tests/compute/test_transform.py::test_module_gcnnorm[idtype0] - TypeEr...
FAILED tests/compute/test_transform.py::test_module_gcnnorm[idtype1] - TypeEr...
FAILED tests/compute/test_transform.py::test_module_sign[g0] - TypeError: emp...
==== 1474 failed, 911 passed, 73 skipped, 322 warnings in 566.50s (0:09:26) ====
Server (0) shutdown.
Server is exiting...
Server (1) shutdown.
Server is exiting...
Server (2) shutdown.
Server is exiting...
FAIL: compute
PASSED [ 38%]
tests/distributed/test_mp_dataloader.py::test_dist_dataloader[1-True-False-4-3]   NumNodes: 2708
  NumEdges: 10556
  NumFeats: 1433
  NumClasses: 7
  NumTrainingSamples: 140
  NumValidationSamples: 500
  NumTestSamples: 1000
Done loading data from cached files.
torch.int64
Converting to homogeneous graph takes 0.001s, peak mem: 1.529 GB
Convert a graph into a bidirected graph: 0.001 seconds, peak memory: 1.529 GB
Construct multi-constraint weights: 0.000 seconds, peak memory: 1.529 GB
[05:33:51] /root/jenkins/workspace/dgl_PR-4648/src/graph/transform/metis_partition_hetero.cc:87: Partition a graph with 2708 nodes and 10556 edges into 3 parts and get 325 edge cuts
Metis partitioning: 0.002 seconds, peak memory: 1.529 GB
Assigning nodes to METIS partitions takes 0.003s, peak mem: 1.529 GB
Reshuffle nodes and edges: 0.001 seconds
Split the graph: 0.002 seconds
Construct subgraphs: 0.003 seconds
Splitting the graph into partitions takes 0.007s, peak mem: 1.529 GB
part 0 has 1046 nodes and 903 are inside the partition
part 0 has 3490 edges and 3317 are inside the partition
part 1 has 1064 nodes and 902 are inside the partition
part 1 has 4006 edges and 3739 are inside the partition
part 2 has 1056 nodes and 903 are inside the partition
part 2 has 3710 edges and 3500 are inside the partition
Save partitions: 0.018 seconds, peak memory: 1.529 GB
There are 10556 edges in the graph and 0 edge cuts for 3 partitions.
[Pipeline] }
[Pipeline] // timeout
[Pipeline] }
[Pipeline] // stage
[Pipeline] stage
[Pipeline] { (Torch CPU Example test)
Stage "Torch CPU Example test" skipped due to earlier failure(s)
[Pipeline] }
[Pipeline] // stage
[Pipeline] stage
[Pipeline] { (Torch CPU Tutorial test)
Stage "Torch CPU Tutorial test" skipped due to earlier failure(s)
[Pipeline] }
[Pipeline] // stage
Post stage
[Pipeline] cleanWs
[WS-CLEANUP] Deleting project workspace...
[WS-CLEANUP] Deferred wipeout is disabled by the job configuration...
server: #clients=5
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_sampling
start graph service on server 0 for part 0
[05:33:52] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:33:52] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:15094]...
server: #clients=5
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_sampling
start graph service on server 1 for part 1
[05:33:53] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:33:53] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:15097]...
[WS-CLEANUP] done
[Pipeline] }
$ docker stop --time=1 af5aa5eeb7433ab669abbd955ede40baf0b228908472383bbed72d33561f7092
server: #clients=5
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_sampling
start graph service on server 2 for part 2
[05:33:54] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:33:54] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:15100]...
[05:33:55] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:33:55] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
$ docker rm -f af5aa5eeb7433ab669abbd955ede40baf0b228908472383bbed72d33561f7092
[Pipeline] // withDockerContainer
[Pipeline] }
[Pipeline] // withEnv
[Pipeline] }
Running on dgl-manual-large-cpu in /root/jenkins/workspace/dgl_PR-4648@3
[Pipeline] // node
[Pipeline] {
[Pipeline] }
[Pipeline] // stage
[Pipeline] }
Failed in branch Torch CPU
[Pipeline] checkout
The recommended git tool is: git
using credential 150de63f-189c-4717-bcaf-010460d2f51a
Cloning the remote Git repository
Cloning with configured refspecs honoured and without tags
[05:33:56] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:33:56] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
[05:33:56] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:33:56] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
[05:33:56] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:33:56] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
[05:33:56] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:33:56] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Cloning repository https://github.com/dmlc/dgl.git
 > git init /root/jenkins/workspace/dgl_PR-4648@3 # timeout=10
Fetching upstream changes from https://github.com/dmlc/dgl.git
 > git --version # timeout=10
 > git --version # 'git version 2.17.1'
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git fetch --no-tags --progress -- https://github.com/dmlc/dgl.git +refs/pull/4648/head:refs/remotes/origin/PR-4648 +refs/heads/master:refs/remotes/origin/master # timeout=10
Cleaning workspace
Fetching without tags
Merging remotes/origin/master commit d78a3a4baf611b90871a849f58647160c7cd9ab4 into PR head commit 37ed78035f406940ed88a4b60cadd2b7cfde5fea
Merge succeeded, producing 37ed78035f406940ed88a4b60cadd2b7cfde5fea
Checking out Revision 37ed78035f406940ed88a4b60cadd2b7cfde5fea (PR-4648)
Client [1890] waits on 172.17.0.3:60767
Client [1892] waits on 172.17.0.3:58815
Client [1889] waits on 172.17.0.3:33889
Client [1885] waits on 172.17.0.3:59449
Client [1891] waits on 172.17.0.3:42885
Machine (0) group (0) client (0) connect to server successfuly!
Machine (0) group (0) client (4) connect to server successfuly!
Machine (0) group (0) client (2) connect to server successfuly!
Machine (0) group (0) client (1) connect to server successfuly!
Machine (0) group (0) client (3) connect to server successfuly!
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
  NumNodes: 2708
  NumEdges: 10556
  NumFeats: 1433
  NumClasses: 7
  NumTrainingSamples: 140
  NumValidationSamples: 500
  NumTestSamples: 1000
Done loading data from cached files.
Commit message: "fix for pytorch < 1.12"
Cleaning workspace
  NumNodes: 2708
  NumEdges: 10556
  NumFeats: 1433
  NumClasses: 7
  NumTrainingSamples: 140
  NumValidationSamples: 500
  NumTestSamples: 1000
Done loading data from cached files.
Client[3] in group[0] is exiting...
Client[0] in group[0] is exiting...
 > git config remote.origin.url https://github.com/dmlc/dgl.git # timeout=10
 > git config --add remote.origin.fetch +refs/pull/4648/head:refs/remotes/origin/PR-4648 # timeout=10
 > git config --add remote.origin.fetch +refs/heads/master:refs/remotes/origin/master # timeout=10
 > git config remote.origin.url https://github.com/dmlc/dgl.git # timeout=10
 > git rev-parse --verify HEAD # timeout=10
No valid HEAD. Skipping the resetting
 > git clean -fdx # timeout=10
Fetching upstream changes from https://github.com/dmlc/dgl.git
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git fetch --no-tags --progress -- https://github.com/dmlc/dgl.git +refs/pull/4648/head:refs/remotes/origin/PR-4648 +refs/heads/master:refs/remotes/origin/master # timeout=10
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git remote # timeout=10
 > git config --get remote.origin.url # timeout=10
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git merge d78a3a4baf611b90871a849f58647160c7cd9ab4 # timeout=10
 > git rev-parse HEAD^{commit} # timeout=10
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git branch -a -v --no-abbrev # timeout=10
 > git checkout -b PR-4648 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git rev-parse --verify HEAD # timeout=10
Resetting working tree
 > git reset --hard # timeout=10
 > git clean -fdx # timeout=10
[Pipeline] withEnv
[Pipeline] {
[Pipeline] isUnix
[Pipeline] withEnv
[Pipeline] {
[Pipeline] sh
Client[4] in group[0] is exiting...
+ docker pull dgllib/dgl-ci-cpu:v220816
Client[1] in group[0] is exiting...
Client[2] in group[0] is exiting...
v220816: Pulling from dgllib/dgl-ci-cpu
Digest: sha256:64b385c33b44dc57cb96ff264a84d8dfb8ced0caa9b30fbc4cec6d5ee511b099
Status: Image is up to date for dgllib/dgl-ci-cpu:v220816
docker.io/dgllib/dgl-ci-cpu:v220816
[Pipeline] }
[Pipeline] // withEnv
[Pipeline] isUnix
[Pipeline] withEnv
[Pipeline] {
[Pipeline] sh
Server (0) shutdown.
Server is exiting...
Server (1) shutdown.
Server is exiting...
+ docker inspect -f . dgllib/dgl-ci-cpu:v220816
.
[Pipeline] }
[Pipeline] // withEnv
[Pipeline] withDockerContainer
dgl-manual-large-cpu does not seem to be running inside a container
$ docker run -t -d -u 0:0 -w /root/jenkins/workspace/dgl_PR-4648@3 -v /root/jenkins/workspace/dgl_PR-4648@3:/root/jenkins/workspace/dgl_PR-4648@3:rw,z -v /root/jenkins/workspace/dgl_PR-4648@3@tmp:/root/jenkins/workspace/dgl_PR-4648@3@tmp:rw,z -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** dgllib/dgl-ci-cpu:v220816 cat
$ docker top 39027e0569c91f60d197e31fb8f36454bc8a908bd28bdc225aa7c1adb21f2213 -eo pid,comm
[Pipeline] {
[Pipeline] stage
[Pipeline] { (DGL-Go CPU test)
[Pipeline] sh
Server (2) shutdown.
Server is exiting...
+ rm -rf CMakeLists.txt CONTRIBUTORS.md Jenkinsfile LICENSE NEWS.md README.md apps benchmarks cmake conda dglgo docker docs examples featgraph include pyproject.toml python readthedocs.yml src tensoradapter tests third_party tools tutorials
[Pipeline] checkout
The recommended git tool is: git
using credential 150de63f-189c-4717-bcaf-010460d2f51a
Warning: JENKINS-30600: special launcher org.jenkinsci.plugins.docker.workflow.WithContainerStep$Decorator$1@3255d550; decorates RemoteLauncher[hudson.remoting.Channel@1d9b9638:dgl-manual-large-cpu] will be ignored (a typical symptom is the Git executable not being run inside a designated container)
Fetching changes from the remote Git repository
Cleaning workspace
Fetching without tags
Merging remotes/origin/master commit d78a3a4baf611b90871a849f58647160c7cd9ab4 into PR head commit 37ed78035f406940ed88a4b60cadd2b7cfde5fea
Merge succeeded, producing 37ed78035f406940ed88a4b60cadd2b7cfde5fea
Checking out Revision 37ed78035f406940ed88a4b60cadd2b7cfde5fea (PR-4648)
Commit message: "fix for pytorch < 1.12"
Cleaning workspace
 > git rev-parse --resolve-git-dir /root/jenkins/workspace/dgl_PR-4648@3/.git # timeout=10
 > git config remote.origin.url https://github.com/dmlc/dgl.git # timeout=10
 > git rev-parse --verify HEAD # timeout=10
Resetting working tree
 > git reset --hard # timeout=10
 > git clean -fdx # timeout=10
Fetching upstream changes from https://github.com/dmlc/dgl.git
 > git --version # timeout=10
 > git --version # 'git version 2.17.1'
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git fetch --no-tags --progress -- https://github.com/dmlc/dgl.git +refs/pull/4648/head:refs/remotes/origin/PR-4648 +refs/heads/master:refs/remotes/origin/master # timeout=10
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git remote # timeout=10
 > git config --get remote.origin.url # timeout=10
using GIT_ASKPASS to set credentials dgl-bot-personal-access-key220906
 > git merge d78a3a4baf611b90871a849f58647160c7cd9ab4 # timeout=10
 > git rev-parse HEAD^{commit} # timeout=10
 > git config core.sparsecheckout # timeout=10
 > git checkout -f 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git branch -a -v --no-abbrev # timeout=10
 > git branch -D PR-4648 # timeout=10
 > git checkout -b PR-4648 37ed78035f406940ed88a4b60cadd2b7cfde5fea # timeout=10
 > git rev-parse --verify HEAD # timeout=10
Resetting working tree
 > git reset --hard # timeout=10
PASSED [ 40%]
tests/distributed/test_mp_dataloader.py::test_dist_dataloader[1-False-True-0-3]   NumNodes: 2708
  NumEdges: 10556
  NumFeats: 1433
  NumClasses: 7
  NumTrainingSamples: 140
  NumValidationSamples: 500
  NumTestSamples: 1000
Done loading data from cached files.
torch.int64
Converting to homogeneous graph takes 0.001s, peak mem: 1.529 GB
Convert a graph into a bidirected graph: 0.001 seconds, peak memory: 1.529 GB
Construct multi-constraint weights: 0.000 seconds, peak memory: 1.529 GB
[05:34:05] /root/jenkins/workspace/dgl_PR-4648/src/graph/transform/metis_partition_hetero.cc:87: Partition a graph with 2708 nodes and 10556 edges into 3 parts and get 325 edge cuts
Metis partitioning: 0.002 seconds, peak memory: 1.529 GB
Assigning nodes to METIS partitions takes 0.003s, peak mem: 1.529 GB
Split the graph: 0.002 seconds
Construct subgraphs: 0.001 seconds
Splitting the graph into partitions takes 0.003s, peak mem: 1.529 GB
Calculate edge assignment: 0.000 seconds
part 0 has 1046 nodes and 903 are inside the partition
part 0 has 3490 edges and 3317 are inside the partition
part 1 has 1064 nodes and 902 are inside the partition
part 1 has 4006 edges and 3739 are inside the partition
part 2 has 1056 nodes and 903 are inside the partition
part 2 has 3710 edges and 3500 are inside the partition
Save partitions: 0.015 seconds, peak memory: 1.529 GB
There are 10556 edges in the graph and 0 edge cuts for 3 partitions.
[Pipeline] sh
+ git submodule update --recursive --init
 > git clean -fdx # timeout=10
Submodule 'third_party/METIS' (https://github.com/KarypisLab/METIS.git) registered for path 'third_party/METIS'
Submodule 'third_party/dlpack' (https://github.com/dmlc/dlpack.git) registered for path 'third_party/dlpack'
Submodule 'third_party/dmlc-core' (https://github.com/dmlc/dmlc-core.git) registered for path 'third_party/dmlc-core'
Submodule 'third_party/googletest' (https://github.com/google/googletest.git) registered for path 'third_party/googletest'
Submodule 'third_party/libxsmm' (https://github.com/hfp/libxsmm.git) registered for path 'third_party/libxsmm'
Submodule 'third_party/nanoflann' (https://github.com/jlblancoc/nanoflann) registered for path 'third_party/nanoflann'
Submodule 'third_party/nccl' (https://github.com/nvidia/nccl) registered for path 'third_party/nccl'
Submodule 'third_party/phmap' (https://github.com/greg7mdp/parallel-hashmap.git) registered for path 'third_party/phmap'
Submodule 'third_party/tensorpipe' (https://github.com/pytorch/tensorpipe) registered for path 'third_party/tensorpipe'
Submodule 'third_party/thrust' (https://github.com/NVIDIA/thrust.git) registered for path 'third_party/thrust'
Submodule 'third_party/tvm' (https://github.com/apache/incubator-tvm) registered for path 'third_party/tvm'
Submodule 'third_party/xbyak' (https://github.com/herumi/xbyak) registered for path 'third_party/xbyak'
Cloning into '/root/jenkins/workspace/dgl_PR-4648@3/third_party/METIS'...
server: #clients=1
load test_sampling
start graph service on server 0 for part 0
[05:34:05] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:34:05] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:17055]...
Cloning into '/root/jenkins/workspace/dgl_PR-4648@3/third_party/dlpack'...
server: #clients=1
load test_sampling
start graph service on server 1 for part 1
[05:34:07] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:34:07] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:17058]...
Cloning into '/root/jenkins/workspace/dgl_PR-4648@3/third_party/dmlc-core'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648@3/third_party/googletest'...
server: #clients=1
load test_sampling
start graph service on server 2 for part 2
[05:34:08] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:34:08] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:17061]...
[05:34:08] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:34:08] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Cloning into '/root/jenkins/workspace/dgl_PR-4648@3/third_party/libxsmm'...
Client [1991] waits on 172.17.0.3:52371
Machine (0) group (0) client (0) connect to server successfuly!
  NumNodes: 2708
  NumEdges: 10556
  NumFeats: 1433
  NumClasses: 7
  NumTrainingSamples: 140
  NumValidationSamples: 500
  NumTestSamples: 1000
Done loading data from cached files.
  NumNodes: 2708
  NumEdges: 10556
  NumFeats: 1433
  NumClasses: 7
  NumTrainingSamples: 140
  NumValidationSamples: 500
  NumTestSamples: 1000
Done loading data from cached files.
Client[0] in group[0] is exiting...
Server (2) shutdown.
Server is exiting...
Server (0) shutdown.
Server is exiting...
Server (1) shutdown.
Server is exiting...
PASSED [ 41%]
tests/distributed/test_mp_dataloader.py::test_dist_dataloader[1-False-True-4-3]   NumNodes: 2708
  NumEdges: 10556
  NumFeats: 1433
  NumClasses: 7
  NumTrainingSamples: 140
  NumValidationSamples: 500
  NumTestSamples: 1000
Done loading data from cached files.
torch.int64
Converting to homogeneous graph takes 0.001s, peak mem: 1.529 GB
Convert a graph into a bidirected graph: 0.001 seconds, peak memory: 1.529 GB
Construct multi-constraint weights: 0.000 seconds, peak memory: 1.529 GB
[05:34:13] /root/jenkins/workspace/dgl_PR-4648/src/graph/transform/metis_partition_hetero.cc:87: Partition a graph with 2708 nodes and 10556 edges into 3 parts and get 325 edge cuts
Metis partitioning: 0.003 seconds, peak memory: 1.529 GB
Assigning nodes to METIS partitions takes 0.004s, peak mem: 1.529 GB
Split the graph: 0.003 seconds
Construct subgraphs: 0.002 seconds
Splitting the graph into partitions takes 0.004s, peak mem: 1.529 GB
Calculate edge assignment: 0.000 seconds
part 0 has 1046 nodes and 903 are inside the partition
part 0 has 3490 edges and 3317 are inside the partition
part 1 has 1064 nodes and 902 are inside the partition
part 1 has 4006 edges and 3739 are inside the partition
part 2 has 1056 nodes and 903 are inside the partition
part 2 has 3710 edges and 3500 are inside the partition
Save partitions: 0.019 seconds, peak memory: 1.529 GB
There are 10556 edges in the graph and 0 edge cuts for 3 partitions.
server: #clients=5
load test_sampling
start graph service on server 0 for part 0
[05:34:14] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:34:14] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:21029]...
server: #clients=5
load test_sampling
start graph service on server 1 for part 1
[05:34:15] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:34:15] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:21032]...
server: #clients=5
load test_sampling
start graph service on server 2 for part 2
[05:34:16] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:34:16] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:21035]...
[05:34:17] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:34:17] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
[05:34:18] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:34:18] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
[05:34:18] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:34:18] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
[05:34:18] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:34:18] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
[05:34:18] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:34:18] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Client [2026] waits on 172.17.0.3:42933
Client [2025] waits on 172.17.0.3:58535Client [2027] waits on 172.17.0.3:56369

Client [2028] waits on 172.17.0.3:45837
Client [2021] waits on 172.17.0.3:35301
Machine (0) group (0) client (1) connect to server successfuly!
Machine (0) group (0) client (3) connect to server successfuly!
Machine (0) group (0) client (0) connect to server successfuly!
Machine (0) group (0) client (2) connect to server successfuly!Machine (0) group (0) client (4) connect to server successfuly!

  NumNodes: 2708
  NumEdges: 10556
  NumFeats: 1433
  NumClasses: 7
  NumTrainingSamples: 140
  NumValidationSamples: 500
  NumTestSamples: 1000
Done loading data from cached files.
  NumNodes: 2708
  NumEdges: 10556
  NumFeats: 1433
  NumClasses: 7
  NumTrainingSamples: 140
  NumValidationSamples: 500
  NumTestSamples: 1000
Done loading data from cached files.
Client[0] in group[0] is exiting...
Client[4] in group[0] is exiting...
Client[1] in group[0] is exiting...
Client[3] in group[0] is exiting...
Client[2] in group[0] is exiting...
Server (0) shutdown.
Server is exiting...
Server (1) shutdown.
Server is exiting...
Cloning into '/root/jenkins/workspace/dgl_PR-4648@3/third_party/nanoflann'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648@3/third_party/nccl'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648@3/third_party/phmap'...
Server (2) shutdown.
Server is exiting...
PASSED [ 43%]
tests/distributed/test_mp_dataloader.py::test_dist_dataloader[1-False-False-0-3]   NumNodes: 2708
  NumEdges: 10556
  NumFeats: 1433
  NumClasses: 7
  NumTrainingSamples: 140
  NumValidationSamples: 500
  NumTestSamples: 1000
Done loading data from cached files.
torch.int64
Converting to homogeneous graph takes 0.001s, peak mem: 1.529 GB
Convert a graph into a bidirected graph: 0.001 seconds, peak memory: 1.529 GB
Construct multi-constraint weights: 0.000 seconds, peak memory: 1.529 GB
[05:34:25] /root/jenkins/workspace/dgl_PR-4648/src/graph/transform/metis_partition_hetero.cc:87: Partition a graph with 2708 nodes and 10556 edges into 3 parts and get 325 edge cuts
Metis partitioning: 0.002 seconds, peak memory: 1.529 GB
Assigning nodes to METIS partitions takes 0.003s, peak mem: 1.529 GB
Split the graph: 0.002 seconds
Construct subgraphs: 0.001 seconds
Splitting the graph into partitions takes 0.004s, peak mem: 1.529 GB
Calculate edge assignment: 0.000 seconds
part 0 has 1046 nodes and 903 are inside the partition
part 0 has 3490 edges and 3317 are inside the partition
part 1 has 1064 nodes and 902 are inside the partition
part 1 has 4006 edges and 3739 are inside the partition
part 2 has 1056 nodes and 903 are inside the partition
part 2 has 3710 edges and 3500 are inside the partition
Save partitions: 0.017 seconds, peak memory: 1.529 GB
There are 10556 edges in the graph and 0 edge cuts for 3 partitions.
Cloning into '/root/jenkins/workspace/dgl_PR-4648@3/third_party/tensorpipe'...
server: #clients=1
load test_sampling
start graph service on server 0 for part 0
[05:34:26] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:34:26] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:13349]...
Cloning into '/root/jenkins/workspace/dgl_PR-4648@3/third_party/thrust'...
server: #clients=1
load test_sampling
start graph service on server 1 for part 1
[05:34:27] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:34:27] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:13352]...
server: #clients=1
load test_sampling
start graph service on server 2 for part 2
[05:34:28] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:34:28] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:13355]...
[05:34:29] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:34:29] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Cloning into '/root/jenkins/workspace/dgl_PR-4648@3/third_party/tvm'...
Client [2127] waits on 172.17.0.3:47999
Machine (0) group (0) client (0) connect to server successfuly!
  NumNodes: 2708
  NumEdges: 10556
  NumFeats: 1433
  NumClasses: 7
  NumTrainingSamples: 140
  NumValidationSamples: 500
  NumTestSamples: 1000
Done loading data from cached files.
  NumNodes: 2708
  NumEdges: 10556
  NumFeats: 1433
  NumClasses: 7
  NumTrainingSamples: 140
  NumValidationSamples: 500
  NumTestSamples: 1000
Done loading data from cached files.
Client[0] in group[0] is exiting...
Server (0) shutdown.
Server is exiting...
Server (1) shutdown.
Server is exiting...
Server (2) shutdown.
Server is exiting...
PASSED [ 45%]
tests/distributed/test_mp_dataloader.py::test_dist_dataloader[1-False-False-4-3]   NumNodes: 2708
  NumEdges: 10556
  NumFeats: 1433
  NumClasses: 7
  NumTrainingSamples: 140
  NumValidationSamples: 500
  NumTestSamples: 1000
Done loading data from cached files.
torch.int64
Converting to homogeneous graph takes 0.001s, peak mem: 1.529 GB
Convert a graph into a bidirected graph: 0.001 seconds, peak memory: 1.529 GB
Construct multi-constraint weights: 0.000 seconds, peak memory: 1.529 GB
[05:34:34] /root/jenkins/workspace/dgl_PR-4648/src/graph/transform/metis_partition_hetero.cc:87: Partition a graph with 2708 nodes and 10556 edges into 3 parts and get 325 edge cuts
Metis partitioning: 0.002 seconds, peak memory: 1.529 GB
Assigning nodes to METIS partitions takes 0.003s, peak mem: 1.529 GB
Split the graph: 0.002 seconds
Construct subgraphs: 0.002 seconds
Splitting the graph into partitions takes 0.004s, peak mem: 1.529 GB
Calculate edge assignment: 0.000 seconds
part 0 has 1046 nodes and 903 are inside the partition
part 0 has 3490 edges and 3317 are inside the partition
part 1 has 1064 nodes and 902 are inside the partition
part 1 has 4006 edges and 3739 are inside the partition
part 2 has 1056 nodes and 903 are inside the partition
part 2 has 3710 edges and 3500 are inside the partition
Save partitions: 0.017 seconds, peak memory: 1.529 GB
There are 10556 edges in the graph and 0 edge cuts for 3 partitions.
Cloning into '/root/jenkins/workspace/dgl_PR-4648@3/third_party/xbyak'...
Submodule path 'third_party/METIS': checked out '10603482cc055626ff99d2ac3ab7e73d1119439d'
Submodule 'GKlib' (https://github.com/KarypisLab/GKlib.git) registered for path 'third_party/METIS/GKlib'
Cloning into '/root/jenkins/workspace/dgl_PR-4648@3/third_party/METIS/GKlib'...
Submodule path 'third_party/METIS/GKlib': checked out '62de20c7f97c719abdc96d7c3e266a1f7cb52bc5'
Submodule path 'third_party/dlpack': checked out 'e2bdd3bee8cb6501558042633fa59144cc8b7f5f'
Submodule path 'third_party/dmlc-core': checked out 'bfad207b448480783a1f428ae3d93d87032d8349'
server: #clients=5
load test_sampling
start graph service on server 0 for part 0
[05:34:35] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:34:35] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:13039]...
Submodule path 'third_party/googletest': checked out 'f71fb4f9a912ec945401cc49a287a759b6131026'
server: #clients=5
load test_sampling
start graph service on server 1 for part 1
[05:34:36] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:34:36] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:13042]...
Submodule path 'third_party/libxsmm': checked out 'fa687556130b6298430f1c0555a14cf79ab6101c'
Submodule path 'third_party/nanoflann': checked out '4c47ca200209550c5628c89803591f8a753c8181'
Submodule path 'third_party/nccl': checked out 'e11238b3029795d33f958b5868d47c90c4f22628'
Submodule path 'third_party/phmap': checked out '25293cefd8b85491b45600c03fe8edf07647553f'
Submodule path 'third_party/tensorpipe': checked out '6042f1a4cbce8eef997f11ed0012de137b317361'
Submodule 'third_party/googletest' (https://github.com/google/googletest.git) registered for path 'third_party/tensorpipe/third_party/googletest'
Submodule 'third_party/libnop' (https://github.com/google/libnop.git) registered for path 'third_party/tensorpipe/third_party/libnop'
Submodule 'third_party/libuv' (https://github.com/libuv/libuv.git) registered for path 'third_party/tensorpipe/third_party/libuv'
Submodule 'third_party/pybind11' (https://github.com/pybind/pybind11.git) registered for path 'third_party/tensorpipe/third_party/pybind11'
Cloning into '/root/jenkins/workspace/dgl_PR-4648@3/third_party/tensorpipe/third_party/googletest'...
server: #clients=5
load test_sampling
start graph service on server 2 for part 2
[05:34:37] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:34:37] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:13045]...
[05:34:39] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:34:39] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Cloning into '/root/jenkins/workspace/dgl_PR-4648@3/third_party/tensorpipe/third_party/libnop'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648@3/third_party/tensorpipe/third_party/libuv'...
[05:34:39] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:34:39] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
[05:34:39] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:34:39] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
[05:34:39] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:34:39] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
[05:34:40] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:34:40] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Cloning into '/root/jenkins/workspace/dgl_PR-4648@3/third_party/tensorpipe/third_party/pybind11'...
Submodule path 'third_party/tensorpipe/third_party/googletest': checked out 'aee0f9d9b5b87796ee8a0ab26b7587ec30e8858e'
Submodule path 'third_party/tensorpipe/third_party/libnop': checked out 'aa95422ea8c409e3f078d2ee7708a5f59a8b9fa2'
Submodule path 'third_party/tensorpipe/third_party/libuv': checked out '1dff88e5161cba5c59276d2070d2e304e4dcb242'
Submodule path 'third_party/tensorpipe/third_party/pybind11': checked out 'a23996fce38ff6ccfbcdc09f1e63f2c4be5ea2ef'
Submodule 'tools/clang' (https://github.com/wjakob/clang-cindex-python3) registered for path 'third_party/tensorpipe/third_party/pybind11/tools/clang'
Cloning into '/root/jenkins/workspace/dgl_PR-4648@3/third_party/tensorpipe/third_party/pybind11/tools/clang'...
Client [2163] waits on 172.17.0.3:37605Client [2165] waits on 172.17.0.3:39021

Client [2166] waits on 172.17.0.3:54221
Client [2157] waits on 172.17.0.3:55859
Client [2164] waits on 172.17.0.3:57233
Machine (0) group (0) client (1) connect to server successfuly!
Machine (0) group (0) client (0) connect to server successfuly!
Machine (0) group (0) client (2) connect to server successfuly!
Machine (0) group (0) client (4) connect to server successfuly!
Machine (0) group (0) client (3) connect to server successfuly!
  NumNodes: 2708
  NumEdges: 10556
  NumFeats: 1433
  NumClasses: 7
  NumTrainingSamples: 140
  NumValidationSamples: 500
  NumTestSamples: 1000
Done loading data from cached files.
  NumNodes: 2708
  NumEdges: 10556
  NumFeats: 1433
  NumClasses: 7
  NumTrainingSamples: 140
  NumValidationSamples: 500
  NumTestSamples: 1000
Done loading data from cached files.
Submodule path 'third_party/tensorpipe/third_party/pybind11/tools/clang': checked out '6a00cbc4a9b8e68b71caf7f774b3f9c753ae84d5'
Client[3] in group[0] is exiting...
Client[0] in group[0] is exiting...
Submodule path 'third_party/thrust': checked out '6a3078c64cab0e2f276340fa5dcafa0d758ed890'
Submodule 'cub' (https://github.com/NVIDIA/cub.git) registered for path 'third_party/thrust/dependencies/cub'
Cloning into '/root/jenkins/workspace/dgl_PR-4648@3/third_party/thrust/dependencies/cub'...
Client[4] in group[0] is exiting...
Client[1] in group[0] is exiting...
Client[2] in group[0] is exiting...
Server (0) shutdown.
Server is exiting...
Submodule path 'third_party/thrust/dependencies/cub': checked out 'cdaa9558a85e45d849016e5fe7b6e4ee79113f95'
Submodule path 'third_party/tvm': checked out 'b2e418cb109df4cd1f17a2cf2894a1b396a6b838'
Submodule 'dlpack' (https://github.com/dmlc/dlpack) registered for path 'third_party/tvm/3rdparty/dlpack'
Submodule 'dmlc-core' (https://github.com/dmlc/dmlc-core) registered for path 'third_party/tvm/3rdparty/dmlc-core'
Submodule '3rdparty/rang' (https://github.com/agauniyal/rang) registered for path 'third_party/tvm/3rdparty/rang'
Submodule '3rdparty/vta-hw' (https://github.com/apache/incubator-tvm-vta) registered for path 'third_party/tvm/3rdparty/vta-hw'
Cloning into '/root/jenkins/workspace/dgl_PR-4648@3/third_party/tvm/3rdparty/dlpack'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648@3/third_party/tvm/3rdparty/dmlc-core'...
Server (1) shutdown.
Server is exiting...
Server (2) shutdown.
Server is exiting...
Cloning into '/root/jenkins/workspace/dgl_PR-4648@3/third_party/tvm/3rdparty/rang'...
Cloning into '/root/jenkins/workspace/dgl_PR-4648@3/third_party/tvm/3rdparty/vta-hw'...
PASSED [ 47%]
tests/distributed/test_mp_dataloader.py::test_dataloader[node-0-3]   NumNodes: 2708
  NumEdges: 10556
  NumFeats: 1433
  NumClasses: 7
  NumTrainingSamples: 140
  NumValidationSamples: 500
  NumTestSamples: 1000
Done loading data from cached files.
Converting to homogeneous graph takes 0.002s, peak mem: 1.529 GB
Convert a graph into a bidirected graph: 0.001 seconds, peak memory: 1.529 GB
Construct multi-constraint weights: 0.000 seconds, peak memory: 1.529 GB
[05:34:48] /root/jenkins/workspace/dgl_PR-4648/src/graph/transform/metis_partition_hetero.cc:87: Partition a graph with 2708 nodes and 10556 edges into 3 parts and get 325 edge cuts
Metis partitioning: 0.003 seconds, peak memory: 1.529 GB
Assigning nodes to METIS partitions takes 0.004s, peak mem: 1.529 GB
Reshuffle nodes and edges: 0.001 seconds
Split the graph: 0.003 seconds
Construct subgraphs: 0.005 seconds
Splitting the graph into partitions takes 0.009s, peak mem: 1.529 GB
part 0 has 1046 nodes and 903 are inside the partition
part 0 has 3490 edges and 3317 are inside the partition
part 1 has 1064 nodes and 902 are inside the partition
part 1 has 4006 edges and 3739 are inside the partition
part 2 has 1056 nodes and 903 are inside the partition
part 2 has 3710 edges and 3500 are inside the partition
Save partitions: 0.029 seconds, peak memory: 1.529 GB
There are 10556 edges in the graph and 0 edge cuts for 3 partitions.
Submodule path 'third_party/tvm/3rdparty/dlpack': checked out '3ec04430e89a6834e5a1b99471f415fa939bf642'
Submodule path 'third_party/tvm/3rdparty/dmlc-core': checked out '6c401e242c59a1f4c913918246591bb13fd714e7'
Submodule path 'third_party/tvm/3rdparty/rang': checked out 'cabe04d6d6b05356fa8f9741704924788f0dd762'
Submodule path 'third_party/tvm/3rdparty/vta-hw': checked out '87ce9acfae550d1a487746e9d06c2e250076e54c'
Submodule path 'third_party/xbyak': checked out '757e4063f6464740b8ff4a2cae9136d2f8458020'
[Pipeline] unstash
[Pipeline] echo
Unpacked build/libdgl.so, build/runUnitTests, python/dgl/_ffi/_cy3/core.cpython-*-x86_64-linux-gnu.so, build/tensoradapter/pytorch/*.so from dgl-cpu-linux
[Pipeline] timeout
Timeout set to expire in 20 min
[Pipeline] {
[Pipeline] sh
+ bash tests/scripts/task_go_test.sh
~/jenkins/workspace/dgl_PR-4648@3/dglgo ~/jenkins/workspace/dgl_PR-4648@3
server: #clients=1
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_sampling
start graph service on server 0 for part 0
[05:34:49] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:34:49] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:22449]...
WARNING: Skipping dglgo as it is not installed.
WARNING: Running pip as the 'root' user can result in broken permissions and conflicting behaviour with the system package manager. It is recommended to use a virtual environment instead: https://pip.pypa.io/warnings/venv
PASSED            [ 11%]
tests/compute/test_kernel.py::test_mean_zero_degree[g0-int32] PASSED     [ 11%]
tests/compute/test_kernel.py::test_mean_zero_degree[g0-int64] PASSED     [ 11%]
tests/compute/test_merge.py::test_heterograph_merge[int32] PASSED        [ 11%]
tests/compute/test_merge.py::test_heterograph_merge[int64] PASSED        [ 11%]
tests/compute/test_nccl.py::test_nccl_id SKIPPED (NCCL only runs on ...) [ 11%]
tests/compute/test_nccl.py::test_nccl_sparse_push_single_remainder SKIPPED [ 11%]
tests/compute/test_nccl.py::test_nccl_sparse_pull_single_remainder SKIPPED [ 11%]
tests/compute/test_nccl.py::test_nccl_sparse_push_single_range SKIPPED   [ 11%]
tests/compute/test_nccl.py::test_nccl_sparse_pull_single_range SKIPPED   [ 11%]
tests/compute/test_nccl.py::test_nccl_support SKIPPED (NCCL only run...) [ 11%]
tests/compute/test_new_update_all_hetero.py::test_unary_copy_u[int32] SKIPPED [ 11%]
tests/compute/test_new_update_all_hetero.py::test_unary_copy_u[int64] SKIPPED [ 11%]
tests/compute/test_new_update_all_hetero.py::test_unary_copy_e[int32] SKIPPED [ 11%]
tests/compute/test_new_update_all_hetero.py::test_unary_copy_e[int64] SKIPPED [ 11%]
tests/compute/test_new_update_all_hetero.py::test_binary_op[int32] SKIPPED [ 11%]
tests/compute/test_new_update_all_hetero.py::test_binary_op[int64] SKIPPED [ 11%]
tests/compute/test_partition.py::test_get_node_partition_from_book[int32] SKIPPED [ 11%]
tests/compute/test_partition.py::test_get_node_partition_from_book[int64] SKIPPED [ 11%]
tests/compute/test_pickle.py::test_pickling_index PASSED                 [ 11%]
tests/compute/test_pickle.py::test_pickling_graph_index PASSED           [ 11%]
tests/compute/test_pickle.py::test_pickling_graph[g0-int32] PASSED       [ 11%]
tests/compute/test_pickle.py::test_pickling_graph[g0-int64] PASSED       [ 11%]
tests/compute/test_pickle.py::test_pickling_graph[g1-int32] PASSED       [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g1-int64] PASSED       [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g2-int32] PASSED       [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g2-int64] PASSED       [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g3-int32] PASSED       [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g3-int64] PASSED       [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g4-int32] PASSED       [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g4-int64] PASSED       [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g5-int32] PASSED       [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g5-int64] PASSED       [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g6-int32] PASSED       [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g6-int64] PASSED       [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g7-int32] PASSED       [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g7-int64] PASSED       [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g8-int32] PASSED       [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g8-int64] PASSED       [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g9-int32] PASSED       [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g9-int64] PASSED       [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g10-int32] PASSED      [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g10-int64] PASSED      [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g11-int32] PASSED      [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g11-int64] PASSED      [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g12-int32] PASSED      [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g12-int64] PASSED      [ 12%]
tests/compute/test_pickle.py::test_pickling_batched_heterograph PASSED   [ 13%]
tests/compute/test_pickle.py::test_pickling_subgraph PASSED              [ 13%]
tests/compute/test_pickle.py::test_pickling_is_pinned[int32] SKIPPED     [ 13%]
tests/compute/test_pickle.py::test_pickling_is_pinned[int64] SKIPPED     [ 13%]
tests/compute/test_pin_memory.py::test_pin_unpin SKIPPED (Need gpu f...) [ 13%]
tests/compute/test_propagate.py::test_prop_nodes_bfs[int32] PASSED       [ 13%]
tests/compute/test_propagate.py::test_prop_nodes_bfs[int64] PASSED       [ 13%]
tests/compute/test_propagate.py::test_prop_edges_dfs[int32] PASSED       [ 13%]
tests/compute/test_propagate.py::test_prop_edges_dfs[int64] PASSED       [ 13%]
tests/compute/test_propagate.py::test_prop_nodes_topo[int32] PASSED      [ 13%]
tests/compute/test_propagate.py::test_prop_nodes_topo[int64] PASSED      [ 13%]
tests/compute/test_random.py::test_random_choice PASSED                  [ 13%]
tests/compute/test_readout.py::test_sum_case1[int32] PASSED              [ 13%]
tests/compute/test_readout.py::test_sum_case1[int64] PASSED              [ 13%]
tests/compute/test_readout.py::test_reduce_readout[sum-g0-int32] PASSED  [ 13%]
tests/compute/test_readout.py::test_reduce_readout[sum-g0-int64] PASSED  [ 13%]
tests/compute/test_readout.py::test_reduce_readout[sum-g1-int32] PASSED  [ 13%]
tests/compute/test_readout.py::test_reduce_readout[sum-g1-int64] PASSED  [ 13%]
tests/compute/test_readout.py::test_reduce_readout[sum-g2-int32] PASSED  [ 13%]
tests/compute/test_readout.py::test_reduce_readout[sum-g2-int64] PASSED  [ 13%]
tests/compute/test_readout.py::test_reduce_readout[sum-g3-int32] PASSED  [ 13%]
tests/compute/test_readout.py::test_reduce_readout[sum-g3-int64] PASSED  [ 13%]
tests/compute/test_readout.py::test_reduce_readout[sum-g4-int32] PASSED  [ 13%]
tests/compute/test_readout.py::test_reduce_readout[sum-g4-int64] PASSED  [ 13%]
tests/compute/test_readout.py::test_reduce_readout[sum-g5-int32] PASSED  [ 13%]
tests/compute/test_readout.py::test_reduce_readout[sum-g5-int64] PASSED  [ 14%]
tests/compute/test_readout.py::test_reduce_readout[sum-g6-int32] PASSED  [ 14%]
tests/compute/test_readout.py::test_reduce_readout[sum-g6-int64] PASSED  [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g0-int32] PASSED  [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g0-int64] PASSED  [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g1-int32] PASSED  [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g1-int64] PASSED  [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g2-int32] PASSED  [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g2-int64] PASSED  [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g3-int32] PASSED  [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g3-int64] PASSED  [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g4-int32] PASSED  [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g4-int64] PASSED  [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g5-int32] PASSED  [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g5-int64] PASSED  [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g6-int32] PASSED  [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g6-int64] PASSED  [ 14%]
tests/compute/test_readout.py::test_reduce_readout[mean-g0-int32] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[mean-g0-int64] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[mean-g1-int32] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[mean-g1-int64] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[mean-g2-int32] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[mean-g2-int64] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[mean-g3-int32] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[mean-g3-int64] PASSED [ 15%]
tests/compute/test_readout.py::test_reduce_readout[mean-g4-int32] PASSED [ 15%]
tests/compute/test_readout.py::test_reduce_readout[mean-g4-int64] PASSED [ 15%]
tests/compute/test_readout.py::test_reduce_readout[mean-g5-int32] PASSED [ 15%]
tests/compute/test_readout.py::test_reduce_readout[mean-g5-int64] PASSED [ 15%]
tests/compute/test_readout.py::test_reduce_readout[mean-g6-int32] PASSED [ 15%]
tests/compute/test_readout.py::test_reduce_readout[mean-g6-int64] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g0-int32] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g0-int64] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g1-int32] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g1-int64] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g2-int32] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g2-int64] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g3-int32] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g3-int64] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g4-int32] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g4-int64] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g5-int32] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g5-int64] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g6-int32] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g6-int64] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g0-int32] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g0-int64] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g1-int32] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g1-int64] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g2-int32] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g2-int64] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g3-int32] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g3-int64] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g4-int32] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g4-int64] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g5-int32] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g5-int64] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g6-int32] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g6-int64] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g0-int32] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g0-int64] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g1-int32] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g1-int64] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g2-int32] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g2-int64] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g3-int32] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g3-int64] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g4-int32] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g4-int64] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g5-int32] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g5-int64] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g6-int32] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g6-int64] server: #clients=1
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_sampling
start graph service on server 1 for part 1
[05:34:50] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:34:50] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:22452]...
PASSED [ 16%]
tests/compute/test_readout.py::test_topk[True-g0-int32] PASSED           [ 16%]
tests/compute/test_readout.py::test_topk[True-g0-int64] PASSED           [ 17%]
tests/compute/test_readout.py::test_topk[True-g1-int32] PASSED           [ 17%]
tests/compute/test_readout.py::test_topk[True-g1-int64] PASSED           [ 17%]
tests/compute/test_readout.py::test_topk[True-g2-int32] running install
PASSED           [ 17%]
tests/compute/test_readout.py::test_topk[True-g2-int64] PASSED           [ 17%]
tests/compute/test_readout.py::test_topk[True-g3-int32] PASSED           [ 17%]
tests/compute/test_readout.py::test_topk[True-g3-int64] running bdist_egg
running egg_info
creating dglgo.egg-info
PASSED           [ 17%]
tests/compute/test_readout.py::test_topk[True-g4-int32] PASSED           [ 17%]
tests/compute/test_readout.py::test_topk[True-g4-int64] PASSED           [ 17%]
tests/compute/test_readout.py::test_topk[True-g5-int32] writing dglgo.egg-info/PKG-INFO
writing dependency_links to dglgo.egg-info/dependency_links.txt
server: #clients=1
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_sampling
start graph service on server 2 for part 2
[05:34:51] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:34:51] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:22455]...
PASSED           [ 17%]
tests/compute/test_readout.py::test_topk[True-g5-int64] PASSED           [ 17%]
tests/compute/test_readout.py::test_topk[True-g6-int32] PASSED           [ 17%]
tests/compute/test_readout.py::test_topk[True-g6-int64] writing entry points to dglgo.egg-info/entry_points.txt
writing requirements to dglgo.egg-info/requires.txt
PASSED           [ 17%]
tests/compute/test_readout.py::test_topk[False-g0-int32] PASSED          [ 17%]
tests/compute/test_readout.py::test_topk[False-g0-int64] PASSED          [ 17%]
tests/compute/test_readout.py::test_topk[False-g1-int32] writing top-level names to dglgo.egg-info/top_level.txt
writing manifest file 'dglgo.egg-info/SOURCES.txt'
reading manifest file 'dglgo.egg-info/SOURCES.txt'
writing manifest file 'dglgo.egg-info/SOURCES.txt'
installing library code to build/bdist.linux-x86_64/egg
running install_lib
running build_py
creating build
creating build/lib
creating build/lib/recipes
copying recipes/__init__.py -> build/lib/recipes
creating build/lib/dglgo
copying dglgo/__init__.py -> build/lib/dglgo
creating build/lib/dglgo/pipeline
copying dglgo/pipeline/__init__.py -> build/lib/dglgo/pipeline
creating build/lib/dglgo/utils
copying dglgo/utils/early_stop.py -> build/lib/dglgo/utils
copying dglgo/utils/optimizer_config.py -> build/lib/dglgo/utils
copying dglgo/utils/enter_config.py -> build/lib/dglgo/utils
copying dglgo/utils/yaml_dump.py -> build/lib/dglgo/utils
copying dglgo/utils/__init__.py -> build/lib/dglgo/utils
copying dglgo/utils/factory.py -> build/lib/dglgo/utils
copying dglgo/utils/base_model.py -> build/lib/dglgo/utils
creating build/lib/dglgo/apply_pipeline
copying dglgo/apply_pipeline/__init__.py -> build/lib/dglgo/apply_pipeline
creating build/lib/dglgo/model
copying dglgo/model/__init__.py -> build/lib/dglgo/model
creating build/lib/dglgo/cli
copying dglgo/cli/config_apply_cli.py -> build/lib/dglgo/cli
copying dglgo/cli/config_cli.py -> build/lib/dglgo/cli
copying dglgo/cli/train_cli.py -> build/lib/dglgo/cli
copying dglgo/cli/cli.py -> build/lib/dglgo/cli
copying dglgo/cli/export_cli.py -> build/lib/dglgo/cli
copying dglgo/cli/recipe_cli.py -> build/lib/dglgo/cli
copying dglgo/cli/__init__.py -> build/lib/dglgo/cli
copying dglgo/cli/apply_cli.py -> build/lib/dglgo/cli
creating build/lib/dglgo/pipeline/nodepred_sample
copying dglgo/pipeline/nodepred_sample/__init__.py -> build/lib/dglgo/pipeline/nodepred_sample
copying dglgo/pipeline/nodepred_sample/gen.py -> build/lib/dglgo/pipeline/nodepred_sample
creating build/lib/dglgo/pipeline/graphpred
copying dglgo/pipeline/graphpred/__init__.py -> build/lib/dglgo/pipeline/graphpred
copying dglgo/pipeline/graphpred/gen.py -> build/lib/dglgo/pipeline/graphpred
creating build/lib/dglgo/pipeline/linkpred
copying dglgo/pipeline/linkpred/__init__.py -> build/lib/dglgo/pipeline/linkpred
copying dglgo/pipeline/linkpred/gen.py -> build/lib/dglgo/pipeline/linkpred
creating build/lib/dglgo/pipeline/nodepred
copying dglgo/pipeline/nodepred/__init__.py -> build/lib/dglgo/pipeline/nodepred
copying dglgo/pipeline/nodepred/gen.py -> build/lib/dglgo/pipeline/nodepred
creating build/lib/dglgo/apply_pipeline/nodepred_sample
copying dglgo/apply_pipeline/nodepred_sample/__init__.py -> build/lib/dglgo/apply_pipeline/nodepred_sample
copying dglgo/apply_pipeline/nodepred_sample/gen.py -> build/lib/dglgo/apply_pipeline/nodepred_sample
creating build/lib/dglgo/apply_pipeline/graphpred
copying dglgo/apply_pipeline/graphpred/__init__.py -> build/lib/dglgo/apply_pipeline/graphpred
copying dglgo/apply_pipeline/graphpred/gen.py -> build/lib/dglgo/apply_pipeline/graphpred
creating build/lib/dglgo/apply_pipeline/nodepred
copying dglgo/apply_pipeline/nodepred/__init__.py -> build/lib/dglgo/apply_pipeline/nodepred
copying dglgo/apply_pipeline/nodepred/gen.py -> build/lib/dglgo/apply_pipeline/nodepred
creating build/lib/dglgo/model/node_encoder
copying dglgo/model/node_encoder/sage.py -> build/lib/dglgo/model/node_encoder
copying dglgo/model/node_encoder/sgc.py -> build/lib/dglgo/model/node_encoder
copying dglgo/model/node_encoder/gcn.py -> build/lib/dglgo/model/node_encoder
copying dglgo/model/node_encoder/gin.py -> build/lib/dglgo/model/node_encoder
copying dglgo/model/node_encoder/__init__.py -> build/lib/dglgo/model/node_encoder
copying dglgo/model/node_encoder/gat.py -> build/lib/dglgo/model/node_encoder
creating build/lib/dglgo/model/graph_encoder
copying dglgo/model/graph_encoder/gin_ogbg.py -> build/lib/dglgo/model/graph_encoder
copying dglgo/model/graph_encoder/__init__.py -> build/lib/dglgo/model/graph_encoder
copying dglgo/model/graph_encoder/pna.py -> build/lib/dglgo/model/graph_encoder
creating build/lib/dglgo/model/edge_encoder
copying dglgo/model/edge_encoder/bilinear.py -> build/lib/dglgo/model/edge_encoder
copying dglgo/model/edge_encoder/dot.py -> build/lib/dglgo/model/edge_encoder
copying dglgo/model/edge_encoder/ele.py -> build/lib/dglgo/model/edge_encoder
copying dglgo/model/edge_encoder/__init__.py -> build/lib/dglgo/model/edge_encoder
copying recipes/graphpred_hiv_gin.yaml -> build/lib/recipes
copying recipes/graphpred_hiv_pna.yaml -> build/lib/recipes
copying recipes/graphpred_pcba_gin.yaml -> build/lib/recipes
copying recipes/linkpred_citation2_sage.yaml -> build/lib/recipes
copying recipes/linkpred_collab_sage.yaml -> build/lib/recipes
copying recipes/linkpred_cora_sage.yaml -> build/lib/recipes
copying recipes/nodepred-ns_arxiv_gcn.yaml -> build/lib/recipes
copying recipes/nodepred-ns_product_sage.yaml -> build/lib/recipes
copying recipes/nodepred_citeseer_gat.yaml -> build/lib/recipes
copying recipes/nodepred_citeseer_gcn.yaml -> build/lib/recipes
copying recipes/nodepred_citeseer_sage.yaml -> build/lib/recipes
copying recipes/nodepred_cora_gat.yaml -> build/lib/recipes
copying recipes/nodepred_cora_gcn.yaml -> build/lib/recipes
copying recipes/nodepred_cora_sage.yaml -> build/lib/recipes
copying recipes/nodepred_pubmed_gat.yaml -> build/lib/recipes
copying recipes/nodepred_pubmed_gcn.yaml -> build/lib/recipes
copying recipes/nodepred_pubmed_sage.yaml -> build/lib/recipes
copying dglgo/pipeline/nodepred_sample/nodepred-ns.jinja-py -> build/lib/dglgo/pipeline/nodepred_sample
copying dglgo/pipeline/graphpred/graphpred.jinja-py -> build/lib/dglgo/pipeline/graphpred
copying dglgo/pipeline/linkpred/linkpred.jinja-py -> build/lib/dglgo/pipeline/linkpred
copying dglgo/pipeline/nodepred/nodepred.jinja-py -> build/lib/dglgo/pipeline/nodepred
copying dglgo/apply_pipeline/nodepred_sample/nodepred-ns.jinja-py -> build/lib/dglgo/apply_pipeline/nodepred_sample
copying dglgo/apply_pipeline/graphpred/graphpred.jinja-py -> build/lib/dglgo/apply_pipeline/graphpred
copying dglgo/apply_pipeline/nodepred/nodepred.jinja-py -> build/lib/dglgo/apply_pipeline/nodepred
creating build/bdist.linux-x86_64
creating build/bdist.linux-x86_64/egg
creating build/bdist.linux-x86_64/egg/recipes
copying build/lib/recipes/graphpred_hiv_pna.yaml -> build/bdist.linux-x86_64/egg/recipes
copying build/lib/recipes/nodepred_cora_gat.yaml -> build/bdist.linux-x86_64/egg/recipes
copying build/lib/recipes/nodepred_cora_gcn.yaml -> build/bdist.linux-x86_64/egg/recipes
copying build/lib/recipes/nodepred_citeseer_sage.yaml -> build/bdist.linux-x86_64/egg/recipes
copying build/lib/recipes/nodepred_pubmed_sage.yaml -> build/bdist.linux-x86_64/egg/recipes
copying build/lib/recipes/nodepred_pubmed_gcn.yaml -> build/bdist.linux-x86_64/egg/recipes
copying build/lib/recipes/linkpred_citation2_sage.yaml -> build/bdist.linux-x86_64/egg/recipes
copying build/lib/recipes/linkpred_cora_sage.yaml -> build/bdist.linux-x86_64/egg/recipes
copying build/lib/recipes/nodepred-ns_product_sage.yaml -> build/bdist.linux-x86_64/egg/recipes
copying build/lib/recipes/nodepred_pubmed_gat.yaml -> build/bdist.linux-x86_64/egg/recipes
copying build/lib/recipes/graphpred_pcba_gin.yaml -> build/bdist.linux-x86_64/egg/recipes
copying build/lib/recipes/graphpred_hiv_gin.yaml -> build/bdist.linux-x86_64/egg/recipes
copying build/lib/recipes/nodepred_citeseer_gcn.yaml -> build/bdist.linux-x86_64/egg/recipes
copying build/lib/recipes/nodepred_cora_sage.yaml -> build/bdist.linux-x86_64/egg/recipes
copying build/lib/recipes/nodepred-ns_arxiv_gcn.yaml -> build/bdist.linux-x86_64/egg/recipes
copying build/lib/recipes/__init__.py -> build/bdist.linux-x86_64/egg/recipes
copying build/lib/recipes/linkpred_collab_sage.yaml -> build/bdist.linux-x86_64/egg/recipes
copying build/lib/recipes/nodepred_citeseer_gat.yaml -> build/bdist.linux-x86_64/egg/recipes
creating build/bdist.linux-x86_64/egg/dglgo
creating build/bdist.linux-x86_64/egg/dglgo/pipeline
creating build/bdist.linux-x86_64/egg/dglgo/pipeline/nodepred_sample
copying build/lib/dglgo/pipeline/nodepred_sample/nodepred-ns.jinja-py -> build/bdist.linux-x86_64/egg/dglgo/pipeline/nodepred_sample
copying build/lib/dglgo/pipeline/nodepred_sample/__init__.py -> build/bdist.linux-x86_64/egg/dglgo/pipeline/nodepred_sample
copying build/lib/dglgo/pipeline/nodepred_sample/gen.py -> build/bdist.linux-x86_64/egg/dglgo/pipeline/nodepred_sample
creating build/bdist.linux-x86_64/egg/dglgo/pipeline/graphpred
copying build/lib/dglgo/pipeline/graphpred/graphpred.jinja-py -> build/bdist.linux-x86_64/egg/dglgo/pipeline/graphpred
copying build/lib/dglgo/pipeline/graphpred/__init__.py -> build/bdist.linux-x86_64/egg/dglgo/pipeline/graphpred
copying build/lib/dglgo/pipeline/graphpred/gen.py -> build/bdist.linux-x86_64/egg/dglgo/pipeline/graphpred
creating build/bdist.linux-x86_64/egg/dglgo/pipeline/linkpred
copying build/lib/dglgo/pipeline/linkpred/__init__.py -> build/bdist.linux-x86_64/egg/dglgo/pipeline/linkpred
copying build/lib/dglgo/pipeline/linkpred/linkpred.jinja-py -> build/bdist.linux-x86_64/egg/dglgo/pipeline/linkpred
copying build/lib/dglgo/pipeline/linkpred/gen.py -> build/bdist.linux-x86_64/egg/dglgo/pipeline/linkpred
creating build/bdist.linux-x86_64/egg/dglgo/pipeline/nodepred
copying build/lib/dglgo/pipeline/nodepred/__init__.py -> build/bdist.linux-x86_64/egg/dglgo/pipeline/nodepred
copying build/lib/dglgo/pipeline/nodepred/nodepred.jinja-py -> build/bdist.linux-x86_64/egg/dglgo/pipeline/nodepred
copying build/lib/dglgo/pipeline/nodepred/gen.py -> build/bdist.linux-x86_64/egg/dglgo/pipeline/nodepred
copying build/lib/dglgo/pipeline/__init__.py -> build/bdist.linux-x86_64/egg/dglgo/pipeline
creating build/bdist.linux-x86_64/egg/dglgo/utils
copying build/lib/dglgo/utils/early_stop.py -> build/bdist.linux-x86_64/egg/dglgo/utils
copying build/lib/dglgo/utils/optimizer_config.py -> build/bdist.linux-x86_64/egg/dglgo/utils
copying build/lib/dglgo/utils/enter_config.py -> build/bdist.linux-x86_64/egg/dglgo/utils
copying build/lib/dglgo/utils/yaml_dump.py -> build/bdist.linux-x86_64/egg/dglgo/utils
copying build/lib/dglgo/utils/__init__.py -> build/bdist.linux-x86_64/egg/dglgo/utils
copying build/lib/dglgo/utils/factory.py -> build/bdist.linux-x86_64/egg/dglgo/utils
copying build/lib/dglgo/utils/base_model.py -> build/bdist.linux-x86_64/egg/dglgo/utils
creating build/bdist.linux-x86_64/egg/dglgo/apply_pipeline
creating build/bdist.linux-x86_64/egg/dglgo/apply_pipeline/nodepred_sample
copying build/lib/dglgo/apply_pipeline/nodepred_sample/nodepred-ns.jinja-py -> build/bdist.linux-x86_64/egg/dglgo/apply_pipeline/nodepred_sample
copying build/lib/dglgo/apply_pipeline/nodepred_sample/__init__.py -> build/bdist.linux-x86_64/egg/dglgo/apply_pipeline/nodepred_sample
copying build/lib/dglgo/apply_pipeline/nodepred_sample/gen.py -> build/bdist.linux-x86_64/egg/dglgo/apply_pipeline/nodepred_sample
creating build/bdist.linux-x86_64/egg/dglgo/apply_pipeline/graphpred
copying build/lib/dglgo/apply_pipeline/graphpred/graphpred.jinja-py -> build/bdist.linux-x86_64/egg/dglgo/apply_pipeline/graphpred
copying build/lib/dglgo/apply_pipeline/graphpred/__init__.py -> build/bdist.linux-x86_64/egg/dglgo/apply_pipeline/graphpred
copying build/lib/dglgo/apply_pipeline/graphpred/gen.py -> build/bdist.linux-x86_64/egg/dglgo/apply_pipeline/graphpred
creating build/bdist.linux-x86_64/egg/dglgo/apply_pipeline/nodepred
copying build/lib/dglgo/apply_pipeline/nodepred/__init__.py -> build/bdist.linux-x86_64/egg/dglgo/apply_pipeline/nodepred
copying build/lib/dglgo/apply_pipeline/nodepred/nodepred.jinja-py -> build/bdist.linux-x86_64/egg/dglgo/apply_pipeline/nodepred
copying build/lib/dglgo/apply_pipeline/nodepred/gen.py -> build/bdist.linux-x86_64/egg/dglgo/apply_pipeline/nodepred
copying build/lib/dglgo/apply_pipeline/__init__.py -> build/bdist.linux-x86_64/egg/dglgo/apply_pipeline
creating build/bdist.linux-x86_64/egg/dglgo/model
creating build/bdist.linux-x86_64/egg/dglgo/model/node_encoder
copying build/lib/dglgo/model/node_encoder/sage.py -> build/bdist.linux-x86_64/egg/dglgo/model/node_encoder
copying build/lib/dglgo/model/node_encoder/sgc.py -> build/bdist.linux-x86_64/egg/dglgo/model/node_encoder
copying build/lib/dglgo/model/node_encoder/gcn.py -> build/bdist.linux-x86_64/egg/dglgo/model/node_encoder
copying build/lib/dglgo/model/node_encoder/gin.py -> build/bdist.linux-x86_64/egg/dglgo/model/node_encoder
copying build/lib/dglgo/model/node_encoder/__init__.py -> build/bdist.linux-x86_64/egg/dglgo/model/node_encoder
copying build/lib/dglgo/model/node_encoder/gat.py -> build/bdist.linux-x86_64/egg/dglgo/model/node_encoder
creating build/bdist.linux-x86_64/egg/dglgo/model/graph_encoder
copying build/lib/dglgo/model/graph_encoder/gin_ogbg.py -> build/bdist.linux-x86_64/egg/dglgo/model/graph_encoder
copying build/lib/dglgo/model/graph_encoder/__init__.py -> build/bdist.linux-x86_64/egg/dglgo/model/graph_encoder
copying build/lib/dglgo/model/graph_encoder/pna.py -> build/bdist.linux-x86_64/egg/dglgo/model/graph_encoder
creating build/bdist.linux-x86_64/egg/dglgo/model/edge_encoder
copying build/lib/dglgo/model/edge_encoder/bilinear.py -> build/bdist.linux-x86_64/egg/dglgo/model/edge_encoder
copying build/lib/dglgo/model/edge_encoder/dot.py -> build/bdist.linux-x86_64/egg/dglgo/model/edge_encoder
copying build/lib/dglgo/model/edge_encoder/ele.py -> build/bdist.linux-x86_64/egg/dglgo/model/edge_encoder
copying build/lib/dglgo/model/edge_encoder/__init__.py -> build/bdist.linux-x86_64/egg/dglgo/model/edge_encoder
copying build/lib/dglgo/model/__init__.py -> build/bdist.linux-x86_64/egg/dglgo/model
creating build/bdist.linux-x86_64/egg/dglgo/cli
copying build/lib/dglgo/cli/config_apply_cli.py -> build/bdist.linux-x86_64/egg/dglgo/cli
copying build/lib/dglgo/cli/config_cli.py -> build/bdist.linux-x86_64/egg/dglgo/cli
copying build/lib/dglgo/cli/train_cli.py -> build/bdist.linux-x86_64/egg/dglgo/cli
copying build/lib/dglgo/cli/cli.py -> build/bdist.linux-x86_64/egg/dglgo/cli
copying build/lib/dglgo/cli/export_cli.py -> build/bdist.linux-x86_64/egg/dglgo/cli
copying build/lib/dglgo/cli/recipe_cli.py -> build/bdist.linux-x86_64/egg/dglgo/cli
copying build/lib/dglgo/cli/__init__.py -> build/bdist.linux-x86_64/egg/dglgo/cli
copying build/lib/dglgo/cli/apply_cli.py -> build/bdist.linux-x86_64/egg/dglgo/cli
copying build/lib/dglgo/__init__.py -> build/bdist.linux-x86_64/egg/dglgo
byte-compiling build/bdist.linux-x86_64/egg/recipes/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dglgo/pipeline/nodepred_sample/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dglgo/pipeline/nodepred_sample/gen.py to gen.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dglgo/pipeline/graphpred/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dglgo/pipeline/graphpred/gen.py to gen.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dglgo/pipeline/linkpred/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dglgo/pipeline/linkpred/gen.py to gen.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dglgo/pipeline/nodepred/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dglgo/pipeline/nodepred/gen.py to gen.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dglgo/pipeline/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dglgo/utils/early_stop.py to early_stop.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dglgo/utils/optimizer_config.py to optimizer_config.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dglgo/utils/enter_config.py to enter_config.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dglgo/utils/yaml_dump.py to yaml_dump.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dglgo/utils/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dglgo/utils/factory.py to factory.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dglgo/utils/base_model.py to base_model.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dglgo/apply_pipeline/nodepred_sample/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dglgo/apply_pipeline/nodepred_sample/gen.py to gen.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dglgo/apply_pipeline/graphpred/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dglgo/apply_pipeline/graphpred/gen.py to gen.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dglgo/apply_pipeline/nodepred/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dglgo/apply_pipeline/nodepred/gen.py to gen.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dglgo/apply_pipeline/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dglgo/model/node_encoder/sage.py to sage.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dglgo/model/node_encoder/sgc.py to sgc.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dglgo/model/node_encoder/gcn.py to gcn.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dglgo/model/node_encoder/gin.py to gin.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dglgo/model/node_encoder/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dglgo/model/node_encoder/gat.py to gat.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dglgo/model/graph_encoder/gin_ogbg.py to gin_ogbg.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dglgo/model/graph_encoder/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dglgo/model/graph_encoder/pna.py to pna.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dglgo/model/edge_encoder/bilinear.py to bilinear.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dglgo/model/edge_encoder/dot.py to dot.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dglgo/model/edge_encoder/ele.py to ele.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dglgo/model/edge_encoder/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dglgo/model/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dglgo/cli/config_apply_cli.py to config_apply_cli.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dglgo/cli/config_cli.py to config_cli.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dglgo/cli/train_cli.py to train_cli.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dglgo/cli/cli.py to cli.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dglgo/cli/export_cli.py to export_cli.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dglgo/cli/recipe_cli.py to recipe_cli.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dglgo/cli/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dglgo/cli/apply_cli.py to apply_cli.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/dglgo/__init__.py to __init__.cpython-37.pyc
creating build/bdist.linux-x86_64/egg/EGG-INFO
copying dglgo.egg-info/PKG-INFO -> build/bdist.linux-x86_64/egg/EGG-INFO
copying dglgo.egg-info/SOURCES.txt -> build/bdist.linux-x86_64/egg/EGG-INFO
copying dglgo.egg-info/dependency_links.txt -> build/bdist.linux-x86_64/egg/EGG-INFO
copying dglgo.egg-info/entry_points.txt -> build/bdist.linux-x86_64/egg/EGG-INFO
copying dglgo.egg-info/requires.txt -> build/bdist.linux-x86_64/egg/EGG-INFO
copying dglgo.egg-info/top_level.txt -> build/bdist.linux-x86_64/egg/EGG-INFO
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/setuptools/command/install.py:37: SetuptoolsDeprecationWarning: setup.py install is deprecated. Use build and pip and other standards-based tools.
  setuptools.SetuptoolsDeprecationWarning,
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/setuptools/command/easy_install.py:147: EasyInstallDeprecationWarning: easy_install command is deprecated. Use build and pip and other standards-based tools.
  EasyInstallDeprecationWarning,
zip_safe flag not set; analyzing archive contents...
dglgo.apply_pipeline.graphpred.__pycache__.gen.cpython-37: module references __file__
dglgo.apply_pipeline.nodepred.__pycache__.gen.cpython-37: module references __file__
dglgo.apply_pipeline.nodepred_sample.__pycache__.gen.cpython-37: module references __file__
dglgo.cli.__pycache__.recipe_cli.cpython-37: module references __file__
dglgo.pipeline.graphpred.__pycache__.gen.cpython-37: module references __file__
dglgo.pipeline.linkpred.__pycache__.gen.cpython-37: module references __file__
dglgo.pipeline.nodepred.__pycache__.gen.cpython-37: module references __file__
dglgo.pipeline.nodepred_sample.__pycache__.gen.cpython-37: module references __file__
dglgo.utils.__pycache__.factory.cpython-37: module references __file__
creating dist
creating 'dist/dglgo-0.0.2-py3.7.egg' and adding 'build/bdist.linux-x86_64/egg' to it
removing 'build/bdist.linux-x86_64/egg' (and everything under it)
Processing dglgo-0.0.2-py3.7.egg
creating /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/dglgo-0.0.2-py3.7.egg
Extracting dglgo-0.0.2-py3.7.egg to /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages
PASSED          [ 17%]
tests/compute/test_readout.py::test_topk[False-g1-int64] PASSED          [ 17%]
tests/compute/test_readout.py::test_topk[False-g2-int32] PASSED          [ 17%]
tests/compute/test_readout.py::test_topk[False-g2-int64] Adding dglgo 0.0.2 to easy-install.pth file
Installing dgl script to /opt/conda/envs/pytorch-ci/bin

Installed /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/dglgo-0.0.2-py3.7.egg
Processing dependencies for dglgo==0.0.2
Searching for rdkit-pypi
Reading https://pypi.org/simple/rdkit-pypi/
PASSED          [ 17%]
tests/compute/test_readout.py::test_topk[False-g3-int32] PASSED          [ 17%]
tests/compute/test_readout.py::test_topk[False-g3-int64] PASSED          [ 17%]
tests/compute/test_readout.py::test_topk[False-g4-int32] [05:34:52] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:34:52] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
PASSED          [ 17%]
tests/compute/test_readout.py::test_topk[False-g4-int64] PASSED          [ 17%]
tests/compute/test_readout.py::test_topk[False-g5-int32] PASSED          [ 17%]
tests/compute/test_readout.py::test_topk[False-g5-int64] PASSED          [ 18%]
tests/compute/test_readout.py::test_topk[False-g6-int32] PASSED          [ 18%]
tests/compute/test_readout.py::test_topk[False-g6-int64] PASSED          [ 18%]
tests/compute/test_readout.py::test_softmax[g0-int32] PASSED             [ 18%]
tests/compute/test_readout.py::test_softmax[g0-int64] PASSED             [ 18%]
tests/compute/test_readout.py::test_softmax[g1-int32] PASSED             [ 18%]
tests/compute/test_readout.py::test_softmax[g1-int64] PASSED             [ 18%]
tests/compute/test_readout.py::test_softmax[g2-int32] PASSED             [ 18%]
tests/compute/test_readout.py::test_softmax[g2-int64] PASSED             [ 18%]
tests/compute/test_readout.py::test_softmax[g3-int32] PASSED             [ 18%]
tests/compute/test_readout.py::test_softmax[g3-int64] PASSED             [ 18%]
tests/compute/test_readout.py::test_softmax[g4-int32] PASSED             [ 18%]
tests/compute/test_readout.py::test_softmax[g4-int64] PASSED             [ 18%]
tests/compute/test_readout.py::test_softmax[g5-int32] PASSED             [ 18%]
tests/compute/test_readout.py::test_softmax[g5-int64] PASSED             [ 18%]
tests/compute/test_readout.py::test_softmax[g6-int32] PASSED             [ 18%]
tests/compute/test_readout.py::test_softmax[g6-int64] PASSED             [ 18%]
tests/compute/test_readout.py::test_broadcast[g0-int32] PASSED           [ 18%]
tests/compute/test_readout.py::test_broadcast[g0-int64] PASSED           [ 18%]
tests/compute/test_readout.py::test_broadcast[g1-int32] PASSED           [ 18%]
tests/compute/test_readout.py::test_broadcast[g1-int64] PASSED           [ 18%]
tests/compute/test_readout.py::test_broadcast[g2-int32] PASSED           [ 18%]
tests/compute/test_readout.py::test_broadcast[g2-int64] PASSED           [ 18%]
tests/compute/test_readout.py::test_broadcast[g3-int32] PASSED           [ 18%]
tests/compute/test_readout.py::test_broadcast[g3-int64] PASSED           [ 18%]
tests/compute/test_readout.py::test_broadcast[g4-int32] PASSED           [ 19%]
tests/compute/test_readout.py::test_broadcast[g4-int64] PASSED           [ 19%]
tests/compute/test_readout.py::test_broadcast[g5-int32] PASSED           [ 19%]
tests/compute/test_readout.py::test_broadcast[g5-int64] PASSED           [ 19%]
tests/compute/test_readout.py::test_broadcast[g6-int32] PASSED           [ 19%]
tests/compute/test_readout.py::test_broadcast[g6-int64] PASSED           [ 19%]
tests/compute/test_removal.py::test_node_removal[int32] PASSED           [ 19%]
tests/compute/test_removal.py::test_node_removal[int64] PASSED           [ 19%]
tests/compute/test_removal.py::test_multigraph_node_removal[int32] PASSED [ 19%]
tests/compute/test_removal.py::test_multigraph_node_removal[int64] PASSED [ 19%]
tests/compute/test_removal.py::test_multigraph_edge_removal[int32] PASSED [ 19%]
tests/compute/test_removal.py::test_multigraph_edge_removal[int64] PASSED [ 19%]
tests/compute/test_removal.py::test_edge_removal[int32] PASSED           [ 19%]
tests/compute/test_removal.py::test_edge_removal[int64] PASSED           [ 19%]
tests/compute/test_removal.py::test_node_and_edge_removal[int32] Downloading https://files.pythonhosted.org/packages/8f/82/f26ed8d8c943235a1700b3e09f98d2c18f7e6b9eb0839460d8b3b5c26671/rdkit_pypi-2022.3.5-cp37-cp37m-manylinux_2_17_x86_64.manylinux2014_x86_64.whl#sha256=de4a4492a812724037fea086b4e6677124ca91a10fa196d83b9c8acac7c64d5d
Best match: rdkit-pypi 2022.3.5
Processing rdkit_pypi-2022.3.5-cp37-cp37m-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
Installing rdkit_pypi-2022.3.5-cp37-cp37m-manylinux_2_17_x86_64.manylinux2014_x86_64.whl to /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages
Client [2265] waits on 172.17.0.3:39213
Machine (0) group (0) client (0) connect to server successfuly!
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
PASSED  [ 19%]
tests/compute/test_removal.py::test_node_and_edge_removal[int64] Client[0] in group[0] is exiting...
Server (0) shutdown.
Server is exiting...
Server (2) shutdown.Server (1) shutdown.

Server is exiting...
Server is exiting...
PASSED  [ 19%]
tests/compute/test_removal.py::test_node_frame[int32] PASSED             [ 19%]
tests/compute/test_removal.py::test_node_frame[int64] PASSED             [ 19%]
tests/compute/test_removal.py::test_edge_frame[int32] PASSED             [ 19%]
tests/compute/test_removal.py::test_edge_frame[int64] PASSED             [ 19%]
tests/compute/test_removal.py::test_issue1287[int32] PASSED              [ 19%]
tests/compute/test_removal.py::test_issue1287[int64] PASSED              [ 19%]
tests/compute/test_sampler.py::test_create_full PASSED                   [ 19%]
tests/compute/test_sampler.py::test_1neighbor_sampler_all Adding rdkit-pypi 2022.3.5 to easy-install.pth file

Installed /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/rdkit_pypi-2022.3.5-py3.7-linux-x86_64.egg
Searching for PyYAML>=5.1
Reading https://pypi.org/simple/PyYAML/
PASSED         [ 19%]
tests/compute/test_sampler.py::test_1neighbor_sampler PASSED             [ 20%]
tests/compute/test_sampler.py::test_prefetch_neighbor_sampler Downloading https://files.pythonhosted.org/packages/eb/5f/6e6fe6904e1a9c67bc2ca5629a69e7a5a0b17f079da838bab98a1e548b25/PyYAML-6.0-cp37-cp37m-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_12_x86_64.manylinux2010_x86_64.whl#sha256=231710d57adfd809ef5d34183b8ed1eeae3f76459c18fb4a0b373ad56bedcdd9
Best match: PyYAML 6.0
Processing PyYAML-6.0-cp37-cp37m-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_12_x86_64.manylinux2010_x86_64.whl
Installing PyYAML-6.0-cp37-cp37m-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_12_x86_64.manylinux2010_x86_64.whl to /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages
Adding PyYAML 6.0 to easy-install.pth file

Installed /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/PyYAML-6.0-py3.7-linux-x86_64.egg
Searching for ruamel.yaml>=0.17.20
Reading https://pypi.org/simple/ruamel.yaml/
PASSED     [ 20%]
tests/compute/test_sampler.py::test_10neighbor_sampler_all PASSED        [ 20%]
tests/compute/test_sampler.py::test_10neighbor_sampler PASSED            [ 20%]
tests/compute/test_sampler.py::test_layer_sampler PASSED                 [ 20%]
tests/compute/test_sampler.py::test_nonuniform_neighbor_sampler PASSED   [ 20%]
tests/compute/test_sampler.py::test_setseed PASSED                       [ 20%]
tests/compute/test_sampler.py::test_negative_sampler PASSED            [ 11%]
tests/compute/test_kernel.py::test_mean_zero_degree[g0-idtype0] PASSED   [ 11%]
tests/compute/test_kernel.py::test_mean_zero_degree[g0-idtype1] PASSED   [ 11%]
tests/compute/test_merge.py::test_heterograph_merge[idtype0] PASSED      [ 11%]
tests/compute/test_merge.py::test_heterograph_merge[idtype1] PASSED      [ 11%]
tests/compute/test_nccl.py::test_nccl_id SKIPPED (NCCL only runs on ...) [ 11%]
tests/compute/test_nccl.py::test_nccl_sparse_push_single_remainder SKIPPED [ 11%]
tests/compute/test_nccl.py::test_nccl_sparse_pull_single_remainder SKIPPED [ 11%]
tests/compute/test_nccl.py::test_nccl_sparse_push_single_range SKIPPED   [ 11%]
tests/compute/test_nccl.py::test_nccl_sparse_pull_single_range SKIPPED   [ 11%]
tests/compute/test_nccl.py::test_nccl_support SKIPPED (NCCL only run...) [ 11%]
tests/compute/test_new_update_all_hetero.py::test_unary_copy_u[idtype0] SKIPPED [ 11%]
tests/compute/test_new_update_all_hetero.py::test_unary_copy_u[idtype1] SKIPPED [ 11%]
tests/compute/test_new_update_all_hetero.py::test_unary_copy_e[idtype0] SKIPPED [ 11%]
tests/compute/test_new_update_all_hetero.py::test_unary_copy_e[idtype1] SKIPPED [ 11%]
tests/compute/test_new_update_all_hetero.py::test_binary_op[idtype0] SKIPPED [ 11%]
tests/compute/test_new_update_all_hetero.py::test_binary_op[idtype1] SKIPPED [ 11%]
tests/compute/test_partition.py::test_get_node_partition_from_book[idtype0] SKIPPED [ 11%]
tests/compute/test_partition.py::test_get_node_partition_from_book[idtype1] SKIPPED [ 12%]
tests/compute/test_pickle.py::test_pickling_index PASSED                 [ 12%]
tests/compute/test_pickle.py::test_pickling_graph_index PASSED           [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g0-idtype0] PASSED     [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g0-idtype1] PASSED     [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g1-idtype0] PASSED     [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g1-idtype1] PASSED     [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g2-idtype0] PASSED     [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g2-idtype1] PASSED     [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g3-idtype0] PASSED     [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g3-idtype1] PASSED     [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g4-idtype0] PASSED     [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g4-idtype1] PASSED     [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g5-idtype0] PASSED     [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g5-idtype1] PASSED     [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g6-idtype0] PASSED     [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g6-idtype1] PASSED     [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g7-idtype0] PASSED     [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g7-idtype1] PASSED     [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g8-idtype0] PASSED     [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g8-idtype1] PASSED     [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g9-idtype0] PASSED     [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g9-idtype1] PASSED     [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g10-idtype0] PASSED    [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g10-idtype1] PASSED    [ 12%]
tests/compute/test_pickle.py::test_pickling_graph[g11-idtype0] PASSED    [ 13%]
tests/compute/test_pickle.py::test_pickling_graph[g11-idtype1] PASSED    [ 13%]
tests/compute/test_pickle.py::test_pickling_graph[g12-idtype0] PASSED    [ 13%]
tests/compute/test_pickle.py::test_pickling_graph[g12-idtype1] PASSED    [ 13%]
tests/compute/test_pickle.py::test_pickling_batched_heterograph PASSED   [ 13%]
tests/compute/test_pickle.py::test_pickling_subgraph PASSED              [ 13%]
tests/compute/test_pickle.py::test_pickling_is_pinned[idtype0] SKIPPED   [ 13%]
tests/compute/test_pickle.py::test_pickling_is_pinned[idtype1] SKIPPED   [ 13%]
tests/compute/test_pin_memory.py::test_pin_unpin SKIPPED (Need gpu f...) [ 13%]
tests/compute/test_propagate.py::test_prop_nodes_bfs[idtype0] PASSED     [ 13%]
tests/compute/test_propagate.py::test_prop_nodes_bfs[idtype1] PASSED     [ 13%]
tests/compute/test_propagate.py::test_prop_edges_dfs[idtype0] PASSED     [ 13%]
tests/compute/test_propagate.py::test_prop_edges_dfs[idtype1] PASSED     [ 13%]
tests/compute/test_propagate.py::test_prop_nodes_topo[idtype0] PASSED    [ 13%]
tests/compute/test_propagate.py::test_prop_nodes_topo[idtype1] PASSED    [ 13%]
tests/compute/test_random.py::test_random_choice PASSED                  [ 13%]
tests/compute/test_readout.py::test_sum_case1[idtype0] PASSED            [ 13%]
tests/compute/test_readout.py::test_sum_case1[idtype1] PASSED            [ 13%]
tests/compute/test_readout.py::test_reduce_readout[sum-g0-idtype0] PASSED [ 13%]
tests/compute/test_readout.py::test_reduce_readout[sum-g0-idtype1] PASSED [ 13%]
tests/compute/test_readout.py::test_reduce_readout[sum-g1-idtype0] PASSED [ 13%]
tests/compute/test_readout.py::test_reduce_readout[sum-g1-idtype1] PASSED [ 13%]
tests/compute/test_readout.py::test_reduce_readout[sum-g2-idtype0] PASSED [ 13%]
tests/compute/test_readout.py::test_reduce_readout[sum-g2-idtype1] PASSED [ 13%]
tests/compute/test_readout.py::test_reduce_readout[sum-g3-idtype0] PASSED [ 13%]
tests/compute/test_readout.py::test_reduce_readout[sum-g3-idtype1] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[sum-g4-idtype0] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[sum-g4-idtype1] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[sum-g5-idtype0] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[sum-g5-idtype1] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[sum-g6-idtype0] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[sum-g6-idtype1] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g0-idtype0] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g0-idtype1] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g1-idtype0] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g1-idtype1] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g2-idtype0] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g2-idtype1] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g3-idtype0] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g3-idtype1] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g4-idtype0] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g4-idtype1] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g5-idtype0] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g5-idtype1] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g6-idtype0] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[max-g6-idtype1] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[mean-g0-idtype0] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[mean-g0-idtype1] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[mean-g1-idtype0] PASSED [ 14%]
tests/compute/test_readout.py::test_reduce_readout[mean-g1-idtype1] PASSED [ 15%]
tests/compute/test_readout.py::test_reduce_readout[mean-g2-idtype0] PASSED [ 15%]
tests/compute/test_readout.py::test_reduce_readout[mean-g2-idtype1] PASSED [ 15%]
tests/compute/test_readout.py::test_reduce_readout[mean-g3-idtype0] PASSED [ 15%]
tests/compute/test_readout.py::test_reduce_readout[mean-g3-idtype1] PASSED [ 15%]
tests/compute/test_readout.py::test_reduce_readout[mean-g4-idtype0] PASSED [ 15%]
tests/compute/test_readout.py::test_reduce_readout[mean-g4-idtype1] PASSED [ 15%]
tests/compute/test_readout.py::test_reduce_readout[mean-g5-idtype0] PASSED [ 15%]
tests/compute/test_readout.py::test_reduce_readout[mean-g5-idtype1] PASSED [ 15%]
tests/compute/test_readout.py::test_reduce_readout[mean-g6-idtype0] PASSED [ 15%]
tests/compute/test_readout.py::test_reduce_readout[mean-g6-idtype1] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g0-idtype0] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g0-idtype1] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g1-idtype0] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g1-idtype1] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g2-idtype0] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g2-idtype1] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g3-idtype0] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g3-idtype1] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g4-idtype0] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g4-idtype1] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g5-idtype0] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g5-idtype1] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g6-idtype0] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[sum-g6-idtype1] PASSED [ 15%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g0-idtype0] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g0-idtype1] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g1-idtype0] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g1-idtype1] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g2-idtype0] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g2-idtype1] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g3-idtype0] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g3-idtype1] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g4-idtype0] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g4-idtype1] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g5-idtype0] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g5-idtype1] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g6-idtype0] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[max-g6-idtype1] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g0-idtype0] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g0-idtype1] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g1-idtype0] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g1-idtype1] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g2-idtype0] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g2-idtype1] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g3-idtype0] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g3-idtype1] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g4-idtype0] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g4-idtype1] PASSED [ 16%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g5-idtype0] PASSED [ 17%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g5-idtype1] PASSED [ 17%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g6-idtype0] PASSED [ 17%]
tests/compute/test_readout.py::test_weighted_reduce_readout[mean-g6-idtype1] PASSED [ 17%]
tests/compute/test_readout.py::test_topk[True-g0-idtype0] PASSED         [ 17%]
tests/compute/test_readout.py::test_topk[True-g0-idtype1] PASSED         [ 17%]
tests/compute/test_readout.py::test_topk[True-g1-idtype0] PASSED         [ 17%]
tests/compute/test_readout.py::test_topk[True-g1-idtype1] PASSED         [ 17%]
tests/compute/test_readout.py::test_topk[True-g2-idtype0] PASSED         [ 17%]
tests/compute/test_readout.py::test_topk[True-g2-idtype1] PASSED         [ 17%]
tests/compute/test_readout.py::test_topk[True-g3-idtype0] PASSED         [ 17%]
tests/compute/test_readout.py::test_topk[True-g3-idtype1] PASSED         [ 17%]
tests/compute/test_readout.py::test_topk[True-g4-idtype0] PASSED         [ 17%]
tests/compute/test_readout.py::test_topk[True-g4-idtype1] PASSED         [ 17%]
tests/compute/test_readout.py::test_topk[True-g5-idtype0] PASSED         [ 17%]
tests/compute/test_readout.py::test_topk[True-g5-idtype1] PASSED         [ 17%]
tests/compute/test_readout.py::test_topk[True-g6-idtype0] PASSED         [ 17%]
tests/compute/test_readout.py::test_topk[True-g6-idtype1] PASSED         [ 17%]
tests/compute/test_readout.py::test_topk[False-g0-idtype0] PASSED        [ 17%]
tests/compute/test_readout.py::test_topk[False-g0-idtype1] PASSED        [ 17%]
tests/compute/test_readout.py::test_topk[False-g1-idtype0] PASSED        [ 17%]
tests/compute/test_readout.py::test_topk[False-g1-idtype1] PASSED        [ 17%]
tests/compute/test_readout.py::test_topk[False-g2-idtype0] PASSED        [ 17%]
tests/compute/test_readout.py::test_topk[False-g2-idtype1] PASSED        [ 17%]
tests/compute/test_readout.py::test_topk[False-g3-idtype0] PASSED        [ 17%]
tests/compute/test_readout.py::test_topk[False-g3-idtype1] PASSED        [ 18%]
tests/compute/test_readout.py::test_topk[False-g4-idtype0] PASSED        [ 18%]
tests/compute/test_readout.py::test_topk[False-g4-idtype1] PASSED        [ 18%]
tests/compute/test_readout.py::test_topk[False-g5-idtype0] PASSED        [ 18%]
tests/compute/test_readout.py::test_topk[False-g5-idtype1] PASSED        [ 18%]
tests/compute/test_readout.py::test_topk[False-g6-idtype0] PASSED        [ 18%]
tests/compute/test_readout.py::test_topk[False-g6-idtype1] PASSED        [ 18%]
tests/compute/test_readout.py::test_softmax[g0-idtype0] PASSED           [ 18%]
tests/compute/test_readout.py::test_softmax[g0-idtype1] PASSED           [ 18%]
tests/compute/test_readout.py::test_softmax[g1-idtype0] PASSED           [ 18%]
tests/compute/test_readout.py::test_softmax[g1-idtype1] PASSED           [ 18%]
tests/compute/test_readout.py::test_softmax[g2-idtype0] PASSED           [ 18%]
tests/compute/test_readout.py::test_softmax[g2-idtype1] PASSED           [ 18%]
tests/compute/test_readout.py::test_softmax[g3-idtype0] PASSED           [ 18%]
tests/compute/test_readout.py::test_softmax[g3-idtype1] PASSED           [ 18%]
tests/compute/test_readout.py::test_softmax[g4-idtype0] PASSED           [ 18%]
tests/compute/test_readout.py::test_softmax[g4-idtype1] PASSED           [ 18%]
tests/compute/test_readout.py::test_softmax[g5-idtype0] PASSED           [ 18%]
tests/compute/test_readout.py::test_softmax[g5-idtype1] PASSED           [ 18%]
tests/compute/test_readout.py::test_softmax[g6-idtype0] PASSED           [ 18%]
tests/compute/test_readout.py::test_softmax[g6-idtype1] PASSED           [ 18%]
tests/compute/test_readout.py::test_broadcast[g0-idtype0] PASSED         [ 18%]
tests/compute/test_readout.py::test_broadcast[g0-idtype1] PASSED         [ 18%]
tests/compute/test_readout.py::test_broadcast[g1-idtype0] PASSED         [ 18%]
tests/compute/test_readout.py::test_broadcast[g1-idtype1] PASSED         [ 18%]
tests/compute/test_readout.py::test_broadcast[g2-idtype0] PASSED         [ 19%]
tests/compute/test_readout.py::test_broadcast[g2-idtype1] PASSED         [ 19%]
tests/compute/test_readout.py::test_broadcast[g3-idtype0] PASSED         [ 19%]
tests/compute/test_readout.py::test_broadcast[g3-idtype1] PASSED         [ 19%]
tests/compute/test_readout.py::test_broadcast[g4-idtype0] PASSED         [ 19%]
tests/compute/test_readout.py::test_broadcast[g4-idtype1] PASSED         [ 19%]
tests/compute/test_readout.py::test_broadcast[g5-idtype0] PASSED         [ 19%]
tests/compute/test_readout.py::test_broadcast[g5-idtype1] PASSED         [ 19%]
tests/compute/test_readout.py::test_broadcast[g6-idtype0] PASSED         [ 19%]
tests/compute/test_readout.py::test_broadcast[g6-idtype1] PASSED         [ 19%]
tests/compute/test_removal.py::test_node_removal[idtype0] PASSED         [ 19%]
tests/compute/test_removal.py::test_node_removal[idtype1] PASSED         [ 19%]
tests/compute/test_removal.py::test_multigraph_node_removal[idtype0] PASSED [ 19%]
tests/compute/test_removal.py::test_multigraph_node_removal[idtype1] PASSED [ 19%]
tests/compute/test_removal.py::test_multigraph_edge_removal[idtype0] PASSED [ 19%]
tests/compute/test_removal.py::test_multigraph_edge_removal[idtype1] PASSED [ 19%]
tests/compute/test_removal.py::test_edge_removal[idtype0] PASSED         [ 19%]
tests/compute/test_removal.py::test_edge_removal[idtype1] PASSED         [ 19%]
tests/compute/test_removal.py::test_node_and_edge_removal[idtype0] PASSED [ 19%]
tests/compute/test_removal.py::test_node_and_edge_removal[idtype1] PASSED [ 19%]
tests/compute/test_removal.py::test_node_frame[idtype0] PASSED           [ 19%]
tests/compute/test_removal.py::test_node_frame[idtype1] PASSED           [ 19%]
tests/compute/test_removal.py::test_edge_frame[idtype0] PASSED           [ 19%]
tests/compute/test_removal.py::test_edge_frame[idtype1] PASSED           [ 19%]
tests/compute/test_removal.py::test_issue1287[idtype0] PASSED            [ 20%]
tests/compute/test_removal.py::test_issue1287[idtype1] PASSED            [ 20%]
tests/compute/test_sampler.py::test_create_full PASSED                   [ 20%]
tests/compute/test_sampler.py::test_1neighbor_sampler_all PASSED         [ 20%]
tests/compute/test_sampler.py::test_1neighbor_sampler PASSED             [ 20%]
tests/compute/test_sampler.py::test_prefetch_neighbor_sampler PASSED     [ 20%]
tests/compute/test_sampler.py::test_10neighbor_sampler_all PASSED        [ 20%]
tests/compute/test_sampler.py::test_10neighbor_sampler PASSED            [ 20%]
tests/compute/test_sampler.py::test_layer_sampler PASSED                 [ 20%]
tests/compute/test_sampler.py::test_nonuniform_neighbor_sampler SKIPPED  [ 20%]
tests/compute/test_sampler.py::test_setseed PASSED                       [ 20%]
tests/compute/test_sampler.py::test_negative_sampler SKIPPED (TF doe...) [ 20%]
tests/compute/test_sampling.py::test_non_uniform_random_walk[True] SKIPPED [ 20%]
tests/compute/test_sampling.py::test_non_uniform_random_walk[False] PASSED [ 20%]
tests/compute/test_sampling.py::test_uniform_random_walk[True] SKIPPED   [ 20%]
tests/compute/test_sampling.py::test_uniform_random_walk[False] PASSED   [ 20%]
tests/compute/test_sampling.py::test_node2vec PASSED                     [ 20%]
tests/compute/test_sampling.py::test_pack_traces PASSED                  [ 20%]
tests/compute/test_sampling.py::test_pinsage_sampling[True] SKIPPED      [ 20%]
tests/compute/test_sampling.py::test_pinsage_sampling[False] PASSED      [ 20%]
tests/compute/test_sampling.py::test_sample_neighbors_noprob Downloading https://files.pythonhosted.org/packages/9e/cb/938214ac358fbef7058343b3765c79a1b7ed0c366f7f992ce7ff38335652/ruamel.yaml-0.17.21-py3-none-any.whl#sha256=742b35d3d665023981bd6d16b3d24248ce5df75fdb4e2924e93a05c1f8b61ca7
Best match: ruamel.yaml 0.17.21
Processing ruamel.yaml-0.17.21-py3-none-any.whl
Installing ruamel.yaml-0.17.21-py3-none-any.whl to /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages
Converting to homogeneous graph takes 0.010s, peak mem: 1.841 GB
Convert a graph into a bidirected graph: 0.017 seconds, peak memory: 1.841 GB
Construct multi-constraint weights: 0.001 seconds, peak memory: 1.841 GB
[05:35:13] /root/jenkins/workspace/dgl_PR-4648/src/graph/transform/metis_partition_hetero.cc:87: Partition a graph with 30030 nodes and 601200 edges into 3 parts and get 151889 edge cuts
Metis partitioning: 0.177 seconds, peak memory: 1.841 GB
Assigning nodes to METIS partitions takes 0.195s, peak mem: 1.841 GB
Reshuffle nodes and edges: 0.016 seconds
Split the graph: 0.045 seconds
Construct subgraphs: 0.043 seconds
Splitting the graph into partitions takes 0.104s, peak mem: 1.841 GB
part 0 has 9942 nodes of type n1 and 3143 are inside the partition
part 0 has 9963 nodes of type n2 and 3139 are inside the partition
part 0 has 9973 nodes of type n3 and 3142 are inside the partition
part 0 has 47618 edges of type r1 and 31517 are inside the partition
part 0 has 47779 edges of type r2 and 31406 are inside the partition
part 0 has 47492 edges of type r3 and 31197 are inside the partition
part 1 has 9967 nodes of type n1 and 3424 are inside the partition
part 1 has 9984 nodes of type n2 and 3436 are inside the partition
part 1 has 9989 nodes of type n3 and 3438 are inside the partition
part 1 has 51567 edges of type r1 and 34410 are inside the partition
part 1 has 51514 edges of type r2 and 34297 are inside the partition
part 1 has 51767 edges of type r3 and 34580 are inside the partition
part 2 has 9964 nodes of type n1 and 3433 are inside the partition
part 2 has 9985 nodes of type n2 and 3435 are inside the partition
part 2 has 9988 nodes of type n3 and 3440 are inside the partition
part 2 has 51379 edges of type r1 and 34173 are inside the partition
part 2 has 51847 edges of type r2 and 34497 are inside the partition
part 2 has 51526 edges of type r3 and 34523 are inside the partition
Save partitions: 0.037 seconds, peak memory: 1.841 GB
There are 300600 edges in the graph and 0 edge cuts for 3 partitions.
server: #clients=1
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_sampling
start graph service on server 0 for part 0
[05:35:14] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:35:14] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:13169]...
server: #clients=1
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_sampling
start graph service on server 1 for part 1
[05:35:15] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:35:15] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:13172]...
Adding ruamel.yaml 0.17.21 to easy-install.pth file

Installed /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/ruamel.yaml-0.17.21-py3.7.egg
Searching for pydantic>=1.9.0
Reading https://pypi.org/simple/pydantic/
server: #clients=1
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_sampling
start graph service on server 2 for part 2
[05:35:16] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:35:16] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:13175]...
[05:35:17] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:35:17] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
PASSED      [ 20%]
tests/compute/test_sampling.py::test_sample_neighbors_prob Client [2305] waits on 172.17.0.3:34813
Machine (0) group (0) client (0) connect to server successfuly!
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
Client[0] in group[0] is exiting...
Server (1) shutdown.
Server is exiting...
Server (0) shutdown.
Server is exiting...
Server (2) shutdown.
Server is exiting...
PASSED [ 49%]
tests/distributed/test_mp_dataloader.py::test_dataloader[node-4-3]   NumNodes: 2708
  NumEdges: 10556
  NumFeats: 1433
  NumClasses: 7
  NumTrainingSamples: 140
  NumValidationSamples: 500
  NumTestSamples: 1000
Done loading data from cached files.
Converting to homogeneous graph takes 0.001s, peak mem: 1.841 GB
Convert a graph into a bidirected graph: 0.001 seconds, peak memory: 1.841 GB
Construct multi-constraint weights: 0.000 seconds, peak memory: 1.841 GB
[05:35:21] /root/jenkins/workspace/dgl_PR-4648/src/graph/transform/metis_partition_hetero.cc:87: Partition a graph with 2708 nodes and 10556 edges into 3 parts and get 325 edge cuts
Metis partitioning: 0.003 seconds, peak memory: 1.841 GB
Assigning nodes to METIS partitions takes 0.004s, peak mem: 1.841 GB
Reshuffle nodes and edges: 0.001 seconds
Split the graph: 0.002 seconds
Construct subgraphs: 0.004 seconds
Splitting the graph into partitions takes 0.008s, peak mem: 1.841 GB
part 0 has 1046 nodes and 903 are inside the partition
part 0 has 3490 edges and 3317 are inside the partition
part 1 has 1064 nodes and 902 are inside the partition
part 1 has 4006 edges and 3739 are inside the partition
part 2 has 1056 nodes and 903 are inside the partition
part 2 has 3710 edges and 3500 are inside the partition
Save partitions: 0.020 seconds, peak memory: 1.841 GB
There are 10556 edges in the graph and 0 edge cuts for 3 partitions.
Downloading https://files.pythonhosted.org/packages/d4/ec/230ab377c457cd68cfda78759e2a57f8c08a9e9adb4cd53c4d2fc9100b15/pydantic-1.10.2-py3-none-any.whl#sha256=1b6ee725bd6e83ec78b1aa32c5b1fa67a3a65badddde3976bca5fe4568f27709
Best match: pydantic 1.10.2
Processing pydantic-1.10.2-py3-none-any.whl
Installing pydantic-1.10.2-py3-none-any.whl to /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages
Adding pydantic 1.10.2 to easy-install.pth file

Installed /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/pydantic-1.10.2-py3.7.egg
Searching for numpydoc>=1.1.0
Reading https://pypi.org/simple/numpydoc/
Downloading https://files.pythonhosted.org/packages/e7/1a/9e3c2a34aae5bd1ab8988b238aafeb4c8d3ab312b8aa5a8c37be6c6d869d/numpydoc-1.4.0-py3-none-any.whl#sha256=fd26258868ebcc75c816fe68e1d41e3b55bd410941acfb969dee3eef6e5cf260
Best match: numpydoc 1.4.0
Processing numpydoc-1.4.0-py3-none-any.whl
Installing numpydoc-1.4.0-py3-none-any.whl to /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages
Adding numpydoc 1.4.0 to easy-install.pth file

Installed /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/numpydoc-1.4.0-py3.7.egg
Searching for autopep8>=1.6.0
Reading https://pypi.org/simple/autopep8/
Downloading https://files.pythonhosted.org/packages/5d/9b/1ed75f8c9086fafe0e9bbb379a70c43b1aa9dff6154ddcfb818f78cb0736/autopep8-1.7.0-py2.py3-none-any.whl#sha256=6f09e90a2be784317e84dc1add17ebfc7abe3924239957a37e5040e27d812087
Best match: autopep8 1.7.0
Processing autopep8-1.7.0-py2.py3-none-any.whl
Installing autopep8-1.7.0-py2.py3-none-any.whl to /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages
server: #clients=5
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_sampling
start graph service on server 0 for part 0
[05:35:22] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:35:22] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:21763]...
Adding autopep8 1.7.0 to easy-install.pth file
Installing autopep8 script to /opt/conda/envs/pytorch-ci/bin

Installed /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/autopep8-1.7.0-py3.7.egg
Searching for isort>=5.10.1
Reading https://pypi.org/simple/isort/
server: #clients=5
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_sampling
start graph service on server 1 for part 1
[05:35:23] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:35:23] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:21766]...
PASSED        [ 20%]
tests/compute/test_sampling.py::test_sample_neighbors_outedge Downloading https://files.pythonhosted.org/packages/b8/5b/f18e227df38b94b4ee30d2502fd531bebac23946a2497e5595067a561274/isort-5.10.1-py3-none-any.whl#sha256=6f62d78e2f89b4500b080fe3a81690850cd254227f27f75c3a0c491a1f351ba7
Best match: isort 5.10.1
Processing isort-5.10.1-py3-none-any.whl
Installing isort-5.10.1-py3-none-any.whl to /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages
server: #clients=5
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
Adding isort 5.10.1 to easy-install.pth file
Installing isort script to /opt/conda/envs/pytorch-ci/bin
Installing isort-identify-imports script to /opt/conda/envs/pytorch-ci/bin

Installed /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/isort-5.10.1-py3.7.egg
Searching for typer>=0.4.0
Reading https://pypi.org/simple/typer/
load test_sampling
start graph service on server 2 for part 2
[05:35:24] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:35:24] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:21769]...
Downloading https://files.pythonhosted.org/packages/e8/9b/7470461c68588ed09c2e53cbb16b802815232796d95f7b4744cc8842f19d/typer-0.6.1-py3-none-any.whl#sha256=54b19e5df18654070a82f8c2aa1da456a4ac16a2a83e6dcd9f170e291c56338e
Best match: typer 0.6.1
Processing typer-0.6.1-py3-none-any.whl
Installing typer-0.6.1-py3-none-any.whl to /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages
Adding typer 0.6.1 to easy-install.pth file

Installed /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/typer-0.6.1-py3.7.egg
Searching for ruamel.yaml.clib>=0.2.6
Reading https://pypi.org/simple/ruamel.yaml.clib/
[05:35:25] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:35:25] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
[05:35:26] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:35:26] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
[05:35:26] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:35:26] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Downloading https://files.pythonhosted.org/packages/98/8a/ba37489b423916162b086b01c7c18001cf297350694180468e1698085c58/ruamel.yaml.clib-0.2.6-cp37-cp37m-manylinux1_x86_64.whl#sha256=78988ed190206672da0f5d50c61afef8f67daa718d614377dcd5e3ed85ab4a99
Best match: ruamel.yaml.clib 0.2.6
Processing ruamel.yaml.clib-0.2.6-cp37-cp37m-manylinux1_x86_64.whl
Installing ruamel.yaml.clib-0.2.6-cp37-cp37m-manylinux1_x86_64.whl to /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages
[05:35:26] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:35:26] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
[05:35:26] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:35:26] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Adding ruamel.yaml.clib 0.2.6 to easy-install.pth file

Installed /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/ruamel.yaml.clib-0.2.6-py3.7-linux-x86_64.egg
Searching for Jinja2>=2.10
Reading https://pypi.org/simple/Jinja2/
Downloading https://files.pythonhosted.org/packages/bc/c3/f068337a370801f372f2f8f6bad74a5c140f6fda3d9de154052708dd3c65/Jinja2-3.1.2-py3-none-any.whl#sha256=6088930bfe239f0e6710546ab9c19c9ef35e29792895fed6e6e31a023a182a61
Best match: Jinja2 3.1.2
Processing Jinja2-3.1.2-py3-none-any.whl
Installing Jinja2-3.1.2-py3-none-any.whl to /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages
Adding Jinja2 3.1.2 to easy-install.pth file

Installed /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/Jinja2-3.1.2-py3.7.egg
Searching for sphinx>=3.0
Reading https://pypi.org/simple/sphinx/
Downloading https://files.pythonhosted.org/packages/00/5c/27480aeb398ab549cbb0840d154464d4854ac49f6fff0b452e106bc54f8e/sphinx-5.2.1-py3-none-any.whl#sha256=3dcf00fcf82cf91118db9b7177edea4fc01998976f893928d0ab0c58c54be2ca
Best match: sphinx 5.2.1
Processing sphinx-5.2.1-py3-none-any.whl
Installing sphinx-5.2.1-py3-none-any.whl to /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages
Adding sphinx 5.2.1 to easy-install.pth file
Installing sphinx-apidoc script to /opt/conda/envs/pytorch-ci/bin
Installing sphinx-autogen script to /opt/conda/envs/pytorch-ci/bin
Installing sphinx-build script to /opt/conda/envs/pytorch-ci/bin
Installing sphinx-quickstart script to /opt/conda/envs/pytorch-ci/bin

Installed /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/sphinx-5.2.1-py3.7.egg
Searching for toml
Reading https://pypi.org/simple/toml/
Downloading https://files.pythonhosted.org/packages/44/6f/7120676b6d73228c96e17f1f794d8ab046fc910d781c8d151120c3f1569e/toml-0.10.2-py2.py3-none-any.whl#sha256=806143ae5bfb6a3c6e736a764057db0e6a0e05e338b5630894a5f779cabb4f9b
Client [2335] waits on 172.17.0.3:55517
Client [2340] waits on 172.17.0.3:40105
Client [2341] waits on 172.17.0.3:50781Client [2339] waits on 172.17.0.3:47191Client [2342] waits on 172.17.0.3:46505


Machine (0) group (0) client (4) connect to server successfuly!
Machine (0) group (0) client (0) connect to server successfuly!
Machine (0) group (0) client (2) connect to server successfuly!
Machine (0) group (0) client (3) connect to server successfuly!
Machine (0) group (0) client (1) connect to server successfuly!
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
Best match: toml 0.10.2
Processing toml-0.10.2-py2.py3-none-any.whl
Installing toml-0.10.2-py2.py3-none-any.whl to /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages
Adding toml 0.10.2 to easy-install.pth file

Installed /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/toml-0.10.2-py3.7.egg
Searching for pycodestyle>=2.9.1
Reading https://pypi.org/simple/pycodestyle/
Client[4] in group[0] is exiting...
Client[2] in group[0] is exiting...
Downloading https://files.pythonhosted.org/packages/67/e4/fc77f1039c34b3612c4867b69cbb2b8a4e569720b1f19b0637002ee03aff/pycodestyle-2.9.1-py2.py3-none-any.whl#sha256=d1735fc58b418fd7c5f658d28d943854f8a849b01a5d0a1e6f3f3fdd0166804b
Best match: pycodestyle 2.9.1
Processing pycodestyle-2.9.1-py2.py3-none-any.whl
Installing pycodestyle-2.9.1-py2.py3-none-any.whl to /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages
Client[0] in group[0] is exiting...
Adding pycodestyle 2.9.1 to easy-install.pth file
Installing pycodestyle script to /opt/conda/envs/pytorch-ci/bin

Installed /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/pycodestyle-2.9.1-py3.7.egg
Searching for MarkupSafe>=2.0
Reading https://pypi.org/simple/MarkupSafe/
Client[3] in group[0] is exiting...
Client[1] in group[0] is exiting...
Server (2) shutdown.
Server (1) shutdown.Server is exiting...

Server is exiting...
Server (0) shutdown.
Server is exiting...
Downloading https://files.pythonhosted.org/packages/9f/83/b221ce5a0224f409b9f02b0dc6cb0b921c46033f4870d64fa3e8a96af701/MarkupSafe-2.1.1-cp37-cp37m-manylinux_2_17_x86_64.manylinux2014_x86_64.whl#sha256=e72591e9ecd94d7feb70c1cbd7be7b3ebea3f548870aa91e2732960fa4d57a37
PASSED     [ 20%]
tests/compute/test_sampling.py::test_sample_neighbors_topk Best match: MarkupSafe 2.1.1
Processing MarkupSafe-2.1.1-cp37-cp37m-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
Installing MarkupSafe-2.1.1-cp37-cp37m-manylinux_2_17_x86_64.manylinux2014_x86_64.whl to /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages
Adding MarkupSafe 2.1.1 to easy-install.pth file

Installed /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/MarkupSafe-2.1.1-py3.7-linux-x86_64.egg
Searching for imagesize>=1.3
Reading https://pypi.org/simple/imagesize/
Downloading https://files.pythonhosted.org/packages/ff/62/85c4c919272577931d407be5ba5d71c20f0b616d31a0befe0ae45bb79abd/imagesize-1.4.1-py2.py3-none-any.whl#sha256=0d8d18d08f840c19d0ee7ca1fd82490fdc3729b7ac93f49870406ddde8ef8d8b
Best match: imagesize 1.4.1
Processing imagesize-1.4.1-py2.py3-none-any.whl
Installing imagesize-1.4.1-py2.py3-none-any.whl to /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages
Adding imagesize 1.4.1 to easy-install.pth file

Installed /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/imagesize-1.4.1-py3.7.egg
Searching for alabaster<0.8,>=0.7
Reading https://pypi.org/simple/alabaster/
Downloading https://files.pythonhosted.org/packages/10/ad/00b090d23a222943eb0eda509720a404f531a439e803f6538f35136cae9e/alabaster-0.7.12-py2.py3-none-any.whl#sha256=446438bdcca0e05bd45ea2de1668c1d9b032e1a9154c2c259092d77031ddd359
PASSED        [ 20%]
tests/compute/test_sampling.py::test_sample_neighbors_topk_outedge Best match: alabaster 0.7.12
Processing alabaster-0.7.12-py2.py3-none-any.whl
Installing alabaster-0.7.12-py2.py3-none-any.whl to /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages
Adding alabaster 0.7.12 to easy-install.pth file
PASSED [ 20%]
tests/compute/test_sampling.py::test_sample_neighbors_with_0deg 
Installed /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/alabaster-0.7.12-py3.7.egg
Searching for babel>=2.9
Reading https://pypi.org/simple/babel/
Downloading https://files.pythonhosted.org/packages/2e/57/a4177e24f8ed700c037e1eca7620097fdfbb1c9b358601e40169adf6d364/Babel-2.10.3-py3-none-any.whl#sha256=ff56f4892c1c4bf0d814575ea23471c230d544203c7748e8c68f0089478d48eb
PASSED   [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_biased_homogeneous Best match: Babel 2.10.3
Processing Babel-2.10.3-py3-none-any.whl
Installing Babel-2.10.3-py3-none-any.whl to /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages
PASSED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_biased_bipartite Adding Babel 2.10.3 to easy-install.pth file
Installing pybabel script to /opt/conda/envs/pytorch-ci/bin

Installed /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/Babel-2.10.3-py3.7.egg
Searching for snowballstemmer>=2.0
Reading https://pypi.org/simple/snowballstemmer/
Downloading https://files.pythonhosted.org/packages/ed/dc/c02e01294f7265e63a7315fe086dd1df7dacb9f840a804da846b96d01b96/snowballstemmer-2.2.0-py2.py3-none-any.whl#sha256=c8e1716e83cc398ae16824e5572ae04e0d9fc2c6b985fb0f900f5f0c96ecba1a
Best match: snowballstemmer 2.2.0
Processing snowballstemmer-2.2.0-py2.py3-none-any.whl
Installing snowballstemmer-2.2.0-py2.py3-none-any.whl to /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages
PASSED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_homogeneous[False-in-coo] PASSED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_homogeneous[False-in-csr] Adding snowballstemmer 2.2.0 to easy-install.pth file

Installed /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/snowballstemmer-2.2.0-py3.7.egg
Searching for docutils<0.20,>=0.14
Reading https://pypi.org/simple/docutils/
PASSED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_homogeneous[False-in-csc] PASSED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_homogeneous[False-out-coo] Downloading https://files.pythonhosted.org/packages/93/69/e391bd51bc08ed9141ecd899a0ddb61ab6465309f1eb470905c0c8868081/docutils-0.19-py3-none-any.whl#sha256=5e1de4d849fee02c63b040a4a3fd567f4ab104defd8a5511fbbc24a8a017efbc
Best match: docutils 0.19
Processing docutils-0.19-py3-none-any.whl
Installing docutils-0.19-py3-none-any.whl to /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages
PASSED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_homogeneous[False-out-csr] PASSED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_homogeneous[False-out-csc] PASSED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_homogeneous[True-in-coo] Adding docutils 0.19 to easy-install.pth file
Installing rst2html4.py script to /opt/conda/envs/pytorch-ci/bin
Installing rstpep2html.py script to /opt/conda/envs/pytorch-ci/bin
Installing rst2man.py script to /opt/conda/envs/pytorch-ci/bin
Installing rst2xetex.py script to /opt/conda/envs/pytorch-ci/bin
Installing rst2s5.py script to /opt/conda/envs/pytorch-ci/bin
Installing rst2pseudoxml.py script to /opt/conda/envs/pytorch-ci/bin
Installing rst2html5.py script to /opt/conda/envs/pytorch-ci/bin
Installing rst2odt.py script to /opt/conda/envs/pytorch-ci/bin
Installing rst2latex.py script to /opt/conda/envs/pytorch-ci/bin
Installing rst2html.py script to /opt/conda/envs/pytorch-ci/bin
Installing rst2odt_prepstyles.py script to /opt/conda/envs/pytorch-ci/bin
Installing rst2xml.py script to /opt/conda/envs/pytorch-ci/bin
Installing docutils script to /opt/conda/envs/pytorch-ci/bin

Installed /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/docutils-0.19-py3.7.egg
Searching for Pygments>=2.12
Reading https://pypi.org/simple/Pygments/
PASSED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_homogeneous[True-in-csr] Downloading https://files.pythonhosted.org/packages/4f/82/672cd382e5b39ab1cd422a672382f08a1fb3d08d9e0c0f3707f33a52063b/Pygments-2.13.0-py3-none-any.whl#sha256=f643f331ab57ba3c9d89212ee4a2dabc6e94f117cf4eefde99a0574720d14c42
Best match: Pygments 2.13.0
Processing Pygments-2.13.0-py3-none-any.whl
PASSED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_homogeneous[True-in-csc] PASSED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_homogeneous[True-out-coo] Installing Pygments-2.13.0-py3-none-any.whl to /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages
PASSED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_homogeneous[True-out-csr] PASSED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_homogeneous[True-out-csc] PASSED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_sorted_homogeneous[in-csr] Adding Pygments 2.13.0 to easy-install.pth file
Installing pygmentize script to /opt/conda/envs/pytorch-ci/bin

Installed /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/Pygments-2.13.0-py3.7.egg
Searching for sphinxcontrib-qthelp
Reading https://pypi.org/simple/sphinxcontrib-qthelp/
PASSED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_sorted_homogeneous[in-csc] PASSED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_sorted_homogeneous[out-csr] PASSED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_sorted_homogeneous[out-csc] PASSED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_exclude_edges_heteroG[int32] Downloading https://files.pythonhosted.org/packages/2b/14/05f9206cf4e9cfca1afb5fd224c7cd434dcc3a433d6d9e4e0264d29c6cdb/sphinxcontrib_qthelp-1.0.3-py2.py3-none-any.whl#sha256=bd9fc24bcb748a8d51fd4ecaade681350aa63009a347a8c14e637895444dfab6
Best match: sphinxcontrib-qthelp 1.0.3
Processing sphinxcontrib_qthelp-1.0.3-py2.py3-none-any.whl
Installing sphinxcontrib_qthelp-1.0.3-py2.py3-none-any.whl to /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages
PASSED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_exclude_edges_heteroG[int64] Adding sphinxcontrib-qthelp 1.0.3 to easy-install.pth file

Installed /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/sphinxcontrib_qthelp-1.0.3-py3.7.egg
Searching for sphinxcontrib-serializinghtml>=1.1.5
Reading https://pypi.org/simple/sphinxcontrib-serializinghtml/
Downloading https://files.pythonhosted.org/packages/c6/77/5464ec50dd0f1c1037e3c93249b040c8fc8078fdda97530eeb02424b6eea/sphinxcontrib_serializinghtml-1.1.5-py2.py3-none-any.whl#sha256=352a9a00ae864471d3a7ead8d7d79f5fc0b57e8b3f95e9867eb9eb28999b92fd
Best match: sphinxcontrib-serializinghtml 1.1.5
Processing sphinxcontrib_serializinghtml-1.1.5-py2.py3-none-any.whl
Installing sphinxcontrib_serializinghtml-1.1.5-py2.py3-none-any.whl to /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages
PASSED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_exclude_edges_homoG[int32] PASSED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_exclude_edges_homoG[int64] PASSED [ 21%]
tests/compute/test_sampling.py::test_global_uniform_negative_sampling[int32] Adding sphinxcontrib-serializinghtml 1.1.5 to easy-install.pth file

Installed /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/sphinxcontrib_serializinghtml-1.1.5-py3.7.egg
Searching for sphinxcontrib-htmlhelp>=2.0.0
Reading https://pypi.org/simple/sphinxcontrib-htmlhelp/
Downloading https://files.pythonhosted.org/packages/63/40/c854ef09500e25f6432dcbad0f37df87fd7046d376272292d8654cc71c95/sphinxcontrib_htmlhelp-2.0.0-py2.py3-none-any.whl#sha256=d412243dfb797ae3ec2b59eca0e52dac12e75a241bf0e4eb861e450d06c6ed07
Best match: sphinxcontrib-htmlhelp 2.0.0
Processing sphinxcontrib_htmlhelp-2.0.0-py2.py3-none-any.whl
Installing sphinxcontrib_htmlhelp-2.0.0-py2.py3-none-any.whl to /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages
PASSED [ 21%]
tests/compute/test_sampling.py::test_global_uniform_negative_sampling[int64] Adding sphinxcontrib-htmlhelp 2.0.0 to easy-install.pth file

Installed /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/sphinxcontrib_htmlhelp-2.0.0-py3.7.egg
Searching for sphinxcontrib-jsmath
Reading https://pypi.org/simple/sphinxcontrib-jsmath/
PASSED [ 22%]
tests/compute/test_serialize.py::test_graph_serialize_with_feature[True] PASSED [ 22%]
tests/compute/test_serialize.py::test_graph_serialize_with_feature[False] Downloading https://files.pythonhosted.org/packages/c2/42/4c8646762ee83602e3fb3fbe774c2fac12f317deb0b5dbeeedd2d3ba4b77/sphinxcontrib_jsmath-1.0.1-py2.py3-none-any.whl#sha256=2ec2eaebfb78f3f2078e73666b1415417a116cc848b72e5172e596c871103178
Best match: sphinxcontrib-jsmath 1.0.1
Processing sphinxcontrib_jsmath-1.0.1-py2.py3-none-any.whl
Installing sphinxcontrib_jsmath-1.0.1-py2.py3-none-any.whl to /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages
PASSED [ 22%]
tests/compute/test_serialize.py::test_graph_serialize_without_feature[True] PASSED [ 22%]
tests/compute/test_serialize.py::test_graph_serialize_without_feature[False] PASSED [ 22%]
tests/compute/test_serialize.py::test_graph_serialize_with_labels[True] PASSED [ 22%]
tests/compute/test_serialize.py::test_graph_serialize_with_labels[False] PASSED [ 22%]
tests/compute/test_serialize.py::test_serialize_tensors PASSED           [ 22%]
tests/compute/test_serialize.py::test_serialize_empty_dict PASSED        [ 22%]
tests/compute/test_serialize.py::test_load_old_files1 PASSED             [ 22%]
tests/compute/test_serialize.py::test_load_old_files2 PASSED             [ 22%]
tests/compute/test_serialize.py::test_deserialize_old_heterograph_file PASSED [ 22%]
tests/compute/test_serialize.py::test_serialize_heterograph Adding sphinxcontrib-jsmath 1.0.1 to easy-install.pth file

Installed /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/sphinxcontrib_jsmath-1.0.1-py3.7.egg
Searching for sphinxcontrib-devhelp
Reading https://pypi.org/simple/sphinxcontrib-devhelp/
Downloading https://files.pythonhosted.org/packages/c5/09/5de5ed43a521387f18bdf5f5af31d099605c992fd25372b2b9b825ce48ee/sphinxcontrib_devhelp-1.0.2-py2.py3-none-any.whl#sha256=8165223f9a335cc1af7ffe1ed31d2871f325254c0423bc0c4c7cd1c1e4734a2e
Best match: sphinxcontrib-devhelp 1.0.2
Processing sphinxcontrib_devhelp-1.0.2-py2.py3-none-any.whl
Installing sphinxcontrib_devhelp-1.0.2-py2.py3-none-any.whl to /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages
PASSED       [ 22%]
tests/compute/test_serialize.py::test_serialize_heterograph_s3 SKIPPED   [ 22%]
tests/compute/test_shared_mem.py::test_single_process[idtype0] SKIPPED   [ 22%]
tests/compute/test_shared_mem.py::test_single_process[idtype1] SKIPPED   [ 22%]
tests/compute/test_shared_mem.py::test_multi_process[idtype0] SKIPPED    [ 22%]
tests/compute/test_shared_mem.py::test_multi_process[idtype1] SKIPPED    [ 22%]
tests/compute/test_shared_mem.py::test_copy_from_gpu SKIPPED (Need g...) [ 22%]
tests/compute/test_sort.py::test_sort_with_tag[idtype0] PASSED           [ 22%]
tests/compute/test_sort.py::test_sort_with_tag[idtype1] Adding sphinxcontrib-devhelp 1.0.2 to easy-install.pth file

Installed /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/sphinxcontrib_devhelp-1.0.2-py3.7.egg
Searching for sphinxcontrib-applehelp
Reading https://pypi.org/simple/sphinxcontrib-applehelp/
Downloading https://files.pythonhosted.org/packages/dc/47/86022665a9433d89a66f5911b558ddff69861766807ba685de2e324bd6ed/sphinxcontrib_applehelp-1.0.2-py2.py3-none-any.whl#sha256=806111e5e962be97c29ec4c1e7fe277bfd19e9652fb1a4392105b43e01af885a
Best match: sphinxcontrib-applehelp 1.0.2
Processing sphinxcontrib_applehelp-1.0.2-py2.py3-none-any.whl
Installing sphinxcontrib_applehelp-1.0.2-py2.py3-none-any.whl to /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages
PASSED           [ 22%]
tests/compute/test_sort.py::test_sort_with_tag_bipartite[idtype0] PASSED [ 22%]
tests/compute/test_sort.py::test_sort_with_tag_bipartite[idtype1] PASSED [ 22%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-add-shp0-g0] Adding sphinxcontrib-applehelp 1.0.2 to easy-install.pth file

Installed /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/sphinxcontrib_applehelp-1.0.2-py3.7.egg
Searching for scikit-learn==1.0.2
Best match: scikit-learn 1.0.2
Adding scikit-learn 1.0.2 to easy-install.pth file

Using /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages
Searching for ogb==1.3.3
Best match: ogb 1.3.3
Adding ogb 1.3.3 to easy-install.pth file

Using /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages
Searching for threadpoolctl==3.1.0
Best match: threadpoolctl 3.1.0
Adding threadpoolctl 3.1.0 to easy-install.pth file

Using /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages
Searching for joblib==1.1.0
Best match: joblib 1.1.0
Adding joblib 1.1.0 to easy-install.pth file

Using /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages
Searching for scipy==1.7.3
Best match: scipy 1.7.3
Adding scipy 1.7.3 to easy-install.pth file

Using /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages
Searching for numpy==1.21.6
Best match: numpy 1.21.6
Adding numpy 1.21.6 to easy-install.pth file
Installing f2py script to /opt/conda/envs/pytorch-ci/bin
Installing f2py3 script to /opt/conda/envs/pytorch-ci/bin
Installing f2py3.7 script to /opt/conda/envs/pytorch-ci/bin

Using /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages
Searching for Pillow==9.2.0
Best match: Pillow 9.2.0
Adding Pillow 9.2.0 to easy-install.pth file

Using /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages
Searching for outdated==0.2.1
Best match: outdated 0.2.1
Adding outdated 0.2.1 to easy-install.pth file

Using /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages
Searching for urllib3==1.26.11
Best match: urllib3 1.26.11
Adding urllib3 1.26.11 to easy-install.pth file

Using /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages
Searching for six==1.16.0
Best match: six 1.16.0
Adding six 1.16.0 to easy-install.pth file

Using /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages
Searching for pandas==1.1.5
Best match: pandas 1.1.5
Adding pandas 1.1.5 to easy-install.pth file

Using /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages
Searching for tqdm==4.64.0
Best match: tqdm 4.64.0
Adding tqdm 4.64.0 to easy-install.pth file
Installing tqdm script to /opt/conda/envs/pytorch-ci/bin

Using /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages
Searching for torch==1.9.0+cpu
Best match: torch 1.9.0+cpu
Adding torch 1.9.0+cpu to easy-install.pth file
Installing convert-caffe2-to-onnx script to /opt/conda/envs/pytorch-ci/bin
Installing convert-onnx-to-caffe2 script to /opt/conda/envs/pytorch-ci/bin

Using /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages
Searching for typing-extensions==4.3.0
Best match: typing-extensions 4.3.0
Adding typing-extensions 4.3.0 to easy-install.pth file

Using /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages
Searching for click==8.1.3
Best match: click 8.1.3
Adding click 8.1.3 to easy-install.pth file

Using /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages
Searching for requests==2.28.1
Best match: requests 2.28.1
Adding requests 2.28.1 to easy-install.pth file

Using /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages
Searching for littleutils==0.2.2
Best match: littleutils 0.2.2
Adding littleutils 0.2.2 to easy-install.pth file

Using /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages
Searching for pytz==2022.2.1
Best match: pytz 2022.2.1
Adding pytz 2022.2.1 to easy-install.pth file

Using /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages
Searching for python-dateutil==2.8.2
Best match: python-dateutil 2.8.2
Adding python-dateutil 2.8.2 to easy-install.pth file

Using /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages
Searching for importlib-metadata==4.12.0
Best match: importlib-metadata 4.12.0
Adding importlib-metadata 4.12.0 to easy-install.pth file

Using /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages
Searching for packaging==21.3
Best match: packaging 21.3
Adding packaging 21.3 to easy-install.pth file

Using /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages
Searching for certifi==2022.6.15
Best match: certifi 2022.6.15
Adding certifi 2022.6.15 to easy-install.pth file

Using /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages
Searching for idna==3.3
Best match: idna 3.3
Adding idna 3.3 to easy-install.pth file

Using /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages
Searching for charset-normalizer==2.1.0
Best match: charset-normalizer 2.1.0
Adding charset-normalizer 2.1.0 to easy-install.pth file
Installing normalizer script to /opt/conda/envs/pytorch-ci/bin

Using /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages
Searching for zipp==3.8.1
Best match: zipp 3.8.1
Adding zipp 3.8.1 to easy-install.pth file

Using /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages
Searching for pyparsing==3.0.9
Best match: pyparsing 3.0.9
Adding pyparsing 3.0.9 to easy-install.pth file

Using /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages
Finished processing dependencies for dglgo==0.0.2
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/pkg_resources/__init__.py:125: PkgResourcesDeprecationWarning:  is an invalid version and will not be supported in a future release
  PkgResourcesDeprecationWarning,
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/pkg_resources/__init__.py:125: PkgResourcesDeprecationWarning: post1 is an invalid version and will not be supported in a future release
  PkgResourcesDeprecationWarning,
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/pkg_resources/__init__.py:125: PkgResourcesDeprecationWarning: qthelp-1.0.0 is an invalid version and will not be supported in a future release
  PkgResourcesDeprecationWarning,
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/pkg_resources/__init__.py:125: PkgResourcesDeprecationWarning: qthelp-1.0.1 is an invalid version and will not be supported in a future release
  PkgResourcesDeprecationWarning,
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/pkg_resources/__init__.py:125: PkgResourcesDeprecationWarning: qthelp-1.0.2 is an invalid version and will not be supported in a future release
  PkgResourcesDeprecationWarning,
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/pkg_resources/__init__.py:125: PkgResourcesDeprecationWarning: qthelp-1.0.3 is an invalid version and will not be supported in a future release
  PkgResourcesDeprecationWarning,
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/pkg_resources/__init__.py:125: PkgResourcesDeprecationWarning: serializinghtml-1.0.0 is an invalid version and will not be supported in a future release
  PkgResourcesDeprecationWarning,
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/pkg_resources/__init__.py:125: PkgResourcesDeprecationWarning: serializinghtml-1.1.0 is an invalid version and will not be supported in a future release
  PkgResourcesDeprecationWarning,
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/pkg_resources/__init__.py:125: PkgResourcesDeprecationWarning: serializinghtml-1.1.1 is an invalid version and will not be supported in a future release
  PkgResourcesDeprecationWarning,
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/pkg_resources/__init__.py:125: PkgResourcesDeprecationWarning: serializinghtml-1.1.3 is an invalid version and will not be supported in a future release
  PkgResourcesDeprecationWarning,
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/pkg_resources/__init__.py:125: PkgResourcesDeprecationWarning: serializinghtml-1.1.4 is an invalid version and will not be supported in a future release
  PkgResourcesDeprecationWarning,
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/pkg_resources/__init__.py:125: PkgResourcesDeprecationWarning: serializinghtml-1.1.5 is an invalid version and will not be supported in a future release
  PkgResourcesDeprecationWarning,
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/pkg_resources/__init__.py:125: PkgResourcesDeprecationWarning: htmlhelp-1.0.0 is an invalid version and will not be supported in a future release
  PkgResourcesDeprecationWarning,
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/pkg_resources/__init__.py:125: PkgResourcesDeprecationWarning: htmlhelp-1.0.1 is an invalid version and will not be supported in a future release
  PkgResourcesDeprecationWarning,
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/pkg_resources/__init__.py:125: PkgResourcesDeprecationWarning: htmlhelp-1.0.2 is an invalid version and will not be supported in a future release
  PkgResourcesDeprecationWarning,
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/pkg_resources/__init__.py:125: PkgResourcesDeprecationWarning: htmlhelp-1.0.3 is an invalid version and will not be supported in a future release
  PkgResourcesDeprecationWarning,
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/pkg_resources/__init__.py:125: PkgResourcesDeprecationWarning: htmlhelp-2.0.0 is an invalid version and will not be supported in a future release
  PkgResourcesDeprecationWarning,
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/pkg_resources/__init__.py:125: PkgResourcesDeprecationWarning: jsmath-1.0.0 is an invalid version and will not be supported in a future release
  PkgResourcesDeprecationWarning,
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/pkg_resources/__init__.py:125: PkgResourcesDeprecationWarning: jsmath-1.0.1 is an invalid version and will not be supported in a future release
  PkgResourcesDeprecationWarning,
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/pkg_resources/__init__.py:125: PkgResourcesDeprecationWarning: devhelp-1.0.0 is an invalid version and will not be supported in a future release
  PkgResourcesDeprecationWarning,
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/pkg_resources/__init__.py:125: PkgResourcesDeprecationWarning: devhelp-1.0.1 is an invalid version and will not be supported in a future release
  PkgResourcesDeprecationWarning,
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/pkg_resources/__init__.py:125: PkgResourcesDeprecationWarning: devhelp-1.0.2 is an invalid version and will not be supported in a future release
  PkgResourcesDeprecationWarning,
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/pkg_resources/__init__.py:125: PkgResourcesDeprecationWarning: applehelp-1.0.0 is an invalid version and will not be supported in a future release
  PkgResourcesDeprecationWarning,
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/pkg_resources/__init__.py:125: PkgResourcesDeprecationWarning: applehelp-1.0.1 is an invalid version and will not be supported in a future release
  PkgResourcesDeprecationWarning,
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/pkg_resources/__init__.py:125: PkgResourcesDeprecationWarning: applehelp-1.0.2 is an invalid version and will not be supported in a future release
  PkgResourcesDeprecationWarning,
~/jenkins/workspace/dgl_PR-4648@3
PASSED  [ 22%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-add-shp0-g1] Collecting psutil
  Downloading psutil-5.9.2-cp37-cp37m-manylinux_2_12_x86_64.manylinux2010_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl (281 kB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 281.3/281.3 kB 10.4 MB/s eta 0:00:00
Installing collected packages: psutil
PASSED  [ 22%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-add-shp1-g0] Successfully installed psutil-5.9.2
WARNING: Running pip as the 'root' user can result in broken permissions and conflicting behaviour with the system package manager. It is recommended to use a virtual environment instead: https://pip.pypa.io/warnings/venv
============================= test session starts ==============================
platform linux -- Python 3.7.0, pytest-7.1.2, pluggy-1.0.0 -- /opt/conda/envs/pytorch-ci/bin/python3
cachedir: .pytest_cache
rootdir: /root/jenkins/workspace/dgl_PR-4648@3
collecting ... PASSED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-add-shp1-g1] PASSED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-add-shp2-g0] PASSED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-add-shp2-g1] collected 81 items

tests/go/test_model.py::test_gcn[g0] FAILED                              [  1%]
tests/go/test_model.py::test_gcn_block[g0] FAILED                        [  2%]
tests/go/test_model.py::test_gat[g0] FAILED                              [  3%]
tests/go/test_model.py::test_gat_block[g0] FAILED                        [  4%]
tests/go/test_model.py::test_gin[g0] FAILED                              [  6%]
tests/go/test_model.py::test_sage[g0] PASSED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-add-shp3-g0] FAILED                             [  7%]
tests/go/test_model.py::test_sage_block[g0] FAILED                       [  8%]
tests/go/test_model.py::test_sgc[g0] FAILED                              [  9%]
tests/go/test_model.py::test_bilinear PASSED                             [ 11%]
tests/go/test_model.py::test_ele PASSED                                  [ 12%]
tests/go/test_model.py::test_ogbg_gin[True] FAILED                       [ 13%]
tests/go/test_model.py::test_ogbg_gin[False] PASSED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-add-shp3-g1] FAILED                      [ 14%]
tests/go/test_model.py::test_pna FAILED                                  [ 16%]
tests/go/test_pipeline.py::test_nodepred_data[cora] PASSED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-add-shp4-g0] PASSED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-add-shp4-g1] Converting to homogeneous graph takes 0.009s, peak mem: 1.841 GB
Convert a graph into a bidirected graph: 0.018 seconds, peak memory: 1.841 GB
Construct multi-constraint weights: 0.001 seconds, peak memory: 1.841 GB
[05:35:49] /root/jenkins/workspace/dgl_PR-4648/src/graph/transform/metis_partition_hetero.cc:87: Partition a graph with 30030 nodes and 601200 edges into 3 parts and get 152443 edge cuts
Metis partitioning: 0.212 seconds, peak memory: 1.841 GB
Assigning nodes to METIS partitions takes 0.232s, peak mem: 1.841 GB
Reshuffle nodes and edges: 0.018 seconds
Split the graph: 0.058 seconds
Construct subgraphs: 0.056 seconds
Splitting the graph into partitions takes 0.132s, peak mem: 1.841 GB
part 0 has 9953 nodes of type n1 and 3136 are inside the partition
part 0 has 9963 nodes of type n2 and 3140 are inside the partition
part 0 has 9959 nodes of type n3 and 3142 are inside the partition
part 0 has 47353 edges of type r1 and 31289 are inside the partition
part 0 has 47698 edges of type r2 and 31241 are inside the partition
part 0 has 47467 edges of type r3 and 31243 are inside the partition
part 1 has 9968 nodes of type n1 and 3432 are inside the partition
part 1 has 9983 nodes of type n2 and 3435 are inside the partition
part 1 has 9990 nodes of type n3 and 3440 are inside the partition
part 1 has 51615 edges of type r1 and 34407 are inside the partition
part 1 has 51895 edges of type r2 and 34446 are inside the partition
part 1 has 51968 edges of type r3 and 34664 are inside the partition
part 2 has 9975 nodes of type n1 and 3432 are inside the partition
part 2 has 9988 nodes of type n2 and 3435 are inside the partition
part 2 has 9983 nodes of type n3 and 3438 are inside the partition
part 2 has 51626 edges of type r1 and 34404 are inside the partition
part 2 has 51781 edges of type r2 and 34513 are inside the partition
part 2 has 51640 edges of type r3 and 34393 are inside the partition
Save partitions: 0.047 seconds, peak memory: 1.841 GB
There are 300600 edges in the graph and 0 edge cuts for 3 partitions.
server: #clients=5
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_sampling
start graph service on server 0 for part 0
[05:35:50] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:35:50] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:21270]...
PASSED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-add-shp5-g0] server: #clients=5
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_sampling
start graph service on server 1 for part 1
[05:35:51] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:35:51] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:21273]...
PASSED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-add-shp5-g1] PASSED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-sub-shp0-g0] server: #clients=5
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_sampling
start graph service on server 2 for part 2
[05:35:52] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:35:52] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:21276]...
PASSED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-sub-shp0-g1] PASSED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-sub-shp1-g0] [05:35:53] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:35:53] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
PASSED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-sub-shp1-g1] [05:35:54] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:35:54] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
[05:35:54] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:35:54] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
[05:35:54] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:35:54] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
[05:35:54] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:35:54] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
PASSED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-sub-shp2-g0] PASSED               [ 17%]
tests/go/test_pipeline.py::test_nodepred_data[citeseer] PASSED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-sub-shp2-g1] PASSED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-sub-shp3-g0] PASSED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-sub-shp3-g1] Client [2460] waits on 172.17.0.3:57381Client [2457] waits on 172.17.0.3:55343

Client [2458] waits on 172.17.0.3:57261
Client [2459] waits on 172.17.0.3:46473
Client [2451] waits on 172.17.0.3:49357
Machine (0) group (0) client (0) connect to server successfuly!
Machine (0) group (0) client (2) connect to server successfuly!
Machine (0) group (0) client (4) connect to server successfuly!
Machine (0) group (0) client (3) connect to server successfuly!
Machine (0) group (0) client (1) connect to server successfuly!
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
PASSED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-sub-shp4-g0] PASSED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-sub-shp4-g1] Client[1] in group[0] is exiting...
Client[2] in group[0] is exiting...
Client[3] in group[0] is exiting...
PASSED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-sub-shp5-g0] Client[0] in group[0] is exiting...
PASSED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-sub-shp5-g1] Client[4] in group[0] is exiting...
PASSED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-mul-shp0-g0] Server (0) shutdown.
Server is exiting...
PASSED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-mul-shp0-g1] Server (2) shutdown.
Server is exiting...
Server (1) shutdown.
Server is exiting...
PASSED  [ 23%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-mul-shp1-g0] PASSED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-mul-shp1-g1] PASSED           [ 18%]
tests/go/test_pipeline.py::test_nodepred_data[pubmed] PASSED [ 50%]
tests/distributed/test_mp_dataloader.py::test_dataloader[edge-0-3]   NumNodes: 2708
  NumEdges: 10556
  NumFeats: 1433
  NumClasses: 7
  NumTrainingSamples: 140
  NumValidationSamples: 500
  NumTestSamples: 1000
Done loading data from cached files.
Converting to homogeneous graph takes 0.002s, peak mem: 1.841 GB
Convert a graph into a bidirected graph: 0.001 seconds, peak memory: 1.841 GB
Construct multi-constraint weights: 0.000 seconds, peak memory: 1.841 GB
[05:36:03] /root/jenkins/workspace/dgl_PR-4648/src/graph/transform/metis_partition_hetero.cc:87: Partition a graph with 2708 nodes and 10556 edges into 3 parts and get 325 edge cuts
Metis partitioning: 0.003 seconds, peak memory: 1.841 GB
Assigning nodes to METIS partitions takes 0.004s, peak mem: 1.841 GB
Reshuffle nodes and edges: 0.001 seconds
Split the graph: 0.002 seconds
Construct subgraphs: 0.004 seconds
Splitting the graph into partitions takes 0.008s, peak mem: 1.841 GB
part 0 has 1046 nodes and 903 are inside the partition
part 0 has 3490 edges and 3317 are inside the partition
part 1 has 1064 nodes and 902 are inside the partition
part 1 has 4006 edges and 3739 are inside the partition
part 2 has 1056 nodes and 903 are inside the partition
part 2 has 3710 edges and 3500 are inside the partition
Save partitions: 0.028 seconds, peak memory: 1.841 GB
There are 10556 edges in the graph and 0 edge cuts for 3 partitions.
PASSED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-mul-shp2-g0] server: #clients=1
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_sampling
start graph service on server 0 for part 0
[05:36:05] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:36:05] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:29782]...
PASSED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-mul-shp2-g1] PASSED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-mul-shp3-g0] server: #clients=1
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_sampling
start graph service on server 1 for part 1
[05:36:06] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:36:06] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:29785]...
PASSED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-mul-shp3-g1] server: #clients=1
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_sampling
start graph service on server 2 for part 2
[05:36:07] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:36:07] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:29788]...
PASSED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-mul-shp4-g0] [05:36:08] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:36:08] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
PASSED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-mul-shp4-g1] PASSED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-mul-shp5-g0] PASSED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-mul-shp5-g1] PASSED              [ 20%]
tests/compute/test_sampling.py::test_non_uniform_random_walk[True] SKIPPED [ 20%]
tests/compute/test_sampling.py::test_non_uniform_random_walk[False] PASSED [ 20%]
tests/compute/test_sampling.py::test_uniform_random_walk[True] SKIPPED   [ 20%]
tests/compute/test_sampling.py::test_uniform_random_walk[False] PASSED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-div-shp0-g0] PASSED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-div-shp0-g1] Client [2559] waits on 172.17.0.3:41249
Machine (0) group (0) client (0) connect to server successfuly!
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
PASSED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-div-shp1-g0] PASSED             [ 19%]
tests/go/test_pipeline.py::test_nodepred_data[csv] Client[0] in group[0] is exiting...
PASSED   [ 20%]
tests/compute/test_sampling.py::test_node2vec PASSED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-div-shp1-g1] PASSED                     [ 20%]
tests/compute/test_sampling.py::test_pack_traces PASSED                  [ 20%]
tests/compute/test_sampling.py::test_pinsage_sampling[True] SKIPPED      [ 20%]
tests/compute/test_sampling.py::test_pinsage_sampling[False] Server (1) shutdown.
Server is exiting...
Server (0) shutdown.
Server is exiting...
Server (2) shutdown.
Server is exiting...
PASSED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-div-shp2-g0] PASSED      [ 20%]
tests/compute/test_sampling.py::test_sample_neighbors_noprob PASSED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-div-shp2-g1] PASSED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-div-shp3-g0] PASSED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-div-shp3-g1] PASSED      [ 20%]
tests/compute/test_sampling.py::test_sample_neighbors_prob PASSED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-div-shp4-g0] PASSED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-div-shp4-g1] PASSED        [ 20%]
tests/compute/test_sampling.py::test_sample_neighbors_outedge PASSED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-div-shp5-g0] PASSED                [ 20%]
tests/go/test_pipeline.py::test_nodepred_data[reddit] PASSED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-div-shp5-g1] PASSED  [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_lhs-shp0-g0] PASSED [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_lhs-shp0-g1] PASSED     [ 20%]
tests/compute/test_sampling.py::test_sample_neighbors_topk PASSED        [ 20%]
tests/compute/test_sampling.py::test_sample_neighbors_topk_outedge PASSED [ 20%]
tests/compute/test_sampling.py::test_sample_neighbors_with_0deg PASSED   [ 20%]
tests/compute/test_sampling.py::test_sample_neighbors_biased_homogeneous PASSED [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_lhs-shp1-g0] PASSED [ 20%]
tests/compute/test_sampling.py::test_sample_neighbors_biased_bipartite PASSED [ 24%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_lhs-shp1-g1] PASSED [ 20%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_homogeneous[False-in-coo] PASSED [ 20%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_homogeneous[False-in-csr] PASSED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_homogeneous[False-in-csc] PASSED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_homogeneous[False-out-coo] PASSED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_homogeneous[False-out-csr] PASSED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_homogeneous[False-out-csc] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_lhs-shp2-g0] PASSED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_homogeneous[True-in-coo] PASSED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_homogeneous[True-in-csr] PASSED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_homogeneous[True-in-csc] PASSED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_homogeneous[True-out-coo] PASSED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_homogeneous[True-out-csr] PASSED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_homogeneous[True-out-csc] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_lhs-shp2-g1] PASSED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_sorted_homogeneous[in-csr] PASSED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_sorted_homogeneous[in-csc] PASSED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_sorted_homogeneous[out-csr] PASSED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_etype_sorted_homogeneous[out-csc] PASSED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_exclude_edges_heteroG[int32] PASSED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_exclude_edges_heteroG[int64] PASSED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_exclude_edges_homoG[int32] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_lhs-shp3-g0] PASSED [ 21%]
tests/compute/test_sampling.py::test_sample_neighbors_exclude_edges_homoG[int64] PASSED [ 21%]
tests/compute/test_sampling.py::test_global_uniform_negative_sampling[int32] PASSED [ 21%]
tests/compute/test_sampling.py::test_global_uniform_negative_sampling[int64] PASSED [ 21%]
tests/compute/test_serialize.py::test_graph_serialize_with_feature[True] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_lhs-shp3-g1] PASSED [ 21%]
tests/compute/test_serialize.py::test_graph_serialize_with_feature[False] PASSED [ 21%]
tests/compute/test_serialize.py::test_graph_serialize_without_feature[True] PASSED [ 21%]
tests/compute/test_serialize.py::test_graph_serialize_without_feature[False] PASSED [ 22%]
tests/compute/test_serialize.py::test_graph_serialize_with_labels[True] PASSED             [ 22%]
tests/go/test_pipeline.py::test_nodepred_data[co-buy-computer] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_lhs-shp4-g0] PASSED [ 22%]
tests/compute/test_serialize.py::test_graph_serialize_with_labels[False] PASSED [ 22%]
tests/compute/test_serialize.py::test_serialize_tensors PASSED           [ 22%]
tests/compute/test_serialize.py::test_serialize_empty_dict PASSED        [ 22%]
tests/compute/test_serialize.py::test_load_old_files1 PASSED             [ 22%]
tests/compute/test_serialize.py::test_load_old_files2 PASSED             [ 22%]
tests/compute/test_serialize.py::test_deserialize_old_heterograph_file PASSED [ 22%]
tests/compute/test_serialize.py::test_serialize_heterograph PASSED       [ 22%]
tests/compute/test_serialize.py::test_serialize_heterograph_s3 SKIPPED   [ 22%]
tests/compute/test_shared_mem.py::test_single_process[int32] PASSED      [ 22%]
tests/compute/test_shared_mem.py::test_single_process[int64] PASSED      [ 22%]
tests/compute/test_shared_mem.py::test_multi_process[int32] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_lhs-shp4-g1] PASSED       [ 22%]
tests/compute/test_shared_mem.py::test_multi_process[int64] PASSED       [ 22%]
tests/compute/test_shared_mem.py::test_copy_from_gpu SKIPPED (Need g...) [ 22%]
tests/compute/test_sort.py::test_sort_with_tag[int32] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_lhs-shp5-g0] PASSED             [ 22%]
tests/compute/test_sort.py::test_sort_with_tag[int64] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_lhs-shp5-g1] PASSED             [ 22%]
tests/compute/test_sort.py::test_sort_with_tag_bipartite[int32] PASSED   [ 22%]
tests/compute/test_sort.py::test_sort_with_tag_bipartite[int64] PASSED   [ 22%]
tests/compute/test_sparse.py::test_spmm[int32-sum-add-shp0-g0] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_rhs-shp0-g0] PASSED    [ 22%]
tests/compute/test_sparse.py::test_spmm[int32-sum-add-shp0-g1] Converting to homogeneous graph takes 0.011s, peak mem: 1.841 GB
Convert a graph into a bidirected graph: 0.019 seconds, peak memory: 1.841 GB
Construct multi-constraint weights: 0.001 seconds, peak memory: 1.841 GB
[05:36:28] /root/jenkins/workspace/dgl_PR-4648/src/graph/transform/metis_partition_hetero.cc:87: Partition a graph with 30030 nodes and 601200 edges into 3 parts and get 152443 edge cuts
Metis partitioning: 0.185 seconds, peak memory: 1.841 GB
Assigning nodes to METIS partitions takes 0.205s, peak mem: 1.841 GB
Reshuffle nodes and edges: 0.021 seconds
Split the graph: 0.055 seconds
Construct subgraphs: 0.048 seconds
Splitting the graph into partitions takes 0.124s, peak mem: 1.841 GB
part 0 has 9953 nodes of type n1 and 3136 are inside the partition
part 0 has 9963 nodes of type n2 and 3140 are inside the partition
part 0 has 9959 nodes of type n3 and 3142 are inside the partition
part 0 has 47353 edges of type r1 and 31289 are inside the partition
part 0 has 47698 edges of type r2 and 31241 are inside the partition
part 0 has 47467 edges of type r3 and 31243 are inside the partition
part 1 has 9968 nodes of type n1 and 3432 are inside the partition
part 1 has 9983 nodes of type n2 and 3435 are inside the partition
part 1 has 9990 nodes of type n3 and 3440 are inside the partition
part 1 has 51615 edges of type r1 and 34407 are inside the partition
part 1 has 51895 edges of type r2 and 34446 are inside the partition
part 1 has 51968 edges of type r3 and 34664 are inside the partition
part 2 has 9975 nodes of type n1 and 3432 are inside the partition
part 2 has 9988 nodes of type n2 and 3435 are inside the partition
part 2 has 9983 nodes of type n3 and 3438 are inside the partition
part 2 has 51626 edges of type r1 and 34404 are inside the partition
part 2 has 51781 edges of type r2 and 34513 are inside the partition
part 2 has 51640 edges of type r3 and 34393 are inside the partition
Save partitions: 0.033 seconds, peak memory: 1.841 GB
There are 300600 edges in the graph and 0 edge cuts for 3 partitions.
PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_rhs-shp0-g1] PASSED    [ 22%]
tests/compute/test_sparse.py::test_spmm[int32-sum-add-shp1-g0] PASSED    [ 22%]
tests/compute/test_sparse.py::test_spmm[int32-sum-add-shp1-g1] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_rhs-shp1-g0] server: #clients=1
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_sampling
start graph service on server 0 for part 0
[05:36:29] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:36:29] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:18667]...
PASSED    [ 22%]
tests/compute/test_sparse.py::test_spmm[int32-sum-add-shp2-g0] PASSED    [ 22%]
tests/compute/test_sparse.py::test_spmm[int32-sum-add-shp2-g1] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_rhs-shp1-g1] server: #clients=1
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_sampling
start graph service on server 1 for part 1
[05:36:30] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:36:30] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:18670]...
PASSED    [ 22%]
tests/compute/test_sparse.py::test_spmm[int32-sum-add-shp3-g0] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_rhs-shp2-g0] PASSED    [ 23%]
tests/compute/test_sparse.py::test_spmm[int32-sum-add-shp3-g1] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_rhs-shp2-g1] server: #clients=1
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
PASSED    [ 23%]
tests/compute/test_sparse.py::test_spmm[int32-sum-add-shp4-g0] load test_sampling
start graph service on server 2 for part 2
[05:36:31] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:36:31] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:18673]...
PASSED    [ 23%]
tests/go/test_pipeline.py::test_nodepred_data[ogbn-arxiv] PASSED    [ 23%]
tests/compute/test_sparse.py::test_spmm[int32-sum-add-shp4-g1] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_rhs-shp3-g0] PASSED    [ 23%]
tests/compute/test_sparse.py::test_spmm[int32-sum-add-shp5-g0] PASSED    [ 23%]
tests/compute/test_sparse.py::test_spmm[int32-sum-add-shp5-g1] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_rhs-shp3-g1] [05:36:32] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:36:32] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
PASSED    [ 23%]
tests/compute/test_sparse.py::test_spmm[int32-sum-sub-shp0-g0] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_rhs-shp4-g0] PASSED    [ 23%]
tests/compute/test_sparse.py::test_spmm[int32-sum-sub-shp0-g1] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_rhs-shp4-g1] PASSED    [ 23%]
tests/compute/test_sparse.py::test_spmm[int32-sum-sub-shp1-g0] PASSED    [ 23%]
tests/compute/test_sparse.py::test_spmm[int32-sum-sub-shp1-g1] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_rhs-shp5-g0] PASSED    [ 23%]
tests/compute/test_sparse.py::test_spmm[int32-sum-sub-shp2-g0] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_rhs-shp5-g1] PASSED    [ 23%]
tests/compute/test_sparse.py::test_spmm[int32-sum-sub-shp2-g1] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-add-shp0-g0] PASSED    [ 23%]
tests/compute/test_sparse.py::test_spmm[int32-sum-sub-shp3-g0] PASSED    [ 23%]
tests/compute/test_sparse.py::test_spmm[int32-sum-sub-shp3-g1] Client [2599] waits on 172.17.0.3:59965
Machine (0) group (0) client (0) connect to server successfuly!
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
PASSED  [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-add-shp0-g1] PASSED    [ 23%]
tests/compute/test_sparse.py::test_spmm[int32-sum-sub-shp4-g0] PASSED    [ 23%]
tests/compute/test_sparse.py::test_spmm[int32-sum-sub-shp4-g1] Client[0] in group[0] is exiting...
PASSED  [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-add-shp1-g0] PASSED    [ 23%]
tests/compute/test_sparse.py::test_spmm[int32-sum-sub-shp5-g0] PASSED    [ 23%]
tests/compute/test_sparse.py::test_spmm[int32-sum-sub-shp5-g1] PASSED  [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-add-shp1-g1] Server (2) shutdown.
Server is exiting...
Server (0) shutdown.
Server is exiting...
Server (1) shutdown.
Server is exiting...
PASSED    [ 23%]
tests/compute/test_sparse.py::test_spmm[int32-sum-mul-shp0-g0] PASSED  [ 25%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-add-shp2-g0] PASSED [ 52%]
tests/distributed/test_mp_dataloader.py::test_dataloader[edge-4-3]   NumNodes: 2708
  NumEdges: 10556
  NumFeats: 1433
  NumClasses: 7
  NumTrainingSamples: 140
  NumValidationSamples: 500
  NumTestSamples: 1000
Done loading data from cached files.
Converting to homogeneous graph takes 0.001s, peak mem: 1.841 GB
Convert a graph into a bidirected graph: 0.001 seconds, peak memory: 1.841 GB
Construct multi-constraint weights: 0.000 seconds, peak memory: 1.841 GB
[05:36:37] /root/jenkins/workspace/dgl_PR-4648/src/graph/transform/metis_partition_hetero.cc:87: Partition a graph with 2708 nodes and 10556 edges into 3 parts and get 325 edge cuts
Metis partitioning: 0.003 seconds, peak memory: 1.841 GB
Assigning nodes to METIS partitions takes 0.004s, peak mem: 1.841 GB
Reshuffle nodes and edges: 0.001 seconds
Split the graph: 0.002 seconds
Construct subgraphs: 0.004 seconds
Splitting the graph into partitions takes 0.008s, peak mem: 1.841 GB
PASSED    [ 23%]
tests/compute/test_sparse.py::test_spmm[int32-sum-mul-shp0-g1] part 0 has 1046 nodes and 903 are inside the partition
part 0 has 3490 edges and 3317 are inside the partition
part 1 has 1064 nodes and 902 are inside the partition
part 1 has 4006 edges and 3739 are inside the partition
part 2 has 1056 nodes and 903 are inside the partition
part 2 has 3710 edges and 3500 are inside the partition
Save partitions: 0.019 seconds, peak memory: 1.841 GB
There are 10556 edges in the graph and 0 edge cuts for 3 partitions.
PASSED         [ 24%]
tests/go/test_pipeline.py::test_nodepred_data[ogbn-products] PASSED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-add-shp2-g1] PASSED    [ 23%]
tests/compute/test_sparse.py::test_spmm[int32-sum-mul-shp1-g0] PASSED    [ 23%]
tests/compute/test_sparse.py::test_spmm[int32-sum-mul-shp1-g1] PASSED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-add-shp3-g0] server: #clients=5
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_sampling
start graph service on server 0 for part 0
[05:36:38] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:36:38] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:11423]...
PASSED    [ 23%]
tests/compute/test_sparse.py::test_spmm[int32-sum-mul-shp2-g0] PASSED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-add-shp3-g1] PASSED    [ 23%]
tests/compute/test_sparse.py::test_spmm[int32-sum-mul-shp2-g1] server: #clients=5
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_sampling
start graph service on server 1 for part 1
[05:36:39] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:36:39] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:11426]...
PASSED    [ 23%]
tests/compute/test_sparse.py::test_spmm[int32-sum-mul-shp3-g0] PASSED    [ 24%]
tests/compute/test_sparse.py::test_spmm[int32-sum-mul-shp3-g1] PASSED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-add-shp4-g0] PASSED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-add-shp4-g1] server: #clients=5
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_sampling
start graph service on server 2 for part 2
[05:36:40] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:36:40] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:11429]...
PASSED    [ 24%]
tests/compute/test_sparse.py::test_spmm[int32-sum-mul-shp4-g0] PASSED    [ 24%]
tests/compute/test_sparse.py::test_spmm[int32-sum-mul-shp4-g1] PASSED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-add-shp5-g0] PASSED    [ 24%]
tests/compute/test_sparse.py::test_spmm[int32-sum-mul-shp5-g0] PASSED    [ 24%]
tests/compute/test_sparse.py::test_spmm[int32-sum-mul-shp5-g1] PASSED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-add-shp5-g1] [05:36:41] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:36:41] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
PASSED    [ 24%]
tests/compute/test_sparse.py::test_spmm[int32-sum-div-shp0-g0] PASSED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-sub-shp0-g0] [[05:36:42] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc05:36:42:] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc140:: Sender with NetType~140socket:  is created.Sender with NetType~
socket is created.
[05:36:42] [/root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc05:36:42:] 159/root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc: :Receiver with NetType~159socket:  is created.Receiver with NetType~
socket is created.
[05:36:42] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:36:42] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
PASSED    [ 24%]
tests/compute/test_sparse.py::test_spmm[int32-sum-div-shp0-g1] [05:36:43] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:36:43] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
PASSED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-sub-shp0-g1] PASSED    [ 24%]
tests/compute/test_sparse.py::test_spmm[int32-sum-div-shp1-g0] PASSED    [ 24%]
tests/compute/test_sparse.py::test_spmm[int32-sum-div-shp1-g1] PASSED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-sub-shp1-g0] PASSED    [ 24%]
tests/compute/test_sparse.py::test_spmm[int32-sum-div-shp2-g0] PASSED    [ 24%]
tests/compute/test_sparse.py::test_spmm[int32-sum-div-shp2-g1] PASSED      [ 25%]
tests/go/test_pipeline.py::test_nodepred_model[gcn] PASSED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-sub-shp1-g1] PASSED    [ 24%]
tests/compute/test_sparse.py::test_spmm[int32-sum-div-shp3-g0] PASSED    [ 24%]
tests/compute/test_sparse.py::test_spmm[int32-sum-div-shp3-g1] PASSED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-sub-shp2-g0] PASSED    [ 24%]
tests/compute/test_sparse.py::test_spmm[int32-sum-div-shp4-g0] PASSED    [ 24%]
tests/compute/test_sparse.py::test_spmm[int32-sum-div-shp4-g1] PASSED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-sub-shp2-g1] Client [2635] waits on 172.17.0.3:43935Client [2636] waits on 172.17.0.3:34133

Client [2637] waits on 172.17.0.3:58657
Client [2629] waits on 172.17.0.3:55007
Client [2638] waits on 172.17.0.3:46993
Machine (0) group (0) client (1) connect to server successfuly!Machine (0) group (0) client (0) connect to server successfuly!

Machine (0) group (0) client (2) connect to server successfuly!
Machine (0) group (0) client (4) connect to server successfuly!
Machine (0) group (0) client (3) connect to server successfuly!
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
PASSED    [ 24%]
tests/compute/test_sparse.py::test_spmm[int32-sum-div-shp5-g0] PASSED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-sub-shp3-g0] Client[3] in group[0] is exiting...
Client[1] in group[0] is exiting...
PASSED    [ 24%]
tests/compute/test_sparse.py::test_spmm[int32-sum-div-shp5-g1] PASSED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-sub-shp3-g1] Client[0] in group[0] is exiting...
PASSED    [ 24%]
tests/compute/test_sparse.py::test_spmm[int32-sum-copy_lhs-shp0-g0] PASSED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-sub-shp4-g0] Client[4] in group[0] is exiting...
PASSED [ 24%]
tests/compute/test_sparse.py::test_spmm[int32-sum-copy_lhs-shp0-g1] PASSED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-sub-shp4-g1] Client[2] in group[0] is exiting...
PASSED [ 24%]
tests/compute/test_sparse.py::test_spmm[int32-sum-copy_lhs-shp1-g0] PASSED [ 24%]
tests/compute/test_sparse.py::test_spmm[int32-sum-copy_lhs-shp1-g1] PASSED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-sub-shp5-g0] PASSED [ 24%]
tests/compute/test_sparse.py::test_spmm[int32-sum-copy_lhs-shp2-g0] PASSED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-sub-shp5-g1] PASSED [ 24%]
tests/compute/test_sparse.py::test_spmm[int32-sum-copy_lhs-shp2-g1] Server (2) shutdown.
Server (0) shutdown.
Server is exiting...Server is exiting...

Server (1) shutdown.
Server is exiting...
PASSED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-mul-shp0-g0] PASSED [ 24%]
tests/compute/test_sparse.py::test_spmm[int32-sum-copy_lhs-shp3-g0] PASSED               [ 27%]
tests/go/test_pipeline.py::test_nodepred_model[gat] PASSED [ 24%]
tests/compute/test_sparse.py::test_spmm[int32-sum-copy_lhs-shp3-g1] PASSED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-mul-shp0-g1] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[int32-sum-copy_lhs-shp4-g0] PASSED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-mul-shp1-g0] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[int32-sum-copy_lhs-shp4-g1] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[int32-sum-copy_lhs-shp5-g0] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[int32-sum-copy_lhs-shp5-g1] PASSED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-mul-shp1-g1] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[int32-sum-copy_rhs-shp0-g0] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[int32-sum-copy_rhs-shp0-g1] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[int32-sum-copy_rhs-shp1-g0] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[int32-sum-copy_rhs-shp1-g1] PASSED  [ 26%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-mul-shp2-g0] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[int32-sum-copy_rhs-shp2-g0] PASSED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-mul-shp2-g1] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[int32-sum-copy_rhs-shp2-g1] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[int32-sum-copy_rhs-shp3-g0] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[int32-sum-copy_rhs-shp3-g1] PASSED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-mul-shp3-g0] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[int32-sum-copy_rhs-shp4-g0] PASSED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-mul-shp3-g1] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[int32-sum-copy_rhs-shp4-g1] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[int32-sum-copy_rhs-shp5-g0] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[int32-sum-copy_rhs-shp5-g1] PASSED [ 25%]
tests/compute/test_sparse.py::test_spmm[int32-min-add-shp0-g0] PASSED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-mul-shp4-g0] PASSED    [ 25%]
tests/compute/test_sparse.py::test_spmm[int32-min-add-shp0-g1] PASSED    [ 25%]
tests/compute/test_sparse.py::test_spmm[int32-min-add-shp1-g0] PASSED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-mul-shp4-g1] PASSED    [ 25%]
tests/compute/test_sparse.py::test_spmm[int32-min-add-shp1-g1] PASSED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-mul-shp5-g0] PASSED    [ 25%]
tests/compute/test_sparse.py::test_spmm[int32-min-add-shp2-g0] PASSED    [ 25%]
tests/compute/test_sparse.py::test_spmm[int32-min-add-shp2-g1] PASSED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-mul-shp5-g1] PASSED               [ 28%]
tests/go/test_pipeline.py::test_nodepred_model[sage] PASSED    [ 25%]
tests/compute/test_sparse.py::test_spmm[int32-min-add-shp3-g0] PASSED    [ 25%]
tests/compute/test_sparse.py::test_spmm[int32-min-add-shp3-g1] PASSED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-div-shp0-g0] PASSED    [ 26%]
tests/compute/test_sparse.py::test_spmm[int32-min-add-shp4-g0] PASSED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-div-shp0-g1] PASSED    [ 26%]
tests/compute/test_sparse.py::test_spmm[int32-min-add-shp4-g1] PASSED    [ 26%]
tests/compute/test_sparse.py::test_spmm[int32-min-add-shp5-g0] PASSED    [ 26%]
tests/compute/test_sparse.py::test_spmm[int32-min-add-shp5-g1] PASSED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-div-shp1-g0] PASSED    [ 26%]
tests/compute/test_sparse.py::test_spmm[int32-min-sub-shp0-g0] PASSED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-div-shp1-g1] PASSED    [ 26%]
tests/compute/test_sparse.py::test_spmm[int32-min-sub-shp0-g1] PASSED    [ 26%]
tests/compute/test_sparse.py::test_spmm[int32-min-sub-shp1-g0] PASSED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-div-shp2-g0] PASSED    [ 26%]
tests/compute/test_sparse.py::test_spmm[int32-min-sub-shp1-g1] PASSED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-div-shp2-g1] PASSED    [ 26%]
tests/compute/test_sparse.py::test_spmm[int32-min-sub-shp2-g0] PASSED    [ 26%]
tests/compute/test_sparse.py::test_spmm[int32-min-sub-shp2-g1] PASSED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-div-shp3-g0] PASSED    [ 26%]
tests/compute/test_sparse.py::test_spmm[int32-min-sub-shp3-g0] PASSED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-div-shp3-g1] PASSED    [ 26%]
tests/compute/test_sparse.py::test_spmm[int32-min-sub-shp3-g1] PASSED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-div-shp4-g0] PASSED    [ 26%]
tests/compute/test_sparse.py::test_spmm[int32-min-sub-shp4-g0] PASSED              [ 29%]
tests/go/test_pipeline.py::test_nodepred_model[sgc] PASSED    [ 26%]
tests/compute/test_sparse.py::test_spmm[int32-min-sub-shp4-g1] PASSED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-div-shp4-g1] PASSED    [ 26%]
tests/compute/test_sparse.py::test_spmm[int32-min-sub-shp5-g0] PASSED    [ 26%]
tests/compute/test_sparse.py::test_spmm[int32-min-sub-shp5-g1] PASSED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-div-shp5-g0] PASSED    [ 26%]
tests/compute/test_sparse.py::test_spmm[int32-min-mul-shp0-g0] PASSED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-div-shp5-g1] PASSED    [ 26%]
tests/compute/test_sparse.py::test_spmm[int32-min-mul-shp0-g1] PASSED  [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_lhs-shp0-g0] PASSED    [ 26%]
tests/compute/test_sparse.py::test_spmm[int32-min-mul-shp1-g0] Converting to homogeneous graph takes 0.008s, peak mem: 1.841 GB
Convert a graph into a bidirected graph: 0.017 seconds, peak memory: 1.841 GB
Construct multi-constraint weights: 0.001 seconds, peak memory: 1.841 GB
[05:37:06] /root/jenkins/workspace/dgl_PR-4648/src/graph/transform/metis_partition_hetero.cc:87: Partition a graph with 30030 nodes and 601200 edges into 3 parts and get 152327 edge cuts
Metis partitioning: 0.178 seconds, peak memory: 1.841 GB
Assigning nodes to METIS partitions takes 0.197s, peak mem: 1.841 GB
Reshuffle nodes and edges: 0.019 seconds
Split the graph: 0.054 seconds
Construct subgraphs: 0.048 seconds
Splitting the graph into partitions takes 0.121s, peak mem: 1.841 GB
part 0 has 9943 nodes of type n1 and 3134 are inside the partition
part 0 has 9950 nodes of type n2 and 3140 are inside the partition
part 0 has 9967 nodes of type n3 and 3142 are inside the partition
part 0 has 47569 edges of type r1 and 31233 are inside the partition
part 0 has 47527 edges of type r2 and 31191 are inside the partition
part 0 has 47507 edges of type r3 and 31390 are inside the partition
part 1 has 9977 nodes of type n1 and 3433 are inside the partition
part 1 has 9976 nodes of type n2 and 3434 are inside the partition
part 1 has 9993 nodes of type n3 and 3438 are inside the partition
part 1 has 51495 edges of type r1 and 34423 are inside the partition
part 1 has 51685 edges of type r2 and 34458 are inside the partition
part 1 has 51892 edges of type r3 and 34467 are inside the partition
part 2 has 9967 nodes of type n1 and 3433 are inside the partition
part 2 has 9984 nodes of type n2 and 3436 are inside the partition
part 2 has 9990 nodes of type n3 and 3440 are inside the partition
part 2 has 51628 edges of type r1 and 34444 are inside the partition
part 2 has 51884 edges of type r2 and 34551 are inside the partition
part 2 has 51740 edges of type r3 and 34443 are inside the partition
Save partitions: 0.041 seconds, peak memory: 1.841 GB
There are 300600 edges in the graph and 0 edge cuts for 3 partitions.
server: #clients=5
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_sampling
start graph service on server 0 for part 0
[05:37:07] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:37:07] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:25054]...
PASSED    [ 26%]
tests/compute/test_sparse.py::test_spmm[int32-min-mul-shp1-g1] server: #clients=5
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_sampling
start graph service on server 1 for part 1
[05:37:08] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:37:08] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:25057]...
PASSED [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_lhs-shp0-g1] PASSED    [ 26%]
tests/compute/test_sparse.py::test_spmm[int32-min-mul-shp2-g0] PASSED    [ 26%]
tests/compute/test_sparse.py::test_spmm[int32-min-mul-shp2-g1] PASSED [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_lhs-shp1-g0] server: #clients=5
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_sampling
start graph service on server 2 for part 2
[05:37:09] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:37:09] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:25060]...
PASSED    [ 26%]
tests/compute/test_sparse.py::test_spmm[int32-min-mul-shp3-g0] PASSED [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_lhs-shp1-g1] PASSED    [ 26%]
tests/compute/test_sparse.py::test_spmm[int32-min-mul-shp3-g1] [05:37:10] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:37:10] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
PASSED [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_lhs-shp2-g0] PASSED    [ 26%]
tests/compute/test_sparse.py::test_spmm[int32-min-mul-shp4-g0] PASSED    [ 27%]
tests/compute/test_sparse.py::test_spmm[int32-min-mul-shp4-g1] PASSED [ 27%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_lhs-shp2-g1] PASSED    [ 27%]
tests/compute/test_sparse.py::test_spmm[int32-min-mul-shp5-g0] [05:37:11] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:37:11] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
[05:37:11] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:37:11] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
[05:37:11] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[[05:37:11] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc05:37:11:] 159/root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc: :Receiver with NetType~140socket:  is created.Sender with NetType~
socket is created.
[05:37:11] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
PASSED    [ 27%]
tests/compute/test_sparse.py::test_spmm[int32-min-mul-shp5-g1] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_lhs-shp3-g0] PASSED    [ 27%]
tests/compute/test_sparse.py::test_spmm[int32-min-div-shp0-g0] PASSED    [ 27%]
tests/compute/test_sparse.py::test_spmm[int32-min-div-shp0-g1] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_lhs-shp3-g1] PASSED               [ 30%]
tests/go/test_pipeline.py::test_nodepred_model[gin] PASSED    [ 27%]
tests/compute/test_sparse.py::test_spmm[int32-min-div-shp1-g0] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_lhs-shp4-g0] PASSED    [ 27%]
tests/compute/test_sparse.py::test_spmm[int32-min-div-shp1-g1] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_lhs-shp4-g1] PASSED    [ 27%]
tests/compute/test_sparse.py::test_spmm[int32-min-div-shp2-g0] PASSED    [ 27%]
tests/compute/test_sparse.py::test_spmm[int32-min-div-shp2-g1] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_lhs-shp5-g0] PASSED    [ 27%]
tests/compute/test_sparse.py::test_spmm[int32-min-div-shp3-g0] Client [2751] waits on 172.17.0.3:59413
Client [2754] waits on 172.17.0.3:59931
Client [2747] waits on 172.17.0.3:43589
Client [2753] waits on 172.17.0.3:40831Client [2752] waits on 172.17.0.3:51987

Machine (0) group (0) client (0) connect to server successfuly!
Machine (0) group (0) client (1) connect to server successfuly!
Machine (0) group (0) client (2) connect to server successfuly!
Machine (0) group (0) client (3) connect to server successfuly!
Machine (0) group (0) client (4) connect to server successfuly!
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
PASSED    [ 27%]
tests/compute/test_sparse.py::test_spmm[int32-min-div-shp3-g1] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_lhs-shp5-g1] PASSED    [ 27%]
tests/compute/test_sparse.py::test_spmm[int32-min-div-shp4-g0] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_rhs-shp0-g0] PASSED    [ 27%]
tests/compute/test_sparse.py::test_spmm[int32-min-div-shp4-g1] Client[1] in group[0] is exiting...
Client[3] in group[0] is exiting...
PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_rhs-shp0-g1] PASSED    [ 27%]
tests/compute/test_sparse.py::test_spmm[int32-min-div-shp5-g0] Client[2] in group[0] is exiting...
PASSED    [ 27%]
tests/compute/test_sparse.py::test_spmm[int32-min-div-shp5-g1] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_rhs-shp1-g0] Client[0] in group[0] is exiting...
PASSED    [ 27%]
tests/compute/test_sparse.py::test_spmm[int32-min-copy_lhs-shp0-g0] PASSED [ 27%]
tests/compute/test_sparse.py::test_spmm[int32-min-copy_lhs-shp0-g1] Client[4] in group[0] is exiting...
PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_rhs-shp1-g1] PASSED [ 27%]
tests/compute/test_sparse.py::test_spmm[int32-min-copy_lhs-shp1-g0] Server (0) shutdown.
Server is exiting...
PASSED [ 27%]
tests/compute/test_sparse.py::test_spmm[int32-min-copy_lhs-shp1-g1] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_rhs-shp2-g0] PASSED [ 27%]
tests/compute/test_sparse.py::test_spmm[int32-min-copy_lhs-shp2-g0] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_rhs-shp2-g1] PASSED [ 27%]
tests/compute/test_sparse.py::test_spmm[int32-min-copy_lhs-shp2-g1] Server (1) shutdown.Server (2) shutdown.

Server is exiting...Server is exiting...

PASSED               [ 32%]
tests/go/test_pipeline.py::test_nodepred_ns_data[cora] PASSED [ 27%]
tests/compute/test_sparse.py::test_spmm[int32-min-copy_lhs-shp3-g0] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_rhs-shp3-g0] PASSED [ 27%]
tests/compute/test_sparse.py::test_spmm[int32-min-copy_lhs-shp3-g1] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_rhs-shp3-g1] PASSED [ 27%]
tests/compute/test_sparse.py::test_spmm[int32-min-copy_lhs-shp4-g0] PASSED [ 54%]
tests/distributed/test_mp_dataloader.py::test_neg_dataloader[0-3]   NumNodes: 2708
  NumEdges: 10556
  NumFeats: 1433
  NumClasses: 7
  NumTrainingSamples: 140
  NumValidationSamples: 500
  NumTestSamples: 1000
Done loading data from cached files.
Converting to homogeneous graph takes 0.001s, peak mem: 1.841 GB
Convert a graph into a bidirected graph: 0.001 seconds, peak memory: 1.841 GB
Construct multi-constraint weights: 0.000 seconds, peak memory: 1.841 GB
[05:37:20] /root/jenkins/workspace/dgl_PR-4648/src/graph/transform/metis_partition_hetero.cc:87: Partition a graph with 2708 nodes and 10556 edges into 3 parts and get 325 edge cuts
Metis partitioning: 0.002 seconds, peak memory: 1.841 GB
Assigning nodes to METIS partitions takes 0.003s, peak mem: 1.841 GB
Reshuffle nodes and edges: 0.001 seconds
Split the graph: 0.002 seconds
Construct subgraphs: 0.003 seconds
Splitting the graph into partitions takes 0.006s, peak mem: 1.841 GB
part 0 has 1046 nodes and 903 are inside the partition
part 0 has 3490 edges and 3317 are inside the partition
part 1 has 1064 nodes and 902 are inside the partition
part 1 has 4006 edges and 3739 are inside the partition
part 2 has 1056 nodes and 903 are inside the partition
part 2 has 3710 edges and 3500 are inside the partition
Save partitions: 0.017 seconds, peak memory: 1.841 GB
There are 10556 edges in the graph and 0 edge cuts for 3 partitions.
PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[int32-min-copy_lhs-shp4-g1] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_rhs-shp4-g0] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[int32-min-copy_lhs-shp5-g0] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[int32-min-copy_lhs-shp5-g1] server: #clients=1
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_sampling
start graph service on server 0 for part 0
[05:37:21] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:37:21] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:27571]...
PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_rhs-shp4-g1] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[int32-min-copy_rhs-shp0-g0] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[int32-min-copy_rhs-shp0-g1] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_rhs-shp5-g0] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[int32-min-copy_rhs-shp1-g0] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[int32-min-copy_rhs-shp1-g1] server: #clients=1
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_sampling
start graph service on server 1 for part 1
[05:37:22] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:37:22] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:27574]...
PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-min-copy_rhs-shp5-g1] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[int32-min-copy_rhs-shp2-g0] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[int32-min-copy_rhs-shp2-g1] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-add-shp0-g0] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[int32-min-copy_rhs-shp3-g0] server: #clients=1
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_sampling
start graph service on server 2 for part 2
[05:37:23] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:37:23] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:27577]...
PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[int32-min-copy_rhs-shp3-g1] PASSED  [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-add-shp0-g1] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[int32-min-copy_rhs-shp4-g0] [05:37:24] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:37:24] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[int32-min-copy_rhs-shp4-g1] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[int32-min-copy_rhs-shp5-g0] PASSED  [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-add-shp1-g0] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[int32-min-copy_rhs-shp5-g1] PASSED  [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-add-shp1-g1] PASSED            [ 33%]
tests/go/test_pipeline.py::test_nodepred_ns_data[citeseer] PASSED [ 28%]
tests/compute/test_sparse.py::test_spmm[int32-max-add-shp0-g0] PASSED    [ 28%]
tests/compute/test_sparse.py::test_spmm[int32-max-add-shp0-g1] PASSED  [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-add-shp2-g0] PASSED    [ 28%]
tests/compute/test_sparse.py::test_spmm[int32-max-add-shp1-g0] PASSED    [ 28%]
tests/compute/test_sparse.py::test_spmm[int32-max-add-shp1-g1] PASSED  [ 28%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-add-shp2-g1] PASSED    [ 28%]
tests/compute/test_sparse.py::test_spmm[int32-max-add-shp2-g0] PASSED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-add-shp3-g0] PASSED    [ 28%]
tests/compute/test_sparse.py::test_spmm[int32-max-add-shp2-g1] Client [2853] waits on 172.17.0.3:50351
Machine (0) group (0) client (0) connect to server successfuly!
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
PASSED    [ 28%]
tests/compute/test_sparse.py::test_spmm[int32-max-add-shp3-g0] Client[0] in group[0] is exiting...
Server (0) shutdown.
Server is exiting...
Server (1) shutdown.
Server (2) shutdown.Server is exiting...

Server is exiting...
PASSED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-add-shp3-g1] PASSED    [ 28%]
tests/compute/test_sparse.py::test_spmm[int32-max-add-shp3-g1] PASSED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-add-shp4-g0] PASSED    [ 28%]
tests/compute/test_sparse.py::test_spmm[int32-max-add-shp4-g0] PASSED    [ 28%]
tests/compute/test_sparse.py::test_spmm[int32-max-add-shp4-g1] PASSED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-add-shp4-g1] PASSED    [ 29%]
tests/compute/test_sparse.py::test_spmm[int32-max-add-shp5-g0] PASSED    [ 29%]
tests/compute/test_sparse.py::test_spmm[int32-max-add-shp5-g1] PASSED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-add-shp5-g0] PASSED    [ 29%]
tests/compute/test_sparse.py::test_spmm[int32-max-sub-shp0-g0] PASSED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-add-shp5-g1] PASSED    [ 29%]
tests/compute/test_sparse.py::test_spmm[int32-max-sub-shp0-g1] PASSED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-sub-shp0-g0] PASSED    [ 29%]
tests/compute/test_sparse.py::test_spmm[int32-max-sub-shp1-g0] PASSED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-sub-shp0-g1] PASSED    [ 29%]
tests/compute/test_sparse.py::test_spmm[int32-max-sub-shp1-g1] PASSED    [ 29%]
tests/compute/test_sparse.py::test_spmm[int32-max-sub-shp2-g0] PASSED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-sub-shp1-g0] PASSED    [ 29%]
tests/compute/test_sparse.py::test_spmm[int32-max-sub-shp2-g1] PASSED        [ 34%]
tests/go/test_pipeline.py::test_nodepred_ns_data[pubmed] PASSED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-sub-shp1-g1] PASSED    [ 29%]
tests/compute/test_sparse.py::test_spmm[int32-max-sub-shp3-g0] PASSED    [ 29%]
tests/compute/test_sparse.py::test_spmm[int32-max-sub-shp3-g1] PASSED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-sub-shp2-g0] PASSED    [ 29%]
tests/compute/test_sparse.py::test_spmm[int32-max-sub-shp4-g0] PASSED    [ 29%]
tests/compute/test_sparse.py::test_spmm[int32-max-sub-shp4-g1] PASSED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-sub-shp2-g1] PASSED    [ 29%]
tests/compute/test_sparse.py::test_spmm[int32-max-sub-shp5-g0] PASSED    [ 29%]
tests/compute/test_sparse.py::test_spmm[int32-max-sub-shp5-g1] PASSED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-sub-shp3-g0] PASSED    [ 29%]
tests/compute/test_sparse.py::test_spmm[int32-max-mul-shp0-g0] PASSED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-sub-shp3-g1] PASSED    [ 29%]
tests/compute/test_sparse.py::test_spmm[int32-max-mul-shp0-g1] PASSED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-sub-shp4-g0] PASSED    [ 29%]
tests/compute/test_sparse.py::test_spmm[int32-max-mul-shp1-g0] PASSED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-sub-shp4-g1] PASSED    [ 29%]
tests/compute/test_sparse.py::test_spmm[int32-max-mul-shp1-g1] PASSED    [ 29%]
tests/compute/test_sparse.py::test_spmm[int32-max-mul-shp2-g0] PASSED    [ 29%]
tests/compute/test_sparse.py::test_spmm[int32-max-mul-shp2-g1] PASSED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-sub-shp5-g0] PASSED    [ 29%]
tests/compute/test_sparse.py::test_spmm[int32-max-mul-shp3-g0] PASSED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-sub-shp5-g1] PASSED    [ 29%]
tests/compute/test_sparse.py::test_spmm[int32-max-mul-shp3-g1] PASSED          [ 35%]
tests/go/test_pipeline.py::test_nodepred_ns_data[csv] PASSED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-mul-shp0-g0] PASSED    [ 29%]
tests/compute/test_sparse.py::test_spmm[int32-max-mul-shp4-g0] PASSED    [ 29%]
tests/compute/test_sparse.py::test_spmm[int32-max-mul-shp4-g1] PASSED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-mul-shp0-g1] PASSED    [ 30%]
tests/compute/test_sparse.py::test_spmm[int32-max-mul-shp5-g0] PASSED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-mul-shp1-g0] PASSED    [ 30%]
tests/compute/test_sparse.py::test_spmm[int32-max-mul-shp5-g1] PASSED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-mul-shp1-g1] PASSED    [ 30%]
tests/compute/test_sparse.py::test_spmm[int32-max-div-shp0-g0] PASSED    [ 30%]
tests/compute/test_sparse.py::test_spmm[int32-max-div-shp0-g1] Converting to homogeneous graph takes 0.010s, peak mem: 1.841 GB
Convert a graph into a bidirected graph: 0.017 seconds, peak memory: 1.841 GB
Construct multi-constraint weights: 0.001 seconds, peak memory: 1.841 GB
PASSED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-mul-shp2-g0] PASSED    [ 30%]
tests/compute/test_sparse.py::test_spmm[int32-max-div-shp1-g0] [05:37:43] /root/jenkins/workspace/dgl_PR-4648/src/graph/transform/metis_partition_hetero.cc:87: Partition a graph with 30030 nodes and 601200 edges into 3 parts and get 150005 edge cuts
Metis partitioning: 0.205 seconds, peak memory: 1.841 GB
Assigning nodes to METIS partitions takes 0.224s, peak mem: 1.841 GB
Reshuffle nodes and edges: 0.018 seconds
Split the graph: 0.053 seconds
Construct subgraphs: 0.047 seconds
Splitting the graph into partitions takes 0.118s, peak mem: 1.841 GB
part 0 has 9954 nodes of type n1 and 3134 are inside the partition
part 0 has 9953 nodes of type n2 and 3138 are inside the partition
part 0 has 9962 nodes of type n3 and 3141 are inside the partition
part 0 has 47034 edges of type r1 and 31192 are inside the partition
part 0 has 47153 edges of type r2 and 31126 are inside the partition
part 0 has 47225 edges of type r3 and 31379 are inside the partition
part 1 has 9978 nodes of type n1 and 3433 are inside the partition
part 1 has 9976 nodes of type n2 and 3436 are inside the partition
part 1 has 9984 nodes of type n3 and 3439 are inside the partition
part 1 has 51442 edges of type r1 and 34436 are inside the partition
part 1 has 51701 edges of type r2 and 34684 are inside the partition
part 1 has 51404 edges of type r3 and 34408 are inside the partition
part 2 has 9964 nodes of type n1 and 3433 are inside the partition
part 2 has 9974 nodes of type n2 and 3436 are inside the partition
part 2 has 9981 nodes of type n3 and 3440 are inside the partition
part 2 has 51573 edges of type r1 and 34472 are inside the partition
part 2 has 51494 edges of type r2 and 34390 are inside the partition
part 2 has 51579 edges of type r3 and 34513 are inside the partition
Save partitions: 0.035 seconds, peak memory: 1.841 GB
There are 300600 edges in the graph and 0 edge cuts for 3 partitions.
PASSED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-mul-shp2-g1] PASSED    [ 30%]
tests/compute/test_sparse.py::test_spmm[int32-max-div-shp1-g1] PASSED    [ 30%]
tests/compute/test_sparse.py::test_spmm[int32-max-div-shp2-g0] PASSED  [ 29%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-mul-shp3-g0] server: #clients=1
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_sampling
start graph service on server 0 for part 0
[05:37:44] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:37:44] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:14090]...
PASSED    [ 30%]
tests/compute/test_sparse.py::test_spmm[int32-max-div-shp2-g1] PASSED    [ 30%]
tests/compute/test_sparse.py::test_spmm[int32-max-div-shp3-g0] PASSED  [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-mul-shp3-g1] PASSED    [ 30%]
tests/compute/test_sparse.py::test_spmm[int32-max-div-shp3-g1] server: #clients=1
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_sampling
start graph service on server 1 for part 1
[05:37:45] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:37:45] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:14093]...
PASSED  [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-mul-shp4-g0] PASSED    [ 30%]
tests/compute/test_sparse.py::test_spmm[int32-max-div-shp4-g0] PASSED    [ 30%]
tests/compute/test_sparse.py::test_spmm[int32-max-div-shp4-g1] server: #clients=1
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_sampling
start graph service on server 2 for part 2
[05:37:46] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:37:46] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:14096]...
PASSED  [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-mul-shp4-g1] PASSED             [ 37%]
tests/go/test_pipeline.py::test_nodepred_ns_data[reddit] PASSED    [ 30%]
tests/compute/test_sparse.py::test_spmm[int32-max-div-shp5-g0] PASSED  [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-mul-shp5-g0] PASSED    [ 30%]
tests/compute/test_sparse.py::test_spmm[int32-max-div-shp5-g1] [05:37:47] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:37:47] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
PASSED  [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-mul-shp5-g1] PASSED    [ 30%]
tests/compute/test_sparse.py::test_spmm[int32-max-copy_lhs-shp0-g0] PASSED [ 30%]
tests/compute/test_sparse.py::test_spmm[int32-max-copy_lhs-shp0-g1] PASSED  [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-div-shp0-g0] PASSED [ 30%]
tests/compute/test_sparse.py::test_spmm[int32-max-copy_lhs-shp1-g0] PASSED  [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-div-shp0-g1] PASSED [ 30%]
tests/compute/test_sparse.py::test_spmm[int32-max-copy_lhs-shp1-g1] PASSED [ 30%]
tests/compute/test_sparse.py::test_spmm[int32-max-copy_lhs-shp2-g0] PASSED [ 30%]
tests/compute/test_sparse.py::test_spmm[int32-max-copy_lhs-shp2-g1] PASSED  [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-div-shp1-g0] PASSED [ 30%]
tests/compute/test_sparse.py::test_spmm[int32-max-copy_lhs-shp3-g0] PASSED [ 30%]PASSED  [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-div-shp1-g1] 
tests/compute/test_sparse.py::test_spmm[int32-max-copy_lhs-shp3-g1] PASSED [ 30%]
tests/compute/test_sparse.py::test_spmm[int32-max-copy_lhs-shp4-g0] Client [2893] waits on 172.17.0.3:47041
Machine (0) group (0) client (0) connect to server successfuly!
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
PASSED  [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-div-shp2-g0] PASSED [ 30%]
tests/compute/test_sparse.py::test_spmm[int32-max-copy_lhs-shp4-g1] PASSED  [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-div-shp2-g1] PASSED [ 30%]
tests/compute/test_sparse.py::test_spmm[int32-max-copy_lhs-shp5-g0] Client[0] in group[0] is exiting...
Server (0) shutdown.
Server is exiting...
Server (2) shutdown.
Server is exiting...
Server (1) shutdown.
Server is exiting...
PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[int32-max-copy_lhs-shp5-g1] PASSED [ 56%]
tests/distributed/test_mp_dataloader.py::test_neg_dataloader[4-3]   NumNodes: 2708
  NumEdges: 10556
  NumFeats: 1433
  NumClasses: 7
  NumTrainingSamples: 140
  NumValidationSamples: 500
  NumTestSamples: 1000
Done loading data from cached files.
Converting to homogeneous graph takes 0.001s, peak mem: 1.841 GB
Convert a graph into a bidirected graph: 0.001 seconds, peak memory: 1.841 GB
Construct multi-constraint weights: 0.000 seconds, peak memory: 1.841 GB
[05:37:52] /root/jenkins/workspace/dgl_PR-4648/src/graph/transform/metis_partition_hetero.cc:87: Partition a graph with 2708 nodes and 10556 edges into 3 parts and get 325 edge cuts
Metis partitioning: 0.003 seconds, peak memory: 1.841 GB
Assigning nodes to METIS partitions takes 0.004s, peak mem: 1.841 GB
Reshuffle nodes and edges: 0.001 seconds
Split the graph: 0.002 seconds
Construct subgraphs: 0.004 seconds
Splitting the graph into partitions takes 0.007s, peak mem: 1.841 GB
part 0 has 1046 nodes and 903 are inside the partition
part 0 has 3490 edges and 3317 are inside the partition
part 1 has 1064 nodes and 902 are inside the partition
part 1 has 4006 edges and 3739 are inside the partition
part 2 has 1056 nodes and 903 are inside the partition
part 2 has 3710 edges and 3500 are inside the partition
Save partitions: 0.020 seconds, peak memory: 1.841 GB
There are 10556 edges in the graph and 0 edge cuts for 3 partitions.
PASSED  [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-div-shp3-g0] PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[int32-max-copy_rhs-shp0-g0] PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[int32-max-copy_rhs-shp0-g1] PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[int32-max-copy_rhs-shp1-g0] PASSED  [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-div-shp3-g1] PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[int32-max-copy_rhs-shp1-g1] PASSED          [ 38%]
tests/go/test_pipeline.py::test_nodepred_ns_data[co-buy-computer] server: #clients=5
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_sampling
start graph service on server 0 for part 0
[05:37:53] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:37:53] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:22403]...
PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[int32-max-copy_rhs-shp2-g0] PASSED  [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-div-shp4-g0] PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[int32-max-copy_rhs-shp2-g1] PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[int32-max-copy_rhs-shp3-g0] PASSED  [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-div-shp4-g1] PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[int32-max-copy_rhs-shp3-g1] server: #clients=5
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_sampling
start graph service on server 1 for part 1
[05:37:54] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:37:54] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:22406]...
PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[int32-max-copy_rhs-shp4-g0] PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[int32-max-copy_rhs-shp4-g1] PASSED  [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-div-shp5-g0] server: #clients=5
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_sampling
start graph service on server 2 for part 2
[05:37:55] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:37:55] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:22409]...
PASSED  [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-div-shp5-g1] PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[int32-max-copy_rhs-shp5-g0] PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[int32-max-copy_rhs-shp5-g1] PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[int64-sum-add-shp0-g0] PASSED    [ 31%]
tests/compute/test_sparse.py::test_spmm[int64-sum-add-shp0-g1] [05:37:56] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:37:56] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
PASSED  [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_lhs-shp0-g0] PASSED    [ 31%]
tests/compute/test_sparse.py::test_spmm[int64-sum-add-shp1-g0] PASSED [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_lhs-shp0-g1] PASSED    [ 31%]
tests/compute/test_sparse.py::test_spmm[int64-sum-add-shp1-g1] [05:37:57] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:37:57] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
[05:37:57] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:37:57] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
[05:37:57] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:37:57] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
[05:37:57] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:37:57] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
PASSED [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_lhs-shp1-g0] PASSED    [ 31%]
tests/compute/test_sparse.py::test_spmm[int64-sum-add-shp2-g0] PASSED    [ 31%]
tests/compute/test_sparse.py::test_spmm[int64-sum-add-shp2-g1] PASSED [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_lhs-shp1-g1] PASSED    [ 31%]
tests/compute/test_sparse.py::test_spmm[int64-sum-add-shp3-g0] PASSED [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_lhs-shp2-g0] PASSED    [ 31%]
tests/compute/test_sparse.py::test_spmm[int64-sum-add-shp3-g1] PASSED [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_lhs-shp2-g1] PASSED    [ 31%]
tests/compute/test_sparse.py::test_spmm[int64-sum-add-shp4-g0] PASSED    [ 31%]
tests/compute/test_sparse.py::test_spmm[int64-sum-add-shp4-g1] PASSED [ 30%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_lhs-shp3-g0] PASSED    [ 31%]
tests/compute/test_sparse.py::test_spmm[int64-sum-add-shp5-g0] PASSED    [ 32%]
tests/compute/test_sparse.py::test_spmm[int64-sum-add-shp5-g1] Client [2928] waits on 172.17.0.3:38919Client [2930] waits on 172.17.0.3:45477Client [2923] waits on 172.17.0.3:56727


Client [2929] waits on 172.17.0.3:54889
Client [2927] waits on 172.17.0.3:41805
Machine (0) group (0) client (3) connect to server successfuly!
Machine (0) group (0) client (0) connect to server successfuly!
Machine (0) group (0) client (2) connect to server successfuly!
Machine (0) group (0) client (1) connect to server successfuly!Machine (0) group (0) client (4) connect to server successfuly!

/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_lhs-shp3-g1] Client[4] in group[0] is exiting...
Client[1] in group[0] is exiting...
PASSED    [ 32%]
tests/compute/test_sparse.py::test_spmm[int64-sum-sub-shp0-g0] PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_lhs-shp4-g0] PASSED    [ 32%]
tests/compute/test_sparse.py::test_spmm[int64-sum-sub-shp0-g1] Client[0] in group[0] is exiting...
PASSED [ 39%]
tests/go/test_pipeline.py::test_nodepred_ns_data[ogbn-arxiv] PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_lhs-shp4-g1] PASSED    [ 32%]
tests/compute/test_sparse.py::test_spmm[int64-sum-sub-shp1-g0] Client[3] in group[0] is exiting...
PASSED    [ 32%]
tests/compute/test_sparse.py::test_spmm[int64-sum-sub-shp1-g1] Client[2] in group[0] is exiting...
PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_lhs-shp5-g0] PASSED    [ 32%]
tests/compute/test_sparse.py::test_spmm[int64-sum-sub-shp2-g0] Server (2) shutdown.Server (0) shutdown.

Server is exiting...
Server is exiting...
Server (1) shutdown.
Server is exiting...
PASSED    [ 32%]
tests/compute/test_sparse.py::test_spmm[int64-sum-sub-shp2-g1] PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_lhs-shp5-g1] PASSED    [ 32%]
tests/compute/test_sparse.py::test_spmm[int64-sum-sub-shp3-g0] PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_rhs-shp0-g0] PASSED    [ 32%]
tests/compute/test_sparse.py::test_spmm[int64-sum-sub-shp3-g1] PASSED    [ 32%]
tests/compute/test_sparse.py::test_spmm[int64-sum-sub-shp4-g0] PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_rhs-shp0-g1] PASSED    [ 32%]
tests/compute/test_sparse.py::test_spmm[int64-sum-sub-shp4-g1] PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_rhs-shp1-g0] PASSED    [ 32%]
tests/compute/test_sparse.py::test_spmm[int64-sum-sub-shp5-g0] PASSED    [ 32%]
tests/compute/test_sparse.py::test_spmm[int64-sum-sub-shp5-g1] PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_rhs-shp1-g1] PASSED    [ 32%]
tests/compute/test_sparse.py::test_spmm[int64-sum-mul-shp0-g0] PASSED    [ 32%]
tests/compute/test_sparse.py::test_spmm[int64-sum-mul-shp0-g1] PASSED      [ 40%]
tests/go/test_pipeline.py::test_nodepred_ns_data[ogbn-products] PASSED    [ 32%]
tests/compute/test_sparse.py::test_spmm[int64-sum-mul-shp1-g0] PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_rhs-shp2-g0] PASSED    [ 32%]
tests/compute/test_sparse.py::test_spmm[int64-sum-mul-shp1-g1] PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_rhs-shp2-g1] PASSED    [ 32%]
tests/compute/test_sparse.py::test_spmm[int64-sum-mul-shp2-g0] PASSED    [ 32%]
tests/compute/test_sparse.py::test_spmm[int64-sum-mul-shp2-g1] PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_rhs-shp3-g0] PASSED    [ 32%]
tests/compute/test_sparse.py::test_spmm[int64-sum-mul-shp3-g0] PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_rhs-shp3-g1] PASSED    [ 32%]
tests/compute/test_sparse.py::test_spmm[int64-sum-mul-shp3-g1] PASSED    [ 32%]
tests/compute/test_sparse.py::test_spmm[int64-sum-mul-shp4-g0] PASSED    [ 32%]
tests/compute/test_sparse.py::test_spmm[int64-sum-mul-shp4-g1] PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_rhs-shp4-g0] PASSED    [ 32%]
tests/compute/test_sparse.py::test_spmm[int64-sum-mul-shp5-g0] PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_rhs-shp4-g1] PASSED    [ 32%]
tests/compute/test_sparse.py::test_spmm[int64-sum-mul-shp5-g1] PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_rhs-shp5-g0] PASSED    [ 33%]
tests/compute/test_sparse.py::test_spmm[int64-sum-div-shp0-g0] PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_rhs-shp5-g1] PASSED    [ 33%]
tests/compute/test_sparse.py::test_spmm[int64-sum-div-shp0-g1] PASSED    [ 33%]
tests/compute/test_sparse.py::test_spmm[int64-sum-div-shp1-g0] PASSED [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-add-shp0-g0] PASSED    [ 33%]
tests/compute/test_sparse.py::test_spmm[int64-sum-div-shp1-g1] PASSED   [ 41%]
tests/go/test_pipeline.py::test_nodepred_ns_model[gcn] PASSED  [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-add-shp0-g1] PASSED    [ 33%]
tests/compute/test_sparse.py::test_spmm[int64-sum-div-shp2-g0] PASSED    [ 33%]
tests/compute/test_sparse.py::test_spmm[int64-sum-div-shp2-g1] PASSED  [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-add-shp1-g0] PASSED    [ 33%]
tests/compute/test_sparse.py::test_spmm[int64-sum-div-shp3-g0] PASSED    [ 33%]
tests/compute/test_sparse.py::test_spmm[int64-sum-div-shp3-g1] PASSED  [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-add-shp1-g1] PASSED    [ 33%]
tests/compute/test_sparse.py::test_spmm[int64-sum-div-shp4-g0] PASSED  [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-add-shp2-g0] PASSED    [ 33%]
tests/compute/test_sparse.py::test_spmm[int64-sum-div-shp4-g1] PASSED  [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-add-shp2-g1] PASSED    [ 33%]
tests/compute/test_sparse.py::test_spmm[int64-sum-div-shp5-g0] PASSED    [ 33%]
tests/compute/test_sparse.py::test_spmm[int64-sum-div-shp5-g1] PASSED  [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-add-shp3-g0] PASSED    [ 33%]
tests/compute/test_sparse.py::test_spmm[int64-sum-copy_lhs-shp0-g0] PASSED  [ 31%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-add-shp3-g1] PASSED [ 33%]
tests/compute/test_sparse.py::test_spmm[int64-sum-copy_lhs-shp0-g1] Converting to homogeneous graph takes 0.008s, peak mem: 1.841 GB
Convert a graph into a bidirected graph: 0.017 seconds, peak memory: 1.841 GB
Construct multi-constraint weights: 0.001 seconds, peak memory: 1.841 GB
[05:38:16] /root/jenkins/workspace/dgl_PR-4648/src/graph/transform/metis_partition_hetero.cc:87: Partition a graph with 30030 nodes and 601200 edges into 3 parts and get 150005 edge cuts
Metis partitioning: 0.198 seconds, peak memory: 1.841 GB
Assigning nodes to METIS partitions takes 0.216s, peak mem: 1.841 GB
Reshuffle nodes and edges: 0.016 seconds
Split the graph: 0.045 seconds
Construct subgraphs: 0.048 seconds
Splitting the graph into partitions takes 0.109s, peak mem: 1.841 GB
part 0 has 9954 nodes of type n1 and 3134 are inside the partition
part 0 has 9953 nodes of type n2 and 3138 are inside the partition
part 0 has 9962 nodes of type n3 and 3141 are inside the partition
part 0 has 47034 edges of type r1 and 31192 are inside the partition
part 0 has 47153 edges of type r2 and 31126 are inside the partition
part 0 has 47225 edges of type r3 and 31379 are inside the partition
part 1 has 9978 nodes of type n1 and 3433 are inside the partition
part 1 has 9976 nodes of type n2 and 3436 are inside the partition
part 1 has 9984 nodes of type n3 and 3439 are inside the partition
part 1 has 51442 edges of type r1 and 34436 are inside the partition
part 1 has 51701 edges of type r2 and 34684 are inside the partition
part 1 has 51404 edges of type r3 and 34408 are inside the partition
part 2 has 9964 nodes of type n1 and 3433 are inside the partition
part 2 has 9974 nodes of type n2 and 3436 are inside the partition
part 2 has 9981 nodes of type n3 and 3440 are inside the partition
part 2 has 51573 edges of type r1 and 34472 are inside the partition
part 2 has 51494 edges of type r2 and 34390 are inside the partition
part 2 has 51579 edges of type r3 and 34513 are inside the partition
Save partitions: 0.037 seconds, peak memory: 1.841 GB
There are 300600 edges in the graph and 0 edge cuts for 3 partitions.
server: #clients=5
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_sampling
start graph service on server 0 for part 0
[05:38:18] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:38:18] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:12582]...
PASSED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-add-shp4-g0] PASSED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-add-shp4-g1] server: #clients=5
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_sampling
start graph service on server 1 for part 1
[05:38:19] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:38:19] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:12585]...
PASSED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-add-shp5-g0] PASSED [ 33%]
tests/compute/test_sparse.py::test_spmm[int64-sum-copy_lhs-shp1-g0] PASSED [ 33%]
tests/compute/test_sparse.py::test_spmm[int64-sum-copy_lhs-shp1-g1] PASSED            [ 43%]
tests/go/test_pipeline.py::test_nodepred_ns_model[gat] PASSED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-add-shp5-g1] PASSED [ 33%]
tests/compute/test_sparse.py::test_spmm[int64-sum-copy_lhs-shp2-g0] server: #clients=5
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
load test_sampling
start graph service on server 2 for part 2
[05:38:20] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:38:20] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:12588]...
PASSED [ 33%]
tests/compute/test_sparse.py::test_spmm[int64-sum-copy_lhs-shp2-g1] PASSED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-sub-shp0-g0] [05:38:21] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:38:21] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
PASSED [ 33%]
tests/compute/test_sparse.py::test_spmm[int64-sum-copy_lhs-shp3-g0] PASSED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-sub-shp0-g1] PASSED [ 33%]
tests/compute/test_sparse.py::test_spmm[int64-sum-copy_lhs-shp3-g1] [05:38:21] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:38:21] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
[05:38:21] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:38:21] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
PASSED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-sub-shp1-g0] PASSED [ 33%]
tests/compute/test_sparse.py::test_spmm[int64-sum-copy_lhs-shp4-g0] [05:38:22] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:38:22] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
[05:38:22] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:38:22] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
PASSED [ 33%]
tests/compute/test_sparse.py::test_spmm[int64-sum-copy_lhs-shp4-g1] PASSED [ 33%]
tests/compute/test_sparse.py::test_spmm[int64-sum-copy_lhs-shp5-g0] PASSED [ 33%]
tests/compute/test_sparse.py::test_spmm[int64-sum-copy_lhs-shp5-g1] PASSED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-sub-shp1-g1] PASSED [ 33%]
tests/compute/test_sparse.py::test_spmm[int64-sum-copy_rhs-shp0-g0] PASSED [ 34%]
tests/compute/test_sparse.py::test_spmm[int64-sum-copy_rhs-shp0-g1] PASSED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-sub-shp2-g0] PASSED [ 34%]
tests/compute/test_sparse.py::test_spmm[int64-sum-copy_rhs-shp1-g0] PASSED [ 34%]
tests/compute/test_sparse.py::test_spmm[int64-sum-copy_rhs-shp1-g1] PASSED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-sub-shp2-g1] PASSED [ 34%]
tests/compute/test_sparse.py::test_spmm[int64-sum-copy_rhs-shp2-g0] PASSED [ 34%]
tests/compute/test_sparse.py::test_spmm[int64-sum-copy_rhs-shp2-g1] PASSED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-sub-shp3-g0] PASSED [ 34%]
tests/compute/test_sparse.py::test_spmm[int64-sum-copy_rhs-shp3-g0] PASSED [ 34%]
tests/compute/test_sparse.py::test_spmm[int64-sum-copy_rhs-shp3-g1] PASSED [ 34%]
tests/compute/test_sparse.py::test_spmm[int64-sum-copy_rhs-shp4-g0] PASSED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-sub-shp3-g1] Client [3043] waits on 172.17.0.3:38427Client [3044] waits on 172.17.0.3:59635

Client [3045] waits on 172.17.0.3:44859Client [3046] waits on 172.17.0.3:46803

Client [3037] waits on 172.17.0.3:39485
Machine (0) group (0) client (0) connect to server successfuly!
Machine (0) group (0) client (4) connect to server successfuly!Machine (0) group (0) client (3) connect to server successfuly!

Machine (0) group (0) client (2) connect to server successfuly!
Machine (0) group (0) client (1) connect to server successfuly!
/root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
  "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")
PASSED [ 34%]
tests/compute/test_sparse.py::test_spmm[int64-sum-copy_rhs-shp4-g1] PASSED [ 34%]
tests/compute/test_sparse.py::test_spmm[int64-sum-copy_rhs-shp5-g0] PASSED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-sub-shp4-g0] PASSED [ 34%]
tests/compute/test_sparse.py::test_spmm[int64-sum-copy_rhs-shp5-g1] Client[1] in group[0] is exiting...
Client[0] in group[0] is exiting...
PASSED [ 34%]
tests/compute/test_sparse.py::test_spmm[int64-min-add-shp0-g0] Client[4] in group[0] is exiting...
PASSED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-sub-shp4-g1] PASSED    [ 34%]
tests/compute/test_sparse.py::test_spmm[int64-min-add-shp0-g1] Client[2] in group[0] is exiting...
PASSED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-sub-shp5-g0] PASSED    [ 34%]
tests/compute/test_sparse.py::test_spmm[int64-min-add-shp1-g0] PASSED    [ 34%]
tests/compute/test_sparse.py::test_spmm[int64-min-add-shp1-g1] Client[3] in group[0] is exiting...
PASSED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-sub-shp5-g1] PASSED    [ 34%]
tests/compute/test_sparse.py::test_spmm[int64-min-add-shp2-g0] Server (0) shutdown.
Server is exiting...
Server (1) shutdown.
Server is exiting...
PASSED            [ 44%]
tests/go/test_pipeline.py::test_nodepred_ns_model[sage] PASSED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-mul-shp0-g0] PASSED    [ 34%]
tests/compute/test_sparse.py::test_spmm[int64-min-add-shp2-g1] PASSED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-mul-shp0-g1] PASSED    [ 34%]
tests/compute/test_sparse.py::test_spmm[int64-min-add-shp3-g0] Server (2) shutdown.
Server is exiting...
PASSED    [ 34%]
tests/compute/test_sparse.py::test_spmm[int64-min-add-shp3-g1] PASSED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-mul-shp1-g0] PASSED    [ 34%]
tests/compute/test_sparse.py::test_spmm[int64-min-add-shp4-g0] PASSED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-mul-shp1-g1] PASSED    [ 34%]
tests/compute/test_sparse.py::test_spmm[int64-min-add-shp4-g1] PASSED [ 58%]
tests/distributed/test_new_kvstore.py::test_partition_policy PASSED      [ 60%]
tests/distributed/test_new_kvstore.py::test_kv_store PASSED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-mul-shp2-g0] PASSED    [ 34%]
tests/compute/test_sparse.py::test_spmm[int64-min-add-shp5-g0] PASSED    [ 34%]
tests/compute/test_sparse.py::test_spmm[int64-min-add-shp5-g1] PASSED    [ 34%]
tests/compute/test_sparse.py::test_spmm[int64-min-sub-shp0-g0] PASSED    [ 35%]
tests/compute/test_sparse.py::test_spmm[int64-min-sub-shp0-g1] /root/jenkins/workspace/dgl_PR-4648/python/dgl/heterograph.py:72: DGLWarning: Recommend creating graphs by `dgl.graph(data)` instead of `dgl.DGLGraph(data)`.
  dgl_warning('Recommend creating graphs by `dgl.graph(data)`'
/root/jenkins/workspace/dgl_PR-4648/python/dgl/heterograph.py:72: DGLWarning: Recommend creating graphs by `dgl.graph(data)` instead of `dgl.DGLGraph(data)`.
  dgl_warning('Recommend creating graphs by `dgl.graph(data)`'
Sleep 5 seconds to test client re-connect.
Sleep 5 seconds to test client re-connect.
/root/jenkins/workspace/dgl_PR-4648/python/dgl/heterograph.py:72: DGLWarning: Recommend creating graphs by `dgl.graph(data)` instead of `dgl.DGLGraph(data)`.
  dgl_warning('Recommend creating graphs by `dgl.graph(data)`'
[05:38:31] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:38:31] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
/root/jenkins/workspace/dgl_PR-4648/python/dgl/heterograph.py:72: DGLWarning: Recommend creating graphs by `dgl.graph(data)` instead of `dgl.DGLGraph(data)`.
  dgl_warning('Recommend creating graphs by `dgl.graph(data)`'
[05:38:31] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:38:31] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
PASSED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-mul-shp2-g1] PASSED    [ 35%]
tests/compute/test_sparse.py::test_spmm[int64-min-sub-shp1-g0] PASSED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-mul-shp3-g0] PASSED    [ 35%]
tests/compute/test_sparse.py::test_spmm[int64-min-sub-shp1-g1] PASSED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-mul-shp3-g1] PASSED    [ 35%]
tests/compute/test_sparse.py::test_spmm[int64-min-sub-shp2-g0] PASSED    [ 35%]
tests/compute/test_sparse.py::test_spmm[int64-min-sub-shp2-g1] PASSED  [ 32%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-mul-shp4-g0] PASSED           [ 45%]
tests/go/test_pipeline.py::test_linkpred_data[cora] PASSED    [ 35%]
tests/compute/test_sparse.py::test_spmm[int64-min-sub-shp3-g0] PASSED  [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-mul-shp4-g1] PASSED    [ 35%]
tests/compute/test_sparse.py::test_spmm[int64-min-sub-shp3-g1] PASSED  [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-mul-shp5-g0] PASSED    [ 35%]
tests/compute/test_sparse.py::test_spmm[int64-min-sub-shp4-g0] PASSED    [ 35%]
tests/compute/test_sparse.py::test_spmm[int64-min-sub-shp4-g1] PASSED  [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-mul-shp5-g1] PASSED    [ 35%]
tests/compute/test_sparse.py::test_spmm[int64-min-sub-shp5-g0] PASSED  [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-div-shp0-g0] PASSED    [ 35%]
tests/compute/test_sparse.py::test_spmm[int64-min-sub-shp5-g1] PASSED  [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-div-shp0-g1] PASSED    [ 35%]
tests/compute/test_sparse.py::test_spmm[int64-min-mul-shp0-g0] PASSED    [ 35%]
tests/compute/test_sparse.py::test_spmm[int64-min-mul-shp0-g1] PASSED    [ 35%]
tests/compute/test_sparse.py::test_spmm[int64-min-mul-shp1-g0] PASSED  [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-div-shp1-g0] [05:38:36] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:38:36] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:28089]...
[05:38:36] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:38:36] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:28090]...
PASSED    [ 35%]
tests/compute/test_sparse.py::test_spmm[int64-min-mul-shp1-g1] PASSED  [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-div-shp1-g1] PASSED    [ 35%]
tests/compute/test_sparse.py::test_spmm[int64-min-mul-shp2-g0] PASSED    [ 35%]
tests/compute/test_sparse.py::test_spmm[int64-min-mul-shp2-g1] PASSED  [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-div-shp2-g0] PASSED    [ 35%]
tests/compute/test_sparse.py::test_spmm[int64-min-mul-shp3-g0] PASSED  [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-div-shp2-g1] PASSED    [ 35%]
tests/compute/test_sparse.py::test_spmm[int64-min-mul-shp3-g1] PASSED               [ 46%]
tests/go/test_pipeline.py::test_linkpred_data[citeseer] PASSED  [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-div-shp3-g0] PASSED    [ 35%]
tests/compute/test_sparse.py::test_spmm[int64-min-mul-shp4-g0] PASSED    [ 35%]
tests/compute/test_sparse.py::test_spmm[int64-min-mul-shp4-g1] PASSED  [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-div-shp3-g1] PASSED    [ 35%]
tests/compute/test_sparse.py::test_spmm[int64-min-mul-shp5-g0] PASSED  [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-div-shp4-g0] PASSED    [ 35%]
tests/compute/test_sparse.py::test_spmm[int64-min-mul-shp5-g1] PASSED  [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-div-shp4-g1] PASSED    [ 35%]
tests/compute/test_sparse.py::test_spmm[int64-min-div-shp0-g0] Client [3137] waits on 172.17.0.3:54393
Client [3136] waits on 172.17.0.3:41255
Machine (0) group (0) client (1) connect to server successfuly!
Machine (0) group (0) client (0) connect to server successfuly!
['data_2', 'data_0_1', 'data_0_3', 'data_1', 'data_0', 'data_0_2']
['data_0', 'data_1', 'data_0_2', 'data_0_3', 'data_2', 'data_0_1']
PASSED    [ 35%]
tests/compute/test_sparse.py::test_spmm[int64-min-div-shp0-g1] PASSED  [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-div-shp5-g0] PASSED    [ 36%]
tests/compute/test_sparse.py::test_spmm[int64-min-div-shp1-g0] PASSED  [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-div-shp5-g1] PASSED    [ 36%]
tests/compute/test_sparse.py::test_spmm[int64-min-div-shp1-g1] add...
PASSED  [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_lhs-shp0-g0] PASSED    [ 36%]
tests/compute/test_sparse.py::test_spmm[int64-min-div-shp2-g0] PASSED    [ 36%]
tests/compute/test_sparse.py::test_spmm[int64-min-div-shp2-g1] PASSED [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_lhs-shp0-g1] add...
Client[0] in group[0] is exiting...
Client[1] in group[0] is exiting...
Server (0) shutdown.
Server is exiting...
Server (1) shutdown.
Server is exiting...
PASSED    [ 36%]
tests/compute/test_sparse.py::test_spmm[int64-min-div-shp3-g0] PASSED              [ 61%]
tests/distributed/test_new_kvstore.py::test_kv_multi_role PASSED    [ 36%]
tests/compute/test_sparse.py::test_spmm[int64-min-div-shp3-g1] PASSED [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_lhs-shp1-g0] PASSED    [ 36%]
tests/compute/test_sparse.py::test_spmm[int64-min-div-shp4-g0] /root/jenkins/workspace/dgl_PR-4648/python/dgl/heterograph.py:72: DGLWarning: Recommend creating graphs by `dgl.graph(data)` instead of `dgl.DGLGraph(data)`.
  dgl_warning('Recommend creating graphs by `dgl.graph(data)`'
[05:38:43] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:38:43] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
/root/jenkins/workspace/dgl_PR-4648/python/dgl/heterograph.py:72: DGLWarning: Recommend creating graphs by `dgl.graph(data)` instead of `dgl.DGLGraph(data)`.
  dgl_warning('Recommend creating graphs by `dgl.graph(data)`'
[05:38:43] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:38:43] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:19607]...
/root/jenkins/workspace/dgl_PR-4648/python/dgl/heterograph.py:72: DGLWarning: Recommend creating graphs by `dgl.graph(data)` instead of `dgl.DGLGraph(data)`.
  dgl_warning('Recommend creating graphs by `dgl.graph(data)`'
PASSED           [ 48%]
tests/go/test_pipeline.py::test_linkpred_data[pubmed] PASSED    [ 36%]
tests/compute/test_sparse.py::test_spmm[int64-min-div-shp4-g1] PASSED [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_lhs-shp1-g1] [05:38:43] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:38:43] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
/root/jenkins/workspace/dgl_PR-4648/python/dgl/heterograph.py:72: DGLWarning: Recommend creating graphs by `dgl.graph(data)` instead of `dgl.DGLGraph(data)`.
  dgl_warning('Recommend creating graphs by `dgl.graph(data)`'
[05:38:43] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:38:43] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:19606]...
PASSED    [ 36%]
tests/compute/test_sparse.py::test_spmm[int64-min-div-shp5-g0] PASSED [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_lhs-shp2-g0] [05:38:44] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:38:44] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
[05:38:44] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:38:44] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
PASSED    [ 36%]
tests/compute/test_sparse.py::test_spmm[int64-min-div-shp5-g1] [05:38:44] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:38:44] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
[05:38:44] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:38:44] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
PASSED [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_lhs-shp2-g1] PASSED    [ 36%]
tests/compute/test_sparse.py::test_spmm[int64-min-copy_lhs-shp0-g0] PASSED [ 36%]
tests/compute/test_sparse.py::test_spmm[int64-min-copy_lhs-shp0-g1] PASSED [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_lhs-shp3-g0] PASSED [ 36%]
tests/compute/test_sparse.py::test_spmm[int64-min-copy_lhs-shp1-g0] PASSED [ 36%]
tests/compute/test_sparse.py::test_spmm[int64-min-copy_lhs-shp1-g1] PASSED [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_lhs-shp3-g1] PASSED [ 36%]
tests/compute/test_sparse.py::test_spmm[int64-min-copy_lhs-shp2-g0] PASSED [ 33%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_lhs-shp4-g0] PASSED [ 36%]
tests/compute/test_sparse.py::test_spmm[int64-min-copy_lhs-shp2-g1] PASSED [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_lhs-shp4-g1] PASSED [ 36%]
tests/compute/test_sparse.py::test_spmm[int64-min-copy_lhs-shp3-g0] PASSED [ 36%]
tests/compute/test_sparse.py::test_spmm[int64-min-copy_lhs-shp3-g1] PASSED [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_lhs-shp5-g0] PASSED [ 36%]
tests/compute/test_sparse.py::test_spmm[int64-min-copy_lhs-shp4-g0] PASSED [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_lhs-shp5-g1] PASSED [ 36%]
tests/compute/test_sparse.py::test_spmm[int64-min-copy_lhs-shp4-g1] PASSED [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_rhs-shp0-g0] PASSED             [ 49%]
tests/go/test_pipeline.py::test_linkpred_data[csv] PASSED [ 36%]
tests/compute/test_sparse.py::test_spmm[int64-min-copy_lhs-shp5-g0] PASSED [ 36%]
tests/compute/test_sparse.py::test_spmm[int64-min-copy_lhs-shp5-g1] PASSED [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_rhs-shp0-g1] PASSED [ 36%]
tests/compute/test_sparse.py::test_spmm[int64-min-copy_rhs-shp0-g0] PASSED [ 36%]
tests/compute/test_sparse.py::test_spmm[int64-min-copy_rhs-shp0-g1] Client [3184] waits on 172.17.0.3:60381Client [3191] waits on 172.17.0.3:52669

Client [3183] waits on 172.17.0.3:36033
Client [3173] waits on 172.17.0.3:47463Client [3172] waits on 172.17.0.3:47927
Client [3190] waits on 172.17.0.3:40983

Machine (0) group (0) client (0) connect to server successfuly!
Machine (0) group (0) client (1) connect to server successfuly!
Machine (0) group (0) client (2) connect to server successfuly!
Machine (0) group (0) client (3) connect to server successfuly!
Machine (0) group (0) client (4) connect to server successfuly!
Machine (0) group (0) client (5) connect to server successfuly!
PASSED [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_rhs-shp1-g0] PASSED [ 37%]
tests/compute/test_sparse.py::test_spmm[int64-min-copy_rhs-shp1-g0] PASSED [ 37%]
tests/compute/test_sparse.py::test_spmm[int64-min-copy_rhs-shp1-g1] PASSED [ 37%]
tests/compute/test_sparse.py::test_spmm[int64-min-copy_rhs-shp2-g0] PASSED [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_rhs-shp1-g1] PASSED [ 37%]
tests/compute/test_sparse.py::test_spmm[int64-min-copy_rhs-shp2-g1] PASSED [ 37%]
tests/compute/test_sparse.py::test_spmm[int64-min-copy_rhs-shp3-g0] PASSED [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_rhs-shp2-g0] PASSED [ 37%]
tests/compute/test_sparse.py::test_spmm[int64-min-copy_rhs-shp3-g1] PASSED [ 37%]
tests/compute/test_sparse.py::test_spmm[int64-min-copy_rhs-shp4-g0] PASSED [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_rhs-shp2-g1] PASSED [ 37%]
tests/compute/test_sparse.py::test_spmm[int64-min-copy_rhs-shp4-g1] PASSED [ 37%]
tests/compute/test_sparse.py::test_spmm[int64-min-copy_rhs-shp5-g0] PASSED [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_rhs-shp3-g0] PASSED [ 37%]
tests/compute/test_sparse.py::test_spmm[int64-min-copy_rhs-shp5-g1] PASSED [ 37%]
tests/compute/test_sparse.py::test_spmm[int64-max-add-shp0-g0] PASSED [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_rhs-shp3-g1] PASSED                [ 50%]
tests/go/test_pipeline.py::test_linkpred_data[reddit] PASSED    [ 37%]
tests/compute/test_sparse.py::test_spmm[int64-max-add-shp0-g1] PASSED [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_rhs-shp4-g0] PASSED    [ 37%]
tests/compute/test_sparse.py::test_spmm[int64-max-add-shp1-g0] PASSED    [ 37%]
tests/compute/test_sparse.py::test_spmm[int64-max-add-shp1-g1] PASSED [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_rhs-shp4-g1] PASSED    [ 37%]
tests/compute/test_sparse.py::test_spmm[int64-max-add-shp2-g0] PASSED    [ 37%]
tests/compute/test_sparse.py::test_spmm[int64-max-add-shp2-g1] PASSED [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_rhs-shp5-g0] PASSED    [ 37%]
tests/compute/test_sparse.py::test_spmm[int64-max-add-shp3-g0] PASSED [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-sum-copy_rhs-shp5-g1] [05:38:54] /root/jenkins/workspace/dgl_PR-4648/src/runtime/semaphore_wrapper.cc:83: sem_timedwait timed out after 5000 milliseconds.
[05:38:54] /root/jenkins/workspace/dgl_PR-4648/src/rpc/network/socket_communicator.cc:283: Timed out when trying to receive rpc meta data after 5000 milliseconds.
[05:38:54] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:84: Recv RPCMessage timeout in 5000 ms. Retrying ...
i: 0 role: default
trainer rank: 1, global rank: 1
Client[3] in group[0] is exiting...
i: 1 role: default
trainer rank: 0, global rank: 0
Client[2] in group[0] is exiting...
Client[0] in group[0] is exiting...
Client[1] in group[0] is exiting...
Client[5] in group[0] is exiting...
Client[4] in group[0] is exiting...
PASSED    [ 37%]
tests/compute/test_sparse.py::test_spmm[int64-max-add-shp3-g1] PASSED [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-add-shp0-g0] Server (0) shutdown.
Server is exiting...
Server (1) shutdown.
Server is exiting...
PASSED    [ 37%]
tests/compute/test_sparse.py::test_spmm[int64-max-add-shp4-g0] PASSED    [ 37%]
tests/compute/test_sparse.py::test_spmm[int64-max-add-shp4-g1] PASSED  [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-add-shp0-g1] PASSED    [ 37%]
tests/compute/test_sparse.py::test_spmm[int64-max-add-shp5-g0] PASSED  [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-add-shp1-g0] PASSED    [ 37%]
tests/compute/test_sparse.py::test_spmm[int64-max-add-shp5-g1] PASSED         [ 63%]
tests/distributed/test_partition.py::test_partition PASSED  [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-add-shp1-g1] FAILED               [ 65%]
tests/distributed/test_partition.py::test_hetero_partition Converting to homogeneous graph takes 0.001s, peak mem: 1.841 GB
Convert a graph into a bidirected graph: 0.000 seconds, peak memory: 1.841 GB
Construct multi-constraint weights: 0.000 seconds, peak memory: 1.841 GB
[05:38:57] /root/jenkins/workspace/dgl_PR-4648/src/graph/transform/metis_partition_hetero.cc:87: Partition a graph with 3000 nodes and 6000 edges into 2 parts and get 216 edge cuts
Metis partitioning: 0.003 seconds, peak memory: 1.841 GB
Assigning nodes to METIS partitions takes 0.004s, peak mem: 1.841 GB
Reshuffle nodes and edges: 0.001 seconds
Split the graph: 0.002 seconds
Construct subgraphs: 0.002 seconds
Splitting the graph into partitions takes 0.005s, peak mem: 1.841 GB
part 0 has 1142 nodes of type item and 1030 are inside the partition
part 0 has 607 nodes of type user and 514 are inside the partition
part 0 has 1693 edges of type like and 1575 are inside the partition
part 1 has 1064 nodes of type item and 970 are inside the partition
part 1 has 589 nodes of type user and 486 are inside the partition
part 1 has 1523 edges of type like and 1425 are inside the partition
Save partitions: 0.002 seconds, peak memory: 1.841 GB
There are 3000 edges in the graph and 0 edge cuts for 2 partitions.
Converting to homogeneous graph takes 0.001s, peak mem: 1.841 GB
Convert a graph into a bidirected graph: 0.000 seconds, peak memory: 1.841 GB
Construct multi-constraint weights: 0.000 seconds, peak memory: 1.841 GB
[05:38:57] /root/jenkins/workspace/dgl_PR-4648/src/graph/transform/metis_partition_hetero.cc:87: Partition a graph with 3000 nodes and 6000 edges into 8 parts and get 449 edge cuts
Metis partitioning: 0.007 seconds, peak memory: 1.841 GB
Assigning nodes to METIS partitions takes 0.009s, peak mem: 1.841 GB
Reshuffle nodes and edges: 0.001 seconds
Split the graph: 0.002 seconds
Construct subgraphs: 0.002 seconds
Splitting the graph into partitions takes 0.005s, peak mem: 1.841 GB
part 0 has 1096 nodes of type item and 998 are inside the partition
part 0 has 623 nodes of type user and 502 are inside the partition
part 0 has 1587 edges of type like and 1475 are inside the partition
part 1 has 1134 nodes of type item and 1002 are inside the partition
part 1 has 597 nodes of type user and 498 are inside the partition
part 1 has 1665 edges of type like and 1525 are inside the partition
Save partitions: 0.002 seconds, peak memory: 1.841 GB
There are 3000 edges in the graph and 0 edge cuts for 2 partitions.
PASSED    [ 37%]
tests/compute/test_sparse.py::test_spmm[int64-max-sub-shp0-g0] Converting to homogeneous graph takes 0.001s, peak mem: 1.841 GB
Convert a graph into a bidirected graph: 0.000 seconds, peak memory: 1.841 GB
Construct multi-constraint weights: 0.000 seconds, peak memory: 1.841 GB
[05:38:57] /root/jenkins/workspace/dgl_PR-4648/src/graph/transform/metis_partition_hetero.cc:87: Partition a graph with 3030 nodes and 6120 edges into 4 parts and get 336 edge cuts
Metis partitioning: 0.004 seconds, peak memory: 1.841 GB
Assigning nodes to METIS partitions takes 0.005s, peak mem: 1.841 GB
Reshuffle nodes and edges: 0.001 seconds
Split the graph: 0.001 seconds
Construct subgraphs: 0.004 seconds
Splitting the graph into partitions takes 0.006s, peak mem: 1.841 GB
part 0 has 295 nodes of type n1 and 248 are inside the partition
part 0 has 306 nodes of type n2 and 252 are inside the partition
part 0 has 312 nodes of type n3 and 258 are inside the partition
part 0 has 284 edges of type r1 and 255 are inside the partition
part 0 has 265 edges of type r2 and 240 are inside the partition
part 0 has 275 edges of type r3 and 244 are inside the partition
part 1 has 316 nodes of type n1 and 257 are inside the partition
part 1 has 317 nodes of type n2 and 257 are inside the partition
part 1 has 321 nodes of type n3 and 262 are inside the partition
part 1 has 274 edges of type r1 and 247 are inside the partition
part 1 has 305 edges of type r2 and 277 are inside the partition
part 1 has 320 edges of type r3 and 287 are inside the partition
part 2 has 307 nodes of type n1 and 257 are inside the partition
part 2 has 315 nodes of type n2 and 259 are inside the partition
part 2 has 323 nodes of type n3 and 261 are inside the partition
part 2 has 272 edges of type r1 and 250 are inside the partition
part 2 has 325 edges of type r2 and 291 are inside the partition
part 2 has 289 edges of type r3 and 259 are inside the partition
part 3 has 288 nodes of type n1 and 238 are inside the partition
part 3 has 293 nodes of type n2 and 242 are inside the partition
part 3 has 288 nodes of type n3 and 239 are inside the partition
part 3 has 286 edges of type r1 and 258 are inside the partition
part 3 has 234 edges of type r2 and 212 are inside the partition
part 3 has 267 edges of type r3 and 240 are inside the partition
Save partitions: 0.005 seconds, peak memory: 1.841 GB
There are 3060 edges in the graph and 0 edge cuts for 4 partitions.
node n1: 1000, 1000
node n2: 1010, 1010
node n3: 1020, 1020
edge r1: 1010, 1010
edge r2: 1020, 1020
edge r3: 1030, 1030
Converting to homogeneous graph takes 0.001s, peak mem: 1.841 GB
Convert a graph into a bidirected graph: 0.000 seconds, peak memory: 1.841 GB
Construct multi-constraint weights: 0.000 seconds, peak memory: 1.841 GB
[05:38:57] /root/jenkins/workspace/dgl_PR-4648/src/graph/transform/metis_partition_hetero.cc:87: Partition a graph with 3030 nodes and 6120 edges into 8 parts and get 443 edge cuts
Metis partitioning: 0.008 seconds, peak memory: 1.841 GB
Assigning nodes to METIS partitions takes 0.010s, peak mem: 1.841 GB
Save partitions: 0.002 seconds, peak memory: 1.841 GB
There are 3060 edges in the graph and 0 edge cuts for 1 partitions.
node n1: 1000, 1000
node n2: 1010, 1010
node n3: 1020, 1020
edge r1: 1010, 1010
edge r2: 1020, 1020
edge r3: 1030, 1030
Converting to homogeneous graph takes 0.001s, peak mem: 1.841 GB
Convert a graph into a bidirected graph: 0.000 seconds, peak memory: 1.841 GB
Construct multi-constraint weights: 0.000 seconds, peak memory: 1.841 GB
[05:38:57] /root/jenkins/workspace/dgl_PR-4648/src/graph/transform/metis_partition_hetero.cc:87: Partition a graph with 3030 nodes and 6120 edges into 32 parts and get 794 edge cuts
Metis partitioning: 0.030 seconds, peak memory: 1.841 GB
Assigning nodes to METIS partitions takes 0.031s, peak mem: 1.841 GB
Reshuffle nodes and edges: 0.001 seconds
Split the graph: 0.002 seconds
Construct subgraphs: 0.004 seconds
Splitting the graph into partitions takes 0.007s, peak mem: 1.841 GB
part 0 has 329 nodes of type n1 and 248 are inside the partition
part 0 has 317 nodes of type n2 and 249 are inside the partition
part 0 has 334 nodes of type n3 and 255 are inside the partition
part 0 has 262 edges of type r1 and 227 are inside the partition
part 0 has 302 edges of type r2 and 264 are inside the partition
part 0 has 303 edges of type r3 and 257 are inside the partition
part 1 has 324 nodes of type n1 and 248 are inside the partition
part 1 has 343 nodes of type n2 and 252 are inside the partition
part 1 has 347 nodes of type n3 and 255 are inside the partition
part 1 has 310 edges of type r1 and 262 are inside the partition
part 1 has 321 edges of type r2 and 270 are inside the partition
part 1 has 316 edges of type r3 and 267 are inside the partition
part 2 has 335 nodes of type n1 and 256 are inside the partition
part 2 has 328 nodes of type n2 and 254 are inside the partition
part 2 has 352 nodes of type n3 and 255 are inside the partition
part 2 has 310 edges of type r1 and 277 are inside the partition
part 2 has 274 edges of type r2 and 227 are inside the partition
part 2 has 291 edges of type r3 and 238 are inside the partition
part 3 has 328 nodes of type n1 and 248 are inside the partition
part 3 has 339 nodes of type n2 and 255 are inside the partition
part 3 has 351 nodes of type n3 and 255 are inside the partition
part 3 has 283 edges of type r1 and 244 are inside the partition
part 3 has 314 edges of type r2 and 259 are inside the partition
part 3 has 320 edges of type r3 and 268 are inside the partition
Save partitions: 0.006 seconds, peak memory: 1.841 GB
There are 3060 edges in the graph and 0 edge cuts for 4 partitions.
node n1: 1000, 1000
node n2: 1010, 1010
node n3: 1020, 1020
edge r1: 1010, 1010
edge r2: 1020, 1020
edge r3: 1030, 1030
Converting to homogeneous graph takes 0.001s, peak mem: 1.841 GB
Reshuffle nodes and edges: 0.001 seconds
Split the graph: 0.002 seconds
Construct subgraphs: 0.004 seconds
Splitting the graph into partitions takes 0.007s, peak mem: 1.841 GB
part 0 has 531 nodes of type n1 and 238 are inside the partition
part 0 has 545 nodes of type n2 and 267 are inside the partition
part 0 has 563 nodes of type n3 and 244 are inside the partition
part 0 has 445 edges of type r1 and 272 are inside the partition
part 0 has 448 edges of type r2 and 240 are inside the partition
part 0 has 435 edges of type r3 and 230 are inside the partition
part 1 has 553 nodes of type n1 and 263 are inside the partition
part 1 has 566 nodes of type n2 and 251 are inside the partition
part 1 has 552 nodes of type n3 and 260 are inside the partition
part 1 has 432 edges of type r1 and 235 are inside the partition
part 1 has 452 edges of type r2 and 274 are inside the partition
part 1 has 469 edges of type r3 and 270 are inside the partition
part 2 has 545 nodes of type n1 and 263 are inside the partition
part 2 has 534 nodes of type n2 and 248 are inside the partition
part 2 has 557 nodes of type n3 and 255 are inside the partition
part 2 has 425 edges of type r1 and 250 are inside the partition
part 2 has 429 edges of type r2 and 233 are inside the partition
part 2 has 439 edges of type r3 and 247 are inside the partition
part 3 has 551 nodes of type n1 and 236 are inside the partition
part 3 has 559 nodes of type n2 and 244 are inside the partition
part 3 has 542 nodes of type n3 and 261 are inside the partition
part 3 has 449 edges of type r1 and 253 are inside the partition
part 3 has 453 edges of type r2 and 273 are inside the partition
part 3 has 463 edges of type r3 and 283 are inside the partition
Save partitions: 0.006 seconds, peak memory: 1.841 GB
There are 3060 edges in the graph and 0 edge cuts for 4 partitions.
PASSED    [ 37%]
tests/compute/test_sparse.py::test_spmm[int64-max-sub-shp0-g1] node n1: 1000, 1000
node n2: 1010, 1010
node n3: 1020, 1020
edge r1: 1010, 1010
edge r2: 1020, 1020
edge r3: 1030, 1030
Converting to homogeneous graph takes 0.001s, peak mem: 1.841 GB
Convert a graph into a bidirected graph: 0.000 seconds, peak memory: 1.841 GB
Construct multi-constraint weights: 0.000 seconds, peak memory: 1.841 GB
[05:38:58] /root/jenkins/workspace/dgl_PR-4648/src/graph/transform/metis_partition_hetero.cc:87: Partition a graph with 3030 nodes and 6120 edges into 32 parts and get 794 edge cuts
Metis partitioning: 0.026 seconds, peak memory: 1.841 GB
Assigning nodes to METIS partitions takes 0.028s, peak mem: 1.841 GB
Reshuffle nodes and edges: 0.001 seconds
Split the graph: 0.001 seconds
Construct subgraphs: 0.003 seconds
Splitting the graph into partitions takes 0.006s, peak mem: 1.841 GB
part 0 has 329 nodes of type n1 and 248 are inside the partition
part 0 has 317 nodes of type n2 and 249 are inside the partition
part 0 has 334 nodes of type n3 and 255 are inside the partition
part 0 has 262 edges of type r1 and 227 are inside the partition
part 0 has 302 edges of type r2 and 264 are inside the partition
part 0 has 303 edges of type r3 and 257 are inside the partition
part 1 has 324 nodes of type n1 and 248 are inside the partition
part 1 has 343 nodes of type n2 and 252 are inside the partition
part 1 has 347 nodes of type n3 and 255 are inside the partition
part 1 has 310 edges of type r1 and 262 are inside the partition
part 1 has 321 edges of type r2 and 270 are inside the partition
part 1 has 316 edges of type r3 and 267 are inside the partition
part 2 has 335 nodes of type n1 and 256 are inside the partition
part 2 has 328 nodes of type n2 and 254 are inside the partition
part 2 has 352 nodes of type n3 and 255 are inside the partition
part 2 has 310 edges of type r1 and 277 are inside the partition
part 2 has 274 edges of type r2 and 227 are inside the partition
part 2 has 291 edges of type r3 and 238 are inside the partition
part 3 has 328 nodes of type n1 and 248 are inside the partition
part 3 has 339 nodes of type n2 and 255 are inside the partition
part 3 has 351 nodes of type n3 and 255 are inside the partition
part 3 has 283 edges of type r1 and 244 are inside the partition
part 3 has 314 edges of type r2 and 259 are inside the partition
part 3 has 320 edges of type r3 and 268 are inside the partition
Save partitions: 0.006 seconds, peak memory: 1.841 GB
There are 3060 edges in the graph and 0 edge cuts for 4 partitions.
node n1: 1000, 1000
node n2: 1010, 1010
node n3: 1020, 1020
edge r1: 1010, 1010
edge r2: 1020, 1020
edge r3: 1030, 1030
PASSED        [ 67%]
tests/distributed/test_partition.py::test_BasicPartitionBook PASSED      [ 69%]
tests/distributed/test_partition.py::test_RangePartitionBook PASSED      [ 70%]
tests/distributed/test_rpc.py::test_rpc_timeout[socket] PASSED  [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-add-shp2-g0] PASSED             [ 51%]
tests/go/test_pipeline.py::test_linkpred_data[co-buy-computer] PASSED    [ 37%]
tests/compute/test_sparse.py::test_spmm[int64-max-sub-shp1-g0] PASSED  [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-add-shp2-g1] PASSED    [ 38%]
tests/compute/test_sparse.py::test_spmm[int64-max-sub-shp1-g1] Sleep 1 seconds to test client re-connect.
[05:38:59] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:38:59] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
PASSED  [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-add-shp3-g0] PASSED    [ 38%]
tests/compute/test_sparse.py::test_spmm[int64-max-sub-shp2-g0] PASSED    [ 38%]
tests/compute/test_sparse.py::test_spmm[int64-max-sub-shp2-g1] PASSED  [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-add-shp3-g1] Start server 0
[05:39:00] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:39:00] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:21850]...
PASSED    [ 38%]
tests/compute/test_sparse.py::test_spmm[int64-max-sub-shp3-g0] PASSED    [ 38%]
tests/compute/test_sparse.py::test_spmm[int64-max-sub-shp3-g1] PASSED  [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-add-shp4-g0] PASSED    [ 38%]
tests/compute/test_sparse.py::test_spmm[int64-max-sub-shp4-g0] PASSED  [ 34%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-add-shp4-g1] PASSED    [ 38%]
tests/compute/test_sparse.py::test_spmm[int64-max-sub-shp4-g1] PASSED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-add-shp5-g0] PASSED    [ 38%]
tests/compute/test_sparse.py::test_spmm[int64-max-sub-shp5-g0] PASSED    [ 38%]
tests/compute/test_sparse.py::test_spmm[int64-max-sub-shp5-g1] PASSED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-add-shp5-g1] PASSED    [ 38%]
tests/compute/test_sparse.py::test_spmm[int64-max-mul-shp0-g0] PASSED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-sub-shp0-g0] PASSED    [ 38%]
tests/compute/test_sparse.py::test_spmm[int64-max-mul-shp0-g1] PASSED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-sub-shp0-g1] PASSED    [ 38%]
tests/compute/test_sparse.py::test_spmm[int64-max-mul-shp1-g0] PASSED    [ 38%]
tests/compute/test_sparse.py::test_spmm[int64-max-mul-shp1-g1] PASSED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-sub-shp1-g0] PASSED    [ 53%]
tests/go/test_pipeline.py::test_linkpred_data[ogbn-arxiv] PASSED    [ 38%]
tests/compute/test_sparse.py::test_spmm[int64-max-mul-shp2-g0] PASSED    [ 38%]
tests/compute/test_sparse.py::test_spmm[int64-max-mul-shp2-g1] PASSED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-sub-shp1-g1] PASSED    [ 38%]
tests/compute/test_sparse.py::test_spmm[int64-max-mul-shp3-g0] PASSED    [ 38%]
tests/compute/test_sparse.py::test_spmm[int64-max-mul-shp3-g1] PASSED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-sub-shp2-g0] PASSED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-sub-shp2-g1] Client [3259] waits on 172.17.0.3:34383
Machine (0) group (0) client (0) connect to server successfuly!
PASSED    [ 38%]
tests/compute/test_sparse.py::test_spmm[int64-max-mul-shp4-g0] PASSED    [ 38%]
tests/compute/test_sparse.py::test_spmm[int64-max-mul-shp4-g1] PASSED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-sub-shp3-g0] [05:39:05] /root/jenkins/workspace/dgl_PR-4648/src/runtime/semaphore_wrapper.cc:83: sem_timedwait timed out after 500 milliseconds.
[05:39:05] /root/jenkins/workspace/dgl_PR-4648/src/rpc/network/socket_communicator.cc:283: Timed out when trying to receive rpc meta data after 500 milliseconds.
[05:39:05] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:84: Recv RPCMessage timeout in 500 ms.
PASSED    [ 38%]
tests/compute/test_sparse.py::test_spmm[int64-max-mul-shp5-g0] PASSED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-sub-shp3-g1] PASSED    [ 38%]
tests/compute/test_sparse.py::test_spmm[int64-max-mul-shp5-g1] [05:39:06] /root/jenkins/workspace/dgl_PR-4648/src/runtime/semaphore_wrapper.cc:83: sem_timedwait timed out after 500 milliseconds.
[05:39:06] /root/jenkins/workspace/dgl_PR-4648/src/rpc/network/socket_communicator.cc:283: Timed out when trying to receive rpc meta data after 500 milliseconds.
[05:39:06] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:84: Recv RPCMessage timeout in 500 ms.
PASSED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-sub-shp4-g0] PASSED    [ 38%]
tests/compute/test_sparse.py::test_spmm[int64-max-div-shp0-g0] [05:39:07] /root/jenkins/workspace/dgl_PR-4648/src/runtime/semaphore_wrapper.cc:83: sem_timedwait timed out after 500 milliseconds.
[05:39:07] /root/jenkins/workspace/dgl_PR-4648/src/rpc/network/socket_communicator.cc:283: Timed out when trying to receive rpc meta data after 500 milliseconds.
[05:39:07] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:84: Recv RPCMessage timeout in 500 ms.
PASSED    [ 38%]
tests/compute/test_sparse.py::test_spmm[int64-max-div-shp0-g1] PASSED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-sub-shp4-g1] PASSED    [ 38%]
tests/compute/test_sparse.py::test_spmm[int64-max-div-shp1-g0] PASSED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-sub-shp5-g0] PASSED    [ 39%]
tests/compute/test_sparse.py::test_spmm[int64-max-div-shp1-g1] PASSED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-sub-shp5-g1] PASSED    [ 39%]
tests/compute/test_sparse.py::test_spmm[int64-max-div-shp2-g0] PASSED    [ 39%]
tests/compute/test_sparse.py::test_spmm[int64-max-div-shp2-g1] PASSED         [ 54%]
tests/go/test_pipeline.py::test_linkpred_data[ogbn-products] PASSED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-mul-shp0-g0] PASSED    [ 39%]
tests/compute/test_sparse.py::test_spmm[int64-max-div-shp3-g0] PASSED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-mul-shp0-g1] PASSED    [ 39%]
tests/compute/test_sparse.py::test_spmm[int64-max-div-shp3-g1] PASSED    [ 39%]
tests/compute/test_sparse.py::test_spmm[int64-max-div-shp4-g0] PASSED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-mul-shp1-g0] PASSED    [ 39%]
tests/compute/test_sparse.py::test_spmm[int64-max-div-shp4-g1] [05:39:10] /root/jenkins/workspace/dgl_PR-4648/src/runtime/semaphore_wrapper.cc:83: sem_timedwait timed out after 500 milliseconds.
[05:39:10] /root/jenkins/workspace/dgl_PR-4648/src/rpc/network/socket_communicator.cc:283: Timed out when trying to receive rpc meta data after 500 milliseconds.
[05:39:10] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:84: Recv RPCMessage timeout in 500 ms.
Client[0] in group[0] is exiting...
PASSED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-mul-shp1-g1] PASSED    [ 39%]
tests/compute/test_sparse.py::test_spmm[int64-max-div-shp5-g0] PASSED    [ 39%]
tests/compute/test_sparse.py::test_spmm[int64-max-div-shp5-g1] PASSED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-mul-shp2-g0] PASSED    [ 39%]
tests/compute/test_sparse.py::test_spmm[int64-max-copy_lhs-shp0-g0] PASSED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-mul-shp2-g1] PASSED [ 39%]
tests/compute/test_sparse.py::test_spmm[int64-max-copy_lhs-shp0-g1] PASSED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-mul-shp3-g0] PASSED [ 39%]
tests/compute/test_sparse.py::test_spmm[int64-max-copy_lhs-shp1-g0] PASSED [ 39%]
tests/compute/test_sparse.py::test_spmm[int64-max-copy_lhs-shp1-g1] PASSED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-mul-shp3-g1] PASSED [ 39%]
tests/compute/test_sparse.py::test_spmm[int64-max-copy_lhs-shp2-g0] PASSED [ 39%]
tests/compute/test_sparse.py::test_spmm[int64-max-copy_lhs-shp2-g1] PASSED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-mul-shp4-g0] PASSED [ 39%]
tests/compute/test_sparse.py::test_spmm[int64-max-copy_lhs-shp3-g0] PASSED [ 39%]
tests/compute/test_sparse.py::test_spmm[int64-max-copy_lhs-shp3-g1] PASSED  [ 35%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-mul-shp4-g1] PASSED      [ 55%]
tests/go/test_pipeline.py::test_linkpred_data[ogbl-collab] Server (0) shutdown.
Server is exiting...
PASSED           [ 72%]
tests/distributed/test_rpc.py::test_rpc_timeout[tensorpipe] PASSED [ 39%]
tests/compute/test_sparse.py::test_spmm[int64-max-copy_lhs-shp4-g0] PASSED  [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-mul-shp5-g0] PASSED [ 39%]
tests/compute/test_sparse.py::test_spmm[int64-max-copy_lhs-shp4-g1] PASSED  [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-mul-shp5-g1] [05:39:15] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~tensorpipe is created.
[05:39:15] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~tensorpipe is created.
[05:39:15] /root/jenkins/workspace/dgl_PR-4648/src/rpc/tensorpipe/tp_communicator.cc:37: Failed to connect to receiver[tcp://172.17.0.3:28918].
Sleep 1 seconds to test client re-connect.
PASSED [ 39%]
tests/compute/test_sparse.py::test_spmm[int64-max-copy_lhs-shp5-g0] PASSED [ 39%]
tests/compute/test_sparse.py::test_spmm[int64-max-copy_lhs-shp5-g1] PASSED  [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-div-shp0-g0] Start server 0
[05:39:16] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~tensorpipe is created.
[05:39:16] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~tensorpipe is created.
Server is waiting for connections on [172.17.0.3:28918]...
[05:39:16] /root/jenkins/workspace/dgl_PR-4648/src/rpc/tensorpipe/tp_communicator.cc:108: TPReceiver starts to wait on [tcp://172.17.0.3:28918].
PASSED  [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-div-shp0-g1] PASSED [ 39%]
tests/compute/test_sparse.py::test_spmm[int64-max-copy_rhs-shp0-g0] PASSED [ 39%]
tests/compute/test_sparse.py::test_spmm[int64-max-copy_rhs-shp0-g1] PASSED  [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-div-shp1-g0] PASSED [ 39%]
tests/compute/test_sparse.py::test_spmm[int64-max-copy_rhs-shp1-g0] PASSED [ 39%]
tests/compute/test_sparse.py::test_spmm[int64-max-copy_rhs-shp1-g1] PASSED  [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-div-shp1-g1] PASSED [ 40%]
tests/compute/test_sparse.py::test_spmm[int64-max-copy_rhs-shp2-g0] PASSED [ 40%]
tests/compute/test_sparse.py::test_spmm[int64-max-copy_rhs-shp2-g1] PASSED [ 40%]
tests/compute/test_sparse.py::test_spmm[int64-max-copy_rhs-shp3-g0] PASSED  [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-div-shp2-g0] PASSED [ 40%]
tests/compute/test_sparse.py::test_spmm[int64-max-copy_rhs-shp3-g1] PASSED [ 40%]
tests/compute/test_sparse.py::test_spmm[int64-max-copy_rhs-shp4-g0] PASSED        [ 56%]
tests/go/test_pipeline.py::test_linkpred_data[ogbl-citation2] PASSED  [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-div-shp2-g1] PASSED [ 40%]
tests/compute/test_sparse.py::test_spmm[int64-max-copy_rhs-shp4-g1] [05:39:18] /root/jenkins/workspace/dgl_PR-4648/src/rpc/tensorpipe/tp_communicator.cc:108: TPReceiver starts to wait on [tcp://172.17.0.3:41135].
Client [3279] waits on 172.17.0.3:41135
PASSED [ 40%]
tests/compute/test_sparse.py::test_spmm[int64-max-copy_rhs-shp5-g0] PASSED  [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-div-shp3-g0] PASSED [ 40%]
tests/compute/test_sparse.py::test_spmm[int64-max-copy_rhs-shp5-g1] PASSED [ 40%]
tests/compute/test_sparse.py::test_sddmm[int32-add-u-u-shp0-g0] PASSED   [ 40%]
tests/compute/test_sparse.py::test_sddmm[int32-add-u-u-shp0-g1] PASSED   [ 40%]
tests/compute/test_sparse.py::test_sddmm[int32-add-u-u-shp1-g0] PASSED   [ 40%]
tests/compute/test_sparse.py::test_sddmm[int32-add-u-u-shp1-g1] PASSED   [ 40%]
tests/compute/test_sparse.py::test_sddmm[int32-add-u-u-shp2-g0] PASSED   [ 40%]
tests/compute/test_sparse.py::test_sddmm[int32-add-u-u-shp2-g1] PASSED   [ 40%]
tests/compute/test_sparse.py::test_sddmm[int32-add-u-u-shp3-g0] PASSED   [ 40%]
tests/compute/test_sparse.py::test_sddmm[int32-add-u-u-shp3-g1] PASSED   [ 40%]
tests/compute/test_sparse.py::test_sddmm[int32-add-u-u-shp4-g0] PASSED   [ 40%]
tests/compute/test_sparse.py::test_sddmm[int32-add-u-u-shp4-g1] PASSED   [ 40%]
tests/compute/test_sparse.py::test_sddmm[int32-add-u-v-shp0-g0] PASSED   [ 40%]
tests/compute/test_sparse.py::test_sddmm[int32-add-u-v-shp0-g1] PASSED  [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-div-shp3-g1] PASSED   [ 40%]
tests/compute/test_sparse.py::test_sddmm[int32-add-u-v-shp1-g0] PASSED   [ 40%]
tests/compute/test_sparse.py::test_sddmm[int32-add-u-v-shp1-g1] PASSED  [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-div-shp4-g0] PASSED   [ 40%]
tests/compute/test_sparse.py::test_sddmm[int32-add-u-v-shp2-g0] PASSED   [ 40%]
tests/compute/test_sparse.py::test_sddmm[int32-add-u-v-shp2-g1] PASSED  [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-div-shp4-g1] PASSED   [ 41%]
tests/compute/test_sparse.py::test_sddmm[int32-add-u-v-shp3-g0] PASSED   [ 41%]
tests/compute/test_sparse.py::test_sddmm[int32-add-u-v-shp3-g1] Machine (0) group (0) client (0) connect to server successfuly!
PASSED  [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-div-shp5-g0] PASSED   [ 41%]
tests/compute/test_sparse.py::test_sddmm[int32-add-u-v-shp4-g0] PASSED   [ 41%]
tests/compute/test_sparse.py::test_sddmm[int32-add-u-v-shp4-g1] [05:39:21] /root/jenkins/workspace/dgl_PR-4648/src/rpc/tensorpipe/./queue.h:44: Times out for popping message after 500 milliseconds.
[05:39:21] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:84: Recv RPCMessage timeout in 500 ms.
PASSED  [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-div-shp5-g1] PASSED   [ 41%]
tests/compute/test_sparse.py::test_sddmm[int32-add-u-e-shp0-g0] PASSED   [ 41%]
tests/compute/test_sparse.py::test_sddmm[int32-add-u-e-shp0-g1] PASSED   [ 41%]
tests/compute/test_sparse.py::test_sddmm[int32-add-u-e-shp1-g0] PASSED   [ 41%]
tests/compute/test_sparse.py::test_sddmm[int32-add-u-e-shp1-g1] PASSED  [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_lhs-shp0-g0] PASSED   [ 41%]
tests/compute/test_sparse.py::test_sddmm[int32-add-u-e-shp2-g0] PASSED   [ 41%]
tests/compute/test_sparse.py::test_sddmm[int32-add-u-e-shp2-g1] [05:39:22] /root/jenkins/workspace/dgl_PR-4648/src/rpc/tensorpipe/./queue.h:44: Times out for popping message after 500 milliseconds.
[05:39:22] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:84: Recv RPCMessage timeout in 500 ms.
PASSED   [ 41%]
tests/compute/test_sparse.py::test_sddmm[int32-add-u-e-shp3-g0] PASSED   [ 41%]
tests/compute/test_sparse.py::test_sddmm[int32-add-u-e-shp3-g1] PASSED [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_lhs-shp0-g1] PASSED   [ 41%]
tests/compute/test_sparse.py::test_sddmm[int32-add-u-e-shp4-g0] PASSED   [ 41%][05:39:23] /root/jenkins/workspace/dgl_PR-4648/src/rpc/tensorpipe/./queue.h:44: Times out for popping message after 500 milliseconds.
[05:39:23] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:84: Recv RPCMessage timeout in 500 ms.

tests/compute/test_sparse.py::test_sddmm[int32-add-u-e-shp4-g1] PASSED   [ 41%]
tests/compute/test_sparse.py::test_sddmm[int32-add-v-u-shp0-g0] PASSED [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_lhs-shp1-g0] PASSED     [ 58%]
tests/go/test_pipeline.py::test_linkpred_node_model[gcn] PASSED   [ 41%]
tests/compute/test_sparse.py::test_sddmm[int32-add-v-u-shp0-g1] PASSED   [ 41%]
tests/compute/test_sparse.py::test_sddmm[int32-add-v-u-shp1-g0] PASSED [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_lhs-shp1-g1] PASSED   [ 41%]
tests/compute/test_sparse.py::test_sddmm[int32-add-v-u-shp1-g1] PASSED   [ 41%]
tests/compute/test_sparse.py::test_sddmm[int32-add-v-u-shp2-g0] PASSED   [ 41%]
tests/compute/test_sparse.py::test_sddmm[int32-add-v-u-shp2-g1] PASSED   [ 41%]
tests/compute/test_sparse.py::test_sddmm[int32-add-v-u-shp3-g0] PASSED [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_lhs-shp2-g0] PASSED   [ 41%]
tests/compute/test_sparse.py::test_sddmm[int32-add-v-u-shp3-g1] PASSED [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_lhs-shp2-g1] PASSED   [ 41%]
tests/compute/test_sparse.py::test_sddmm[int32-add-v-u-shp4-g0] PASSED   [ 41%]
tests/compute/test_sparse.py::test_sddmm[int32-add-v-u-shp4-g1] PASSED [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_lhs-shp3-g0] PASSED   [ 41%]
tests/compute/test_sparse.py::test_sddmm[int32-add-v-v-shp0-g0] PASSED   [ 42%]
tests/compute/test_sparse.py::test_sddmm[int32-add-v-v-shp0-g1] PASSED   [ 42%]
tests/compute/test_sparse.py::test_sddmm[int32-add-v-v-shp1-g0] PASSED   [ 42%]
tests/compute/test_sparse.py::test_sddmm[int32-add-v-v-shp1-g1] PASSED   [ 42%]
tests/compute/test_sparse.py::test_sddmm[int32-add-v-v-shp2-g0] PASSED   [ 42%]
tests/compute/test_sparse.py::test_sddmm[int32-add-v-v-shp2-g1] PASSED   [ 42%]
tests/compute/test_sparse.py::test_sddmm[int32-add-v-v-shp3-g0] PASSED   [ 42%]
tests/compute/test_sparse.py::test_sddmm[int32-add-v-v-shp3-g1] PASSED   [ 42%]
tests/compute/test_sparse.py::test_sddmm[int32-add-v-v-shp4-g0] PASSED   [ 42%]
tests/compute/test_sparse.py::test_sddmm[int32-add-v-v-shp4-g1] PASSED   [ 42%]
tests/compute/test_sparse.py::test_sddmm[int32-add-v-e-shp0-g0] PASSED   [ 42%]
tests/compute/test_sparse.py::test_sddmm[int32-add-v-e-shp0-g1] PASSED   [ 42%]
tests/compute/test_sparse.py::test_sddmm[int32-add-v-e-shp1-g0] PASSED [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_lhs-shp3-g1] PASSED   [ 42%]
tests/compute/test_sparse.py::test_sddmm[int32-add-v-e-shp1-g1] PASSED   [ 42%]
tests/compute/test_sparse.py::test_sddmm[int32-add-v-e-shp2-g0] PASSED   [ 42%]
tests/compute/test_sparse.py::test_sddmm[int32-add-v-e-shp2-g1] PASSED [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_lhs-shp4-g0] PASSED   [ 42%]
tests/compute/test_sparse.py::test_sddmm[int32-add-v-e-shp3-g0] PASSED   [ 42%]
tests/compute/test_sparse.py::test_sddmm[int32-add-v-e-shp3-g1] PASSED   [ 42%]
tests/compute/test_sparse.py::test_sddmm[int32-add-v-e-shp4-g0] PASSED   [ 42%]
tests/compute/test_sparse.py::test_sddmm[int32-add-v-e-shp4-g1] [05:39:26] /root/jenkins/workspace/dgl_PR-4648/src/rpc/tensorpipe/./queue.h:44: Times out for popping message after 500 milliseconds.
[05:39:26] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:84: Recv RPCMessage timeout in 500 ms.
Client[0] in group[0] is exiting...
PASSED   [ 42%]
tests/compute/test_sparse.py::test_sddmm[int32-add-e-u-shp0-g0] PASSED   [ 42%]
tests/compute/test_sparse.py::test_sddmm[int32-add-e-u-shp0-g1] PASSED [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_lhs-shp4-g1] PASSED   [ 42%]
tests/compute/test_sparse.py::test_sddmm[int32-add-e-u-shp1-g0] PASSED   [ 42%]
tests/compute/test_sparse.py::test_sddmm[int32-add-e-u-shp1-g1] PASSED   [ 42%]
tests/compute/test_sparse.py::test_sddmm[int32-add-e-u-shp2-g0] PASSED [ 36%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_lhs-shp5-g0] PASSED   [ 43%]
tests/compute/test_sparse.py::test_sddmm[int32-add-e-u-shp2-g1] PASSED   [ 43%]
tests/compute/test_sparse.py::test_sddmm[int32-add-e-u-shp3-g0] PASSED   [ 43%]
tests/compute/test_sparse.py::test_sddmm[int32-add-e-u-shp3-g1] PASSED [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_lhs-shp5-g1] PASSED   [ 43%]
tests/compute/test_sparse.py::test_sddmm[int32-add-e-u-shp4-g0] PASSED   [ 43%]
tests/compute/test_sparse.py::test_sddmm[int32-add-e-u-shp4-g1] PASSED          [ 59%]
tests/go/test_pipeline.py::test_linkpred_node_model[gat] PASSED   [ 43%]
tests/compute/test_sparse.py::test_sddmm[int32-add-e-v-shp0-g0] PASSED   [ 43%]
tests/compute/test_sparse.py::test_sddmm[int32-add-e-v-shp0-g1] Server (0) shutdown.
Server is exiting...
PASSED       [ 74%]
tests/distributed/test_rpc.py::test_serialize PASSED                     [ 76%]
tests/distributed/test_rpc.py::test_rpc_msg PASSED                       [ 78%]
tests/distributed/test_rpc.py::test_rpc[tensorpipe] PASSED   [ 43%]
tests/compute/test_sparse.py::test_sddmm[int32-add-e-v-shp1-g0] PASSED   [ 43%]
tests/compute/test_sparse.py::test_sddmm[int32-add-e-v-shp1-g1] PASSED   [ 43%]
tests/compute/test_sparse.py::test_sddmm[int32-add-e-v-shp2-g0] PASSED   [ 43%]
tests/compute/test_sparse.py::test_sddmm[int32-add-e-v-shp2-g1] PASSED [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_rhs-shp0-g0] PASSED   [ 43%]
tests/compute/test_sparse.py::test_sddmm[int32-add-e-v-shp3-g0] PASSED   [ 43%]
tests/compute/test_sparse.py::test_sddmm[int32-add-e-v-shp3-g1] PASSED   [ 43%]
tests/compute/test_sparse.py::test_sddmm[int32-add-e-v-shp4-g0] PASSED [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_rhs-shp0-g1] [05:39:30] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~tensorpipe is created.
[05:39:30] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~tensorpipe is created.
[05:39:30] /root/jenkins/workspace/dgl_PR-4648/src/rpc/tensorpipe/tp_communicator.cc:37: Failed to connect to receiver[tcp://172.17.0.3:16300].
Sleep 1 seconds to test client re-connect.
PASSED   [ 43%]
tests/compute/test_sparse.py::test_sddmm[int32-add-e-v-shp4-g1] PASSED   [ 43%]
tests/compute/test_sparse.py::test_sddmm[int32-add-e-e-shp0-g0] PASSED   [ 43%]
tests/compute/test_sparse.py::test_sddmm[int32-add-e-e-shp0-g1] PASSED   [ 43%]
tests/compute/test_sparse.py::test_sddmm[int32-add-e-e-shp1-g0] PASSED   [ 43%]
tests/compute/test_sparse.py::test_sddmm[int32-add-e-e-shp1-g1] PASSED   [ 43%]
tests/compute/test_sparse.py::test_sddmm[int32-add-e-e-shp2-g0] PASSED   [ 43%]
tests/compute/test_sparse.py::test_sddmm[int32-add-e-e-shp2-g1] PASSED   [ 43%]
tests/compute/test_sparse.py::test_sddmm[int32-add-e-e-shp3-g0] PASSED   [ 43%]
tests/compute/test_sparse.py::test_sddmm[int32-add-e-e-shp3-g1] PASSED   [ 43%]
tests/compute/test_sparse.py::test_sddmm[int32-add-e-e-shp4-g0] PASSED   [ 43%]
tests/compute/test_sparse.py::test_sddmm[int32-add-e-e-shp4-g1] PASSED   [ 44%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-u-u-shp0-g0] PASSED   [ 44%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-u-u-shp0-g1] PASSED   [ 44%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-u-u-shp1-g0] PASSED   [ 44%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-u-u-shp1-g1] PASSED   [ 44%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-u-u-shp2-g0] PASSED   [ 44%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-u-u-shp2-g1] PASSED   [ 44%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-u-u-shp3-g0] PASSED   [ 44%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-u-u-shp3-g1] PASSED   [ 44%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-u-u-shp4-g0] PASSED   [ 44%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-u-u-shp4-g1] PASSED   [ 44%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-u-v-shp0-g0] PASSED   [ 44%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-u-v-shp0-g1] PASSED [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_rhs-shp1-g0] PASSED   [ 44%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-u-v-shp1-g0] Start server 0
PASSED   [ 44%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-u-v-shp1-g1] PASSED [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_rhs-shp1-g1] [05:39:31] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~tensorpipe is created.
[05:39:31] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~tensorpipe is created.
Server is waiting for connections on [172.17.0.3:16300]...
[05:39:31] /root/jenkins/workspace/dgl_PR-4648/src/rpc/tensorpipe/tp_communicator.cc:108: TPReceiver starts to wait on [tcp://172.17.0.3:16300].
PASSED   [ 44%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-u-v-shp2-g0] PASSED   [ 44%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-u-v-shp2-g1] PASSED [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_rhs-shp2-g0] PASSED   [ 44%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-u-v-shp3-g0] PASSED   [ 44%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-u-v-shp3-g1] PASSED [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_rhs-shp2-g1] PASSED   [ 44%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-u-v-shp4-g0] PASSED   [ 44%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-u-v-shp4-g1] PASSED   [ 44%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-u-e-shp0-g0] PASSED   [ 44%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-u-e-shp0-g1] PASSED   [ 44%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-u-e-shp1-g0] PASSED   [ 44%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-u-e-shp1-g1] [05:39:33] /root/jenkins/workspace/dgl_PR-4648/src/rpc/tensorpipe/tp_communicator.cc:108: TPReceiver starts to wait on [tcp://172.17.0.3:48855].
Client [3301] waits on 172.17.0.3:48855
PASSED [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_rhs-shp3-g0] PASSED   [ 45%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-u-e-shp2-g0] PASSED   [ 45%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-u-e-shp2-g1] PASSED   [ 45%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-u-e-shp3-g0] PASSED [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_rhs-shp3-g1] PASSED   [ 45%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-u-e-shp3-g1] PASSED   [ 45%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-u-e-shp4-g0] PASSED   [ 45%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-u-e-shp4-g1] PASSED          [ 60%]
tests/go/test_pipeline.py::test_linkpred_node_model[sage] PASSED   [ 45%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-v-u-shp0-g0] PASSED [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_rhs-shp4-g0] PASSED   [ 45%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-v-u-shp0-g1] PASSED   [ 45%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-v-u-shp1-g0] PASSED [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_rhs-shp4-g1] PASSED   [ 45%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-v-u-shp1-g1] PASSED   [ 45%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-v-u-shp2-g0] PASSED   [ 45%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-v-u-shp2-g1] PASSED [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_rhs-shp5-g0] PASSED   [ 45%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-v-u-shp3-g0] PASSED   [ 45%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-v-u-shp3-g1] PASSED [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-min-copy_rhs-shp5-g1] PASSED   [ 45%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-v-u-shp4-g0] Machine (0) group (0) client (0) connect to server successfuly!
PASSED   [ 45%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-v-u-shp4-g1] PASSED [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-add-shp0-g0] Client[0] in group[0] is exiting...
PASSED   [ 45%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-v-v-shp0-g0] PASSED   [ 45%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-v-v-shp0-g1] PASSED   [ 45%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-v-v-shp1-g0] PASSED   [ 45%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-v-v-shp1-g1] PASSED   [ 45%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-v-v-shp2-g0] PASSED   [ 45%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-v-v-shp2-g1] PASSED   [ 45%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-v-v-shp3-g0] PASSED   [ 45%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-v-v-shp3-g1] PASSED   [ 45%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-v-v-shp4-g0] PASSED   [ 46%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-v-v-shp4-g1] PASSED   [ 46%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-v-e-shp0-g0] Server (0) shutdown.
Server is exiting...
PASSED   [ 46%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-v-e-shp0-g1] PASSED   [ 46%]PASSED               [ 80%]
tests/distributed/test_rpc.py::test_multi_client[socket] 
tests/compute/test_sparse.py::test_sddmm[int32-sub-v-e-shp1-g0] PASSED   [ 46%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-v-e-shp1-g1] PASSED   [ 46%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-v-e-shp2-g0] PASSED   [ 46%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-v-e-shp2-g1] PASSED  [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-add-shp0-g1] PASSED   [ 46%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-v-e-shp3-g0] PASSED   [ 46%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-v-e-shp3-g1] PASSED   [ 46%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-v-e-shp4-g0] PASSED   [ 46%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-v-e-shp4-g1] [05:39:38] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:39:38] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Sleep 1 seconds to test client re-connect.
[05:39:38] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:39:38] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
[05:39:38] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:39:38] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
PASSED   [ 46%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-e-u-shp0-g0] [05:39:38] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:39:38] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
[05:39:38] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:39:38] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
[05:39:38] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:39:38] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
[05:39:38] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:39:38] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
[05:39:38] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:39:38] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
[05:39:38] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:39:38] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
[05:39:38] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:39:38] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
[[05:39:38[] 05:39:38/root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc] 05:39:38:/root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc] :140/root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc: 140:Sender with NetType~: 140socketSender with NetType~:  is created.socketSender with NetType~
 is created.socket
 is created.
[05:39:38] [/root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc05:39:38[:] 05:39:38159/root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc] : :/root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.ccReceiver with NetType~159:socket: 159 is created.Receiver with NetType~: 
socketReceiver with NetType~ is created.socket
 is created.
[[05:39:38] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:05:39:38140] : /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.ccSender with NetType~:socket140 is created.: 
Sender with NetType~socket is created.
[05:39:38] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:[15905:39:38: ] Receiver with NetType~/root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.ccsocket: is created.159
: Receiver with NetType~socket is created.
[05:39:38] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:39:38] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
[05:39:38] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:39:38] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
[05:39:38] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:39:38] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
[05:39:38] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:39:38] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
[05:39:38] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:39:38] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
PASSED  [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-add-shp1-g0] PASSED   [ 46%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-e-u-shp0-g1] PASSED   [ 46%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-e-u-shp1-g0] PASSED  [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-add-shp1-g1] PASSED   [ 46%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-e-u-shp1-g1] PASSED   [ 46%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-e-u-shp2-g0] PASSED   [ 46%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-e-u-shp2-g1] PASSED   [ 46%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-e-u-shp3-g0] PASSED         [ 61%]
tests/go/test_pipeline.py::test_linkpred_node_model[sgc] PASSED  [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-add-shp2-g0] PASSED   [ 46%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-e-u-shp3-g1] PASSED   [ 46%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-e-u-shp4-g0] Start server 0
[05:39:39] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:39:39] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:12279]...
PASSED   [ 46%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-e-u-shp4-g1] PASSED   [ 46%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-e-v-shp0-g0] PASSED   [ 46%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-e-v-shp0-g1] PASSED   [ 46%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-e-v-shp1-g0] PASSED  [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-add-shp2-g1] PASSED   [ 47%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-e-v-shp1-g1] PASSED   [ 47%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-e-v-shp2-g0] PASSED   [ 47%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-e-v-shp2-g1] PASSED  [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-add-shp3-g0] PASSED   [ 47%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-e-v-shp3-g0] PASSED   [ 47%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-e-v-shp3-g1] PASSED   [ 47%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-e-v-shp4-g0] PASSED   [ 47%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-e-v-shp4-g1] PASSED  [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-add-shp3-g1] PASSED   [ 47%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-e-e-shp0-g0] PASSED   [ 47%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-e-e-shp0-g1] PASSED   [ 47%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-e-e-shp1-g0] PASSED   [ 47%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-e-e-shp1-g1] PASSED   [ 47%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-e-e-shp2-g0] PASSED   [ 47%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-e-e-shp2-g1] PASSED   [ 47%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-e-e-shp3-g0] PASSED   [ 47%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-e-e-shp3-g1] PASSED   [ 47%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-e-e-shp4-g0] PASSED   [ 47%]
tests/compute/test_sparse.py::test_sddmm[int32-sub-e-e-shp4-g1] PASSED   [ 47%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-u-u-shp0-g0] PASSED   [ 47%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-u-u-shp0-g1] PASSED   [ 47%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-u-u-shp1-g0] PASSED   [ 47%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-u-u-shp1-g1] PASSED   [ 47%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-u-u-shp2-g0] PASSED   [ 47%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-u-u-shp2-g1] PASSED   [ 47%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-u-u-shp3-g0] PASSED   [ 47%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-u-u-shp3-g1] PASSED   [ 48%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-u-u-shp4-g0] PASSED   [ 48%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-u-u-shp4-g1] PASSED   [ 48%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-u-v-shp0-g0] PASSED   [ 48%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-u-v-shp0-g1] PASSED  [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-add-shp4-g0] PASSED   [ 48%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-u-v-shp1-g0] PASSED   [ 48%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-u-v-shp1-g1] PASSED  [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-add-shp4-g1] PASSED   [ 48%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-u-v-shp2-g0] PASSED   [ 48%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-u-v-shp2-g1] PASSED  [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-add-shp5-g0] PASSED   [ 48%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-u-v-shp3-g0] PASSED   [ 48%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-u-v-shp3-g1] PASSED  [ 37%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-add-shp5-g1] PASSED   [ 48%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-u-v-shp4-g0] PASSED   [ 48%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-u-v-shp4-g1] PASSED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-sub-shp0-g0] PASSED   [ 48%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-u-e-shp0-g0] PASSED   [ 48%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-u-e-shp0-g1] PASSED          [ 62%]
tests/go/test_pipeline.py::test_linkpred_node_model[gin] PASSED   [ 48%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-u-e-shp1-g0] PASSED   [ 48%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-u-e-shp1-g1] Client [3324] waits on 172.17.0.3:60811
Client [3322] waits on 172.17.0.3:60539Client [3350] waits on 172.17.0.3:60133

Client [3339] waits on 172.17.0.3:59875
Client [3321] waits on 172.17.0.3:42499
Client [3340] waits on 172.17.0.3:43691
Client [3331] waits on 172.17.0.3:41001Client [3320] waits on 172.17.0.3:38685

Client [3327] waits on 172.17.0.3:34399
Client [3345] waits on 172.17.0.3:33877
Client [3333] waits on 172.17.0.3:44893
Client [3326] waits on 172.17.0.3:45687
Client [3319] waits on 172.17.0.3:48661
Client [3325] waits on 172.17.0.3:47561
Client [3329] waits on 172.17.0.3:53529
Client [3349] waits on 172.17.0.3:48929
Client [3328] waits on 172.17.0.3:53689
Client [3346] waits on 172.17.0.3:56611
Client [3323] waits on 172.17.0.3:57303
Client [3330] waits on 172.17.0.3:58351
Machine (0) group (0) client (3) connect to server successfuly!
Machine (0) group (0) client (5) connect to server successfuly!
Machine (0) group (0) client (6) connect to server successfuly!
Machine (0) group (0) client (7) connect to server successfuly!
Machine (0) group (0) client (2) connect to server successfuly!Machine (0) group (0) client (11) connect to server successfuly!

Machine (0) group (0) client (4) connect to server successfuly!
Machine (0) group (0) client (1) connect to server successfuly!
Machine (0) group (0) client (15) connect to server successfuly!
Machine (0) group (0) client (14) connect to server successfuly!
Machine (0) group (0) client (17) connect to server successfuly!
Machine (0) group (0) client (12) connect to server successfuly!Machine (0) group (0) client (8) connect to server successfuly!

Machine (0) group (0) client (16) connect to server successfuly!
Machine (0) group (0) client (9) connect to server successfuly!
Machine (0) group (0) client (0) connect to server successfuly!
Machine (0) group (0) client (19) connect to server successfuly!
Machine (0) group (0) client (13) connect to server successfuly!
Machine (0) group (0) client (18) connect to server successfuly!
Machine (0) group (0) client (10) connect to server successfuly!
PASSED   [ 48%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-u-e-shp2-g0] PASSED   [ 48%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-u-e-shp2-g1] PASSED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-sub-shp0-g1] Client[11] in group[0] is exiting...
Client[4] in group[0] is exiting...
PASSED   [ 48%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-u-e-shp3-g0] PASSED   [ 48%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-u-e-shp3-g1] Client[16] in group[0] is exiting...
Client[8] in group[0] is exiting...
Client[12] in group[0] is exiting...
Client[18] in group[0] is exiting...
Client[14] in group[0] is exiting...
Client[3] in group[0] is exiting...
Client[1] in group[0] is exiting...
Client[13] in group[0] is exiting...
Client[19] in group[0] is exiting...
Client[9] in group[0] is exiting...
Client[10] in group[0] is exiting...
Client[17] in group[0] is exiting...
Client[7] in group[0] is exiting...
Client[6] in group[0] is exiting...
Client[5] in group[0] is exiting...
Client[0] in group[0] is exiting...
Client[2] in group[0] is exiting...
Client[15] in group[0] is exiting...
PASSED   [ 48%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-u-e-shp4-g0] PASSED   [ 48%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-u-e-shp4-g1] PASSED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-sub-shp1-g0] PASSED   [ 48%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-v-u-shp0-g0] PASSED   [ 48%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-v-u-shp0-g1] PASSED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-sub-shp1-g1] PASSED   [ 49%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-v-u-shp1-g0] PASSED   [ 49%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-v-u-shp1-g1] Server (0) shutdown.
Server is exiting...
PASSED          [ 81%]
tests/distributed/test_rpc.py::test_multi_client[tensorpipe] PASSED   [ 49%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-v-u-shp2-g0] PASSED   [ 49%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-v-u-shp2-g1] PASSED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-sub-shp2-g0] PASSED   [ 49%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-v-u-shp3-g0] PASSED   [ 49%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-v-u-shp3-g1] PASSED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-sub-shp2-g1] PASSED   [ 49%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-v-u-shp4-g0] PASSED   [ 49%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-v-u-shp4-g1] [05:39:48] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~tensorpipe is created.
[05:39:48] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~tensorpipe is created.
[05:39:48] /root/jenkins/workspace/dgl_PR-4648/src/rpc/tensorpipe/tp_communicator.cc:37: Failed to connect to receiver[tcp://172.17.0.3:11501].
[05:39:48] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~tensorpipe is created.
[05:39:48] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~tensorpipe is created.
[05:39:48] /root/jenkins/workspace/dgl_PR-4648/src/rpc/tensorpipe/tp_communicator.cc:37: Failed to connect to receiver[tcp://172.17.0.3:11501].
[05:39:48] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~tensorpipe is created.
[[05:39:4805:39:48] ] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc/root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc::140159: : Sender with NetType~Receiver with NetType~tensorpipetensorpipe is created. is created.

[05:39:48] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~tensorpipe is created.
[05:39:48] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~tensorpipe is created.
[05:39:48] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~tensorpipe is created.
[05:39:48] /root/jenkins/workspace/dgl_PR-4648/src/rpc/tensorpipe/tp_communicator.cc:37: Failed to connect to receiver[tcp://172.17.0.3:11501].
[05:39:48] /root/jenkins/workspace/dgl_PR-4648/src/rpc/tensorpipe/tp_communicator.cc:37: Failed to connect to receiver[tcp://172.17.0.3:11501].
[05:39:48] /root/jenkins/workspace/dgl_PR-4648/src/rpc/tensorpipe/tp_communicator.cc:37: Failed to connect to receiver[tcp://172.17.0.3:11501].
[05:39:48] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~tensorpipe is created.
[05:39:48] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~tensorpipe is created.
[05:39:48] /root/jenkins/workspace/dgl_PR-4648/src/rpc/tensorpipe/tp_communicator.cc:37: Failed to connect to receiver[tcp://172.17.0.3:11501].
[05:39:48] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~tensorpipe is created.
[05:39:48] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~tensorpipe is created.
[05:39:48] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~tensorpipe is created.
[05:39:48] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~[tensorpipe is created.
05:39:48] /root/jenkins/workspace/dgl_PR-4648/src/rpc/tensorpipe/tp_communicator.cc:37: Failed to connect to receiver[tcp://172.17.0.3:11501].
[05:39:48] /root/jenkins/workspace/dgl_PR-4648/src/rpc/tensorpipe/tp_communicator.cc:37: Failed to connect to receiver[tcp://172.17.0.3:11501].
Sleep 1 seconds to test client re-connect.
[05:39:48] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~tensorpipe is created.
[05:39:48] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~tensorpipe is created.
[05:39:48] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~tensorpipe is created.
[05:39:48] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~tensorpipe is created.
[05:39:48] /root/jenkins/workspace/dgl_PR-4648/src/rpc/tensorpipe/tp_communicator.cc:37: Failed to connect to receiver[tcp://172.17.0.3:11501].
[05:39:48] /root/jenkins/workspace/dgl_PR-4648/src/rpc/tensorpipe/tp_communicator.cc:37: Failed to connect to receiver[tcp://172.17.0.3:11501].
[05:39:48] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~tensorpipe is created.
[05:39:48] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~tensorpipe is created.
[05:39:48] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~tensorpipe is created.
[05:39:48] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~tensorpipe is created.
[05:39:48] /root/jenkins/workspace/dgl_PR-4648/src/rpc/tensorpipe/tp_communicator.cc:37: Failed to connect to receiver[tcp://172.17.0.3:11501].
[05:39:48] /root/jenkins/workspace/dgl_PR-4648/src/rpc/tensorpipe/tp_communicator.cc:37: Failed to connect to receiver[tcp://172.17.0.3:11501].
[05:39:48] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~tensorpipe is created.
[05:39:48] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~tensorpipe is created.
[05:39:48] /root/jenkins/workspace/dgl_PR-4648/src/rpc/tensorpipe/tp_communicator.cc:37: Failed to connect to receiver[tcp://172.17.0.3:11501].
[05:39:48] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~tensorpipe is created.
[05:39:48] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~tensorpipe is created.
[05:39:48] /root/jenkins/workspace/dgl_PR-4648/src/rpc/tensorpipe/tp_communicator.cc:37: Failed to connect to receiver[tcp://172.17.0.3:11501].
[05:39:48] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~tensorpipe is created.
[05:39:48] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~tensorpipe is created.
[05:39:48] /root/jenkins/workspace/dgl_PR-4648/src/rpc/tensorpipe/tp_communicator.cc:37: Failed to connect to receiver[tcp://172.17.0.3:11501].
[05:39:48] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~tensorpipe is created.
[05:39:48] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~tensorpipe is created.
[05:39:48] /root/jenkins/workspace/dgl_PR-4648/src/rpc/tensorpipe/tp_communicator.cc:37: Failed to connect to receiver[tcp://172.17.0.3:11501].
[05:39:48] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~tensorpipe is created.
[05:39:48] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~tensorpipe is created.
[05:39:48] /root/jenkins/workspace/dgl_PR-4648/src/rpc/tensorpipe/tp_communicator.cc:37: Failed to connect to receiver[tcp://172.17.0.3:11501].
[05:39:48] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~tensorpipe is created.
[05:39:48] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~tensorpipe is created.
[05:39:48] /root/jenkins/workspace/dgl_PR-4648/src/rpc/tensorpipe/tp_communicator.cc:37: Failed to connect to receiver[tcp://172.17.0.3:11501].
[05:39:48] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~tensorpipe is created.
[05:39:48] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~tensorpipe is created.
[05:39:48] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~tensorpipe is created.
[05:39:48] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~tensorpipe is created.
[[05:39:48] /root/jenkins/workspace/dgl_PR-4648/src/rpc/tensorpipe/tp_communicator.cc:05:39:4837] : /root/jenkins/workspace/dgl_PR-4648/src/rpc/tensorpipe/tp_communicator.ccFailed to connect to receiver[:tcp://172.17.0.3:11501].37
: Failed to connect to receiver[tcp://172.17.0.3:11501].
PASSED   [ 49%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-v-v-shp0-g0] PASSED   [ 49%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-v-v-shp0-g1] PASSED   [ 49%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-v-v-shp1-g0] PASSED   [ 49%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-v-v-shp1-g1] PASSED   [ 49%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-v-v-shp2-g0] PASSED   [ 49%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-v-v-shp2-g1] PASSED   [ 49%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-v-v-shp3-g0] PASSED   [ 49%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-v-v-shp3-g1] PASSED   [ 49%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-v-v-shp4-g0] PASSED   [ 49%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-v-v-shp4-g1] PASSED   [ 49%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-v-e-shp0-g0] PASSED   [ 49%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-v-e-shp0-g1] PASSED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-sub-shp3-g0] PASSED   [ 49%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-v-e-shp1-g0] PASSED   [ 49%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-v-e-shp1-g1] PASSED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-sub-shp3-g1] Start server 0
[05:39:49] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~tensorpipe is created.
[05:39:49] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~tensorpipe is created.
Server is waiting for connections on [172.17.0.3:11501]...
[05:39:49] /root/jenkins/workspace/dgl_PR-4648/src/rpc/tensorpipe/tp_communicator.cc:108: TPReceiver starts to wait on [tcp://172.17.0.3:11501].
PASSED   [ 49%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-v-e-shp2-g0] PASSED   [ 49%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-v-e-shp2-g1] PASSED   [ 49%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-v-e-shp3-g0] PASSED   [ 50%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-v-e-shp3-g1] PASSED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-sub-shp4-g0] PASSED   [ 50%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-v-e-shp4-g0] PASSED          [ 64%]
tests/go/test_pipeline.py::test_linkpred_edge_model[ele] PASSED   [ 50%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-v-e-shp4-g1] PASSED   [ 50%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-e-u-shp0-g0] PASSED   [ 50%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-e-u-shp0-g1] PASSED   [ 50%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-e-u-shp1-g0] PASSED   [ 50%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-e-u-shp1-g1] PASSED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-sub-shp4-g1] PASSED   [ 50%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-e-u-shp2-g0] PASSED   [ 50%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-e-u-shp2-g1] PASSED   [ 50%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-e-u-shp3-g0] PASSED   [ 50%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-e-u-shp3-g1] [05:39:51] /root/jenkins/workspace/dgl_PR-4648/src/rpc/tensorpipe/tp_communicator.cc:108: TPReceiver starts to wait on [tcp://172.17.0.3:33563].
Client [3468] waits on 172.17.0.3:33563
[05:39:51] /root/jenkins/workspace/dgl_PR-4648/src/rpc/tensorpipe/tp_communicator.cc:108: TPReceiver starts to wait on [tcp://172.17.0.3:48785].
Client [3469] waits on 172.17.0.3:48785
[05:39:51] /root/jenkins/workspace/dgl_PR-4648/src/rpc/tensorpipe/tp_communicator.cc:108: TPReceiver starts to wait on [tcp://172.17.0.3:55543].
[05:39:51] /root/jenkins/workspace/dgl_PR-4648/src/rpc/tensorpipe/tp_communicator.cc:108: TPReceiver starts to wait on [tcp://172.17.0.3:51795].
[05:39:51] /root/jenkins/workspace/dgl_PR-4648/src/rpc/tensorpipe/tp_communicator.cc:108: TPReceiver starts to wait on [tcp://172.17.0.3:51813].
Client [3476] waits on 172.17.0.3:55543
Client [3483] waits on 172.17.0.3:51795
Client [3471] waits on 172.17.0.3:51813
[05:39:51] /root/jenkins/workspace/dgl_PR-4648/src/rpc/tensorpipe/tp_communicator.cc:108: TPReceiver starts to wait on [tcp://172.17.0.3:36643].
Client [3473] waits on 172.17.0.3:36643
[05:39:51] /root/jenkins/workspace/dgl_PR-4648/src/rpc/tensorpipe/tp_communicator.cc:108: TPReceiver starts to wait on [tcp://172.17.0.3:35533].
[05:39:51] /root/jenkins/workspace/dgl_PR-4648/src/rpc/tensorpipe/tp_communicator.cc:108: TPReceiver starts to wait on [tcp://172.17.0.3:55301].
Client [3475] waits on 172.17.0.3:35533
Client [3477] waits on 172.17.0.3:55301
[05:39:51] /root/jenkins/workspace/dgl_PR-4648/src/rpc/tensorpipe/tp_communicator.cc:108: TPReceiver starts to wait on [tcp://172.17.0.3:47747].
[05:39:51] /root/jenkins/workspace/dgl_PR-4648/src/rpc/tensorpipe/tp_communicator.cc:108: TPReceiver starts to wait on [tcp://172.17.0.3:53047].
Client [3496] waits on 172.17.0.3:47747
Client [3472] waits on 172.17.0.3:53047
[05:39:51] /root/jenkins/workspace/dgl_PR-4648/src/rpc/tensorpipe/tp_communicator.cc:108: TPReceiver starts to wait on [tcp://172.17.0.3:41577].
Client [3478] waits on 172.17.0.3:41577
[05:39:51] /root/jenkins/workspace/dgl_PR-4648/src/rpc/tensorpipe/tp_communicator.cc:108: TPReceiver starts to wait on [tcp://172.17.0.3:49799].
Client [3474] waits on 172.17.0.3:49799
[05:39:51] /root/jenkins/workspace/dgl_PR-4648/src/rpc/tensorpipe/tp_communicator.cc:108: TPReceiver starts to wait on [tcp://172.17.0.3:34857].
Client [3470] waits on 172.17.0.3:34857
[05:39:51] /root/jenkins/workspace/dgl_PR-4648/src/rpc/tensorpipe/tp_communicator.cc:108: TPReceiver starts to wait on [tcp://172.17.0.3:41643].
[05:39:51] /root/jenkins/workspace/dgl_PR-4648/src/rpc/tensorpipe/tp_communicator.cc:108: TPReceiver starts to wait on [tcp://172.17.0.3:58793].
Client [3494] waits on 172.17.0.3:41643
Client [3499] waits on 172.17.0.3:58793
[05:39:51] /root/jenkins/workspace/dgl_PR-4648/src/rpc/tensorpipe/tp_communicator.cc:108: TPReceiver starts to wait on [tcp://172.17.0.3:50777].
Client [3502] waits on 172.17.0.3:50777
[05:39:51] /root/jenkins/workspace/dgl_PR-4648/src/rpc/tensorpipe/tp_communicator.cc:108: TPReceiver starts to wait on [tcp://172.17.0.3:55745].
Client [3497] waits on 172.17.0.3:55745
[05:39:51] /root/jenkins/workspace/dgl_PR-4648/src/rpc/tensorpipe/tp_communicator.cc:108: TPReceiver starts to wait on [tcp://172.17.0.3:50827].
Client [3479] waits on 172.17.0.3:50827
PASSED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-sub-shp5-g0] PASSED   [ 50%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-e-u-shp4-g0] PASSED   [ 50%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-e-u-shp4-g1] [[05:39:51] 05:39:51/root/jenkins/workspace/dgl_PR-4648/src/rpc/tensorpipe/tp_communicator.cc] :/root/jenkins/workspace/dgl_PR-4648/src/rpc/tensorpipe/tp_communicator.cc108:: 108TPReceiver starts to wait on [: tcp://172.17.0.3:53589TPReceiver starts to wait on [].tcp://172.17.0.3:59807
].
Client [3481] waits on 172.17.0.3:53589Client [3489] waits on 172.17.0.3:59807

PASSED   [ 50%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-e-v-shp0-g0] PASSED   [ 50%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-e-v-shp0-g1] PASSED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-sub-shp5-g1] PASSED   [ 50%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-e-v-shp1-g0] PASSED   [ 50%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-e-v-shp1-g1] PASSED   [ 50%]PASSED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-mul-shp0-g0] 
tests/compute/test_sparse.py::test_sddmm[int32-mul-e-v-shp2-g0] PASSED   [ 50%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-e-v-shp2-g1] PASSED   [ 50%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-e-v-shp3-g0] PASSED   [ 50%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-e-v-shp3-g1] PASSED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-mul-shp0-g1] PASSED   [ 50%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-e-v-shp4-g0] PASSED   [ 50%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-e-v-shp4-g1] PASSED   [ 50%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-e-e-shp0-g0] PASSED   [ 50%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-e-e-shp0-g1] PASSED   [ 51%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-e-e-shp1-g0] PASSED   [ 51%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-e-e-shp1-g1] PASSED   [ 51%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-e-e-shp2-g0] PASSED   [ 51%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-e-e-shp2-g1] PASSED   [ 51%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-e-e-shp3-g0] PASSED   [ 51%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-e-e-shp3-g1] PASSED   [ 51%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-e-e-shp4-g0] PASSED   [ 51%]
tests/compute/test_sparse.py::test_sddmm[int32-mul-e-e-shp4-g1] PASSED   [ 51%]
tests/compute/test_sparse.py::test_sddmm[int32-div-u-u-shp0-g0] PASSED   [ 51%]
tests/compute/test_sparse.py::test_sddmm[int32-div-u-u-shp0-g1] PASSED   [ 51%]
tests/compute/test_sparse.py::test_sddmm[int32-div-u-u-shp1-g0] PASSED   [ 51%]
tests/compute/test_sparse.py::test_sddmm[int32-div-u-u-shp1-g1] PASSED   [ 51%]
tests/compute/test_sparse.py::test_sddmm[int32-div-u-u-shp2-g0] PASSED   [ 51%]
tests/compute/test_sparse.py::test_sddmm[int32-div-u-u-shp2-g1] PASSED   [ 51%]
tests/compute/test_sparse.py::test_sddmm[int32-div-u-u-shp3-g0] PASSED   [ 51%]
tests/compute/test_sparse.py::test_sddmm[int32-div-u-u-shp3-g1] PASSED   [ 51%]
tests/compute/test_sparse.py::test_sddmm[int32-div-u-u-shp4-g0] PASSED   [ 51%]
tests/compute/test_sparse.py::test_sddmm[int32-div-u-u-shp4-g1] PASSED   [ 51%]
tests/compute/test_sparse.py::test_sddmm[int32-div-u-v-shp0-g0] PASSED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-mul-shp1-g0] PASSED   [ 51%]
tests/compute/test_sparse.py::test_sddmm[int32-div-u-v-shp0-g1] PASSED   [ 51%]
tests/compute/test_sparse.py::test_sddmm[int32-div-u-v-shp1-g0] PASSED   [ 51%]
tests/compute/test_sparse.py::test_sddmm[int32-div-u-v-shp1-g1] PASSED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-mul-shp1-g1] PASSED   [ 51%]
tests/compute/test_sparse.py::test_sddmm[int32-div-u-v-shp2-g0] PASSED   [ 51%]
tests/compute/test_sparse.py::test_sddmm[int32-div-u-v-shp2-g1] PASSED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-mul-shp2-g0] Machine (0) group (0) client (1) connect to server successfuly!
Machine (0) group (0) client (0) connect to server successfuly!
Machine (0) group (0) client (8) connect to server successfuly!
Machine (0) group (0) client (7) connect to server successfuly!
Machine (0) group (0) client (4) connect to server successfuly!
Machine (0) group (0) client (2) connect to server successfuly!Machine (0) group (0) client (5) connect to server successfuly!Machine (0) group (0) client (3) connect to server successfuly!

Machine (0) group (0) client (10) connect to server successfuly!
Machine (0) group (0) client (11) connect to server successfuly!Machine (0) group (0) client (14) connect to server successfuly!
Machine (0) group (0) client (9) connect to server successfuly!


Machine (0) group (0) client (6) connect to server successfuly!
Machine (0) group (0) client (15) connect to server successfuly!Machine (0) group (0) client (19) connect to server successfuly!Machine (0) group (0) client (13) connect to server successfuly!


Machine (0) group (0) client (17) connect to server successfuly!
Machine (0) group (0) client (16) connect to server successfuly!
Machine (0) group (0) client (12) connect to server successfuly!
Machine (0) group (0) client (18) connect to server successfuly!
PASSED   [ 52%]
tests/compute/test_sparse.py::test_sddmm[int32-div-u-v-shp3-g0] PASSED   [ 52%]
tests/compute/test_sparse.py::test_sddmm[int32-div-u-v-shp3-g1] PASSED          [ 65%]
tests/go/test_pipeline.py::test_linkpred_edge_model[bilinear] PASSED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-mul-shp2-g1] PASSED   [ 52%]
tests/compute/test_sparse.py::test_sddmm[int32-div-u-v-shp4-g0] PASSED   [ 52%]
tests/compute/test_sparse.py::test_sddmm[int32-div-u-v-shp4-g1] PASSED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-mul-shp3-g0] PASSED   [ 52%]
tests/compute/test_sparse.py::test_sddmm[int32-div-u-e-shp0-g0] PASSED   [ 52%]
tests/compute/test_sparse.py::test_sddmm[int32-div-u-e-shp0-g1] PASSED   [ 52%]
tests/compute/test_sparse.py::test_sddmm[int32-div-u-e-shp1-g0] PASSED   [ 52%]
tests/compute/test_sparse.py::test_sddmm[int32-div-u-e-shp1-g1] PASSED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-mul-shp3-g1] PASSED   [ 52%]
tests/compute/test_sparse.py::test_sddmm[int32-div-u-e-shp2-g0] PASSED   [ 52%]
tests/compute/test_sparse.py::test_sddmm[int32-div-u-e-shp2-g1] PASSED   [ 52%]
tests/compute/test_sparse.py::test_sddmm[int32-div-u-e-shp3-g0] PASSED   [ 52%]
tests/compute/test_sparse.py::test_sddmm[int32-div-u-e-shp3-g1] PASSED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-mul-shp4-g0] PASSED   [ 52%]
tests/compute/test_sparse.py::test_sddmm[int32-div-u-e-shp4-g0] PASSED   [ 52%]
tests/compute/test_sparse.py::test_sddmm[int32-div-u-e-shp4-g1] PASSED   [ 52%]
tests/compute/test_sparse.py::test_sddmm[int32-div-v-u-shp0-g0] PASSED   [ 52%]
tests/compute/test_sparse.py::test_sddmm[int32-div-v-u-shp0-g1] PASSED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-mul-shp4-g1] PASSED   [ 52%]
tests/compute/test_sparse.py::test_sddmm[int32-div-v-u-shp1-g0] PASSED   [ 52%]
tests/compute/test_sparse.py::test_sddmm[int32-div-v-u-shp1-g1] PASSED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-mul-shp5-g0] PASSED   [ 52%]
tests/compute/test_sparse.py::test_sddmm[int32-div-v-u-shp2-g0] PASSED   [ 52%]
tests/compute/test_sparse.py::test_sddmm[int32-div-v-u-shp2-g1] PASSED  [ 38%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-mul-shp5-g1] PASSED   [ 52%]
tests/compute/test_sparse.py::test_sddmm[int32-div-v-u-shp3-g0] PASSED   [ 52%]
tests/compute/test_sparse.py::test_sddmm[int32-div-v-u-shp3-g1] PASSED  [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-div-shp0-g0] Client[8] in group[0] is exiting...
Client[7] in group[0] is exiting...
PASSED   [ 52%]
tests/compute/test_sparse.py::test_sddmm[int32-div-v-u-shp4-g0] PASSED   [ 52%]
tests/compute/test_sparse.py::test_sddmm[int32-div-v-u-shp4-g1] PASSED  [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-div-shp0-g1] Client[1] in group[0] is exiting...
Client[0] in group[0] is exiting...
Client[4] in group[0] is exiting...
PASSED   [ 52%]
tests/compute/test_sparse.py::test_sddmm[int32-div-v-v-shp0-g0] PASSED   [ 53%]
tests/compute/test_sparse.py::test_sddmm[int32-div-v-v-shp0-g1] PASSED   [ 53%]
tests/compute/test_sparse.py::test_sddmm[int32-div-v-v-shp1-g0] PASSED   [ 53%]
tests/compute/test_sparse.py::test_sddmm[int32-div-v-v-shp1-g1] PASSED   [ 53%]
tests/compute/test_sparse.py::test_sddmm[int32-div-v-v-shp2-g0] PASSED   [ 53%]
tests/compute/test_sparse.py::test_sddmm[int32-div-v-v-shp2-g1] PASSED   [ 53%]
tests/compute/test_sparse.py::test_sddmm[int32-div-v-v-shp3-g0] PASSED   [ 53%]
tests/compute/test_sparse.py::test_sddmm[int32-div-v-v-shp3-g1] PASSED   [ 53%]
tests/compute/test_sparse.py::test_sddmm[int32-div-v-v-shp4-g0] PASSED   [ 53%]
tests/compute/test_sparse.py::test_sddmm[int32-div-v-v-shp4-g1] PASSED   [ 53%]
tests/compute/test_sparse.py::test_sddmm[int32-div-v-e-shp0-g0] PASSED     [ 66%]
tests/go/test_pipeline.py::test_linkpred_neg_sampler[global] Client[9] in group[0] is exiting...
Client[10] in group[0] is exiting...
Client[3] in group[0] is exiting...
Client[14] in group[0] is exiting...
Client[19] in group[0] is exiting...
Client[13] in group[0] is exiting...
PASSED   [ 53%]
tests/compute/test_sparse.py::test_sddmm[int32-div-v-e-shp0-g1] Client[5] in group[0] is exiting...
Client[17] in group[0] is exiting...
Client[15] in group[0] is exiting...
Client[2] in group[0] is exiting...
Client[11] in group[0] is exiting...
PASSED   [ 53%]
tests/compute/test_sparse.py::test_sddmm[int32-div-v-e-shp1-g0] PASSED   [ 53%]
tests/compute/test_sparse.py::test_sddmm[int32-div-v-e-shp1-g1] Client[6] in group[0] is exiting...
Client[12] in group[0] is exiting...
PASSED  [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-div-shp1-g0] PASSED   [ 53%]
tests/compute/test_sparse.py::test_sddmm[int32-div-v-e-shp2-g0] PASSED   [ 53%]
tests/compute/test_sparse.py::test_sddmm[int32-div-v-e-shp2-g1] Client[16] in group[0] is exiting...
Client[18] in group[0] is exiting...
Server (0) shutdown.
Server is exiting...
PASSED   [ 53%]
tests/compute/test_sparse.py::test_sddmm[int32-div-v-e-shp3-g0] PASSED      [ 83%]
tests/distributed/test_rpc.py::test_multi_thread_rpc[socket] PASSED  [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-div-shp1-g1] PASSED   [ 53%]
tests/compute/test_sparse.py::test_sddmm[int32-div-v-e-shp3-g1] PASSED   [ 53%]
tests/compute/test_sparse.py::test_sddmm[int32-div-v-e-shp4-g0] [05:40:02] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:40:02] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
PASSED   [ 53%]
tests/compute/test_sparse.py::test_sddmm[int32-div-v-e-shp4-g1] PASSED   [ 53%]
tests/compute/test_sparse.py::test_sddmm[int32-div-e-u-shp0-g0] PASSED   [ 53%]
tests/compute/test_sparse.py::test_sddmm[int32-div-e-u-shp0-g1] PASSED  [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-div-shp2-g0] PASSED   [ 53%]
tests/compute/test_sparse.py::test_sddmm[int32-div-e-u-shp1-g0] PASSED   [ 53%]
tests/compute/test_sparse.py::test_sddmm[int32-div-e-u-shp1-g1] PASSED   [ 53%]
tests/compute/test_sparse.py::test_sddmm[int32-div-e-u-shp2-g0] PASSED   [ 54%]
tests/compute/test_sparse.py::test_sddmm[int32-div-e-u-shp2-g1] Sleep 1 seconds to test client re-connect.
Sleep 1 seconds to test client re-connect.
PASSED  [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-div-shp2-g1] PASSED   [ 54%]
tests/compute/test_sparse.py::test_sddmm[int32-div-e-u-shp3-g0] PASSED   [ 54%]
tests/compute/test_sparse.py::test_sddmm[int32-div-e-u-shp3-g1] PASSED   [ 54%]
tests/compute/test_sparse.py::test_sddmm[int32-div-e-u-shp4-g0] PASSED  [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-div-shp3-g0] PASSED   [ 54%]
tests/compute/test_sparse.py::test_sddmm[int32-div-e-u-shp4-g1] PASSED   [ 54%]
tests/compute/test_sparse.py::test_sddmm[int32-div-e-v-shp0-g0] Start server 1
[05:40:04] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:40:04] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:17469]...
Start server 0
[05:40:04] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:40:04] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:17467]...
PASSED   [ 54%]
tests/compute/test_sparse.py::test_sddmm[int32-div-e-v-shp0-g1] PASSED  [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-div-shp3-g1] PASSED   [ 54%]
tests/compute/test_sparse.py::test_sddmm[int32-div-e-v-shp1-g0] PASSED   [ 54%]
tests/compute/test_sparse.py::test_sddmm[int32-div-e-v-shp1-g1] PASSED   [ 54%]
tests/compute/test_sparse.py::test_sddmm[int32-div-e-v-shp2-g0] PASSED   [ 54%]
tests/compute/test_sparse.py::test_sddmm[int32-div-e-v-shp2-g1] PASSED  [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-div-shp4-g0] PASSED   [ 54%]
tests/compute/test_sparse.py::test_sddmm[int32-div-e-v-shp3-g0] PASSED   [ 54%]
tests/compute/test_sparse.py::test_sddmm[int32-div-e-v-shp3-g1] PASSED  [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-div-shp4-g1] PASSED   [ 54%]
tests/compute/test_sparse.py::test_sddmm[int32-div-e-v-shp4-g0] PASSED   [ 54%]
tests/compute/test_sparse.py::test_sddmm[int32-div-e-v-shp4-g1] PASSED   [ 54%]
tests/compute/test_sparse.py::test_sddmm[int32-div-e-e-shp0-g0] PASSED   [ 54%]
tests/compute/test_sparse.py::test_sddmm[int32-div-e-e-shp0-g1] PASSED   [ 54%]
tests/compute/test_sparse.py::test_sddmm[int32-div-e-e-shp1-g0] PASSED   [ 54%]
tests/compute/test_sparse.py::test_sddmm[int32-div-e-e-shp1-g1] PASSED   [ 54%]
tests/compute/test_sparse.py::test_sddmm[int32-div-e-e-shp2-g0] PASSED   [ 54%]
tests/compute/test_sparse.py::test_sddmm[int32-div-e-e-shp2-g1] PASSED   [ 54%]
tests/compute/test_sparse.py::test_sddmm[int32-div-e-e-shp3-g0] PASSED   [ 54%]
tests/compute/test_sparse.py::test_sddmm[int32-div-e-e-shp3-g1] PASSED   [ 54%]
tests/compute/test_sparse.py::test_sddmm[int32-div-e-e-shp4-g0] PASSED   [ 54%]
tests/compute/test_sparse.py::test_sddmm[int32-div-e-e-shp4-g1] PASSED   [ 55%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-u-u-shp0-g0] PASSED   [ 55%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-u-u-shp0-g1] PASSED   [ 55%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-u-u-shp1-g0] PASSED   [ 55%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-u-u-shp1-g1] PASSED   [ 55%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-u-u-shp2-g0] PASSED   [ 55%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-u-u-shp2-g1] PASSED   [ 55%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-u-u-shp3-g0] PASSED   [ 55%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-u-u-shp3-g1] PASSED   [ 55%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-u-u-shp4-g0] PASSED   [ 55%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-u-u-shp4-g1] PASSED   [ 55%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-u-v-shp0-g0] PASSED   [ 55%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-u-v-shp0-g1] PASSED  [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-div-shp5-g0] PASSED      [ 67%]
tests/go/test_pipeline.py::test_linkpred_neg_sampler[persource] PASSED   [ 55%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-u-v-shp1-g0] PASSED   [ 55%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-u-v-shp1-g1] PASSED   [ 55%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-u-v-shp2-g0] PASSED  [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-div-shp5-g1] PASSED   [ 55%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-u-v-shp2-g1] PASSED   [ 55%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-u-v-shp3-g0] PASSED  [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_lhs-shp0-g0] PASSED   [ 55%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-u-v-shp3-g1] PASSED [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_lhs-shp0-g1] PASSED   [ 55%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-u-v-shp4-g0] PASSED   [ 55%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-u-v-shp4-g1] PASSED [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_lhs-shp1-g0] PASSED   [ 55%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-u-e-shp0-g0] Client [1090] waits on 172.17.0.3:44973
Machine (0) group (0) client (0) connect to server successfuly!
Client[0] in group[0] is exiting...
Server (0) shutdown.
Server is exiting...
Server (1) shutdown.
Server is exiting...
PASSED   [ 55%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-u-e-shp0-g1] PASSED   [ 55%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-u-e-shp1-g0] PASSED      [ 85%]
tests/distributed/test_rpc.py::test_multi_thread_rpc[tensorpipe] [05:40:08] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~tensorpipe is created.
[05:40:08] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~tensorpipe is created.
[05:40:08] /root/jenkins/workspace/dgl_PR-4648/src/rpc/tensorpipe/tp_communicator.cc:37: Failed to connect to receiver[tcp://172.17.0.3:19482].
PASSED   [ 55%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-u-e-shp1-g1] PASSED   [ 56%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-u-e-shp2-g0] PASSED   [ 56%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-u-e-shp2-g1] PASSED [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_lhs-shp1-g1] PASSED   [ 56%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-u-e-shp3-g0] PASSED   [ 56%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-u-e-shp3-g1] PASSED [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_lhs-shp2-g0] PASSED   [ 56%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-u-e-shp4-g0] PASSED   [ 56%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-u-e-shp4-g1] Sleep 1 seconds to test client re-connect.
Sleep 1 seconds to test client re-connect.
PASSED   [ 56%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-v-u-shp0-g0] PASSED [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_lhs-shp2-g1] PASSED   [ 56%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-v-u-shp0-g1] PASSED   [ 56%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-v-u-shp1-g0] PASSED [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_lhs-shp3-g0] PASSED   [ 56%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-v-u-shp1-g1] Start server 0
[05:40:10] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~tensorpipe is created.
[05:40:10] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~tensorpipe is created.
Server is waiting for connections on [172.17.0.3:19482]...
[05:40:10] /root/jenkins/workspace/dgl_PR-4648/src/rpc/tensorpipe/tp_communicator.cc:108: TPReceiver starts to wait on [tcp://172.17.0.3:19482].
Start server 1
PASSED   [ 56%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-v-u-shp2-g0] [05:40:10] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~tensorpipe is created.
[05:40:10] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~tensorpipe is created.
Server is waiting for connections on [172.17.0.3:19484]...
[05:40:10] /root/jenkins/workspace/dgl_PR-4648/src/rpc/tensorpipe/tp_communicator.cc:108: TPReceiver starts to wait on [tcp://172.17.0.3:19484].
PASSED   [ 56%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-v-u-shp2-g1] PASSED [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_lhs-shp3-g1] PASSED   [ 56%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-v-u-shp3-g0] PASSED   [ 69%]
tests/go/test_pipeline.py::test_linkpred_neg_sampler[] PASSED   [ 56%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-v-u-shp3-g1] PASSED [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_lhs-shp4-g0] [05:40:11] /root/jenkins/workspace/dgl_PR-4648/src/rpc/tensorpipe/tp_communicator.cc:108: TPReceiver starts to wait on [tcp://172.17.0.3:59207].
Client [1090] waits on 172.17.0.3:59207
PASSED   [ 56%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-v-u-shp4-g0] PASSED   [ 56%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-v-u-shp4-g1] PASSED [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_lhs-shp4-g1] PASSED   [ 56%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-v-v-shp0-g0] PASSED   [ 56%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-v-v-shp0-g1] PASSED   [ 56%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-v-v-shp1-g0] PASSED   [ 56%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-v-v-shp1-g1] PASSED   [ 56%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-v-v-shp2-g0] PASSED   [ 56%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-v-v-shp2-g1] PASSED   [ 56%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-v-v-shp3-g0] PASSED   [ 56%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-v-v-shp3-g1] PASSED   [ 56%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-v-v-shp4-g0] PASSED   [ 57%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-v-v-shp4-g1] PASSED   [ 57%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-v-e-shp0-g0] PASSED   [ 57%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-v-e-shp0-g1] PASSED   [ 57%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-v-e-shp1-g0] PASSED   [ 57%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-v-e-shp1-g1] PASSED [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_lhs-shp5-g0] PASSED   [ 57%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-v-e-shp2-g0] PASSED   [ 57%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-v-e-shp2-g1] PASSED   [ 57%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-v-e-shp3-g0] PASSED   [ 57%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-v-e-shp3-g1] PASSED   [ 57%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-v-e-shp4-g0] PASSED   [ 57%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-v-e-shp4-g1] PASSED   [ 57%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-e-u-shp0-g0] PASSED [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_lhs-shp5-g1] PASSED   [ 57%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-e-u-shp0-g1] PASSED   [ 57%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-e-u-shp1-g0] PASSED [ 39%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_rhs-shp0-g0] PASSED   [ 57%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-e-u-shp1-g1] PASSED   [ 57%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-e-u-shp2-g0] PASSED   [ 57%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-e-u-shp2-g1] PASSED [ 40%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_rhs-shp0-g1] PASSED   [ 57%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-e-u-shp3-g0] PASSED   [ 57%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-e-u-shp3-g1] PASSED   [ 57%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-e-u-shp4-g0] Machine (0) group (0) client (0) connect to server successfuly!
Client[0] in group[0] is exiting...
Server (0) shutdown.
Server is exiting...
Server (1) shutdown.
Server is exiting...
PASSED  [ 87%]
tests/distributed/test_rpc.py::test_multi_client_groups SKIPPED (Tes...) [ 89%]
tests/distributed/test_rpc.py::test_multi_client_connect[socket] [05:40:14] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:40:14] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
PASSED [ 40%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_rhs-shp1-g0] PASSED   [ 57%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-e-u-shp4-g1] PASSED   [ 57%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-e-v-shp0-g0] PASSED   [ 57%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-e-v-shp0-g1] PASSED [ 40%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_rhs-shp1-g1] PASSED   [ 57%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-e-v-shp1-g0] PASSED   [ 58%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-e-v-shp1-g1] PASSED   [ 58%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-e-v-shp2-g0] PASSED   [ 58%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-e-v-shp2-g1] PASSED [ 40%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_rhs-shp2-g0] PASSED   [ 58%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-e-v-shp3-g0] PASSED   [ 58%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-e-v-shp3-g1] PASSED   [ 58%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-e-v-shp4-g0] PASSED            [ 70%]
tests/go/test_pipeline.py::test_graphpred[gin-csv] PASSED   [ 58%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-e-v-shp4-g1] PASSED   [ 58%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-e-e-shp0-g0] PASSED   [ 58%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-e-e-shp0-g1] PASSED   [ 58%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-e-e-shp1-g0] PASSED   [ 58%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-e-e-shp1-g1] PASSED   [ 58%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-e-e-shp2-g0] PASSED   [ 58%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-e-e-shp2-g1] PASSED   [ 58%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-e-e-shp3-g0] PASSED   [ 58%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-e-e-shp3-g1] PASSED   [ 58%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-e-e-shp4-g0] PASSED   [ 58%]
tests/compute/test_sparse.py::test_sddmm[int32-dot-e-e-shp4-g1] PASSED   [ 58%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-u-u-shp0-g0] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-u-u-shp0-g1] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-u-u-shp1-g0] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-u-u-shp1-g1] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-u-u-shp2-g0] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-u-u-shp2-g1] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-u-u-shp3-g0] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-u-u-shp3-g1] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-u-u-shp4-g0] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-u-u-shp4-g1] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-u-v-shp0-g0] PASSED [ 40%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_rhs-shp2-g1] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-u-v-shp0-g1] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-u-v-shp1-g0] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-u-v-shp1-g1] PASSED [ 40%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_rhs-shp3-g0] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-u-v-shp2-g0] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-u-v-shp2-g1] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-u-v-shp3-g0] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-u-v-shp3-g1] PASSED [ 40%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_rhs-shp3-g1] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-u-v-shp4-g0] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-u-v-shp4-g1] Expected error: Failed to build conncetion with peer after 1 retries. Please check network availability or increase max try times via 'DGL_DIST_MAX_TRY_TIMES'.
PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-u-e-shp0-g0] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-u-e-shp0-g1] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-u-e-shp1-g0] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-u-e-shp1-g1] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-u-e-shp2-g0] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-u-e-shp2-g1] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-u-e-shp3-g0] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-u-e-shp3-g1] PASSED [ 40%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_rhs-shp4-g0] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-u-e-shp4-g0] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-u-e-shp4-g1] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-v-u-shp0-g0] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-v-u-shp0-g1] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-v-u-shp1-g0] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-v-u-shp1-g1] [05:40:18] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:40:18] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Sleep 1 seconds to test client re-connect.
PASSED [ 40%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_rhs-shp4-g1] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-v-u-shp2-g0] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-v-u-shp2-g1] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-v-u-shp3-g0] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-v-u-shp3-g1] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-v-u-shp4-g0] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-v-u-shp4-g1] PASSED [ 40%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_rhs-shp5-g0] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-v-v-shp0-g0] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-v-v-shp0-g1] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-v-v-shp1-g0] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-v-v-shp1-g1] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-v-v-shp2-g0] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-v-v-shp2-g1] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-v-v-shp3-g0] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-v-v-shp3-g1] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-v-v-shp4-g0] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-v-v-shp4-g1] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-v-e-shp0-g0] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-v-e-shp0-g1] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-v-e-shp1-g0] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-v-e-shp1-g1] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-v-e-shp2-g0] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-v-e-shp2-g1] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-v-e-shp3-g0] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-v-e-shp3-g1] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-v-e-shp4-g0] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-v-e-shp4-g1] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-e-u-shp0-g0] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-e-u-shp0-g1] PASSED [ 40%]
tests/compute/test_sparse.py::test_spmm[idtype1-max-copy_rhs-shp5-g1] Start server 0
[05:40:19] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~socket is created.
[05:40:19] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~socket is created.
Server is waiting for connections on [172.17.0.3:12614]...
PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-e-u-shp1-g0] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-e-u-shp1-g1] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-e-u-shp2-g0] PASSED [ 40%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-u-shp0-g0] PASSED [ 40%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-u-shp0-g1] PASSED [ 40%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-u-shp1-g0] PASSED [ 40%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-u-shp1-g1] PASSED [ 40%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-u-shp2-g0] PASSED [ 40%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-u-shp2-g1] PASSED [ 40%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-u-shp3-g0] PASSED [ 40%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-u-shp3-g1] PASSED [ 40%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-u-shp4-g0] PASSED [ 40%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-u-shp4-g1] PASSED [ 40%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-v-shp0-g0] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-e-u-shp2-g1] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-e-u-shp3-g0] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-e-u-shp3-g1] PASSED [ 40%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-v-shp0-g1] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-e-u-shp4-g0] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-e-u-shp4-g1] PASSED [ 40%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-v-shp1-g0] PASSED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-v-shp1-g1] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-e-v-shp0-g0] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-e-v-shp0-g1] PASSED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-v-shp2-g0] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-e-v-shp1-g0] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-e-v-shp1-g1] PASSED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-v-shp2-g1] PASSED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-v-shp3-g0] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-e-v-shp2-g0] PASSED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-v-shp3-g1] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-e-v-shp2-g1] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-e-v-shp3-g0] PASSED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-v-shp4-g0] PASSED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-v-shp4-g1] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-e-v-shp3-g1] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-e-v-shp4-g0] PASSED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-e-shp0-g0] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-e-v-shp4-g1] PASSED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-e-shp0-g1] PASSED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-e-shp1-g0] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-e-e-shp0-g0] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-e-e-shp0-g1] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-e-e-shp1-g0] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-e-e-shp1-g1] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-e-e-shp2-g0] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-e-e-shp2-g1] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-e-e-shp3-g0] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-e-e-shp3-g1] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-e-e-shp4-g0] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_lhs-e-e-shp4-g1] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-u-u-shp0-g0] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-u-u-shp0-g1] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-u-u-shp1-g0] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-u-u-shp1-g1] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-u-u-shp2-g0] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-u-u-shp2-g1] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-u-u-shp3-g0] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-u-u-shp3-g1] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-u-u-shp4-g0] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-u-u-shp4-g1] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-u-v-shp0-g0] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-u-v-shp0-g1] PASSED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-e-shp1-g1] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-u-v-shp1-g0] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-u-v-shp1-g1] PASSED                [ 71%]
tests/go/test_pipeline.py::test_graphpred[gin-ogbg-molhiv] PASSED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-e-shp2-g0] PASSED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-e-shp2-g1] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-u-v-shp2-g0] PASSED [ 62%]PASSED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-e-shp3-g0] 
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-u-v-shp2-g1] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-u-v-shp3-g0] PASSED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-e-shp3-g1] PASSED [ 41%]PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-u-v-shp3-g1] 
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-e-shp4-g0] PASSED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-u-e-shp4-g1] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-u-v-shp4-g0] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-u-v-shp4-g1] PASSED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-u-shp0-g0] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-u-e-shp0-g0] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-u-e-shp0-g1] PASSED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-u-shp0-g1] PASSED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-u-shp1-g0] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-u-e-shp1-g0] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-u-e-shp1-g1] PASSED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-u-shp1-g1] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-u-e-shp2-g0] PASSED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-u-shp2-g0] PASSED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-u-shp2-g1] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-u-e-shp2-g1] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-u-e-shp3-g0] PASSED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-u-shp3-g0] Client [3644] waits on 172.17.0.3:46127
Machine (0) group (0) client (0) connect to server successfuly!
PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-u-e-shp3-g1] PASSED [ 41%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-u-shp3-g1] Client[0] in group[0] is exiting...
PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-u-e-shp4-g0] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-u-e-shp4-g1] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-u-shp4-g0] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-u-shp4-g1] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-v-u-shp0-g0] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-v-u-shp0-g1] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-v-shp0-g0] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-v-shp0-g1] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-v-shp1-g0] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-v-shp1-g1] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-v-shp2-g0] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-v-shp2-g1] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-v-shp3-g0] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-v-shp3-g1] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-v-shp4-g0] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-v-shp4-g1] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-e-shp0-g0] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-v-u-shp1-g0] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-v-u-shp1-g1] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-e-shp0-g1] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-v-u-shp2-g0] Server (0) shutdown.
Server is exiting...
PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-e-shp1-g0] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-e-shp1-g1] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-v-u-shp2-g1] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-v-u-shp3-g0] PASSED  [ 90%]
tests/distributed/test_rpc.py::test_multi_client_connect[tensorpipe] [05:40:27] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~tensorpipe is created.
[05:40:27] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~tensorpipe is created.
[05:40:27] /root/jenkins/workspace/dgl_PR-4648/src/rpc/tensorpipe/tp_communicator.cc:37: Failed to connect to receiver[tcp://172.17.0.3:17628].
Expected error: Failed to build conncetion with peer[172.17.0.3:17628] after 1 retries. Please check network availability or increase max try times via 'DGL_DIST_MAX_TRY_TIMES'.
PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-e-shp2-g0] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-v-u-shp3-g1] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-e-shp2-g1] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-v-u-shp4-g0] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-v-u-shp4-g1] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-e-shp3-g0] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-e-shp3-g1] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-v-v-shp0-g0] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-v-v-shp0-g1] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-v-v-shp1-g0] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-v-v-shp1-g1] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-v-v-shp2-g0] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-v-v-shp2-g1] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-v-v-shp3-g0] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-v-v-shp3-g1] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-v-v-shp4-g0] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-v-v-shp4-g1] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-v-e-shp0-g0] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-v-e-shp0-g1] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-e-shp4-g0] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-v-e-shp1-g0] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-v-e-shp4-g1] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-v-e-shp1-g1] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-v-e-shp2-g0] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-u-shp0-g0] PASSED [ 42%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-u-shp0-g1] [05:40:28] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~tensorpipe is created.
[05:40:28] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~tensorpipe is created.
[05:40:28] /root/jenkins/workspace/dgl_PR-4648/src/rpc/tensorpipe/tp_communicator.cc:37: Failed to connect to receiver[tcp://172.17.0.3:17628].
Sleep 1 seconds to test client re-connect.
PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-v-e-shp2-g1] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-u-shp1-g0] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-v-e-shp3-g0] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-v-e-shp3-g1] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-u-shp1-g1] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-u-shp2-g0] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-v-e-shp4-g0] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-v-e-shp4-g1] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-u-shp2-g1] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-e-u-shp0-g0] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-e-u-shp0-g1] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-e-u-shp1-g0] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-e-u-shp1-g1] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-e-u-shp2-g0] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-e-u-shp2-g1] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-e-u-shp3-g0] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-e-u-shp3-g1] Start server 0
[05:40:29] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:140: Sender with NetType~tensorpipe is created.
[05:40:29] /root/jenkins/workspace/dgl_PR-4648/src/rpc/rpc.cc:159: Receiver with NetType~tensorpipe is created.
Server is waiting for connections on [172.17.0.3:17628]...
[05:40:29] /root/jenkins/workspace/dgl_PR-4648/src/rpc/tensorpipe/tp_communicator.cc:108: TPReceiver starts to wait on [tcp://172.17.0.3:17628].
PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-u-shp3-g0] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-e-u-shp4-g0] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-e-u-shp4-g1] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-e-v-shp0-g0] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-e-v-shp0-g1] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-e-v-shp1-g0] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-e-v-shp1-g1] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-e-v-shp2-g0] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-e-v-shp2-g1] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-e-v-shp3-g0] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-e-v-shp3-g1] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-e-v-shp4-g0] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-e-v-shp4-g1] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-e-e-shp0-g0] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-e-e-shp0-g1] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-e-e-shp1-g0] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-e-e-shp1-g1] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-e-e-shp2-g0] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-e-e-shp2-g1] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-e-e-shp3-g0] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-e-e-shp3-g1] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-e-e-shp4-g0] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[int32-copy_rhs-e-e-shp4-g1] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-u-shp3-g1] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-u-shp4-g0] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[int64-add-u-u-shp0-g0] PASSED   [ 66%]
tests/compute/test_sparse.py::test_sddmm[int64-add-u-u-shp0-g1] PASSED   [ 66%]
tests/compute/test_sparse.py::test_sddmm[int64-add-u-u-shp1-g0] PASSED   [ 66%]
tests/compute/test_sparse.py::test_sddmm[int64-add-u-u-shp1-g1] PASSED   [ 66%]
tests/compute/test_sparse.py::test_sddmm[int64-add-u-u-shp2-g0] PASSED   [ 66%]
tests/compute/test_sparse.py::test_sddmm[int64-add-u-u-shp2-g1] PASSED   [ 66%]
tests/compute/test_sparse.py::test_sddmm[int64-add-u-u-shp3-g0] PASSED   [ 66%]
tests/compute/test_sparse.py::test_sddmm[int64-add-u-u-shp3-g1] PASSED   [ 66%]
tests/compute/test_sparse.py::test_sddmm[int64-add-u-u-shp4-g0] PASSED   [ 66%]
tests/compute/test_sparse.py::test_sddmm[int64-add-u-u-shp4-g1] PASSED   [ 66%]
tests/compute/test_sparse.py::test_sddmm[int64-add-u-v-shp0-g0] PASSED   [ 66%]
tests/compute/test_sparse.py::test_sddmm[int64-add-u-v-shp0-g1] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-u-shp4-g1] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-v-shp0-g0] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-v-shp0-g1] PASSED   [ 66%]
tests/compute/test_sparse.py::test_sddmm[int64-add-u-v-shp1-g0] PASSED   [ 66%]
tests/compute/test_sparse.py::test_sddmm[int64-add-u-v-shp1-g1] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-v-shp1-g0] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-v-shp1-g1] PASSED   [ 66%]
tests/compute/test_sparse.py::test_sddmm[int64-add-u-v-shp2-g0] PASSED   [ 66%]
tests/compute/test_sparse.py::test_sddmm[int64-add-u-v-shp2-g1] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-v-shp2-g0] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-v-shp2-g1] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-v-shp3-g0] [05:40:31] /root/jenkins/workspace/dgl_PR-4648/src/rpc/tensorpipe/tp_communicator.cc:108: TPReceiver starts to wait on [tcp://172.17.0.3:48147].
Client [3662] waits on 172.17.0.3:48147
PASSED        [ 72%]
tests/go/test_pipeline.py::test_graphpred[gin-ogbg-molpcba] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-v-shp3-g1] PASSED   [ 66%]
tests/compute/test_sparse.py::test_sddmm[int64-add-u-v-shp3-g0] PASSED   [ 66%]
tests/compute/test_sparse.py::test_sddmm[int64-add-u-v-shp3-g1] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-v-shp4-g0] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-v-shp4-g1] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-e-shp0-g0] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-e-shp0-g1] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-e-shp1-g0] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-e-shp1-g1] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-e-shp2-g0] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-e-shp2-g1] PASSED [ 43%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-e-shp3-g0] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-e-shp3-g1] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-e-shp4-g0] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-add-e-e-shp4-g1] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-u-shp0-g0] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-u-shp0-g1] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-u-shp1-g0] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-u-shp1-g1] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-u-shp2-g0] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-u-shp2-g1] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-u-shp3-g0] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-u-shp3-g1] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-u-shp4-g0] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-u-shp4-g1] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-v-shp0-g0] PASSED   [ 66%]
tests/compute/test_sparse.py::test_sddmm[int64-add-u-v-shp4-g0] PASSED   [ 66%]
tests/compute/test_sparse.py::test_sddmm[int64-add-u-v-shp4-g1] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-v-shp0-g1] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-v-shp1-g0] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-v-shp1-g1] PASSED   [ 66%]
tests/compute/test_sparse.py::test_sddmm[int64-add-u-e-shp0-g0] PASSED   [ 66%]
tests/compute/test_sparse.py::test_sddmm[int64-add-u-e-shp0-g1] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-v-shp2-g0] PASSED   [ 66%]
tests/compute/test_sparse.py::test_sddmm[int64-add-u-e-shp1-g0] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-v-shp2-g1] PASSED   [ 66%]
tests/compute/test_sparse.py::test_sddmm[int64-add-u-e-shp1-g1] PASSED   [ 67%]
tests/compute/test_sparse.py::test_sddmm[int64-add-u-e-shp2-g0] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-v-shp3-g0] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-v-shp3-g1] PASSED   [ 67%]
tests/compute/test_sparse.py::test_sddmm[int64-add-u-e-shp2-g1] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-v-shp4-g0] PASSED   [ 67%]
tests/compute/test_sparse.py::test_sddmm[int64-add-u-e-shp3-g0] PASSED   [ 67%]
tests/compute/test_sparse.py::test_sddmm[int64-add-u-e-shp3-g1] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-v-shp4-g1] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-e-shp0-g0] PASSED   [ 67%]
tests/compute/test_sparse.py::test_sddmm[int64-add-u-e-shp4-g0] PASSED   [ 67%]
tests/compute/test_sparse.py::test_sddmm[int64-add-u-e-shp4-g1] PASSED [ 44%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-e-shp0-g1] PASSED   [ 67%]
tests/compute/test_sparse.py::test_sddmm[int64-add-v-u-shp0-g0] PASSED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-e-shp1-g0] PASSED   [ 67%]
tests/compute/test_sparse.py::test_sddmm[int64-add-v-u-shp0-g1] Machine (0) group (0) client (0) connect to server successfuly!
PASSED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-e-shp1-g1] PASSED   [ 67%]
tests/compute/test_sparse.py::test_sddmm[int64-add-v-u-shp1-g0] PASSED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-e-shp2-g0] PASSED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-e-shp2-g1] PASSED   [ 67%]
tests/compute/test_sparse.py::test_sddmm[int64-add-v-u-shp1-g1] Client[0] in group[0] is exiting...
PASSED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-e-shp3-g0] Server (0) shutdown.
Server is exiting...
PASSED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-e-shp3-g1] PASSED   [ 67%]
tests/compute/test_sparse.py::test_sddmm[int64-add-v-u-shp2-g0] PASSED   [ 67%]
tests/compute/test_sparse.py::test_sddmm[int64-add-v-u-shp2-g1] PASSED [ 92%]
tests/distributed/test_shared_mem_store.py::test_init SKIPPED (skip ...) [ 94%]
tests/distributed/test_shared_mem_store.py::test_compute SKIPPED (sk...) [ 96%]
tests/distributed/test_shared_mem_store.py::test_sync_barrier SKIPPED    [ 98%]
tests/distributed/test_shared_mem_store.py::test_copy_shared_mem SKIPPED [100%]

=================================== FAILURES ===================================
________________________________ test_partition ________________________________

    @unittest.skipIf(os.name == 'nt', reason='Do not support windows yet')
    def test_partition():
        g = create_random_graph(1000)
>       check_partition(g, 'metis', False)

tests/distributed/test_partition.py:407: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
tests/distributed/test_partition.py:247: in check_partition
    g.update_all(fn.copy_src('feats', 'msg'), fn.sum('msg', 'h'))
python/dgl/heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python/dgl/core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python/dgl/core.py:332: in invoke_gspmm
    z = op(graph, x)
python/dgl/ops/spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7fd40cd42120>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.8143, -1.1574,  0.6893,  ..., -0.5411,  0.6006,  1.0566],
        [-0.3864, -0.4807, -0.8135,  ...,  1.400...0168,  0.9815,  ...,  0.7116,  0.4885, -0.7186],
        [-0.9336, -1.3287,  0.1880,  ..., -0.3666, -0.7204,  1.1899]])
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
=============================== warnings summary ===============================
python/dgl/backend/backend.py:1741
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/backend/backend.py:1741: DeprecationWarning: invalid escape sequence \P
    """

python/dgl/backend/pytorch/tensor.py:16
python/dgl/backend/pytorch/tensor.py:16
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/backend/pytorch/tensor.py:16: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    if LooseVersion(th.__version__) < LooseVersion("1.9.0"):

python/dgl/backend/pytorch/tensor.py:340
python/dgl/backend/pytorch/tensor.py:340
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/backend/pytorch/tensor.py:340: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    if LooseVersion(th.__version__) >= LooseVersion("1.10.0"):

python/dgl/dataloading/dataloader.py:33
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/dataloading/dataloader.py:33: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    PYTORCH_VER = LooseVersion(torch.__version__)

python/dgl/_dataloading/pytorch/dataloader.py:23
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/_dataloading/pytorch/dataloader.py:23: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    PYTORCH_VER = LooseVersion(th.__version__)

python/dgl/_dataloading/pytorch/dataloader.py:24
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/_dataloading/pytorch/dataloader.py:24: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    PYTORCH_16 = PYTORCH_VER >= LooseVersion("1.6.0")

python/dgl/_dataloading/pytorch/dataloader.py:25
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/_dataloading/pytorch/dataloader.py:25: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    PYTORCH_17 = PYTORCH_VER >= LooseVersion("1.7.0")

python/dgl/heterograph.py:72
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/heterograph.py:72: DGLWarning: Recommend creating graphs by `dgl.graph(data)` instead of `dgl.DGLGraph(data)`.
    dgl_warning('Recommend creating graphs by `dgl.graph(data)`'

tests/distributed/test_dist_graph_store.py: 5 warnings
tests/distributed/test_distributed_sampling.py: 9 warnings
tests/distributed/test_mp_dataloader.py: 1 warning
tests/distributed/test_partition.py: 2 warnings
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/graph_partition_book.py:733: DGLWarning: Etype with 'str' format is deprecated. Please use '(str, str, str)'.
    "Etype with 'str' format is deprecated. Please use '(str, str, str)'.")

tests/distributed/test_dist_graph_store.py::test_split[True]
  /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/torch/_tensor.py:575: UserWarning: floor_divide is deprecated, and will be removed in a future version of pytorch. It currently rounds toward 0 (like the 'trunc' function NOT 'floor'). This results in incorrect rounding for negative values.
  To keep the current behavior, use torch.div(a, b, rounding_mode='trunc'), or for actual floor division, use torch.div(a, b, rounding_mode='floor'). (Triggered internally at  /pytorch/aten/src/ATen/native/BinaryOps.cpp:467.)
    return torch.floor_divide(self, other)

tests/distributed/test_distributed_sampling.py::test_rpc_find_edges_shuffle[1]
tests/distributed/test_distributed_sampling.py::test_rpc_find_edges_shuffle[2]
tests/distributed/test_distributed_sampling.py::test_rpc_get_degree_shuffle[1]
tests/distributed/test_distributed_sampling.py::test_rpc_get_degree_shuffle[2]
tests/distributed/test_distributed_sampling.py::test_rpc_sampling_shuffle[1]
tests/distributed/test_distributed_sampling.py::test_rpc_sampling_shuffle[2]
tests/distributed/test_distributed_sampling.py::test_rpc_in_subgraph
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/heterograph.py:6168: DGLWarning: DGLGraph.readonly is deprecated in v0.5.
  DGLGraph now always supports mutable operations like add_nodes and add_edges.
    dgl_warning('DGLGraph.readonly is deprecated in v0.5.\n'

tests/distributed/test_distributed_sampling.py::test_standalone_sampling
tests/distributed/test_distributed_sampling.py::test_standalone_etype_sampling
tests/distributed/test_mp_dataloader.py::test_dist_dataloader[1-False-True-0-3]
tests/distributed/test_mp_dataloader.py::test_dist_dataloader[1-False-True-4-3]
tests/distributed/test_mp_dataloader.py::test_dist_dataloader[1-False-False-0-3]
tests/distributed/test_mp_dataloader.py::test_dist_dataloader[1-False-False-4-3]
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/distributed/partition.py:578: DGLWarning: The argument reshuffle will be deprecated in the next release. For heterogeneous graphs, reshuffle must be enabled.
    dgl_warning("The argument reshuffle will be deprecated in the next release. "

tests/distributed/test_distributed_sampling.py::test_standalone_sampling
tests/distributed/test_distributed_sampling.py::test_standalone_etype_sampling
tests/distributed/test_mp_dataloader.py::test_dist_dataloader[1-False-True-0-3]
tests/distributed/test_mp_dataloader.py::test_dist_dataloader[1-False-True-4-3]
tests/distributed/test_mp_dataloader.py::test_dist_dataloader[1-False-False-0-3]
tests/distributed/test_mp_dataloader.py::test_dist_dataloader[1-False-False-4-3]
  /root/jenkins/workspace/dgl_PR-4648/python/dgl/backend/pytorch/tensor.py:277: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).
    mask = th.tensor(mask, dtype=th.bool)

-- Docs: https://docs.pytest.org/en/stable/how-to/capture-warnings.html
- generated xml file: /root/jenkins/workspace/dgl_PR-4648/pytest_distributed.xml -
============================ slowest 100 durations =============================
90.56s call     tests/distributed/test_dist_graph_store.py::test_server_client
63.46s call     tests/distributed/test_distributed_sampling.py::test_rpc_sampling_shuffle[2]
45.08s call     tests/distributed/test_distributed_sampling.py::test_rpc_sampling_shuffle[1]
42.43s call     tests/distributed/test_mp_dataloader.py::test_dataloader[edge-4-3]
42.26s call     tests/distributed/test_mp_dataloader.py::test_dataloader[node-4-3]
38.21s call     tests/distributed/test_mp_dataloader.py::test_neg_dataloader[4-3]
33.95s call     tests/distributed/test_mp_dataloader.py::test_dataloader[edge-0-3]
33.13s call     tests/distributed/test_mp_dataloader.py::test_dataloader[node-0-3]
31.83s call     tests/distributed/test_mp_dataloader.py::test_neg_dataloader[0-3]
19.06s call     tests/distributed/test_dist_graph_store.py::test_dist_emb_server_client
16.29s call     tests/distributed/test_rpc.py::test_rpc_timeout[socket]
15.18s call     tests/distributed/test_rpc.py::test_rpc_timeout[tensorpipe]
14.69s call     tests/distributed/test_rpc.py::test_multi_client[tensorpipe]
14.64s call     tests/distributed/test_new_kvstore.py::test_kv_multi_role
13.54s call     tests/distributed/test_mp_dataloader.py::test_dist_dataloader[1-False-False-4-3]
13.33s call     tests/distributed/test_mp_dataloader.py::test_dist_dataloader[1-True-False-4-3]
12.52s call     tests/distributed/test_new_kvstore.py::test_kv_store
12.46s call     tests/distributed/test_mp_dataloader.py::test_dist_dataloader[1-False-True-4-3]
12.21s call     tests/distributed/test_rpc.py::test_multi_client_connect[socket]
12.08s call     tests/distributed/test_mp_dataloader.py::test_dist_dataloader[1-True-True-4-3]
11.29s call     tests/distributed/test_dist_graph_store.py::test_split[True]
10.54s call     tests/distributed/test_distributed_sampling.py::test_rpc_find_edges_shuffle[2]
9.79s call     tests/distributed/test_rpc.py::test_multi_client[socket]
9.33s call     tests/distributed/test_distributed_sampling.py::test_rpc_in_subgraph
9.28s call     tests/distributed/test_mp_dataloader.py::test_dist_dataloader[1-False-False-0-3]
9.14s call     tests/distributed/test_mp_dataloader.py::test_dist_dataloader[1-True-False-0-3]
8.94s call     tests/distributed/test_mp_dataloader.py::test_dist_dataloader[1-True-True-0-3]
8.65s call     tests/distributed/test_distributed_sampling.py::test_rpc_find_edges_shuffle[1]
8.41s call     tests/distributed/test_rpc.py::test_multi_client_connect[tensorpipe]
8.20s call     tests/distributed/test_mp_dataloader.py::test_dist_dataloader[1-False-True-0-3]
8.02s call     tests/distributed/test_rpc.py::test_rpc[tensorpipe]
6.38s call     tests/distributed/test_rpc.py::test_multi_thread_rpc[socket]
6.35s call     tests/distributed/test_rpc.py::test_multi_thread_rpc[tensorpipe]
5.59s call     tests/distributed/test_distributed_sampling.py::test_rpc_get_degree_shuffle[2]
5.02s call     tests/distributed/test_dist_graph_store.py::test_split_even
4.47s call     tests/distributed/test_distributed_sampling.py::test_rpc_get_degree_shuffle[1]
4.16s call     tests/distributed/test_dist_graph_store.py::test_split[False]
3.46s call     tests/distributed/test_dist_graph_store.py::test_standalone
3.43s call     tests/distributed/test_dist_graph_store.py::test_standalone_node_emb
0.42s call     tests/distributed/test_partition.py::test_hetero_partition
0.32s call     tests/distributed/test_mp_dataloader.py::test_standalone
0.32s call     tests/distributed/test_distributed_sampling.py::test_standalone_etype_sampling
0.24s call     tests/distributed/test_distributed_sampling.py::test_standalone_sampling
0.03s call     tests/distributed/test_partition.py::test_partition
0.01s setup    tests/distributed/test_mp_dataloader.py::test_dataloader[node-0-3]

(55 durations < 0.005s hidden.  Use -vv to show these durations.)
=========================== short test summary info ============================
FAILED tests/distributed/test_partition.py::test_partition - TypeError: empty...
======= 1 failed, 48 passed, 6 skipped, 47 warnings in 720.74s (0:12:00) =======
PASSED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-e-shp4-g0] PASSED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-u-e-shp4-g1] FAIL: distributed
[Pipeline] }
PASSED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-u-shp0-g0] [Pipeline] // timeout
[Pipeline] }
[Pipeline] // stage
Post stage
[Pipeline] cleanWs
[WS-CLEANUP] Deleting project workspace...
[WS-CLEANUP] Deferred wipeout is disabled by the job configuration...
PASSED   [ 67%]
tests/compute/test_sparse.py::test_sddmm[int64-add-v-u-shp3-g0] PASSED   [ 67%]
tests/compute/test_sparse.py::test_sddmm[int64-add-v-u-shp3-g1] PASSED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-u-shp0-g1] PASSED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-u-shp1-g0] PASSED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-u-shp1-g1] PASSED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-u-shp2-g0] PASSED   [ 67%]
tests/compute/test_sparse.py::test_sddmm[int64-add-v-u-shp4-g0] PASSED   [ 67%]
tests/compute/test_sparse.py::test_sddmm[int64-add-v-u-shp4-g1] [WS-CLEANUP] done
[Pipeline] }
$ docker stop --time=1 9d11908e8a1623fe2cb3ee8ecb53c2f89a04f190533b91026267498f2decd76b
PASSED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-u-shp2-g1] PASSED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-u-shp3-g0] PASSED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-u-shp3-g1] PASSED   [ 67%]
tests/compute/test_sparse.py::test_sddmm[int64-add-v-v-shp0-g0] PASSED   [ 67%]
tests/compute/test_sparse.py::test_sddmm[int64-add-v-v-shp0-g1] PASSED   [ 67%]
tests/compute/test_sparse.py::test_sddmm[int64-add-v-v-shp1-g0] PASSED   [ 67%]
tests/compute/test_sparse.py::test_sddmm[int64-add-v-v-shp1-g1] PASSED   [ 67%]
tests/compute/test_sparse.py::test_sddmm[int64-add-v-v-shp2-g0] PASSED   [ 67%]
tests/compute/test_sparse.py::test_sddmm[int64-add-v-v-shp2-g1] PASSED   [ 67%]
tests/compute/test_sparse.py::test_sddmm[int64-add-v-v-shp3-g0] PASSED   [ 67%]
tests/compute/test_sparse.py::test_sddmm[int64-add-v-v-shp3-g1] PASSED   [ 67%]
tests/compute/test_sparse.py::test_sddmm[int64-add-v-v-shp4-g0] PASSED   [ 68%]
tests/compute/test_sparse.py::test_sddmm[int64-add-v-v-shp4-g1] PASSED   [ 68%]
tests/compute/test_sparse.py::test_sddmm[int64-add-v-e-shp0-g0] PASSED   [ 68%]
tests/compute/test_sparse.py::test_sddmm[int64-add-v-e-shp0-g1] PASSED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-u-shp4-g0] PASSED   [ 68%]
tests/compute/test_sparse.py::test_sddmm[int64-add-v-e-shp1-g0] PASSED   [ 68%]
tests/compute/test_sparse.py::test_sddmm[int64-add-v-e-shp1-g1] PASSED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-u-shp4-g1] PASSED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-v-shp0-g0] PASSED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-v-shp0-g1] PASSED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-v-shp1-g0] PASSED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-v-shp1-g1] PASSED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-v-shp2-g0] PASSED [ 45%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-v-shp2-g1] PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-v-shp3-g0] PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-v-shp3-g1] PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-v-shp4-g0] PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-v-shp4-g1] PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-e-shp0-g0] PASSED   [ 68%]
tests/compute/test_sparse.py::test_sddmm[int64-add-v-e-shp2-g0] PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-e-shp0-g1] PASSED   [ 68%]
tests/compute/test_sparse.py::test_sddmm[int64-add-v-e-shp2-g1] PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-e-shp1-g0] PASSED   [ 68%]
tests/compute/test_sparse.py::test_sddmm[int64-add-v-e-shp3-g0] PASSED   [ 68%]
tests/compute/test_sparse.py::test_sddmm[int64-add-v-e-shp3-g1] $ docker rm -f 9d11908e8a1623fe2cb3ee8ecb53c2f89a04f190533b91026267498f2decd76b
PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-e-shp1-g1] PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-e-shp2-g0] PASSED   [ 68%]
tests/compute/test_sparse.py::test_sddmm[int64-add-v-e-shp4-g0] PASSED   [ 68%]
tests/compute/test_sparse.py::test_sddmm[int64-add-v-e-shp4-g1] PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-e-shp2-g1] PASSED   [ 68%]
tests/compute/test_sparse.py::test_sddmm[int64-add-e-u-shp0-g0] [Pipeline] // withDockerContainer
[Pipeline] }
[Pipeline] // withEnv
[Pipeline] }
PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-e-shp3-g0] [Pipeline] // node
[Pipeline] }
PASSED   [ 68%]
tests/compute/test_sparse.py::test_sddmm[int64-add-e-u-shp0-g1] [Pipeline] // stage
[Pipeline] }
Failed in branch Distributed
PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-e-shp3-g1] PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-e-shp4-g0] PASSED   [ 68%]
tests/compute/test_sparse.py::test_sddmm[int64-add-e-u-shp1-g0] PASSED   [ 68%]
tests/compute/test_sparse.py::test_sddmm[int64-add-e-u-shp1-g1] PASSED       [ 74%]
tests/go/test_pipeline.py::test_graphpred[pna-csv] PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-v-e-shp4-g1] PASSED   [ 68%]
tests/compute/test_sparse.py::test_sddmm[int64-add-e-u-shp2-g0] PASSED   [ 68%]
tests/compute/test_sparse.py::test_sddmm[int64-add-e-u-shp2-g1] PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-u-shp0-g0] PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-u-shp0-g1] PASSED   [ 68%]
tests/compute/test_sparse.py::test_sddmm[int64-add-e-u-shp3-g0] PASSED   [ 68%]
tests/compute/test_sparse.py::test_sddmm[int64-add-e-u-shp3-g1] PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-u-shp1-g0] PASSED   [ 68%]
tests/compute/test_sparse.py::test_sddmm[int64-add-e-u-shp4-g0] PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-u-shp1-g1] PASSED   [ 68%]
tests/compute/test_sparse.py::test_sddmm[int64-add-e-u-shp4-g1] PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-u-shp2-g0] PASSED   [ 68%]
tests/compute/test_sparse.py::test_sddmm[int64-add-e-v-shp0-g0] PASSED   [ 68%]
tests/compute/test_sparse.py::test_sddmm[int64-add-e-v-shp0-g1] PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-u-shp2-g1] PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-u-shp3-g0] PASSED   [ 68%]
tests/compute/test_sparse.py::test_sddmm[int64-add-e-v-shp1-g0] PASSED   [ 69%]
tests/compute/test_sparse.py::test_sddmm[int64-add-e-v-shp1-g1] PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-u-shp3-g1] PASSED   [ 69%]
tests/compute/test_sparse.py::test_sddmm[int64-add-e-v-shp2-g0] PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-u-shp4-g0] PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-u-shp4-g1] PASSED   [ 69%]
tests/compute/test_sparse.py::test_sddmm[int64-add-e-v-shp2-g1] PASSED   [ 69%]
tests/compute/test_sparse.py::test_sddmm[int64-add-e-v-shp3-g0] PASSED [ 46%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-v-shp0-g0] PASSED   [ 69%]
tests/compute/test_sparse.py::test_sddmm[int64-add-e-v-shp3-g1] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-v-shp0-g1] PASSED   [ 69%]
tests/compute/test_sparse.py::test_sddmm[int64-add-e-v-shp4-g0] PASSED   [ 69%]
tests/compute/test_sparse.py::test_sddmm[int64-add-e-v-shp4-g1] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-v-shp1-g0] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-v-shp1-g1] PASSED   [ 69%]
tests/compute/test_sparse.py::test_sddmm[int64-add-e-e-shp0-g0] PASSED   [ 69%]
tests/compute/test_sparse.py::test_sddmm[int64-add-e-e-shp0-g1] PASSED   [ 69%]
tests/compute/test_sparse.py::test_sddmm[int64-add-e-e-shp1-g0] PASSED   [ 69%]
tests/compute/test_sparse.py::test_sddmm[int64-add-e-e-shp1-g1] PASSED   [ 69%]
tests/compute/test_sparse.py::test_sddmm[int64-add-e-e-shp2-g0] PASSED   [ 69%]
tests/compute/test_sparse.py::test_sddmm[int64-add-e-e-shp2-g1] PASSED   [ 69%]
tests/compute/test_sparse.py::test_sddmm[int64-add-e-e-shp3-g0] PASSED   [ 69%]
tests/compute/test_sparse.py::test_sddmm[int64-add-e-e-shp3-g1] PASSED   [ 69%]
tests/compute/test_sparse.py::test_sddmm[int64-add-e-e-shp4-g0] PASSED   [ 69%]
tests/compute/test_sparse.py::test_sddmm[int64-add-e-e-shp4-g1] PASSED   [ 69%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-u-u-shp0-g0] PASSED   [ 69%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-u-u-shp0-g1] PASSED   [ 69%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-u-u-shp1-g0] PASSED   [ 69%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-u-u-shp1-g1] PASSED   [ 69%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-u-u-shp2-g0] PASSED   [ 69%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-u-u-shp2-g1] PASSED   [ 69%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-u-u-shp3-g0] PASSED   [ 69%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-u-u-shp3-g1] PASSED   [ 70%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-u-u-shp4-g0] PASSED   [ 70%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-u-u-shp4-g1] PASSED   [ 70%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-u-v-shp0-g0] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-v-shp2-g0] PASSED   [ 70%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-u-v-shp0-g1] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-v-shp2-g1] PASSED   [ 70%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-u-v-shp1-g0] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-v-shp3-g0] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-v-shp3-g1] PASSED   [ 70%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-u-v-shp1-g1] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-v-shp4-g0] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-v-shp4-g1] PASSED   [ 70%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-u-v-shp2-g0] PASSED   [ 70%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-u-v-shp2-g1] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-e-shp0-g0] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-e-shp0-g1] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-e-shp1-g0] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-e-shp1-g1] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-e-shp2-g0] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-e-shp2-g1] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-e-shp3-g0] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-e-shp3-g1] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-e-shp4-g0] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-sub-e-e-shp4-g1] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-u-shp0-g0] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-u-shp0-g1] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-u-shp1-g0] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-u-shp1-g1] PASSED [ 47%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-u-shp2-g0] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-u-shp2-g1] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-u-shp3-g0] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-u-shp3-g1] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-u-shp4-g0] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-u-shp4-g1] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-v-shp0-g0] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-v-shp0-g1] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-v-shp1-g0] PASSED   [ 70%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-u-v-shp3-g0] PASSED   [ 70%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-u-v-shp3-g1] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-v-shp1-g1] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-v-shp2-g0] PASSED   [ 70%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-u-v-shp4-g0] PASSED   [ 70%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-u-v-shp4-g1] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-v-shp2-g1] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-v-shp3-g0] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-v-shp3-g1] PASSED   [ 70%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-u-e-shp0-g0] PASSED   [ 70%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-u-e-shp0-g1] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-v-shp4-g0] PASSED   [ 70%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-u-e-shp1-g0] PASSED   [ 70%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-u-e-shp1-g1] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-v-shp4-g1] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-e-shp0-g0] PASSED   [ 70%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-u-e-shp2-g0] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-e-shp0-g1] PASSED   [ 70%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-u-e-shp2-g1] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-e-shp1-g0] PASSED   [ 70%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-u-e-shp3-g0] PASSED   [ 70%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-u-e-shp3-g1] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-e-shp1-g1] PASSED   [ 70%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-u-e-shp4-g0] PASSED   [ 70%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-u-e-shp4-g1] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-e-shp2-g0] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-e-shp2-g1] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-e-shp3-g0] PASSED   [ 70%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-v-u-shp0-g0] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-e-shp3-g1] PASSED   [ 70%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-v-u-shp0-g1] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-e-shp4-g0] PASSED   [ 71%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-v-u-shp1-g0] PASSED [ 48%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-u-e-shp4-g1] PASSED   [ 71%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-v-u-shp1-g1] PASSED                [ 75%]
tests/go/test_pipeline.py::test_graphpred[pna-ogbg-molhiv] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-u-shp0-g0] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-u-shp0-g1] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-u-shp1-g0] PASSED   [ 71%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-v-u-shp2-g0] PASSED   [ 71%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-v-u-shp2-g1] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-u-shp1-g1] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-u-shp2-g0] PASSED   [ 71%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-v-u-shp3-g0] PASSED   [ 71%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-v-u-shp3-g1] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-u-shp2-g1] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-u-shp3-g0] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-u-shp3-g1] PASSED   [ 71%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-v-u-shp4-g0] PASSED   [ 71%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-v-u-shp4-g1] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-u-shp4-g0] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-u-shp4-g1] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-v-shp0-g0] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-v-shp0-g1] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-v-shp1-g0] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-v-shp1-g1] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-v-shp2-g0] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-v-shp2-g1] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-v-shp3-g0] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-v-shp3-g1] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-v-shp4-g0] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-v-shp4-g1] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-e-shp0-g0] PASSED   [ 71%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-v-v-shp0-g0] PASSED   [ 71%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-v-v-shp0-g1] PASSED   [ 71%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-v-v-shp1-g0] PASSED   [ 71%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-v-v-shp1-g1] PASSED   [ 71%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-v-v-shp2-g0] PASSED   [ 71%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-v-v-shp2-g1] PASSED   [ 71%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-v-v-shp3-g0] PASSED   [ 71%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-v-v-shp3-g1] PASSED   [ 71%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-v-v-shp4-g0] PASSED   [ 71%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-v-v-shp4-g1] PASSED   [ 71%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-v-e-shp0-g0] PASSED   [ 71%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-v-e-shp0-g1] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-e-shp0-g1] PASSED   [ 71%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-v-e-shp1-g0] PASSED   [ 71%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-v-e-shp1-g1] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-e-shp1-g0] PASSED   [ 71%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-v-e-shp2-g0] PASSED [ 49%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-e-shp1-g1] PASSED   [ 71%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-v-e-shp2-g1] PASSED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-e-shp2-g0] PASSED   [ 71%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-v-e-shp3-g0] PASSED   [ 72%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-v-e-shp3-g1] PASSED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-e-shp2-g1] PASSED   [ 72%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-v-e-shp4-g0] PASSED   [ 72%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-v-e-shp4-g1] PASSED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-e-shp3-g0] PASSED   [ 72%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-e-u-shp0-g0] PASSED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-e-shp3-g1] PASSED   [ 72%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-e-u-shp0-g1] PASSED   [ 72%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-e-u-shp1-g0] PASSED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-e-shp4-g0] PASSED   [ 72%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-e-u-shp1-g1] PASSED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-v-e-shp4-g1] PASSED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-u-shp0-g0] PASSED   [ 72%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-e-u-shp2-g0] PASSED   [ 72%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-e-u-shp2-g1] PASSED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-u-shp0-g1] PASSED   [ 72%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-e-u-shp3-g0] PASSED   [ 72%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-e-u-shp3-g1] PASSED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-u-shp1-g0] PASSED   [ 72%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-e-u-shp4-g0] PASSED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-u-shp1-g1] PASSED   [ 72%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-e-u-shp4-g1] PASSED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-u-shp2-g0] PASSED   [ 72%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-e-v-shp0-g0] PASSED   [ 72%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-e-v-shp0-g1] PASSED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-u-shp2-g1] PASSED   [ 72%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-e-v-shp1-g0] PASSED   [ 72%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-e-v-shp1-g1] PASSED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-u-shp3-g0] PASSED   [ 72%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-e-v-shp2-g0] PASSED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-u-shp3-g1] PASSED   [ 72%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-e-v-shp2-g1] PASSED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-u-shp4-g0] PASSED   [ 72%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-e-v-shp3-g0] PASSED   [ 72%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-e-v-shp3-g1] PASSED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-u-shp4-g1] PASSED   [ 72%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-e-v-shp4-g0] PASSED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-v-shp0-g0] PASSED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-v-shp0-g1] PASSED   [ 72%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-e-v-shp4-g1] PASSED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-v-shp1-g0] PASSED   [ 72%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-e-e-shp0-g0] PASSED   [ 73%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-e-e-shp0-g1] PASSED   [ 73%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-e-e-shp1-g0] PASSED   [ 73%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-e-e-shp1-g1] PASSED   [ 73%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-e-e-shp2-g0] PASSED   [ 73%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-e-e-shp2-g1] PASSED   [ 73%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-e-e-shp3-g0] PASSED   [ 73%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-e-e-shp3-g1] PASSED   [ 73%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-e-e-shp4-g0] PASSED   [ 73%]
tests/compute/test_sparse.py::test_sddmm[int64-sub-e-e-shp4-g1] PASSED   [ 73%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-u-u-shp0-g0] PASSED   [ 73%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-u-u-shp0-g1] PASSED   [ 73%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-u-u-shp1-g0] PASSED   [ 73%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-u-u-shp1-g1] PASSED   [ 73%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-u-u-shp2-g0] PASSED   [ 73%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-u-u-shp2-g1] PASSED   [ 73%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-u-u-shp3-g0] PASSED   [ 73%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-u-u-shp3-g1] PASSED   [ 73%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-u-u-shp4-g0] PASSED   [ 73%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-u-u-shp4-g1] PASSED   [ 73%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-u-v-shp0-g0] PASSED   [ 73%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-u-v-shp0-g1] PASSED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-v-shp1-g1] PASSED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-v-shp2-g0] PASSED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-v-shp2-g1] PASSED   [ 73%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-u-v-shp1-g0] PASSED        [ 76%]
tests/go/test_pipeline.py::test_graphpred[pna-ogbg-molpcba] PASSED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-v-shp3-g0] PASSED   [ 73%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-u-v-shp1-g1] PASSED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-v-shp3-g1] PASSED [ 50%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-v-shp4-g0] PASSED   [ 73%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-u-v-shp2-g0] PASSED   [ 73%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-u-v-shp2-g1] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-v-shp4-g1] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-e-shp0-g0] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-e-shp0-g1] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-e-shp1-g0] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-e-shp1-g1] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-e-shp2-g0] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-e-shp2-g1] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-e-shp3-g0] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-e-shp3-g1] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-e-shp4-g0] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-mul-e-e-shp4-g1] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-u-shp0-g0] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-u-shp0-g1] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-u-shp1-g0] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-u-shp1-g1] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-u-shp2-g0] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-u-shp2-g1] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-u-shp3-g0] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-u-shp3-g1] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-u-shp4-g0] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-u-shp4-g1] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-v-shp0-g0] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-v-shp0-g1] PASSED   [ 74%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-u-v-shp3-g0] PASSED   [ 74%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-u-v-shp3-g1] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-v-shp1-g0] PASSED [ 51%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-v-shp1-g1] PASSED   [ 74%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-u-v-shp4-g0] PASSED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-v-shp2-g0] PASSED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-v-shp2-g1] PASSED   [ 74%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-u-v-shp4-g1] PASSED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-v-shp3-g0] PASSED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-v-shp3-g1] PASSED   [ 74%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-u-e-shp0-g0] PASSED   [ 74%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-u-e-shp0-g1] PASSED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-v-shp4-g0] PASSED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-v-shp4-g1] PASSED   [ 74%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-u-e-shp1-g0] PASSED   [ 74%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-u-e-shp1-g1] PASSED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-e-shp0-g0] PASSED   [ 74%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-u-e-shp2-g0] PASSED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-e-shp0-g1] PASSED   [ 74%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-u-e-shp2-g1] PASSED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-e-shp1-g0] PASSED   [ 74%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-u-e-shp3-g0] PASSED   [ 74%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-u-e-shp3-g1] PASSED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-e-shp1-g1] PASSED   [ 74%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-u-e-shp4-g0] PASSED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-e-shp2-g0] PASSED   [ 74%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-u-e-shp4-g1] PASSED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-e-shp2-g1] PASSED   [ 74%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-v-u-shp0-g0] PASSED   [ 74%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-v-u-shp0-g1] PASSED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-e-shp3-g0] PASSED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-e-shp3-g1] PASSED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-e-shp4-g0] PASSED   [ 74%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-v-u-shp1-g0] PASSED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-u-e-shp4-g1] PASSED   [ 74%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-v-u-shp1-g1] PASSED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-u-shp0-g0] PASSED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-u-shp0-g1] PASSED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-u-shp1-g0] PASSED   [ 74%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-v-u-shp2-g0] PASSED   [ 74%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-v-u-shp2-g1] PASSED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-u-shp1-g1] PASSED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-u-shp2-g0] PASSED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-u-shp2-g1] PASSED   [ 74%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-v-u-shp3-g0] PASSED   [ 74%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-v-u-shp3-g1] PASSED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-u-shp3-g0] PASSED [ 52%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-u-shp3-g1] PASSED   [ 74%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-v-u-shp4-g0] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-u-shp4-g0] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-u-shp4-g1] PASSED   [ 74%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-v-u-shp4-g1] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-v-shp0-g0] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-v-shp0-g1] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-v-shp1-g0] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-v-shp1-g1] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-v-shp2-g0] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-v-shp2-g1] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-v-shp3-g0] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-v-shp3-g1] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-v-shp4-g0] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-v-shp4-g1] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-e-shp0-g0] PASSED   [ 75%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-v-v-shp0-g0] PASSED   [ 75%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-v-v-shp0-g1] PASSED   [ 75%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-v-v-shp1-g0] PASSED   [ 75%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-v-v-shp1-g1] PASSED   [ 75%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-v-v-shp2-g0] PASSED   [ 75%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-v-v-shp2-g1] PASSED   [ 75%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-v-v-shp3-g0] PASSED   [ 75%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-v-v-shp3-g1] PASSED   [ 75%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-v-v-shp4-g0] PASSED   [ 75%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-v-v-shp4-g1] PASSED   [ 75%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-v-e-shp0-g0] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-e-shp0-g1] PASSED   [ 75%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-v-e-shp0-g1] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-e-shp1-g0] PASSED   [ 75%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-v-e-shp1-g0] PASSED   [ 75%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-v-e-shp1-g1] PASSED       [ 77%]
tests/go/test_pipeline.py::test_recipe[graphpred_hiv_gin.yaml] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-e-shp1-g1] PASSED   [ 75%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-v-e-shp2-g0] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-e-shp2-g0] PASSED   [ 75%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-v-e-shp2-g1] PASSED   [ 75%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-v-e-shp3-g0] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-e-shp2-g1] PASSED   [ 75%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-v-e-shp3-g1] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-e-shp3-g0] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-e-shp3-g1] PASSED   [ 75%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-v-e-shp4-g0] PASSED   [ 75%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-v-e-shp4-g1] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-e-shp4-g0] PASSED   [ 75%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-e-u-shp0-g0] PASSED   [ 75%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-e-u-shp0-g1] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-v-e-shp4-g1] PASSED   [ 75%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-e-u-shp1-g0] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-u-shp0-g0] PASSED   [ 75%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-e-u-shp1-g1] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-u-shp0-g1] PASSED   [ 75%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-e-u-shp2-g0] PASSED   [ 76%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-e-u-shp2-g1] PASSED [ 53%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-u-shp1-g0] PASSED   [ 76%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-e-u-shp3-g0] PASSED   [ 76%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-e-u-shp3-g1] PASSED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-u-shp1-g1] PASSED    [ 79%]
tests/go/test_pipeline.py::test_recipe[graphpred_hiv_pna.yaml] PASSED   [ 76%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-e-u-shp4-g0] PASSED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-u-shp2-g0] PASSED   [ 76%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-e-u-shp4-g1] PASSED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-u-shp2-g1] PASSED   [ 76%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-e-v-shp0-g0] PASSED   [ 76%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-e-v-shp0-g1] PASSED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-u-shp3-g0] PASSED   [ 76%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-e-v-shp1-g0] PASSED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-u-shp3-g1] PASSED   [ 76%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-e-v-shp1-g1] PASSED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-u-shp4-g0] PASSED   [ 76%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-e-v-shp2-g0] PASSED   [ 76%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-e-v-shp2-g1] PASSED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-u-shp4-g1] PASSED   [ 76%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-e-v-shp3-g0] PASSED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-v-shp0-g0] PASSED   [ 76%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-e-v-shp3-g1] PASSED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-v-shp0-g1] PASSED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-v-shp1-g0] PASSED   [ 76%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-e-v-shp4-g0] PASSED   [ 76%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-e-v-shp4-g1] PASSED   [ 76%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-e-e-shp0-g0] PASSED   [ 76%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-e-e-shp0-g1] PASSED   [ 76%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-e-e-shp1-g0] PASSED   [ 76%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-e-e-shp1-g1] PASSED   [ 76%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-e-e-shp2-g0] PASSED   [ 76%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-e-e-shp2-g1] PASSED   [ 76%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-e-e-shp3-g0] PASSED   [ 76%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-e-e-shp3-g1] PASSED   [ 76%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-e-e-shp4-g0] PASSED   [ 77%]
tests/compute/test_sparse.py::test_sddmm[int64-mul-e-e-shp4-g1] PASSED   [ 77%]
tests/compute/test_sparse.py::test_sddmm[int64-div-u-u-shp0-g0] PASSED   [ 77%]
tests/compute/test_sparse.py::test_sddmm[int64-div-u-u-shp0-g1] PASSED   [ 77%]
tests/compute/test_sparse.py::test_sddmm[int64-div-u-u-shp1-g0] PASSED   [ 77%]
tests/compute/test_sparse.py::test_sddmm[int64-div-u-u-shp1-g1] PASSED   [ 77%]
tests/compute/test_sparse.py::test_sddmm[int64-div-u-u-shp2-g0] PASSED   [ 77%]
tests/compute/test_sparse.py::test_sddmm[int64-div-u-u-shp2-g1] PASSED   [ 77%]
tests/compute/test_sparse.py::test_sddmm[int64-div-u-u-shp3-g0] PASSED   [ 77%]
tests/compute/test_sparse.py::test_sddmm[int64-div-u-u-shp3-g1] PASSED   [ 77%]
tests/compute/test_sparse.py::test_sddmm[int64-div-u-u-shp4-g0] PASSED   [ 77%]
tests/compute/test_sparse.py::test_sddmm[int64-div-u-u-shp4-g1] PASSED   [ 77%]
tests/compute/test_sparse.py::test_sddmm[int64-div-u-v-shp0-g0] PASSED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-v-shp1-g1] PASSED [ 54%]PASSED    [ 80%]
tests/go/test_pipeline.py::test_recipe[graphpred_pcba_gin.yaml] PASSED   [ 77%]
tests/compute/test_sparse.py::test_sddmm[int64-div-u-v-shp0-g1] 
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-v-shp2-g0] PASSED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-v-shp2-g1] PASSED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-v-shp3-g0] PASSED   [ 77%]
tests/compute/test_sparse.py::test_sddmm[int64-div-u-v-shp1-g0] PASSED   [ 77%]
tests/compute/test_sparse.py::test_sddmm[int64-div-u-v-shp1-g1] PASSED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-v-shp3-g1] PASSED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-v-shp4-g0] PASSED   [ 77%]
tests/compute/test_sparse.py::test_sddmm[int64-div-u-v-shp2-g0] PASSED   [ 77%]
tests/compute/test_sparse.py::test_sddmm[int64-div-u-v-shp2-g1] PASSED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-v-shp4-g1] PASSED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-e-shp0-g0] PASSED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-e-shp0-g1] PASSED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-e-shp1-g0] PASSED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-e-shp1-g1] PASSED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-e-shp2-g0] PASSED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-e-shp2-g1] PASSED [ 54%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-e-shp3-g0] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-e-shp3-g1] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-e-shp4-g0] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-div-e-e-shp4-g1] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-u-shp0-g0] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-u-shp0-g1] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-u-shp1-g0] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-u-shp1-g1] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-u-shp2-g0] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-u-shp2-g1] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-u-shp3-g0] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-u-shp3-g1] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-u-shp4-g0] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-u-shp4-g1] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-v-shp0-g0] PASSED   [ 77%]
tests/compute/test_sparse.py::test_sddmm[int64-div-u-v-shp3-g0] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-v-shp0-g1] PASSED   [ 81%]
tests/go/test_pipeline.py::test_recipe[linkpred_cora_sage.yaml] PASSED   [ 77%]
tests/compute/test_sparse.py::test_sddmm[int64-div-u-v-shp3-g1] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-v-shp1-g0] PASSED   [ 77%]
tests/compute/test_sparse.py::test_sddmm[int64-div-u-v-shp4-g0] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-v-shp1-g1] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-v-shp2-g0] PASSED   [ 77%]
tests/compute/test_sparse.py::test_sddmm[int64-div-u-v-shp4-g1] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-v-shp2-g1] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-v-shp3-g0] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-v-shp3-g1] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-v-shp4-g0] PASSED   [ 77%]
tests/compute/test_sparse.py::test_sddmm[int64-div-u-e-shp0-g0] PASSED   [ 77%]
tests/compute/test_sparse.py::test_sddmm[int64-div-u-e-shp0-g1] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-v-shp4-g1] PASSED   [ 77%]
tests/compute/test_sparse.py::test_sddmm[int64-div-u-e-shp1-g0] PASSED   [ 77%]
tests/compute/test_sparse.py::test_sddmm[int64-div-u-e-shp1-g1] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-e-shp0-g0] PASSED   [ 78%]
tests/compute/test_sparse.py::test_sddmm[int64-div-u-e-shp2-g0] PASSED [ 55%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-e-shp0-g1] PASSED   [ 78%]
tests/compute/test_sparse.py::test_sddmm[int64-div-u-e-shp2-g1] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-e-shp1-g0] PASSED   [ 78%]
tests/compute/test_sparse.py::test_sddmm[int64-div-u-e-shp3-g0] PASSED   [ 78%]
tests/compute/test_sparse.py::test_sddmm[int64-div-u-e-shp3-g1] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-e-shp1-g1] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-e-shp2-g0] PASSED   [ 82%]
tests/go/test_pipeline.py::test_recipe[linkpred_citation2_sage.yaml] PASSED   [ 78%]
tests/compute/test_sparse.py::test_sddmm[int64-div-u-e-shp4-g0] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-e-shp2-g1] PASSED   [ 78%]
tests/compute/test_sparse.py::test_sddmm[int64-div-u-e-shp4-g1] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-e-shp3-g0] PASSED   [ 78%]
tests/compute/test_sparse.py::test_sddmm[int64-div-v-u-shp0-g0] PASSED   [ 78%]
tests/compute/test_sparse.py::test_sddmm[int64-div-v-u-shp0-g1] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-e-shp3-g1] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-e-shp4-g0] PASSED   [ 78%]
tests/compute/test_sparse.py::test_sddmm[int64-div-v-u-shp1-g0] PASSED   [ 78%]
tests/compute/test_sparse.py::test_sddmm[int64-div-v-u-shp1-g1] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-u-e-shp4-g1] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-u-shp0-g0] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-u-shp0-g1] PASSED   [ 78%]
tests/compute/test_sparse.py::test_sddmm[int64-div-v-u-shp2-g0] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-u-shp1-g0] PASSED   [ 78%]
tests/compute/test_sparse.py::test_sddmm[int64-div-v-u-shp2-g1] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-u-shp1-g1] PASSED   [ 78%]
tests/compute/test_sparse.py::test_sddmm[int64-div-v-u-shp3-g0] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-u-shp2-g0] PASSED   [ 78%]
tests/compute/test_sparse.py::test_sddmm[int64-div-v-u-shp3-g1] PASSED [ 83%]
tests/go/test_pipeline.py::test_recipe[linkpred_collab_sage.yaml] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-u-shp2-g1] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-u-shp3-g0] PASSED   [ 78%]PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-u-shp3-g1] 
tests/compute/test_sparse.py::test_sddmm[int64-div-v-u-shp4-g0] PASSED   [ 78%]
tests/compute/test_sparse.py::test_sddmm[int64-div-v-u-shp4-g1] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-u-shp4-g0] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-u-shp4-g1] PASSED   [ 78%]
tests/compute/test_sparse.py::test_sddmm[int64-div-v-v-shp0-g0] PASSED   [ 78%]
tests/compute/test_sparse.py::test_sddmm[int64-div-v-v-shp0-g1] PASSED   [ 78%]
tests/compute/test_sparse.py::test_sddmm[int64-div-v-v-shp1-g0] PASSED   [ 78%]
tests/compute/test_sparse.py::test_sddmm[int64-div-v-v-shp1-g1] PASSED   [ 78%]
tests/compute/test_sparse.py::test_sddmm[int64-div-v-v-shp2-g0] PASSED   [ 78%]
tests/compute/test_sparse.py::test_sddmm[int64-div-v-v-shp2-g1] PASSED   [ 78%]
tests/compute/test_sparse.py::test_sddmm[int64-div-v-v-shp3-g0] PASSED   [ 78%]
tests/compute/test_sparse.py::test_sddmm[int64-div-v-v-shp3-g1] PASSED   [ 79%]
tests/compute/test_sparse.py::test_sddmm[int64-div-v-v-shp4-g0] PASSED   [ 79%]
tests/compute/test_sparse.py::test_sddmm[int64-div-v-v-shp4-g1] PASSED   [ 79%]
tests/compute/test_sparse.py::test_sddmm[int64-div-v-e-shp0-g0] PASSED   [ 79%]
tests/compute/test_sparse.py::test_sddmm[int64-div-v-e-shp0-g1] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-v-shp0-g0] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-v-shp0-g1] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-v-shp1-g0] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-v-shp1-g1] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-v-shp2-g0] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-v-shp2-g1] PASSED [ 56%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-v-shp3-g0] PASSED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-v-shp3-g1] PASSED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-v-shp4-g0] PASSED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-v-shp4-g1] PASSED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-e-shp0-g0] PASSED   [ 79%]
tests/compute/test_sparse.py::test_sddmm[int64-div-v-e-shp1-g0] PASSED   [ 79%]
tests/compute/test_sparse.py::test_sddmm[int64-div-v-e-shp1-g1] PASSED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-e-shp0-g1] PASSED   [ 79%]PASSED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-e-shp1-g0] 
tests/compute/test_sparse.py::test_sddmm[int64-div-v-e-shp2-g0] PASSED   [ 79%]
tests/compute/test_sparse.py::test_sddmm[int64-div-v-e-shp2-g1] PASSED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-e-shp1-g1] PASSED   [ 79%]
tests/compute/test_sparse.py::test_sddmm[int64-div-v-e-shp3-g0] PASSED   [ 79%]
tests/compute/test_sparse.py::test_sddmm[int64-div-v-e-shp3-g1] PASSED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-e-shp2-g0] PASSED   [ 79%]
tests/compute/test_sparse.py::test_sddmm[int64-div-v-e-shp4-g0] PASSED   [ 79%]
tests/compute/test_sparse.py::test_sddmm[int64-div-v-e-shp4-g1] PASSED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-e-shp2-g1] PASSED [ 85%]
tests/go/test_pipeline.py::test_recipe[nodepred_citeseer_gat.yaml] PASSED   [ 79%]
tests/compute/test_sparse.py::test_sddmm[int64-div-e-u-shp0-g0] PASSED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-e-shp3-g0] PASSED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-e-shp3-g1] PASSED   [ 79%]
tests/compute/test_sparse.py::test_sddmm[int64-div-e-u-shp0-g1] PASSED   [ 79%]
tests/compute/test_sparse.py::test_sddmm[int64-div-e-u-shp1-g0] PASSED   [ 79%]
tests/compute/test_sparse.py::test_sddmm[int64-div-e-u-shp1-g1] PASSED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-e-shp4-g0] PASSED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-v-e-shp4-g1] PASSED   [ 79%]
tests/compute/test_sparse.py::test_sddmm[int64-div-e-u-shp2-g0] PASSED   [ 79%]
tests/compute/test_sparse.py::test_sddmm[int64-div-e-u-shp2-g1] PASSED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-u-shp0-g0] PASSED   [ 79%]
tests/compute/test_sparse.py::test_sddmm[int64-div-e-u-shp3-g0] PASSED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-u-shp0-g1] PASSED   [ 79%]
tests/compute/test_sparse.py::test_sddmm[int64-div-e-u-shp3-g1] PASSED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-u-shp1-g0] PASSED   [ 79%]
tests/compute/test_sparse.py::test_sddmm[int64-div-e-u-shp4-g0] PASSED   [ 79%]
tests/compute/test_sparse.py::test_sddmm[int64-div-e-u-shp4-g1] PASSED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-u-shp1-g1] PASSED   [ 79%]
tests/compute/test_sparse.py::test_sddmm[int64-div-e-v-shp0-g0] PASSED [ 86%]
tests/go/test_pipeline.py::test_recipe[nodepred_citeseer_gcn.yaml] PASSED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-u-shp2-g0] PASSED   [ 79%]
tests/compute/test_sparse.py::test_sddmm[int64-div-e-v-shp0-g1] PASSED   [ 79%]
tests/compute/test_sparse.py::test_sddmm[int64-div-e-v-shp1-g0] PASSED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-u-shp2-g1] PASSED   [ 80%]
tests/compute/test_sparse.py::test_sddmm[int64-div-e-v-shp1-g1] PASSED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-u-shp3-g0] PASSED   [ 80%]
tests/compute/test_sparse.py::test_sddmm[int64-div-e-v-shp2-g0] PASSED   [ 80%]
tests/compute/test_sparse.py::test_sddmm[int64-div-e-v-shp2-g1] PASSED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-u-shp3-g1] PASSED   [ 80%]
tests/compute/test_sparse.py::test_sddmm[int64-div-e-v-shp3-g0] PASSED   [ 80%]
tests/compute/test_sparse.py::test_sddmm[int64-div-e-v-shp3-g1] PASSED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-u-shp4-g0] PASSED   [ 80%]
tests/compute/test_sparse.py::test_sddmm[int64-div-e-v-shp4-g0] PASSED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-u-shp4-g1] PASSED   [ 80%]
tests/compute/test_sparse.py::test_sddmm[int64-div-e-v-shp4-g1] PASSED [ 57%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-v-shp0-g0] PASSED   [ 80%]
tests/compute/test_sparse.py::test_sddmm[int64-div-e-e-shp0-g0] PASSED   [ 80%]
tests/compute/test_sparse.py::test_sddmm[int64-div-e-e-shp0-g1] PASSED   [ 80%]
tests/compute/test_sparse.py::test_sddmm[int64-div-e-e-shp1-g0] PASSED   [ 80%]
tests/compute/test_sparse.py::test_sddmm[int64-div-e-e-shp1-g1] PASSED   [ 80%]
tests/compute/test_sparse.py::test_sddmm[int64-div-e-e-shp2-g0] PASSED   [ 80%]
tests/compute/test_sparse.py::test_sddmm[int64-div-e-e-shp2-g1] PASSED   [ 80%]
tests/compute/test_sparse.py::test_sddmm[int64-div-e-e-shp3-g0] PASSED   [ 80%]
tests/compute/test_sparse.py::test_sddmm[int64-div-e-e-shp3-g1] PASSED   [ 80%]
tests/compute/test_sparse.py::test_sddmm[int64-div-e-e-shp4-g0] PASSED   [ 80%]
tests/compute/test_sparse.py::test_sddmm[int64-div-e-e-shp4-g1] PASSED   [ 80%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-u-u-shp0-g0] PASSED   [ 80%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-u-u-shp0-g1] PASSED   [ 80%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-u-u-shp1-g0] PASSED   [ 80%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-u-u-shp1-g1] PASSED   [ 80%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-u-u-shp2-g0] PASSED   [ 80%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-u-u-shp2-g1] PASSED   [ 80%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-u-u-shp3-g0] PASSED   [ 81%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-u-u-shp3-g1] PASSED   [ 81%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-u-u-shp4-g0] PASSED   [ 81%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-u-u-shp4-g1] PASSED   [ 81%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-u-v-shp0-g0] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-v-shp0-g1] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-v-shp1-g0] PASSED   [ 81%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-u-v-shp0-g1] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-v-shp1-g1] PASSED   [ 81%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-u-v-shp1-g0] PASSED   [ 81%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-u-v-shp1-g1] PASSED [ 87%]
tests/go/test_pipeline.py::test_recipe[nodepred_citeseer_sage.yaml] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-v-shp2-g0] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-v-shp2-g1] PASSED   [ 81%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-u-v-shp2-g0] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-v-shp3-g0] PASSED   [ 81%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-u-v-shp2-g1] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-v-shp3-g1] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-v-shp4-g0] PASSED   [ 81%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-u-v-shp3-g0] PASSED   [ 81%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-u-v-shp3-g1] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-v-shp4-g1] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-e-shp0-g0] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-e-shp0-g1] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-e-shp1-g0] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-e-shp1-g1] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-e-shp2-g0] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-e-shp2-g1] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-e-shp3-g0] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-e-shp3-g1] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-e-shp4-g0] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-dot-e-e-shp4-g1] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-u-shp0-g0] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-u-shp0-g1] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-u-shp1-g0] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-u-shp1-g1] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-u-shp2-g0] PASSED [ 58%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-u-shp2-g1] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-u-shp3-g0] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-u-shp3-g1] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-u-shp4-g0] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-u-shp4-g1] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-v-shp0-g0] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-v-shp0-g1] PASSED   [ 81%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-u-v-shp4-g0] PASSED   [ 81%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-u-v-shp4-g1] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-v-shp1-g0] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-v-shp1-g1] PASSED   [ 81%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-u-e-shp0-g0] PASSED   [ 81%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-u-e-shp0-g1] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-v-shp2-g0] PASSED   [ 81%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-u-e-shp1-g0] PASSED   [ 81%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-u-e-shp1-g1] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-v-shp2-g1] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-v-shp3-g0] PASSED [ 88%]
tests/go/test_pipeline.py::test_recipe[nodepred_cora_gat.yaml] PASSED   [ 81%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-u-e-shp2-g0] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-v-shp3-g1] PASSED   [ 81%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-u-e-shp2-g1] PASSED   [ 81%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-u-e-shp3-g0] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-v-shp4-g0] PASSED   [ 81%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-u-e-shp3-g1] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-v-shp4-g1] PASSED   [ 81%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-u-e-shp4-g0] PASSED   [ 81%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-u-e-shp4-g1] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-e-shp0-g0] PASSED   [ 81%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-v-u-shp0-g0] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-e-shp0-g1] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-e-shp1-g0] PASSED   [ 81%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-v-u-shp0-g1] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-e-shp1-g1] PASSED   [ 82%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-v-u-shp1-g0] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-e-shp2-g0] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-e-shp2-g1] PASSED   [ 82%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-v-u-shp1-g1] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-e-shp3-g0] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-e-shp3-g1] PASSED    [ 90%]
tests/go/test_pipeline.py::test_recipe[nodepred_cora_gcn.yaml] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-e-shp4-g0] PASSED [ 59%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-u-e-shp4-g1] PASSED   [ 82%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-v-u-shp2-g0] PASSED   [ 82%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-v-u-shp2-g1] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-u-shp0-g0] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-u-shp0-g1] PASSED   [ 82%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-v-u-shp3-g0] PASSED   [ 82%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-v-u-shp3-g1] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-u-shp1-g0] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-u-shp1-g1] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-u-shp2-g0] PASSED   [ 82%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-v-u-shp4-g0] PASSED   [ 82%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-v-u-shp4-g1] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-u-shp2-g1] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-u-shp3-g0] PASSED   [ 82%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-v-v-shp0-g0] PASSED   [ 82%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-v-v-shp0-g1] PASSED   [ 82%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-v-v-shp1-g0] PASSED   [ 82%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-v-v-shp1-g1] PASSED   [ 82%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-v-v-shp2-g0] PASSED   [ 82%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-v-v-shp2-g1] PASSED   [ 82%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-v-v-shp3-g0] PASSED   [ 82%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-v-v-shp3-g1] PASSED   [ 82%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-v-v-shp4-g0] PASSED   [ 82%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-v-v-shp4-g1] PASSED   [ 82%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-v-e-shp0-g0] PASSED   [ 82%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-v-e-shp0-g1] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-u-shp3-g1] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-u-shp4-g0] PASSED   [ 82%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-v-e-shp1-g0] PASSED   [ 82%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-v-e-shp1-g1] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-u-shp4-g1] PASSED   [ 82%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-v-e-shp2-g0] PASSED   [ 82%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-v-e-shp2-g1] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-v-shp0-g0] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-v-shp0-g1] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-v-shp1-g0] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-v-shp1-g1] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-v-shp2-g0] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-v-shp2-g1] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-v-shp3-g0] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-v-shp3-g1] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-v-shp4-g0] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-v-shp4-g1] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-e-shp0-g0] PASSED    [ 91%]
tests/go/test_pipeline.py::test_recipe[nodepred_cora_sage.yaml] PASSED   [ 83%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-v-e-shp3-g0] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-e-shp0-g1] PASSED   [ 83%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-v-e-shp3-g1] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-e-shp1-g0] PASSED   [ 83%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-v-e-shp4-g0] PASSED   [ 83%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-v-e-shp4-g1] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-e-shp1-g1] PASSED [ 60%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-e-shp2-g0] PASSED   [ 83%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-e-u-shp0-g0] PASSED   [ 83%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-e-u-shp0-g1] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-e-shp2-g1] PASSED   [ 83%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-e-u-shp1-g0] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-e-shp3-g0] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-e-shp3-g1] PASSED   [ 83%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-e-u-shp1-g1] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-e-shp4-g0] PASSED   [ 83%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-e-u-shp2-g0] PASSED   [ 83%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-e-u-shp2-g1] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-v-e-shp4-g1] PASSED   [ 83%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-e-u-shp3-g0] PASSED   [ 83%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-e-u-shp3-g1] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-u-shp0-g0] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-u-shp0-g1] PASSED   [ 83%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-e-u-shp4-g0] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-u-shp1-g0] PASSED   [ 83%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-e-u-shp4-g1] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-u-shp1-g1] PASSED   [ 92%]
tests/go/test_pipeline.py::test_recipe[nodepred_pubmed_gat.yaml] PASSED   [ 83%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-e-v-shp0-g0] PASSED   [ 83%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-e-v-shp0-g1] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-u-shp2-g0] PASSED   [ 83%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-e-v-shp1-g0] PASSED   [ 83%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-e-v-shp1-g1] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-u-shp2-g1] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-u-shp3-g0] PASSED   [ 83%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-e-v-shp2-g0] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-u-shp3-g1] PASSED   [ 83%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-e-v-shp2-g1] PASSED   [ 83%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-e-v-shp3-g0] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-u-shp4-g0] PASSED   [ 83%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-e-v-shp3-g1] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-u-shp4-g1] PASSED   [ 83%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-e-v-shp4-g0] PASSED   [ 83%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-e-v-shp4-g1] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-v-shp0-g0] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-v-shp0-g1] PASSED   [ 83%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-e-e-shp0-g0] PASSED   [ 84%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-e-e-shp0-g1] PASSED   [ 84%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-e-e-shp1-g0] PASSED   [ 84%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-e-e-shp1-g1] PASSED   [ 84%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-e-e-shp2-g0] PASSED   [ 84%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-e-e-shp2-g1] PASSED   [ 84%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-e-e-shp3-g0] PASSED   [ 84%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-e-e-shp3-g1] PASSED   [ 84%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-e-e-shp4-g0] PASSED   [ 84%]
tests/compute/test_sparse.py::test_sddmm[int64-dot-e-e-shp4-g1] PASSED   [ 84%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-u-u-shp0-g0] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-u-u-shp0-g1] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-u-u-shp1-g0] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-u-u-shp1-g1] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-u-u-shp2-g0] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-u-u-shp2-g1] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-u-u-shp3-g0] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-u-u-shp3-g1] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-u-u-shp4-g0] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-u-u-shp4-g1] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-u-v-shp0-g0] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-v-shp1-g0] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-u-v-shp0-g1] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-u-v-shp1-g0] PASSED  [ 93%]
tests/go/test_pipeline.py::test_recipe[nodepred_pubmed_gcn.yaml] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-v-shp1-g1] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-u-v-shp1-g1] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-v-shp2-g0] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-u-v-shp2-g0] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-u-v-shp2-g1] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-v-shp2-g1] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-u-v-shp3-g0] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-u-v-shp3-g1] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-v-shp3-g0] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-u-v-shp4-g0] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-u-v-shp4-g1] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-v-shp3-g1] PASSED [ 61%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-v-shp4-g0] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-u-e-shp0-g0] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-u-e-shp0-g1] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-u-e-shp1-g0] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-u-e-shp1-g1] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-u-e-shp2-g0] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-u-e-shp2-g1] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-v-shp4-g1] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-u-e-shp3-g0] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-u-e-shp3-g1] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-u-e-shp4-g0] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-u-e-shp4-g1] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-v-u-shp0-g0] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-v-u-shp0-g1] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-e-shp0-g0] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-e-shp0-g1] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-e-shp1-g0] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-e-shp1-g1] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-e-shp2-g0] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-e-shp2-g1] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-e-shp3-g0] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-e-shp3-g1] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-e-shp4-g0] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_lhs-e-e-shp4-g1] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-u-shp0-g0] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-u-shp0-g1] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-u-shp1-g0] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-u-shp1-g1] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-u-shp2-g0] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-u-shp2-g1] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-u-shp3-g0] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-u-shp3-g1] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-u-shp4-g0] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-u-shp4-g1] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-v-shp0-g0] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-v-u-shp1-g0] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-v-u-shp1-g1] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-v-shp0-g1] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-v-u-shp2-g0] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-v-shp1-g0] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-v-u-shp2-g1] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-v-u-shp3-g0] PASSED [ 62%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-v-shp1-g1] PASSED  [ 95%]
tests/go/test_pipeline.py::test_recipe[nodepred_pubmed_sage.yaml] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-v-u-shp3-g1] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-v-shp2-g0] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-v-shp2-g1] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-v-u-shp4-g0] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-v-u-shp4-g1] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-v-shp3-g0] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-v-v-shp0-g0] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-v-v-shp0-g1] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-v-v-shp1-g0] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-v-v-shp1-g1] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-v-v-shp2-g0] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-v-v-shp2-g1] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-v-v-shp3-g0] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-v-v-shp3-g1] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-v-v-shp4-g0] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-v-v-shp4-g1] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-v-e-shp0-g0] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-v-e-shp0-g1] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-v-e-shp1-g0] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-v-e-shp1-g1] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-v-e-shp2-g0] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-v-e-shp2-g1] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-v-e-shp3-g0] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-v-e-shp3-g1] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-v-e-shp4-g0] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-v-shp3-g1] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-v-e-shp4-g1] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-e-u-shp0-g0] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-e-u-shp0-g1] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-v-shp4-g0] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-e-u-shp1-g0] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-e-u-shp1-g1] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-v-shp4-g1] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-e-shp0-g0] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-e-u-shp2-g0] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-e-u-shp2-g1] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-e-shp0-g1] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-e-u-shp3-g0] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-e-shp1-g0] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-e-u-shp3-g1] PASSED [ 87%]PASSED [ 96%]
tests/go/test_pipeline.py::test_recipe[nodepred-ns_arxiv_gcn.yaml] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-e-shp1-g1] 
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-e-u-shp4-g0] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-e-u-shp4-g1] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-e-shp2-g0] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-e-v-shp0-g0] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-e-v-shp0-g1] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-e-shp2-g1] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-e-v-shp1-g0] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-e-shp3-g0] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-e-v-shp1-g1] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-e-shp3-g1] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-e-v-shp2-g0] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-e-v-shp2-g1] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-e-shp4-g0] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-e-v-shp3-g0] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-u-e-shp4-g1] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-e-v-shp3-g1] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-u-shp0-g0] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-e-v-shp4-g0] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-e-v-shp4-g1] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-u-shp0-g1] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-e-e-shp0-g0] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-e-e-shp0-g1] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-e-e-shp1-g0] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-e-e-shp1-g1] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-e-e-shp2-g0] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-e-e-shp2-g1] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-e-e-shp3-g0] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-e-e-shp3-g1] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-e-e-shp4-g0] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_lhs-e-e-shp4-g1] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-u-u-shp0-g0] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-u-u-shp0-g1] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-u-u-shp1-g0] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-u-u-shp1-g1] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-u-u-shp2-g0] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-u-u-shp2-g1] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-u-u-shp3-g0] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-u-u-shp3-g1] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-u-u-shp4-g0] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-u-u-shp4-g1] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-u-v-shp0-g0] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-u-shp1-g0] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-u-v-shp0-g1] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-u-shp1-g1] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-u-v-shp1-g0] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-u-v-shp1-g1] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-u-shp2-g0] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-u-v-shp2-g0] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-u-shp2-g1] PASSED [ 97%]
tests/go/test_pipeline.py::test_recipe[nodepred-ns_product_sage.yaml] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-u-v-shp2-g1] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-u-shp3-g0] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-u-v-shp3-g0] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-u-v-shp3-g1] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-u-shp3-g1] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-u-v-shp4-g0] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-u-v-shp4-g1] PASSED [ 63%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-u-shp4-g0] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-u-e-shp0-g0] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-u-shp4-g1] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-u-e-shp0-g1] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-v-shp0-g0] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-v-shp0-g1] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-v-shp1-g0] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-v-shp1-g1] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-v-shp2-g0] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-v-shp2-g1] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-v-shp3-g0] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-v-shp3-g1] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-v-shp4-g0] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-v-shp4-g1] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-e-shp0-g0] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-e-shp0-g1] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-u-e-shp1-g0] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-u-e-shp1-g1] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-e-shp1-g0] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-u-e-shp2-g0] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-u-e-shp2-g1] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-e-shp1-g1] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-e-shp2-g0] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-u-e-shp3-g0] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-u-e-shp3-g1] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-e-shp2-g1] PASSED [ 98%]
tests/go/test_pipeline.py::test_node_cora PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-u-e-shp4-g0] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-u-e-shp4-g1] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-e-shp3-g0] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-v-u-shp0-g0] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-v-u-shp0-g1] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-e-shp3-g1] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-v-u-shp1-g0] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-e-shp4-g0] PASSED [ 64%]PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-v-u-shp1-g1] 
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-v-e-shp4-g1] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-u-shp0-g0] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-v-u-shp2-g0] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-v-u-shp2-g1] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-u-shp0-g1] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-v-u-shp3-g0] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-v-u-shp3-g1] PASSED [ 64%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-u-shp1-g0] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-v-u-shp4-g0] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-u-shp1-g1] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-v-u-shp4-g1] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-u-shp2-g0] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-v-v-shp0-g0] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-v-v-shp0-g1] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-v-v-shp1-g0] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-v-v-shp1-g1] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-v-v-shp2-g0] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-v-v-shp2-g1] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-v-v-shp3-g0] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-v-v-shp3-g1] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-v-v-shp4-g0] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-v-v-shp4-g1] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-v-e-shp0-g0] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-v-e-shp0-g1] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-u-shp2-g1] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-v-e-shp1-g0] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-v-e-shp1-g1] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-u-shp3-g0] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-u-shp3-g1] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-v-e-shp2-g0] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-u-shp4-g0] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-v-e-shp2-g1] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-u-shp4-g1] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-v-e-shp3-g0] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-v-e-shp3-g1] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-v-shp0-g0] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-v-e-shp4-g0] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-v-e-shp4-g1] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-v-shp0-g1] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-e-u-shp0-g0] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-e-u-shp0-g1] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-e-u-shp1-g0] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-e-u-shp1-g1] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-e-u-shp2-g0] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-e-u-shp2-g1] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-e-u-shp3-g0] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-e-u-shp3-g1] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-e-u-shp4-g0] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-e-u-shp4-g1] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-e-v-shp0-g0] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-v-shp1-g0] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-e-v-shp0-g1] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-e-v-shp1-g0] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-e-v-shp1-g1] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-e-v-shp2-g0] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-e-v-shp2-g1] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-e-v-shp3-g0] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-e-v-shp3-g1] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-e-v-shp4-g0] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-e-v-shp4-g1] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-e-e-shp0-g0] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-e-e-shp0-g1] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-e-e-shp1-g0] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-e-e-shp1-g1] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-e-e-shp2-g0] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-e-e-shp2-g1] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-e-e-shp3-g0] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-e-e-shp3-g1] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-e-e-shp4-g0] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-e-e-shp4-g1] PASSED [ 91%]
tests/compute/test_sparse.py::test_edge_softmax[int32-shp0-src-g0] PASSED [ 91%]
tests/compute/test_sparse.py::test_edge_softmax[int32-shp0-dst-g0] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-v-shp1-g1] PASSED [ 91%]
tests/compute/test_sparse.py::test_edge_softmax[int32-shp1-src-g0] PASSED [ 91%]
tests/compute/test_sparse.py::test_edge_softmax[int32-shp1-dst-g0] PASSED [ 91%]
tests/compute/test_sparse.py::test_edge_softmax[int32-shp2-src-g0] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-v-shp2-g0] PASSED [ 91%]
tests/compute/test_sparse.py::test_edge_softmax[int32-shp2-dst-g0] PASSED [ 91%]
tests/compute/test_sparse.py::test_edge_softmax[int64-shp0-src-g0] PASSED [ 92%]
tests/compute/test_sparse.py::test_edge_softmax[int64-shp0-dst-g0] PASSED [ 92%]
tests/compute/test_sparse.py::test_edge_softmax[int64-shp1-src-g0] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-v-shp2-g1] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-v-shp3-g0] PASSED [ 92%]
tests/compute/test_sparse.py::test_edge_softmax[int64-shp1-dst-g0] PASSED [ 92%]
tests/compute/test_sparse.py::test_edge_softmax[int64-shp2-src-g0] PASSED [ 92%]
tests/compute/test_sparse.py::test_edge_softmax[int64-shp2-dst-g0] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-v-shp3-g1] PASSED [ 92%]
tests/compute/test_sparse.py::test_segment_reduce[sum] PASSED            [ 92%]
tests/compute/test_sparse.py::test_segment_reduce[max] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-v-shp4-g0] PASSED            [ 92%]
tests/compute/test_sparse.py::test_segment_reduce[min] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-v-shp4-g1] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-e-shp0-g0] PASSED [ 65%]PASSED            [ 92%]
tests/compute/test_sparse.py::test_segment_reduce[mean] 
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-e-shp0-g1] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-e-shp1-g0] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-e-shp1-g1] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-e-shp2-g0] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-e-shp2-g1] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-e-shp3-g0] PASSED [ 65%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-e-shp3-g1] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-e-shp4-g0] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype0-copy_rhs-e-e-shp4-g1] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-u-shp0-g0] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-u-shp0-g1] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-u-shp1-g0] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-u-shp1-g1] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-u-shp2-g0] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-u-shp2-g1] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-u-shp3-g0] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-u-shp3-g1] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-u-shp4-g0] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-u-shp4-g1] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-v-shp0-g0] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-v-shp0-g1] PASSED           [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype0-0.01-1-int32] SKIPPED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype0-0.01-1-int64] SKIPPED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype0-0.01-8-int32] SKIPPED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype0-0.01-8-int64] SKIPPED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype0-0.01-16-int32] SKIPPED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype0-0.01-16-int64] SKIPPED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype0-0.01-64-int32] SKIPPED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype0-0.01-64-int64] SKIPPED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype0-0.01-256-int32] SKIPPED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype0-0.01-256-int64] SKIPPED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype1-0.01-1-int32] SKIPPED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype1-0.01-1-int64] SKIPPED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype1-0.01-8-int32] SKIPPED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype1-0.01-8-int64] SKIPPED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype1-0.01-16-int32] SKIPPED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype1-0.01-16-int64] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype1-0.01-64-int32] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype1-0.01-64-int64] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype1-0.01-256-int32] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype1-0.01-256-int64] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype2-0.003-1-int32] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype2-0.003-1-int64] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype2-0.003-8-int32] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype2-0.003-8-int64] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype2-0.003-16-int32] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype2-0.003-16-int64] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype2-0.003-64-int32] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype2-0.003-64-int64] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype2-0.003-256-int32] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype2-0.003-256-int64] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype3-0.0001-1-int32] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype3-0.0001-1-int64] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype3-0.0001-8-int32] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype3-0.0001-8-int64] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype3-0.0001-16-int32] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype3-0.0001-16-int64] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype3-0.0001-64-int32] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype3-0.0001-64-int64] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype3-0.0001-256-int32] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype3-0.0001-256-int64] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype0-0.01-1] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype0-0.01-8] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype0-0.01-16] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype0-0.01-64] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype0-0.01-256] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype1-0.02-1] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype1-0.02-8] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype1-0.02-16] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype1-0.02-64] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype1-0.02-256] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype2-0.003-1] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype2-0.003-8] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype2-0.003-16] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype2-0.003-64] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype2-0.003-256] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype3-0.0001-1] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype3-0.0001-8] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype3-0.0001-16] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype3-0.0001-64] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype3-0.0001-256] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_use_libxsmm_switch SKIPPED (Only ...) [ 94%]
tests/compute/test_specialization.py::test_v2v_update_all[int32] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-v-shp1-g0] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-v-shp1-g1] FAILED                         [100%]

=================================== FAILURES ===================================
_________________________________ test_gcn[g0] _________________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(7,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})

    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature']))
    def test_gcn(g):
        data_info = {
            'num_nodes': g.num_nodes(),
            'out_size': 7
        }
        node_feat = None
        edge_feat = g.edata['scalar_w']
    
        # node embedding + not use_edge_weight
        model = GCN(data_info, embed_size=10, use_edge_weight=False)
>       model(g, node_feat)

tests/go/test_model.py:18: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/torch/nn/modules/module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/dglgo-0.0.2-py3.7.egg/dglgo/model/node_encoder/gcn.py:65: in forward
    h = layer(g, h, edge_weight=edge_weight)
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/torch/nn/modules/module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python/dgl/nn/pytorch/conv/graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python/dgl/heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python/dgl/core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python/dgl/core.py:332: in invoke_gspmm
    z = op(graph, x)
python/dgl/ops/spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7f1fec0ed630>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-2.6910e-01, -7.8042e-01, -2.7359e-03, -3.2366e-02,  5.7685e-01,
         -4.6568e-01,  2.5890e-01],
        ...586e+00, -9.3698e-01,  2.8112e-01, -3.2321e-01, -8.1060e-01,
         -5.0542e-01,  1.1881e+00]], grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
______________________________ test_gcn_block[g0] ______________________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)

    @pytest.mark.parametrize('g', get_cases(['block-bipartite']))
    def test_gcn_block(g):
        data_info = {
            'in_size': 10,
            'out_size': 7
        }
    
        blocks = [g]
        node_feat = torch.randn(g.num_src_nodes(), data_info['in_size'])
        edge_feat = torch.abs(torch.randn(g.num_edges()))
        # not use_edge_weight
        model = GCN(data_info, use_edge_weight=False)
>       model.forward_block(blocks, node_feat)

tests/go/test_model.py:47: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/dglgo-0.0.2-py3.7.egg/dglgo/model/node_encoder/gcn.py:75: in forward_block
    h = layer(block, h, edge_weight=edge_weight)
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/torch/nn/modules/module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python/dgl/nn/pytorch/conv/graphconv.py:423: in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))
python/dgl/heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python/dgl/core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python/dgl/core.py:332: in invoke_gspmm
    z = op(graph, x)
python/dgl/ops/spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7f1fcd5a7900>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.1589,  1.8696,  2.4331, -0.3975, -2.2170, -0.2013, -0.0745],
        [ 1.6567, -1.0327, -0.8260, -0.4221, ....0788,  0.8304],
        [-0.7473,  0.4214,  0.4989, -0.3357,  0.6086, -1.8190, -1.4260]],
       grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
_________________________________ test_gat[g0] _________________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(2,), dtype=torch.float32), 'ft': Scheme(shap...r': Scheme(shape=(8, 1), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})

    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature']))
    def test_gat(g):
        data_info = {
            'num_nodes': g.num_nodes(),
            'out_size': 7
        }
        node_feat = None
    
        # node embedding
        model = GAT(data_info, embed_size=10)
>       model(g, node_feat)

tests/go/test_model.py:63: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/torch/nn/modules/module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/dglgo-0.0.2-py3.7.egg/dglgo/model/node_encoder/gat.py:77: in forward
    h = self.gat_layers[l](graph, h).flatten(1)
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/torch/nn/modules/module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python/dgl/nn/pytorch/conv/gatconv.py:303: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python/dgl/heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python/dgl/core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python/dgl/ops/sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7f1fcd4a0ca8>
op = 'add'
lhs_data = tensor([[[-0.6326],
         [-0.1831],
         [-0.6531],
         [ 0.0632],
         [ 0.3255],
         [-1.1950]....1747],
         [ 0.0026],
         [-0.7224],
         [-1.0509],
         [ 0.2529]]], grad_fn=<UnsqueezeBackward0>)
rhs_data = tensor([[[-0.3539],
         [ 0.0378],
         [ 1.2875],
         [ 1.1894],
         [ 1.7702],
         [ 0.9975]....0768],
         [-0.6095],
         [-0.3911],
         [-0.1075],
         [ 1.4653]]], grad_fn=<UnsqueezeBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
______________________________ test_gat_block[g0] ______________________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)

    @pytest.mark.parametrize('g', get_cases(['block-bipartite']))
    def test_gat_block(g):
        data_info = {
            'in_size': 10,
            'out_size': 7
        }
    
        blocks = [g]
        node_feat = torch.randn(g.num_src_nodes(), data_info['in_size'])
        model = GAT(data_info, num_layers=1, heads=[8])
>       model.forward_block(blocks, node_feat)

tests/go/test_model.py:81: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/dglgo-0.0.2-py3.7.egg/dglgo/model/node_encoder/gat.py:86: in forward_block
    logits = self.gat_layers[-1](blocks[-1], h).mean(1)
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/torch/nn/modules/module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python/dgl/nn/pytorch/conv/gatconv.py:303: in forward
    graph.apply_edges(fn.u_add_v('el', 'er', 'e'))
python/dgl/heterograph.py:4458: in apply_edges
    edata = core.invoke_gsddmm(g, func)
python/dgl/core.py:266: in invoke_gsddmm
    z = op(graph, x, y)
python/dgl/ops/sddmm.py:128: in func
    lhs_target=lhs_target, rhs_target=rhs_target)
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7f1fcd6b1480>
op = 'add'
lhs_data = tensor([[[ 0.4385],
         [-0.1202],
         [ 0.6981],
         [ 0.8333],
         [ 0.7433],
         [ 0.2976]....4548],
         [ 0.1070],
         [ 0.2085],
         [-0.5366],
         [ 0.5327]]], grad_fn=<UnsqueezeBackward0>)
rhs_data = tensor([[[ 1.7303],
         [-0.2018],
         [ 1.4550],
         [ 0.8330],
         [ 0.5960],
         [ 0.8471]....2173],
         [ 0.3000],
         [ 0.0604],
         [ 0.9013],
         [ 0.8341]]], grad_fn=<UnsqueezeBackward0>)
lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
_________________________________ test_gin[g0] _________________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(10,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})

    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature']))
    def test_gin(g):
        data_info = {
            'num_nodes': g.num_nodes(),
            'out_size': 7
        }
        node_feat = None
    
        # node embedding
        model = GIN(data_info, embed_size=10)
>       model(g, node_feat)

tests/go/test_model.py:93: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/torch/nn/modules/module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/dglgo-0.0.2-py3.7.egg/dglgo/model/node_encoder/gin.py:57: in forward
    h = self.conv_list[i](graph, h)
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/torch/nn/modules/module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python/dgl/nn/pytorch/conv/ginconv.py:147: in forward
    graph.update_all(aggregate_fn, _reducer('m', 'neigh'))
python/dgl/heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python/dgl/core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python/dgl/core.py:332: in invoke_gspmm
    z = op(graph, x)
python/dgl/ops/spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7f1fcd6babd0>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = Parameter containing:
tensor([[-1.1347, -0.3137, -1.1602, -0.8782, -0.1786, -0.5248,  0.2633,  0.5876,
          0.931...1.1028,  0.5693,  1.6440, -0.4924, -1.3252, -0.2387, -0.2325, -0.4017,
          0.2927, -1.5606]], requires_grad=True)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
________________________________ test_sage[g0] _________________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(7,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})

    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature']))
    def test_sage(g):
        data_info = {
            'num_nodes': g.num_nodes(),
            'out_size': 7
        }
        node_feat = None
        edge_feat = g.edata['scalar_w']
    
        # node embedding
        model = GraphSAGE(data_info, embed_size=10)
>       model(g, node_feat)

tests/go/test_model.py:112: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/torch/nn/modules/module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/dglgo-0.0.2-py3.7.egg/dglgo/model/node_encoder/sage.py:58: in forward
    h = layer(graph, h, edge_feat)
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/torch/nn/modules/module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python/dgl/nn/pytorch/conv/sageconv.py:249: in forward
    graph.update_all(msg_fn, fn.sum('m', 'neigh'))
python/dgl/heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python/dgl/core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python/dgl/core.py:332: in invoke_gspmm
    z = op(graph, x)
python/dgl/ops/spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7f1fcd650318>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[-1.0455,  0.1407, -3.6295, -0.0526, -2.7218, -2.7691,  4.4690],
        [ 0.9919,  0.8407,  1.4997,  0.4836, ....4458,  1.5004],
        [-0.7763,  2.2160,  0.0870, -0.4196,  0.8794,  2.8239, -1.2195]],
       grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
_____________________________ test_sage_block[g0] ______________________________

g = Block(num_src_nodes=6, num_dst_nodes=3, num_edges=3)

    @pytest.mark.parametrize('g', get_cases(['block-bipartite']))
    def test_sage_block(g):
        data_info = {
            'in_size': 10,
            'out_size': 7
        }
    
        blocks = [g]
        node_feat = torch.randn(g.num_src_nodes(), data_info['in_size'])
        edge_feat = torch.abs(torch.randn(g.num_edges()))
        model = GraphSAGE(data_info, embed_size=-1)
>       model.forward_block(blocks, node_feat)

tests/go/test_model.py:133: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/dglgo-0.0.2-py3.7.egg/dglgo/model/node_encoder/sage.py:67: in forward_block
    h = layer(block, h, edge_feat)
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/torch/nn/modules/module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python/dgl/nn/pytorch/conv/sageconv.py:249: in forward
    graph.update_all(msg_fn, fn.sum('m', 'neigh'))
python/dgl/heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python/dgl/core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python/dgl/core.py:332: in invoke_gspmm
    z = op(graph, x)
python/dgl/ops/spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7f1fcd5692d0>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.5972,  1.5512,  0.4784,  3.6568,  3.2279, -0.6815,  1.2453],
        [ 2.3165,  2.3200,  2.0049,  1.9059, ....2783, -2.2715],
        [-0.6206,  0.4980,  1.8651, -0.6186, -0.2495, -0.8035,  1.8651]],
       grad_fn=<MmBackward>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
_________________________________ test_sgc[g0] _________________________________

g = Graph(num_nodes=10, num_edges=17,
      ndata_schemes={'h': Scheme(shape=(10,), dtype=torch.float32)}
      edata_schemes={'scalar_w': Scheme(shape=(), dtype=torch.float32)})

    @pytest.mark.parametrize('g', get_cases(['has_scalar_e_feature']))
    def test_sgc(g):
        data_info = {
            'num_nodes': g.num_nodes(),
            'out_size': 7
        }
        node_feat = None
    
        # node embedding
        model = SGC(data_info, embed_size=10)
>       model(g, node_feat)

tests/go/test_model.py:146: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/torch/nn/modules/module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/dglgo-0.0.2-py3.7.egg/dglgo/model/node_encoder/sgc.py:46: in forward
    return self.sgc(g, h)
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/torch/nn/modules/module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python/dgl/nn/pytorch/conv/sgconv.py:203: in forward
    fn.sum('m', 'h'))
python/dgl/heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python/dgl/core.py:357: in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)
python/dgl/core.py:332: in invoke_gspmm
    z = op(graph, x)
python/dgl/ops/spmm.py:189: in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7f1fcd4c5d80>
op = 'copy_lhs', reduce_op = 'sum'
lhs_data = tensor([[ 0.1245,  0.2361,  0.5253, -0.4005,  0.4957,  1.0950, -0.9002,  0.0527,
         -0.6777, -0.3907],
        [...55, -0.6413, -0.4899,  0.4337,  0.1490, -0.8998, -0.1092,  0.0496,
          0.4883, -0.1608]], grad_fn=<MulBackward0>)
rhs_data = None

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
_____________________________ test_ogbg_gin[True] ______________________________

virtual_node = True

    @pytest.mark.parametrize('virtual_node', [True, False])
    def test_ogbg_gin(virtual_node):
        # Test for ogbg-mol datasets
        data_info = {
            'name': 'ogbg-molhiv',
            'out_size': 1
        }
        model = OGBGGIN(data_info,
                        embed_size=10,
                        num_layers=2,
                        virtual_node=virtual_node)
        num_nodes = 5
        num_edges = 15
        g1 = dgl.rand_graph(num_nodes, num_edges)
        g2 = dgl.rand_graph(num_nodes, num_edges)
        g = dgl.batch([g1, g2])
        num_nodes = g.num_nodes()
        num_edges = g.num_edges()
        nfeat = torch.zeros(num_nodes, 9).long()
        efeat = torch.zeros(num_edges, 3).long()
>       model(g, nfeat, efeat)

tests/go/test_model.py:196: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/torch/nn/modules/module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/dglgo-0.0.2-py3.7.egg/dglgo/model/graph_encoder/gin_ogbg.py:91: in forward
    hn = self.conv_layers[layer](graph, hn, he)
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/torch/nn/modules/module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python/dgl/nn/pytorch/conv/gineconv.py:94: in forward
    graph.update_all(self.message, fn.sum('m', 'neigh'))
python/dgl/heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python/dgl/core.py:369: in message_passing
    ndata = invoke_gspmm(g, fn.copy_e(msg, msg), rfunc, edata=msgdata)
python/dgl/core.py:332: in invoke_gspmm
    z = op(graph, x)
python/dgl/ops/spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7f1fcd3eca68>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([[0.0000, 0.6656, 0.0000, 0.3696, 0.0000, 0.1483, 0.0000, 0.0000, 0.0039,
         0.0000],
        [0.0000, 0....   [0.0000, 0.6656, 0.0000, 0.3696, 0.0000, 0.1483, 0.0000, 0.0000, 0.0039,
         0.0000]], grad_fn=<ReluBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
_____________________________ test_ogbg_gin[False] _____________________________

virtual_node = False

    @pytest.mark.parametrize('virtual_node', [True, False])
    def test_ogbg_gin(virtual_node):
        # Test for ogbg-mol datasets
        data_info = {
            'name': 'ogbg-molhiv',
            'out_size': 1
        }
        model = OGBGGIN(data_info,
                        embed_size=10,
                        num_layers=2,
                        virtual_node=virtual_node)
        num_nodes = 5
        num_edges = 15
        g1 = dgl.rand_graph(num_nodes, num_edges)
        g2 = dgl.rand_graph(num_nodes, num_edges)
        g = dgl.batch([g1, g2])
        num_nodes = g.num_nodes()
        num_edges = g.num_edges()
        nfeat = torch.zeros(num_nodes, 9).long()
        efeat = torch.zeros(num_edges, 3).long()
>       model(g, nfeat, efeat)

tests/go/test_model.py:196: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/torch/nn/modules/module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/dglgo-0.0.2-py3.7.egg/dglgo/model/graph_encoder/gin_ogbg.py:91: in forward
    hn = self.conv_layers[layer](graph, hn, he)
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/torch/nn/modules/module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
python/dgl/nn/pytorch/conv/gineconv.py:94: in forward
    graph.update_all(self.message, fn.sum('m', 'neigh'))
python/dgl/heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python/dgl/core.py:369: in message_passing
    ndata = invoke_gspmm(g, fn.copy_e(msg, msg), rfunc, edata=msgdata)
python/dgl/core.py:332: in invoke_gspmm
    z = op(graph, x)
python/dgl/ops/spmm.py:191: in func
    return gspmm(g, 'copy_rhs', reduce_op, None, x)
python/dgl/ops/spmm.py:77: in gspmm
    lhs_data, rhs_data)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7f1fcd3f7cf0>
op = 'copy_rhs', reduce_op = 'sum', lhs_data = None
rhs_data = tensor([[0.0000, 1.4420, 0.0000, 1.7451, 0.0000, 0.4129, 0.0000, 1.5422, 1.2173,
         0.2543],
        [0.0000, 1....   [0.0000, 1.4420, 0.0000, 1.7451, 0.0000, 0.4129, 0.0000, 1.5422, 1.2173,
         0.2543]], grad_fn=<ReluBackward0>)

    def gspmm(gidx, op, reduce_op, lhs_data, rhs_data):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, reduce_op, lhs_data, rhs_data)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:720: TypeError
___________________________________ test_pna ___________________________________

    def test_pna():
        # Test for ogbg-mol datasets
        data_info = {
            'name': 'ogbg-molhiv',
            'delta': 1,
            'out_size': 1
        }
        model = PNA(data_info,
                    embed_size=10,
                    num_layers=2)
        num_nodes = 5
        num_edges = 15
        g = dgl.rand_graph(num_nodes, num_edges)
        nfeat = torch.zeros(num_nodes, 9).long()
>       model(g, nfeat)

tests/go/test_model.py:227: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/torch/nn/modules/module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/dglgo-0.0.2-py3.7.egg/dglgo/model/graph_encoder/pna.py:227: in forward
    hn = conv(graph, hn)
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/torch/nn/modules/module.py:1051: in _call_impl
    return forward_call(*input, **kwargs)
/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/dglgo-0.0.2-py3.7.egg/dglgo/model/graph_encoder/pna.py:123: in forward
    g.update_all(fn.copy_u('h', 'm'), self.reduce)
python/dgl/heterograph.py:4895: in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)
python/dgl/core.py:362: in message_passing
    msgdata = invoke_gsddmm(g, mfunc)
python/dgl/core.py:276: in invoke_gsddmm
    z = op(graph, x)
python/dgl/ops/sddmm.py:164: in copy_u
    return gsddmm(g, 'copy_lhs', x, None)
python/dgl/ops/sddmm.py:75: in gsddmm
    g._graph, op, lhs_data, rhs_data, lhs_target, rhs_target)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

gidx = <dgl.heterograph_index.HeteroGraphIndex object at 0x7f1fccd12c18>
op = 'copy_lhs'
lhs_data = tensor([[-1.4288, -0.2643, -0.6708, -0.3757, -0.2398, -0.4878,  1.0679, -0.7565,
         -1.4319, -1.2884],
        [...88, -0.2643, -0.6708, -0.3757, -0.2398, -0.4878,  1.0679, -0.7565,
         -1.4319, -1.2884]], grad_fn=<AddBackward0>)
rhs_data = None, lhs_target = 'u', rhs_target = 'v'

    def gsddmm(gidx, op, lhs_data, rhs_data, lhs_target='u', rhs_target='v'):
        if op == 'sub':
            op = 'add'
            rhs_data = -rhs_data
        if op == 'div':
            op = 'mul'
            rhs_data = 1. / rhs_data
        args = _cast_if_autocast_enabled(gidx, op, lhs_data, rhs_data, lhs_target, rhs_target)
>       with autocast(enabled=False):
E       TypeError: empty_context() got an unexpected keyword argument 'enabled'

python/dgl/backend/pytorch/sparse.py:731: TypeError
________________________________ test_node_cora ________________________________

    def test_node_cora():
        os.system('dgl configure nodepred --data cora --model gcn')
        os.system('dgl train --cfg nodepred_cora_gcn.yaml')
        assert os.path.exists('results')
>       assert os.path.exists('results/run_0.pth')
E       AssertionError: assert False
E        +  where False = <function exists at 0x7f20c8a51730>('results/run_0.pth')
E        +    where <function exists at 0x7f20c8a51730> = <module 'posixpath' from '/opt/conda/envs/pytorch-ci/lib/python3.7/posixpath.py'>.exists
E        +      where <module 'posixpath' from '/opt/conda/envs/pytorch-ci/lib/python3.7/posixpath.py'> = os.path

tests/go/test_pipeline.py:170: AssertionError
----------------------------- Captured stdout call -----------------------------
Configuration file is generated at /root/jenkins/workspace/dgl_PR-4648@3/nodepred_cora_gcn.yaml
Downloading /root/jenkins/workspace/dgl_PR-4648@3/cora_v2.zip from https://data.dgl.ai/dataset/cora_v2.zip...
Extracting file to /root/jenkins/workspace/dgl_PR-4648@3/cora_v2
Finished data loading and preprocessing.
  NumNodes: 2708
  NumEdges: 10556
  NumFeats: 1433
  NumClasses: 7
  NumTrainingSamples: 140
  NumValidationSamples: 500
  NumTestSamples: 1000
Done saving data into cached files.
Run experiment #0
----------------------------- Captured stderr call -----------------------------
WARNING:root:The OGB package is out of date. Your version is 1.3.3, while the latest version is 1.3.4.
WARNING:root:The OGB package is out of date. Your version is 1.3.3, while the latest version is 1.3.4.
Traceback (most recent call last):

  File "/opt/conda/envs/pytorch-ci/bin/dgl", line 33, in <module>
    sys.exit(load_entry_point('dglgo==0.0.2', 'console_scripts', 'dgl')())

  File "/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/dglgo-0.0.2-py3.7.egg/dglgo/cli/cli.py", line 21, in main
    app()

  File "/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/dglgo-0.0.2-py3.7.egg/dglgo/cli/train_cli.py", line 20, in train
    exec(code,  {'__name__': '__main__'})

  File "dglgo_tmp.py", line 248, in <module>

  File "dglgo_tmp.py", line 192, in main

  File "dglgo_tmp.py", line 156, in train

  File "/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/torch/nn/modules/module.py", line 1051, in _call_impl
    return forward_call(*input, **kwargs)

  File "dglgo_tmp.py", line 78, in forward

  File "/opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/torch/nn/modules/module.py", line 1051, in _call_impl
    return forward_call(*input, **kwargs)

  File "/root/jenkins/workspace/dgl_PR-4648@3/python/dgl/nn/pytorch/conv/graphconv.py", line 423, in forward
    graph.update_all(aggregate_fn, fn.sum(msg='m', out='h'))

  File "/root/jenkins/workspace/dgl_PR-4648@3/python/dgl/heterograph.py", line 4895, in update_all
    ndata = core.message_passing(g, message_func, reduce_func, apply_node_func)

  File "/root/jenkins/workspace/dgl_PR-4648@3/python/dgl/core.py", line 357, in message_passing
    ndata = invoke_gspmm(g, mfunc, rfunc)

  File "/root/jenkins/workspace/dgl_PR-4648@3/python/dgl/core.py", line 332, in invoke_gspmm
    z = op(graph, x)

  File "/root/jenkins/workspace/dgl_PR-4648@3/python/dgl/ops/spmm.py", line 189, in func
    return gspmm(g, 'copy_lhs', reduce_op, x, None)

  File "/root/jenkins/workspace/dgl_PR-4648@3/python/dgl/ops/spmm.py", line 77, in gspmm
    lhs_data, rhs_data)

  File "/root/jenkins/workspace/dgl_PR-4648@3/python/dgl/backend/pytorch/sparse.py", line 720, in gspmm
    with autocast(enabled=False):

TypeError: empty_context() got an unexpected keyword argument 'enabled'

=============================== warnings summary ===============================
python/dgl/backend/backend.py:1741
  /root/jenkins/workspace/dgl_PR-4648@3/python/dgl/backend/backend.py:1741: DeprecationWarning: invalid escape sequence \P
    """

python/dgl/backend/pytorch/tensor.py:16
python/dgl/backend/pytorch/tensor.py:16
  /root/jenkins/workspace/dgl_PR-4648@3/python/dgl/backend/pytorch/tensor.py:16: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    if LooseVersion(th.__version__) < LooseVersion("1.9.0"):

python/dgl/backend/pytorch/tensor.py:340
python/dgl/backend/pytorch/tensor.py:340
  /root/jenkins/workspace/dgl_PR-4648@3/python/dgl/backend/pytorch/tensor.py:340: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    if LooseVersion(th.__version__) >= LooseVersion("1.10.0"):

python/dgl/dataloading/dataloader.py:33
  /root/jenkins/workspace/dgl_PR-4648@3/python/dgl/dataloading/dataloader.py:33: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    PYTORCH_VER = LooseVersion(torch.__version__)

python/dgl/_dataloading/pytorch/dataloader.py:23
  /root/jenkins/workspace/dgl_PR-4648@3/python/dgl/_dataloading/pytorch/dataloader.py:23: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    PYTORCH_VER = LooseVersion(th.__version__)

python/dgl/_dataloading/pytorch/dataloader.py:24
  /root/jenkins/workspace/dgl_PR-4648@3/python/dgl/_dataloading/pytorch/dataloader.py:24: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    PYTORCH_16 = PYTORCH_VER >= LooseVersion("1.6.0")

python/dgl/_dataloading/pytorch/dataloader.py:25
  /root/jenkins/workspace/dgl_PR-4648@3/python/dgl/_dataloading/pytorch/dataloader.py:25: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    PYTORCH_17 = PYTORCH_VER >= LooseVersion("1.7.0")

../../../../opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:10
  /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:10: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    _nlv = LooseVersion(_np_version)

../../../../opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:11
  /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:11: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    _np_version_under1p16 = _nlv < LooseVersion("1.16")

../../../../opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:12
  /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:12: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    _np_version_under1p17 = _nlv < LooseVersion("1.17")

../../../../opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:13
  /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:13: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    _np_version_under1p18 = _nlv < LooseVersion("1.18")

../../../../opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:14
  /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:14: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    _np_version_under1p19 = _nlv < LooseVersion("1.19")

../../../../opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:15
  /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:15: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    _np_version_under1p20 = _nlv < LooseVersion("1.20")

../../../../opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/setuptools/_distutils/version.py:351
  /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/setuptools/_distutils/version.py:351: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    other = LooseVersion(other)

../../../../opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/pandas/compat/numpy/function.py:125
../../../../opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/pandas/compat/numpy/function.py:125
  /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/pandas/compat/numpy/function.py:125: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    if LooseVersion(_np_version) >= LooseVersion("1.17.0"):

tests/go/test_model.py::test_gcn[g0]
  /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/dglgo-0.0.2-py3.7.egg/dglgo/model/node_encoder/gcn.py:59: DGLWarning: The embedding for node feature is used, and input node_feat is ignored, due to the provided embed_size.
    dgl_warning("The embedding for node feature is used, and input node_feat is ignored, due to the provided embed_size.")

tests/go/test_model.py::test_gat[g0]
  /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/dglgo-0.0.2-py3.7.egg/dglgo/model/node_encoder/gat.py:72: DGLWarning: The embedding for node feature is used, and input node_feat is ignored, due to the provided embed_size.
    "The embedding for node feature is used, and input node_feat is ignored, due to the provided embed_size.")

tests/go/test_model.py::test_gin[g0]
  /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/dglgo-0.0.2-py3.7.egg/dglgo/model/node_encoder/gin.py:52: DGLWarning: The embedding for node feature is used, and input node_feat is ignored, due to the provided embed_size.
    "The embedding for node feature is used, and input node_feat is ignored, due to the provided embed_size.")

tests/go/test_model.py::test_sage[g0]
  /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/dglgo-0.0.2-py3.7.egg/dglgo/model/node_encoder/sage.py:52: DGLWarning: The embedding for node feature is used, and input node_feat is ignored, due to the provided embed_size.
    dgl_warning("The embedding for node feature is used, and input node_feat is ignored, due to the provided embed_size.")

tests/go/test_model.py::test_sgc[g0]
  /opt/conda/envs/pytorch-ci/lib/python3.7/site-packages/dglgo-0.0.2-py3.7.egg/dglgo/model/node_encoder/sgc.py:42: DGLWarning: The embedding for node feature is used, and input node_feat is ignored, due to the provided embed_size.
    dgl_warning("The embedding for node feature is used, and input node_feat is ignored, due to the provided embed_size.")

-- Docs: https://docs.pytest.org/en/stable/how-to/capture-warnings.html
--- generated xml file: /root/jenkins/workspace/dgl_PR-4648@3/pytest_go.xml ----
============================ slowest 100 durations =============================
8.67s call     tests/go/test_pipeline.py::test_graphpred[pna-csv]
8.22s call     tests/go/test_pipeline.py::test_graphpred[pna-ogbg-molhiv]
8.08s call     tests/go/test_pipeline.py::test_graphpred[pna-ogbg-molpcba]
7.81s call     tests/go/test_pipeline.py::test_graphpred[gin-ogbg-molpcba]
7.79s call     tests/go/test_pipeline.py::test_nodepred_ns_data[citeseer]
7.73s call     tests/go/test_pipeline.py::test_graphpred[gin-ogbg-molhiv]
7.56s call     tests/go/test_pipeline.py::test_nodepred_data[csv]
7.53s call     tests/go/test_pipeline.py::test_nodepred_data[citeseer]
7.44s call     tests/go/test_pipeline.py::test_nodepred_model[gat]
7.37s call     tests/go/test_pipeline.py::test_nodepred_data[pubmed]
7.32s call     tests/go/test_pipeline.py::test_nodepred_ns_data[cora]
7.28s call     tests/go/test_pipeline.py::test_graphpred[gin-csv]
6.89s call     tests/go/test_pipeline.py::test_nodepred_ns_data[csv]
6.81s call     tests/go/test_pipeline.py::test_nodepred_model[sage]
6.80s call     tests/go/test_pipeline.py::test_nodepred_ns_data[ogbn-arxiv]
6.70s call     tests/go/test_pipeline.py::test_nodepred_ns_model[gat]
6.66s call     tests/go/test_pipeline.py::test_nodepred_ns_data[co-buy-computer]
6.62s call     tests/go/test_pipeline.py::test_nodepred_ns_data[pubmed]
6.62s call     tests/go/test_pipeline.py::test_nodepred_model[gcn]
6.60s call     tests/go/test_pipeline.py::test_nodepred_ns_model[gcn]
6.59s call     tests/go/test_pipeline.py::test_nodepred_model[gin]
6.59s call     tests/go/test_pipeline.py::test_nodepred_data[ogbn-arxiv]
6.55s call     tests/go/test_pipeline.py::test_nodepred_data[reddit]
6.55s call     tests/go/test_pipeline.py::test_nodepred_ns_data[reddit]
6.54s call     tests/go/test_pipeline.py::test_nodepred_data[cora]
6.52s call     tests/go/test_pipeline.py::test_nodepred_data[ogbn-products]
6.51s call     tests/go/test_pipeline.py::test_nodepred_ns_model[sage]
6.50s call     tests/go/test_pipeline.py::test_node_cora
6.50s call     tests/go/test_pipeline.py::test_nodepred_model[sgc]
6.47s call     tests/go/test_pipeline.py::test_nodepred_data[co-buy-computer]
6.47s call     tests/go/test_pipeline.py::test_nodepred_ns_data[ogbn-products]
5.51s call     tests/go/test_pipeline.py::test_linkpred_edge_model[bilinear]
5.42s call     tests/go/test_pipeline.py::test_linkpred_node_model[gin]
5.32s call     tests/go/test_pipeline.py::test_linkpred_neg_sampler[global]
5.32s call     tests/go/test_pipeline.py::test_linkpred_neg_sampler[persource]
5.26s call     tests/go/test_pipeline.py::test_linkpred_node_model[sage]
5.23s call     tests/go/test_pipeline.py::test_linkpred_data[ogbl-citation2]
5.22s call     tests/go/test_pipeline.py::test_linkpred_neg_sampler[]
5.19s call     tests/go/test_pipeline.py::test_linkpred_node_model[sgc]
5.13s call     tests/go/test_pipeline.py::test_linkpred_node_model[gat]
5.09s call     tests/go/test_pipeline.py::test_linkpred_data[ogbn-products]
5.09s call     tests/go/test_pipeline.py::test_linkpred_data[citeseer]
5.08s call     tests/go/test_pipeline.py::test_linkpred_data[ogbn-arxiv]
5.07s call     tests/go/test_pipeline.py::test_linkpred_data[pubmed]
5.02s call     tests/go/test_pipeline.py::test_linkpred_data[reddit]
5.00s call     tests/go/test_pipeline.py::test_linkpred_node_model[gcn]
4.99s call     tests/go/test_pipeline.py::test_linkpred_data[csv]
4.98s call     tests/go/test_pipeline.py::test_linkpred_data[cora]
4.96s call     tests/go/test_pipeline.py::test_linkpred_edge_model[ele]
4.96s call     tests/go/test_pipeline.py::test_linkpred_data[co-buy-computer]
4.91s call     tests/go/test_pipeline.py::test_linkpred_data[ogbl-collab]
3.02s call     tests/go/test_pipeline.py::test_recipe[nodepred-ns_product_sage.yaml]
2.79s call     tests/go/test_pipeline.py::test_recipe[nodepred-ns_arxiv_gcn.yaml]
2.71s call     tests/go/test_pipeline.py::test_recipe[nodepred_citeseer_sage.yaml]
2.45s call     tests/go/test_pipeline.py::test_recipe[graphpred_pcba_gin.yaml]
2.43s call     tests/go/test_pipeline.py::test_recipe[linkpred_collab_sage.yaml]
2.43s call     tests/go/test_pipeline.py::test_recipe[nodepred_cora_gcn.yaml]
2.41s call     tests/go/test_pipeline.py::test_recipe[nodepred_citeseer_gcn.yaml]
2.41s call     tests/go/test_pipeline.py::test_recipe[linkpred_cora_sage.yaml]
2.37s call     tests/go/test_pipeline.py::test_recipe[graphpred_hiv_gin.yaml]
2.35s call     tests/go/test_pipeline.py::test_recipe[nodepred_citeseer_gat.yaml]
2.35s call     tests/go/test_pipeline.py::test_recipe[nodepred_cora_sage.yaml]
2.34s call     tests/go/test_pipeline.py::test_recipe[nodepred_cora_gat.yaml]
2.31s call     tests/go/test_pipeline.py::test_recipe[nodepred_pubmed_sage.yaml]
2.30s call     tests/go/test_pipeline.py::test_recipe[graphpred_hiv_pna.yaml]
2.30s call     tests/go/test_pipeline.py::test_recipe[nodepred_pubmed_gat.yaml]
2.29s call     tests/go/test_pipeline.py::test_recipe[nodepred_pubmed_gcn.yaml]
2.22s call     tests/go/test_pipeline.py::test_recipe[linkpred_citation2_sage.yaml]
0.03s call     tests/go/test_model.py::test_gat_block[g0]
0.03s call     tests/go/test_model.py::test_gcn_block[g0]
0.03s call     tests/go/test_model.py::test_gcn[g0]
0.01s call     tests/go/test_model.py::test_ogbg_gin[True]
0.01s call     tests/go/test_model.py::test_gat[g0]
0.01s call     tests/go/test_model.py::test_bilinear

(26 durations < 0.005s hidden.  Use -vv to show these durations.)
=========================== short test summary info ============================
FAILED tests/go/test_model.py::test_gcn[g0] - TypeError: empty_context() got ...
FAILED tests/go/test_model.py::test_gcn_block[g0] - TypeError: empty_context(...
FAILED tests/go/test_model.py::test_gat[g0] - TypeError: empty_context() got ...
FAILED tests/go/test_model.py::test_gat_block[g0] - TypeError: empty_context(...
FAILED tests/go/test_model.py::test_gin[g0] - TypeError: empty_context() got ...
FAILED tests/go/test_model.py::test_sage[g0] - TypeError: empty_context() got...
FAILED tests/go/test_model.py::test_sage_block[g0] - TypeError: empty_context...
FAILED tests/go/test_model.py::test_sgc[g0] - TypeError: empty_context() got ...
FAILED tests/go/test_model.py::test_ogbg_gin[True] - TypeError: empty_context...
FAILED tests/go/test_model.py::test_ogbg_gin[False] - TypeError: empty_contex...
FAILED tests/go/test_model.py::test_pna - TypeError: empty_context() got an u...
FAILED tests/go/test_pipeline.py::test_node_cora - AssertionError: assert False
============ 12 failed, 69 passed, 23 warnings in 366.88s (0:06:06) ============
PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-v-shp2-g0] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-v-shp2-g1] PASSED  [ 94%]
tests/compute/test_specialization.py::test_v2v_update_all[int64] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-v-shp3-g0] FAIL: go
[Pipeline] }
[Pipeline] // timeout
[Pipeline] }
[Pipeline] // stage
Post stage
[Pipeline] cleanWs
[WS-CLEANUP] Deleting project workspace...
[WS-CLEANUP] Deferred wipeout is disabled by the job configuration...
PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-v-shp3-g1] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-v-shp4-g0] [WS-CLEANUP] done
[Pipeline] }
$ docker stop --time=1 39027e0569c91f60d197e31fb8f36454bc8a908bd28bdc225aa7c1adb21f2213
PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-v-shp4-g1] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-e-shp0-g0] PASSED [ 66%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-e-shp0-g1] PASSED  [ 94%]
tests/compute/test_specialization.py::test_v2v_snr[int32] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-e-shp1-g0] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-e-shp1-g1] $ docker rm -f 39027e0569c91f60d197e31fb8f36454bc8a908bd28bdc225aa7c1adb21f2213
PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-e-shp2-g0] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-e-shp2-g1] [Pipeline] // withDockerContainer
[Pipeline] }
[Pipeline] // withEnv
[Pipeline] }
[Pipeline] // node
[Pipeline] }
[Pipeline] // stage
[Pipeline] }
Failed in branch DGL-Go
PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-e-shp3-g0] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-e-shp3-g1] PASSED         [ 94%]
tests/compute/test_specialization.py::test_v2v_snr[int64] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-e-shp4-g0] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-u-e-shp4-g1] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-u-shp0-g0] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-u-shp0-g1] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-u-shp1-g0] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-u-shp1-g1] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-u-shp2-g0] PASSED         [ 95%]
tests/compute/test_specialization.py::test_v2v_pull[int32] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-u-shp2-g1] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-u-shp3-g0] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-u-shp3-g1] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-u-shp4-g0] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-u-shp4-g1] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-v-shp0-g0] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-v-shp0-g1] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-v-shp1-g0] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-v-shp1-g1] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-v-shp2-g0] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-v-shp2-g1] PASSED [ 67%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-v-shp3-g0] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-v-shp3-g1] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-v-shp4-g0] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-v-shp4-g1] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-e-shp0-g0] PASSED        [ 95%]
tests/compute/test_specialization.py::test_v2v_pull[int64] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-e-shp0-g1] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-e-shp1-g0] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-e-shp1-g1] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-e-shp2-g0] PASSED        [ 95%]
tests/compute/test_specialization.py::test_update_all_multi_fallback[int32] PASSED [ 95%]
tests/compute/test_specialization.py::test_update_all_multi_fallback[int64] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-e-shp2-g1] PASSED [ 95%]
tests/compute/test_specialization.py::test_pull_multi_fallback[int32] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-e-shp3-g0] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-e-shp3-g1] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-e-shp4-g0] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-v-e-shp4-g1] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-u-shp0-g0] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-u-shp0-g1] PASSED [ 95%]
tests/compute/test_specialization.py::test_pull_multi_fallback[int64] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-u-shp1-g0] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-u-shp1-g1] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-u-shp2-g0] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-u-shp2-g1] PASSED [ 95%]
tests/compute/test_specialization.py::test_spmv_3d_feat[int32] PASSED    [ 95%]
tests/compute/test_specialization.py::test_spmv_3d_feat[int64] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-u-shp3-g0] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-u-shp3-g1] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-u-shp4-g0] PASSED    [ 95%]
tests/compute/test_subgraph.py::test_edge_subgraph PASSED                [ 95%]
tests/compute/test_subgraph.py::test_subgraph PASSED                     [ 95%]
tests/compute/test_subgraph.py::test_subgraph_mask[int32] SKIPPED (M...) [ 95%]
tests/compute/test_subgraph.py::test_subgraph_mask[int64] SKIPPED (M...) [ 95%]
tests/compute/test_subgraph.py::test_subgraph1[int32] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-u-shp4-g1] PASSED             [ 95%]
tests/compute/test_subgraph.py::test_subgraph1[int64] PASSED             [ 95%]
tests/compute/test_subgraph.py::test_in_subgraph[int32] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-v-shp0-g0] PASSED           [ 95%]
tests/compute/test_subgraph.py::test_in_subgraph[int64] PASSED           [ 95%]
tests/compute/test_subgraph.py::test_out_subgraph[int32] PASSED [ 68%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-v-shp0-g1] PASSED          [ 95%]
tests/compute/test_subgraph.py::test_out_subgraph[int64] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-v-shp1-g0] PASSED          [ 95%]
tests/compute/test_subgraph.py::test_subgraph_message_passing PASSED     [ 95%]
tests/compute/test_subgraph.py::test_khop_in_subgraph[int32] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-v-shp1-g1] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-v-shp2-g0] PASSED      [ 95%]
tests/compute/test_subgraph.py::test_khop_in_subgraph[int64] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-v-shp2-g1] PASSED      [ 95%]
tests/compute/test_subgraph.py::test_khop_out_subgraph[int32] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-v-shp3-g0] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-v-shp3-g1] PASSED     [ 95%]
tests/compute/test_subgraph.py::test_khop_out_subgraph[int64] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-v-shp4-g0] PASSED     [ 95%]
tests/compute/test_subgraph.py::test_subframes[child_device0-parent_idx_device0] SKIPPED [ 96%]
tests/compute/test_subgraph.py::test_subframes[child_device0-parent_idx_device1] SKIPPED [ 96%]
tests/compute/test_subgraph.py::test_subframes[child_device0-parent_idx_device2] SKIPPED [ 96%]
tests/compute/test_subgraph.py::test_subframes[child_device0-parent_idx_device3] SKIPPED [ 96%]
tests/compute/test_subgraph.py::test_subframes[child_device1-parent_idx_device0] SKIPPED [ 96%]
tests/compute/test_subgraph.py::test_subframes[child_device1-parent_idx_device1] SKIPPED [ 96%]
tests/compute/test_subgraph.py::test_subframes[child_device1-parent_idx_device2] SKIPPED [ 96%]
tests/compute/test_subgraph.py::test_subframes[child_device1-parent_idx_device3] SKIPPED [ 96%]
tests/compute/test_subgraph.py::test_uva_subgraph[int32-device0] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-v-shp4-g1] SKIPPED [ 96%]
tests/compute/test_subgraph.py::test_uva_subgraph[int32-device1] SKIPPED [ 96%]
tests/compute/test_subgraph.py::test_uva_subgraph[int64-device0] SKIPPED [ 96%]
tests/compute/test_subgraph.py::test_uva_subgraph[int64-device1] SKIPPED [ 96%]
tests/compute/test_transform.py::test_line_graph1 PASSED                 [ 96%]
tests/compute/test_transform.py::test_line_graph2[int32] PASSED          [ 96%]
tests/compute/test_transform.py::test_line_graph2[int64] PASSED          [ 96%]
tests/compute/test_transform.py::test_no_backtracking PASSED             [ 96%]
tests/compute/test_transform.py::test_reverse[int32] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-e-shp0-g0] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-e-shp0-g1] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-e-shp1-g0] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-e-shp1-g1] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-e-shp2-g0] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-e-shp2-g1] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-e-shp3-g0] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-e-shp3-g1] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-e-shp4-g0] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-add-e-e-shp4-g1] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-u-shp0-g0] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-u-shp0-g1] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-u-shp1-g0] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-u-shp1-g1] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-u-shp2-g0] PASSED [ 69%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-u-shp2-g1] PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-u-shp3-g0] PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-u-shp3-g1] PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-u-shp4-g0] PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-u-shp4-g1] PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-v-shp0-g0] PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-v-shp0-g1] PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-v-shp1-g0] PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-v-shp1-g1] PASSED              [ 96%]
tests/compute/test_transform.py::test_reverse[int64] PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-v-shp2-g0] PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-v-shp2-g1] PASSED              [ 96%]
tests/compute/test_transform.py::test_reverse_shared_frames[int32] PASSED [ 96%]
tests/compute/test_transform.py::test_reverse_shared_frames[int64] PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-v-shp3-g0] PASSED [ 96%]
tests/compute/test_transform.py::test_to_bidirected PASSED               [ 96%]
tests/compute/test_transform.py::test_add_reverse_edges PASSED           [ 96%]
tests/compute/test_transform.py::test_simple_graph PASSED                [ 96%]
tests/compute/test_transform.py::test_khop_graph PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-v-shp3-g1] PASSED                  [ 96%]
tests/compute/test_transform.py::test_khop_adj PASSED                    [ 96%]
tests/compute/test_transform.py::test_laplacian_lambda_max PASSED        [ 97%]
tests/compute/test_transform.py::test_partition_with_halo PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-v-shp4-g0] PASSED         [ 97%]
tests/compute/test_transform.py::test_metis_partition[int32] PASSED      [ 97%]
tests/compute/test_transform.py::test_metis_partition[int64] PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-v-shp4-g1] PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-e-shp0-g0] PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-e-shp0-g1] PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-e-shp1-g0] PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-e-shp1-g1] PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-e-shp2-g0] PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-e-shp2-g1] PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-e-shp3-g0] PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-e-shp3-g1] PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-e-shp4-g0] PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-u-e-shp4-g1] PASSED [ 70%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-u-shp0-g0] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-u-shp0-g1] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-u-shp1-g0] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-u-shp1-g1] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-u-shp2-g0] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-u-shp2-g1] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-u-shp3-g0] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-u-shp3-g1] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-u-shp4-g0] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-u-shp4-g1] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-v-shp0-g0] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-v-shp0-g1] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-v-shp1-g0] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-v-shp1-g1] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-v-shp2-g0] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-v-shp2-g1] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-v-shp3-g0] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-v-shp3-g1] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-v-shp4-g0] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-v-shp4-g1] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-e-shp0-g0] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-e-shp0-g1] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-e-shp1-g0] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-e-shp1-g1] PASSED [ 71%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-e-shp2-g0] PASSED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-e-shp2-g1] PASSED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-e-shp3-g0] PASSED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-e-shp3-g1] PASSED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-e-shp4-g0] PASSED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-v-e-shp4-g1] PASSED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-u-shp0-g0] PASSED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-u-shp0-g1] PASSED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-u-shp1-g0] PASSED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-u-shp1-g1] PASSED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-u-shp2-g0] PASSED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-u-shp2-g1] PASSED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-u-shp3-g0] PASSED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-u-shp3-g1] PASSED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-u-shp4-g0] PASSED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-u-shp4-g1] PASSED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-v-shp0-g0] PASSED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-v-shp0-g1] PASSED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-v-shp1-g0] PASSED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-v-shp1-g1] PASSED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-v-shp2-g0] PASSED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-v-shp2-g1] PASSED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-v-shp3-g0] PASSED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-v-shp3-g1] PASSED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-v-shp4-g0] PASSED [ 72%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-v-shp4-g1] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-e-shp0-g0] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-e-shp0-g1] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-e-shp1-g0] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-e-shp1-g1] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-e-shp2-g0] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-e-shp2-g1] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-e-shp3-g0] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-e-shp3-g1] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-e-shp4-g0] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-sub-e-e-shp4-g1] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-u-shp0-g0] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-u-shp0-g1] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-u-shp1-g0] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-u-shp1-g1] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-u-shp2-g0] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-u-shp2-g1] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-u-shp3-g0] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-u-shp3-g1] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-u-shp4-g0] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-u-shp4-g1] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-v-shp0-g0] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-v-shp0-g1] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-v-shp1-g0] PASSED [ 73%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-v-shp1-g1] PASSED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-v-shp2-g0] PASSED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-v-shp2-g1] PASSED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-v-shp3-g0] PASSED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-v-shp3-g1] PASSED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-v-shp4-g0] PASSED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-v-shp4-g1] PASSED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-e-shp0-g0] PASSED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-e-shp0-g1] PASSED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-e-shp1-g0] PASSED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-e-shp1-g1] PASSED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-e-shp2-g0] PASSED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-e-shp2-g1] PASSED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-e-shp3-g0] PASSED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-e-shp3-g1] PASSED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-e-shp4-g0] PASSED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-u-e-shp4-g1] PASSED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-u-shp0-g0] PASSED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-u-shp0-g1] PASSED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-u-shp1-g0] PASSED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-u-shp1-g1] PASSED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-u-shp2-g0] PASSED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-u-shp2-g1] PASSED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-u-shp3-g0] PASSED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-u-shp3-g1] PASSED [ 74%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-u-shp4-g0] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-u-shp4-g1] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-v-shp0-g0] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-v-shp0-g1] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-v-shp1-g0] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-v-shp1-g1] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-v-shp2-g0] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-v-shp2-g1] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-v-shp3-g0] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-v-shp3-g1] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-v-shp4-g0] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-v-shp4-g1] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-e-shp0-g0] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-e-shp0-g1] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-e-shp1-g0] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-e-shp1-g1] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-e-shp2-g0] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-e-shp2-g1] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-e-shp3-g0] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-e-shp3-g1] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-e-shp4-g0] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-v-e-shp4-g1] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-u-shp0-g0] PASSED      [ 97%]
tests/compute/test_transform.py::test_reorder_nodes PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-u-shp0-g1] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-u-shp1-g0] PASSED [ 75%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-u-shp1-g1] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-u-shp2-g0] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-u-shp2-g1] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-u-shp3-g0] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-u-shp3-g1] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-u-shp4-g0] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-u-shp4-g1] PASSED               [ 97%]
tests/compute/test_transform.py::test_compact[int32] PASSED              [ 97%]
tests/compute/test_transform.py::test_compact[int64] PASSED              [ 97%]
tests/compute/test_transform.py::test_to_simple[int32] PASSED            [ 97%]
tests/compute/test_transform.py::test_to_simple[int64] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-v-shp0-g0] PASSED            [ 97%]
tests/compute/test_transform.py::test_to_block[int32] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-v-shp0-g1] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-v-shp1-g0] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-v-shp1-g1] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-v-shp2-g0] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-v-shp2-g1] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-v-shp3-g0] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-v-shp3-g1] PASSED             [ 97%]
tests/compute/test_transform.py::test_to_block[int64] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-v-shp4-g0] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-v-shp4-g1] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-e-shp0-g0] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-e-shp0-g1] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-e-shp1-g0] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-e-shp1-g1] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-e-shp2-g0] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-e-shp2-g1] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-e-shp3-g0] PASSED [ 76%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-e-shp3-g1] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-e-shp4-g0] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-mul-e-e-shp4-g1] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-u-shp0-g0] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-u-shp0-g1] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-u-shp1-g0] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-u-shp1-g1] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-u-shp2-g0] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-u-shp2-g1] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-u-shp3-g0] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-u-shp3-g1] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-u-shp4-g0] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-u-shp4-g1] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-v-shp0-g0] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-v-shp0-g1] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-v-shp1-g0] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-v-shp1-g1] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-v-shp2-g0] PASSED             [ 97%]
tests/compute/test_transform.py::test_remove_edges[int32] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-v-shp2-g1] PASSED         [ 97%]
tests/compute/test_transform.py::test_remove_edges[int64] PASSED         [ 97%]
tests/compute/test_transform.py::test_add_edges[int32] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-v-shp3-g0] PASSED            [ 97%]
tests/compute/test_transform.py::test_add_edges[int64] PASSED            [ 97%]
tests/compute/test_transform.py::test_add_nodes[int32] PASSED            [ 97%]
tests/compute/test_transform.py::test_add_nodes[int64] PASSED            [ 97%]
tests/compute/test_transform.py::test_remove_nodes[int32] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-v-shp3-g1] PASSED         [ 97%]
tests/compute/test_transform.py::test_remove_nodes[int64] PASSED         [ 97%]
tests/compute/test_transform.py::test_add_selfloop[int32] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-v-shp4-g0] PASSED         [ 97%]
tests/compute/test_transform.py::test_add_selfloop[int64] PASSED         [ 97%]
tests/compute/test_transform.py::test_remove_selfloop[int32] PASSED      [ 97%]
tests/compute/test_transform.py::test_remove_selfloop[int64] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-v-shp4-g1] PASSED      [ 97%]
tests/compute/test_transform.py::test_reorder_graph[int32] PASSED        [ 97%]
tests/compute/test_transform.py::test_reorder_graph[int64] PASSED        [ 98%]
tests/compute/test_transform.py::test_norm_by_dst[int32] PASSED          [ 98%]
tests/compute/test_transform.py::test_norm_by_dst[int64] PASSED          [ 98%]
tests/compute/test_transform.py::test_module_add_self_loop[int32] PASSED [ 98%]
tests/compute/test_transform.py::test_module_add_self_loop[int64] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-e-shp0-g0] PASSED [ 98%]
tests/compute/test_transform.py::test_module_remove_self_loop[int32] PASSED [ 98%]
tests/compute/test_transform.py::test_module_remove_self_loop[int64] PASSED [ 98%]
tests/compute/test_transform.py::test_module_add_reverse[int32] PASSED   [ 98%]
tests/compute/test_transform.py::test_module_add_reverse[int64] PASSED   [ 98%]
tests/compute/test_transform.py::test_module_to_simple[int32] PASSED     [ 98%]
tests/compute/test_transform.py::test_module_to_simple[int64] PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-e-shp0-g1] PASSED     [ 98%]
tests/compute/test_transform.py::test_module_line_graph[int32] PASSED    [ 98%]
tests/compute/test_transform.py::test_module_line_graph[int64] PASSED    [ 98%]
tests/compute/test_transform.py::test_module_khop_graph[int32] PASSED    [ 98%]
tests/compute/test_transform.py::test_module_khop_graph[int64] PASSED    [ 98%]
tests/compute/test_transform.py::test_module_add_metapaths[int32] PASSED [ 98%]
tests/compute/test_transform.py::test_module_add_metapaths[int64] PASSED [ 98%]
tests/compute/test_transform.py::test_module_compose[int32] PASSED       [ 98%]
tests/compute/test_transform.py::test_module_compose[int64] PASSED       [ 98%]PASSED [ 77%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-e-shp1-g0] 
tests/compute/test_transform.py::test_module_gcnnorm[int32] PASSED       [ 98%]
tests/compute/test_transform.py::test_module_gcnnorm[int64] PASSED       [ 98%]
tests/compute/test_transform.py::test_module_ppr[int32] SKIPPED (Onl...) [ 98%]
tests/compute/test_transform.py::test_module_ppr[int64] SKIPPED (Onl...) [ 98%]
tests/compute/test_transform.py::test_module_heat_kernel[int32] SKIPPED  [ 98%]
tests/compute/test_transform.py::test_module_heat_kernel[int64] SKIPPED  [ 98%]
tests/compute/test_transform.py::test_module_gdc[int32] SKIPPED (Onl...) [ 99%]
tests/compute/test_transform.py::test_module_gdc[int64] SKIPPED (Onl...) [ 99%]
tests/compute/test_transform.py::test_module_node_shuffle[int32] PASSED  [ 99%]
tests/compute/test_transform.py::test_module_node_shuffle[int64] PASSED  [ 99%]
tests/compute/test_transform.py::test_module_drop_node[int32] SKIPPED    [ 99%]
tests/compute/test_transform.py::test_module_drop_node[int64] SKIPPED    [ 99%]
tests/compute/test_transform.py::test_module_drop_edge[int32] SKIPPED    [ 99%]
tests/compute/test_transform.py::test_module_drop_edge[int64] SKIPPED    [ 99%]
tests/compute/test_transform.py::test_module_add_edge[int32] PASSED      [ 99%]
tests/compute/test_transform.py::test_module_add_edge[int64] PASSED      [ 99%]
tests/compute/test_transform.py::test_module_random_walk_pe[int32] PASSED [ 99%]
tests/compute/test_transform.py::test_module_random_walk_pe[int64] PASSED [ 99%]
tests/compute/test_transform.py::test_module_laplacian_pe[int32] PASSED  [ 99%]
tests/compute/test_transform.py::test_module_laplacian_pe[int64] PASSED  [ 99%]
tests/compute/test_transform.py::test_module_sign[g0] SKIPPED (Only ...) [ 99%]
tests/compute/test_transform.py::test_module_row_feat_normalizer[int32] SKIPPED [ 99%]
tests/compute/test_transform.py::test_module_row_feat_normalizer[int64] SKIPPED [ 99%]
tests/compute/test_transform.py::test_module_feat_mask[int32] SKIPPED    [ 99%]
tests/compute/test_transform.py::test_module_feat_mask[int64] SKIPPED    [ 99%]
tests/compute/test_traversal.py::test_bfs[int32] PASSED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-e-shp1-g1] PASSED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-e-shp2-g0] PASSED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-e-shp2-g1] PASSED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-e-shp3-g0] PASSED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-e-shp3-g1] PASSED                  [ 99%]
tests/compute/test_traversal.py::test_bfs[int64] PASSED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-e-shp4-g0] PASSED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-u-e-shp4-g1] PASSED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-u-shp0-g0] PASSED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-u-shp0-g1] PASSED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-u-shp1-g0] PASSED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-u-shp1-g1] PASSED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-u-shp2-g0] PASSED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-u-shp2-g1] PASSED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-u-shp3-g0] PASSED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-u-shp3-g1] PASSED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-u-shp4-g0] PASSED                  [ 99%]
tests/compute/test_traversal.py::test_topological_nodes[int32] PASSED    [ 99%]
tests/compute/test_traversal.py::test_topological_nodes[int64] PASSED    [ 99%]
tests/compute/test_traversal.py::test_dfs_labeled_edges[int32] PASSED    [ 99%]
tests/compute/test_traversal.py::test_dfs_labeled_edges[int64] PASSED    [100%]

=============================== warnings summary ===============================
python/dgl/backend/backend.py:1741
  /root/jenkins/workspace/dgl_PR-4648@4/python/dgl/backend/backend.py:1741: DeprecationWarning: invalid escape sequence \P
    """

tests/compute/test_basics.py: 2 warnings
tests/compute/test_filter.py: 1 warning
tests/compute/test_graph.py: 9 warnings
tests/compute/test_kernel.py: 3 warnings
tests/compute/test_propagate.py: 2 warnings
tests/compute/test_removal.py: 16 warnings
tests/compute/test_serialize.py: 3 warnings
tests/compute/test_specialization.py: 12 warnings
tests/compute/test_subgraph.py: 2 warnings
tests/compute/test_transform.py: 10 warnings
tests/compute/test_traversal.py: 2 warnings
  /root/jenkins/workspace/dgl_PR-4648@4/python/dgl/heterograph.py:72: DGLWarning: Recommend creating graphs by `dgl.graph(data)` instead of `dgl.DGLGraph(data)`.
    dgl_warning('Recommend creating graphs by `dgl.graph(data)`'

tests/compute/test_basics.py: 2 warnings
tests/compute/test_batched_graph.py: 10 warnings
tests/compute/test_graph.py: 2 warnings
tests/compute/test_kernel.py: 1 warning
tests/compute/test_propagate.py: 2 warnings
tests/compute/test_removal.py: 10 warnings
tests/compute/test_specialization.py: 10 warnings
tests/compute/test_subgraph.py: 2 warnings
  /root/jenkins/workspace/dgl_PR-4648@4/python/dgl/heterograph.py:354: DGLWarning: DGLGraph.add_edge is deprecated. Please use DGLGraph.add_edges
    dgl_warning("DGLGraph.add_edge is deprecated. Please use DGLGraph.add_edges")

tests/compute/test_basics.py::test_update_all_0deg[int32]
tests/compute/test_basics.py::test_update_all_0deg[int64]
tests/compute/test_basics.py::test_pull_0deg[int32]
tests/compute/test_basics.py::test_pull_0deg[int64]
tests/compute/test_propagate.py::test_prop_nodes_topo[int32]
tests/compute/test_propagate.py::test_prop_nodes_topo[int64]
  /root/jenkins/workspace/dgl_PR-4648@4/python/dgl/core.py:79: DGLWarning: The input graph for the user-defined edge function does not contain valid edges
    dgl_warning('The input graph for the user-defined edge function ' \

tests/compute/test_batched_graph.py::test_batched_edge_ordering[int32]
tests/compute/test_batched_graph.py::test_batched_edge_ordering[int64]
tests/compute/test_graph.py::test_query
tests/compute/test_graph.py::test_hypersparse_query
tests/compute/test_transform.py::test_no_backtracking
tests/compute/test_transform.py::test_reverse[int32]
tests/compute/test_transform.py::test_reverse[int64]
  /root/jenkins/workspace/dgl_PR-4648@4/python/dgl/heterograph.py:2978: DGLWarning: DGLGraph.edge_id is deprecated. Please use DGLGraph.edge_ids.
    dgl_warning("DGLGraph.edge_id is deprecated. Please use DGLGraph.edge_ids.")

tests/compute/test_batched_heterograph.py::test_features[int32]
tests/compute/test_batched_heterograph.py::test_features[int64]
  /root/jenkins/workspace/dgl_PR-4648@4/python/dgl/batch.py:159: DGLWarning: Arguments edge_attrs has been deprecated. Please use edata instead.
    dgl_warning('Arguments edge_attrs has been deprecated. Please use'

tests/compute/test_csrmm.py::test_csrmm[float32-int32]
tests/compute/test_csrmm.py::test_csrmm[float32-int64]
tests/compute/test_csrmm.py::test_csrmm[float64-int32]
tests/compute/test_csrmm.py::test_csrmm[float64-int64]
tests/compute/test_csrmm.py::test_csrsum[float32-int32]
tests/compute/test_csrmm.py::test_csrsum[float32-int64]
tests/compute/test_csrmm.py::test_csrsum[float64-int32]
tests/compute/test_csrmm.py::test_csrsum[float64-int64]
  /root/jenkins/workspace/dgl_PR-4648@4/python/dgl/heterograph_index.py:797: FutureWarning: Adjacency matrix by default currently returns edge IDs.  As a result there is one 0 entry which is not eliminated.  In the next release it will return 1s by default, and 0 will be eliminated otherwise.
    FutureWarning)

tests/compute/test_graph.py::test_query
tests/compute/test_graph.py::test_hypersparse_query
tests/compute/test_heterograph.py::test_query[int32]
tests/compute/test_heterograph.py::test_query[int64]
  /root/jenkins/workspace/dgl_PR-4648@4/python/dgl/heterograph.py:2753: DGLWarning: DGLGraph.has_node is deprecated. Please use DGLGraph.has_nodes
    dgl_warning("DGLGraph.has_node is deprecated. Please use DGLGraph.has_nodes")

tests/compute/test_graph.py::test_query
tests/compute/test_graph.py::test_hypersparse_query
  /root/jenkins/workspace/dgl_PR-4648@4/python/dgl/heterograph.py:2687: DGLWarning: DGLGraph.__contains__ is deprecated. Please directly call has_nodes.
    dgl_warning('DGLGraph.__contains__ is deprecated.'

tests/compute/test_graph.py::test_query
tests/compute/test_sampling.py::test_non_uniform_random_walk[False]
tests/compute/test_sampling.py::test_uniform_random_walk[False]
tests/compute/test_sampling.py::test_node2vec
tests/compute/test_transform.py::test_no_backtracking
  /root/jenkins/workspace/dgl_PR-4648@4/python/dgl/heterograph.py:2851: DGLWarning: DGLGraph.has_edge_between is deprecated. Please use DGLGraph.has_edges_between
    dgl_warning("DGLGraph.has_edge_between is deprecated. "

tests/compute/test_graph.py::test_query
tests/compute/test_graph.py::test_hypersparse_query
  /root/jenkins/workspace/dgl_PR-4648@4/python/dgl/heterograph.py:3432: DGLWarning: DGLGraph.in_degree is deprecated. Please use DGLGraph.in_degrees
    dgl_warning("DGLGraph.in_degree is deprecated. Please use DGLGraph.in_degrees")

tests/compute/test_graph.py::test_query
tests/compute/test_graph.py::test_hypersparse_query
  /root/jenkins/workspace/dgl_PR-4648@4/python/dgl/heterograph.py:3516: DGLWarning: DGLGraph.out_degree is deprecated. Please use DGLGraph.out_degrees
    dgl_warning("DGLGraph.out_degree is deprecated. Please use DGLGraph.out_degrees")

tests/compute/test_graph.py::test_query
  /root/jenkins/workspace/dgl_PR-4648@4/python/dgl/heterograph.py:84: DGLWarning: Keyword arguments ['readonly', 'sort_csr'] are deprecated in v0.5, and can be safely removed in all cases.
    ' removed in all cases.'.format(list(deprecate_kwargs.keys())))

tests/compute/test_heterograph.py: 20 warnings
  /root/jenkins/workspace/dgl_PR-4648@4/tests/compute/test_heterograph.py:1128: DeprecationWarning: np.asscalar(a) is deprecated since NumPy v1.16, use a.item() instead
    assert np.asscalar(F.asnumpy(src_i)) == nid[src[i]]

tests/compute/test_heterograph.py: 20 warnings
  /root/jenkins/workspace/dgl_PR-4648@4/tests/compute/test_heterograph.py:1129: DeprecationWarning: np.asscalar(a) is deprecated since NumPy v1.16, use a.item() instead
    assert np.asscalar(F.asnumpy(dst_i)) == nid[dst[i]]

tests/compute/test_heterograph.py::test_invertible_conversion[int32]
tests/compute/test_heterograph.py::test_invertible_conversion[int64]
tests/compute/test_shared_mem.py::test_single_process[int32]
tests/compute/test_shared_mem.py::test_single_process[int64]
  /root/jenkins/workspace/dgl_PR-4648@4/python/dgl/heterograph.py:2635: DGLWarning: DGLGraph.is_readonly is deprecated in v0.5.
  DGLGraph now always supports mutable operations like add_nodes and add_edges.
    dgl_warning('DGLGraph.is_readonly is deprecated in v0.5.\n'

tests/compute/test_kernel.py: 3 warnings
tests/compute/test_sparse.py: 290 warnings
  /root/jenkins/workspace/dgl_PR-4648@4/python/dgl/backend/mxnet/sparse.py:17: DGLWarning: MXNet do not support scatter_add, fallback to numpy.
    dgl_warning("MXNet do not support scatter_add, fallback to numpy.")

tests/compute/test_pickle.py::test_pickling_batched_heterograph
  /root/jenkins/workspace/dgl_PR-4648@4/python/dgl/batch.py:511: DGLWarning: From v0.5, DGLHeteroGraph is merged into DGLGraph. You can safely replace dgl.batch_hetero with dgl.batch
    dgl_warning('From v0.5, DGLHeteroGraph is merged into DGLGraph. You can safely'

tests/compute/test_sampler.py::test_create_full
tests/compute/test_sampler.py::test_1neighbor_sampler_all
tests/compute/test_sampler.py::test_1neighbor_sampler
tests/compute/test_sampler.py::test_prefetch_neighbor_sampler
tests/compute/test_sampler.py::test_10neighbor_sampler_all
tests/compute/test_sampler.py::test_10neighbor_sampler
tests/compute/test_sampler.py::test_nonuniform_neighbor_sampler
tests/compute/test_sampler.py::test_setseed
  /root/jenkins/workspace/dgl_PR-4648@4/python/dgl/contrib/sampling/sampler.py:317: DGLWarning: dgl.contrib.sampling.NeighborSampler is deprecated starting from v0.5. Please read our guide<link> for how to use the new sampling APIs.
    dgl_warning('dgl.contrib.sampling.NeighborSampler is deprecated starting from v0.5.'

tests/compute/test_sampler.py::test_create_full
tests/compute/test_sampler.py::test_1neighbor_sampler_all
tests/compute/test_sampler.py::test_1neighbor_sampler
tests/compute/test_sampler.py::test_prefetch_neighbor_sampler
tests/compute/test_sampler.py::test_10neighbor_sampler_all
tests/compute/test_sampler.py::test_10neighbor_sampler
tests/compute/test_sampler.py::test_layer_sampler
tests/compute/test_sampler.py::test_nonuniform_neighbor_sampler
tests/compute/test_sampler.py::test_setseed
  /root/jenkins/workspace/dgl_PR-4648@4/python/dgl/_deprecate/nodeflow.py:99: DGLWarning: NodeFlow APIs are deprecated starting from v0.5. Please read our guide<link> for how to use the new sampling APIs.
    dgl_warning('NodeFlow APIs are deprecated starting from v0.5. Please read our'

tests/compute/test_sampling.py::test_global_uniform_negative_sampling[int32]
tests/compute/test_sampling.py::test_global_uniform_negative_sampling[int64]
  /root/jenkins/workspace/dgl_PR-4648@4/python/dgl/_ffi/_ctypes/function.py:116: ComplexWarning: Casting complex values to real discards the imaginary part
    values[i].v_float64 = arg

tests/compute/test_serialize.py::test_graph_serialize_with_feature[False]
tests/compute/test_serialize.py::test_graph_serialize_without_feature[False]
tests/compute/test_serialize.py::test_graph_serialize_with_labels[False]
tests/compute/test_transform.py::test_simple_graph
  /root/jenkins/workspace/dgl_PR-4648@4/python/dgl/heterograph.py:84: DGLWarning: Keyword arguments ['readonly'] are deprecated in v0.5, and can be safely removed in all cases.
    ' removed in all cases.'.format(list(deprecate_kwargs.keys())))

tests/compute/test_serialize.py::test_load_old_files1
tests/compute/test_serialize.py::test_load_old_files2
  /root/jenkins/workspace/dgl_PR-4648@4/python/dgl/data/graph_serialize.py:179: DGLWarning: You are loading a graph file saved by old version of dgl.              Please consider saving it again with the current format.
    Please consider saving it again with the current format.")

tests/compute/test_transform.py::test_reverse_shared_frames[int32]
tests/compute/test_transform.py::test_reverse_shared_frames[int64]
  /root/jenkins/workspace/dgl_PR-4648@4/python/dgl/transforms/functional.py:1267: DGLWarning: share_ndata argument has been renamed to copy_ndata.
    dgl_warning('share_ndata argument has been renamed to copy_ndata.')

tests/compute/test_transform.py::test_reverse_shared_frames[int32]
tests/compute/test_transform.py::test_reverse_shared_frames[int64]
  /root/jenkins/workspace/dgl_PR-4648@4/python/dgl/transforms/functional.py:1270: DGLWarning: share_edata argument has been renamed to copy_edata.
    dgl_warning('share_edata argument has been renamed to copy_edata.')

tests/compute/test_transform.py::test_simple_graph
  /root/jenkins/workspace/dgl_PR-4648@4/python/dgl/transforms/functional.py:1319: DGLWarning: dgl.to_simple_graph is renamed to dgl.to_simple in v0.5.
    dgl_warning('dgl.to_simple_graph is renamed to dgl.to_simple in v0.5.')

-- Docs: https://docs.pytest.org/en/stable/how-to/capture-warnings.html
- generated xml file: /root/jenkins/workspace/dgl_PR-4648@4/pytest_compute.xml -
============================ slowest 100 durations =============================
133.49s call     tests/compute/test_kernel.py::test_all_binary_builtins
62.39s call     tests/compute/test_sampler.py::test_negative_sampler
20.05s call     tests/compute/test_transform.py::test_metis_partition[int64]
6.88s call     tests/compute/test_heterograph.py::test_query[int64]
5.04s call     tests/compute/test_kernel.py::test_copy_src_reduce
4.42s call     tests/compute/test_sampling.py::test_non_uniform_random_walk[False]
4.36s call     tests/compute/test_heterograph.py::test_level2[int64]
4.31s call     tests/compute/test_transform.py::test_reorder_nodes
4.04s call     tests/compute/test_graph.py::test_query
3.73s call     tests/compute/test_kernel.py::test_copy_edge_reduce
3.07s call     tests/compute/test_sampling.py::test_sample_neighbors_outedge
2.76s call     tests/compute/test_sampling.py::test_uniform_random_walk[False]
2.62s call     tests/compute/test_heterograph.py::test_query[int32]
2.58s call     tests/compute/test_heterograph.py::test_level2[int32]
2.47s call     tests/compute/test_specialization.py::test_v2v_snr[int32]
2.44s call     tests/compute/test_heterograph.py::test_updates[int64]
2.36s call     tests/compute/test_heterograph.py::test_view1[int32]
2.13s call     tests/compute/test_heterograph.py::test_view1[int64]
2.11s call     tests/compute/test_sampling.py::test_sample_neighbors_prob
2.09s call     tests/compute/test_sampling.py::test_sample_neighbors_noprob
2.08s call     tests/compute/test_heterograph.py::test_updates[int32]
2.04s call     tests/compute/test_traversal.py::test_bfs[int64]
1.94s call     tests/compute/test_specialization.py::test_v2v_update_all[int64]
1.93s call     tests/compute/test_transform.py::test_to_block[int32]
1.86s call     tests/compute/test_specialization.py::test_v2v_snr[int64]
1.78s call     tests/compute/test_transform.py::test_to_block[int64]
1.63s call     tests/compute/test_sparse.py::test_spmm[int64-sum-copy_lhs-shp0-g1]
1.51s call     tests/compute/test_sampler.py::test_10neighbor_sampler
1.48s call     tests/compute/test_specialization.py::test_pull_multi_fallback[int32]
1.39s call     tests/compute/test_traversal.py::test_bfs[int32]
1.25s call     tests/compute/test_heterograph.py::test_flatten[int64]
1.21s call     tests/compute/test_removal.py::test_node_and_edge_removal[int64]
1.14s call     tests/compute/test_specialization.py::test_v2v_pull[int32]
1.06s call     tests/compute/test_sampler.py::test_prefetch_neighbor_sampler
0.96s call     tests/compute/test_specialization.py::test_v2v_update_all[int32]
0.96s call     tests/compute/test_sparse.py::test_spmm[int32-sum-copy_lhs-shp2-g1]
0.92s call     tests/compute/test_sampling.py::test_node2vec
0.92s call     tests/compute/test_removal.py::test_node_and_edge_removal[int32]
0.91s call     tests/compute/test_sparse.py::test_spmm[int64-sum-mul-shp2-g1]
0.90s call     tests/compute/test_specialization.py::test_v2v_pull[int64]
0.89s call     tests/compute/test_sparse.py::test_spmm[int32-sum-copy_lhs-shp1-g1]
0.87s call     tests/compute/test_batched_graph.py::test_batch_propagate[int64]
0.87s call     tests/compute/test_sparse.py::test_spmm[int32-sum-copy_lhs-shp3-g1]
0.85s call     tests/compute/test_sampler.py::test_1neighbor_sampler
0.84s call     tests/compute/test_sparse.py::test_sddmm[int32-mul-u-v-shp3-g1]
0.76s call     tests/compute/test_sampler.py::test_1neighbor_sampler_all
0.76s call     tests/compute/test_heterograph.py::test_convert[int32]
0.74s call     tests/compute/test_sparse.py::test_spmm[int64-sum-copy_lhs-shp2-g1]
0.72s call     tests/compute/test_sparse.py::test_spmm[int32-max-sub-shp2-g1]
0.69s call     tests/compute/test_sparse.py::test_spmm[int64-sum-mul-shp5-g1]
0.69s call     tests/compute/test_sparse.py::test_spmm[int32-max-div-shp2-g1]
0.69s call     tests/compute/test_sparse.py::test_spmm[int32-max-add-shp3-g1]
0.68s call     tests/compute/test_transform.py::test_reverse[int32]
0.68s call     tests/compute/test_sparse.py::test_spmm[int32-max-sub-shp0-g1]
0.67s call     tests/compute/test_specialization.py::test_pull_multi_fallback[int64]
0.66s call     tests/compute/test_heterograph.py::test_flatten[int32]
0.65s call     tests/compute/test_shared_mem.py::test_multi_process[int64]
0.64s call     tests/compute/test_propagate.py::test_prop_edges_dfs[int64]
0.64s call     tests/compute/test_sparse.py::test_spmm[int64-sum-copy_lhs-shp3-g1]
0.63s call     tests/compute/test_sparse.py::test_spmm[int64-sum-mul-shp2-g0]
0.63s call     tests/compute/test_sparse.py::test_spmm[int32-max-add-shp2-g1]
0.63s call     tests/compute/test_sparse.py::test_spmm[int32-max-mul-shp2-g1]
0.62s call     tests/compute/test_sparse.py::test_spmm[int32-max-add-shp4-g1]
0.61s call     tests/compute/test_sparse.py::test_spmm[int32-min-sub-shp4-g1]
0.60s call     tests/compute/test_sparse.py::test_spmm[int32-min-add-shp4-g1]
0.60s call     tests/compute/test_sparse.py::test_spmm[int32-min-sub-shp2-g1]
0.60s call     tests/compute/test_sparse.py::test_spmm[int32-max-mul-shp0-g1]
0.60s call     tests/compute/test_sparse.py::test_spmm[int32-max-sub-shp1-g1]
0.60s call     tests/compute/test_sparse.py::test_spmm[int64-max-sub-shp0-g1]
0.60s call     tests/compute/test_sparse.py::test_spmm[int64-sum-add-shp2-g1]
0.59s call     tests/compute/test_sparse.py::test_spmm[int32-min-div-shp4-g1]
0.59s call     tests/compute/test_sparse.py::test_spmm[int32-max-mul-shp1-g1]
0.59s call     tests/compute/test_sparse.py::test_spmm[int64-sum-div-shp2-g1]
0.59s call     tests/compute/test_sparse.py::test_spmm[int32-sum-add-shp0-g1]
0.59s call     tests/compute/test_sparse.py::test_spmm[int32-max-div-shp0-g1]
0.58s call     tests/compute/test_sparse.py::test_spmm[int32-min-div-shp3-g1]
0.58s call     tests/compute/test_propagate.py::test_prop_edges_dfs[int32]
0.58s call     tests/compute/test_sparse.py::test_spmm[int32-sum-add-shp2-g0]
0.58s call     tests/compute/test_sparse.py::test_spmm[int32-max-add-shp1-g1]
0.58s call     tests/compute/test_sparse.py::test_spmm[int32-min-add-shp2-g1]
0.58s call     tests/compute/test_sparse.py::test_spmm[int64-min-add-shp2-g1]
0.57s call     tests/compute/test_sparse.py::test_spmm[int32-max-mul-shp5-g1]
0.57s call     tests/compute/test_sparse.py::test_spmm[int64-max-mul-shp1-g1]
0.57s call     tests/compute/test_sparse.py::test_spmm[int32-min-sub-shp3-g1]
0.56s call     tests/compute/test_heterograph.py::test_convert[int64]
0.56s call     tests/compute/test_sparse.py::test_spmm[int64-sum-div-shp4-g1]
0.55s call     tests/compute/test_sparse.py::test_spmm[int32-min-mul-shp0-g1]
0.55s call     tests/compute/test_sparse.py::test_spmm[int32-sum-mul-shp2-g1]
0.55s call     tests/compute/test_sparse.py::test_spmm[int64-min-add-shp5-g1]
0.55s call     tests/compute/test_sparse.py::test_spmm[int64-min-mul-shp5-g1]
0.55s call     tests/compute/test_sparse.py::test_spmm[int64-max-copy_lhs-shp4-g1]
0.55s call     tests/compute/test_sparse.py::test_spmm[int64-min-mul-shp4-g1]
0.55s call     tests/compute/test_sparse.py::test_spmm[int32-min-sub-shp5-g1]
0.55s call     tests/compute/test_sparse.py::test_spmm[int32-min-sub-shp1-g1]
0.55s call     tests/compute/test_sparse.py::test_spmm[int32-max-div-shp1-g1]
0.55s call     tests/compute/test_sparse.py::test_spmm[int64-sum-div-shp0-g1]
0.55s call     tests/compute/test_sparse.py::test_sddmm[int64-copy_rhs-u-e-shp2-g1]
0.54s call     tests/compute/test_batched_graph.py::test_batch_propagate[int32]
0.54s call     tests/compute/test_basics.py::test_update_routines[int32]
0.54s call     tests/compute/test_batched_graph.py::test_batch_send_and_recv[int32]
========= 2284 passed, 169 skipped, 509 warnings in 684.92s (0:11:24) ==========
Error in sys.excepthook:

Original exception was:
Error in sys.excepthook:

Original exception was:
Error in sys.excepthook:

Original exception was:
Error in sys.excepthook:

Original exception was:
Error in sys.excepthook:

Original exception was:
Error in sys.excepthook:

Original exception was:
Error in sys.excepthook:

Original exception was:
Error in sys.excepthook:

Original exception was:
Error in sys.excepthook:

Original exception was:
Error in sys.excepthook:

Original exception was:
Error in sys.excepthook:

Original exception was:
Error in sys.excepthook:

Original exception was:
PASSED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-u-shp4-g1] Error in sys.excepthook:

Original exception was:
Error in sys.excepthook:

Original exception was:
Error in sys.excepthook:

Original exception was:
Error in sys.excepthook:

Original exception was:
PASSED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-v-shp0-g0] PASSED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-v-shp0-g1] PASSED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-v-shp1-g0] PASSED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-v-shp1-g1] PASSED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-v-shp2-g0] PASSED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-v-shp2-g1] PASSED [ 78%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-v-shp3-g0] PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-v-shp3-g1] PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-v-shp4-g0] PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-v-shp4-g1] PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-e-shp0-g0] ============================= test session starts ==============================
platform linux -- Python 3.6.9, pytest-7.0.1, pluggy-1.0.0 -- /opt/conda/envs/mxnet-ci/bin/python3
cachedir: .pytest_cache
rootdir: /root/jenkins/workspace/dgl_PR-4648@4
collecting ... PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-e-shp0-g1] PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-e-shp1-g0] PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-e-shp1-g1] PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-e-shp2-g0] PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-e-shp2-g1] PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-e-shp3-g0] PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-e-shp3-g1] PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-e-shp4-g0] PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-v-e-shp4-g1] PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-u-shp0-g0] PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-u-shp0-g1] PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-u-shp1-g0] PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-u-shp1-g1] PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-u-shp2-g0] collected 806 items

tests/mxnet/test_geometry.py::test_fps PASSED                            [  0%]
tests/mxnet/test_nn.py::test_graph_conv[1-int32] PASSED                  [  0%]
tests/mxnet/test_nn.py::test_graph_conv[1-int64] PASSED                  [  0%]
tests/mxnet/test_nn.py::test_graph_conv[2-int32] PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-u-shp2-g1] PASSED                  [  0%]
tests/mxnet/test_nn.py::test_graph_conv[2-int64] PASSED                  [  0%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-none-g0-int32] PASSED [  0%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-none-g0-int64] PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-u-shp3-g0] PASSED [  0%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-none-g1-int32] PASSED [  0%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-none-g1-int64] PASSED [  1%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-none-g2-int32] PASSED [  1%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-none-g2-int64] PASSED [  1%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-none-g3-int32] PASSED [  1%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-none-g3-int64] PASSED [  1%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-none-g4-int32] PASSED [  1%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-none-g4-int64] PASSED [  1%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-none-g5-int32] PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-u-shp3-g1] PASSED [  1%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-none-g5-int64] PASSED [  2%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-none-g6-int32] PASSED [  2%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-none-g6-int64] PASSED [  2%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-both-g0-int32] PASSED [  2%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-both-g0-int64] PASSED [  2%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-both-g1-int32] PASSED [  2%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-both-g1-int64] PASSED [  2%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-both-g2-int32] PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-u-shp4-g0] PASSED [  2%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-both-g2-int64] PASSED [  3%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-both-g3-int32] PASSED [  3%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-both-g3-int64] PASSED [  3%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-both-g4-int32] PASSED [  3%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-both-g4-int64] PASSED [  3%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-both-g5-int32] PASSED [  3%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-both-g5-int64] PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-u-shp4-g1] PASSED [  3%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-both-g6-int32] PASSED [  3%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-both-g6-int64] PASSED [  4%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-right-g0-int32] PASSED [  4%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-right-g0-int64] PASSED [  4%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-right-g1-int32] PASSED [  4%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-right-g1-int64] PASSED [  4%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-right-g2-int32] PASSED [  4%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-right-g2-int64] PASSED [  4%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-right-g3-int32] PASSED [  4%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-right-g3-int64] PASSED [  5%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-right-g4-int32] PASSED [  5%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-right-g4-int64] PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-v-shp0-g0] PASSED [  5%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-right-g5-int32] PASSED [  5%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-right-g5-int64] PASSED [  5%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-right-g6-int32] PASSED [  5%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-right-g6-int64] PASSED [  5%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-left-g0-int32] PASSED [  5%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-left-g0-int64] PASSED [  6%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-left-g1-int32] PASSED [  6%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-left-g1-int64] PASSED [  6%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-left-g2-int32] PASSED [  6%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-left-g2-int64] PASSED [  6%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-left-g3-int32] PASSED [  6%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-left-g3-int64] PASSED [  6%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-left-g4-int32] PASSED [  6%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-left-g4-int64] PASSED [  7%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-left-g5-int32] PASSED [  7%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-left-g5-int64] PASSED [  7%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-left-g6-int32] PASSED [  7%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-True-left-g6-int64] PASSED [  7%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-none-g0-int32] PASSED [  7%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-none-g0-int64] PASSED [  7%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-none-g1-int32] PASSED [  7%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-none-g1-int64] PASSED [  8%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-none-g2-int32] PASSED [  8%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-none-g2-int64] PASSED [  8%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-none-g3-int32] PASSED [  8%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-none-g3-int64] PASSED [  8%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-none-g4-int32] PASSED [  8%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-none-g4-int64] PASSED [  8%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-none-g5-int32] PASSED [  8%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-none-g5-int64] PASSED [  9%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-none-g6-int32] PASSED [  9%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-none-g6-int64] PASSED [  9%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-both-g0-int32] PASSED [  9%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-both-g0-int64] PASSED [ 79%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-v-shp0-g1] PASSED [  9%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-both-g1-int32] PASSED [  9%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-both-g1-int64] PASSED [  9%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-both-g2-int32] PASSED [  9%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-both-g2-int64] PASSED [ 10%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-both-g3-int32] PASSED [ 10%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-both-g3-int64] PASSED [ 10%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-both-g4-int32] PASSED [ 10%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-both-g4-int64] PASSED [ 10%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-both-g5-int32] PASSED [ 10%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-both-g5-int64] PASSED [ 10%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-both-g6-int32] PASSED [ 10%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-both-g6-int64] PASSED [ 11%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-right-g0-int32] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-v-shp1-g0] PASSED [ 11%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-right-g0-int64] PASSED [ 11%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-right-g1-int32] PASSED [ 11%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-right-g1-int64] PASSED [ 11%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-right-g2-int32] PASSED [ 11%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-right-g2-int64] PASSED [ 11%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-right-g3-int32] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-v-shp1-g1] PASSED [ 11%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-right-g3-int64] PASSED [ 12%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-right-g4-int32] PASSED [ 12%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-right-g4-int64] PASSED [ 12%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-right-g5-int32] PASSED [ 12%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-right-g5-int64] PASSED [ 12%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-right-g6-int32] PASSED [ 12%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-right-g6-int64] PASSED [ 12%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-left-g0-int32] PASSED [ 12%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-left-g0-int64] PASSED [ 13%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-left-g1-int32] PASSED [ 13%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-left-g1-int64] PASSED [ 13%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-left-g2-int32] PASSED [ 13%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-left-g2-int64] PASSED [ 13%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-left-g3-int32] PASSED [ 13%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-left-g3-int64] PASSED [ 13%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-left-g4-int32] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-v-shp2-g0] PASSED [ 13%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-left-g4-int64] PASSED [ 14%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-left-g5-int32] PASSED [ 14%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-left-g5-int64] PASSED [ 14%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-left-g6-int32] PASSED [ 14%]
tests/mxnet/test_nn.py::test_graph_conv2[1-False-False-left-g6-int64] PASSED [ 14%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-none-g0-int32] PASSED [ 14%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-none-g0-int64] PASSED [ 14%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-none-g1-int32] PASSED [ 14%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-none-g1-int64] PASSED [ 15%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-none-g2-int32] PASSED [ 15%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-none-g2-int64] PASSED [ 15%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-none-g3-int32] PASSED [ 15%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-none-g3-int64] PASSED [ 15%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-none-g4-int32] PASSED [ 15%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-none-g4-int64] PASSED [ 15%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-none-g5-int32] PASSED [ 15%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-none-g5-int64] PASSED [ 16%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-none-g6-int32] PASSED [ 16%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-none-g6-int64] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-v-shp2-g1] PASSED [ 16%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-both-g0-int32] PASSED [ 16%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-both-g0-int64] PASSED [ 16%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-both-g1-int32] PASSED [ 16%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-both-g1-int64] PASSED [ 16%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-both-g2-int32] PASSED [ 16%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-both-g2-int64] PASSED [ 16%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-both-g3-int32] PASSED [ 17%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-both-g3-int64] PASSED [ 17%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-both-g4-int32] PASSED [ 17%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-both-g4-int64] PASSED [ 17%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-both-g5-int32] PASSED [ 17%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-both-g5-int64] PASSED [ 17%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-both-g6-int32] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-v-shp3-g0] PASSED [ 17%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-both-g6-int64] PASSED [ 17%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-right-g0-int32] PASSED [ 18%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-right-g0-int64] PASSED [ 18%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-right-g1-int32] PASSED [ 18%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-right-g1-int64] PASSED [ 18%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-right-g2-int32] PASSED [ 18%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-right-g2-int64] PASSED [ 18%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-right-g3-int32] PASSED [ 18%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-right-g3-int64] PASSED [ 18%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-right-g4-int32] PASSED [ 19%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-right-g4-int64] PASSED [ 19%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-right-g5-int32] PASSED [ 19%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-right-g5-int64] PASSED [ 19%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-right-g6-int32] PASSED [ 19%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-right-g6-int64] PASSED [ 19%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-left-g0-int32] PASSED [ 19%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-left-g0-int64] PASSED [ 19%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-left-g1-int32] PASSED [ 20%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-left-g1-int64] PASSED [ 20%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-left-g2-int32] PASSED [ 20%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-left-g2-int64] PASSED [ 20%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-left-g3-int32] PASSED [ 20%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-left-g3-int64] PASSED [ 20%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-left-g4-int32] PASSED [ 20%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-left-g4-int64] PASSED [ 20%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-left-g5-int32] PASSED [ 21%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-left-g5-int64] PASSED [ 21%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-left-g6-int32] PASSED [ 21%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-True-left-g6-int64] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-v-shp3-g1] PASSED [ 21%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-none-g0-int32] PASSED [ 21%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-none-g0-int64] PASSED [ 21%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-none-g1-int32] PASSED [ 21%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-none-g1-int64] PASSED [ 21%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-none-g2-int32] PASSED [ 22%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-none-g2-int64] PASSED [ 22%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-none-g3-int32] PASSED [ 22%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-none-g3-int64] PASSED [ 22%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-none-g4-int32] PASSED [ 22%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-none-g4-int64] PASSED [ 22%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-none-g5-int32] PASSED [ 22%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-none-g5-int64] PASSED [ 22%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-none-g6-int32] PASSED [ 23%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-none-g6-int64] PASSED [ 23%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-both-g0-int32] PASSED [ 23%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-both-g0-int64] PASSED [ 23%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-both-g1-int32] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-v-shp4-g0] PASSED [ 23%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-both-g1-int64] PASSED [ 23%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-both-g2-int32] PASSED [ 23%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-both-g2-int64] PASSED [ 23%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-both-g3-int32] PASSED [ 24%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-both-g3-int64] PASSED [ 24%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-both-g4-int32] PASSED [ 24%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-both-g4-int64] PASSED [ 24%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-both-g5-int32] PASSED [ 24%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-both-g5-int64] PASSED [ 24%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-both-g6-int32] PASSED [ 24%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-both-g6-int64] PASSED [ 24%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-right-g0-int32] PASSED [ 25%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-right-g0-int64] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-v-shp4-g1] PASSED [ 25%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-right-g1-int32] PASSED [ 25%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-right-g1-int64] PASSED [ 25%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-right-g2-int32] PASSED [ 25%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-right-g2-int64] PASSED [ 25%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-right-g3-int32] PASSED [ 25%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-right-g3-int64] PASSED [ 25%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-right-g4-int32] PASSED [ 26%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-right-g4-int64] PASSED [ 26%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-right-g5-int32] PASSED [ 26%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-right-g5-int64] PASSED [ 26%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-right-g6-int32] PASSED [ 26%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-right-g6-int64] PASSED [ 26%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-left-g0-int32] PASSED [ 26%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-left-g0-int64] PASSED [ 26%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-left-g1-int32] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-e-shp0-g0] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-e-shp0-g1] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-e-shp1-g0] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-e-shp1-g1] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-e-shp2-g0] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-e-shp2-g1] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-e-shp3-g0] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-e-shp3-g1] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-e-shp4-g0] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-div-e-e-shp4-g1] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-u-shp0-g0] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-u-shp0-g1] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-u-shp1-g0] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-u-shp1-g1] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-u-shp2-g0] PASSED [ 80%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-u-shp2-g1] PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-u-shp3-g0] PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-u-shp3-g1] PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-u-shp4-g0] PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-u-shp4-g1] PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-v-shp0-g0] PASSED [ 27%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-left-g1-int64] PASSED [ 27%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-left-g2-int32] PASSED [ 27%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-left-g2-int64] PASSED [ 27%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-left-g3-int32] PASSED [ 27%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-left-g3-int64] PASSED [ 27%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-left-g4-int32] PASSED [ 27%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-left-g4-int64] PASSED [ 27%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-left-g5-int32] PASSED [ 28%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-left-g5-int64] PASSED [ 28%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-left-g6-int32] PASSED [ 28%]
tests/mxnet/test_nn.py::test_graph_conv2[2-False-False-left-g6-int64] PASSED [ 28%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[1-False-True-none-g0-int32] PASSED [ 28%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[1-False-True-none-g0-int64] PASSED [ 28%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[1-False-True-none-g1-int32] PASSED [ 28%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[1-False-True-none-g1-int64] PASSED [ 28%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[1-False-True-both-g0-int32] PASSED [ 29%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[1-False-True-both-g0-int64] PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-v-shp0-g1] PASSED [ 29%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[1-False-True-both-g1-int32] PASSED [ 29%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[1-False-True-both-g1-int64] PASSED [ 29%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[1-False-True-right-g0-int32] PASSED [ 29%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[1-False-True-right-g0-int64] PASSED [ 29%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[1-False-True-right-g1-int32] PASSED [ 29%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[1-False-True-right-g1-int64] PASSED [ 29%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[1-False-False-none-g0-int32] PASSED [ 30%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[1-False-False-none-g0-int64] PASSED [ 30%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[1-False-False-none-g1-int32] PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-v-shp1-g0] PASSED [ 30%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[1-False-False-none-g1-int64] PASSED [ 30%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[1-False-False-both-g0-int32] PASSED [ 30%]PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-v-shp1-g1] 
tests/mxnet/test_nn.py::test_graph_conv2_bi[1-False-False-both-g0-int64] PASSED [ 30%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[1-False-False-both-g1-int32] PASSED [ 30%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[1-False-False-both-g1-int64] PASSED [ 30%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[1-False-False-right-g0-int32] PASSED [ 31%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[1-False-False-right-g0-int64] PASSED [ 31%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[1-False-False-right-g1-int32] PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-v-shp2-g0] PASSED [ 31%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[1-False-False-right-g1-int64] PASSED [ 31%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[2-False-True-none-g0-int32] PASSED [ 31%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[2-False-True-none-g0-int64] PASSED [ 31%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[2-False-True-none-g1-int32] PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-v-shp2-g1] PASSED [ 31%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[2-False-True-none-g1-int64] PASSED [ 31%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[2-False-True-both-g0-int32] PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-v-shp3-g0] PASSED [ 32%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[2-False-True-both-g0-int64] PASSED [ 32%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[2-False-True-both-g1-int32] PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-v-shp3-g1] PASSED [ 32%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[2-False-True-both-g1-int64] PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-v-shp4-g0] PASSED [ 32%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[2-False-True-right-g0-int32] PASSED [ 32%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[2-False-True-right-g0-int64] PASSED [ 32%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[2-False-True-right-g1-int32] PASSED [ 32%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[2-False-True-right-g1-int64] PASSED [ 32%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[2-False-False-none-g0-int32] PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-v-shp4-g1] PASSED [ 33%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[2-False-False-none-g0-int64] PASSED [ 33%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[2-False-False-none-g1-int32] PASSED [ 33%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[2-False-False-none-g1-int64] PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-e-shp0-g0] PASSED [ 33%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[2-False-False-both-g0-int32] PASSED [ 33%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[2-False-False-both-g0-int64] PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-e-shp0-g1] PASSED [ 33%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[2-False-False-both-g1-int32] PASSED [ 33%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[2-False-False-both-g1-int64] PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-e-shp1-g0] PASSED [ 33%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[2-False-False-right-g0-int32] PASSED [ 33%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[2-False-False-right-g0-int64] PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-e-shp1-g1] PASSED [ 34%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[2-False-False-right-g1-int32] PASSED [ 34%]
tests/mxnet/test_nn.py::test_graph_conv2_bi[2-False-False-right-g1-int64] PASSED [ 34%]
tests/mxnet/test_nn.py::test_tagconv[1] PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-e-shp2-g0] PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-e-shp2-g1] PASSED                           [ 34%]
tests/mxnet/test_nn.py::test_tagconv[2] PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-e-shp3-g0] PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-e-shp3-g1] PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-e-shp4-g0] PASSED                           [ 34%]
tests/mxnet/test_nn.py::test_gat_conv[1-1-g0-int32] PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-u-e-shp4-g1] PASSED [ 81%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-u-shp0-g0] PASSED               [ 34%]
tests/mxnet/test_nn.py::test_gat_conv[1-1-g0-int64] PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-u-shp0-g1] PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-u-shp1-g0] PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-u-shp1-g1] PASSED               [ 34%]
tests/mxnet/test_nn.py::test_gat_conv[1-1-g1-int32] PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-u-shp2-g0] PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-u-shp2-g1] PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-u-shp3-g0] PASSED               [ 34%]
tests/mxnet/test_nn.py::test_gat_conv[1-1-g1-int64] PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-u-shp3-g1] PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-u-shp4-g0] PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-u-shp4-g1] PASSED               [ 35%]
tests/mxnet/test_nn.py::test_gat_conv[1-1-g2-int32] PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-v-shp0-g0] PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-v-shp0-g1] PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-v-shp1-g0] PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-v-shp1-g1] PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-v-shp2-g0] PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-v-shp2-g1] PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-v-shp3-g0] PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-v-shp3-g1] PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-v-shp4-g0] PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-v-shp4-g1] PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-e-shp0-g0] PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-e-shp0-g1] PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-e-shp1-g0] PASSED               [ 35%]
tests/mxnet/test_nn.py::test_gat_conv[1-1-g2-int64] PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-e-shp1-g1] PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-e-shp2-g0] PASSED [ 82%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-e-shp2-g1] PASSED [ 83%]PASSED               [ 35%]
tests/mxnet/test_nn.py::test_gat_conv[1-1-g3-int32] 
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-e-shp3-g0] PASSED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-e-shp3-g1] PASSED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-e-shp4-g0] PASSED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-v-e-shp4-g1] PASSED               [ 35%]
tests/mxnet/test_nn.py::test_gat_conv[1-1-g3-int64] PASSED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-u-shp0-g0] PASSED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-u-shp0-g1] PASSED               [ 35%]
tests/mxnet/test_nn.py::test_gat_conv[1-1-g4-int32] PASSED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-u-shp1-g0] PASSED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-u-shp1-g1] PASSED               [ 35%]
tests/mxnet/test_nn.py::test_gat_conv[1-1-g4-int64] PASSED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-u-shp2-g0] PASSED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-u-shp2-g1] PASSED               [ 35%]
tests/mxnet/test_nn.py::test_gat_conv[1-1-g5-int32] PASSED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-u-shp3-g0] PASSED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-u-shp3-g1] PASSED               [ 35%]
tests/mxnet/test_nn.py::test_gat_conv[1-1-g5-int64] PASSED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-u-shp4-g0] PASSED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-u-shp4-g1] PASSED               [ 36%]
tests/mxnet/test_nn.py::test_gat_conv[1-1-g6-int32] PASSED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-v-shp0-g0] PASSED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-v-shp0-g1] PASSED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-v-shp1-g0] PASSED               [ 36%]
tests/mxnet/test_nn.py::test_gat_conv[1-1-g6-int64] PASSED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-v-shp1-g1] PASSED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-v-shp2-g0] PASSED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-v-shp2-g1] PASSED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-v-shp3-g0] PASSED               [ 36%]
tests/mxnet/test_nn.py::test_gat_conv[1-20-g0-int32] PASSED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-v-shp3-g1] PASSED              [ 36%]
tests/mxnet/test_nn.py::test_gat_conv[1-20-g0-int64] PASSED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-v-shp4-g0] PASSED [ 83%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-v-shp4-g1] PASSED              [ 36%]
tests/mxnet/test_nn.py::test_gat_conv[1-20-g1-int32] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-e-shp0-g0] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-e-shp0-g1] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-e-shp1-g0] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-e-shp1-g1] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-e-shp2-g0] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-e-shp2-g1] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-e-shp3-g0] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-e-shp3-g1] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-e-shp4-g0] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-dot-e-e-shp4-g1] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-u-shp0-g0] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-u-shp0-g1] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-u-shp1-g0] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-u-shp1-g1] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-u-shp2-g0] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-u-shp2-g1] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-u-shp3-g0] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-u-shp3-g1] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-u-shp4-g0] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-u-shp4-g1] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-v-shp0-g0] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-v-shp0-g1] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-v-shp1-g0] PASSED              [ 36%]
tests/mxnet/test_nn.py::test_gat_conv[1-20-g1-int64] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-v-shp1-g1] PASSED [ 84%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-v-shp2-g0] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-v-shp2-g1] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-v-shp3-g0] PASSED              [ 36%]
tests/mxnet/test_nn.py::test_gat_conv[1-20-g2-int32] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-v-shp3-g1] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-v-shp4-g0] PASSED              [ 36%]
tests/mxnet/test_nn.py::test_gat_conv[1-20-g2-int64] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-v-shp4-g1] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-e-shp0-g0] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-e-shp0-g1] PASSED              [ 37%]
tests/mxnet/test_nn.py::test_gat_conv[1-20-g3-int32] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-e-shp1-g0] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-e-shp1-g1] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-e-shp2-g0] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-e-shp2-g1] PASSED              [ 37%]
tests/mxnet/test_nn.py::test_gat_conv[1-20-g3-int64] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-e-shp3-g0] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-e-shp3-g1] PASSED              [ 37%]
tests/mxnet/test_nn.py::test_gat_conv[1-20-g4-int32] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-e-shp4-g0] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-u-e-shp4-g1] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-u-shp0-g0] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-u-shp0-g1] PASSED              [ 37%]
tests/mxnet/test_nn.py::test_gat_conv[1-20-g4-int64] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-u-shp1-g0] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-u-shp1-g1] PASSED              [ 37%]
tests/mxnet/test_nn.py::test_gat_conv[1-20-g5-int32] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-u-shp2-g0] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-u-shp2-g1] PASSED              [ 37%]
tests/mxnet/test_nn.py::test_gat_conv[1-20-g5-int64] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-u-shp3-g0] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-u-shp3-g1] PASSED [ 85%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-u-shp4-g0] PASSED              [ 37%]
tests/mxnet/test_nn.py::test_gat_conv[1-20-g6-int32] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-u-shp4-g1] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-v-shp0-g0] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-v-shp0-g1] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-v-shp1-g0] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-v-shp1-g1] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-v-shp2-g0] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-v-shp2-g1] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-v-shp3-g0] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-v-shp3-g1] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-v-shp4-g0] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-v-shp4-g1] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-e-shp0-g0] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-e-shp0-g1] PASSED              [ 37%]
tests/mxnet/test_nn.py::test_gat_conv[1-20-g6-int64] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-e-shp1-g0] PASSED              [ 38%]
tests/mxnet/test_nn.py::test_gat_conv[5-1-g0-int32] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-e-shp1-g1] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-e-shp2-g0] PASSED               [ 38%]
tests/mxnet/test_nn.py::test_gat_conv[5-1-g0-int64] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-e-shp2-g1] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-e-shp3-g0] PASSED               [ 38%]
tests/mxnet/test_nn.py::test_gat_conv[5-1-g1-int32] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-e-shp3-g1] PASSED               [ 38%]
tests/mxnet/test_nn.py::test_gat_conv[5-1-g1-int64] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-e-shp4-g0] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-v-e-shp4-g1] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-u-shp0-g0] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-u-shp0-g1] PASSED               [ 38%]
tests/mxnet/test_nn.py::test_gat_conv[5-1-g2-int32] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-u-shp1-g0] PASSED               [ 38%]
tests/mxnet/test_nn.py::test_gat_conv[5-1-g2-int64] PASSED [ 86%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-u-shp1-g1] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-u-shp2-g0] PASSED               [ 38%]
tests/mxnet/test_nn.py::test_gat_conv[5-1-g3-int32] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-u-shp2-g1] PASSED               [ 38%]
tests/mxnet/test_nn.py::test_gat_conv[5-1-g3-int64] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-u-shp3-g0] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-u-shp3-g1] PASSED               [ 39%]
tests/mxnet/test_nn.py::test_gat_conv[5-1-g4-int32] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-u-shp4-g0] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-u-shp4-g1] PASSED               [ 39%]
tests/mxnet/test_nn.py::test_gat_conv[5-1-g4-int64] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-v-shp0-g0] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-v-shp0-g1] PASSED               [ 39%]
tests/mxnet/test_nn.py::test_gat_conv[5-1-g5-int32] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-v-shp1-g0] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-v-shp1-g1] PASSED               [ 39%]
tests/mxnet/test_nn.py::test_gat_conv[5-1-g5-int64] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-v-shp2-g0] PASSED               [ 39%]
tests/mxnet/test_nn.py::test_gat_conv[5-1-g6-int32] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-v-shp2-g1] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-v-shp3-g0] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-v-shp3-g1] PASSED               [ 39%]
tests/mxnet/test_nn.py::test_gat_conv[5-1-g6-int64] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-v-shp4-g0] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-v-shp4-g1] PASSED               [ 39%]
tests/mxnet/test_nn.py::test_gat_conv[5-20-g0-int32] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-e-shp0-g0] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-e-shp0-g1] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-e-shp1-g0] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-e-shp1-g1] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-e-shp2-g0] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-e-shp2-g1] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-e-shp3-g0] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-e-shp3-g1] PASSED [ 87%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-e-shp4-g0] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_lhs-e-e-shp4-g1] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-u-shp0-g0] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-u-shp0-g1] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-u-shp1-g0] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-u-shp1-g1] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-u-shp2-g0] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-u-shp2-g1] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-u-shp3-g0] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-u-shp3-g1] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-u-shp4-g0] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-u-shp4-g1] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-v-shp0-g0] PASSED              [ 39%]
tests/mxnet/test_nn.py::test_gat_conv[5-20-g0-int64] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-v-shp0-g1] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-v-shp1-g0] PASSED              [ 40%]
tests/mxnet/test_nn.py::test_gat_conv[5-20-g1-int32] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-v-shp1-g1] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-v-shp2-g0] PASSED              [ 40%]
tests/mxnet/test_nn.py::test_gat_conv[5-20-g1-int64] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-v-shp2-g1] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-v-shp3-g0] PASSED              [ 40%]
tests/mxnet/test_nn.py::test_gat_conv[5-20-g2-int32] PASSED              [ 40%]
tests/mxnet/test_nn.py::test_gat_conv[5-20-g2-int64] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-v-shp3-g1] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-v-shp4-g0] PASSED              [ 40%]
tests/mxnet/test_nn.py::test_gat_conv[5-20-g3-int32] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-v-shp4-g1] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-e-shp0-g0] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-e-shp0-g1] PASSED              [ 40%]
tests/mxnet/test_nn.py::test_gat_conv[5-20-g3-int64] PASSED [ 88%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-e-shp1-g0] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-e-shp1-g1] PASSED              [ 40%]
tests/mxnet/test_nn.py::test_gat_conv[5-20-g4-int32] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-e-shp2-g0] PASSED              [ 40%]
tests/mxnet/test_nn.py::test_gat_conv[5-20-g4-int64] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-e-shp2-g1] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-e-shp3-g0] PASSED              [ 41%]
tests/mxnet/test_nn.py::test_gat_conv[5-20-g5-int32] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-e-shp3-g1] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-e-shp4-g0] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-u-e-shp4-g1] PASSED              [ 41%]
tests/mxnet/test_nn.py::test_gat_conv[5-20-g5-int64] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-u-shp0-g0] PASSED              [ 41%]
tests/mxnet/test_nn.py::test_gat_conv[5-20-g6-int32] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-u-shp0-g1] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-u-shp1-g0] PASSED              [ 41%]
tests/mxnet/test_nn.py::test_gat_conv[5-20-g6-int64] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-u-shp1-g1] PASSED              [ 41%]
tests/mxnet/test_nn.py::test_gat_conv_bi[1-1-g0-int32] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-u-shp2-g0] PASSED            [ 41%]
tests/mxnet/test_nn.py::test_gat_conv_bi[1-1-g0-int64] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-u-shp2-g1] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-u-shp3-g0] PASSED            [ 41%]
tests/mxnet/test_nn.py::test_gat_conv_bi[1-1-g1-int32] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-u-shp3-g1] PASSED            [ 41%]
tests/mxnet/test_nn.py::test_gat_conv_bi[1-1-g1-int64] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-u-shp4-g0] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-u-shp4-g1] PASSED            [ 42%]
tests/mxnet/test_nn.py::test_gat_conv_bi[1-2-g0-int32] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-v-shp0-g0] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-v-shp0-g1] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-v-shp1-g0] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-v-shp1-g1] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-v-shp2-g0] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-v-shp2-g1] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-v-shp3-g0] PASSED [ 89%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-v-shp3-g1] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-v-shp4-g0] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-v-shp4-g1] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-e-shp0-g0] PASSED            [ 42%]
tests/mxnet/test_nn.py::test_gat_conv_bi[1-2-g0-int64] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-e-shp0-g1] PASSED            [ 42%]
tests/mxnet/test_nn.py::test_gat_conv_bi[1-2-g1-int32] PASSED            [ 42%]
tests/mxnet/test_nn.py::test_gat_conv_bi[1-2-g1-int64] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-e-shp1-g0] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-e-shp1-g1] PASSED            [ 42%]
tests/mxnet/test_nn.py::test_gat_conv_bi[4-1-g0-int32] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-e-shp2-g0] PASSED            [ 42%]
tests/mxnet/test_nn.py::test_gat_conv_bi[4-1-g0-int64] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-e-shp2-g1] PASSED            [ 42%]
tests/mxnet/test_nn.py::test_gat_conv_bi[4-1-g1-int32] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-e-shp3-g0] PASSED            [ 42%]
tests/mxnet/test_nn.py::test_gat_conv_bi[4-1-g1-int64] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-e-shp3-g1] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-e-shp4-g0] PASSED            [ 43%]
tests/mxnet/test_nn.py::test_gat_conv_bi[4-2-g0-int32] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-v-e-shp4-g1] PASSED            [ 43%]
tests/mxnet/test_nn.py::test_gat_conv_bi[4-2-g0-int64] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-u-shp0-g0] PASSED            [ 43%]
tests/mxnet/test_nn.py::test_gat_conv_bi[4-2-g1-int32] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-u-shp0-g1] PASSED            [ 43%]
tests/mxnet/test_nn.py::test_gat_conv_bi[4-2-g1-int64] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-u-shp1-g0] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-u-shp1-g1] PASSED            [ 43%]
tests/mxnet/test_nn.py::test_sage_conv[1-mean-g0-int32] PASSED           [ 43%]
tests/mxnet/test_nn.py::test_sage_conv[1-mean-g0-int64] PASSED           [ 43%]
tests/mxnet/test_nn.py::test_sage_conv[1-mean-g1-int32] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-u-shp2-g0] PASSED           [ 43%]
tests/mxnet/test_nn.py::test_sage_conv[1-mean-g1-int64] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-u-shp2-g1] PASSED           [ 44%]
tests/mxnet/test_nn.py::test_sage_conv[1-mean-g2-int32] PASSED           [ 44%]
tests/mxnet/test_nn.py::test_sage_conv[1-mean-g2-int64] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-u-shp3-g0] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-u-shp3-g1] PASSED           [ 44%]
tests/mxnet/test_nn.py::test_sage_conv[1-mean-g3-int32] PASSED           [ 44%]
tests/mxnet/test_nn.py::test_sage_conv[1-mean-g3-int64] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-u-shp4-g0] PASSED           [ 44%]
tests/mxnet/test_nn.py::test_sage_conv[1-mean-g4-int32] PASSED           [ 44%]
tests/mxnet/test_nn.py::test_sage_conv[1-mean-g4-int64] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-u-shp4-g1] PASSED           [ 44%]
tests/mxnet/test_nn.py::test_sage_conv[1-mean-g5-int32] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-v-shp0-g0] PASSED           [ 44%]
tests/mxnet/test_nn.py::test_sage_conv[1-mean-g5-int64] PASSED           [ 45%]
tests/mxnet/test_nn.py::test_sage_conv[1-mean-g6-int32] PASSED [ 90%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-v-shp0-g1] PASSED           [ 45%]
tests/mxnet/test_nn.py::test_sage_conv[1-mean-g6-int64] PASSED           [ 45%]
tests/mxnet/test_nn.py::test_sage_conv[1-mean-g7-int32] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-v-shp1-g0] PASSED           [ 45%]
tests/mxnet/test_nn.py::test_sage_conv[1-mean-g7-int64] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-v-shp1-g1] PASSED           [ 45%]
tests/mxnet/test_nn.py::test_sage_conv[1-pool-g0-int32] PASSED           [ 45%]
tests/mxnet/test_nn.py::test_sage_conv[1-pool-g0-int64] PASSED           [ 45%]
tests/mxnet/test_nn.py::test_sage_conv[1-pool-g1-int32] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-v-shp2-g0] PASSED           [ 45%]
tests/mxnet/test_nn.py::test_sage_conv[1-pool-g1-int64] PASSED           [ 46%]
tests/mxnet/test_nn.py::test_sage_conv[1-pool-g2-int32] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-v-shp2-g1] PASSED           [ 46%]
tests/mxnet/test_nn.py::test_sage_conv[1-pool-g2-int64] PASSED           [ 46%]
tests/mxnet/test_nn.py::test_sage_conv[1-pool-g3-int32] PASSED           [ 46%]
tests/mxnet/test_nn.py::test_sage_conv[1-pool-g3-int64] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-v-shp3-g0] PASSED           [ 46%]
tests/mxnet/test_nn.py::test_sage_conv[1-pool-g4-int32] PASSED           [ 46%]
tests/mxnet/test_nn.py::test_sage_conv[1-pool-g4-int64] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-v-shp3-g1] PASSED           [ 46%]
tests/mxnet/test_nn.py::test_sage_conv[1-pool-g5-int32] PASSED           [ 46%]
tests/mxnet/test_nn.py::test_sage_conv[1-pool-g5-int64] PASSED           [ 47%]
tests/mxnet/test_nn.py::test_sage_conv[1-pool-g6-int32] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-v-shp4-g0] PASSED           [ 47%]
tests/mxnet/test_nn.py::test_sage_conv[1-pool-g6-int64] PASSED           [ 47%]
tests/mxnet/test_nn.py::test_sage_conv[1-pool-g7-int32] PASSED           [ 47%]
tests/mxnet/test_nn.py::test_sage_conv[1-pool-g7-int64] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-v-shp4-g1] PASSED           [ 47%]
tests/mxnet/test_nn.py::test_sage_conv[1-gcn-g0-int32] PASSED            [ 47%]
tests/mxnet/test_nn.py::test_sage_conv[1-gcn-g0-int64] PASSED            [ 47%]
tests/mxnet/test_nn.py::test_sage_conv[1-gcn-g1-int32] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-e-shp0-g0] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-e-shp0-g1] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-e-shp1-g0] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-e-shp1-g1] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-e-shp2-g0] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-e-shp2-g1] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-e-shp3-g0] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-e-shp3-g1] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-e-shp4-g0] PASSED [ 91%]
tests/compute/test_sparse.py::test_sddmm[idtype1-copy_rhs-e-e-shp4-g1] PASSED [ 91%]
tests/compute/test_sparse.py::test_edge_softmax[idtype0-shp0-src-g0] PASSED            [ 47%]
tests/mxnet/test_nn.py::test_sage_conv[1-gcn-g1-int64] PASSED            [ 48%]
tests/mxnet/test_nn.py::test_sage_conv[1-gcn-g2-int32] PASSED            [ 48%]
tests/mxnet/test_nn.py::test_sage_conv[1-gcn-g2-int64] PASSED            [ 48%]
tests/mxnet/test_nn.py::test_sage_conv[1-gcn-g3-int32] PASSED            [ 48%]
tests/mxnet/test_nn.py::test_sage_conv[1-gcn-g3-int64] PASSED            [ 48%]
tests/mxnet/test_nn.py::test_sage_conv[1-gcn-g4-int32] PASSED [ 91%]
tests/compute/test_sparse.py::test_edge_softmax[idtype0-shp0-dst-g0] PASSED            [ 48%]
tests/mxnet/test_nn.py::test_sage_conv[1-gcn-g4-int64] PASSED            [ 48%]
tests/mxnet/test_nn.py::test_sage_conv[1-gcn-g5-int32] PASSED            [ 48%]
tests/mxnet/test_nn.py::test_sage_conv[1-gcn-g5-int64] PASSED            [ 49%]
tests/mxnet/test_nn.py::test_sage_conv[1-gcn-g6-int32] PASSED            [ 49%]
tests/mxnet/test_nn.py::test_sage_conv[1-gcn-g6-int64] PASSED            [ 49%]
tests/mxnet/test_nn.py::test_sage_conv[1-gcn-g7-int32] PASSED [ 91%]
tests/compute/test_sparse.py::test_edge_softmax[idtype0-shp1-src-g0] PASSED            [ 49%]
tests/mxnet/test_nn.py::test_sage_conv[1-gcn-g7-int64] PASSED            [ 49%]
tests/mxnet/test_nn.py::test_sage_conv[10-mean-g0-int32] PASSED          [ 49%]
tests/mxnet/test_nn.py::test_sage_conv[10-mean-g0-int64] PASSED          [ 49%]
tests/mxnet/test_nn.py::test_sage_conv[10-mean-g1-int32] PASSED          [ 49%]
tests/mxnet/test_nn.py::test_sage_conv[10-mean-g1-int64] PASSED [ 91%]
tests/compute/test_sparse.py::test_edge_softmax[idtype0-shp1-dst-g0] PASSED          [ 50%]
tests/mxnet/test_nn.py::test_sage_conv[10-mean-g2-int32] PASSED          [ 50%]
tests/mxnet/test_nn.py::test_sage_conv[10-mean-g2-int64] PASSED          [ 50%]
tests/mxnet/test_nn.py::test_sage_conv[10-mean-g3-int32] PASSED          [ 50%]
tests/mxnet/test_nn.py::test_sage_conv[10-mean-g3-int64] PASSED          [ 50%]
tests/mxnet/test_nn.py::test_sage_conv[10-mean-g4-int32] PASSED          [ 50%]
tests/mxnet/test_nn.py::test_sage_conv[10-mean-g4-int64] PASSED          [ 50%]
tests/mxnet/test_nn.py::test_sage_conv[10-mean-g5-int32] PASSED [ 91%]
tests/compute/test_sparse.py::test_edge_softmax[idtype0-shp2-src-g0] PASSED          [ 50%]
tests/mxnet/test_nn.py::test_sage_conv[10-mean-g5-int64] PASSED          [ 50%]
tests/mxnet/test_nn.py::test_sage_conv[10-mean-g6-int32] PASSED          [ 51%]
tests/mxnet/test_nn.py::test_sage_conv[10-mean-g6-int64] PASSED          [ 51%]
tests/mxnet/test_nn.py::test_sage_conv[10-mean-g7-int32] PASSED [ 91%]
tests/compute/test_sparse.py::test_edge_softmax[idtype0-shp2-dst-g0] PASSED          [ 51%]
tests/mxnet/test_nn.py::test_sage_conv[10-mean-g7-int64] PASSED          [ 51%]
tests/mxnet/test_nn.py::test_sage_conv[10-pool-g0-int32] PASSED          [ 51%]
tests/mxnet/test_nn.py::test_sage_conv[10-pool-g0-int64] PASSED          [ 51%]
tests/mxnet/test_nn.py::test_sage_conv[10-pool-g1-int32] PASSED [ 91%]
tests/compute/test_sparse.py::test_edge_softmax[idtype1-shp0-src-g0] PASSED          [ 51%]
tests/mxnet/test_nn.py::test_sage_conv[10-pool-g1-int64] PASSED          [ 51%]
tests/mxnet/test_nn.py::test_sage_conv[10-pool-g2-int32] PASSED          [ 52%]
tests/mxnet/test_nn.py::test_sage_conv[10-pool-g2-int64] PASSED [ 92%]
tests/compute/test_sparse.py::test_edge_softmax[idtype1-shp0-dst-g0] PASSED          [ 52%]
tests/mxnet/test_nn.py::test_sage_conv[10-pool-g3-int32] PASSED          [ 52%]
tests/mxnet/test_nn.py::test_sage_conv[10-pool-g3-int64] PASSED          [ 52%]
tests/mxnet/test_nn.py::test_sage_conv[10-pool-g4-int32] PASSED          [ 52%]
tests/mxnet/test_nn.py::test_sage_conv[10-pool-g4-int64] PASSED          [ 52%]
tests/mxnet/test_nn.py::test_sage_conv[10-pool-g5-int32] PASSED [ 92%]
tests/compute/test_sparse.py::test_edge_softmax[idtype1-shp1-src-g0] PASSED          [ 52%]
tests/mxnet/test_nn.py::test_sage_conv[10-pool-g5-int64] PASSED          [ 52%]
tests/mxnet/test_nn.py::test_sage_conv[10-pool-g6-int32] PASSED          [ 53%]
tests/mxnet/test_nn.py::test_sage_conv[10-pool-g6-int64] PASSED [ 92%]
tests/compute/test_sparse.py::test_edge_softmax[idtype1-shp1-dst-g0] PASSED          [ 53%]
tests/mxnet/test_nn.py::test_sage_conv[10-pool-g7-int32] PASSED          [ 53%]
tests/mxnet/test_nn.py::test_sage_conv[10-pool-g7-int64] PASSED          [ 53%]
tests/mxnet/test_nn.py::test_sage_conv[10-gcn-g0-int32] PASSED           [ 53%]
tests/mxnet/test_nn.py::test_sage_conv[10-gcn-g0-int64] PASSED           [ 53%]
tests/mxnet/test_nn.py::test_sage_conv[10-gcn-g1-int32] PASSED [ 92%]
tests/compute/test_sparse.py::test_edge_softmax[idtype1-shp2-src-g0] PASSED           [ 53%]
tests/mxnet/test_nn.py::test_sage_conv[10-gcn-g1-int64] PASSED           [ 53%]
tests/mxnet/test_nn.py::test_sage_conv[10-gcn-g2-int32] PASSED           [ 54%]
tests/mxnet/test_nn.py::test_sage_conv[10-gcn-g2-int64] PASSED           [ 54%]
tests/mxnet/test_nn.py::test_sage_conv[10-gcn-g3-int32] PASSED           [ 54%]
tests/mxnet/test_nn.py::test_sage_conv[10-gcn-g3-int64] PASSED           [ 54%]
tests/mxnet/test_nn.py::test_sage_conv[10-gcn-g4-int32] PASSED           [ 54%]
tests/mxnet/test_nn.py::test_sage_conv[10-gcn-g4-int64] PASSED [ 92%]
tests/compute/test_sparse.py::test_edge_softmax[idtype1-shp2-dst-g0] PASSED           [ 54%]
tests/mxnet/test_nn.py::test_sage_conv[10-gcn-g5-int32] PASSED           [ 54%]
tests/mxnet/test_nn.py::test_sage_conv[10-gcn-g5-int64] PASSED [ 92%]
tests/compute/test_sparse.py::test_segment_reduce[sum] PASSED           [ 54%]
tests/mxnet/test_nn.py::test_sage_conv[10-gcn-g6-int32] PASSED           [ 55%]
tests/mxnet/test_nn.py::test_sage_conv[10-gcn-g6-int64] PASSED           [ 55%]
tests/mxnet/test_nn.py::test_sage_conv[10-gcn-g7-int32] PASSED           [ 55%]
tests/mxnet/test_nn.py::test_sage_conv[10-gcn-g7-int64] PASSED           [ 55%]
tests/mxnet/test_nn.py::test_sage_conv_bi[1-mean-g0-int32] PASSED        [ 55%]
tests/mxnet/test_nn.py::test_sage_conv_bi[1-mean-g0-int64] PASSED        [ 55%]
tests/mxnet/test_nn.py::test_sage_conv_bi[1-mean-g1-int32] PASSED            [ 92%]
tests/compute/test_sparse.py::test_segment_reduce[max] PASSED        [ 55%]
tests/mxnet/test_nn.py::test_sage_conv_bi[1-mean-g1-int64] PASSED        [ 55%]
tests/mxnet/test_nn.py::test_sage_conv_bi[1-mean-g2-int32] PASSED            [ 92%]
tests/compute/test_sparse.py::test_segment_reduce[min] PASSED        [ 56%]
tests/mxnet/test_nn.py::test_sage_conv_bi[1-mean-g2-int64] PASSED        [ 56%]
tests/mxnet/test_nn.py::test_sage_conv_bi[1-pool-g0-int32] PASSED            [ 92%]
tests/compute/test_sparse.py::test_segment_reduce[mean] PASSED        [ 56%]
tests/mxnet/test_nn.py::test_sage_conv_bi[1-pool-g0-int64] PASSED        [ 56%]
tests/mxnet/test_nn.py::test_sage_conv_bi[1-pool-g1-int32] PASSED        [ 56%]
tests/mxnet/test_nn.py::test_sage_conv_bi[1-pool-g1-int64] PASSED        [ 56%]
tests/mxnet/test_nn.py::test_sage_conv_bi[1-pool-g2-int32] PASSED           [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype0-0.01-1-idtype0] SKIPPED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype0-0.01-1-idtype1] SKIPPED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype0-0.01-8-idtype0] SKIPPED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype0-0.01-8-idtype1] SKIPPED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype0-0.01-16-idtype0] SKIPPED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype0-0.01-16-idtype1] SKIPPED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype0-0.01-64-idtype0] SKIPPED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype0-0.01-64-idtype1] SKIPPED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype0-0.01-256-idtype0] SKIPPED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype0-0.01-256-idtype1] SKIPPED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype1-0.01-1-idtype0] SKIPPED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype1-0.01-1-idtype1] SKIPPED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype1-0.01-8-idtype0] SKIPPED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype1-0.01-8-idtype1] SKIPPED [ 92%]
tests/compute/test_sparse.py::test_segment_mm[dtype1-0.01-16-idtype0] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype1-0.01-16-idtype1] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype1-0.01-64-idtype0] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype1-0.01-64-idtype1] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype1-0.01-256-idtype0] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype1-0.01-256-idtype1] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype2-0.003-1-idtype0] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype2-0.003-1-idtype1] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype2-0.003-8-idtype0] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype2-0.003-8-idtype1] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype2-0.003-16-idtype0] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype2-0.003-16-idtype1] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype2-0.003-64-idtype0] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype2-0.003-64-idtype1] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype2-0.003-256-idtype0] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype2-0.003-256-idtype1] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype3-0.0001-1-idtype0] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype3-0.0001-1-idtype1] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype3-0.0001-8-idtype0] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype3-0.0001-8-idtype1] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype3-0.0001-16-idtype0] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype3-0.0001-16-idtype1] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype3-0.0001-64-idtype0] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype3-0.0001-64-idtype1] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype3-0.0001-256-idtype0] SKIPPED [ 93%]
tests/compute/test_sparse.py::test_segment_mm[dtype3-0.0001-256-idtype1] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype0-0.01-1] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype0-0.01-8] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype0-0.01-16] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype0-0.01-64] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype0-0.01-256] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype1-0.02-1] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype1-0.02-8] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype1-0.02-16] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype1-0.02-64] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype1-0.02-256] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype2-0.003-1] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype2-0.003-8] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype2-0.003-16] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype2-0.003-64] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype2-0.003-256] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype3-0.0001-1] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype3-0.0001-8] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype3-0.0001-16] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype3-0.0001-64] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_gather_mm_idx_b[dtype3-0.0001-256] SKIPPED [ 94%]
tests/compute/test_sparse.py::test_use_libxsmm_switch SKIPPED (Only ...) [ 94%]
tests/compute/test_specialization.py::test_v2v_update_all[idtype0] PASSED        [ 56%]
tests/mxnet/test_nn.py::test_sage_conv_bi[1-pool-g2-int64] PASSED        [ 56%]
tests/mxnet/test_nn.py::test_sage_conv_bi[1-gcn-g0-int32] PASSED         [ 57%]
tests/mxnet/test_nn.py::test_sage_conv_bi[1-gcn-g0-int64] PASSED         [ 57%]
tests/mxnet/test_nn.py::test_sage_conv_bi[1-gcn-g1-int32] PASSED         [ 57%]
tests/mxnet/test_nn.py::test_sage_conv_bi[1-gcn-g1-int64] PASSED         [ 57%]
tests/mxnet/test_nn.py::test_sage_conv_bi[1-gcn-g2-int32] PASSED         [ 57%]
tests/mxnet/test_nn.py::test_sage_conv_bi[1-gcn-g2-int64] PASSED         [ 57%]
tests/mxnet/test_nn.py::test_sage_conv_bi[2-mean-g0-int32] PASSED [ 94%]
tests/compute/test_specialization.py::test_v2v_update_all[idtype1] PASSED        [ 57%]
tests/mxnet/test_nn.py::test_sage_conv_bi[2-mean-g0-int64] PASSED        [ 57%]
tests/mxnet/test_nn.py::test_sage_conv_bi[2-mean-g1-int32] PASSED        [ 58%]
tests/mxnet/test_nn.py::test_sage_conv_bi[2-mean-g1-int64] PASSED        [ 58%]
tests/mxnet/test_nn.py::test_sage_conv_bi[2-mean-g2-int32] PASSED        [ 58%]
tests/mxnet/test_nn.py::test_sage_conv_bi[2-mean-g2-int64] PASSED        [ 58%]
tests/mxnet/test_nn.py::test_sage_conv_bi[2-pool-g0-int32] PASSED        [ 58%]
tests/mxnet/test_nn.py::test_sage_conv_bi[2-pool-g0-int64] PASSED        [ 58%]
tests/mxnet/test_nn.py::test_sage_conv_bi[2-pool-g1-int32] PASSED        [ 58%]
tests/mxnet/test_nn.py::test_sage_conv_bi[2-pool-g1-int64] PASSED [ 94%]
tests/compute/test_specialization.py::test_v2v_snr[idtype0] PASSED        [ 58%]
tests/mxnet/test_nn.py::test_sage_conv_bi[2-pool-g2-int32] PASSED        [ 59%]
tests/mxnet/test_nn.py::test_sage_conv_bi[2-pool-g2-int64] PASSED        [ 59%]
tests/mxnet/test_nn.py::test_sage_conv_bi[2-gcn-g0-int32] PASSED         [ 59%]
tests/mxnet/test_nn.py::test_sage_conv_bi[2-gcn-g0-int64] PASSED         [ 59%]
tests/mxnet/test_nn.py::test_sage_conv_bi[2-gcn-g1-int32] PASSED         [ 59%]
tests/mxnet/test_nn.py::test_sage_conv_bi[2-gcn-g1-int64] PASSED         [ 59%]
tests/mxnet/test_nn.py::test_sage_conv_bi[2-gcn-g2-int32] PASSED         [ 59%]
tests/mxnet/test_nn.py::test_sage_conv_bi[2-gcn-g2-int64] PASSED         [ 59%]
tests/mxnet/test_nn.py::test_sage_conv_bi2[1-mean-int32] PASSED          [ 60%]
tests/mxnet/test_nn.py::test_sage_conv_bi2[1-mean-int64] PASSED          [ 60%]
tests/mxnet/test_nn.py::test_sage_conv_bi2[1-pool-int32] PASSED          [ 60%]
tests/mxnet/test_nn.py::test_sage_conv_bi2[1-pool-int64] PASSED       [ 94%]
tests/compute/test_specialization.py::test_v2v_snr[idtype1] PASSED          [ 60%]
tests/mxnet/test_nn.py::test_sage_conv_bi2[1-gcn-int32] PASSED           [ 60%]
tests/mxnet/test_nn.py::test_sage_conv_bi2[1-gcn-int64] PASSED           [ 60%]
tests/mxnet/test_nn.py::test_sage_conv_bi2[2-mean-int32] PASSED          [ 60%]
tests/mxnet/test_nn.py::test_sage_conv_bi2[2-mean-int64] PASSED          [ 60%]
tests/mxnet/test_nn.py::test_sage_conv_bi2[2-pool-int32] PASSED          [ 61%]
tests/mxnet/test_nn.py::test_sage_conv_bi2[2-pool-int64] PASSED          [ 61%]
tests/mxnet/test_nn.py::test_sage_conv_bi2[2-gcn-int32] PASSED           [ 61%]
tests/mxnet/test_nn.py::test_sage_conv_bi2[2-gcn-int64] PASSED           [ 61%]
tests/mxnet/test_nn.py::test_gg_conv PASSED       [ 95%]
tests/compute/test_specialization.py::test_v2v_pull[idtype0] PASSED                              [ 61%]
tests/mxnet/test_nn.py::test_cheb_conv[1] PASSED                         [ 61%]
tests/mxnet/test_nn.py::test_cheb_conv[20] PASSED      [ 95%]
tests/compute/test_specialization.py::test_v2v_pull[idtype1] PASSED                        [ 61%]
tests/mxnet/test_nn.py::test_agnn_conv[g0-int32] PASSED                  [ 61%]
tests/mxnet/test_nn.py::test_agnn_conv[g0-int64] PASSED                  [ 62%]
tests/mxnet/test_nn.py::test_agnn_conv[g1-int32] PASSED                  [ 62%]
tests/mxnet/test_nn.py::test_agnn_conv[g1-int64] PASSED                  [ 62%]
tests/mxnet/test_nn.py::test_agnn_conv[g2-int32] PASSED                  [ 62%]
tests/mxnet/test_nn.py::test_agnn_conv[g2-int64] PASSED                  [ 62%]
tests/mxnet/test_nn.py::test_agnn_conv[g3-int32] PASSED                  [ 62%]
tests/mxnet/test_nn.py::test_agnn_conv[g3-int64] PASSED                  [ 62%]
tests/mxnet/test_nn.py::test_agnn_conv[g4-int32] PASSED                  [ 62%]
tests/mxnet/test_nn.py::test_agnn_conv[g4-int64] PASSED      [ 95%]
tests/compute/test_specialization.py::test_update_all_multi_fallback[idtype0] PASSED                  [ 63%]
tests/mxnet/test_nn.py::test_agnn_conv[g5-int32] PASSED                  [ 63%]
tests/mxnet/test_nn.py::test_agnn_conv[g5-int64] PASSED [ 95%]
tests/compute/test_specialization.py::test_update_all_multi_fallback[idtype1] PASSED                  [ 63%]
tests/mxnet/test_nn.py::test_agnn_conv[g6-int32] PASSED [ 95%]
tests/compute/test_specialization.py::test_pull_multi_fallback[idtype0] PASSED                  [ 63%]
tests/mxnet/test_nn.py::test_agnn_conv[g6-int64] PASSED                  [ 63%]
tests/mxnet/test_nn.py::test_agnn_conv_bi[g0-int32] PASSED               [ 63%]
tests/mxnet/test_nn.py::test_agnn_conv_bi[g0-int64] PASSED               [ 63%]
tests/mxnet/test_nn.py::test_agnn_conv_bi[g1-int32] PASSED               [ 63%]
tests/mxnet/test_nn.py::test_agnn_conv_bi[g1-int64] PASSED               [ 64%]
tests/mxnet/test_nn.py::test_appnp_conv PASSED                           [ 64%]
tests/mxnet/test_nn.py::test_dense_cheb_conv[1] PASSED                   [ 64%]
tests/mxnet/test_nn.py::test_dense_cheb_conv[2] PASSED [ 95%]
tests/compute/test_specialization.py::test_pull_multi_fallback[idtype1] PASSED                   [ 64%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g0-both-int32] PASSED    [ 64%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g0-both-int64] PASSED    [ 64%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g0-right-int32] PASSED   [ 64%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g0-right-int64] PASSED   [ 64%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g0-none-int32] PASSED    [ 65%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g0-none-int64] PASSED    [ 65%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g1-both-int32] PASSED    [ 65%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g1-both-int64] PASSED    [ 65%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g1-right-int32] PASSED   [ 65%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g1-right-int64] PASSED [ 95%]
tests/compute/test_specialization.py::test_spmv_3d_feat[idtype0] PASSED   [ 65%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g1-none-int32] PASSED    [ 65%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g1-none-int64] PASSED    [ 65%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g2-both-int32] PASSED    [ 66%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g2-both-int64] PASSED    [ 66%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g2-right-int32] PASSED   [ 66%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g2-right-int64] PASSED   [ 66%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g2-none-int32] PASSED    [ 66%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g2-none-int64] PASSED    [ 66%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g3-both-int32] PASSED    [ 66%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g3-both-int64] PASSED    [ 66%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g3-right-int32] PASSED  [ 95%]
tests/compute/test_specialization.py::test_spmv_3d_feat[idtype1] PASSED   [ 66%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g3-right-int64] PASSED   [ 67%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g3-none-int32] PASSED    [ 67%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g3-none-int64] PASSED    [ 67%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g4-both-int32] PASSED    [ 67%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g4-both-int64] PASSED    [ 67%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g4-right-int32] PASSED   [ 67%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g4-right-int64] PASSED   [ 67%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g4-none-int32] PASSED    [ 67%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g4-none-int64] PASSED    [ 68%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g5-both-int32] PASSED    [ 68%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g5-both-int64] PASSED    [ 68%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g5-right-int32] PASSED  [ 95%]
tests/compute/test_subgraph.py::test_edge_subgraph PASSED   [ 68%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g5-right-int64] PASSED   [ 68%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g5-none-int32] PASSED                [ 95%]
tests/compute/test_subgraph.py::test_subgraph PASSED                     [ 95%]
tests/compute/test_subgraph.py::test_subgraph_mask[idtype0] PASSED       [ 95%]
tests/compute/test_subgraph.py::test_subgraph_mask[idtype1] PASSED    [ 68%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g5-none-int64] PASSED    [ 68%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g6-both-int32] PASSED    [ 68%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g6-both-int64] PASSED       [ 95%]
tests/compute/test_subgraph.py::test_subgraph1[idtype0] PASSED    [ 69%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g6-right-int32] PASSED   [ 69%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g6-right-int64] PASSED   [ 69%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g6-none-int32] PASSED    [ 69%]
tests/mxnet/test_nn.py::test_dense_graph_conv[1-g6-none-int64] PASSED    [ 69%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g0-both-int32] PASSED    [ 69%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g0-both-int64] PASSED    [ 69%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g0-right-int32] PASSED   [ 69%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g0-right-int64] PASSED   [ 70%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g0-none-int32] PASSED    [ 70%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g0-none-int64] PASSED    [ 70%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g1-both-int32] PASSED    [ 70%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g1-both-int64] PASSED    [ 70%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g1-right-int32] PASSED           [ 95%]
tests/compute/test_subgraph.py::test_subgraph1[idtype1] PASSED   [ 70%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g1-right-int64] PASSED   [ 70%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g1-none-int32] PASSED    [ 70%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g1-none-int64] PASSED    [ 71%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g2-both-int32] PASSED    [ 71%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g2-both-int64] PASSED    [ 71%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g2-right-int32] PASSED   [ 71%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g2-right-int64] PASSED   [ 71%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g2-none-int32] PASSED    [ 71%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g2-none-int64] PASSED    [ 71%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g3-both-int32] PASSED    [ 71%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g3-both-int64] PASSED    [ 72%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g3-right-int32] PASSED   [ 72%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g3-right-int64] PASSED           [ 95%]
tests/compute/test_subgraph.py::test_in_subgraph[idtype0] PASSED   [ 72%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g3-none-int32] PASSED    [ 72%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g3-none-int64] PASSED    [ 72%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g4-both-int32] PASSED    [ 72%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g4-both-int64] PASSED    [ 72%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g4-right-int32] PASSED   [ 72%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g4-right-int64] PASSED   [ 73%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g4-none-int32] PASSED    [ 73%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g4-none-int64] PASSED    [ 73%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g5-both-int32] PASSED    [ 73%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g5-both-int64] PASSED    [ 73%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g5-right-int32] PASSED   [ 73%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g5-right-int64] PASSED   [ 73%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g5-none-int32] PASSED    [ 73%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g5-none-int64] PASSED    [ 74%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g6-both-int32] PASSED    [ 74%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g6-both-int64] PASSED    [ 74%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g6-right-int32] PASSED   [ 74%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g6-right-int64] PASSED   [ 74%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g6-none-int32] PASSED    [ 74%]
tests/mxnet/test_nn.py::test_dense_graph_conv[2-g6-none-int64] PASSED    [ 74%]
tests/mxnet/test_nn.py::test_dense_sage_conv[1-g0-int32] PASSED          [ 74%]
tests/mxnet/test_nn.py::test_dense_sage_conv[1-g0-int64] PASSED          [ 75%]
tests/mxnet/test_nn.py::test_dense_sage_conv[1-g1-int32] PASSED         [ 95%]
tests/compute/test_subgraph.py::test_in_subgraph[idtype1] PASSED          [ 75%]
tests/mxnet/test_nn.py::test_dense_sage_conv[1-g1-int64] PASSED          [ 75%]
tests/mxnet/test_nn.py::test_dense_sage_conv[1-g2-int32] PASSED          [ 75%]
tests/mxnet/test_nn.py::test_dense_sage_conv[1-g2-int64] PASSED          [ 75%]
tests/mxnet/test_nn.py::test_dense_sage_conv[1-g3-int32] PASSED          [ 75%]
tests/mxnet/test_nn.py::test_dense_sage_conv[1-g3-int64] PASSED          [ 75%]
tests/mxnet/test_nn.py::test_dense_sage_conv[1-g4-int32] PASSED          [ 75%]
tests/mxnet/test_nn.py::test_dense_sage_conv[1-g4-int64] PASSED          [ 76%]
tests/mxnet/test_nn.py::test_dense_sage_conv[1-g5-int32] PASSED          [ 76%]
tests/mxnet/test_nn.py::test_dense_sage_conv[1-g5-int64] PASSED          [ 76%]
tests/mxnet/test_nn.py::test_dense_sage_conv[1-g6-int32] PASSED          [ 76%]
tests/mxnet/test_nn.py::test_dense_sage_conv[1-g6-int64] PASSED          [ 76%]
tests/mxnet/test_nn.py::test_dense_sage_conv[1-g7-int32] PASSED          [ 76%]
tests/mxnet/test_nn.py::test_dense_sage_conv[1-g7-int64] PASSED          [ 76%]
tests/mxnet/test_nn.py::test_dense_sage_conv[1-g8-int32] PASSED          [ 76%]
tests/mxnet/test_nn.py::test_dense_sage_conv[1-g8-int64] PASSED          [ 77%]
tests/mxnet/test_nn.py::test_dense_sage_conv[1-g9-int32] PASSED          [ 77%]
tests/mxnet/test_nn.py::test_dense_sage_conv[1-g9-int64] PASSED          [ 77%]
tests/mxnet/test_nn.py::test_dense_sage_conv[2-g0-int32] PASSED          [ 77%]
tests/mxnet/test_nn.py::test_dense_sage_conv[2-g0-int64] PASSED          [ 77%]
tests/mxnet/test_nn.py::test_dense_sage_conv[2-g1-int32] PASSED          [ 77%]
tests/mxnet/test_nn.py::test_dense_sage_conv[2-g1-int64] PASSED          [ 77%]
tests/mxnet/test_nn.py::test_dense_sage_conv[2-g2-int32] PASSED          [ 77%]
tests/mxnet/test_nn.py::test_dense_sage_conv[2-g2-int64] PASSED          [ 78%]
tests/mxnet/test_nn.py::test_dense_sage_conv[2-g3-int32] PASSED          [ 78%]
tests/mxnet/test_nn.py::test_dense_sage_conv[2-g3-int64] PASSED         [ 95%]
tests/compute/test_subgraph.py::test_out_subgraph[idtype0] PASSED          [ 78%]
tests/mxnet/test_nn.py::test_dense_sage_conv[2-g4-int32] PASSED          [ 78%]
tests/mxnet/test_nn.py::test_dense_sage_conv[2-g4-int64] PASSED          [ 78%]
tests/mxnet/test_nn.py::test_dense_sage_conv[2-g5-int32] PASSED          [ 78%]
tests/mxnet/test_nn.py::test_dense_sage_conv[2-g5-int64] PASSED          [ 78%]
tests/mxnet/test_nn.py::test_dense_sage_conv[2-g6-int32] PASSED          [ 78%]
tests/mxnet/test_nn.py::test_dense_sage_conv[2-g6-int64] PASSED          [ 79%]
tests/mxnet/test_nn.py::test_dense_sage_conv[2-g7-int32] PASSED          [ 79%]
tests/mxnet/test_nn.py::test_dense_sage_conv[2-g7-int64] PASSED          [ 79%]
tests/mxnet/test_nn.py::test_dense_sage_conv[2-g8-int32] PASSED          [ 79%]
tests/mxnet/test_nn.py::test_dense_sage_conv[2-g8-int64] PASSED          [ 79%]
tests/mxnet/test_nn.py::test_dense_sage_conv[2-g9-int32] PASSED          [ 79%]
tests/mxnet/test_nn.py::test_dense_sage_conv[2-g9-int64] PASSED          [ 79%]
tests/mxnet/test_nn.py::test_edge_conv[1-g0-int32] PASSED                [ 79%]
tests/mxnet/test_nn.py::test_edge_conv[1-g0-int64] PASSED                [ 80%]
tests/mxnet/test_nn.py::test_edge_conv[1-g1-int32] PASSED                [ 80%]
tests/mxnet/test_nn.py::test_edge_conv[1-g1-int64] PASSED                [ 80%]
tests/mxnet/test_nn.py::test_edge_conv[1-g2-int32] PASSED                [ 80%]
tests/mxnet/test_nn.py::test_edge_conv[1-g2-int64] PASSED        [ 95%]
tests/compute/test_subgraph.py::test_out_subgraph[idtype1] PASSED                [ 80%]
tests/mxnet/test_nn.py::test_edge_conv[1-g3-int32] PASSED                [ 80%]
tests/mxnet/test_nn.py::test_edge_conv[1-g3-int64] PASSED                [ 80%]
tests/mxnet/test_nn.py::test_edge_conv[1-g4-int32] PASSED                [ 80%]
tests/mxnet/test_nn.py::test_edge_conv[1-g4-int64] PASSED                [ 81%]
tests/mxnet/test_nn.py::test_edge_conv[1-g5-int32] PASSED                [ 81%]
tests/mxnet/test_nn.py::test_edge_conv[1-g5-int64] PASSED        [ 95%]
tests/compute/test_subgraph.py::test_subgraph_message_passing PASSED                [ 81%]
tests/mxnet/test_nn.py::test_edge_conv[1-g6-int32] PASSED                [ 81%]
tests/mxnet/test_nn.py::test_edge_conv[1-g6-int64] PASSED     [ 95%]
tests/compute/test_subgraph.py::test_khop_in_subgraph[idtype0] PASSED                [ 81%]
tests/mxnet/test_nn.py::test_edge_conv[2-g0-int32] PASSED                [ 81%]
tests/mxnet/test_nn.py::test_edge_conv[2-g0-int64] PASSED                [ 81%]
tests/mxnet/test_nn.py::test_edge_conv[2-g1-int32] PASSED                [ 81%]
tests/mxnet/test_nn.py::test_edge_conv[2-g1-int64] PASSED                [ 82%]
tests/mxnet/test_nn.py::test_edge_conv[2-g2-int32] PASSED                [ 82%]
tests/mxnet/test_nn.py::test_edge_conv[2-g2-int64] PASSED                [ 82%]
tests/mxnet/test_nn.py::test_edge_conv[2-g3-int32] PASSED                [ 82%]
tests/mxnet/test_nn.py::test_edge_conv[2-g3-int64] PASSED                [ 82%]
tests/mxnet/test_nn.py::test_edge_conv[2-g4-int32] PASSED                [ 82%]
tests/mxnet/test_nn.py::test_edge_conv[2-g4-int64] PASSED                [ 82%]
tests/mxnet/test_nn.py::test_edge_conv[2-g5-int32] PASSED                [ 82%]
tests/mxnet/test_nn.py::test_edge_conv[2-g5-int64] PASSED                [ 83%]
tests/mxnet/test_nn.py::test_edge_conv[2-g6-int32] PASSED                [ 83%]
tests/mxnet/test_nn.py::test_edge_conv[2-g6-int64] PASSED                [ 83%]
tests/mxnet/test_nn.py::test_edge_conv_bi[1-g0-int32] PASSED             [ 83%]
tests/mxnet/test_nn.py::test_edge_conv_bi[1-g0-int64] PASSED             [ 83%]
tests/mxnet/test_nn.py::test_edge_conv_bi[1-g1-int32] PASSED             [ 83%]
tests/mxnet/test_nn.py::test_edge_conv_bi[1-g1-int64] PASSED             [ 83%]
tests/mxnet/test_nn.py::test_edge_conv_bi[2-g0-int32] PASSED             [ 83%]
tests/mxnet/test_nn.py::test_edge_conv_bi[2-g0-int64] PASSED    [ 95%]
tests/compute/test_subgraph.py::test_khop_in_subgraph[idtype1] PASSED             [ 83%]
tests/mxnet/test_nn.py::test_edge_conv_bi[2-g1-int32] PASSED             [ 84%]
tests/mxnet/test_nn.py::test_edge_conv_bi[2-g1-int64] PASSED             [ 84%]
tests/mxnet/test_nn.py::test_gin_conv[mean-g0-int32] PASSED              [ 84%]
tests/mxnet/test_nn.py::test_gin_conv[mean-g0-int64] PASSED              [ 84%]
tests/mxnet/test_nn.py::test_gin_conv[mean-g1-int32] PASSED              [ 84%]
tests/mxnet/test_nn.py::test_gin_conv[mean-g1-int64] PASSED              [ 84%]
tests/mxnet/test_nn.py::test_gin_conv[mean-g2-int32] PASSED              [ 84%]
tests/mxnet/test_nn.py::test_gin_conv[mean-g2-int64] PASSED              [ 84%]
tests/mxnet/test_nn.py::test_gin_conv[mean-g3-int32] PASSED              [ 85%]
tests/mxnet/test_nn.py::test_gin_conv[mean-g3-int64] PASSED              [ 85%]
tests/mxnet/test_nn.py::test_gin_conv[mean-g4-int32] PASSED              [ 85%]
tests/mxnet/test_nn.py::test_gin_conv[mean-g4-int64] PASSED              [ 85%]
tests/mxnet/test_nn.py::test_gin_conv[mean-g5-int32] PASSED              [ 85%]
tests/mxnet/test_nn.py::test_gin_conv[mean-g5-int64] PASSED              [ 85%]
tests/mxnet/test_nn.py::test_gin_conv[mean-g6-int32] PASSED              [ 85%]
tests/mxnet/test_nn.py::test_gin_conv[mean-g6-int64] PASSED              [ 85%]
tests/mxnet/test_nn.py::test_gin_conv[mean-g7-int32] PASSED              [ 86%]
tests/mxnet/test_nn.py::test_gin_conv[mean-g7-int64] PASSED              [ 86%]
tests/mxnet/test_nn.py::test_gin_conv[max-g0-int32] PASSED               [ 86%]
tests/mxnet/test_nn.py::test_gin_conv[max-g0-int64] PASSED               [ 86%]
tests/mxnet/test_nn.py::test_gin_conv[max-g1-int32] PASSED               [ 86%]
tests/mxnet/test_nn.py::test_gin_conv[max-g1-int64] PASSED               [ 86%]
tests/mxnet/test_nn.py::test_gin_conv[max-g2-int32] PASSED               [ 86%]
tests/mxnet/test_nn.py::test_gin_conv[max-g2-int64] PASSED               [ 86%]
tests/mxnet/test_nn.py::test_gin_conv[max-g3-int32] PASSED               [ 87%]
tests/mxnet/test_nn.py::test_gin_conv[max-g3-int64] PASSED               [ 87%]
tests/mxnet/test_nn.py::test_gin_conv[max-g4-int32] PASSED               [ 87%]
tests/mxnet/test_nn.py::test_gin_conv[max-g4-int64] PASSED               [ 87%]
tests/mxnet/test_nn.py::test_gin_conv[max-g5-int32] PASSED               [ 87%]
tests/mxnet/test_nn.py::test_gin_conv[max-g5-int64] PASSED               [ 87%]
tests/mxnet/test_nn.py::test_gin_conv[max-g6-int32] PASSED               [ 87%]
tests/mxnet/test_nn.py::test_gin_conv[max-g6-int64] PASSED               [ 87%]
tests/mxnet/test_nn.py::test_gin_conv[max-g7-int32] PASSED               [ 88%]
tests/mxnet/test_nn.py::test_gin_conv[max-g7-int64] PASSED               [ 88%]
tests/mxnet/test_nn.py::test_gin_conv[sum-g0-int32] PASSED               [ 88%]
tests/mxnet/test_nn.py::test_gin_conv[sum-g0-int64] PASSED               [ 88%]
tests/mxnet/test_nn.py::test_gin_conv[sum-g1-int32] PASSED               [ 88%]
tests/mxnet/test_nn.py::test_gin_conv[sum-g1-int64] PASSED               [ 88%]
tests/mxnet/test_nn.py::test_gin_conv[sum-g2-int32] PASSED               [ 88%]
tests/mxnet/test_nn.py::test_gin_conv[sum-g2-int64] PASSED               [ 88%]
tests/mxnet/test_nn.py::test_gin_conv[sum-g3-int32] PASSED               [ 89%]
tests/mxnet/test_nn.py::test_gin_conv[sum-g3-int64] PASSED               [ 89%]
tests/mxnet/test_nn.py::test_gin_conv[sum-g4-int32] PASSED               [ 89%]
tests/mxnet/test_nn.py::test_gin_conv[sum-g4-int64] PASSED               [ 89%]
tests/mxnet/test_nn.py::test_gin_conv[sum-g5-int32] PASSED               [ 89%]
tests/mxnet/test_nn.py::test_gin_conv[sum-g5-int64] PASSED               [ 89%]
tests/mxnet/test_nn.py::test_gin_conv[sum-g6-int32] PASSED               [ 89%]
tests/mxnet/test_nn.py::test_gin_conv[sum-g6-int64] PASSED               [ 89%]
tests/mxnet/test_nn.py::test_gin_conv[sum-g7-int32] PASSED               [ 90%]
tests/mxnet/test_nn.py::test_gin_conv[sum-g7-int64] PASSED               [ 90%]
tests/mxnet/test_nn.py::test_gin_conv_bi[mean-g0-int32] PASSED           [ 90%]
tests/mxnet/test_nn.py::test_gin_conv_bi[mean-g0-int64] PASSED           [ 90%]
tests/mxnet/test_nn.py::test_gin_conv_bi[mean-g1-int32] PASSED           [ 90%]
tests/mxnet/test_nn.py::test_gin_conv_bi[mean-g1-int64] PASSED    [ 95%]
tests/compute/test_subgraph.py::test_khop_out_subgraph[idtype0] PASSED           [ 90%]
tests/mxnet/test_nn.py::test_gin_conv_bi[mean-g2-int32] PASSED           [ 90%]
tests/mxnet/test_nn.py::test_gin_conv_bi[mean-g2-int64] PASSED           [ 90%]
tests/mxnet/test_nn.py::test_gin_conv_bi[max-g0-int32] PASSED            [ 91%]
tests/mxnet/test_nn.py::test_gin_conv_bi[max-g0-int64] PASSED            [ 91%]
tests/mxnet/test_nn.py::test_gin_conv_bi[max-g1-int32] PASSED            [ 91%]
tests/mxnet/test_nn.py::test_gin_conv_bi[max-g1-int64] PASSED            [ 91%]
tests/mxnet/test_nn.py::test_gin_conv_bi[max-g2-int32] PASSED            [ 91%]
tests/mxnet/test_nn.py::test_gin_conv_bi[max-g2-int64] PASSED            [ 91%]
tests/mxnet/test_nn.py::test_gin_conv_bi[sum-g0-int32] PASSED            [ 91%]
tests/mxnet/test_nn.py::test_gin_conv_bi[sum-g0-int64] PASSED            [ 91%]
tests/mxnet/test_nn.py::test_gin_conv_bi[sum-g1-int32] PASSED            [ 92%]
tests/mxnet/test_nn.py::test_gin_conv_bi[sum-g1-int64] PASSED            [ 92%]
tests/mxnet/test_nn.py::test_gin_conv_bi[sum-g2-int32] PASSED            [ 92%]
tests/mxnet/test_nn.py::test_gin_conv_bi[sum-g2-int64] PASSED            [ 92%]
tests/mxnet/test_nn.py::test_gmm_conv[g0-int32] PASSED                   [ 92%]
tests/mxnet/test_nn.py::test_gmm_conv[g0-int64] PASSED                   [ 92%]
tests/mxnet/test_nn.py::test_gmm_conv[g1-int32] PASSED                   [ 92%]
tests/mxnet/test_nn.py::test_gmm_conv[g1-int64] PASSED                   [ 92%]
tests/mxnet/test_nn.py::test_gmm_conv[g2-int32] PASSED                   [ 93%]
tests/mxnet/test_nn.py::test_gmm_conv[g2-int64] PASSED                   [ 93%]
tests/mxnet/test_nn.py::test_gmm_conv[g3-int32] PASSED                   [ 93%]
tests/mxnet/test_nn.py::test_gmm_conv[g3-int64] PASSED   [ 95%]
tests/compute/test_subgraph.py::test_khop_out_subgraph[idtype1] PASSED                   [ 93%]
tests/mxnet/test_nn.py::test_gmm_conv[g4-int32] PASSED                   [ 93%]
tests/mxnet/test_nn.py::test_gmm_conv[g4-int64] PASSED                   [ 93%]
tests/mxnet/test_nn.py::test_gmm_conv[g5-int32] PASSED                   [ 93%]
tests/mxnet/test_nn.py::test_gmm_conv[g5-int64] PASSED                   [ 93%]
tests/mxnet/test_nn.py::test_gmm_conv[g6-int32] PASSED                   [ 94%]
tests/mxnet/test_nn.py::test_gmm_conv[g6-int64] PASSED                   [ 94%]
tests/mxnet/test_nn.py::test_gmm_conv_bi[g0-int32] PASSED                [ 94%]
tests/mxnet/test_nn.py::test_gmm_conv_bi[g0-int64] PASSED                [ 94%]
tests/mxnet/test_nn.py::test_gmm_conv_bi[g1-int32] PASSED                [ 94%]
tests/mxnet/test_nn.py::test_gmm_conv_bi[g1-int64] PASSED                [ 94%]
tests/mxnet/test_nn.py::test_nn_conv[g0-int32] PASSED                    [ 94%]
tests/mxnet/test_nn.py::test_nn_conv[g0-int64] PASSED                    [ 94%]
tests/mxnet/test_nn.py::test_nn_conv[g1-int32] PASSED                    [ 95%]
tests/mxnet/test_nn.py::test_nn_conv[g1-int64] PASSED                    [ 95%]
tests/mxnet/test_nn.py::test_nn_conv[g2-int32] PASSED                    [ 95%]
tests/mxnet/test_nn.py::test_nn_conv[g2-int64] PASSED                    [ 95%]
tests/mxnet/test_nn.py::test_nn_conv[g3-int32] PASSED                    [ 95%]
tests/mxnet/test_nn.py::test_nn_conv[g3-int64] PASSED   [ 95%]
tests/compute/test_subgraph.py::test_subframes[/cpu:0-parent_idx_device0] SKIPPED [ 96%]
tests/compute/test_subgraph.py::test_subframes[/cpu:0-parent_idx_device1] SKIPPED [ 96%]
tests/compute/test_subgraph.py::test_subframes[/cpu:0-parent_idx_device2] SKIPPED [ 96%]
tests/compute/test_subgraph.py::test_subframes[/cpu:0-parent_idx_device3] SKIPPED [ 96%]
tests/compute/test_subgraph.py::test_subframes[/gpu:0-parent_idx_device0] SKIPPED [ 96%]
tests/compute/test_subgraph.py::test_subframes[/gpu:0-parent_idx_device1] SKIPPED [ 96%]
tests/compute/test_subgraph.py::test_subframes[/gpu:0-parent_idx_device2] SKIPPED [ 96%]
tests/compute/test_subgraph.py::test_subframes[/gpu:0-parent_idx_device3] SKIPPED [ 96%]
tests/compute/test_subgraph.py::test_uva_subgraph[idtype0-/cpu:0] SKIPPED [ 96%]
tests/compute/test_subgraph.py::test_uva_subgraph[idtype0-/gpu:0] SKIPPED [ 96%]
tests/compute/test_subgraph.py::test_uva_subgraph[idtype1-/cpu:0] SKIPPED [ 96%]
tests/compute/test_subgraph.py::test_uva_subgraph[idtype1-/gpu:0] SKIPPED [ 96%]
tests/compute/test_transform.py::test_line_graph1 PASSED                 [ 96%]
tests/compute/test_transform.py::test_line_graph2[idtype0] PASSED        [ 96%]
tests/compute/test_transform.py::test_line_graph2[idtype1] PASSED        [ 96%]
tests/compute/test_transform.py::test_no_backtracking PASSED                    [ 95%]
tests/mxnet/test_nn.py::test_nn_conv[g4-int32] PASSED                    [ 95%]
tests/mxnet/test_nn.py::test_nn_conv[g4-int64] PASSED                    [ 95%]
tests/mxnet/test_nn.py::test_nn_conv[g5-int32] PASSED             [ 96%]
tests/compute/test_transform.py::test_reverse[idtype0] PASSED                    [ 96%]
tests/mxnet/test_nn.py::test_nn_conv[g5-int64] PASSED                    [ 96%]
tests/mxnet/test_nn.py::test_nn_conv[g6-int32] PASSED                    [ 96%]
tests/mxnet/test_nn.py::test_nn_conv[g6-int64] PASSED                    [ 96%]
tests/mxnet/test_nn.py::test_nn_conv[g7-int32] PASSED                    [ 96%]
tests/mxnet/test_nn.py::test_nn_conv[g7-int64] PASSED            [ 96%]
tests/compute/test_transform.py::test_reverse[idtype1] PASSED            [ 96%]
tests/compute/test_transform.py::test_reverse_shared_frames[idtype0] PASSED                    [ 96%]
tests/mxnet/test_nn.py::test_nn_conv_bi[g0-int32] PASSED                 [ 96%]
tests/mxnet/test_nn.py::test_nn_conv_bi[g0-int64] PASSED                 [ 96%]
tests/mxnet/test_nn.py::test_nn_conv_bi[g1-int32] PASSED                 [ 97%]
tests/mxnet/test_nn.py::test_nn_conv_bi[g1-int64] PASSED                 [ 97%]
tests/mxnet/test_nn.py::test_nn_conv_bi[g2-int32] PASSED                 [ 97%]
tests/mxnet/test_nn.py::test_nn_conv_bi[g2-int64] PASSED                 [ 97%]
tests/mxnet/test_nn.py::test_sg_conv[1] PASSED [ 96%]
tests/compute/test_transform.py::test_reverse_shared_frames[idtype1] PASSED                           [ 97%]
tests/mxnet/test_nn.py::test_sg_conv[2] PASSED                           [ 97%]
tests/mxnet/test_nn.py::test_set2set PASSED [ 96%]
tests/compute/test_transform.py::test_to_bidirected PASSED               [ 96%]
tests/compute/test_transform.py::test_add_reverse_edges PASSED           [ 96%]
tests/compute/test_transform.py::test_simple_graph PASSED                [ 96%]
tests/compute/test_transform.py::test_khop_graph PASSED                              [ 97%]
tests/mxnet/test_nn.py::test_glob_att_pool PASSED                        [ 97%]
tests/mxnet/test_nn.py::test_simple_pool PASSED                          [ 98%]
tests/mxnet/test_nn.py::test_rgcn[1] PASSED                              [ 98%]
tests/mxnet/test_nn.py::test_rgcn[2] PASSED                              [ 98%]
tests/mxnet/test_nn.py::test_rgcn[8] PASSED                              [ 98%]
tests/mxnet/test_nn.py::test_sequential PASSED                           [ 98%]
tests/mxnet/test_nn.py::test_hetero_conv[sum-int32] PASSED                  [ 96%]
tests/compute/test_transform.py::test_khop_adj PASSED                    [ 96%]
tests/compute/test_transform.py::test_laplacian_lambda_max PASSED        [ 97%]
tests/compute/test_transform.py::test_partition_with_halo PASSED               [ 98%]
tests/mxnet/test_nn.py::test_hetero_conv[sum-int64] PASSED         [ 97%]
tests/compute/test_transform.py::test_metis_partition[idtype0] PASSED    [ 97%]
tests/compute/test_transform.py::test_metis_partition[idtype1] PASSED               [ 98%]
tests/mxnet/test_nn.py::test_hetero_conv[max-int32] PASSED               [ 98%]
tests/mxnet/test_nn.py::test_hetero_conv[max-int64] PASSED               [ 99%]
tests/mxnet/test_nn.py::test_hetero_conv[min-int32] PASSED               [ 99%]
tests/mxnet/test_nn.py::test_hetero_conv[min-int64] PASSED    [ 97%]
tests/compute/test_transform.py::test_reorder_nodes PASSED               [ 97%]
tests/compute/test_transform.py::test_compact[idtype0] PASSED            [ 97%]
tests/compute/test_transform.py::test_compact[idtype1] PASSED            [ 97%]
tests/compute/test_transform.py::test_to_simple[idtype0] PASSED          [ 97%]
tests/compute/test_transform.py::test_to_simple[idtype1] PASSED          [ 97%]
tests/compute/test_transform.py::test_to_block[idtype0] PASSED           [ 97%]
tests/compute/test_transform.py::test_to_block[idtype1] PASSED           [ 97%]
tests/compute/test_transform.py::test_remove_edges[idtype0] PASSED       [ 97%]
tests/compute/test_transform.py::test_remove_edges[idtype1] PASSED       [ 97%]
tests/compute/test_transform.py::test_add_edges[idtype0] PASSED          [ 97%]
tests/compute/test_transform.py::test_add_edges[idtype1] PASSED          [ 97%]
tests/compute/test_transform.py::test_add_nodes[idtype0] PASSED          [ 97%]
tests/compute/test_transform.py::test_add_nodes[idtype1] PASSED          [ 97%]
tests/compute/test_transform.py::test_remove_nodes[idtype0] PASSED       [ 97%]
tests/compute/test_transform.py::test_remove_nodes[idtype1] PASSED               [ 99%]
tests/mxnet/test_nn.py::test_hetero_conv[mean-int32] PASSED       [ 97%]
tests/compute/test_transform.py::test_add_selfloop[idtype0] PASSED       [ 97%]
tests/compute/test_transform.py::test_add_selfloop[idtype1] PASSED       [ 97%]
tests/compute/test_transform.py::test_remove_selfloop[idtype0] PASSED    [ 97%]
tests/compute/test_transform.py::test_remove_selfloop[idtype1] PASSED    [ 97%]
tests/compute/test_transform.py::test_reorder_graph[idtype0] PASSED      [ 97%]
tests/compute/test_transform.py::test_reorder_graph[idtype1] PASSED      [ 98%]
tests/compute/test_transform.py::test_norm_by_dst[idtype0] SKIPPED (...) [ 98%]
tests/compute/test_transform.py::test_norm_by_dst[idtype1] SKIPPED (...) [ 98%]
tests/compute/test_transform.py::test_module_add_self_loop[idtype0] PASSED [ 98%]
tests/compute/test_transform.py::test_module_add_self_loop[idtype1] PASSED [ 98%]
tests/compute/test_transform.py::test_module_remove_self_loop[idtype0] PASSED [ 98%]
tests/compute/test_transform.py::test_module_remove_self_loop[idtype1] PASSED [ 98%]
tests/compute/test_transform.py::test_module_add_reverse[idtype0] PASSED [ 98%]
tests/compute/test_transform.py::test_module_add_reverse[idtype1] PASSED [ 98%]
tests/compute/test_transform.py::test_module_to_simple[idtype0] PASSED   [ 98%]
tests/compute/test_transform.py::test_module_to_simple[idtype1] PASSED   [ 98%]
tests/compute/test_transform.py::test_module_line_graph[idtype0] PASSED  [ 98%]
tests/compute/test_transform.py::test_module_line_graph[idtype1] PASSED  [ 98%]
tests/compute/test_transform.py::test_module_khop_graph[idtype0] PASSED  [ 98%]
tests/compute/test_transform.py::test_module_khop_graph[idtype1] PASSED  [ 98%]
tests/compute/test_transform.py::test_module_add_metapaths[idtype0] PASSED [ 98%]
tests/compute/test_transform.py::test_module_add_metapaths[idtype1] PASSED [ 98%]
tests/compute/test_transform.py::test_module_compose[idtype0] PASSED     [ 98%]
tests/compute/test_transform.py::test_module_compose[idtype1] PASSED     [ 98%]
tests/compute/test_transform.py::test_module_gcnnorm[idtype0] PASSED     [ 98%]
tests/compute/test_transform.py::test_module_gcnnorm[idtype1] PASSED     [ 98%]
tests/compute/test_transform.py::test_module_ppr[idtype0] SKIPPED (O...) [ 98%]
tests/compute/test_transform.py::test_module_ppr[idtype1] SKIPPED (O...) [ 98%]
tests/compute/test_transform.py::test_module_heat_kernel[idtype0] SKIPPED [ 98%]
tests/compute/test_transform.py::test_module_heat_kernel[idtype1] SKIPPED [ 98%]
tests/compute/test_transform.py::test_module_gdc[idtype0] SKIPPED (O...) [ 99%]
tests/compute/test_transform.py::test_module_gdc[idtype1] SKIPPED (O...) [ 99%]
tests/compute/test_transform.py::test_module_node_shuffle[idtype0] PASSED [ 99%]
tests/compute/test_transform.py::test_module_node_shuffle[idtype1] PASSED [ 99%]
tests/compute/test_transform.py::test_module_drop_node[idtype0] SKIPPED  [ 99%]
tests/compute/test_transform.py::test_module_drop_node[idtype1] SKIPPED  [ 99%]
tests/compute/test_transform.py::test_module_drop_edge[idtype0] SKIPPED  [ 99%]
tests/compute/test_transform.py::test_module_drop_edge[idtype1] SKIPPED  [ 99%]
tests/compute/test_transform.py::test_module_add_edge[idtype0] PASSED    [ 99%]
tests/compute/test_transform.py::test_module_add_edge[idtype1] PASSED    [ 99%]
tests/compute/test_transform.py::test_module_random_walk_pe[idtype0] PASSED [ 99%]
tests/compute/test_transform.py::test_module_random_walk_pe[idtype1] PASSED [ 99%]
tests/compute/test_transform.py::test_module_laplacian_pe[idtype0] PASSED [ 99%]
tests/compute/test_transform.py::test_module_laplacian_pe[idtype1] PASSED [ 99%]
tests/compute/test_transform.py::test_module_sign[g0] SKIPPED (Only ...) [ 99%]
tests/compute/test_transform.py::test_module_row_feat_normalizer[idtype0] SKIPPED [ 99%]
tests/compute/test_transform.py::test_module_row_feat_normalizer[idtype1] SKIPPED [ 99%]
tests/compute/test_transform.py::test_module_feat_mask[idtype0] SKIPPED  [ 99%]
tests/compute/test_transform.py::test_module_feat_mask[idtype1] SKIPPED  [ 99%]
tests/compute/test_traversal.py::test_bfs[idtype0] PASSED                [ 99%]
tests/compute/test_traversal.py::test_bfs[idtype1] PASSED                [ 99%]
tests/compute/test_traversal.py::test_topological_nodes[idtype0] PASSED  [ 99%]
tests/compute/test_traversal.py::test_topological_nodes[idtype1] PASSED  [ 99%]
tests/compute/test_traversal.py::test_dfs_labeled_edges[idtype0] PASSED  [ 99%]
tests/compute/test_traversal.py::test_dfs_labeled_edges[idtype1] PASSED              [ 99%]
tests/mxnet/test_nn.py::test_hetero_conv[mean-int64] PASSED  [100%]

=============================== warnings summary ===============================
python/dgl/backend/backend.py:1741
  /root/jenkins/workspace/dgl_PR-4648@2/python/dgl/backend/backend.py:1741: DeprecationWarning: invalid escape sequence \P
    """

../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/tensorflow/python/data/ops/iterator_ops.py:546
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/tensorflow/python/data/ops/iterator_ops.py:546: DeprecationWarning: Using or importing the ABCs from 'collections' instead of from 'collections.abc' is deprecated, and in 3.8 it will stop working
    class IteratorBase(collections.Iterator, trackable.Trackable,

../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/tensorflow/python/data/ops/dataset_ops.py:106
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/tensorflow/python/data/ops/dataset_ops.py:106: DeprecationWarning: Using or importing the ABCs from 'collections' instead of from 'collections.abc' is deprecated, and in 3.8 it will stop working
    class DatasetV2(collections.Iterable, tracking_base.Trackable,

../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/tensorflow/python/autograph/utils/testing.py:21
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/tensorflow/python/autograph/utils/testing.py:21: DeprecationWarning: the imp module is deprecated in favour of importlib; see the module's documentation for alternative uses
    import imp

../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:10
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:10: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    _nlv = LooseVersion(_np_version)

../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:11
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:11: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    _np_version_under1p16 = _nlv < LooseVersion("1.16")

../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:12
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:12: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    _np_version_under1p17 = _nlv < LooseVersion("1.17")

../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:13
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:13: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    _np_version_under1p18 = _nlv < LooseVersion("1.18")

../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:14
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:14: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    _np_version_under1p19 = _nlv < LooseVersion("1.19")

../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:15
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:15: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    _np_version_under1p20 = _nlv < LooseVersion("1.20")

../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/setuptools/_distutils/version.py:351
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/setuptools/_distutils/version.py:351: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    other = LooseVersion(other)

../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/pandas/compat/numpy/function.py:125
../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/pandas/compat/numpy/function.py:125
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/pandas/compat/numpy/function.py:125: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    if LooseVersion(_np_version) >= LooseVersion("1.17.0"):

../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/keras_preprocessing/image/utils.py:23
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/keras_preprocessing/image/utils.py:23: DeprecationWarning: NEAREST is deprecated and will be removed in Pillow 10 (2023-07-01). Use Resampling.NEAREST or Dither.NONE instead.
    'nearest': pil_image.NEAREST,

../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/keras_preprocessing/image/utils.py:24
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/keras_preprocessing/image/utils.py:24: DeprecationWarning: BILINEAR is deprecated and will be removed in Pillow 10 (2023-07-01). Use Resampling.BILINEAR instead.
    'bilinear': pil_image.BILINEAR,

../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/keras_preprocessing/image/utils.py:25
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/keras_preprocessing/image/utils.py:25: DeprecationWarning: BICUBIC is deprecated and will be removed in Pillow 10 (2023-07-01). Use Resampling.BICUBIC instead.
    'bicubic': pil_image.BICUBIC,

../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/keras_preprocessing/image/utils.py:28
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/keras_preprocessing/image/utils.py:28: DeprecationWarning: HAMMING is deprecated and will be removed in Pillow 10 (2023-07-01). Use Resampling.HAMMING instead.
    if hasattr(pil_image, 'HAMMING'):

../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/keras_preprocessing/image/utils.py:29
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/keras_preprocessing/image/utils.py:29: DeprecationWarning: HAMMING is deprecated and will be removed in Pillow 10 (2023-07-01). Use Resampling.HAMMING instead.
    _PIL_INTERPOLATION_METHODS['hamming'] = pil_image.HAMMING

../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/keras_preprocessing/image/utils.py:30
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/keras_preprocessing/image/utils.py:30: DeprecationWarning: BOX is deprecated and will be removed in Pillow 10 (2023-07-01). Use Resampling.BOX instead.
    if hasattr(pil_image, 'BOX'):

../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/keras_preprocessing/image/utils.py:31
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/keras_preprocessing/image/utils.py:31: DeprecationWarning: BOX is deprecated and will be removed in Pillow 10 (2023-07-01). Use Resampling.BOX instead.
    _PIL_INTERPOLATION_METHODS['box'] = pil_image.BOX

../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/keras_preprocessing/image/utils.py:33
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/keras_preprocessing/image/utils.py:33: DeprecationWarning: LANCZOS is deprecated and will be removed in Pillow 10 (2023-07-01). Use Resampling.LANCZOS instead.
    if hasattr(pil_image, 'LANCZOS'):

../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/keras_preprocessing/image/utils.py:34
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/keras_preprocessing/image/utils.py:34: DeprecationWarning: LANCZOS is deprecated and will be removed in Pillow 10 (2023-07-01). Use Resampling.LANCZOS instead.
    _PIL_INTERPOLATION_METHODS['lanczos'] = pil_image.LANCZOS

python/dgl/backend/tensorflow/tensor.py:15
python/dgl/backend/tensorflow/tensor.py:15
  /root/jenkins/workspace/dgl_PR-4648@2/python/dgl/backend/tensorflow/tensor.py:15: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    if LooseVersion(tf.__version__) < LooseVersion("2.3.0"):

tests/compute/test_basics.py: 2 warnings
tests/compute/test_filter.py: 1 warning
tests/compute/test_graph.py: 9 warnings
tests/compute/test_kernel.py: 3 warnings
tests/compute/test_propagate.py: 2 warnings
tests/compute/test_removal.py: 16 warnings
tests/compute/test_serialize.py: 3 warnings
tests/compute/test_specialization.py: 12 warnings
tests/compute/test_subgraph.py: 2 warnings
tests/compute/test_transform.py: 10 warnings
tests/compute/test_traversal.py: 2 warnings
  /root/jenkins/workspace/dgl_PR-4648@2/python/dgl/heterograph.py:72: DGLWarning: Recommend creating graphs by `dgl.graph(data)` instead of `dgl.DGLGraph(data)`.
    dgl_warning('Recommend creating graphs by `dgl.graph(data)`'

tests/compute/test_basics.py: 2 warnings
tests/compute/test_batched_graph.py: 8 warnings
tests/compute/test_graph.py: 2 warnings
tests/compute/test_kernel.py: 1 warning
tests/compute/test_propagate.py: 2 warnings
tests/compute/test_removal.py: 10 warnings
tests/compute/test_specialization.py: 10 warnings
tests/compute/test_subgraph.py: 2 warnings
  /root/jenkins/workspace/dgl_PR-4648@2/python/dgl/heterograph.py:354: DGLWarning: DGLGraph.add_edge is deprecated. Please use DGLGraph.add_edges
    dgl_warning("DGLGraph.add_edge is deprecated. Please use DGLGraph.add_edges")

tests/compute/test_basics.py::test_update_all_0deg[idtype0]
tests/compute/test_basics.py::test_update_all_0deg[idtype1]
tests/compute/test_basics.py::test_pull_0deg[idtype0]
tests/compute/test_basics.py::test_pull_0deg[idtype1]
tests/compute/test_propagate.py::test_prop_nodes_topo[idtype0]
tests/compute/test_propagate.py::test_prop_nodes_topo[idtype1]
  /root/jenkins/workspace/dgl_PR-4648@2/python/dgl/core.py:79: DGLWarning: The input graph for the user-defined edge function does not contain valid edges
    dgl_warning('The input graph for the user-defined edge function ' \

tests/compute/test_batched_graph.py::test_batched_edge_ordering[idtype0]
tests/compute/test_batched_graph.py::test_batched_edge_ordering[idtype1]
tests/compute/test_graph.py::test_query
tests/compute/test_graph.py::test_hypersparse_query
tests/compute/test_transform.py::test_no_backtracking
tests/compute/test_transform.py::test_reverse[idtype0]
tests/compute/test_transform.py::test_reverse[idtype1]
  /root/jenkins/workspace/dgl_PR-4648@2/python/dgl/heterograph.py:2978: DGLWarning: DGLGraph.edge_id is deprecated. Please use DGLGraph.edge_ids.
    dgl_warning("DGLGraph.edge_id is deprecated. Please use DGLGraph.edge_ids.")

tests/compute/test_batched_heterograph.py::test_features[idtype0]
tests/compute/test_batched_heterograph.py::test_features[idtype1]
  /root/jenkins/workspace/dgl_PR-4648@2/python/dgl/batch.py:159: DGLWarning: Arguments edge_attrs has been deprecated. Please use edata instead.
    dgl_warning('Arguments edge_attrs has been deprecated. Please use'

tests/compute/test_csrmm.py::test_csrmm[dtype0-idtype0]
tests/compute/test_csrmm.py::test_csrmm[dtype0-idtype1]
tests/compute/test_csrmm.py::test_csrmm[dtype1-idtype0]
tests/compute/test_csrmm.py::test_csrmm[dtype1-idtype1]
tests/compute/test_csrmm.py::test_csrsum[dtype0-idtype0]
tests/compute/test_csrmm.py::test_csrsum[dtype0-idtype1]
tests/compute/test_csrmm.py::test_csrsum[dtype1-idtype0]
tests/compute/test_csrmm.py::test_csrsum[dtype1-idtype1]
  /root/jenkins/workspace/dgl_PR-4648@2/python/dgl/heterograph_index.py:797: FutureWarning: Adjacency matrix by default currently returns edge IDs.  As a result there is one 0 entry which is not eliminated.  In the next release it will return 1s by default, and 0 will be eliminated otherwise.
    FutureWarning)

tests/compute/test_data.py::test_citation_graph
  /root/jenkins/workspace/dgl_PR-4648@2/python/dgl/data/citation_graph.py:287: RuntimeWarning: divide by zero encountered in power
    r_inv = np.power(rowsum, -1).flatten()

tests/compute/test_data.py::test_csvdataset
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/numpy/lib/arraysetops.py:565: FutureWarning: elementwise comparison failed; returning scalar instead, but in the future will perform elementwise comparison
    mask &= (ar1 != a)

tests/compute/test_data.py::test_csvdataset
  /root/jenkins/workspace/dgl_PR-4648@2/python/dgl/data/csv_dataset_base.py:298: DGLWarning: Unamed column is found. Ignored...
    dgl_warning("Unamed column is found. Ignored...")

tests/compute/test_graph.py::test_query
tests/compute/test_graph.py::test_hypersparse_query
tests/compute/test_heterograph.py::test_query[idtype0]
tests/compute/test_heterograph.py::test_query[idtype1]
  /root/jenkins/workspace/dgl_PR-4648@2/python/dgl/heterograph.py:2753: DGLWarning: DGLGraph.has_node is deprecated. Please use DGLGraph.has_nodes
    dgl_warning("DGLGraph.has_node is deprecated. Please use DGLGraph.has_nodes")

tests/compute/test_graph.py::test_query
tests/compute/test_graph.py::test_hypersparse_query
  /root/jenkins/workspace/dgl_PR-4648@2/python/dgl/heterograph.py:2687: DGLWarning: DGLGraph.__contains__ is deprecated. Please directly call has_nodes.
    dgl_warning('DGLGraph.__contains__ is deprecated.'

tests/compute/test_graph.py::test_query
tests/compute/test_sampling.py::test_non_uniform_random_walk[False]
tests/compute/test_sampling.py::test_uniform_random_walk[False]
tests/compute/test_sampling.py::test_node2vec
tests/compute/test_transform.py::test_no_backtracking
  /root/jenkins/workspace/dgl_PR-4648@2/python/dgl/heterograph.py:2851: DGLWarning: DGLGraph.has_edge_between is deprecated. Please use DGLGraph.has_edges_between
    dgl_warning("DGLGraph.has_edge_between is deprecated. "

tests/compute/test_graph.py::test_query
tests/compute/test_graph.py::test_hypersparse_query
  /root/jenkins/workspace/dgl_PR-4648@2/python/dgl/heterograph.py:3432: DGLWarning: DGLGraph.in_degree is deprecated. Please use DGLGraph.in_degrees
    dgl_warning("DGLGraph.in_degree is deprecated. Please use DGLGraph.in_degrees")

tests/compute/test_graph.py::test_query
tests/compute/test_graph.py::test_hypersparse_query
  /root/jenkins/workspace/dgl_PR-4648@2/python/dgl/heterograph.py:3516: DGLWarning: DGLGraph.out_degree is deprecated. Please use DGLGraph.out_degrees
    dgl_warning("DGLGraph.out_degree is deprecated. Please use DGLGraph.out_degrees")

tests/compute/test_graph.py::test_query
  /root/jenkins/workspace/dgl_PR-4648@2/python/dgl/heterograph.py:84: DGLWarning: Keyword arguments ['readonly', 'sort_csr'] are deprecated in v0.5, and can be safely removed in all cases.
    ' removed in all cases.'.format(list(deprecate_kwargs.keys())))

tests/compute/test_heterograph.py: 20 warnings
  /root/jenkins/workspace/dgl_PR-4648@2/tests/compute/test_heterograph.py:1128: DeprecationWarning: np.asscalar(a) is deprecated since NumPy v1.16, use a.item() instead
    assert np.asscalar(F.asnumpy(src_i)) == nid[src[i]]

tests/compute/test_heterograph.py: 20 warnings
  /root/jenkins/workspace/dgl_PR-4648@2/tests/compute/test_heterograph.py:1129: DeprecationWarning: np.asscalar(a) is deprecated since NumPy v1.16, use a.item() instead
    assert np.asscalar(F.asnumpy(dst_i)) == nid[dst[i]]

tests/compute/test_heterograph.py::test_invertible_conversion[idtype0]
tests/compute/test_heterograph.py::test_invertible_conversion[idtype1]
  /root/jenkins/workspace/dgl_PR-4648@2/python/dgl/heterograph.py:2635: DGLWarning: DGLGraph.is_readonly is deprecated in v0.5.
  DGLGraph now always supports mutable operations like add_nodes and add_edges.
    dgl_warning('DGLGraph.is_readonly is deprecated in v0.5.\n'

tests/compute/test_pickle.py::test_pickling_batched_heterograph
  /root/jenkins/workspace/dgl_PR-4648@2/python/dgl/batch.py:511: DGLWarning: From v0.5, DGLHeteroGraph is merged into DGLGraph. You can safely replace dgl.batch_hetero with dgl.batch
    dgl_warning('From v0.5, DGLHeteroGraph is merged into DGLGraph. You can safely'

tests/compute/test_sampler.py::test_create_full
tests/compute/test_sampler.py::test_1neighbor_sampler_all
tests/compute/test_sampler.py::test_1neighbor_sampler
tests/compute/test_sampler.py::test_prefetch_neighbor_sampler
tests/compute/test_sampler.py::test_10neighbor_sampler_all
tests/compute/test_sampler.py::test_10neighbor_sampler
tests/compute/test_sampler.py::test_setseed
  /root/jenkins/workspace/dgl_PR-4648@2/python/dgl/contrib/sampling/sampler.py:317: DGLWarning: dgl.contrib.sampling.NeighborSampler is deprecated starting from v0.5. Please read our guide<link> for how to use the new sampling APIs.
    dgl_warning('dgl.contrib.sampling.NeighborSampler is deprecated starting from v0.5.'

tests/compute/test_sampler.py::test_create_full
tests/compute/test_sampler.py::test_1neighbor_sampler_all
tests/compute/test_sampler.py::test_1neighbor_sampler
tests/compute/test_sampler.py::test_prefetch_neighbor_sampler
tests/compute/test_sampler.py::test_10neighbor_sampler_all
tests/compute/test_sampler.py::test_10neighbor_sampler
tests/compute/test_sampler.py::test_layer_sampler
tests/compute/test_sampler.py::test_setseed
  /root/jenkins/workspace/dgl_PR-4648@2/python/dgl/_deprecate/nodeflow.py:99: DGLWarning: NodeFlow APIs are deprecated starting from v0.5. Please read our guide<link> for how to use the new sampling APIs.
    dgl_warning('NodeFlow APIs are deprecated starting from v0.5. Please read our'

tests/compute/test_sampling.py::test_global_uniform_negative_sampling[int32]
tests/compute/test_sampling.py::test_global_uniform_negative_sampling[int64]
  /root/jenkins/workspace/dgl_PR-4648@2/python/dgl/sampling/negative.py:102: ComplexWarning: Casting complex values to real discards the imaginary part
    g._graph, etype_id, num_samples, 3, exclude_self_loops, replace, redundancy)

tests/compute/test_serialize.py::test_graph_serialize_with_feature[False]
tests/compute/test_serialize.py::test_graph_serialize_without_feature[False]
tests/compute/test_serialize.py::test_graph_serialize_with_labels[False]
tests/compute/test_transform.py::test_simple_graph
  /root/jenkins/workspace/dgl_PR-4648@2/python/dgl/heterograph.py:84: DGLWarning: Keyword arguments ['readonly'] are deprecated in v0.5, and can be safely removed in all cases.
    ' removed in all cases.'.format(list(deprecate_kwargs.keys())))

tests/compute/test_serialize.py::test_load_old_files1
tests/compute/test_serialize.py::test_load_old_files2
  /root/jenkins/workspace/dgl_PR-4648@2/python/dgl/data/graph_serialize.py:179: DGLWarning: You are loading a graph file saved by old version of dgl.              Please consider saving it again with the current format.
    Please consider saving it again with the current format.")

tests/compute/test_transform.py::test_reverse_shared_frames[idtype0]
tests/compute/test_transform.py::test_reverse_shared_frames[idtype1]
  /root/jenkins/workspace/dgl_PR-4648@2/python/dgl/transforms/functional.py:1267: DGLWarning: share_ndata argument has been renamed to copy_ndata.
    dgl_warning('share_ndata argument has been renamed to copy_ndata.')

tests/compute/test_transform.py::test_reverse_shared_frames[idtype0]
tests/compute/test_transform.py::test_reverse_shared_frames[idtype1]
  /root/jenkins/workspace/dgl_PR-4648@2/python/dgl/transforms/functional.py:1270: DGLWarning: share_edata argument has been renamed to copy_edata.
    dgl_warning('share_edata argument has been renamed to copy_edata.')

tests/compute/test_transform.py::test_simple_graph
  /root/jenkins/workspace/dgl_PR-4648@2/python/dgl/transforms/functional.py:1319: DGLWarning: dgl.to_simple_graph is renamed to dgl.to_simple in v0.5.
    dgl_warning('dgl.to_simple_graph is renamed to dgl.to_simple in v0.5.')

-- Docs: https://docs.pytest.org/en/stable/how-to/capture-warnings.html
- generated xml file: /root/jenkins/workspace/dgl_PR-4648@2/pytest_compute.xml -
============================ slowest 100 durations =============================
196.55s call     tests/compute/test_kernel.py::test_all_binary_builtins
132.31s call     tests/compute/test_data.py::test_reddit
64.08s call     tests/compute/test_data.py::test_fakenews
55.00s call     tests/compute/test_data.py::test_gin
36.58s call     tests/compute/test_data.py::test_as_graphpred
15.67s call     tests/compute/test_transform.py::test_metis_partition[idtype1]
11.99s call     tests/compute/test_data.py::test_tudataset_regression
11.59s call     tests/compute/test_data.py::test_gnn_benchmark
11.27s call     tests/compute/test_data.py::test_explain_syn
10.12s call     tests/compute/test_sampling.py::test_sample_neighbors_outedge
8.64s call     tests/compute/test_kernel.py::test_copy_src_reduce
7.50s call     tests/compute/test_data.py::test_as_graphpred_reprocess
6.15s call     tests/compute/test_kernel.py::test_copy_edge_reduce
5.85s call     tests/compute/test_sampling.py::test_sample_neighbors_prob
5.74s call     tests/compute/test_sampling.py::test_sample_neighbors_noprob
5.56s call     tests/compute/test_graph.py::test_query
4.70s call     tests/compute/test_data.py::test_wiki_cs
3.98s call     tests/compute/test_data.py::test_citation_graph
3.95s call     tests/compute/test_heterograph.py::test_query[idtype1]
3.94s call     tests/compute/test_heterograph.py::test_query[idtype0]
3.60s call     tests/compute/test_data.py::test_flickr
3.22s call     tests/compute/test_data.py::test_fraud
3.17s call     tests/compute/test_data.py::test_add_nodepred_split
2.88s call     tests/compute/test_heterograph.py::test_forking_pickler
2.78s call     tests/compute/test_heterograph.py::test_updates[idtype0]
2.73s call     tests/compute/test_transform.py::test_remove_nodes[idtype0]
2.72s call     tests/compute/test_transform.py::test_remove_nodes[idtype1]
2.59s call     tests/compute/test_heterograph.py::test_updates[idtype1]
2.49s call     tests/compute/test_specialization.py::test_pull_multi_fallback[idtype0]
2.48s call     tests/compute/test_specialization.py::test_pull_multi_fallback[idtype1]
2.46s call     tests/compute/test_specialization.py::test_spmv_3d_feat[idtype0]
2.42s call     tests/compute/test_transform.py::test_partition_with_halo
2.09s call     tests/compute/test_transform.py::test_reorder_nodes
2.06s call     tests/compute/test_data.py::test_csvdataset
2.03s call     tests/compute/test_specialization.py::test_v2v_pull[idtype0]
2.02s call     tests/compute/test_specialization.py::test_spmv_3d_feat[idtype1]
2.00s call     tests/compute/test_heterograph.py::test_level2[idtype0]
1.99s call     tests/compute/test_specialization.py::test_v2v_pull[idtype1]
1.83s call     tests/compute/test_heterograph.py::test_level2[idtype1]
1.80s call     tests/compute/test_basics.py::test_update_routines[idtype0]
1.80s call     tests/compute/test_basics.py::test_update_routines[idtype1]
1.68s call     tests/compute/test_propagate.py::test_prop_edges_dfs[idtype1]
1.68s call     tests/compute/test_specialization.py::test_v2v_snr[idtype0]
1.66s call     tests/compute/test_propagate.py::test_prop_edges_dfs[idtype0]
1.60s call     tests/compute/test_specialization.py::test_v2v_snr[idtype1]
1.45s call     tests/compute/test_data.py::test_minigc
1.45s call     tests/compute/test_heterograph.py::test_format[idtype1]
1.43s call     tests/compute/test_subgraph.py::test_khop_in_subgraph[idtype1]
1.28s call     tests/compute/test_subgraph.py::test_khop_in_subgraph[idtype0]
1.28s call     tests/compute/test_transform.py::test_khop_graph
1.25s call     tests/compute/test_heterograph.py::test_format[idtype0]
1.17s call     tests/compute/test_sampler.py::test_10neighbor_sampler_all
1.14s call     tests/compute/test_heterograph.py::test_view1[idtype0]
1.14s call     tests/compute/test_subgraph.py::test_khop_out_subgraph[idtype0]
1.14s call     tests/compute/test_sampling.py::test_sample_neighbors_topk
1.12s call     tests/compute/test_subgraph.py::test_out_subgraph[idtype0]
1.09s call     tests/compute/test_subgraph.py::test_khop_out_subgraph[idtype1]
1.08s call     tests/compute/test_subgraph.py::test_in_subgraph[idtype1]
1.07s call     tests/compute/test_subgraph.py::test_in_subgraph[idtype0]
1.07s call     tests/compute/test_heterograph.py::test_view1[idtype1]
1.04s call     tests/compute/test_subgraph.py::test_out_subgraph[idtype1]
1.00s call     tests/compute/test_specialization.py::test_v2v_update_all[idtype0]
0.99s call     tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_rhs-shp2-g1]
0.98s call     tests/compute/test_sparse.py::test_spmm[idtype0-min-mul-shp0-g1]
0.95s call     tests/compute/test_specialization.py::test_v2v_update_all[idtype1]
0.95s call     tests/compute/test_sparse.py::test_spmm[idtype1-max-add-shp0-g1]
0.94s call     tests/compute/test_heterograph.py::test_types_in_function[idtype0]
0.93s call     tests/compute/test_sampling.py::test_uniform_random_walk[False]
0.90s call     tests/compute/test_heterograph.py::test_types_in_function[idtype1]
0.90s call     tests/compute/test_sparse.py::test_spmm[idtype0-sum-copy_lhs-shp4-g1]
0.90s call     tests/compute/test_sparse.py::test_spmm[idtype0-sum-add-shp2-g1]
0.89s call     tests/compute/test_sparse.py::test_spmm[idtype0-sum-div-shp3-g0]
0.88s call     tests/compute/test_sparse.py::test_spmm[idtype0-min-mul-shp3-g0]
0.88s call     tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_rhs-shp1-g1]
0.88s call     tests/compute/test_sparse.py::test_spmm[idtype0-min-mul-shp1-g1]
0.87s call     tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_rhs-shp0-g1]
0.86s call     tests/compute/test_sparse.py::test_spmm[idtype0-sum-add-shp1-g1]
0.86s call     tests/compute/test_basics.py::test_issue_2484[idtype1]
0.86s call     tests/compute/test_sparse.py::test_spmm[idtype0-sum-add-shp1-g0]
0.86s call     tests/compute/test_sparse.py::test_spmm[idtype0-min-sub-shp4-g1]
0.86s call     tests/compute/test_sparse.py::test_spmm[idtype1-sum-add-shp0-g1]
0.85s call     tests/compute/test_sparse.py::test_spmm[idtype0-sum-mul-shp1-g1]
0.85s call     tests/compute/test_sparse.py::test_spmm[idtype0-min-mul-shp1-g0]
0.85s call     tests/compute/test_sparse.py::test_spmm[idtype0-sum-add-shp3-g0]
0.84s call     tests/compute/test_sparse.py::test_spmm[idtype0-sum-sub-shp3-g1]
0.84s call     tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_rhs-shp3-g0]
0.84s call     tests/compute/test_basics.py::test_issue_2484[idtype0]
0.83s call     tests/compute/test_sparse.py::test_spmm[idtype0-sum-mul-shp0-g1]
0.83s call     tests/compute/test_sparse.py::test_spmm[idtype0-sum-add-shp3-g1]
0.83s call     tests/compute/test_sparse.py::test_spmm[idtype1-max-add-shp5-g0]
0.83s call     tests/compute/test_sparse.py::test_spmm[idtype0-sum-add-shp0-g1]
0.83s call     tests/compute/test_sparse.py::test_spmm[idtype0-sum-sub-shp3-g0]
0.82s call     tests/compute/test_sparse.py::test_spmm[idtype1-max-sub-shp1-g1]
0.82s call     tests/compute/test_sparse.py::test_spmm[idtype0-sum-mul-shp3-g1]
0.82s call     tests/compute/test_sparse.py::test_spmm[idtype0-min-div-shp1-g1]
0.81s call     tests/compute/test_sparse.py::test_spmm[idtype0-sum-div-shp3-g1]
0.81s call     tests/compute/test_sparse.py::test_spmm[idtype0-min-mul-shp3-g1]
0.81s call     tests/compute/test_sparse.py::test_spmm[idtype0-max-copy_rhs-shp1-g0]
0.81s call     tests/compute/test_sparse.py::test_spmm[idtype0-min-mul-shp4-g0]
0.80s call     tests/compute/test_sparse.py::test_spmm[idtype0-max-sub-shp1-g0]
========= 2308 passed, 150 skipped, 236 warnings in 1246.70s (0:20:46) =========
============================= test session starts ==============================
platform linux -- Python 3.7.0, pytest-7.1.2, pluggy-1.0.0 -- /opt/conda/envs/tensorflow-ci/bin/python3
cachedir: .pytest_cache
rootdir: /root/jenkins/workspace/dgl_PR-4648@2
collecting ... PASSED              [ 99%]
tests/mxnet/test_nn.py::test_hetero_conv[stack-int32] collected 920 items

tests/tensorflow/test_basic.py::test PASSED                              [  0%]
tests/tensorflow/test_nn.py::test_graph_conv[1] PASSED                   [  0%]
tests/tensorflow/test_nn.py::test_graph_conv[2] PASSED                   [  0%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-none-g0-idtype0] PASSED [  0%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-none-g0-idtype1] PASSED [  0%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-none-g1-idtype0] PASSED [  0%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-none-g1-idtype1] PASSED [  0%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-none-g2-idtype0] PASSED [  0%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-none-g2-idtype1] PASSED [  0%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-none-g3-idtype0] PASSED [  1%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-none-g3-idtype1] PASSED [  1%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-none-g4-idtype0] PASSED [  1%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-none-g4-idtype1] PASSED [  1%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-none-g5-idtype0] PASSED [  1%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-none-g5-idtype1] PASSED [  1%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-none-g6-idtype0] PASSED [  1%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-none-g6-idtype1] PASSED [  1%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-both-g0-idtype0] PASSED [  1%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-both-g0-idtype1] PASSED [  2%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-both-g1-idtype0] PASSED [  2%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-both-g1-idtype1] PASSED [  2%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-both-g2-idtype0] PASSED [  2%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-both-g2-idtype1] PASSED [  2%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-both-g3-idtype0] PASSED             [ 99%]
tests/mxnet/test_nn.py::test_hetero_conv[stack-int64] PASSED [  2%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-both-g3-idtype1] PASSED [  2%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-both-g4-idtype0] PASSED [  2%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-both-g4-idtype1] PASSED [  2%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-both-g5-idtype0] PASSED [  3%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-both-g5-idtype1] PASSED [  3%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-both-g6-idtype0] PASSED [  3%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-both-g6-idtype1] PASSED [  3%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-right-g0-idtype0] PASSED [  3%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-right-g0-idtype1] PASSED [  3%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-right-g1-idtype0] PASSED [  3%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-right-g1-idtype1] PASSED [  3%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-right-g2-idtype0] PASSED [  3%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-right-g2-idtype1] PASSED [  4%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-right-g3-idtype0] PASSED [  4%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-right-g3-idtype1] PASSED [  4%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-right-g4-idtype0] PASSED [  4%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-right-g4-idtype1] PASSED [  4%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-right-g5-idtype0] PASSED [  4%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-right-g5-idtype1] PASSED [  4%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-right-g6-idtype0] PASSED [  4%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-right-g6-idtype1] PASSED [  4%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-left-g0-idtype0] PASSED [  5%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-left-g0-idtype1] PASSED [  5%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-left-g1-idtype0] PASSED [  5%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-left-g1-idtype1] PASSED [  5%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-left-g2-idtype0] PASSED [  5%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-left-g2-idtype1] PASSED [  5%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-left-g3-idtype0] PASSED [  5%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-left-g3-idtype1] PASSED [  5%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-left-g4-idtype0] PASSED [  5%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-left-g4-idtype1] PASSED [  5%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-left-g5-idtype0] PASSED [  6%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-left-g5-idtype1] PASSED [  6%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-left-g6-idtype0] PASSED [  6%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-left-g6-idtype1] PASSED [  6%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-none-g0-idtype0] PASSED [  6%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-none-g0-idtype1] PASSED [  6%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-none-g1-idtype0] PASSED [  6%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-none-g1-idtype1] PASSED [  6%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-none-g2-idtype0] PASSED [  6%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-none-g2-idtype1] PASSED [  7%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-none-g3-idtype0] PASSED [  7%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-none-g3-idtype1] PASSED [  7%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-none-g4-idtype0] PASSED [  7%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-none-g4-idtype1] PASSED [  7%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-none-g5-idtype0] PASSED [  7%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-none-g5-idtype1] PASSED [  7%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-none-g6-idtype0] PASSED [  7%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-none-g6-idtype1] PASSED [  7%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-both-g0-idtype0] PASSED [  8%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-both-g0-idtype1] PASSED [  8%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-both-g1-idtype0] PASSED [  8%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-both-g1-idtype1] PASSED [  8%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-both-g2-idtype0] PASSED [  8%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-both-g2-idtype1] PASSED [  8%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-both-g3-idtype0] PASSED [  8%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-both-g3-idtype1] PASSED [  8%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-both-g4-idtype0] PASSED [  8%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-both-g4-idtype1] PASSED [  9%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-both-g5-idtype0] PASSED [  9%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-both-g5-idtype1] PASSED             [ 99%]
tests/mxnet/test_nn.py::test_hetero_conv[myagg-int32] PASSED [  9%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-both-g6-idtype0] PASSED [  9%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-both-g6-idtype1] PASSED [  9%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-right-g0-idtype0] PASSED [  9%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-right-g0-idtype1] PASSED [  9%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-right-g1-idtype0] PASSED [  9%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-right-g1-idtype1] PASSED [  9%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-right-g2-idtype0] PASSED [ 10%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-right-g2-idtype1] PASSED [ 10%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-right-g3-idtype0] PASSED [ 10%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-right-g3-idtype1] PASSED [ 10%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-right-g4-idtype0] PASSED [ 10%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-right-g4-idtype1] PASSED [ 10%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-right-g5-idtype0] PASSED [ 10%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-right-g5-idtype1] PASSED [ 10%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-right-g6-idtype0] PASSED [ 10%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-right-g6-idtype1] PASSED [ 10%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-left-g0-idtype0] PASSED [ 11%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-left-g0-idtype1] PASSED [ 11%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-left-g1-idtype0] PASSED [ 11%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-left-g1-idtype1] PASSED [ 11%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-left-g2-idtype0] PASSED [ 11%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-left-g2-idtype1] PASSED [ 11%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-left-g3-idtype0] PASSED [ 11%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-left-g3-idtype1] PASSED [ 11%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-left-g4-idtype0] PASSED [ 11%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-left-g4-idtype1] PASSED [ 12%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-left-g5-idtype0] PASSED [ 12%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-left-g5-idtype1] PASSED [ 12%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-left-g6-idtype0] PASSED [ 12%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-True-False-left-g6-idtype1] PASSED [ 12%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-none-g0-idtype0] PASSED [ 12%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-none-g0-idtype1] PASSED [ 12%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-none-g1-idtype0] PASSED [ 12%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-none-g1-idtype1] PASSED [ 12%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-none-g2-idtype0] PASSED [ 13%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-none-g2-idtype1] PASSED [ 13%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-none-g3-idtype0] PASSED [ 13%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-none-g3-idtype1] PASSED [ 13%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-none-g4-idtype0] PASSED [ 13%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-none-g4-idtype1] PASSED [ 13%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-none-g5-idtype0] PASSED [ 13%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-none-g5-idtype1] PASSED [ 13%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-none-g6-idtype0] PASSED [ 13%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-none-g6-idtype1] PASSED [ 14%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-both-g0-idtype0] PASSED [ 14%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-both-g0-idtype1] PASSED [ 14%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-both-g1-idtype0] PASSED [ 14%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-both-g1-idtype1] PASSED [ 14%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-both-g2-idtype0] PASSED [ 14%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-both-g2-idtype1] PASSED [ 14%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-both-g3-idtype0] PASSED [ 14%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-both-g3-idtype1] PASSED [ 14%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-both-g4-idtype0] PASSED [ 15%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-both-g4-idtype1] PASSED [ 15%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-both-g5-idtype0] PASSED [ 15%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-both-g5-idtype1] PASSED [ 15%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-both-g6-idtype0] PASSED [ 15%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-both-g6-idtype1] PASSED [ 15%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-right-g0-idtype0] PASSED [ 15%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-right-g0-idtype1] PASSED [ 15%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-right-g1-idtype0] PASSED [ 15%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-right-g1-idtype1] PASSED [ 15%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-right-g2-idtype0] PASSED [ 16%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-right-g2-idtype1] PASSED [ 16%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-right-g3-idtype0] PASSED [ 16%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-right-g3-idtype1] PASSED [ 16%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-right-g4-idtype0] PASSED [ 16%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-right-g4-idtype1] PASSED [ 16%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-right-g5-idtype0] PASSED [ 16%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-right-g5-idtype1] PASSED [ 16%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-right-g6-idtype0] PASSED [ 16%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-right-g6-idtype1] PASSED [ 17%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-left-g0-idtype0] PASSED [ 17%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-left-g0-idtype1] PASSED [ 17%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-left-g1-idtype0] PASSED [ 17%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-left-g1-idtype1] PASSED [ 17%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-left-g2-idtype0] PASSED [ 17%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-left-g2-idtype1] PASSED [ 17%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-left-g3-idtype0] PASSED [ 17%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-left-g3-idtype1] PASSED [ 17%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-left-g4-idtype0] PASSED [ 18%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-left-g4-idtype1] PASSED [ 18%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-left-g5-idtype0] PASSED [ 18%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-left-g5-idtype1] PASSED [ 18%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-left-g6-idtype0] PASSED [ 18%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-True-left-g6-idtype1] PASSED [ 18%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-none-g0-idtype0] PASSED [ 18%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-none-g0-idtype1] PASSED [ 18%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-none-g1-idtype0] PASSED [ 18%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-none-g1-idtype1] PASSED [ 19%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-none-g2-idtype0] PASSED [ 19%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-none-g2-idtype1] PASSED [ 19%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-none-g3-idtype0] PASSED [ 19%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-none-g3-idtype1] PASSED             [ 99%]
tests/mxnet/test_nn.py::test_hetero_conv[myagg-int64] PASSED [ 19%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-none-g4-idtype0] PASSED [ 19%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-none-g4-idtype1] PASSED [ 19%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-none-g5-idtype0] PASSED [ 19%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-none-g5-idtype1] PASSED [ 19%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-none-g6-idtype0] PASSED [ 20%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-none-g6-idtype1] PASSED [ 20%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-both-g0-idtype0] PASSED [ 20%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-both-g0-idtype1] PASSED [ 20%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-both-g1-idtype0] PASSED [ 20%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-both-g1-idtype1] PASSED [ 20%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-both-g2-idtype0] PASSED [ 20%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-both-g2-idtype1] PASSED [ 20%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-both-g3-idtype0] PASSED [ 20%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-both-g3-idtype1] PASSED [ 20%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-both-g4-idtype0] PASSED [ 21%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-both-g4-idtype1] PASSED [ 21%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-both-g5-idtype0] PASSED [ 21%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-both-g5-idtype1] PASSED [ 21%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-both-g6-idtype0] PASSED [ 21%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-both-g6-idtype1] PASSED [ 21%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-right-g0-idtype0] PASSED [ 21%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-right-g0-idtype1] PASSED [ 21%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-right-g1-idtype0] PASSED [ 21%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-right-g1-idtype1] PASSED [ 22%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-right-g2-idtype0] PASSED [ 22%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-right-g2-idtype1] PASSED [ 22%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-right-g3-idtype0] PASSED [ 22%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-right-g3-idtype1] PASSED [ 22%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-right-g4-idtype0] PASSED [ 22%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-right-g4-idtype1] PASSED [ 22%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-right-g5-idtype0] PASSED [ 22%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-right-g5-idtype1] PASSED [ 22%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-right-g6-idtype0] PASSED [ 23%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-right-g6-idtype1] PASSED [ 23%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-left-g0-idtype0] PASSED [ 23%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-left-g0-idtype1] PASSED [ 23%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-left-g1-idtype0] PASSED [ 23%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-left-g1-idtype1] PASSED [ 23%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-left-g2-idtype0] PASSED [ 23%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-left-g2-idtype1] PASSED [ 23%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-left-g3-idtype0] PASSED [ 23%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-left-g3-idtype1] PASSED [ 24%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-left-g4-idtype0] PASSED [ 24%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-left-g4-idtype1] PASSED [ 24%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-left-g5-idtype0] PASSED [ 24%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-left-g5-idtype1] PASSED [ 24%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-left-g6-idtype0] PASSED [ 24%]
tests/tensorflow/test_nn.py::test_graph_conv2[1-False-False-left-g6-idtype1] PASSED [ 24%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-none-g0-idtype0] PASSED [ 24%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-none-g0-idtype1] PASSED [ 24%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-none-g1-idtype0] PASSED [ 25%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-none-g1-idtype1] PASSED [ 25%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-none-g2-idtype0] PASSED [ 25%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-none-g2-idtype1] PASSED [ 25%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-none-g3-idtype0] PASSED [ 25%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-none-g3-idtype1] PASSED [ 25%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-none-g4-idtype0] PASSED [ 25%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-none-g4-idtype1] PASSED [ 25%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-none-g5-idtype0] PASSED [ 25%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-none-g5-idtype1] PASSED [ 25%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-none-g6-idtype0] PASSED [ 26%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-none-g6-idtype1] PASSED [ 26%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-both-g0-idtype0] PASSED [ 26%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-both-g0-idtype1] PASSED [ 26%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-both-g1-idtype0] PASSED [ 26%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-both-g1-idtype1] PASSED [ 26%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-both-g2-idtype0] PASSED [ 26%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-both-g2-idtype1] PASSED [ 26%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-both-g3-idtype0] PASSED [ 26%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-both-g3-idtype1] PASSED [ 27%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-both-g4-idtype0] PASSED [ 27%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-both-g4-idtype1] PASSED [ 27%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-both-g5-idtype0] PASSED [ 27%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-both-g5-idtype1] PASSED [ 27%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-both-g6-idtype0] PASSED [ 27%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-both-g6-idtype1] PASSED [ 27%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-right-g0-idtype0] PASSED [ 27%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-right-g0-idtype1] PASSED [ 27%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-right-g1-idtype0] PASSED             [100%]

=============================== warnings summary ===============================
tests/mxnet/test_nn.py::test_cheb_conv[1]
tests/mxnet/test_nn.py::test_cheb_conv[20]
  /root/jenkins/workspace/dgl_PR-4648@4/python/dgl/nn/mxnet/conv/chebconv.py:121: DGLWarning: lambda_max is not provided, using default value of 2.  Please use dgl.laplacian_lambda_max to compute the eigenvalues.
    "lambda_max is not provided, using default value of 2.  "

-- Docs: https://docs.pytest.org/en/stable/how-to/capture-warnings.html
- generated xml file: /root/jenkins/workspace/dgl_PR-4648@4/pytest_backend.xml -
============================ slowest 100 durations =============================
7.56s call     tests/mxnet/test_nn.py::test_hetero_conv[stack-int32]
7.14s call     tests/mxnet/test_nn.py::test_hetero_conv[min-int32]
6.67s call     tests/mxnet/test_nn.py::test_hetero_conv[mean-int32]
6.57s call     tests/mxnet/test_nn.py::test_hetero_conv[min-int64]
6.41s call     tests/mxnet/test_nn.py::test_hetero_conv[stack-int64]
6.07s call     tests/mxnet/test_nn.py::test_hetero_conv[max-int64]
5.94s call     tests/mxnet/test_nn.py::test_hetero_conv[mean-int64]
5.66s call     tests/mxnet/test_nn.py::test_hetero_conv[myagg-int32]
5.35s call     tests/mxnet/test_nn.py::test_hetero_conv[myagg-int64]
3.10s call     tests/mxnet/test_nn.py::test_hetero_conv[sum-int64]
2.41s call     tests/mxnet/test_nn.py::test_hetero_conv[max-int32]
2.03s call     tests/mxnet/test_nn.py::test_hetero_conv[sum-int32]
1.57s call     tests/mxnet/test_nn.py::test_gg_conv
1.50s call     tests/mxnet/test_nn.py::test_gat_conv[1-1-g6-int64]
1.47s call     tests/mxnet/test_nn.py::test_gat_conv[1-1-g5-int64]
1.43s call     tests/mxnet/test_nn.py::test_gat_conv[1-1-g6-int32]
1.39s call     tests/mxnet/test_nn.py::test_gat_conv[1-1-g3-int64]
1.36s call     tests/mxnet/test_nn.py::test_gat_conv[1-1-g3-int32]
1.33s call     tests/mxnet/test_nn.py::test_gat_conv[1-1-g4-int64]
1.32s call     tests/mxnet/test_nn.py::test_gat_conv[1-1-g4-int32]
1.27s call     tests/mxnet/test_nn.py::test_gat_conv[1-1-g2-int64]
1.26s call     tests/mxnet/test_nn.py::test_gat_conv[1-20-g1-int32]
1.24s call     tests/mxnet/test_nn.py::test_gat_conv[1-1-g1-int32]
1.22s call     tests/mxnet/test_nn.py::test_gat_conv[1-20-g1-int64]
1.22s call     tests/mxnet/test_nn.py::test_gat_conv[1-1-g0-int64]
1.20s call     tests/mxnet/test_nn.py::test_gat_conv[1-1-g5-int32]
1.20s call     tests/mxnet/test_nn.py::test_gat_conv[1-20-g0-int32]
1.20s call     tests/mxnet/test_nn.py::test_gat_conv[1-1-g2-int32]
1.19s call     tests/mxnet/test_nn.py::test_gat_conv[1-1-g1-int64]
1.13s call     tests/mxnet/test_nn.py::test_gat_conv[1-20-g2-int32]
1.13s call     tests/mxnet/test_nn.py::test_gat_conv[1-1-g0-int32]
1.12s call     tests/mxnet/test_nn.py::test_gat_conv[1-20-g3-int64]
1.10s call     tests/mxnet/test_nn.py::test_gat_conv[1-20-g0-int64]
1.06s call     tests/mxnet/test_nn.py::test_gat_conv[1-20-g4-int32]
1.04s call     tests/mxnet/test_nn.py::test_gat_conv[1-20-g5-int32]
1.04s call     tests/mxnet/test_nn.py::test_gat_conv[1-20-g4-int64]
1.03s call     tests/mxnet/test_nn.py::test_gat_conv[1-20-g3-int32]
0.99s call     tests/mxnet/test_nn.py::test_gat_conv[1-20-g2-int64]
0.86s call     tests/mxnet/test_nn.py::test_gat_conv[1-20-g6-int32]
0.83s call     tests/mxnet/test_nn.py::test_gat_conv[1-20-g5-int64]
0.83s call     tests/mxnet/test_nn.py::test_dense_cheb_conv[2]
0.83s call     tests/mxnet/test_nn.py::test_dense_cheb_conv[1]
0.70s call     tests/mxnet/test_nn.py::test_tagconv[2]
0.68s call     tests/mxnet/test_nn.py::test_gat_conv[1-20-g6-int64]
0.62s call     tests/mxnet/test_nn.py::test_tagconv[1]
0.59s call     tests/mxnet/test_nn.py::test_gat_conv[5-1-g0-int32]
0.54s call     tests/mxnet/test_nn.py::test_gat_conv[5-1-g0-int64]
0.54s call     tests/mxnet/test_nn.py::test_gat_conv[5-1-g3-int64]
0.53s call     tests/mxnet/test_nn.py::test_gat_conv[5-1-g4-int32]
0.53s call     tests/mxnet/test_nn.py::test_gat_conv[5-20-g0-int64]
0.53s call     tests/mxnet/test_nn.py::test_gat_conv[5-1-g2-int32]
0.53s call     tests/mxnet/test_nn.py::test_gat_conv[5-1-g5-int64]
0.53s call     tests/mxnet/test_nn.py::test_gat_conv[5-1-g3-int32]
0.53s call     tests/mxnet/test_nn.py::test_gat_conv[5-1-g2-int64]
0.52s call     tests/mxnet/test_nn.py::test_gat_conv[5-1-g6-int32]
0.52s call     tests/mxnet/test_nn.py::test_gat_conv[5-1-g1-int32]
0.51s call     tests/mxnet/test_nn.py::test_gat_conv[5-1-g6-int64]
0.51s call     tests/mxnet/test_nn.py::test_gat_conv[5-1-g1-int64]
0.51s call     tests/mxnet/test_nn.py::test_gat_conv[5-1-g4-int64]
0.50s call     tests/mxnet/test_nn.py::test_gat_conv[5-1-g5-int32]
0.50s call     tests/mxnet/test_nn.py::test_gat_conv[5-20-g4-int64]
0.50s call     tests/mxnet/test_nn.py::test_gat_conv[5-20-g5-int64]
0.49s call     tests/mxnet/test_nn.py::test_gat_conv[5-20-g2-int32]
0.49s call     tests/mxnet/test_nn.py::test_gat_conv[5-20-g1-int32]
0.49s call     tests/mxnet/test_nn.py::test_gat_conv[5-20-g0-int32]
0.49s call     tests/mxnet/test_nn.py::test_gat_conv[5-20-g1-int64]
0.49s call     tests/mxnet/test_nn.py::test_gat_conv[5-20-g5-int32]
0.49s call     tests/mxnet/test_nn.py::test_gat_conv[5-20-g3-int64]
0.49s call     tests/mxnet/test_nn.py::test_gat_conv[5-20-g6-int64]
0.48s call     tests/mxnet/test_nn.py::test_gat_conv[5-20-g2-int64]
0.48s call     tests/mxnet/test_nn.py::test_gat_conv[5-20-g4-int32]
0.48s call     tests/mxnet/test_nn.py::test_gat_conv[5-20-g6-int32]
0.48s call     tests/mxnet/test_nn.py::test_gat_conv[5-20-g3-int32]
0.40s call     tests/mxnet/test_nn.py::test_gat_conv_bi[1-1-g0-int64]
0.39s call     tests/mxnet/test_nn.py::test_gat_conv_bi[1-1-g0-int32]
0.38s call     tests/mxnet/test_nn.py::test_gat_conv_bi[1-1-g1-int32]
0.38s call     tests/mxnet/test_nn.py::test_gat_conv_bi[1-1-g1-int64]
0.37s call     tests/mxnet/test_nn.py::test_graph_conv2_bi[1-False-True-both-g0-int32]
0.37s call     tests/mxnet/test_nn.py::test_gat_conv_bi[4-1-g1-int64]
0.35s call     tests/mxnet/test_nn.py::test_gat_conv_bi[4-1-g0-int64]
0.34s call     tests/mxnet/test_nn.py::test_gat_conv_bi[4-1-g0-int32]
0.34s call     tests/mxnet/test_nn.py::test_gat_conv_bi[4-2-g1-int32]
0.33s call     tests/mxnet/test_nn.py::test_gat_conv_bi[4-1-g1-int32]
0.33s call     tests/mxnet/test_nn.py::test_dense_graph_conv[1-g0-both-int32]
0.33s call     tests/mxnet/test_nn.py::test_dense_graph_conv[1-g1-both-int32]
0.33s call     tests/mxnet/test_nn.py::test_dense_graph_conv[1-g2-both-int32]
0.32s call     tests/mxnet/test_nn.py::test_dense_graph_conv[1-g5-both-int32]
0.32s call     tests/mxnet/test_nn.py::test_gat_conv_bi[4-2-g0-int64]
0.32s call     tests/mxnet/test_nn.py::test_dense_graph_conv[1-g1-both-int64]
0.32s call     tests/mxnet/test_nn.py::test_gat_conv_bi[4-2-g0-int32]
0.32s call     tests/mxnet/test_nn.py::test_gat_conv_bi[1-2-g1-int64]
0.32s call     tests/mxnet/test_nn.py::test_dense_graph_conv[1-g0-both-int64]
0.32s call     tests/mxnet/test_nn.py::test_gat_conv_bi[1-2-g0-int32]
0.32s call     tests/mxnet/test_nn.py::test_dense_graph_conv[1-g3-both-int32]
0.32s call     tests/mxnet/test_nn.py::test_gat_conv_bi[4-2-g1-int64]
0.32s call     tests/mxnet/test_nn.py::test_gat_conv_bi[1-2-g1-int32]
0.31s call     tests/mxnet/test_nn.py::test_graph_conv2_bi[1-False-True-both-g0-int64]
0.31s call     tests/mxnet/test_nn.py::test_dense_graph_conv[1-g2-both-int64]
0.31s call     tests/mxnet/test_nn.py::test_dense_graph_conv[1-g4-both-int32]
0.31s call     tests/mxnet/test_nn.py::test_gat_conv_bi[1-2-g0-int64]
================= 806 passed, 2 warnings in 177.66s (0:02:57) ==================
[Pipeline] }
[Pipeline] // timeout
PASSED [ 28%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-right-g1-idtype1] PASSED [ 28%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-right-g2-idtype0] PASSED [ 28%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-right-g2-idtype1] PASSED [ 28%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-right-g3-idtype0] [Pipeline] }
[Pipeline] // stage
Post stage
[Pipeline] cleanWs
[WS-CLEANUP] Deleting project workspace...
[WS-CLEANUP] Deferred wipeout is disabled by the job configuration...
PASSED [ 28%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-right-g3-idtype1] PASSED [ 28%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-right-g4-idtype0] PASSED [ 28%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-right-g4-idtype1] PASSED [ 28%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-right-g5-idtype0] PASSED [ 28%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-right-g5-idtype1] PASSED [ 29%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-right-g6-idtype0] PASSED [ 29%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-right-g6-idtype1] PASSED [ 29%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-left-g0-idtype0] PASSED [ 29%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-left-g0-idtype1] PASSED [ 29%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-left-g1-idtype0] PASSED [ 29%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-left-g1-idtype1] PASSED [ 29%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-left-g2-idtype0] PASSED [ 29%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-left-g2-idtype1] PASSED [ 29%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-left-g3-idtype0] PASSED [ 30%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-left-g3-idtype1] [WS-CLEANUP] done
[Pipeline] }
$ docker stop --time=1 6db97419e6b6b2f1427e0e55271c5f8554e9cd989f9b44da6a3861a1be7f5a89
PASSED [ 30%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-left-g4-idtype0] PASSED [ 30%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-left-g4-idtype1] PASSED [ 30%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-left-g5-idtype0] PASSED [ 30%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-left-g5-idtype1] PASSED [ 30%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-left-g6-idtype0] PASSED [ 30%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-True-left-g6-idtype1] PASSED [ 30%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-none-g0-idtype0] PASSED [ 30%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-none-g0-idtype1] PASSED [ 30%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-none-g1-idtype0] PASSED [ 31%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-none-g1-idtype1] PASSED [ 31%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-none-g2-idtype0] PASSED [ 31%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-none-g2-idtype1] PASSED [ 31%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-none-g3-idtype0] PASSED [ 31%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-none-g3-idtype1] PASSED [ 31%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-none-g4-idtype0] PASSED [ 31%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-none-g4-idtype1] PASSED [ 31%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-none-g5-idtype0] PASSED [ 31%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-none-g5-idtype1] PASSED [ 32%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-none-g6-idtype0] PASSED [ 32%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-none-g6-idtype1] PASSED [ 32%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-both-g0-idtype0] PASSED [ 32%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-both-g0-idtype1] PASSED [ 32%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-both-g1-idtype0] PASSED [ 32%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-both-g1-idtype1] PASSED [ 32%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-both-g2-idtype0] PASSED [ 32%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-both-g2-idtype1] PASSED [ 32%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-both-g3-idtype0] PASSED [ 33%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-both-g3-idtype1] PASSED [ 33%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-both-g4-idtype0] PASSED [ 33%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-both-g4-idtype1] PASSED [ 33%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-both-g5-idtype0] PASSED [ 33%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-both-g5-idtype1] PASSED [ 33%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-both-g6-idtype0] PASSED [ 33%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-both-g6-idtype1] PASSED [ 33%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-right-g0-idtype0] PASSED [ 33%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-right-g0-idtype1] PASSED [ 34%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-right-g1-idtype0] PASSED [ 34%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-right-g1-idtype1] PASSED [ 34%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-right-g2-idtype0] PASSED [ 34%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-right-g2-idtype1] PASSED [ 34%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-right-g3-idtype0] PASSED [ 34%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-right-g3-idtype1] PASSED [ 34%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-right-g4-idtype0] PASSED [ 34%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-right-g4-idtype1] PASSED [ 34%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-right-g5-idtype0] PASSED [ 35%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-right-g5-idtype1] PASSED [ 35%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-right-g6-idtype0] PASSED [ 35%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-right-g6-idtype1] $ docker rm -f 6db97419e6b6b2f1427e0e55271c5f8554e9cd989f9b44da6a3861a1be7f5a89
PASSED [ 35%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-left-g0-idtype0] PASSED [ 35%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-left-g0-idtype1] PASSED [ 35%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-left-g1-idtype0] PASSED [ 35%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-left-g1-idtype1] PASSED [ 35%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-left-g2-idtype0] PASSED [ 35%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-left-g2-idtype1] PASSED [ 35%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-left-g3-idtype0] PASSED [ 36%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-left-g3-idtype1] PASSED [ 36%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-left-g4-idtype0] PASSED [ 36%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-left-g4-idtype1] PASSED [ 36%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-left-g5-idtype0] PASSED [ 36%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-left-g5-idtype1] PASSED [ 36%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-left-g6-idtype0] PASSED [ 36%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-True-False-left-g6-idtype1] PASSED [ 36%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-none-g0-idtype0] PASSED [ 36%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-none-g0-idtype1] PASSED [ 37%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-none-g1-idtype0] PASSED [ 37%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-none-g1-idtype1] PASSED [ 37%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-none-g2-idtype0] [Pipeline] // withDockerContainer
[Pipeline] }
[Pipeline] // withEnv
[Pipeline] }
[Pipeline] // node
[Pipeline] }
PASSED [ 37%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-none-g2-idtype1] PASSED [ 37%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-none-g3-idtype0] PASSED [ 37%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-none-g3-idtype1] PASSED [ 37%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-none-g4-idtype0] PASSED [ 37%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-none-g4-idtype1] [Pipeline] // stage
[Pipeline] }
PASSED [ 37%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-none-g5-idtype0] PASSED [ 38%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-none-g5-idtype1] PASSED [ 38%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-none-g6-idtype0] PASSED [ 38%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-none-g6-idtype1] PASSED [ 38%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-both-g0-idtype0] PASSED [ 38%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-both-g0-idtype1] PASSED [ 38%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-both-g1-idtype0] PASSED [ 38%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-both-g1-idtype1] PASSED [ 38%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-both-g2-idtype0] PASSED [ 38%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-both-g2-idtype1] PASSED [ 39%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-both-g3-idtype0] PASSED [ 39%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-both-g3-idtype1] PASSED [ 39%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-both-g4-idtype0] PASSED [ 39%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-both-g4-idtype1] PASSED [ 39%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-both-g5-idtype0] PASSED [ 39%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-both-g5-idtype1] PASSED [ 39%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-both-g6-idtype0] PASSED [ 39%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-both-g6-idtype1] PASSED [ 39%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-right-g0-idtype0] PASSED [ 40%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-right-g0-idtype1] PASSED [ 40%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-right-g1-idtype0] PASSED [ 40%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-right-g1-idtype1] PASSED [ 40%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-right-g2-idtype0] PASSED [ 40%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-right-g2-idtype1] PASSED [ 40%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-right-g3-idtype0] PASSED [ 40%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-right-g3-idtype1] PASSED [ 40%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-right-g4-idtype0] PASSED [ 40%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-right-g4-idtype1] PASSED [ 40%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-right-g5-idtype0] PASSED [ 41%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-right-g5-idtype1] PASSED [ 41%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-right-g6-idtype0] PASSED [ 41%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-right-g6-idtype1] PASSED [ 41%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-left-g0-idtype0] PASSED [ 41%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-left-g0-idtype1] PASSED [ 41%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-left-g1-idtype0] PASSED [ 41%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-left-g1-idtype1] PASSED [ 41%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-left-g2-idtype0] PASSED [ 41%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-left-g2-idtype1] PASSED [ 42%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-left-g3-idtype0] PASSED [ 42%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-left-g3-idtype1] PASSED [ 42%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-left-g4-idtype0] PASSED [ 42%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-left-g4-idtype1] PASSED [ 42%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-left-g5-idtype0] PASSED [ 42%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-left-g5-idtype1] PASSED [ 42%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-left-g6-idtype0] PASSED [ 42%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-True-left-g6-idtype1] PASSED [ 42%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-none-g0-idtype0] PASSED [ 43%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-none-g0-idtype1] PASSED [ 43%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-none-g1-idtype0] PASSED [ 43%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-none-g1-idtype1] PASSED [ 43%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-none-g2-idtype0] PASSED [ 43%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-none-g2-idtype1] PASSED [ 43%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-none-g3-idtype0] PASSED [ 43%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-none-g3-idtype1] PASSED [ 43%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-none-g4-idtype0] PASSED [ 43%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-none-g4-idtype1] PASSED [ 44%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-none-g5-idtype0] PASSED [ 44%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-none-g5-idtype1] PASSED [ 44%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-none-g6-idtype0] PASSED [ 44%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-none-g6-idtype1] PASSED [ 44%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-both-g0-idtype0] PASSED [ 44%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-both-g0-idtype1] PASSED [ 44%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-both-g1-idtype0] PASSED [ 44%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-both-g1-idtype1] PASSED [ 44%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-both-g2-idtype0] PASSED [ 45%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-both-g2-idtype1] PASSED [ 45%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-both-g3-idtype0] PASSED [ 45%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-both-g3-idtype1] PASSED [ 45%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-both-g4-idtype0] PASSED [ 45%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-both-g4-idtype1] PASSED [ 45%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-both-g5-idtype0] PASSED [ 45%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-both-g5-idtype1] PASSED [ 45%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-both-g6-idtype0] PASSED [ 45%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-both-g6-idtype1] PASSED [ 45%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-right-g0-idtype0] PASSED [ 46%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-right-g0-idtype1] PASSED [ 46%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-right-g1-idtype0] PASSED [ 46%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-right-g1-idtype1] PASSED [ 46%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-right-g2-idtype0] PASSED [ 46%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-right-g2-idtype1] PASSED [ 46%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-right-g3-idtype0] PASSED [ 46%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-right-g3-idtype1] PASSED [ 46%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-right-g4-idtype0] PASSED [ 46%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-right-g4-idtype1] PASSED [ 47%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-right-g5-idtype0] PASSED [ 47%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-right-g5-idtype1] PASSED [ 47%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-right-g6-idtype0] PASSED [ 47%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-right-g6-idtype1] PASSED [ 47%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-left-g0-idtype0] PASSED [ 47%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-left-g0-idtype1] PASSED [ 47%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-left-g1-idtype0] PASSED [ 47%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-left-g1-idtype1] PASSED [ 47%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-left-g2-idtype0] PASSED [ 48%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-left-g2-idtype1] PASSED [ 48%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-left-g3-idtype0] PASSED [ 48%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-left-g3-idtype1] PASSED [ 48%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-left-g4-idtype0] PASSED [ 48%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-left-g4-idtype1] PASSED [ 48%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-left-g5-idtype0] PASSED [ 48%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-left-g5-idtype1] PASSED [ 48%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-left-g6-idtype0] PASSED [ 48%]
tests/tensorflow/test_nn.py::test_graph_conv2[2-False-False-left-g6-idtype1] PASSED [ 49%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-True-True-none-g0-idtype0] PASSED [ 49%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-True-True-none-g0-idtype1] PASSED [ 49%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-True-True-none-g1-idtype0] PASSED [ 49%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-True-True-none-g1-idtype1] PASSED [ 49%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-True-True-both-g0-idtype0] PASSED [ 49%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-True-True-both-g0-idtype1] PASSED [ 49%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-True-True-both-g1-idtype0] PASSED [ 49%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-True-True-both-g1-idtype1] PASSED [ 49%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-True-True-right-g0-idtype0] PASSED [ 50%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-True-True-right-g0-idtype1] PASSED [ 50%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-True-True-right-g1-idtype0] PASSED [ 50%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-True-True-right-g1-idtype1] PASSED [ 50%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-True-False-none-g0-idtype0] PASSED [ 50%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-True-False-none-g0-idtype1] PASSED [ 50%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-True-False-none-g1-idtype0] PASSED [ 50%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-True-False-none-g1-idtype1] PASSED [ 50%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-True-False-both-g0-idtype0] PASSED [ 50%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-True-False-both-g0-idtype1] PASSED [ 50%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-True-False-both-g1-idtype0] PASSED [ 51%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-True-False-both-g1-idtype1] PASSED [ 51%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-True-False-right-g0-idtype0] PASSED [ 51%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-True-False-right-g0-idtype1] PASSED [ 51%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-True-False-right-g1-idtype0] PASSED [ 51%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-True-False-right-g1-idtype1] PASSED [ 51%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-False-True-none-g0-idtype0] PASSED [ 51%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-False-True-none-g0-idtype1] PASSED [ 51%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-False-True-none-g1-idtype0] PASSED [ 51%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-False-True-none-g1-idtype1] PASSED [ 52%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-False-True-both-g0-idtype0] PASSED [ 52%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-False-True-both-g0-idtype1] PASSED [ 52%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-False-True-both-g1-idtype0] PASSED [ 52%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-False-True-both-g1-idtype1] PASSED [ 52%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-False-True-right-g0-idtype0] PASSED [ 52%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-False-True-right-g0-idtype1] PASSED [ 52%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-False-True-right-g1-idtype0] PASSED [ 52%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-False-True-right-g1-idtype1] PASSED [ 52%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-False-False-none-g0-idtype0] PASSED [ 53%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-False-False-none-g0-idtype1] PASSED [ 53%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-False-False-none-g1-idtype0] PASSED [ 53%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-False-False-none-g1-idtype1] PASSED [ 53%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-False-False-both-g0-idtype0] PASSED [ 53%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-False-False-both-g0-idtype1] PASSED [ 53%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-False-False-both-g1-idtype0] PASSED [ 53%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-False-False-both-g1-idtype1] PASSED [ 53%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-False-False-right-g0-idtype0] PASSED [ 53%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-False-False-right-g0-idtype1] PASSED [ 54%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-False-False-right-g1-idtype0] PASSED [ 54%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[1-False-False-right-g1-idtype1] PASSED [ 54%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-True-True-none-g0-idtype0] PASSED [ 54%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-True-True-none-g0-idtype1] PASSED [ 54%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-True-True-none-g1-idtype0] PASSED [ 54%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-True-True-none-g1-idtype1] PASSED [ 54%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-True-True-both-g0-idtype0] PASSED [ 54%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-True-True-both-g0-idtype1] PASSED [ 54%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-True-True-both-g1-idtype0] PASSED [ 55%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-True-True-both-g1-idtype1] PASSED [ 55%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-True-True-right-g0-idtype0] PASSED [ 55%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-True-True-right-g0-idtype1] PASSED [ 55%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-True-True-right-g1-idtype0] PASSED [ 55%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-True-True-right-g1-idtype1] PASSED [ 55%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-True-False-none-g0-idtype0] PASSED [ 55%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-True-False-none-g0-idtype1] PASSED [ 55%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-True-False-none-g1-idtype0] PASSED [ 55%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-True-False-none-g1-idtype1] PASSED [ 55%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-True-False-both-g0-idtype0] PASSED [ 56%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-True-False-both-g0-idtype1] PASSED [ 56%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-True-False-both-g1-idtype0] PASSED [ 56%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-True-False-both-g1-idtype1] PASSED [ 56%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-True-False-right-g0-idtype0] PASSED [ 56%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-True-False-right-g0-idtype1] PASSED [ 56%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-True-False-right-g1-idtype0] PASSED [ 56%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-True-False-right-g1-idtype1] PASSED [ 56%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-False-True-none-g0-idtype0] PASSED [ 56%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-False-True-none-g0-idtype1] PASSED [ 57%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-False-True-none-g1-idtype0] PASSED [ 57%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-False-True-none-g1-idtype1] PASSED [ 57%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-False-True-both-g0-idtype0] PASSED [ 57%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-False-True-both-g0-idtype1] PASSED [ 57%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-False-True-both-g1-idtype0] PASSED [ 57%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-False-True-both-g1-idtype1] PASSED [ 57%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-False-True-right-g0-idtype0] PASSED [ 57%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-False-True-right-g0-idtype1] PASSED [ 57%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-False-True-right-g1-idtype0] PASSED [ 58%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-False-True-right-g1-idtype1] PASSED [ 58%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-False-False-none-g0-idtype0] PASSED [ 58%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-False-False-none-g0-idtype1] PASSED [ 58%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-False-False-none-g1-idtype0] PASSED [ 58%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-False-False-none-g1-idtype1] PASSED [ 58%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-False-False-both-g0-idtype0] PASSED [ 58%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-False-False-both-g0-idtype1] PASSED [ 58%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-False-False-both-g1-idtype0] PASSED [ 58%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-False-False-both-g1-idtype1] PASSED [ 59%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-False-False-right-g0-idtype0] PASSED [ 59%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-False-False-right-g0-idtype1] PASSED [ 59%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-False-False-right-g1-idtype0] PASSED [ 59%]
tests/tensorflow/test_nn.py::test_graph_conv2_bi[2-False-False-right-g1-idtype1] PASSED [ 59%]
tests/tensorflow/test_nn.py::test_simple_pool PASSED                     [ 59%]
tests/tensorflow/test_nn.py::test_glob_att_pool PASSED                   [ 59%]
tests/tensorflow/test_nn.py::test_rgcn[1] PASSED                         [ 59%]
tests/tensorflow/test_nn.py::test_rgcn[2] PASSED                         [ 59%]
tests/tensorflow/test_nn.py::test_rgcn[8] PASSED                         [ 60%]
tests/tensorflow/test_nn.py::test_gat_conv[1-1-g0-idtype0] PASSED        [ 60%]
tests/tensorflow/test_nn.py::test_gat_conv[1-1-g0-idtype1] PASSED        [ 60%]
tests/tensorflow/test_nn.py::test_gat_conv[1-1-g1-idtype0] PASSED        [ 60%]
tests/tensorflow/test_nn.py::test_gat_conv[1-1-g1-idtype1] PASSED        [ 60%]
tests/tensorflow/test_nn.py::test_gat_conv[1-1-g2-idtype0] PASSED        [ 60%]
tests/tensorflow/test_nn.py::test_gat_conv[1-1-g2-idtype1] PASSED        [ 60%]
tests/tensorflow/test_nn.py::test_gat_conv[1-1-g3-idtype0] PASSED        [ 60%]
tests/tensorflow/test_nn.py::test_gat_conv[1-1-g3-idtype1] PASSED        [ 60%]
tests/tensorflow/test_nn.py::test_gat_conv[1-1-g4-idtype0] PASSED        [ 60%]
tests/tensorflow/test_nn.py::test_gat_conv[1-1-g4-idtype1] PASSED        [ 61%]
tests/tensorflow/test_nn.py::test_gat_conv[1-1-g5-idtype0] PASSED        [ 61%]
tests/tensorflow/test_nn.py::test_gat_conv[1-1-g5-idtype1] PASSED        [ 61%]
tests/tensorflow/test_nn.py::test_gat_conv[1-1-g6-idtype0] PASSED        [ 61%]
tests/tensorflow/test_nn.py::test_gat_conv[1-1-g6-idtype1] PASSED        [ 61%]
tests/tensorflow/test_nn.py::test_gat_conv[1-2-g0-idtype0] PASSED        [ 61%]
tests/tensorflow/test_nn.py::test_gat_conv[1-2-g0-idtype1] PASSED        [ 61%]
tests/tensorflow/test_nn.py::test_gat_conv[1-2-g1-idtype0] PASSED        [ 61%]
tests/tensorflow/test_nn.py::test_gat_conv[1-2-g1-idtype1] PASSED        [ 61%]
tests/tensorflow/test_nn.py::test_gat_conv[1-2-g2-idtype0] PASSED        [ 62%]
tests/tensorflow/test_nn.py::test_gat_conv[1-2-g2-idtype1] PASSED        [ 62%]
tests/tensorflow/test_nn.py::test_gat_conv[1-2-g3-idtype0] PASSED        [ 62%]
tests/tensorflow/test_nn.py::test_gat_conv[1-2-g3-idtype1] PASSED        [ 62%]
tests/tensorflow/test_nn.py::test_gat_conv[1-2-g4-idtype0] PASSED        [ 62%]
tests/tensorflow/test_nn.py::test_gat_conv[1-2-g4-idtype1] PASSED        [ 62%]
tests/tensorflow/test_nn.py::test_gat_conv[1-2-g5-idtype0] PASSED        [ 62%]
tests/tensorflow/test_nn.py::test_gat_conv[1-2-g5-idtype1] PASSED        [ 62%]
tests/tensorflow/test_nn.py::test_gat_conv[1-2-g6-idtype0] PASSED        [ 62%]
tests/tensorflow/test_nn.py::test_gat_conv[1-2-g6-idtype1] PASSED        [ 63%]
tests/tensorflow/test_nn.py::test_gat_conv[4-1-g0-idtype0] PASSED        [ 63%]
tests/tensorflow/test_nn.py::test_gat_conv[4-1-g0-idtype1] PASSED        [ 63%]
tests/tensorflow/test_nn.py::test_gat_conv[4-1-g1-idtype0] PASSED        [ 63%]
tests/tensorflow/test_nn.py::test_gat_conv[4-1-g1-idtype1] PASSED        [ 63%]
tests/tensorflow/test_nn.py::test_gat_conv[4-1-g2-idtype0] PASSED        [ 63%]
tests/tensorflow/test_nn.py::test_gat_conv[4-1-g2-idtype1] PASSED        [ 63%]
tests/tensorflow/test_nn.py::test_gat_conv[4-1-g3-idtype0] PASSED        [ 63%]
tests/tensorflow/test_nn.py::test_gat_conv[4-1-g3-idtype1] PASSED        [ 63%]
tests/tensorflow/test_nn.py::test_gat_conv[4-1-g4-idtype0] PASSED        [ 64%]
tests/tensorflow/test_nn.py::test_gat_conv[4-1-g4-idtype1] PASSED        [ 64%]
tests/tensorflow/test_nn.py::test_gat_conv[4-1-g5-idtype0] PASSED        [ 64%]
tests/tensorflow/test_nn.py::test_gat_conv[4-1-g5-idtype1] PASSED        [ 64%]
tests/tensorflow/test_nn.py::test_gat_conv[4-1-g6-idtype0] PASSED        [ 64%]
tests/tensorflow/test_nn.py::test_gat_conv[4-1-g6-idtype1] PASSED        [ 64%]
tests/tensorflow/test_nn.py::test_gat_conv[4-2-g0-idtype0] PASSED        [ 64%]
tests/tensorflow/test_nn.py::test_gat_conv[4-2-g0-idtype1] PASSED        [ 64%]
tests/tensorflow/test_nn.py::test_gat_conv[4-2-g1-idtype0] PASSED        [ 64%]
tests/tensorflow/test_nn.py::test_gat_conv[4-2-g1-idtype1] PASSED        [ 65%]
tests/tensorflow/test_nn.py::test_gat_conv[4-2-g2-idtype0] PASSED        [ 65%]
tests/tensorflow/test_nn.py::test_gat_conv[4-2-g2-idtype1] PASSED        [ 65%]
tests/tensorflow/test_nn.py::test_gat_conv[4-2-g3-idtype0] PASSED        [ 65%]
tests/tensorflow/test_nn.py::test_gat_conv[4-2-g3-idtype1] PASSED        [ 65%]
tests/tensorflow/test_nn.py::test_gat_conv[4-2-g4-idtype0] PASSED        [ 65%]
tests/tensorflow/test_nn.py::test_gat_conv[4-2-g4-idtype1] PASSED        [ 65%]
tests/tensorflow/test_nn.py::test_gat_conv[4-2-g5-idtype0] PASSED        [ 65%]
tests/tensorflow/test_nn.py::test_gat_conv[4-2-g5-idtype1] PASSED        [ 65%]
tests/tensorflow/test_nn.py::test_gat_conv[4-2-g6-idtype0] PASSED        [ 65%]
tests/tensorflow/test_nn.py::test_gat_conv[4-2-g6-idtype1] PASSED        [ 66%]
tests/tensorflow/test_nn.py::test_gat_conv_bi[1-1-g0-idtype0] PASSED     [ 66%]
tests/tensorflow/test_nn.py::test_gat_conv_bi[1-1-g0-idtype1] PASSED     [ 66%]
tests/tensorflow/test_nn.py::test_gat_conv_bi[1-1-g1-idtype0] PASSED     [ 66%]
tests/tensorflow/test_nn.py::test_gat_conv_bi[1-1-g1-idtype1] PASSED     [ 66%]
tests/tensorflow/test_nn.py::test_gat_conv_bi[1-2-g0-idtype0] PASSED     [ 66%]
tests/tensorflow/test_nn.py::test_gat_conv_bi[1-2-g0-idtype1] PASSED     [ 66%]
tests/tensorflow/test_nn.py::test_gat_conv_bi[1-2-g1-idtype0] PASSED     [ 66%]
tests/tensorflow/test_nn.py::test_gat_conv_bi[1-2-g1-idtype1] PASSED     [ 66%]
tests/tensorflow/test_nn.py::test_gat_conv_bi[4-1-g0-idtype0] PASSED     [ 67%]
tests/tensorflow/test_nn.py::test_gat_conv_bi[4-1-g0-idtype1] PASSED     [ 67%]
tests/tensorflow/test_nn.py::test_gat_conv_bi[4-1-g1-idtype0] PASSED     [ 67%]
tests/tensorflow/test_nn.py::test_gat_conv_bi[4-1-g1-idtype1] PASSED     [ 67%]
tests/tensorflow/test_nn.py::test_gat_conv_bi[4-2-g0-idtype0] PASSED     [ 67%]
tests/tensorflow/test_nn.py::test_gat_conv_bi[4-2-g0-idtype1] PASSED     [ 67%]
tests/tensorflow/test_nn.py::test_gat_conv_bi[4-2-g1-idtype0] PASSED     [ 67%]
tests/tensorflow/test_nn.py::test_gat_conv_bi[4-2-g1-idtype1] PASSED     [ 67%]
tests/tensorflow/test_nn.py::test_sage_conv[1-mean-g0-idtype0] PASSED    [ 67%]
tests/tensorflow/test_nn.py::test_sage_conv[1-mean-g0-idtype1] PASSED    [ 68%]
tests/tensorflow/test_nn.py::test_sage_conv[1-mean-g1-idtype0] PASSED    [ 68%]
tests/tensorflow/test_nn.py::test_sage_conv[1-mean-g1-idtype1] PASSED    [ 68%]
tests/tensorflow/test_nn.py::test_sage_conv[1-mean-g2-idtype0] PASSED    [ 68%]
tests/tensorflow/test_nn.py::test_sage_conv[1-mean-g2-idtype1] PASSED    [ 68%]
tests/tensorflow/test_nn.py::test_sage_conv[1-mean-g3-idtype0] PASSED    [ 68%]
tests/tensorflow/test_nn.py::test_sage_conv[1-mean-g3-idtype1] PASSED    [ 68%]
tests/tensorflow/test_nn.py::test_sage_conv[1-mean-g4-idtype0] PASSED    [ 68%]
tests/tensorflow/test_nn.py::test_sage_conv[1-mean-g4-idtype1] PASSED    [ 68%]
tests/tensorflow/test_nn.py::test_sage_conv[1-mean-g5-idtype0] PASSED    [ 69%]
tests/tensorflow/test_nn.py::test_sage_conv[1-mean-g5-idtype1] PASSED    [ 69%]
tests/tensorflow/test_nn.py::test_sage_conv[1-mean-g6-idtype0] PASSED    [ 69%]
tests/tensorflow/test_nn.py::test_sage_conv[1-mean-g6-idtype1] PASSED    [ 69%]
tests/tensorflow/test_nn.py::test_sage_conv[1-mean-g7-idtype0] PASSED    [ 69%]
tests/tensorflow/test_nn.py::test_sage_conv[1-mean-g7-idtype1] PASSED    [ 69%]
tests/tensorflow/test_nn.py::test_sage_conv[1-pool-g0-idtype0] PASSED    [ 69%]
tests/tensorflow/test_nn.py::test_sage_conv[1-pool-g0-idtype1] PASSED    [ 69%]
tests/tensorflow/test_nn.py::test_sage_conv[1-pool-g1-idtype0] PASSED    [ 69%]
tests/tensorflow/test_nn.py::test_sage_conv[1-pool-g1-idtype1] PASSED    [ 70%]
tests/tensorflow/test_nn.py::test_sage_conv[1-pool-g2-idtype0] PASSED    [ 70%]
tests/tensorflow/test_nn.py::test_sage_conv[1-pool-g2-idtype1] PASSED    [ 70%]
tests/tensorflow/test_nn.py::test_sage_conv[1-pool-g3-idtype0] PASSED    [ 70%]
tests/tensorflow/test_nn.py::test_sage_conv[1-pool-g3-idtype1] PASSED    [ 70%]
tests/tensorflow/test_nn.py::test_sage_conv[1-pool-g4-idtype0] PASSED    [ 70%]
tests/tensorflow/test_nn.py::test_sage_conv[1-pool-g4-idtype1] PASSED    [ 70%]
tests/tensorflow/test_nn.py::test_sage_conv[1-pool-g5-idtype0] PASSED    [ 70%]
tests/tensorflow/test_nn.py::test_sage_conv[1-pool-g5-idtype1] PASSED    [ 70%]
tests/tensorflow/test_nn.py::test_sage_conv[1-pool-g6-idtype0] PASSED    [ 70%]
tests/tensorflow/test_nn.py::test_sage_conv[1-pool-g6-idtype1] PASSED    [ 71%]
tests/tensorflow/test_nn.py::test_sage_conv[1-pool-g7-idtype0] PASSED    [ 71%]
tests/tensorflow/test_nn.py::test_sage_conv[1-pool-g7-idtype1] PASSED    [ 71%]
tests/tensorflow/test_nn.py::test_sage_conv[1-gcn-g0-idtype0] PASSED     [ 71%]
tests/tensorflow/test_nn.py::test_sage_conv[1-gcn-g0-idtype1] PASSED     [ 71%]
tests/tensorflow/test_nn.py::test_sage_conv[1-gcn-g1-idtype0] PASSED     [ 71%]
tests/tensorflow/test_nn.py::test_sage_conv[1-gcn-g1-idtype1] PASSED     [ 71%]
tests/tensorflow/test_nn.py::test_sage_conv[1-gcn-g2-idtype0] PASSED     [ 71%]
tests/tensorflow/test_nn.py::test_sage_conv[1-gcn-g2-idtype1] PASSED     [ 71%]
tests/tensorflow/test_nn.py::test_sage_conv[1-gcn-g3-idtype0] PASSED     [ 72%]
tests/tensorflow/test_nn.py::test_sage_conv[1-gcn-g3-idtype1] PASSED     [ 72%]
tests/tensorflow/test_nn.py::test_sage_conv[1-gcn-g4-idtype0] PASSED     [ 72%]
tests/tensorflow/test_nn.py::test_sage_conv[1-gcn-g4-idtype1] PASSED     [ 72%]
tests/tensorflow/test_nn.py::test_sage_conv[1-gcn-g5-idtype0] PASSED     [ 72%]
tests/tensorflow/test_nn.py::test_sage_conv[1-gcn-g5-idtype1] PASSED     [ 72%]
tests/tensorflow/test_nn.py::test_sage_conv[1-gcn-g6-idtype0] PASSED     [ 72%]
tests/tensorflow/test_nn.py::test_sage_conv[1-gcn-g6-idtype1] PASSED     [ 72%]
tests/tensorflow/test_nn.py::test_sage_conv[1-gcn-g7-idtype0] PASSED     [ 72%]
tests/tensorflow/test_nn.py::test_sage_conv[1-gcn-g7-idtype1] PASSED     [ 73%]
tests/tensorflow/test_nn.py::test_sage_conv[10-mean-g0-idtype0] PASSED   [ 73%]
tests/tensorflow/test_nn.py::test_sage_conv[10-mean-g0-idtype1] PASSED   [ 73%]
tests/tensorflow/test_nn.py::test_sage_conv[10-mean-g1-idtype0] PASSED   [ 73%]
tests/tensorflow/test_nn.py::test_sage_conv[10-mean-g1-idtype1] PASSED   [ 73%]
tests/tensorflow/test_nn.py::test_sage_conv[10-mean-g2-idtype0] PASSED   [ 73%]
tests/tensorflow/test_nn.py::test_sage_conv[10-mean-g2-idtype1] PASSED   [ 73%]
tests/tensorflow/test_nn.py::test_sage_conv[10-mean-g3-idtype0] PASSED   [ 73%]
tests/tensorflow/test_nn.py::test_sage_conv[10-mean-g3-idtype1] PASSED   [ 73%]
tests/tensorflow/test_nn.py::test_sage_conv[10-mean-g4-idtype0] PASSED   [ 74%]
tests/tensorflow/test_nn.py::test_sage_conv[10-mean-g4-idtype1] PASSED   [ 74%]
tests/tensorflow/test_nn.py::test_sage_conv[10-mean-g5-idtype0] PASSED   [ 74%]
tests/tensorflow/test_nn.py::test_sage_conv[10-mean-g5-idtype1] PASSED   [ 74%]
tests/tensorflow/test_nn.py::test_sage_conv[10-mean-g6-idtype0] PASSED   [ 74%]
tests/tensorflow/test_nn.py::test_sage_conv[10-mean-g6-idtype1] PASSED   [ 74%]
tests/tensorflow/test_nn.py::test_sage_conv[10-mean-g7-idtype0] PASSED   [ 74%]
tests/tensorflow/test_nn.py::test_sage_conv[10-mean-g7-idtype1] PASSED   [ 74%]
tests/tensorflow/test_nn.py::test_sage_conv[10-pool-g0-idtype0] PASSED   [ 74%]
tests/tensorflow/test_nn.py::test_sage_conv[10-pool-g0-idtype1] PASSED   [ 75%]
tests/tensorflow/test_nn.py::test_sage_conv[10-pool-g1-idtype0] PASSED   [ 75%]
tests/tensorflow/test_nn.py::test_sage_conv[10-pool-g1-idtype1] PASSED   [ 75%]
tests/tensorflow/test_nn.py::test_sage_conv[10-pool-g2-idtype0] PASSED   [ 75%]
tests/tensorflow/test_nn.py::test_sage_conv[10-pool-g2-idtype1] PASSED   [ 75%]
tests/tensorflow/test_nn.py::test_sage_conv[10-pool-g3-idtype0] PASSED   [ 75%]
tests/tensorflow/test_nn.py::test_sage_conv[10-pool-g3-idtype1] PASSED   [ 75%]
tests/tensorflow/test_nn.py::test_sage_conv[10-pool-g4-idtype0] PASSED   [ 75%]
tests/tensorflow/test_nn.py::test_sage_conv[10-pool-g4-idtype1] PASSED   [ 75%]
tests/tensorflow/test_nn.py::test_sage_conv[10-pool-g5-idtype0] PASSED   [ 75%]
tests/tensorflow/test_nn.py::test_sage_conv[10-pool-g5-idtype1] PASSED   [ 76%]
tests/tensorflow/test_nn.py::test_sage_conv[10-pool-g6-idtype0] PASSED   [ 76%]
tests/tensorflow/test_nn.py::test_sage_conv[10-pool-g6-idtype1] PASSED   [ 76%]
tests/tensorflow/test_nn.py::test_sage_conv[10-pool-g7-idtype0] PASSED   [ 76%]
tests/tensorflow/test_nn.py::test_sage_conv[10-pool-g7-idtype1] PASSED   [ 76%]
tests/tensorflow/test_nn.py::test_sage_conv[10-gcn-g0-idtype0] PASSED    [ 76%]
tests/tensorflow/test_nn.py::test_sage_conv[10-gcn-g0-idtype1] PASSED    [ 76%]
tests/tensorflow/test_nn.py::test_sage_conv[10-gcn-g1-idtype0] PASSED    [ 76%]
tests/tensorflow/test_nn.py::test_sage_conv[10-gcn-g1-idtype1] PASSED    [ 76%]
tests/tensorflow/test_nn.py::test_sage_conv[10-gcn-g2-idtype0] PASSED    [ 77%]
tests/tensorflow/test_nn.py::test_sage_conv[10-gcn-g2-idtype1] PASSED    [ 77%]
tests/tensorflow/test_nn.py::test_sage_conv[10-gcn-g3-idtype0] PASSED    [ 77%]
tests/tensorflow/test_nn.py::test_sage_conv[10-gcn-g3-idtype1] PASSED    [ 77%]
tests/tensorflow/test_nn.py::test_sage_conv[10-gcn-g4-idtype0] PASSED    [ 77%]
tests/tensorflow/test_nn.py::test_sage_conv[10-gcn-g4-idtype1] PASSED    [ 77%]
tests/tensorflow/test_nn.py::test_sage_conv[10-gcn-g5-idtype0] PASSED    [ 77%]
tests/tensorflow/test_nn.py::test_sage_conv[10-gcn-g5-idtype1] PASSED    [ 77%]
tests/tensorflow/test_nn.py::test_sage_conv[10-gcn-g6-idtype0] PASSED    [ 77%]
tests/tensorflow/test_nn.py::test_sage_conv[10-gcn-g6-idtype1] PASSED    [ 78%]
tests/tensorflow/test_nn.py::test_sage_conv[10-gcn-g7-idtype0] PASSED    [ 78%]
tests/tensorflow/test_nn.py::test_sage_conv[10-gcn-g7-idtype1] PASSED    [ 78%]
tests/tensorflow/test_nn.py::test_sage_conv_bi[1-mean-g0-idtype0] PASSED [ 78%]
tests/tensorflow/test_nn.py::test_sage_conv_bi[1-mean-g0-idtype1] PASSED [ 78%]
tests/tensorflow/test_nn.py::test_sage_conv_bi[1-mean-g1-idtype0] PASSED [ 78%]
tests/tensorflow/test_nn.py::test_sage_conv_bi[1-mean-g1-idtype1] PASSED [ 78%]
tests/tensorflow/test_nn.py::test_sage_conv_bi[1-mean-g2-idtype0] PASSED [ 78%]
tests/tensorflow/test_nn.py::test_sage_conv_bi[1-mean-g2-idtype1] PASSED [ 78%]
tests/tensorflow/test_nn.py::test_sage_conv_bi[1-pool-g0-idtype0] PASSED [ 79%]
tests/tensorflow/test_nn.py::test_sage_conv_bi[1-pool-g0-idtype1] PASSED [ 79%]
tests/tensorflow/test_nn.py::test_sage_conv_bi[1-pool-g1-idtype0] PASSED [ 79%]
tests/tensorflow/test_nn.py::test_sage_conv_bi[1-pool-g1-idtype1] PASSED [ 79%]
tests/tensorflow/test_nn.py::test_sage_conv_bi[1-pool-g2-idtype0] PASSED [ 79%]
tests/tensorflow/test_nn.py::test_sage_conv_bi[1-pool-g2-idtype1] PASSED [ 79%]
tests/tensorflow/test_nn.py::test_sage_conv_bi[1-gcn-g0-idtype0] PASSED  [ 79%]
tests/tensorflow/test_nn.py::test_sage_conv_bi[1-gcn-g0-idtype1] PASSED  [ 79%]
tests/tensorflow/test_nn.py::test_sage_conv_bi[1-gcn-g1-idtype0] PASSED  [ 79%]
tests/tensorflow/test_nn.py::test_sage_conv_bi[1-gcn-g1-idtype1] PASSED  [ 80%]
tests/tensorflow/test_nn.py::test_sage_conv_bi[1-gcn-g2-idtype0] PASSED  [ 80%]
tests/tensorflow/test_nn.py::test_sage_conv_bi[1-gcn-g2-idtype1] PASSED  [ 80%]
tests/tensorflow/test_nn.py::test_sage_conv_bi[2-mean-g0-idtype0] PASSED [ 80%]
tests/tensorflow/test_nn.py::test_sage_conv_bi[2-mean-g0-idtype1] PASSED [ 80%]
tests/tensorflow/test_nn.py::test_sage_conv_bi[2-mean-g1-idtype0] PASSED [ 80%]
tests/tensorflow/test_nn.py::test_sage_conv_bi[2-mean-g1-idtype1] PASSED [ 80%]
tests/tensorflow/test_nn.py::test_sage_conv_bi[2-mean-g2-idtype0] PASSED [ 80%]
tests/tensorflow/test_nn.py::test_sage_conv_bi[2-mean-g2-idtype1] PASSED [ 80%]
tests/tensorflow/test_nn.py::test_sage_conv_bi[2-pool-g0-idtype0] PASSED [ 80%]
tests/tensorflow/test_nn.py::test_sage_conv_bi[2-pool-g0-idtype1] PASSED [ 81%]
tests/tensorflow/test_nn.py::test_sage_conv_bi[2-pool-g1-idtype0] PASSED [ 81%]
tests/tensorflow/test_nn.py::test_sage_conv_bi[2-pool-g1-idtype1] PASSED [ 81%]
tests/tensorflow/test_nn.py::test_sage_conv_bi[2-pool-g2-idtype0] PASSED [ 81%]
tests/tensorflow/test_nn.py::test_sage_conv_bi[2-pool-g2-idtype1] PASSED [ 81%]
tests/tensorflow/test_nn.py::test_sage_conv_bi[2-gcn-g0-idtype0] PASSED  [ 81%]
tests/tensorflow/test_nn.py::test_sage_conv_bi[2-gcn-g0-idtype1] PASSED  [ 81%]
tests/tensorflow/test_nn.py::test_sage_conv_bi[2-gcn-g1-idtype0] PASSED  [ 81%]
tests/tensorflow/test_nn.py::test_sage_conv_bi[2-gcn-g1-idtype1] PASSED  [ 81%]
tests/tensorflow/test_nn.py::test_sage_conv_bi[2-gcn-g2-idtype0] PASSED  [ 82%]
tests/tensorflow/test_nn.py::test_sage_conv_bi[2-gcn-g2-idtype1] PASSED  [ 82%]
tests/tensorflow/test_nn.py::test_sage_conv_bi_empty[1-mean-idtype0] PASSED [ 82%]
tests/tensorflow/test_nn.py::test_sage_conv_bi_empty[1-mean-idtype1] PASSED [ 82%]
tests/tensorflow/test_nn.py::test_sage_conv_bi_empty[1-pool-idtype0] PASSED [ 82%]
tests/tensorflow/test_nn.py::test_sage_conv_bi_empty[1-pool-idtype1] PASSED [ 82%]
tests/tensorflow/test_nn.py::test_sage_conv_bi_empty[1-gcn-idtype0] PASSED [ 82%]
tests/tensorflow/test_nn.py::test_sage_conv_bi_empty[1-gcn-idtype1] PASSED [ 82%]
tests/tensorflow/test_nn.py::test_sage_conv_bi_empty[2-mean-idtype0] PASSED [ 82%]
tests/tensorflow/test_nn.py::test_sage_conv_bi_empty[2-mean-idtype1] PASSED [ 83%]
tests/tensorflow/test_nn.py::test_sage_conv_bi_empty[2-pool-idtype0] PASSED [ 83%]
tests/tensorflow/test_nn.py::test_sage_conv_bi_empty[2-pool-idtype1] PASSED [ 83%]
tests/tensorflow/test_nn.py::test_sage_conv_bi_empty[2-gcn-idtype0] PASSED [ 83%]
tests/tensorflow/test_nn.py::test_sage_conv_bi_empty[2-gcn-idtype1] PASSED [ 83%]
tests/tensorflow/test_nn.py::test_sgc_conv[1-g0-idtype0] PASSED          [ 83%]
tests/tensorflow/test_nn.py::test_sgc_conv[1-g0-idtype1] PASSED          [ 83%]
tests/tensorflow/test_nn.py::test_sgc_conv[1-g1-idtype0] PASSED          [ 83%]
tests/tensorflow/test_nn.py::test_sgc_conv[1-g1-idtype1] PASSED          [ 83%]
tests/tensorflow/test_nn.py::test_sgc_conv[1-g2-idtype0] PASSED          [ 84%]
tests/tensorflow/test_nn.py::test_sgc_conv[1-g2-idtype1] PASSED          [ 84%]
tests/tensorflow/test_nn.py::test_sgc_conv[1-g3-idtype0] PASSED          [ 84%]
tests/tensorflow/test_nn.py::test_sgc_conv[1-g3-idtype1] PASSED          [ 84%]
tests/tensorflow/test_nn.py::test_sgc_conv[1-g4-idtype0] PASSED          [ 84%]
tests/tensorflow/test_nn.py::test_sgc_conv[1-g4-idtype1] PASSED          [ 84%]
tests/tensorflow/test_nn.py::test_sgc_conv[1-g5-idtype0] PASSED          [ 84%]
tests/tensorflow/test_nn.py::test_sgc_conv[1-g5-idtype1] PASSED          [ 84%]
tests/tensorflow/test_nn.py::test_sgc_conv[2-g0-idtype0] PASSED          [ 84%]
tests/tensorflow/test_nn.py::test_sgc_conv[2-g0-idtype1] PASSED          [ 85%]
tests/tensorflow/test_nn.py::test_sgc_conv[2-g1-idtype0] PASSED          [ 85%]
tests/tensorflow/test_nn.py::test_sgc_conv[2-g1-idtype1] PASSED          [ 85%]
tests/tensorflow/test_nn.py::test_sgc_conv[2-g2-idtype0] PASSED          [ 85%]
tests/tensorflow/test_nn.py::test_sgc_conv[2-g2-idtype1] PASSED          [ 85%]
tests/tensorflow/test_nn.py::test_sgc_conv[2-g3-idtype0] PASSED          [ 85%]
tests/tensorflow/test_nn.py::test_sgc_conv[2-g3-idtype1] PASSED          [ 85%]
tests/tensorflow/test_nn.py::test_sgc_conv[2-g4-idtype0] PASSED          [ 85%]
tests/tensorflow/test_nn.py::test_sgc_conv[2-g4-idtype1] PASSED          [ 85%]
tests/tensorflow/test_nn.py::test_sgc_conv[2-g5-idtype0] PASSED          [ 85%]
tests/tensorflow/test_nn.py::test_sgc_conv[2-g5-idtype1] PASSED          [ 86%]
tests/tensorflow/test_nn.py::test_appnp_conv[g0-idtype0] PASSED          [ 86%]
tests/tensorflow/test_nn.py::test_appnp_conv[g0-idtype1] PASSED          [ 86%]
tests/tensorflow/test_nn.py::test_appnp_conv[g1-idtype0] PASSED          [ 86%]
tests/tensorflow/test_nn.py::test_appnp_conv[g1-idtype1] PASSED          [ 86%]
tests/tensorflow/test_nn.py::test_appnp_conv[g2-idtype0] PASSED          [ 86%]
tests/tensorflow/test_nn.py::test_appnp_conv[g2-idtype1] PASSED          [ 86%]
tests/tensorflow/test_nn.py::test_appnp_conv[g3-idtype0] PASSED          [ 86%]
tests/tensorflow/test_nn.py::test_appnp_conv[g3-idtype1] PASSED          [ 86%]
tests/tensorflow/test_nn.py::test_appnp_conv[g4-idtype0] PASSED          [ 87%]
tests/tensorflow/test_nn.py::test_appnp_conv[g4-idtype1] PASSED          [ 87%]
tests/tensorflow/test_nn.py::test_appnp_conv[g5-idtype0] PASSED          [ 87%]
tests/tensorflow/test_nn.py::test_appnp_conv[g5-idtype1] PASSED          [ 87%]
tests/tensorflow/test_nn.py::test_gin_conv[mean-g0-idtype0] PASSED       [ 87%]
tests/tensorflow/test_nn.py::test_gin_conv[mean-g0-idtype1] PASSED       [ 87%]
tests/tensorflow/test_nn.py::test_gin_conv[mean-g1-idtype0] PASSED       [ 87%]
tests/tensorflow/test_nn.py::test_gin_conv[mean-g1-idtype1] PASSED       [ 87%]
tests/tensorflow/test_nn.py::test_gin_conv[mean-g2-idtype0] PASSED       [ 87%]
tests/tensorflow/test_nn.py::test_gin_conv[mean-g2-idtype1] PASSED       [ 88%]
tests/tensorflow/test_nn.py::test_gin_conv[mean-g3-idtype0] PASSED       [ 88%]
tests/tensorflow/test_nn.py::test_gin_conv[mean-g3-idtype1] PASSED       [ 88%]
tests/tensorflow/test_nn.py::test_gin_conv[mean-g4-idtype0] PASSED       [ 88%]
tests/tensorflow/test_nn.py::test_gin_conv[mean-g4-idtype1] PASSED       [ 88%]
tests/tensorflow/test_nn.py::test_gin_conv[mean-g5-idtype0] PASSED       [ 88%]
tests/tensorflow/test_nn.py::test_gin_conv[mean-g5-idtype1] PASSED       [ 88%]
tests/tensorflow/test_nn.py::test_gin_conv[mean-g6-idtype0] PASSED       [ 88%]
tests/tensorflow/test_nn.py::test_gin_conv[mean-g6-idtype1] PASSED       [ 88%]
tests/tensorflow/test_nn.py::test_gin_conv[mean-g7-idtype0] PASSED       [ 89%]
tests/tensorflow/test_nn.py::test_gin_conv[mean-g7-idtype1] PASSED       [ 89%]
tests/tensorflow/test_nn.py::test_gin_conv[max-g0-idtype0] PASSED        [ 89%]
tests/tensorflow/test_nn.py::test_gin_conv[max-g0-idtype1] PASSED        [ 89%]
tests/tensorflow/test_nn.py::test_gin_conv[max-g1-idtype0] PASSED        [ 89%]
tests/tensorflow/test_nn.py::test_gin_conv[max-g1-idtype1] PASSED        [ 89%]
tests/tensorflow/test_nn.py::test_gin_conv[max-g2-idtype0] PASSED        [ 89%]
tests/tensorflow/test_nn.py::test_gin_conv[max-g2-idtype1] PASSED        [ 89%]
tests/tensorflow/test_nn.py::test_gin_conv[max-g3-idtype0] PASSED        [ 89%]
tests/tensorflow/test_nn.py::test_gin_conv[max-g3-idtype1] PASSED        [ 90%]
tests/tensorflow/test_nn.py::test_gin_conv[max-g4-idtype0] PASSED        [ 90%]
tests/tensorflow/test_nn.py::test_gin_conv[max-g4-idtype1] PASSED        [ 90%]
tests/tensorflow/test_nn.py::test_gin_conv[max-g5-idtype0] PASSED        [ 90%]
tests/tensorflow/test_nn.py::test_gin_conv[max-g5-idtype1] PASSED        [ 90%]
tests/tensorflow/test_nn.py::test_gin_conv[max-g6-idtype0] PASSED        [ 90%]
tests/tensorflow/test_nn.py::test_gin_conv[max-g6-idtype1] PASSED        [ 90%]
tests/tensorflow/test_nn.py::test_gin_conv[max-g7-idtype0] PASSED        [ 90%]
tests/tensorflow/test_nn.py::test_gin_conv[max-g7-idtype1] PASSED        [ 90%]
tests/tensorflow/test_nn.py::test_gin_conv[sum-g0-idtype0] PASSED        [ 90%]
tests/tensorflow/test_nn.py::test_gin_conv[sum-g0-idtype1] PASSED        [ 91%]
tests/tensorflow/test_nn.py::test_gin_conv[sum-g1-idtype0] PASSED        [ 91%]
tests/tensorflow/test_nn.py::test_gin_conv[sum-g1-idtype1] PASSED        [ 91%]
tests/tensorflow/test_nn.py::test_gin_conv[sum-g2-idtype0] PASSED        [ 91%]
tests/tensorflow/test_nn.py::test_gin_conv[sum-g2-idtype1] PASSED        [ 91%]
tests/tensorflow/test_nn.py::test_gin_conv[sum-g3-idtype0] PASSED        [ 91%]
tests/tensorflow/test_nn.py::test_gin_conv[sum-g3-idtype1] PASSED        [ 91%]
tests/tensorflow/test_nn.py::test_gin_conv[sum-g4-idtype0] PASSED        [ 91%]
tests/tensorflow/test_nn.py::test_gin_conv[sum-g4-idtype1] PASSED        [ 91%]
tests/tensorflow/test_nn.py::test_gin_conv[sum-g5-idtype0] PASSED        [ 92%]
tests/tensorflow/test_nn.py::test_gin_conv[sum-g5-idtype1] PASSED        [ 92%]
tests/tensorflow/test_nn.py::test_gin_conv[sum-g6-idtype0] PASSED        [ 92%]
tests/tensorflow/test_nn.py::test_gin_conv[sum-g6-idtype1] PASSED        [ 92%]
tests/tensorflow/test_nn.py::test_gin_conv[sum-g7-idtype0] PASSED        [ 92%]
tests/tensorflow/test_nn.py::test_gin_conv[sum-g7-idtype1] PASSED        [ 92%]
tests/tensorflow/test_nn.py::test_gin_conv_bi[mean-g0-idtype0] PASSED    [ 92%]
tests/tensorflow/test_nn.py::test_gin_conv_bi[mean-g0-idtype1] PASSED    [ 92%]
tests/tensorflow/test_nn.py::test_gin_conv_bi[mean-g1-idtype0] PASSED    [ 92%]
tests/tensorflow/test_nn.py::test_gin_conv_bi[mean-g1-idtype1] PASSED    [ 93%]
tests/tensorflow/test_nn.py::test_gin_conv_bi[mean-g2-idtype0] PASSED    [ 93%]
tests/tensorflow/test_nn.py::test_gin_conv_bi[mean-g2-idtype1] PASSED    [ 93%]
tests/tensorflow/test_nn.py::test_gin_conv_bi[max-g0-idtype0] PASSED     [ 93%]
tests/tensorflow/test_nn.py::test_gin_conv_bi[max-g0-idtype1] PASSED     [ 93%]
tests/tensorflow/test_nn.py::test_gin_conv_bi[max-g1-idtype0] PASSED     [ 93%]
tests/tensorflow/test_nn.py::test_gin_conv_bi[max-g1-idtype1] PASSED     [ 93%]
tests/tensorflow/test_nn.py::test_gin_conv_bi[max-g2-idtype0] PASSED     [ 93%]
tests/tensorflow/test_nn.py::test_gin_conv_bi[max-g2-idtype1] PASSED     [ 93%]
tests/tensorflow/test_nn.py::test_gin_conv_bi[sum-g0-idtype0] PASSED     [ 94%]
tests/tensorflow/test_nn.py::test_gin_conv_bi[sum-g0-idtype1] PASSED     [ 94%]
tests/tensorflow/test_nn.py::test_gin_conv_bi[sum-g1-idtype0] PASSED     [ 94%]
tests/tensorflow/test_nn.py::test_gin_conv_bi[sum-g1-idtype1] PASSED     [ 94%]
tests/tensorflow/test_nn.py::test_gin_conv_bi[sum-g2-idtype0] PASSED     [ 94%]
tests/tensorflow/test_nn.py::test_gin_conv_bi[sum-g2-idtype1] PASSED     [ 94%]
tests/tensorflow/test_nn.py::test_edge_conv[1-g0-idtype0] PASSED         [ 94%]
tests/tensorflow/test_nn.py::test_edge_conv[1-g0-idtype1] PASSED         [ 94%]
tests/tensorflow/test_nn.py::test_edge_conv[1-g1-idtype0] PASSED         [ 94%]
tests/tensorflow/test_nn.py::test_edge_conv[1-g1-idtype1] PASSED         [ 95%]
tests/tensorflow/test_nn.py::test_edge_conv[1-g2-idtype0] PASSED         [ 95%]
tests/tensorflow/test_nn.py::test_edge_conv[1-g2-idtype1] PASSED         [ 95%]
tests/tensorflow/test_nn.py::test_edge_conv[1-g3-idtype0] PASSED         [ 95%]
tests/tensorflow/test_nn.py::test_edge_conv[1-g3-idtype1] PASSED         [ 95%]
tests/tensorflow/test_nn.py::test_edge_conv[1-g4-idtype0] PASSED         [ 95%]
tests/tensorflow/test_nn.py::test_edge_conv[1-g4-idtype1] PASSED         [ 95%]
tests/tensorflow/test_nn.py::test_edge_conv[1-g5-idtype0] PASSED         [ 95%]
tests/tensorflow/test_nn.py::test_edge_conv[1-g5-idtype1] PASSED         [ 95%]
tests/tensorflow/test_nn.py::test_edge_conv[1-g6-idtype0] PASSED         [ 95%]
tests/tensorflow/test_nn.py::test_edge_conv[1-g6-idtype1] PASSED         [ 96%]
tests/tensorflow/test_nn.py::test_edge_conv[2-g0-idtype0] PASSED         [ 96%]
tests/tensorflow/test_nn.py::test_edge_conv[2-g0-idtype1] PASSED         [ 96%]
tests/tensorflow/test_nn.py::test_edge_conv[2-g1-idtype0] PASSED         [ 96%]
tests/tensorflow/test_nn.py::test_edge_conv[2-g1-idtype1] PASSED         [ 96%]
tests/tensorflow/test_nn.py::test_edge_conv[2-g2-idtype0] PASSED         [ 96%]
tests/tensorflow/test_nn.py::test_edge_conv[2-g2-idtype1] PASSED         [ 96%]
tests/tensorflow/test_nn.py::test_edge_conv[2-g3-idtype0] PASSED         [ 96%]
tests/tensorflow/test_nn.py::test_edge_conv[2-g3-idtype1] PASSED         [ 96%]
tests/tensorflow/test_nn.py::test_edge_conv[2-g4-idtype0] PASSED         [ 97%]
tests/tensorflow/test_nn.py::test_edge_conv[2-g4-idtype1] PASSED         [ 97%]
tests/tensorflow/test_nn.py::test_edge_conv[2-g5-idtype0] PASSED         [ 97%]
tests/tensorflow/test_nn.py::test_edge_conv[2-g5-idtype1] PASSED         [ 97%]
tests/tensorflow/test_nn.py::test_edge_conv[2-g6-idtype0] PASSED         [ 97%]
tests/tensorflow/test_nn.py::test_edge_conv[2-g6-idtype1] PASSED         [ 97%]
tests/tensorflow/test_nn.py::test_edge_conv_bi[1-g0-idtype0] PASSED      [ 97%]
tests/tensorflow/test_nn.py::test_edge_conv_bi[1-g0-idtype1] PASSED      [ 97%]
tests/tensorflow/test_nn.py::test_edge_conv_bi[1-g1-idtype0] PASSED      [ 97%]
tests/tensorflow/test_nn.py::test_edge_conv_bi[1-g1-idtype1] PASSED      [ 98%]
tests/tensorflow/test_nn.py::test_edge_conv_bi[2-g0-idtype0] PASSED      [ 98%]
tests/tensorflow/test_nn.py::test_edge_conv_bi[2-g0-idtype1] PASSED      [ 98%]
tests/tensorflow/test_nn.py::test_edge_conv_bi[2-g1-idtype0] PASSED      [ 98%]
tests/tensorflow/test_nn.py::test_edge_conv_bi[2-g1-idtype1] PASSED      [ 98%]
tests/tensorflow/test_nn.py::test_hetero_conv[sum-idtype0] PASSED        [ 98%]
tests/tensorflow/test_nn.py::test_hetero_conv[sum-idtype1] PASSED        [ 98%]
tests/tensorflow/test_nn.py::test_hetero_conv[max-idtype0] PASSED        [ 98%]
tests/tensorflow/test_nn.py::test_hetero_conv[max-idtype1] PASSED        [ 98%]
tests/tensorflow/test_nn.py::test_hetero_conv[min-idtype0] PASSED        [ 99%]
tests/tensorflow/test_nn.py::test_hetero_conv[min-idtype1] PASSED        [ 99%]
tests/tensorflow/test_nn.py::test_hetero_conv[mean-idtype0] PASSED       [ 99%]
tests/tensorflow/test_nn.py::test_hetero_conv[mean-idtype1] PASSED       [ 99%]
tests/tensorflow/test_nn.py::test_hetero_conv[stack-idtype0] PASSED      [ 99%]
tests/tensorflow/test_nn.py::test_hetero_conv[stack-idtype1] PASSED      [ 99%]
tests/tensorflow/test_nn.py::test_hetero_conv[myagg-idtype0] PASSED      [ 99%]
tests/tensorflow/test_nn.py::test_hetero_conv[myagg-idtype1] PASSED      [ 99%]
tests/tensorflow/test_nn.py::test_dense_cheb_conv[1] PASSED              [ 99%]
tests/tensorflow/test_nn.py::test_dense_cheb_conv[2] PASSED              [100%]

=============================== warnings summary ===============================
../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/tensorflow/python/data/ops/iterator_ops.py:546
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/tensorflow/python/data/ops/iterator_ops.py:546: DeprecationWarning: Using or importing the ABCs from 'collections' instead of from 'collections.abc' is deprecated, and in 3.8 it will stop working
    class IteratorBase(collections.Iterator, trackable.Trackable,

../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/tensorflow/python/data/ops/dataset_ops.py:106
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/tensorflow/python/data/ops/dataset_ops.py:106: DeprecationWarning: Using or importing the ABCs from 'collections' instead of from 'collections.abc' is deprecated, and in 3.8 it will stop working
    class DatasetV2(collections.Iterable, tracking_base.Trackable,

../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/tensorflow/python/autograph/utils/testing.py:21
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/tensorflow/python/autograph/utils/testing.py:21: DeprecationWarning: the imp module is deprecated in favour of importlib; see the module's documentation for alternative uses
    import imp

../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:10
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:10: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    _nlv = LooseVersion(_np_version)

../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:11
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:11: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    _np_version_under1p16 = _nlv < LooseVersion("1.16")

../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:12
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:12: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    _np_version_under1p17 = _nlv < LooseVersion("1.17")

../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:13
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:13: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    _np_version_under1p18 = _nlv < LooseVersion("1.18")

../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:14
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:14: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    _np_version_under1p19 = _nlv < LooseVersion("1.19")

../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:15
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/pandas/compat/numpy/__init__.py:15: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    _np_version_under1p20 = _nlv < LooseVersion("1.20")

../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/setuptools/_distutils/version.py:351
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/setuptools/_distutils/version.py:351: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    other = LooseVersion(other)

../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/pandas/compat/numpy/function.py:125
../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/pandas/compat/numpy/function.py:125
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/pandas/compat/numpy/function.py:125: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    if LooseVersion(_np_version) >= LooseVersion("1.17.0"):

../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/keras_preprocessing/image/utils.py:23
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/keras_preprocessing/image/utils.py:23: DeprecationWarning: NEAREST is deprecated and will be removed in Pillow 10 (2023-07-01). Use Resampling.NEAREST or Dither.NONE instead.
    'nearest': pil_image.NEAREST,

../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/keras_preprocessing/image/utils.py:24
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/keras_preprocessing/image/utils.py:24: DeprecationWarning: BILINEAR is deprecated and will be removed in Pillow 10 (2023-07-01). Use Resampling.BILINEAR instead.
    'bilinear': pil_image.BILINEAR,

../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/keras_preprocessing/image/utils.py:25
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/keras_preprocessing/image/utils.py:25: DeprecationWarning: BICUBIC is deprecated and will be removed in Pillow 10 (2023-07-01). Use Resampling.BICUBIC instead.
    'bicubic': pil_image.BICUBIC,

../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/keras_preprocessing/image/utils.py:28
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/keras_preprocessing/image/utils.py:28: DeprecationWarning: HAMMING is deprecated and will be removed in Pillow 10 (2023-07-01). Use Resampling.HAMMING instead.
    if hasattr(pil_image, 'HAMMING'):

../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/keras_preprocessing/image/utils.py:29
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/keras_preprocessing/image/utils.py:29: DeprecationWarning: HAMMING is deprecated and will be removed in Pillow 10 (2023-07-01). Use Resampling.HAMMING instead.
    _PIL_INTERPOLATION_METHODS['hamming'] = pil_image.HAMMING

../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/keras_preprocessing/image/utils.py:30
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/keras_preprocessing/image/utils.py:30: DeprecationWarning: BOX is deprecated and will be removed in Pillow 10 (2023-07-01). Use Resampling.BOX instead.
    if hasattr(pil_image, 'BOX'):

../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/keras_preprocessing/image/utils.py:31
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/keras_preprocessing/image/utils.py:31: DeprecationWarning: BOX is deprecated and will be removed in Pillow 10 (2023-07-01). Use Resampling.BOX instead.
    _PIL_INTERPOLATION_METHODS['box'] = pil_image.BOX

../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/keras_preprocessing/image/utils.py:33
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/keras_preprocessing/image/utils.py:33: DeprecationWarning: LANCZOS is deprecated and will be removed in Pillow 10 (2023-07-01). Use Resampling.LANCZOS instead.
    if hasattr(pil_image, 'LANCZOS'):

../../../../opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/keras_preprocessing/image/utils.py:34
  /opt/conda/envs/tensorflow-ci/lib/python3.7/site-packages/keras_preprocessing/image/utils.py:34: DeprecationWarning: LANCZOS is deprecated and will be removed in Pillow 10 (2023-07-01). Use Resampling.LANCZOS instead.
    _PIL_INTERPOLATION_METHODS['lanczos'] = pil_image.LANCZOS

python/dgl/backend/tensorflow/tensor.py:15
python/dgl/backend/tensorflow/tensor.py:15
  /root/jenkins/workspace/dgl_PR-4648@2/python/dgl/backend/tensorflow/tensor.py:15: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
    if LooseVersion(tf.__version__) < LooseVersion("2.3.0"):

tests/tensorflow/test_nn.py::test_graph_conv[1]
tests/tensorflow/test_nn.py::test_graph_conv[2]
tests/tensorflow/test_nn.py::test_simple_pool
tests/tensorflow/test_nn.py::test_glob_att_pool
tests/tensorflow/test_nn.py::test_rgcn[1]
tests/tensorflow/test_nn.py::test_rgcn[2]
tests/tensorflow/test_nn.py::test_rgcn[8]
tests/tensorflow/test_nn.py::test_dense_cheb_conv[1]
tests/tensorflow/test_nn.py::test_dense_cheb_conv[2]
  /root/jenkins/workspace/dgl_PR-4648@2/python/dgl/heterograph.py:72: DGLWarning: Recommend creating graphs by `dgl.graph(data)` instead of `dgl.DGLGraph(data)`.
    dgl_warning('Recommend creating graphs by `dgl.graph(data)`'

tests/tensorflow/test_nn.py::test_rgcn[1]
tests/tensorflow/test_nn.py::test_rgcn[2]
tests/tensorflow/test_nn.py::test_rgcn[8]
  /root/jenkins/workspace/dgl_PR-4648@2/python/dgl/heterograph.py:84: DGLWarning: Keyword arguments ['readonly'] are deprecated in v0.5, and can be safely removed in all cases.
    ' removed in all cases.'.format(list(deprecate_kwargs.keys())))

-- Docs: https://docs.pytest.org/en/stable/how-to/capture-warnings.html
- generated xml file: /root/jenkins/workspace/dgl_PR-4648@2/pytest_backend.xml -
============================ slowest 100 durations =============================
1.37s call     tests/tensorflow/test_nn.py::test_graph_conv[1]
1.10s call     tests/tensorflow/test_nn.py::test_hetero_conv[sum-idtype0]
1.09s call     tests/tensorflow/test_nn.py::test_hetero_conv[min-idtype0]
1.08s call     tests/tensorflow/test_nn.py::test_hetero_conv[min-idtype1]
1.08s call     tests/tensorflow/test_nn.py::test_hetero_conv[sum-idtype1]
1.06s call     tests/tensorflow/test_nn.py::test_graph_conv[2]
0.80s call     tests/tensorflow/test_nn.py::test_hetero_conv[mean-idtype0]
0.62s call     tests/tensorflow/test_nn.py::test_hetero_conv[stack-idtype1]
0.61s call     tests/tensorflow/test_nn.py::test_hetero_conv[max-idtype0]
0.61s call     tests/tensorflow/test_nn.py::test_hetero_conv[mean-idtype1]
0.60s call     tests/tensorflow/test_nn.py::test_hetero_conv[stack-idtype0]
0.60s call     tests/tensorflow/test_nn.py::test_hetero_conv[myagg-idtype0]
0.59s call     tests/tensorflow/test_nn.py::test_hetero_conv[max-idtype1]
0.59s call     tests/tensorflow/test_nn.py::test_hetero_conv[myagg-idtype1]
0.54s call     tests/tensorflow/test_nn.py::test_gat_conv[1-1-g0-idtype0]
0.54s call     tests/tensorflow/test_nn.py::test_gat_conv[1-1-g0-idtype1]
0.53s call     tests/tensorflow/test_nn.py::test_gat_conv[1-1-g1-idtype0]
0.53s call     tests/tensorflow/test_nn.py::test_gat_conv[1-1-g4-idtype0]
0.53s call     tests/tensorflow/test_nn.py::test_gat_conv[1-1-g1-idtype1]
0.53s call     tests/tensorflow/test_nn.py::test_gat_conv[1-1-g6-idtype0]
0.53s call     tests/tensorflow/test_nn.py::test_gat_conv[1-1-g3-idtype1]
0.53s call     tests/tensorflow/test_nn.py::test_gat_conv[1-1-g3-idtype0]
0.53s call     tests/tensorflow/test_nn.py::test_gat_conv[1-1-g4-idtype1]
0.52s call     tests/tensorflow/test_nn.py::test_gat_conv[1-1-g2-idtype0]
0.52s call     tests/tensorflow/test_nn.py::test_gat_conv[1-1-g5-idtype1]
0.52s call     tests/tensorflow/test_nn.py::test_gat_conv[1-1-g2-idtype1]
0.52s call     tests/tensorflow/test_nn.py::test_gat_conv[1-1-g6-idtype1]
0.52s call     tests/tensorflow/test_nn.py::test_gat_conv[1-1-g5-idtype0]
0.47s call     tests/tensorflow/test_nn.py::test_gat_conv[4-1-g2-idtype1]
0.46s call     tests/tensorflow/test_nn.py::test_gat_conv[4-1-g1-idtype0]
0.46s call     tests/tensorflow/test_nn.py::test_gat_conv[4-1-g1-idtype1]
0.46s call     tests/tensorflow/test_nn.py::test_gat_conv[4-1-g0-idtype0]
0.46s call     tests/tensorflow/test_nn.py::test_gat_conv[4-1-g3-idtype0]
0.46s call     tests/tensorflow/test_nn.py::test_gat_conv[4-1-g2-idtype0]
0.45s call     tests/tensorflow/test_nn.py::test_gat_conv[4-1-g0-idtype1]
0.45s call     tests/tensorflow/test_nn.py::test_gat_conv[4-1-g5-idtype1]
0.45s call     tests/tensorflow/test_nn.py::test_gat_conv[4-1-g3-idtype1]
0.44s call     tests/tensorflow/test_nn.py::test_gat_conv[4-1-g6-idtype1]
0.44s call     tests/tensorflow/test_nn.py::test_gat_conv[4-1-g6-idtype0]
0.43s call     tests/tensorflow/test_nn.py::test_gat_conv[1-2-g1-idtype1]
0.42s call     tests/tensorflow/test_nn.py::test_gat_conv[4-2-g5-idtype1]
0.42s call     tests/tensorflow/test_nn.py::test_gat_conv[1-2-g2-idtype0]
0.42s call     tests/tensorflow/test_nn.py::test_gat_conv[1-2-g0-idtype1]
0.42s call     tests/tensorflow/test_nn.py::test_gat_conv[1-2-g3-idtype0]
0.41s call     tests/tensorflow/test_nn.py::test_gat_conv[1-2-g1-idtype0]
0.41s call     tests/tensorflow/test_nn.py::test_gat_conv[1-2-g3-idtype1]
0.41s call     tests/tensorflow/test_nn.py::test_gat_conv[4-1-g5-idtype0]
0.41s call     tests/tensorflow/test_nn.py::test_gat_conv[1-2-g6-idtype1]
0.41s call     tests/tensorflow/test_nn.py::test_gat_conv[1-2-g2-idtype1]
0.41s call     tests/tensorflow/test_nn.py::test_gat_conv[1-2-g5-idtype0]
0.41s call     tests/tensorflow/test_nn.py::test_gat_conv[4-2-g5-idtype0]
0.41s call     tests/tensorflow/test_nn.py::test_gat_conv[4-2-g6-idtype1]
0.41s call     tests/tensorflow/test_nn.py::test_rgcn[8]
0.41s call     tests/tensorflow/test_nn.py::test_gat_conv[4-2-g6-idtype0]
0.41s call     tests/tensorflow/test_nn.py::test_gat_conv[4-2-g2-idtype0]
0.40s call     tests/tensorflow/test_nn.py::test_gat_conv[4-2-g3-idtype1]
0.40s call     tests/tensorflow/test_nn.py::test_gat_conv[1-2-g0-idtype0]
0.40s call     tests/tensorflow/test_nn.py::test_gat_conv[4-2-g0-idtype1]
0.40s call     tests/tensorflow/test_nn.py::test_gat_conv[4-2-g1-idtype1]
0.40s call     tests/tensorflow/test_nn.py::test_gat_conv[4-2-g0-idtype0]
0.39s call     tests/tensorflow/test_nn.py::test_gat_conv[1-2-g6-idtype0]
0.39s call     tests/tensorflow/test_nn.py::test_gat_conv[1-2-g4-idtype0]
0.39s call     tests/tensorflow/test_nn.py::test_gat_conv[4-2-g4-idtype0]
0.39s call     tests/tensorflow/test_nn.py::test_gat_conv[4-2-g1-idtype0]
0.38s call     tests/tensorflow/test_nn.py::test_rgcn[2]
0.38s call     tests/tensorflow/test_nn.py::test_gat_conv[4-2-g3-idtype0]
0.38s call     tests/tensorflow/test_nn.py::test_gat_conv[4-2-g4-idtype1]
0.37s call     tests/tensorflow/test_nn.py::test_gat_conv_bi[1-1-g1-idtype0]
0.37s call     tests/tensorflow/test_nn.py::test_gat_conv_bi[1-1-g1-idtype1]
0.36s call     tests/tensorflow/test_nn.py::test_gat_conv_bi[1-1-g0-idtype0]
0.36s call     tests/tensorflow/test_nn.py::test_gat_conv[4-2-g2-idtype1]
0.34s call     tests/tensorflow/test_nn.py::test_gat_conv[1-2-g5-idtype1]
0.34s call     tests/tensorflow/test_nn.py::test_gat_conv[4-1-g4-idtype0]
0.33s call     tests/tensorflow/test_nn.py::test_gat_conv_bi[1-1-g0-idtype1]
0.32s call     tests/tensorflow/test_nn.py::test_gat_conv[1-2-g4-idtype1]
0.32s call     tests/tensorflow/test_nn.py::test_rgcn[1]
0.30s call     tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-none-g3-idtype0]
0.30s call     tests/tensorflow/test_nn.py::test_gat_conv_bi[4-1-g0-idtype0]
0.30s call     tests/tensorflow/test_nn.py::test_gat_conv_bi[4-1-g1-idtype0]
0.30s call     tests/tensorflow/test_nn.py::test_gat_conv_bi[4-1-g0-idtype1]
0.29s call     tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-none-g1-idtype1]
0.29s call     tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-none-g3-idtype1]
0.28s call     tests/tensorflow/test_nn.py::test_gat_conv_bi[1-2-g0-idtype1]
0.27s call     tests/tensorflow/test_nn.py::test_gat_conv_bi[1-2-g1-idtype1]
0.27s call     tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-none-g1-idtype0]
0.27s call     tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-none-g0-idtype0]
0.27s call     tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-none-g5-idtype0]
0.27s call     tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-none-g4-idtype0]
0.27s call     tests/tensorflow/test_nn.py::test_gat_conv_bi[4-2-g1-idtype1]
0.27s call     tests/tensorflow/test_nn.py::test_gat_conv_bi[1-2-g1-idtype0]
0.26s call     tests/tensorflow/test_nn.py::test_gat_conv_bi[1-2-g0-idtype0]
0.26s call     tests/tensorflow/test_nn.py::test_sgc_conv[1-g3-idtype0]
0.26s call     tests/tensorflow/test_nn.py::test_sgc_conv[1-g0-idtype0]
0.26s call     tests/tensorflow/test_nn.py::test_gat_conv[4-1-g4-idtype1]
0.26s call     tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-none-g2-idtype1]
0.25s call     tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-both-g2-idtype1]
0.25s call     tests/tensorflow/test_nn.py::test_sgc_conv[1-g1-idtype1]
0.25s call     tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-none-g0-idtype1]
0.25s call     tests/tensorflow/test_nn.py::test_sgc_conv[1-g0-idtype1]
0.25s call     tests/tensorflow/test_nn.py::test_graph_conv2[1-True-True-none-g6-idtype0]
================= 920 passed, 35 warnings in 102.04s (0:01:42) =================
[Pipeline] }
[Pipeline] // timeout
[Pipeline] }
[Pipeline] // stage
Post stage
[Pipeline] cleanWs
[WS-CLEANUP] Deleting project workspace...
[WS-CLEANUP] Deferred wipeout is disabled by the job configuration...
[WS-CLEANUP] done
[Pipeline] }
$ docker stop --time=1 7ac09922888ca0e0d8459513540702ad50c9384fbb02f538e426a164f5d98757
$ docker rm -f 7ac09922888ca0e0d8459513540702ad50c9384fbb02f538e426a164f5d98757
[Pipeline] // withDockerContainer
[Pipeline] }
[Pipeline] // withEnv
[Pipeline] }
[Pipeline] // node
[Pipeline] }
[Pipeline] // stage
[Pipeline] }
[Pipeline] // parallel
[Pipeline] }
[Pipeline] // stage
[Pipeline] }
[Pipeline] // stage
[Pipeline] stage
[Pipeline] { (Declarative: Post Actions)
[Pipeline] script
[Pipeline] {
[Pipeline] node
Running on dglci-post-linux in /home/ubuntu/dgl_ci_post_linux/workspace/dgl_PR-4648
[Pipeline] {
[Pipeline] isUnix
[Pipeline] withEnv
[Pipeline] {
[Pipeline] sh
+ docker inspect -f . dgllib/dgl-ci-awscli:v220418
.
[Pipeline] }
[Pipeline] // withEnv
[Pipeline] withDockerContainer
dglci-post-linux does not seem to be running inside a container
$ docker run -t -d -u 0:0 --pull always --entrypoint= -w /home/ubuntu/dgl_ci_post_linux/workspace/dgl_PR-4648 -v /home/ubuntu/dgl_ci_post_linux/workspace/dgl_PR-4648:/home/ubuntu/dgl_ci_post_linux/workspace/dgl_PR-4648:rw,z -v /home/ubuntu/dgl_ci_post_linux/workspace/dgl_PR-4648@tmp:/home/ubuntu/dgl_ci_post_linux/workspace/dgl_PR-4648@tmp:rw,z -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** dgllib/dgl-ci-awscli:v220418 cat
$ docker top d465420ececd46eddeed628742417712df14465db5d2d5b2af636540c58e3707 -eo pid,comm
[Pipeline] {
[Pipeline] sh
+ rm -rf ci_tmp
[Pipeline] dir
Running in /home/ubuntu/dgl_ci_post_linux/workspace/dgl_PR-4648/ci_tmp
[Pipeline] {
[Pipeline] sh
+ curl -k -o cireport.log https://dgl-jenkins-eksvpc-2136217999.us-west-2.elb.amazonaws.com/job/dgl/job/PR-4648/2/consoleText
  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
                                 Dload  Upload   Total   Spent    Left  Speed